ETH Library

Fostering Innovation in EmergingTechnologiesEssays on the Patent System

Doctoral Thesis

Author(s):Zingg, Raphael

Publication date:2019-03

Permanent link:https://doi.org/10.3929/ethz-b-000329060

Rights / license:In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection.For more information, please consult the Terms of use.

D I S S . E T H N O 2 5 8 5 8

F O S T E R I N G I N N O VAT I O N I N E M E R G I N G T E C H N O L O G I E S

Essays on the Patent System

A thesis submitted to attain the degree ofDOCTOR OF SCIENCES of ETH Zurich

(Dr. sc. ETH Zurich)

presented by

R A P H A E L Z I N G G

Master of Law UZH

born on 23.05.1990

citizen of Switzerland

to be accepted on the recommendation of

Stefan Bechtold

Stuart Graham

David Schwartz

2019

Raphael Zingg: Fostering Innovation in Emerging Technologies, Essayson the Patent System, © March 5, 2019

F O R E W O R D

This text constitutes a cumulative doctoral thesis submitted to attainthe degree of Doctor of Sciences of ETH Zurich. It represents theoutput of four years of research at the ETH Zurich’s Center for Law& Economics, the Max Planck Institute for Innovation and Compe-tition in Munich, the University of California, Berkeley, School ofLaw, and the Waseda University’s Institute for Advanced Study inTokyo. During this time, studies in law and economics, and quantita-tive sciences, were conducted at the Study Center Gerzensee and atthe University of Michigan. Doctoral coursework at the ETH Zurichwas undertaken in the field of pharmaceutical sciences and biome-dical engineering, additional study in biology was completed at theUniversity of California, Berkeley. This research seeks to strengthenour understanding and answer to key questions about how legal in-stitutions impact innovation in emerging technologies. The disserta-tion integrates four stand-alone papers and a number of introductorychapters on the topic.

I thank David Schwartz and Stuart Graham for accepting and actingas mentor in their role as co-supervisors. I am grateful to Robert Mer-ges and Reto Hilty for welcoming me in their respective institutions.The research greatly benefited from comments and suggestions froma variety of scholars. In the internal seminar series at ETH Zurich, pre-cious first feedback was provided by past and present members of theCenter for Law & Economics, in particular Gerard Hertig, ErasmusElsner, Jens Frankenreiter, Gabriel Gertsch, Jérôme Hergueux, HeikoKarle, Vardges Levonyan and Martin Schonger. For the commentsand advice on the research papers, I warmly thank Barton Beebe, Ro-chelle C. Dreyfuss, Michael Frakes, Pedro W. Garcia, Fabian Gässler,John Golden, Dietmar Harhoff, Scott Hemphill, Lisa L. Ouelette, Geer-trui Van Overwalle, Gaétan de Rassenfosse, Bhaven Sampat, Chris-topher Sprigman, Hisashi Sugime, Reinhilde Veugelers and LudvigWier. Family and friends provided for invaluable support over theyears, and a very special word of thanks goes to my fiancée Nicole.

iii

The individual papers were presented at numerous conferences andseminars. For Chapter 3, I thank the participants of the 2018 Associa-tion of American Law Schools Annual Meeting, San Diego, the 2017

Annual Conference of the European Association of Law and Econo-mics, London, the 2016 Munich Summer Institute, Munich, the 2017

Berkeley Visiting Researcher Workshop, Berkeley, the 2017 BrownbagSeminar of the Max Planck Institute for Innovation and Competition,Munich, and the 2016 Roundtable on Empirical Methods in Intellec-tual Property, United States Patent & Trademark Office, WashingtonDC. The paper was granted the Göran Skogh Award 2017 by the Eu-ropean Association of Law and Economics and the Forum for Lawand Economics. It is under review with the European Journal of Lawand Economics. Chapter 4 was presented at the 2017 Munich Sum-mer Institute, Munich, and at the 2016 Workshop for Junior Resear-chers in IP Law, Leuven. A modified version of this Chapter has beensubmitted to the Journal of Law and the Biosciences. Chapter 6 waspresented at the 2018 PatCon 8, The Patent Conference, San Diego.Modified sections of this Chapter were published or are forthcomingwith Nano Today and the Journal of Nanoparticle Research. Additi-onally, further modifications have been made by invitation for publi-cation in WIREs Nanomedicine and Nanobiotechnology. For Chapter7, I thank the participants of the 2018 bioip Faculty Workshop, At-lanta, the 2017 Inaugural Wiet Life Science Law Scholars Conference,Chicago, and the 2017 Entrepreneurship 2.0: Legal, Regulatory, andEconomic Challenges to the Innovation Economy, Ft. Lauderdale, Flo-rida. A short version of the Chapter is under review with NatureBiotechnology.

But mainly, I thank my dissertation supervisor, Stefan Bechtold. Byproviding a carte blanche to his doctoral students, accompanied by astrong methodological training, he created a Center where one mightseek and try research at the frontier. His dedication, willingness tofind and create time, and constructive advice on the thesis is a roleexample for all of us embracing the academic path.

iv

C O N T E N T S

i prologue 1

1 about this text 3

1.1 The Subject: Innovation in Emerging Technologies . . . 8

1.2 The Method: Empirical Legal Studies . . . . . . . . . . 9

ii enforcement of patents in europe 13

2 introductory remarks 15

2.1 Shift Towards Harmonization . . . . . . . . . . . . . . . 15

2.2 Social Benefits of Harmonization . . . . . . . . . . . . . 17

2.3 Dimensions of Heterogeneity . . . . . . . . . . . . . . . 21

2.3.1 Dimensions on Uniformity Level . . . . . . . . . 22

2.3.2 Dimensions on Harmonization Level . . . . . . 25

2.4 Litigation as Evidence of Heterogeneity . . . . . . . . . 28

2.4.1 Litigation as Evidence . . . . . . . . . . . . . . . 28

2.4.2 Empirics on Litigation . . . . . . . . . . . . . . . 30

3 protection heterogeneity in a harmonized system 35

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3.2.1 Litigated Patents: the Tip of the Iceberg . . . . . 40

3.2.2 Microeconomic Model of Patent Disputes . . . . 41

3.3 Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.4 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.4.1 Germany . . . . . . . . . . . . . . . . . . . . . . . 50

3.4.2 France . . . . . . . . . . . . . . . . . . . . . . . . 51

3.4.3 United Kingdom . . . . . . . . . . . . . . . . . . 52

3.5 Data Coding . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.5.1 Decision Coding . . . . . . . . . . . . . . . . . . 53

3.5.2 Patent Quality . . . . . . . . . . . . . . . . . . . . 53

3.5.3 Technology Coding . . . . . . . . . . . . . . . . . 55

3.5.4 Industry Coding . . . . . . . . . . . . . . . . . . 56

3.6 Descriptive Results . . . . . . . . . . . . . . . . . . . . . 56

3.7 Empirical Results . . . . . . . . . . . . . . . . . . . . . . 59

3.7.1 Econometric Specification . . . . . . . . . . . . . 59

v

vi contents

3.7.2 Results . . . . . . . . . . . . . . . . . . . . . . . . 60

3.8 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.8.1 Differences Across Jurisdiction . . . . . . . . . . 62

3.8.2 Differences Across Patent Quality . . . . . . . . 65

3.8.3 Future of Litigation in Europe . . . . . . . . . . 66

3.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 70

3.10 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4 parallel litigation : the case of biotechnology 82

4.1 Biotechnology before the Courts . . . . . . . . . . . . . 86

4.1.1 Data Collection . . . . . . . . . . . . . . . . . . . 88

4.1.2 Methodology . . . . . . . . . . . . . . . . . . . . 89

4.1.3 Current Divide . . . . . . . . . . . . . . . . . . . 91

4.2 The Future . . . . . . . . . . . . . . . . . . . . . . . . . . 102

4.2.1 Transitional Period . . . . . . . . . . . . . . . . . 104

4.2.2 Bifurcation . . . . . . . . . . . . . . . . . . . . . . 105

4.2.3 Non-Participating States . . . . . . . . . . . . . . 106

4.2.4 Enforcement and Opposition . . . . . . . . . . . 107

4.2.5 Double Protection . . . . . . . . . . . . . . . . . 108

4.3 Theory of Parallel Litigation . . . . . . . . . . . . . . . . 109

4.3.1 Model Setting . . . . . . . . . . . . . . . . . . . . 110

4.3.2 Timing . . . . . . . . . . . . . . . . . . . . . . . . 112

4.3.3 Outlining the Solution . . . . . . . . . . . . . . . 114

4.3.4 Conclusions Subgame . . . . . . . . . . . . . . . 120

4.3.5 First Settlement Offer . . . . . . . . . . . . . . . 120

4.3.6 Predictions . . . . . . . . . . . . . . . . . . . . . 122

4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 123

4.5 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

iii funding of science in the united states 130

5 introductory remarks 132

5.1 Impact of Public Funding . . . . . . . . . . . . . . . . . 133

5.2 Rise of Academic Patenting . . . . . . . . . . . . . . . . 137

5.3 Technology Transfer . . . . . . . . . . . . . . . . . . . . 139

5.4 Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

6 patent thickets in nanotechnology 142

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 142

6.2 Nanotechnology and the Patent System . . . . . . . . . 143

6.3 Nanotechnology and Collaborations . . . . . . . . . . . 147

contents vii

6.4 Research Questions . . . . . . . . . . . . . . . . . . . . . 149

6.5 Related Literature . . . . . . . . . . . . . . . . . . . . . . 152

6.6 Data Collection . . . . . . . . . . . . . . . . . . . . . . . 154

6.7 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6.7.1 Increase in Patenting . . . . . . . . . . . . . . . . 155

6.7.2 Clustered Web of Patents . . . . . . . . . . . . . 156

6.7.3 Private-Public Collaboration in Nanotechnology 164

6.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 167

6.9 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

7 federal funding in cancer research 171

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 171

7.2 Research Questions . . . . . . . . . . . . . . . . . . . . . 174

7.3 Data Collection . . . . . . . . . . . . . . . . . . . . . . . 176

7.4 National Institutes of Health . . . . . . . . . . . . . . . 177

7.5 Background on Cancer Patenting . . . . . . . . . . . . . 178

7.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

7.6.1 Decrease in Overall Returns . . . . . . . . . . . . 182

7.6.2 Decrease in Grant Returns . . . . . . . . . . . . . 187

7.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 189

7.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 195

7.9 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

iv epilogue 201

8 concluding remarks 203

L I S T O F F I G U R E S

Figure 1.1 Flow Diagram of the Linear Model of Innovation 3

Figure 2.1 Dimensions of Heterogeneity . . . . . . . . . . 22

Figure 3.1 Litigation vs. Settlement Decision Tree . . . . . 42

Figure 3.2 Litigation vs. Settlement Indifference Curves . 43

Figure 3.3 Patent Infringement Bargaining Area . . . . . . 44

Figure 3.4 Patent Quality Function . . . . . . . . . . . . . 45

Figure 3.5 Technology Distribution in Germany, Franceand in the UK . . . . . . . . . . . . . . . . . . . 58

Figure 3.6 Average Patent Quality. . . . . . . . . . . . . . . 59

Figure 3.7 Distribution Quality. . . . . . . . . . . . . . . . 59

Figure 4.1 Biotechnology Litigation Geographical Coverage. 89

Figure 4.2 Timing of the Game . . . . . . . . . . . . . . . . 113

Figure 4.3 Timing of the Second-Offer Subgame . . . . . . 113

Figure 4.4 Market Profits vs. Litigation Costs for High Of-fer after Loss . . . . . . . . . . . . . . . . . . . . 117

Figure 4.5 Market Profits vs. Litigation Costs for High Of-fer after Win . . . . . . . . . . . . . . . . . . . . 119

Figure 4.6 Market Profits vs. Litigation Costs for High Offer122

Figure A4.1 Biotechnology Litigation Summary . . . . . . . 129

Figure 6.1 Europe. . . . . . . . . . . . . . . . . . . . . . . . 156

Figure 6.2 United States. . . . . . . . . . . . . . . . . . . . 156

Figure 6.3 Backward Citation Networks . . . . . . . . . . 158

Figure 6.4 Forward Citation Networks . . . . . . . . . . . 159

Figure 6.5 Clustering and Density of Prior Art . . . . . . 161

Figure 6.6 Nanomedicine Prior Art Networks . . . . . . . 163

Figure 6.7 Europe. . . . . . . . . . . . . . . . . . . . . . . . 165

Figure 6.8 United State. . . . . . . . . . . . . . . . . . . . . 165

Figure 6.9 Europe. . . . . . . . . . . . . . . . . . . . . . . . 166

Figure 6.10 United States. . . . . . . . . . . . . . . . . . . . 166

Figure A6.1 EU Public and Private Patent Applications. . . 170

Figure A6.2 EU Percentage Patent Applications, by Type. . 170

Figure 7.3 Growth in Cancer Patents Filings. . . . . . . . 180

viii

List of Figures ix

Figure 7.4 Growth in Cancer Patents Filings & NIH. . . . 182

Figure 7.5 Budget of NCI and NIH Patents. . . . . . . . . 183

Figure 7.6 Trapeze Distribution of Patent Quality. . . . . . 185

Figure 7.7 Grants with Linked Publication. . . . . . . . . 186

Figure 7.8 Probability of Grants to Patent . . . . . . . . . 187

Figure 7.9 Probability of Grants to Linked Publication. . . 188

Figure A7.1 Average Size of R01 Grants. . . . . . . . . . . . 200

L I S T O F TA B L E S

Table 3.1 Factor Loadings . . . . . . . . . . . . . . . . . . 55

Table 3.2 Win Rates by Country . . . . . . . . . . . . . . . . 57

Table 3.3 Summary Regression Results . . . . . . . . . . 61

Table A3.1 Summary Caseload by Country . . . . . . . . 74

Table A3.2 Overall Win Rate by Technology . . . . . . . . 75

Table A3.3 Infringement by Technology . . . . . . . . . . . 75

Table A3.4 Invalidity by Technology . . . . . . . . . . . . . 76

Table A3.5 Overall Win Rate by Industry . . . . . . . . . . 76

Table A3.6 Infringement by Industry . . . . . . . . . . . . 77

Table A3.7 Invalidity by Industry . . . . . . . . . . . . . . 77

Table A3.8 Full Regression Results . . . . . . . . . . . . . . 81

Table 4.1 Litigation outcome across Europe, by Country 92

Table A4.1 Duplicate Litigation Summary . . . . . . . . . 128

Table 6.1 Clustering Coefficients . . . . . . . . . . . . . . 160

Table 6.2 Density Coefficients . . . . . . . . . . . . . . . . 161

Table 6.3 EU Forward Citations 5-Y (Up to 2008) . . . . 167

Table 6.4 US Forward Citations 5-Y (Up to 2008) . . . . . 167

Table 6.5 EU Family Size (Up to 2012) . . . . . . . . . . . 167

Table 6.6 US Family Size (Up to 2012) . . . . . . . . . . . 167

Table 7.1 Forward Citations 5-Years Drugs & Chemistry 184

Table A7.1 Quality Proxies Drugs & Chemistry 1976-1990 199

Table A7.2 Quality Proxies Drugs & Chemistry 1991-2013 199

x

A C R O N Y M S

BE BelgiumDE GermanyDF Degrees of FreedomECJ European Court of JusticeEPC European Patent ConventionEU European UnionES SpainEWCA Court of Appeal (England and Wales)EWHC High Court (England and Wales)FDA Food and Drug AdministrationFR FranceIPC International Patent ClassificationIT ItalyMIT Massachusetts Institute of TechnologyNCI National Cancer InstituteNIH National Institutes of HealthNL NetherlandsOECD Organization for Economic Co-operation and DevelopmentR&D Research and DevelopmentSD Standard DeviationTRIPS Trade-Related Aspects of Intellectual Property RightsUK United KingdomUKHL United Kingdom House of LordsUPC Unified Patent CourtUS United States of AmericaUSD United States DollarUSPTO United States Patent and Trademark OfficeWIPO World Intellectual Property OrganizationWTO World Trade Organization

xi

A B S T R A C T

This dissertation seeks to investigate how legal institutions impactinnovation in emerging technologies both in Europe and in the Uni-ted States. It studies three key patterns that have repeatedly beenidentified as instrumental in fostering emerging technologies; publicfunding of science, technology linkage between public and private in-novators and corporate innovation by firms. The thesis follows a lawand economics approach, utilizing interdisciplinary methodologies. Itidentifies current leading legal issues in the field of patent law anddevelops measures of the policies aiming at encouraging the advan-cement of emerging technologies. Amongst a range of mechanisms,intellectual property rights, and patents especially, have occupied apivotal role in promoting – or stalling – innovation. The dissertationtakes the approach of studying the patent system, and the legal fra-mework that contributes to the rise of innovation.

The patent system is especially prone to conflict with disruptive techno-logies that challenge the status quo. The patentability of software orbiotechnology are typical examples in which new scientific advanceschallenged a system originally designed for mechanical inventions.In fact all three stages of innovation in emerging technologies – go-vernmental funding, technology linkage and corporate innovation –ultimately depend upon or interact with patent laws. Consideringthe impact of emerging technologies that are predicted to lead to newareas of scientific research, revolutionize industry processes and re-structure economic value chains, understanding the mechanisms thatgovern different technologies is crucial. Furthermore, studying indi-vidual technologies likely to be disproportionally affected by patentlaws and its effects provides for valuable insight for the overall designof the patent system.

Europe and the United States are investigated, two markets leadingtechnological innovation. For Europe, the innovation pattern of studyis corporate innovation, for the US it is governmental funding andtechnology transfer. The choice results both from the current policy

xii

abstract xiii

debate in the respective jurisdictions and from the state of knowledgein scholarly research. Europe is at the dawn of the largest reform of itspatent system in fifty years. By centralizing the working of the courts,it is hoped that companies may more efficiently enforce their patentsand allocate more resources to innovation. In the US, the innovationpolicy aims at maintaining the position of America as the world le-ader in science, technology and military. The large sums invested inscience resulted in a large number of publicly-funded patented inven-tions, that call for increasing technology transfer mechanisms.

The first part of the dissertation (Part II) studies the functioning ofthe courts in Europe. With the goal of fostering innovation, centraland declared policy has aimed for the harmonization of patent sys-tems. When a patent is filed in the European system, it undergoesuniform examination. However, once the patent is granted, it movesin a country-by-country enforcement regime. Here, differences in pa-tent litigation systems and application of patentability requirementsreflect the lack of an integrated jurisdiction. Intellectual property li-tigation affects corporate innovation by reducing incentive to invest-ment, pooling away resources from research and development to li-tigation monitoring and defense. Part II empirically investigates towhat extent the patent enforcements systems have been harmonizedin Europe. The papers make use of two approaches to capture theheterogeneity between the different patent systems.

Protection Heterogeneity in a Harmonized System, with Erasmus Elsner,(Chapter 3) introduces a game theoretical model for patent infringe-ment disputes, where both litigation and settlement are driven bypatent quality. Under the model, patent quality depends on bothbroadness and definiteness of the patent. The model predicts thattechnologies where the definiteness attribute can be estimated withhigh accuracy will have higher settlement rates. At trial stage, it israther the assessment of the patent quality by the judge which de-cides the outcome. In its empirical section, the paper evaluates overa thousand hand-collected and hand-coded patent infringement andcounterclaim decisions rendered by courts in the three largest patent-granting European countries – Germany, France and the United King-dom. The paper utilizes empirical methods to investigate whetherthe characteristics of the patents or the country of litigation predict

xiv abstract

the outcome of litigation. Examination of the patent characteristicsis conducted in line with our model, in that the patent quality, andunderlying technology and industry are tested. The findings showevidence of the failing harmonization of the patent systems in Eu-rope. Demonstrated was the lack of importance of the characteristicsof the litigated patent; rather, it was the forum to which the case wasbrought that was decisive.

The second approach the dissertation takes is to study the enforce-ment of identical patents in several countries. Systematic variancewould be a further sign of the lack of success in the application of har-monized patent laws. Biotechnology is being used as a larger exampleof such practice. This emerging technology exemplifies the problema-tic nature of litigation across the harmonized European system.

Parallel Litigation: The Case of Biotechnology (Chapter 4) introduces ananalysis of decisions over identical patents by different courts to shedlight on the extent of harmonization in Europe. The qualitative asses-sment of over a hundred biotechnology infringement and revocationdecisions presents the divide between the United Kingdom versusGermany and the Netherlands. The paper investigates whether theUnified Patent Court has the potential to achieve its ambition of coun-tering parallel and duplicate litigation. Temporary and persistent ob-stacles to a uniform patent litigation system are identified and analy-zed. The hope of resolving the issue of duplicate litigation throughunified patent enforcement can succeed only partially; there will re-main potential for dissident decisions over European patents. Centra-lizing the patent enforcement system is, however, necessary to elimi-nate parallel litigation. From a game theoretical standpoint, the paperdemonstrates that even in case of uniform application of harmonizedpatent laws, parallel litigation will exist. Litigation with settlementin two countries is formulated as a game where the patent holder,as uninformed party, proposes the license fee required to settle theclaims. The model shows that it is often optimal for patent holders toproceed to trial. In a system where several courts exist, parties mightlitigate individual national subparts, yet at any moment settle overthe whole group of lawsuits. The expected utility calculation fromthe standpoint of the patent holder is European, not national. Accor-

abstract xv

dingly, it is argued that the elimination of parallel litigation cannotbe reached through the mere harmonization of substantial law.

Part III comprises two papers focusing on the funding of science.After World War II, President Eisenhower laid the foundation for ascience policy for the US to lead technological progress. Governmen-tal funding has since then been a cornerstone of innovation policy,and the US still in 2018 outspends all other countries in terms of re-search and development expenditures. The exponential rise of suchpublicly-funded research has led to various issues. For example, pa-tents may impede rather than encourage innovation. In several indus-tries, such as telecommunications, biotechnology or semiconductors,the increase of patents are thought to have created a patent thicket: aweb of overlapping patents that stifle innovation.

Patent Thickets in Nanotechnology, with Marius Fischer, (Chapter 6) in-vestigates why, despite all predictions and promises, nanotechnologystill seems to be more of an emerging science than a fully-fledgedrevolution. More often than not, the patent system is blamed for this,supposedly tangling up nanotechnology in an impenetrable thicketof exclusionary rights instead of setting it free. Drawing on methodsof complex network analysis, the paper descriptively and empiricallystudies backward and forward citation networks to determine thepresence of such thickets. The findings suggest that there is actuallya patent thicket in the US but not in the Europe. It then shows howthe industry might mitigate the effects of the thickets by presentingfirst data on research output by public-private collaborations in nano-technology.

Amongst the range of innovation policy levers, ex ante or push me-chanisms, such as research grants or tax credits, subsidize researchinput, while pull mechanisms, like innovation prizes or patents re-ward successful research output. Inherently, input rewards are lesslikely to be verifiable than output mechanisms.

Federal Funding in Cancer Research (Chapter 7) exploits the feature thatthe observable funding input might result in observable output inthe form of patents. The paper studies the impact of the funding bythe National Institutes of Health in the field of cancer. Using patentsand publications subsequently cited in patents as proxies, it findsevidence for a productivity slowdown around 1995. Since then, the

xvi abstract

returns to public investment are decreasing both overall and by grant.Furthermore, the technological impact of federally funded patents ison the fall. The paper investigates a range of potential explanationfactors such as a possible saturation of the innovation space or flawsin the grant allocation system. It concludes by suggesting that theresults are in line with an incremental, rather than a high-risk high-rewards funding strategy by the federal agencies.

These studies combine economic theory, empirical methodologies, andtechnical knowledge in natural sciences, to investigate the legal fra-mework under which innovation takes place. By using law and eco-nomics methods that were traditionally developed in the United Sta-tes, and applying them to the European patent system – the mostprevalent example of a harmonized patent system worldwide – thedissertation sheds new light on key legal questions.

This dissertation contains numerous contributions to the study of in-novation and patent law. Chapter II establishes two novel and compre-hensive hand-collected and hand-coded datasets of European patentlitigation. The quantitative and qualitative study of these demonstratethe extent of the current heterogeneity of patent systems in Europe.The forum to which patent lawsuits are brought, rather than the me-rits of the patent, was determinant for the outcome of the ruling attrial. Game theoretical modeling is utilized to determine the empiri-cal strategy and to stress out the importance of the quality of litigatedpatents in litigation and settlement of claims. In light of these studiesfindings, how heterogeneity will play out in the new unified systemof the Unified Patent Court is discussed. The papers aim to contributeto the debate surrounding whether a homogeneous or heterogene-ous patent system is needed in Europe. In doing so, this dissertationcontributes to the general debate of the benefits and drawbacks ofhomogeneous and heterogeneous legal systems.

Chapter III quantifies innovation policies relating to the funding ofscience and technology transfer. The dissertation undertakes the quan-titative study of two large, public datasets in nanotechnology andcancer research. In nanotechnology, this data is used to present evi-dence of the existence of a patent thicket and of the potential forthe private-public partnering motion to solve it. The results supportthe view that the patent system is, partially at least, responsible for

abstract xvii

the lack of progress in nanotechnology. In cancer research, findingsof the decrease in returns from governmental spending demonstratethe decreasing impact of current grant strategies. This supports theclaim that funding agencies provide disincentives to funding transfor-mative research, rather focusing on conform science proposals. Thesestudies underline how policies not optimal in their design can stallrather than foster innovative activity.

The dissertation calls for attention on the mismatch between poli-cies and their realization. Public funding, transfer of knowledge andcorporate innovation in emerging technologies were hindered in bio-technology, cancer research and nanotechnology. The effectiveness ofinnovation incentive mechanisms fluctuates over time and field, andneeds readjustment to account for changes in the state of advance-ment of emerging technologies. Policymakers should therefore favora pluralistic view of innovation and invest more heavily in the asses-sment and monitoring of legal rules and institutions. It is the hopeof this dissertation that it will be one of many works determiningnot only the reading but also the practical impact of legal norms andpolicies.

R E S U M E

La présente thèse doctorale propose une analyse de l’impact des in-stitutions juridiques sur l’innovation dans les technologies émergen-tes en Europe et aux États-Unis. Elle étudie trois composantes clésidentifiées de manière récurrente comme étant essentielles à la pro-motion des technologies émergentes ; le financement public de lascience, les liens technologiques entre innovateurs publics et privéset l’innovation des entreprises. La thèse suit une approche de lawand economics en se basant sur des méthodologies interdisciplinaires.D’importantes problématiques régnant dans le domaine du droit desbrevets sont identifiées et des mesures de politiques publiques vi-sant à encourager l’avancement des technologies émergentes dévelop-pées. Parmi une série de divers mécanismes, les droits de propriétéintellectuelle, et les brevets en particulier, occupent un rôle centraldans la promotion ou le blocage de l’innovation. Cette thèse adoptel’approche de l’étude du système des brevets et du cadre juridiquecontribuant à l’essor de l’innovation.

Le système des brevets est particulièrement susceptible d’entrer enconflit avec des technologies émergentes remettant en question lestatu quo. Les questions de brevetabilité des logiciels ou des bio-technologies sont des exemples classiques de cas où de nouvelles per-cées scientifiques ont remis en question un système conçu à l’originepour des inventions mécaniques. De fait, les trois étapes de l’innovationdans les technologies émergentes – à savoir le financement gouver-nemental, les liens technologiques et l’innovation des entreprises –dépendent ou interagissent avec les lois sur les brevets. Au vu del’impact des technologies émergentes qui devraient mener à une nou-velle ère de recherche scientifique, révolutionner les processus indus-triels et restructurer les chaînes de valeur économiques, il est crucialde comprendre les mécanismes gouvernant le développement des dif-férentes technologies. En outre, l’étude des technologies individuellessusceptibles d’être affectées de manière disproportionnée par les lois

xviii

abstract xix

sur les brevets et leurs effets, fournit des informations précieuses pourla conception globale du système des brevets.

L’Europe et les Etats-Unis, deux marchés à la pointe en innovationtechnologique, font l’objet de la présente étude. Pour l’Europe, la com-posante d’innovation étudiée est l’innovation d’entreprise, tandis quepour les États-Unis, le financement gouvernemental et le transfert detechnologies sont analysés. Ce choix résulte autant du débat politiqueen cours dans les juridictions respectives que de l’état des connaissan-ces au niveau de la recherche académique. L’Europe est à l’aube de laplus grande réforme de son système de brevets depuis cinquante ans.En raison de la centralisation du fonctionnement des tribunaux, il estsupputé que les entreprises pourront faire respecter plus efficacementleurs brevets et allouer davantage de ressources à l’innovation. AuxÉtats-Unis, la politique d’innovation vise à maintenir la position del’Amérique en tant que leader mondial dans les domaines scientifi-ques, technologiques et militaires. Les sommes importantes investiesdans la science génèrent un grand nombre d’inventions financées pardes fonds publics, qui exigent des mécanismes de transfert de techno-logie de plus en plus nombreux.

La première partie de la thèse (Partie II) porte sur le fonctionnementdes tribunaux en Europe. Dans le but d’encourager l’innovation, laprincipale politique déclarée de ces dernières décennies vise l’harmon-isation des systèmes de brevets. Lorsqu’une demande de brevet estdéposée dans le système européen, elle est soumise à un examen uni-forme. Néanmoins, une fois le brevet délivré, ce dernier est soumis àun régime d’application spécifique à chaque pays. Ici, les différencesentre les systèmes de règlement des litiges et l’application des condi-tions en matière de brevetabilité reflètent l’absence d’une juridictionintégrée. Les litiges en matière de propriété intellectuelle affectentl’innovation des entreprises en réduisant l’incitation à l’investissement,détournant des ressources de la recherche et du développement auprofit de la surveillance et de la défense des litiges. La Partie II ex-amine empiriquement dans quelle mesure ces systèmes manquentd’harmonisation. Les articles de cette Partie se fondent sur deux ap-proches pour saisir l’hétérogénéité entre les différents systèmes debrevets.

xx abstract

Le Chapitre 3 (Protection Heterogeneity in a Harmonized System), co-écrit avec Erasmus Elsner, introduit un modèle théorique pour leslitiges en matière de contrefaçon de brevets, où le litige et le règle-ment amiable sont tous deux déterminés par la qualité des brevets.Selon le modèle, la qualité d’un brevet dépend à la fois de l’étendueet de la définition dudit brevet. Le modèle prévoit que les techno-logies pour lesquelles l’attribut de définition peut être estimé avecune grande précision ont un taux de règlements amiables plus élevé.Au stade du procès, l’évaluation de la qualité du brevet par le jugedétermine le résultat. Dans sa section empirique, le chapitre évalueplus d’un millier de décisions de contrefaçons et de demandes re-conventionnelles en nullité de brevets, recueillies et codées à la main,rendues par les tribunaux des trois plus importants pays européensen terme de brevets, soit l’Allemagne, la France et le Royaume-Uni.L’étude fait appel à des méthodes empiriques afin de déterminer siles caractéristiques des brevets, ou le pays du litige, permettent deprédire l’issue du litige. L’examen des caractéristiques du brevet esteffectué conformément à notre modèle, en ce sens que la qualité dubrevet, sa technologie sous-jacente et son industrie sont testées. Lesrésultats montrent l’échec de l’harmonisation des systèmes de brevetsen Europe. L’absence d’importance des caractéristiques du brevet enlitige est démontrée ; au contraire, l’instance à laquelle l’affaire estsoumise est déterminante. La deuxième approche adoptée dans lathèse consiste à étudier l’application de brevets identiques dans plu-sieurs pays. La variance systématique serait un autre signe de l’échecde l’application des lois harmonisées sur les brevets. La biotechno-logie est utilisée comme l’exemple plus large d’une telle pratique.Cette technologie émergente illustre la nature problématique des li-tiges dans l’ensemble du système européen harmonisé.

Le Chapitre 4 (Parallel Litigation: The Case of Biotechnology) introduitune analyse des décisions rendues par différents tribunaux sur desbrevets identiques afin de faire la lumière sur le degré d’harmonisationen Europe. L’évaluation qualitative de plus d’une centaine de décisi-ons judiciaires en matière de contrefaçon et de révocation de brevetsde biotechnologie présente le clivage entre le Royaume-Uni d’un côté,et l’Allemagne et les Pays-Bas de l’autre. L’article examine si la juri-diction unifiée du brevet a le potentiel de réaliser son ambition de

abstract xxi

contrer les litiges parallèles et redondants. Les obstacles temporaireset persistants à un système uniforme de règlement des litiges en ma-tière de brevets sont identifiés et analysés. L’espoir de résoudre leproblème des doubles litiges par une juridiction unifiée des brevetsne peut réussir que partiellement ; il subsistera en effet un risquede décisions dissidentes sur les brevets européens. La centralisationdu système d’application des brevets est toutefois nécessaire pour éli-miner les litiges parallèles. Du point de vue de la théorie des jeux,l’article démontre que même en cas d’harmonisation complète deslois sur les brevets, des litiges parallèles peuvent exister. Le litigeavec règlement amiable dans deux pays est formulé comme un jeuoù le titulaire du brevet, en tant que partie non informée, propose leslicences requises pour régler les revendications. Le modèle montrequ’il est souvent optimal pour les titulaires de brevets de procéder àun procès. Dans un système où il existe plusieurs tribunaux, les par-ties peuvent plaider des sous-parties nationales individuelles, tout enréglant à tout moment l’ensemble des actions en justice. Le calcul del’utilité attendu du point de vue du titulaire du brevet est européenet non national. Par conséquent, il est avancé que l’élimination des li-tiges parallèles ne peut être obtenue par la simple harmonisation dudroit substantiel en soi.

La Partie III comprend deux articles portant sur le financement de lascience aux États-Unis. Après la Seconde Guerre mondiale, le prési-dent américain Eisenhower jeta les bases d’une politique scientifiquepermettant aux États-Unis de se positionner en leader du progrèstechnologique. Depuis lors, le financement public est la pierre angu-laire de la politique d’innovation et, en 2018, les États-Unis dépen-saient encore plus que tout autre pays en matière de recherche etde développement. L’augmentation exponentielle de ces recherchesfinancées par l’État soulève diverses questions. Par exemple, les bre-vets peuvent entraver plutôt qu’encourager l’innovation. Dans plusi-eurs industries, comme les télécommunications, la biotechnologie oules semi-conducteurs, l’augmentation du nombre de brevets a poten-tiellement créé un maquis de brevets, soit un réseau de brevets sechevauchant et étouffant l’innovation.

xxii abstract

Le Chapitre 6 (Patent Thickets in Nanotechnology), co-écrit avec Ma-rius Fischer, examine pourquoi, malgré toutes les prédictions et pro-messes, la nanotechnologie semble encore être davantage une scienceémergente qu’une révolution à part entière. Fréquemment, le systèmedes brevets est blâmé, enchevêtrant supposément la nanotechnologiedans un maquis impénétrable de droits d’exclusion au lieu de la li-bérer. S’appuyant sur des méthodes d’analyse de réseaux complexes,l’article étudie de manière descriptive et empirique les réseaux decitations en amont et en aval pour déterminer la présence de telsmaquis. Les résultats suggèrent qu’il existe effectivement des maquisde brevets aux États-Unis, mais pas en Europe. L’article montre ens-uite comment l’industrie pourrait atténuer les effets des maquis enprésentant des données sur les résultats de la recherche issue de col-laborations publiques et privées en nanotechnologie.

Parmi la gamme des leviers de politique d’innovation, les mécanis-mes ex-ante ou de push, tels que les subventions ou les crédits d’impôtà la recherche, subventionnent les intrants de recherche, tandis que lesmécanismes de pull, tels que les prix de l’innovation ou des brevets,favorisent le succès des résultats de la recherche. En soi, les récom-penses d’intrants sont moins susceptibles d’être vérifiables que lesmécanismes de résultats.

Le Chapitre 7 (Federal Funding in Cancer Research) exploite le fait quel’intrant financier observable peut donner lieu à des résultats observa-bles sous forme de brevets. L’article étudie l’impact du financementdes National Institutes of Health dans le domaine de la recherche surle cancer. À l’aide de brevets et de publications citées dans les bre-vets en tant qu’indicateurs, l’article constate un ralentissement de laproductivité autour de l’année 1995. Depuis lors, le rendement del’investissement public diminue. De plus, l’impact technologique desbrevets financés par le gouvernement fédéral est à la baisse. L’articleexamine un certain nombre de facteurs explicatifs éventuels, commela possible saturation de l’espace de l’innovation ou les lacunes dusystème d’attribution des bourses. Il conclut en suggérant que les ré-sultats sont conformes à une stratégie de financement conservatricedes organismes fédéraux, plutôt qu’à une stratégie de financement àhaut risque et à rendement élevé.

abstract xxiii

Ces études combinent la théorie économique, les méthodologies em-piriques et les connaissances techniques en sciences naturelles pourétudier le cadre juridique dans lequel s’inscrit l’innovation. En utili-sant des méthodes juridiques et économiques traditionnellement dé-veloppées aux Etats-Unis et en les appliquant au système européendes brevets – constituant l’exemple le plus répandu d’un systèmeharmonisé des brevets dans le monde - la présente thèse doctoraleapporte un éclairage nouveau sur des questions juridiques clés.

Cette thèse contient de nombreuses contributions à l’étude de l’innova-tion et du droit des brevets. La Partie II établit deux nouveaux setsde données collectées et codées à la main sur les litiges de brevetseuropéens. L’étude quantitative et qualitative de ces derniers mon-tre l’ampleur de l’hétérogénéité actuelle des systèmes de brevets enEurope. L’instance à laquelle le litige de brevet est soumise est dé-terminante, plutôt que les caractéristiques du brevet en cause. Lamodélisation de la théorie des jeux est utilisée pour déterminer lastratégie empirique et pour souligner l’importance de la qualité desbrevets dans les litiges et le règlement amiable des revendications. Àla lumière des résultats de ces études, la thèse discute de la manièredont l’hétérogénéité se manifestera dans le nouveau système unifiéde la juridiction unifiée des brevets. Cette partie vise à contribuer audébat sur la question de savoir si un système de brevets homogèneou hétérogène est requis en Europe. Ce faisant, cette thèse contribueau débat général sur les avantages et les inconvénients de systèmesjuridiques homogènes et hétérogènes.

La Partie III quantifie les politiques d’innovation relatives au finan-cement de la science. La thèse porte sur l’étude quantitative de deuxgrands sets de données publics en nanotechnologie et en recherchesur le cancer. En nanotechnologie, ces données sont utilisées pourprésenter l’existence d’un maquis de brevets et la possibilité que lespartenariats publics et privés puissent le résoudre. Les résultats con-firment l’opinion selon laquelle le système des brevets est, du moinsen partie, responsable de l’absence de progrès dans le domaine desnanotechnologies. En matière de recherche sur le cancer, les résultatsde la diminution du rendement des dépenses gouvernementales dé-montrent l’impact décroissant des stratégies de subvention actuelles.La thèse appuie l’affirmation selon laquelle les organismes de finan-

xxiv abstract

cement dissuadent le financement de la recherche transformatrice, ense concentrant au contraire sur des propositions scientifiques conser-vatrices. Les articles développés dans le cadre de cette recherche dé-montrent que des politiques dont la conception n’est pas optimalepeuvent ralentir plutôt que favoriser l’activité innovante.

La présente thèse de doctorat attire l’attention sur l’inadéquation en-tre les politiques d’innovation et leur réalisation. Le financement pu-blic, le transfert de technologies et l’innovation des entreprises dansles technologies émergentes sont entravés dans les domaines de la bi-otechnologie, de la recherche sur le cancer et de la nanotechnologie.L’efficacité des mécanismes d’incitation à l’innovation fluctue dans letemps et selon le domaine et doit être réajustée pour tenir compte deschangements dans l’état d’avancement des technologies émergentes.Les décideurs politiques se doivent de favoriser une vision pluralistede l’innovation et d’investir davantage dans l’évaluation et le suivides normes et institutions juridiques. L’espoir et l’ambition de cettethèse est qu’elle présentera l’un des nombreux travaux qui détermine-ront non seulement la lecture mais aussi l’impact pratique des normeset politiques juridiques en matière de droit des brevets.

Part I

P R O L O G U E

1A B O U T T H I S T E X T

Technological innovation is considered a major driver in economicgrowth1. Scholars from different fields have tried to untangle theconnection between research, innovation and economic growth. Ac-cording to the ‘linear innovation model’, long predominant in main-stream economics literature, basic research leads to applied research,followed by product development, and the cycle ends with productionand diffusion (see Figure 1.1)2.

BasicResearch

AppliedResearch

TechnologicalDevelopment Production

Figure 1.1: Flow Diagram of the Linear Model of Innovation

In essence, the linear model of innovation argues that there is a dis-tinction between basic scientific research and applied industrial rese-arch, with the former being the source of technological innovation.The actors in the innovative pipeline are clearly divided: basic rese-arch is led by universities and public institutions while applied rese-arch and development is carried out by corporations. Basic researchshould accordingly be publicly funded, and the knowledge producedby such research be available to the public domain3.

1. The economic growth literature traces back to work by Robert M. Solow,“Technical Change and the Aggregate Production Function,” Review of Economicsand Statistics 39, no. 3 (1957): 312–320 and Trevor W. Swan, “Economic Growth andCapital Accumulation,” The Economic Record 32, no. 2 (1956): 334–361.

2. The linear model is attributed to Vannevar Bush, Science, the Endless Frontier;A Report to the President on a Program for Postwar Scientific Research (National ScienceFoundation, 1960) who outlined a rudiment of this model by establishing causallinks between basic science and socio-economic progress; for the history, see BenoîtGodin, “The Linear Model of Innovation: The Historical Construction of an Analyti-cal Framework,” Science, Technology, & Human Values 31, no. 6 (2006): 639–667.

3. Margherita Balconi, Stefano Brusoni, and Luigi Orsenigo, “In Defence of theLinear Model: An Essay,” Research Policy 39, no. 1 (2010): 1–13.

3

4 about this text

This dissertation follows an analytical framework based upon this li-near concept of innovation. This model is imperfect in capturing thediversity and complexity that may exist in innovation processes4. Ho-wever, it provides the policy maker with a strong framework for ana-lysis, that may be refined as wished. Certain points of criticism holdless weight in emerging technologies than in more mature ones. Theargument that technological improvement can be unrelated to basicresearch is not true for high-tech, where research fundamentals arerequisite building blocks5. It is possible that the traditional divisionof labor has become more interactive overall. Yet, in fields such as bi-otechnology, universities and small specialized firms are the driversof inventions later developed by large pharmaceutical firms6.

Other models complement our understanding of the functioning oftechnological progress. Geographers and evolutionary economists fa-vor ’systems of innovation’ models, in which institutional networksact as catalysts for innovation in social and structural conditions indistinct territories. Policies should aim at improving the linkage amon-gst actors and institutions7. For ‘knowledge spillover’ theories, on theother hand, the diffusion and assimilation of innovation are key. Inno-vation is spatial, with clusters where knowledge externalities reducethe costs of knowledge generation8.

4. For criticism, see William J. Price and Lawrence W. Bass, “Scientific Researchand the Innovative Process,” Science 164, no. 3881 (1969): 802–806; Stephen J. Kline,“Innovation Is Not a Linear Process,” Research Management 28, no. 4 (1985): 36–45.

5. Stephen J. Kline and Nathan Rosenberg, “An Overview of Innovation,” in ThePositive Sum Strategy: Harnessing Technology for Economic Growth, ed. Landau Ralphand Nathan Rosenberg (National Academy of Sciences, 1986), 275–306 arguing that‘[t]he notion that innovation is initiated by research is wrong most of the time’ andthat the source is often design rather than research.

6. Luigi Orsenigo, Fabio Pammolli, and Massimo Riccaboni, “TechnologicalChange and the Dynamics of Networks of Collaborative Relations: The Case of theBio-Pharmaceutical Industry,” Research Policy 30 (2001): 485–508.

7. The systems of innovation literature emerged in the 1980s with ChristopherFreeman, Technology, Policy, and Economic Performance: Lessons from Japan (Pinter, 1987)and Bengt-Ake Lundvall, National Systems of Innovation: Towards a Theory of Innovationand Interactive Learning (Pinter, 1992).

8. Knowledge spillover theories base upon Alfred Marshall, Principles of Economics(MacMillan, 1920), p. 225 stating that ‘when an industry has thus chosen a localityfor itself, it is likely to stay there long: so great are the advantages which peoplefollowing the same skilled trade get from near neighborhood to one another’.

about this text 5

Patents are the most prominent innovation policy mechanism9. Sim-ply put, patents grant the right to an inventor to exclude others fromthe commercial exploitation of his inventions for a limited periodof time, in return for the disclosure of the invention10. Governmen-tal intervention shall address market failures that would arise dueto the public good nature of ideas. A public good has two proper-ties: nonrivalness in consumption and nonexcludability11. For suchgoods, the social value from investing in research and development(R&D) may be higher than the private value for the investing party.Several factors, including sunk costs, risks, uncertainty or difficultyof appropriability induce underinvestment in knowledge below thesocially desired level12. A patent system should remedy to underpro-duction by providing additional incentives for research investments,providing inventors with supracompetitive profits. Without a patentsystem, a generic drug manufacturer would free-ride on the invest-ments made by innovative drug companies. Since it did not incur thecosts of drug discovery and development, the generic company cansell the identical pharmaceutical drug for lower prices. Under suchregime, the innovative company will not invest initially if it cannotrecoup its costs, leading to overall underinvestment in socially valua-ble projects – medical knowledge in this case13.

9. See, e.g., Suzanne Scotchmer, Innovation and Incentives (MIT Press, 2004).10. Rebecca Eisenberg, “Patents and the Progress of Science: Exclusive Rights and

Experimental Use,” University of Chicago Law Review 56 (1989): 1025–1028.11. Paul A. Samuelson, “The Pure Theory of Public Expenditure,” Review of Econo-

mics and Statistics 36, no. 4 (1954): 387–389 and Joseph E. Stiglitz, “Knowledge as aGlobal Public Good,” Global Public Goods 19 (1999): 308–326.

12. See Kenneth W. Dam, “The Economic Underpinnings of Patent Law,” The Jour-nal of Legal Studies 23, no. 1 (1994): 247–271; and Richard R. Nelson, “The SimpleEconomics of Basic Scientific Research,” Journal of Political Economy 67 (1959): 297–306 and Kenneth Arrow, “Economic Welfare and the Allocation of Resources for In-vention,” in The Rate and Direction of Inventive Activity: Economic and Social Factors(National Bureau of Economic Research, 1962), 609–626 for the potential for marketfailures in investments in R&D.

13. See Suzanne Scotchmer and Green Jerry, “Novelty and Disclosure in PatentLaw,” The RAND Journal of Economics 21, no. 1 (1990): 131–146 and Richard C. Le-vin, “A New Look at the Patent System,” The American Economic Review 76, no. 2

(1986): 199–202 for the free rider problem that lies at the foundation of the currentintellectual property laws.

6 about this text

An optimal patent policy can be described as resolving the trade-offbetween such additional incentives to innovate and the cost of dead-weight loss resulting from higher prices14. The question of whetherpatent law promotes or hinders innovation is subject of a heated de-bate. Theoretical economic frameworks15, statistical models of techno-logical innovation16, measurements of technological innovation in asingle state17 or comparison of innovation amongst states with dif-ferent level of patent protection18 have sometimes supported, andsometimes undermined the assumption that patents promote techno-logical innovation. In fact, the patent system is presently subject toan unprecedented level of criticism. Boldrin and Levine (2008) call

14. See William Nordhaus, Invention, Growth, and Welfare: A Theoretical Treatment ofTechnological Change (MIT Press, 1969) and Heidi L. Williams, “Intellectual PropertyRights and Innovation: Evidence from Health Care Markets,” Innovation Policy andthe Economy 16 (2016): p. 54.

15. William M. Landes and Richard A. Posner, The Economic Structure of IntellectualProperty Law (Belknap Press, 2003).

16. See Robert L. Basmann, Michael McAleer, and Daniel Slottje, “Patent Activityand Technical Change,” Journal of Econometrics 139 (2007): 355–375 measuring theeffect of patent grants on technological change; Michael McAleer and Daniel Slottje,“A New Measure of Innovation: The Patent Success Ratio,” Scientometrics 63, no. 3

(2005): 421–429 studying innovation through the ratio of successful patent applicati-ons to total patent applications; and Petra Moser, “How Do Patent Laws InfluenceInnovation? Evidence from Nineteenth-Century World’s Fairs,” American EconomicReview 95, no. 4 (2005): 1214–1236 finding that existence of national patent laws im-pacts the industry focus of inventors.

17. Julie E. Cohen and Mark A. Lemley, “Patent Scope and Innovation in the Soft-ware Industry,” California Law Review 89 (2001): 1–57 investigating the implications ofpatent law for innovation in the software industry; Josh Lerner, “Where Does StateStreet Lead? A First Look at Finance Patents, 1971-2000,” Journal of Finance 57, no. 2

(2002): 901–930 for patents for financial formulas and methods; and Stuart J. Grahamand Matthew Higgins, “The Impact of Patenting on New Product Introductions inthe Pharmaceutical Industry,” MPRA Paper, no. 4574 (2007) for pharmaceutical pa-tents.

18. Lee G. Branstetter, “Do Stronger Patents Induce More Local Innovation?,” Jour-nal of International Economic Law 7, no. 2 (2004): 359–370 comparing the US and Japan;Janice M. Mueller, “The Tiger Awakens: The Tumultuous Transformation of India’sPatent System and the Rise of Indian Pharmaceutical Innovation,” University of Pitts-burgh Law Review 68, no. 3 (2006): 491–641 for India; and Mark Schankerman, “HowValuable is Patent Protection? Estimates by Technology Field,” The RAND Journal ofEconomics 29, no. 1 (1998): 77–107 for France.

about this text 7

IP rights an ‘unnecessary evil’19, Jaffe and Lerner (2004) and Bessenand Meurer (2008) claim the patent system is ‘broken’ 20 respectivelya ‘failure’ and Stallman (2013) concludes that the latter should be‘abolished’21. Answering this question is complicated by the widelyvaried use of patents across technologies and industries. The pharma-ceutical industry is recognized to heavily rely on patents22. In otherindustries, firms protect profits through an array of mechanisms. Ma-nufacturing firms tend to emphasize secrecy and lead time, ratherthan patents23. Large technology companies have altered their policyto embrace open innovation and open source software24. In certainfields, the first mover advantage may sustain investment in inventionand innovation without the need for formalized patent protection25.These differences increase the difficulty in finding consensus aboutthe exact objective of the patent system and the best way to reachit.

19. Michele Boldrin and David K. Levine, Against Intellectual Monopoly (CambridgeUniversity Press, 2008), 11.

20. Adam B. Jaffe and Josh Lerner, Innovation and Its Discontents: How Our BrokenPatent System Is Endangering Innovation and Progress, and What to Do About It (Prin-ceton University Press, 2004) contending that the patent system has ‘become sandrather than lubricant in the wheels of American progress’.

21. Richard Stallman, “Patent Law is, at Best, Not Worth Keeping,” Loyola Univer-sity Chicago Law Journal 45, no. 2 (2013): 389; see Maureen K. Ohlhausen, “PatentRights in a Climate of Intellectual Property Rights Skepticism,” Harvard Journal ofLaw & Technology 30, no. 1 (2016): 1–49 detailing (and defending the system against)further criticism of the patent system.

22. For surveys, see, e.g., Edwin Mansfield, Mark Schwartz, and Samuel Wagner,“Imitation Costs and Patents: An Empirical Study,” Economic Journal 91, no. 364

(1981): 907–918 arguing that 90% of the pharmaceutical innovations would not havebeen introduced without patents.

23. Wesley M. Cohen, Richard R. Nelson, and John P. Walsh, “Protecting TheirIntellectual Assets: Appropriability Conditions and Why U.S. Manufacturing FirmsPatent (or Not),” NBER Working Paper, no. 7552 (2000).

24. See Bronwyn H. Hall, “Open Innovation & Intellectual Property Rights: TheTwo-Edged Sword,” Japan Spotlight Jan/Feb (2010): 18–19 detailing the partnershipbetween IBM, Nokia and Sony to release a number of environmentally related pa-tents; while Microsoft has previously fought the open source program Linux, callingit a ‘cancer’, it joined the Linux Foundation in 2016.

25. Michele Boldrin and David K. Levine, “The Case Against Patents,” The Journalof Economic Perspectives 27, no. 1 (2013): 3–22 and Frederic M. Scherer, “First MoverAdvantages and Optimal Patent Protection,” The Journal of Technology Transfer 40, no.4 (2015): 559–580.

8 about this text

1.1 the subject : innovation in emerging technologies

Emerging technologies provide a particularly interesting frameworkto examine the patent system. When tracing back the historical de-velopment of such technologies, a set of patterns can be identified26.Firstly, government funding is instrumental in driving innovation inemerging technologies. Second, competitive market forces and firmsengaging in follow-on innovation are crucial. Thirdly, cooperationthrough knowledge share between research entities and the privatesector feed innovation. Those characteristics, from the standpoint ofa lawyer, are all impacted either directly or indirectly by patent law.Governmental funding is only one of several incentive schemes – nextto patents, prizes or tax credits – that might be the optimal responseto foster innovation in a particular field27. The capacity of firms toengage in corporate innovation depends on a variety of factors, pa-tent litigation being one of them28. Finally, the sharing of knowledge,while facilitated by patents due to disclosure requirements, can behindered when patents block developments in an industry29. This dis-sertation focuses on these three distinct patterns by targeting specificaspects of the public funding (by studying the impact of grants), the

26. A study by World Intellectual Property Report, Breakthrough Innovation and Eco-nomic Growth, (World Intellectual Property Organization, 2015), p. 11 determinedthese characteristics based on six case studies of breakthrough innovation (airplanes,antibiotics, semiconductors, 3D printing, nanotechnology and robotics).

27. See Daniel J. Hemel and Lisa L. Ouellette, “Beyond the Patents-Prizes Debate,”Texas Law Review 92, no. 2 (2006): 303–382 for a taxonomy of innovation policies.

28. Litigation prospects pool away resources from investment to defensive activi-ties, see Roger Smeets, “Does Patent Litigation Reduce Corporate R&D? An Analysisof US Public Firms,” Working Paper Rutger University, 2014, for a study of the ne-gative impact of patent litigation on corporate R&D intensity in small firms thatare involved in extensive patent lawsuits (in the range of 2.6-4.7%-points); see JamesBessen and Michael J. Meurer, “The Private Costs of Patent Litigation,” Boston Univer-sity School of Law Working Paper, nos. 07-08 (2007) estimating that patent infringementlawsuits cost about half a percentage of the stock market value to alleged infringers,rising to $16.1 billion in 1999 or 19% of these firms R&D spendings.

29. See Rebecca Eisenberg, “Patents: Help or Hindrance to Technology Transfer?,”in Biotechnology: Science, Engineering, and Ethical Challenges for the Twenty-First Century,ed. Frederick B. Rudolph and Larry McIntire (Joseph Henry Press, 1996), 161–174

and Bronwyn H. Hall, “Does Patent Protection Help or Hinder Technology Trans-fer?,” chap. 2 in Intellectual Property for Economic Development, ed. Sanghoon Ahn,Bronwyn H. Hall, and Keun Lee, Chapters (Edward Elgar Publishing, 2014), 11–32.

1.2 the method : empirical legal studies 9

transfer of knowledge (by investigating the rise of patents and howcompanies may remedy to the lack of technology transfer), and cor-porate innovation (by assessing the patent system harmonization).

Since emerging technologies commonly challenge the established setof rules, they appear to be especially prone to conflict with patentlaw. Illustratively, applying laws originally written for mechanical in-ventions brought dissidences to the patent offices and courts whenquestions of patentability of software or biotechnology arose. Whenselecting emerging technologies30, the thesis thus targets technologiespotentially subject to such conflict. Nanotechnology was so selected,as it exemplifies the stagnating effect lack of technology transfer andcommercialization can have on innovation. To understand the impactof large scale governmental spending and its returns, investments incancer research was chosen. With the goal of determining whethercourts treat different technologies differently, an analysis was con-ducted to study the technology underlying litigated patents. The dif-ferences amongst a single technology were then examined by asses-sing biotechnology with greater care. Therefore, the chosen technolo-gies are case studies that illustrate a broader legal question under thecurrent patent regime.

1.2 the method : empirical legal studies

The methodologies used in this dissertation to study innovation inemerging technologies depart from traditional doctrinal scholarship.Traditionally, the study of the law is dogmatic; and seeks to produceinformation about the laws and systematize legal norms31. Doctrinesare to be formulated from an internal perspective: The jurist should

30. The area of study itself lack a clear and accepted definition. An investigationbased on several hundreds publications in social science pinned down the attributesof emerging technologies as follows: (i) radical novelty, (ii) relatively fast growth, (iii)coherence, (iv) prominent impact and, (v) uncertainty and ambiguity, see DanieleRotolo, Diana Hicks, and Ben Martin, “What Is an Emerging Technology?,” ResearchPolicy 44, no. 10 (2015): 1–40.

31. See Aulis Aarnio, Essays on the Doctrinal Study of Law (Springer, 2011) for thedefinition and the quote by Philip Heck, “Was ist diejenige Begriffsjurisprudenz,die wir bekämpfen ?,” Deutsche Juristen-Zeitung, 1909, 1457–1461 ‘In entsprechenderWeise werden die spezielleren Gebotsbegriffe aus immer allgemeineren abgeleitet,so daß das ganze System einen logisch-deduktiven Charakter erhält [...]’ [The special

10 about this text

use the legal system as a subject of inquiry, but also use this systemas normative framework32. Through its focus, studying law from adoctrinal standpoint helps to assess legal relationships, statutes andinstitutions at a conceptual level33.

Beginning in the 1920s, US legal theorists began to adopt methodo-logies from other fields. They gained perspectives from economics,political theory, moral theory, anthropology, to name but a few. Thisapproach, termed legal realism, sought to bring the law closer to re-ality with hypotheses tested against observations of the world34. Theeconomic analysis of law has grown from the early 1960s rapidly toestablish itself as the most influential school of thought35. Law andeconomics operates based on two premises. Human choice is rational,and legal rules are instruments to correct market failures. The forma-tion, structure, processes and impact of law and legal institutions isexamined using economic theory and econometric methods36. An eco-nomic analysis of the law can be positive or normative. The positiveanalysis explains the law, predicts its effects and thereby enables ef-ficiency assessments. The normative analysis uses these results andprescribes policy recommendation maximizing social welfare37.

precepts shall therefore be derived from generalizable precepts, so that the whole system gainsa logical-deductive character.].

32. As stated by Oliver W. Holmes, The Common Law (Barns & Noble, 2004), p. 163

‘[t]he business of the jurist is to make known the content of the law; that is, to workupon it from within.’

33. See Robert Cooter, “Bargaining with the State: Offsets and Mitigation in Deve-loping Land,” in Internationalisierung des Rechts und seine ökonomische Analyse; Inter-naionalization of the Law and its Economic Analysis, ed. Thomas Eger, Claus Ott, andJochen Bigus (Gabler Edition Wissenschaft, 2008), p. 55 arguing that ‘[w]ith respectto the law, economists sometimes wonder what lawyers really study: Is the law abranch of philosophy? Is it a list of famous cases? Is it a collection of rules?’.

34. See, e.g., Lon L. Fuller, “American Legal Realism,” University of PennsylvaniaLaw Review 82 (1934): 429–462.

35. Richard A. Posner, “Legal Scholarship Today,” Harvard Law Review 115, no. 5

(2002): 1314–1326.36. Charles K. Rowley, “Public Choice and the Economic Analysis of Law” (Kluwer

Academic Publishers, 1989), p. 125.37. Heico Kerkmeester, “Methodology: General,” in Encyclopedia of Law and Econo-

mics. Volume I. The History and Methodology of Law and Economics (Edward Elgar, 2000),p. 390 and Robert Cooter and Thomas Ulen, Law & Economics (Berkeley Law Books,2016).

1.2 the method : empirical legal studies 11

Further evolution of law and economics is constituted by the empiri-cal study of law38. This investigates the operations and effects of thelaw by making use of ‘qualitative’ or ‘quantitative’ methods39. Thequalitative approach comprises an in-depth study of a modest num-ber of interaction. The quantitative method employs the methodologyof statistical analysis to pursue systematic knowledge40. In recent ye-ars, law professors have presented statistical evidence on both smal-ler issues (e.g., parking violations in New Haven)41 and systemic ones(e.g., jury versus judge verdicts in criminal trials)42. Law-related em-pirical scholarship has contributed to crucial debates about the rateof error in capital punishment, the decline in civil trials, or outcomesof employment of discrimination cases43.

In Europe, law and economics, has penetrated legal research to va-rious extent. In Switzerland especially, such scholarly work remainssparse. A handful of writings44 and seminars address the analysis

38. See Michael Heise, “An Empirical Analysis of Empirical Legal Scholarship Pro-duction, 1990-200,” University of Illinois Law Review 9 (2011): 1739–1752.

39. Mark Tushnet et al., “Empirical Research in Law,” in Oxford Handbook of LegalStudies (Oxford University Press, 2005), 173–203 and Jason Seawright and John Ger-ring, “Case Selection Techniques in Case Study Research: A Menu of Qualitative andQuantitative Options,” Political Research Quarterly 61, no. 2 (2008): 294–308.

40. See, e.g., Theodore Eisenberg, “The Origins, Nature, and Promise of EmpiricalLegal Studies and a Response to Concerns,” University of Illinois Law Review 5 (2011):1713–1738 and the casebook Peter Cane and Herbert Kritzer, The Oxford Handbookof Empirical Legal Research (Oxford University Press, 2010). Its flagship journal is theJournal of Empirical Legal Studies.

41. Underhill Moore and Charles C. Callahan, “Law and Learning Theory: A Studyin Legal Control,” Yale Law Journal 53 (1943): 1–136.

42. Harry Jr. Kalven and Hans Zeisel, The American Jury (Little, Brown / Company,1966) and Theodore Eisenberg et al., “Judge-Jury Agreement in Criminal Cases: APartial Replication of Kalven and Zeisel’s The American Jury,” Journal of EmpiricalLegal Studies 2, no. 1, 171–207.

43. Theodore Eisenberg, “Why Do Empirical Legal Scholarship?,” San Diego LawReview 41, no. 4 (2004): 1741–1746.

44. See in particular Anne van Aaken, "Rational Choice" in der Rechtswissenschaft:zum Stellenwert der ökonomischen Theorie im Recht (Nomos, 2003) arguing that legalanalysis should include the analysis the effect of the rules, but that the evaluation ofjurisprudence arguments should go first; Klaus Mathis, Effizienz statt Gerechtigkeit?Auf der Suche nach den philosophischen Grundlagen der Ökonomischen Analyse des Rechts(Duncker & Humblot, 2009), p. 205-206 stating that ‘Effizienz ist daher stets auchein Gebot der Gerechtigkeit’ [Efficiency is therefore also a precept of fairness.] and thatthe economic analysis of the law should be included in the Swiss legal education;Peter Nobel, “Einführung – Die Aktiengesellschaft als Kapitalgesellschaft,” in Berner

12 about this text

of the economic effect of the law45. This dissertation seeks to add tothe body of literature utilizing such a law and economics approach.Each paper engages the following structure. A policy issue relating topatent law is identified. The research question and hypothesis are de-fined, and in some cases refined through the use of game theoreticalmodels. For the first section, a lack of available data was remediedthrough considerable hand-collection and hand-coding. For the se-cond section, publicly-available data was pulled over and analyzed.All research is based on quantitative methods, except for Chapter 4

which is qualitative in nature. The interdisciplinary character of thedissertation is further underlined by the choice of two co-authors, onewith a background in physics, the other in economics.

Kommentar zum Aktienrecht, ed. Peter Nobel (Stämpfli, 2017), 42–64 on law and eco-nomics in corporate law; Peter Nobel, “Zur Koexistenz von Recht und Ökonomie:Ökonomische Analyse des Rechts als Aufforderung zum Dialog,” in Der Kanton St.Gallen und seine Hochschule – Beiträge zur Eröffnung des Bibliotheksbaus, ed. Rolf Dubs,Yvo Hangartner, and Alfred Nydegger (Hochschule St. Gallen für Wirtschaftsrechts-und Sozialwissenschaften, 1989), 454–462 for the coexistence between law and econo-mics; Peter Nobel, “Die wirtschaftliche Betrachtungsweise im Recht,” SchweizerischeJuristen-Zeitung 113 (2017): 457–469 for a review of the law and economics movementincluding Swiss academics; Andreas Heinemann, “Recht, Ökonomie und Realität,”in Law & Economics – Festschrift für Peter Nobel zum 70. Geburtstag, ed. Robert Wald-burger et al. (2015), 21–41 for the cultural shock between law and economics as wellas their relation to reality; and Markus Ruffner, “Reformbedarf des Aktienrechts –eine "Law and Economics" – Perspektive,” Schweizerische Juristen-Zeitung 101 (2005):253–261 for a perspective on how to reform Swiss corporate law from a law andeconomics approach.

45. Mainly the Workshop & Lecture Series in Law & Economics of the ETH Zurichand the Universities of Basel, Lucerne, St. Gallen and Zurich and the InternationalConference on Law and Economics at the University of St. Gallen

Part II

E N F O R C E M E N T O F PAT E N T S I N E U R O P E

2I N T R O D U C T O RY R E M A R K S

In the realm of international law, the ‘Westphalian’ concept of sover-eignty grants a nation-state the power monopoly with respect to itsterritories and citizens1. Against the backdrop of their supreme po-litical authority, sovereign nations are thus free to design their legalsystems. Within this context, modern patent law emerged, introdu-ced in Venice in the 15th century, with the ultimate aim of fosteringlocal technological innovation2. Nations historically drew their legalsystems with effect to their borders only, within which a set of rightsand duties were set. Rooted in the territorial limitation of a state’ssovereignty, the monopoly a patent confers has effect only withinthe territorial boundary of the octroying state3. Departing from thatprinciple, nations have begun to coordinate in order to mediate thedifferences amongst the national patent regimes, by unifying or har-monizing their legal systems.

2.1 shift towards harmonization

The desirability of a harmonized system has been stipulated on inter-national and European level. By international agreement, namely the

1. See, e.g., John H. Jackson, “Sovereignty-Modern: A New Approach to an Out-dated Concept,” American Journal of International Law 97, no. 4 (2003): 782–802 revie-wing the traditional westphalian sovereignity concepts and their current use; furtherMichael R. Fowler and Julie M. Bunck, Law, Power, and the Sovereign State: The Evolu-tion and Application of the Concept of Sovereignty (Penn State University Press, 1995).

2. Craig A. Nard and Andrew P. Morris, “Constitutionalizing Patents: From Ve-nice to Philadelphia,” Review of Law & Economics 2, no. 2 (2006): 223–321 for a histo-rical account of the Venetian statute of 1474.

3. See Hanns Ullrich, “TRIPS: Adequate Protection, Inadequate Trade, AdequateCompetition Policy,” Pacific Rim Law & Policy Association 4, no. 1 (1995): 153–210

for the doctrine of territoriality in the field of intellectual property and NorbertKoch and Franz Froschmaier, “The Doctrine Of Territoriality In Patent Law andThe European Common Market,” Idea: The Patent, Trademark, and Copyright Journal ofResearch and Education, 1965, 343–360 for the field of patent law.

15

16 introductory remarks

World Trade Organization (WTO) Agreement on Trade-Related As-pects of Intellectual Property Rights (TRIPS) that has been affirmedby 153 member states, patents shall be available and patent rightsenjoyable without discrimination as to the place of invention, techno-logy field and whether products are imported or locally produced4.

European countries have moved a step further towards homogeneityby harmonizing the national regimes under the roof of the StrasbourgConvention of 1963 and the European Patent Convention in 1973

(EPC)5. These treaties provided for a uniform body of substantivepatentability requirements as well as the establishment of Europeanpatents. In the same vein, the Member States of the EPC achieved abreakthrough agreement in 2012. The Agreement on the Unified Pa-tent Court (UPC), which will enter into force with the ratificationsof at least thirteen states including the three largest patent-grantingcountries Germany, France and the United Kingdom, introduces anew specialized patent court with exclusive jurisdiction for Europeanpatents and European patents with unitary effect. The future of theUPC, however, is uncertain after Brexit6.

In the United States and in Europe, the harmonization of the patentsystem has been challenged by both theoretical and empirical evi-dence within the last years. In the United States, the centralizationof patent appeal by the Court of Appeal for the Federal Circuit inthe 1980s hoped to harmonize patent law across the country and eli-

4. Laurence R. Helfer, “Regime Shifting: The TRIPs Agreement and New Dyna-mics of International Intellectual Property Lawmaking,” Yale Journal of InternationalLaw 29 (2004): 1–83.

5. Convention on the Unification of Certain Points of Substantive Law on Patentsfor Invention, signed November 27, 1963 and Convention on the Grant of EuropeanPatents of 5 October 1973; Kate H. Murashige, “Harmonization of Patent Laws,”Houston Journal of International Law 16 (1994): 591–614 naming it ‘[a] giant step towardharmonization’.

6. It is debated whether the UK may ratify the UPC Agreeement and remain amember after its withdrawal from the EU, see, e.g., Ansgar Ohly and Rudolf Streinz,“Can the UK Stay in the UPC System After Brexit?,” Journal of Intellectual PropertyLaw & Practice 12, no. 3 (2017): 245–258 and Winfried Tilmann, “The Harmonisationof Invalidity and Scope of Protection Practice of the National Courts of EPC MemberStates,” International Review of Intellectual Property and Competition Law, 2006, 62–74.The UK ratified the UPC Agreement on April 26, 2018.

2.2 social benefits of harmonization 17

minate forum shopping7. The debate has not vanished since then. Inrecent years, questions have arisen surrounding the advantages ofharmonization versus diversity8. In particular, a strand of the litera-ture considers that courts should tailor the uniform patent law rulesto the different industries9. In Europe, important differences betweencountries in terms of litigation outcomes, share of appealed cases andcharacteristics of litigants and litigated patents have been revealed inthe last years10. In some instances, one same patent can be foundvalid by the courts of one country, but not by the one of another.

2.2 social benefits of harmonization

Whether Europe should follow the path of a homogeneous or a he-terogeneous patent system is of significant importance, and both sy-stems bear their advantages and limitations. From an economic per-spective, the choice of regulation of patent laws by an union of he-terogeneous countries should be driven by cost and benefit analysis

7. On discussions of the history, rational and effects of the Court of Appeal for theFederal Circuit, see Jonas J. Anderson, “Court Competition for Patent Cases,” Uni-versity of Pennsylvania Law Review 163, no. 3 (1997): 631–698 and Charles W. Adams,“The Court of Appeals for the Federal Circuit: More Than a National Patent Court,”Missouri Law Review 49 (1984): 43–84.

8. For scholars rethinking uniformity, see John F. Duffy, “Harmony and Diver-sity in Global Patent Law,” Berkeley Technology Law Journal 17 (2002): 685–726 andCraig A. Nard and John F. Duffy, “Rethinking Patent Law’s Uniformity Principle,”Northwestern University Law Review 101, no. 4 (2007): 1619–1676; for the benefits ofglobal harmonization, see Randy L. Campbell, “Global Patent Law Harmonization:Benefits and Implementation,” Indiana International & Comparative Law Review 605,no. 13 (2003): 605–638; and for a simulation of the geography of patent litigation, seeJeanne C. Fromer, “Patentography,” New York University Law Review 85 (2010): 1444–1520.

9. Dan L. Burk and Mark A. Lemley, The Patent Crisis and How the Courts Can SolveIt (The University of Chicago Press, 2009); Dan L. Burk and Mark A. Lemley, “PolicyLevers in Patent Law,” Virginia Law Review 89, no. 7 (2003): 1575–1696; Dan L. Burkand Mark A. Lemley, “Is Patent Law Technology-Specific?,” Berkeley Technology LawJournal 17 (2002): 1155–1206; and John R. Allison, Mark A. Lemley, and David L.Schwartz, “Our Divided Patent System,” University of Chicago Law Review 82, no. 3

(2015): 1073–1154.10. Katrin Cremers et al., “Patent Litigation in Europe,” European Journal of Law and

Economics 44, no. 1 (2017): 1–44.

18 introductory remarks

of a harmonized versus heterogeneous legal patent system11. Harmo-nization of patent laws across nation-states is regarded as creatingsocial benefits by policymakers and many patent scholars: reducingcosts for inventors and their assignees while creating a simpler anduniform protection system throughout the world12.

Such a level playing field shall optimize the worldwide pace of techno-logical innovation13. For economists, harmonization is viewed as theresponse to jurisdictional externalities, economies of scale in gover-nance and destructive protectionism. Negative externalities providesupport for harmonization, since companies of a country without apatent system will free ride on the investments made by the firms ofthe neighbor countries with patent protection14. This argument pre-vailed in the negotiations of the TRIPS agreement between developedcountries with stronger patents and free riding developing nations15.If the patent office is considered being a classical natural monopoly,because the entire market can be satisfied at the lowest cost by oneoffice rather than by two or more, its average costs of service de-

11. See Alberto Alesina, Ignazio Angeloni, and Federico Etro, “International Uni-ons,” American Economic Review 95 (2005): 602–615 and Alberto Alesina, Angeloni Ig-nazio, and Schuknecht Ludger, “What Does the European Union Do?,” Public Choice123 (2005): 275–319 for the trade-off between the benefits of centralization (economiesof scale, positive externalities) and its costs (heterogeneity of preferences).

12. See Duffy, “Harmony and Diversity in Global Patent Law” quoting the formerhead of the U.S. Patent and Trademark Office Todd Dickinson in this sense: ‘I thinkmost of us [..] would say that there definitely should be a global patent systemof some sort by 2010. I think we can all list probable benefits of such a system:reduced costs for inventors and for their assignees, dramatically simpler protection,and uniformity of that protection throughout the world.’

13. Or, as Carl Moy, “Patent Harmonization, Protectionism and Legislation,” Jour-nal of the Patent and Trademark Office Society 74 (1992): p. 782 says, at least operate tothe net benefit of US entities.

14. For the classical free-rider problem applied to the international patent system,see Peter K. Yu, “Five Disharmonizing Trends in the International Intellectual Pro-perty Regime,” in Intellectual Property and Information Wealth: Issues and Practices inthe Digital Age, ed. Peter K. Yu (Cambridge University Press, 2004), 73–112.

15. See Duffy, “Harmony and Diversity in Global Patent Law” and Peter S. Menelland Suzanne Scotchmer, “Intellectual Property Law,” in Handbook of Law and Econo-mics, ed. Mitchell A. Polinsky and Steven Shavell (Elsevier, 2007), 1473–1570 for thediscussion that countries may argue for harmonization serving their own interestrather than for the efficient regime, and that countries that promote harmonizationare those with large markets or with more innovation.

2.2 social benefits of harmonization 19

cline as its coverage expands16. Economy of scales in governance, ad-ministration and prosecution of patents support harmonization asefficient solution17. Finally uniformity provides a safeguard againstdestructive protectionism based on the argument that it reduces over-all welfare by impeding innovation, free trade and the stability ofthe international community18. Article 3 of the TRIPS – the generalanti-discrimination clause – imposes in this regard that each WTOMember shall accord to the nationals of other Members treatment noless favorable than it accords to its own nationals with regard to theprotection of intellectual property.

Contrastingly, territoriality and its market-dividing effects providethe benefit of inter-jurisdictional legal diversity. As such, it can helpto match the level of public goods to the unique preferences of theconsumers, the citizens of a local population19. In this sense, territo-riality allows firms to focus their patenting strategies with considera-tion to their markets of interest and competitors20. Interjurisdictional

16. For the standard literature on natural monopolies, see Richard A. Posner, “Na-tural Monopoly and Its Regulation,” Stanford Law Review 21, no. 3 (1969): 548–643;William J. Baumol, “Contestable Markets: An Uprising in the Theory of IndustryStructure,” American Economic Review, 72, no. 1 (1975): 1–15; and Sanford V. Berg andJohn Tschirhart, Natural Monopoly Regulation: Principles and Practice (Cambridge Uni-versity Press, 1988); for the patent office as a case of classical natural monopoloy, seeDuffy, “Harmony and Diversity in Global Patent Law,” p. 700.

17. See Yu, “Five Disharmonizing Trends in the International Intellectual PropertyRegime” and Alesina, Ignazio, and Ludger, “What Does the European Union Do?”stating that ‘the EU [...] should focus exclusively on policy areas where economiesof scale are large, and internalizing externalities is important’.

18. John O. McGinnis and Mark L. Movsesian, “The World Trade Constitution,”Harvard Law Review, 2000, 511–605 contending that the WTO reduces the power ofprotectionist interest groups, while simultaneously promoting international tradeand domestic democracy; see also on this point Nikolaus Thumm, “Introduction,” inIntellectual Property Rights: National Systems and Harmonization in Europe, ed. NikolausThumm (Springer, 2000), 1–4.

19. See Simon Deakin, “Legal Diversity and Regulatory Competition: Which Modelfor Europe?,” European Law Journal 12, no. 4 (2006): 440–454 describing the mid-1950’s history of regulatory competition formalized in relation to the issue of theproduction of local public goods; and Duffy, “Harmony and Diversity in GlobalPatent Law,” p. 704 for strong and weak matching of local preferences.

20. Ullrich, “TRIPS: Adequate Protection, Inadequate Trade, Adequate Competi-tion Policy,” p. 190; Rochelle C. Dreyfuss and Andreas F. Lowenfeld, “Two Achie-vements of the Uruguay Round: Putting TRIPS and Dispute Settlements Together,”Virginia Journal of International Law Association 37 (1997): p. 296: ‘[h]ow that problem

20 introductory remarks

competition can pressure government behavior, checking harmful ten-dencies and encouraging efficiency, for instance when an innovativehigh-spending government pressures others to follow suit21. Legal ex-perimentation also may lead to the discovery of superior practices22.Particularly, transnational standard within the European Union aimat promoting this diversity by leaving room for the emergence of au-tonomous solutions. Whereas downwards regulations are prohibited,Member States are encouraged to improve on the standards that havebeen set23.

The most prevalent example of the harmonization of legal systemsis the formation of the European Union. As such, the Member Statesfollowed the view that harmonization is, at least in certain areas, to bepreferred to legal diversity24. It has been estimated that bilateral tradeof services within the European Union could be boosted by 30 to 60 %by a uniform regulatory system25. Harmonization, even if designed tofit within national context on the surface, may nevertheless give riseto conceptual divergences on the interpretation of harmonized laws.This phenomena may be explained by the fact that Member Statesface conflicting incentives: whereas there is a need for harmonizationfor the regulation of transnational trading flows and relationships,

is best solved can depend on a country’s intellectual and industrial development, itsculture, and the types of creative work in which its citizens are engaged’.

21. See John Shannon, “Federalism’s "Invisible Regulator" – InterjurisdictionalCompetition,” in Competition Among States and Local Governments: Efficiency and Equityin American Federalism, ed. Daphne Kenyon and John Kincaid (Urban Institute Press,1991), 117–126 tracing this acceleration effect back to two factors: the ‘pacesetter phe-nomenon’ and the ‘catch-up imperative’.

22. Lisa L. Ouellette, “Patent Experimentalism,” Virginia Law Review 101 (2015): 65–128 and Alexander Stack, International Patent Law: Cooperation, Harmonization and anInstitutional Analysis of WIPO and the WTO (Edward Elgar, 2011), p. 35, 62-83 for thediffusion of deferred examination policies initiated by the Dutch Patent Office.

23. Deakin, “Legal Diversity and Regulatory Competition: Which Model for Eu-rope?,” p. 452 highlighting that ‘central-level intervention is the precondition forcontinued local-level experimentation’.

24. Nevertheless, see Alesina, Angeloni, and Etro, “International Unions” findingevidence suggesting that the normative policy prediction of harmonization by theEU does not satisfy the broad criteria of optimality in certain areas such as socialprotection where the range of prerogatives has been expanded considerably andbeyond optimality.

25. Henk Kox, Arjan Lejour, and Raymond Montizaan, “The Free Movement ofServices Within the EU,” CPB Document, no. 69 (2004) for estimates.

2.3 dimensions of heterogeneity 21

the costs of harmonization – learning, adapting and risk of failure –may be considerable and are coupled with the loss of the advantagesresulting from territoriality26.

2.3 dimensions of heterogeneity

Within the context of the harmonization of legal systems, this the-sis investigates the harmonization of the European patent system. Inparticular, it studies the harmonization of patent litigation througha set of dimensions that may be tested empirically. Heterogeneity inpatent litigation is defined as variances in distribution and outcomeof court decisions. Deriving from a one-size-fits-all patent system inwhich courts deal with patents granted under harmonized patentabi-lity requirements rules, determinants that may explain differences indistribution and outcome of litigation are sought. To address this, theglobal ‘Uniformity Level’ is distinguished from the European ‘Har-monization Level’ (summarized in Figure. 2.1).

26. See Pierre Larouche, “Legal Emulation Between Regulatory Competition andComparative Law,” in National Legal Systems and Globalization: New Role, ContinuingRelevance, ed. Pierre Larouche and Péter Cserne (Springer, 2013), p. 282 for a dis-cussion of the clash between the strong political will at the superior level and theweaker will in the broader legal order.

22 introductory remarks

Harmonizatio

n Level

Substantive Rules

Procedural Rules

Uniform

ityLevel

Players

Industry

Technology

Figure 2.1: Dimensions of Heterogeneity

The ‘Uniformity Level’ comprises of technology-, industry-, and litigants-heterogeneity dimensions which can be found in any patent system.Contrastingly, the ‘Harmonization Level’ is concerned with the Euro-pean harmonization of substantive and procedural rules27.

2.3.1 Dimensions on Uniformity Level

The TRIPS Agreement praises a global shift towards a monolithictechnology- and industry-neutrality of the patent system. At the timethe TRIPS was drafted in 1994, the Agreement sought for a standardof technological neutrality intended to establish patent protection for

27. Note that distinguishing between the judge’s bias goes beyond the scope of thisdissertation, with a multitude studies to the effects of gender, religion or politicalpreferences on case outcomes within the past two decades, see for a short overviewStephen J. Choi and Mitu G. Gulati, “Bias in Judicial Citations: A Window into theBehavior of Judges?,” Journal of Legal Studies 37 (2008): 87–129.

2.3 dimensions of heterogeneity 23

various inventions for which patents were not available in many coun-tries, such as on pharmaceuticals and agricultural chemistry28. Uni-formity in the set of exclusive rights granted to the inventor washistorically justified by the relative homogeneity of the patented in-ventions and the applicable market conditions29. Further, technologyneutrality is assumed to promote equally old and new technologies30.With the break-in of biotechnology and software in the 1990s, vari-ances across the needs of innovators led to an increasingly dividedpatent system31.

Within the uniform patent system, the social costs and benefits ofpatents vary significantly across technologies and industries32. Theapplication of a socially costly uniform solution to problems whosemagnitude vary imposes uniformity costs, and thus underprotectsthose that would or do invest in costly innovation while overpro-tecting those with low costs of innovation33. Based on this argument,scholars have argued that the current uniform system is adapted tothe chemical and pharmaceutical industry whereas it imposes bar-riers on innovation for information technologies industries34.

28. With countries such as India not granting patent protection for pharmaceuticalsand food, see Martin J. Adelman and Sonia Baldia, “Prospects and Limits of thePatent Provision in the TRIPS Agreement: The Case of India,” Vanderbilt Journal ofTransnational Law 20 (1996): 507–534; and on the new international dynamics afterthe passing of the TRIPS, see Helfer, “Regime Shifting: The TRIPs Agreement andNew Dynamics of International Intellectual Property Lawmaking.”

29. See Michael W. Carroll, “One for All: The Problem of Uniformity Cost in In-tellectual Property Law,” American University Law Review 55, no. 4 (2009): 845–900

contending that reducing the uniformity costs in intellectual property law is thecentral problem for policymaking.

30. Brad A. Greenberg, “Rethinking Technology Neutrality,” Minnesota Law Review100 (2016): p. 1495 arguing that this neutrality is, however, often, suboptimal andself-defeating (and also not neutral).

31. For a systematic review on the differences in the assessment of these twotechnologies by courts, see Burk and Lemley, The Patent Crisis and How the CourtsCan Solve It, p. 51 and Burk and Lemley, “Policy Levers in Patent Law,” p. 1581.

32. This argument is the starting point for the work by Burk and Lemley, The PatentCrisis and How the Courts Can Solve It, p. 256 and James Bessen and Michael J. Meurer,Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovation at Risk (PrincetonUniversity Press, 2008), p. 51.

33. Carroll, “One for All: The Problem of Uniformity Cost in Intellectual PropertyLaw,” p. 847.

34. Bessen and Meurer, Patent Failure: How Judges, Bureaucrats, and Lawyers Put In-novation at Risk, p. 16; Burk and Lemley, The Patent Crisis and How the Courts Can

24 introductory remarks

Determining whether the users of the patent system experience divi-sion or uniformity is of first importance to derive future patent poli-cies. The dimensions of heterogeneity challenging the one-size-fits-allpostulate are twofold, mainly a technology- and industry-specificityand, in the second place, a party- and narrative-specificity:

technology- and industry-specificity — The difficulty, timerequirements and costs of patenting today vary considerablyamongst the technologies the patented invention embody aswell as the industry in which they are implemented. Whereasthe patent system ideologically treats technologies uniformly,the social costs and benefits of patents are heterogeneous acrosstechnologies and industries35. On a doctrinal level, most of thesedifferences can be explained by the fact that patent law lacks asingle theoretical foundation: depending on the theorists andthe industry in cause, patents should not exist, be broad or nar-row; granted to innovators or improvers36. Hence, uniform rulesmay favor some industries over others, and the pharmaceuticalindustry is often considered to drive the patent system37. Thus,the patent system applies the identical set of rules to the con-ception of drugs, and their decade-lasting research, design, andclinical trial process over a regulatory labyrinth costing in the

Solve It, p. 3; and Michael W. Carroll, “One Size Does Not Fit All: A Framework forTailoring Intellectual Property Rights,” Ohio State Law Journal 70, no. 6 (2009): p. 1390

comparing the operation of the pharmaceutical and computer software industry.35. For developments and examples, see Burk and Lemley, The Patent Crisis and How

the Courts Can Solve It, p. 41; James Bessen and Michael J. Meurer, “The Private Costsof Patent Litigation,” Boston University School of Law Working Paper, nos. 07-08 (2007);Burk and Lemley, “Is Patent Law Technology-Specific?”; and Dietmar Harhoff andStefan Wanger, “The Duration of Patent Examination at the European Patent Office,”Management Science 55, no. 12 (2009): p. 1351.

36. For an overview of the different justifications of patent law and their limitationsacross industries, see Burk and Lemley, “Policy Levers in Patent Law,” p. 73.

37. Brian Kahin, “Through the Lens of Intangibles: What Patents on Software andServices Reveal about the System,” in Patents, Innovation and Economic Performance,OECD Conference Proceedings, ed. Adam B. Jaffe, Josh Lerner, and Scott Stern (OECDPublishing, 2004), p. 218 and Claus R. Gawel, “Patent Protection as a Key Driver forPharmaceutical Innovation,” Pharmaceuticals Policy and Law, 18, nos. 1-4 (2006): 45–53

summarizing that ‘a well-functioning patent protection system is a prerequisite forattracting finance for costly pharmaceutical research, given its high failure rates, byensuring that successful innovation is rewarded’.

2.3 dimensions of heterogeneity 25

hundred of millions USD per new drug than to electric ice cubtrays, toothbrushes and umbrellas38.

party- and narrative-specificity — Innovators use the patentsystems in different ways. Most patents are nowadays issued tolarge firms39. Through their standardized R&D program, thesecorporations arguably engage in ‘routine innovation’ directedtowards their competitors40. Despite their smaller absolute sharein patenting, individual inventors and small-and-medium sizedcompanies are about three-times more likely to enforce their pa-tents than large corporations in the US41. Whereas no data onthis phenomenon exists for Europe, there are good reasons tobelieve that the situation is similar. Furthermore, the strategicconstellation of the disputes may be a potential driver of patentlitigation. There are differences when patent assertion entities,pure enforcement companies without manufacturing activity, li-tigate their patents or when large multinationals enforce theirclaim against their competitors42.

2.3.2 Dimensions on Harmonization Level

Within the European patent system, lack of homogeneity on nationaland cross-national levels can result from variances in the substantiveand procedural rules applied by the national courts. Such divergences

38. Pharmaceutical Research and Manufacturers of America estimates the costs oflaunching a new drug of $1.3 billion, whereas further studies supported these num-bers ($802 million, Joseph A. DiMasi, Ronald W. Hansen, and Henry G. Grabowski,“The Price of Innovation: New Estimates of Drug Development Costs,” Journal of He-alth Economics 22 (2003): 151–185) or challenged them ($43 million, Donald W. Lightand Rebecca Warburton, “Demythologizing the High Costs of Pharmaceutical Rese-arch,” Management Science 50 (2004): 804–820).

39. In a sample of 1,000 US patents, 707 are assigned to large entities, see John R.Allison and Mark A. Lemley, “Who’s Patenting What? An Empirical Exploration ofPatent Prosecution,” Vanderbilt Law Review 6 (2000): p. 2117.

40. James Bessen and Michael J. Meurer, “Lessons for Patent Policy from EmpiricalResearch on Patent Litigation,” Lewis and Clark Review 9, no. 1 (2005): p. 12.

41. Allison and Lemley, “Who’s Patenting What? An Empirical Exploration of Pa-tent Prosecution,” p. 2128.

42. See Christopher A. Cotropia, Jay P. Kesan, and David L. Schwartz, “Heteroge-neity Among Patent Plaintiffs: An Empirical Analysis of Patent Case Progression,Settlement, and Adjudication,” Journal of Empirical Legal Studies 15, no. 1, 80–125.

26 introductory remarks

create incentives for forum shopping for the litigants, since a rationalplaintiff will select the court in which the expected value of its claim– less the costs of litigating – is the highest based on the set of appli-cable substantive and procedural rules before the court43.

substantive rules — The application of uniform substantial pa-tent laws may vary across countries. These divergences may beexplained by the fact that although the creation of uniform sub-stantial patent laws was the major objective of the framers of theEuropean Patent Convention, they failed to completely do so44.Before the entry into force of the EPC, illustratively, two oppo-site methodologies of patent claims interpretation prevailed inEurope. A more restrictive methodology was applied by courtsin the UK and in Switzerland, where claims defined both the in-vention and limited it. In Germany and the Netherlands, howe-ver, claims defined the invention only, being the starting pointof the scope of protection45. The EPC hoped to bridge the gapbut ultimately defined a position between these two extremes46.Displaying the divergences in application that followed across

43. Christopher A. Whytock, “The Evolving Forum Shopping System,” Cornell LawReview 96, no. 3 (2011): 481–534 contending that ‘[t]o shop among those legitimatechoices for the forum that offers the potential for the most favorable outcome is theonly rational decision under rational choice theory and game theory because forumshopping maximizes the client’s expected payoff’.

44. Hugh Laddie, “Kirin Amgen - The End of Equivalents in England?,” Internatio-nal Review of Intellectual Property and Competition Law, 2009, 3–38 on how the benefitssecured by the creation of a central granting authority are undermined if the grantedpatents vary in their effects in the different Member States of the EPC.

45. For a summary of these differences, see scholarship by Toshiko Takenaka, Inter-preting Patent Claims: The United States, Germany and Japan (Wiley-VCH, 1995); DonaldS. Chisum, “Common Law and Civil Law Approaches to Patent Claim Interpretation:"Fence Posts" and "Sign Posts",” in Intellectual Property in the New Millennium: Essaysin Honour of William R. Cornish, ed. David Vaver and Lionel Bently (Cambridge Uni-versity Press, 2004); Dan L. Burk and Mark A. Lemley, “Fence Posts or Sign Posts:Rethinking Patent Claim Construction,” University of Pennsylvania Law Review 157

(2009): 1743–1799; and the judicial decision Kirin-Amgen v. Roche, TranskaryoticTherapies et al., House of Lords, [2004] UKHL 46, October 21, 2004.

46. Article 69 EPC states that patent claims determine the extent of the protectionwhile the description and drawings shall be used to inteprete the claims, followingthe peripheral claiming approach. But the Protocole on the Interpretation of Article69 EPC, adopted at the Munich Diplomatic Conference on 5 October 1973 decreedthat the claims were neither to be limited to the literal meaning, neither as a guidelineonly, but that the track shall be a position between the extremes. As such, the article

2.3 dimensions of heterogeneity 27

Europe, in the Epilady litigation, judges in Belgium, Germany,Italy and the Netherlands found the patent at dispute infringed,and courts in Austria, France and the United Kingdom judgedthe opposite47.

procedural rules — In the event of a dispute, the enforcementof patents will be subject to a highly variable institutional setof enforcement rules, whose harmonization is not complete. Onthe institutional court level, the dual German enforcement sy-stem – so-called bifurcation – dissociates infringement and vali-dity claims which are heard by separate courts. In other coun-tries, both issues are tried by the same court and judges. In thesame vein, the court systems are also characterized by substan-tial variances in their cost structure. In the UK, patent litigationcosts are substantially higher than in other states48. Quality andspeed of the courts, as a result of the form of the proceedingsand experience of the judgment-makers with patent cases, alsodiverge. A range of procedural provisions such as provisionalmeasures, language of proceedings or rulings on evidence andthe onus of proof characterize the individual states49. Finally, asto the level of damages, the EU Enforcement Directive has har-monized the calculation methods, but variances in the amountof the level of damages across countries exists50.

defines a position striving a balance between protection for the patentee, reflectingthe German and Dutch approach, and legal certainty for third parties, the UK focus.

47. John P. Hatter, “The Doctrine of Equivalents in Patent Litigation: An Analy-sis of the Epilady,” Indiana International & Comparative Law Review, 1995, 461–494

and Matthew Parker, “Giving Teeth to European Patent Reform, Overcoming Re-cent Legal Challenges,” Emory International Law Review 26 (2012): 1079–1110 arguingthat ‘[i]n actuality, the national courts of Europe have construed patents differentlydespite applying the so-called uniform requirements of the EPC’.

48. With costs in the UK in an aggregate of between £1 million and £6 million, seeChristian Helmers and Luke McDonagh, “Patent Litigation in the U.K.: An EmpiricalSurvey 2000–2008,” Journal of Intellectual Property & Practice 8 (2013): 846–861. Thesefigures are to be compared to costs of EUR 50,000 to 200,000 in France and of EUR25,000 to EUR 91,456 for the court fees and of EUR 40,000 to EUR 100,000 for attorneyfees in Germany, see Cremers et al., “Patent Litigation in Europe.”

49. See the detailed analysis by Stefan Luginbuehl, European Patent Law - Towards aUniform Interpretation (Edward Elgar, 2011), p. 64.

50. See Stuart J. Graham and Nicolas Van Zeebroeck, “Comparing Patent LitigationAcross Europe: A First Look,” Stanford Technology Law Review 17 (2014): p. 667 esti-

28 introductory remarks

Fragmentation in litigation leads to an increase of litigation tacticssuch as forum shopping and wasteful duplication. In the single Euro-pean market, infringement and revocation in multiple countries of auniformly granted patent are therefore subject to the risk of divergentoutcomes. Harhoff (2009) estimated that the total private savings fromhaving access to a unified patent court would range between EUR 147

to 289 million annually.

2.4 litigation as evidence of heterogeneity

Studying the harmonized substantial laws across Europe would notbe sufficient to determine the heterogeneity that may occur in practice.Therefore, litigation is studied, heterogeneity revealing itself in caseoutcomes.

2.4.1 Litigation as Evidence

Only a very small amount of patents are likely to be involved in liti-gation, with estimations varying between 1% and 3% in most patentsystems51. Litigation is relied on, because despite the fact that thisnumber is small, the signals sent by decisions to the patent holders,potential infringers and third parties seeking to avoid patent conflictsare extremely important52. Certainty in the construction of claims andinterpretation of laws by the courts allows a potential infringer to ‘re-asonably foresee’ which actions will potentially infringe a patent53.Uncertainty in the enforceability of a patent leads to an increase of thedivergences in the expectations as well as in the information stake of

mating the awarded damages in Belgium, France and Italy as low, in Germany andthe Netherlands as average, and in the UK as high; see EU Directive 2004/48/EC.

51. Mark A. Lemley, “Rational Ignorance at the Patent Office,” Northwestern Uni-versity Law Review 95 (2001): p. 95 estimating that of about two million US patentsin force, only about 2,000 different patents are subject to court disputes and, of that,about 100 cases per year actually make it to trial.

52. Dietmar Harhoff, “Economic Cost-Benefit Analysis of a Unified and IntegratedEuropean Patent Litigation System,” Final Report, Tender No. Markt/2008/06/D, 2009,

53. Parker, “Giving Teeth to European Patent Reform, Overcoming Recent LegalChallenges,” p. 1091 and Josh Lerner, “Patenting in the Shadow of Competitors,”Journal of Law and Economics 38, no. 2 (1995): 463–495.

2.4 litigation as evidence of heterogeneity 29

the parties, which are, according to the theoretical literature, ultima-tely increasing the likelihood of a dispute of being litigated in courtrather than settled54. Allison et al. (2004) have argued that litigatedpatents are an indicator of the value of patents55. Whereas two-thirdsof the patents expire for lack of payment of the few thousands dol-lars maintenance fee, patent holders will engage in highly expensivelitigation for the defense of a fraction of patents.

In the common wisdom, Germany is considered the most ‘patent-friendly’ forum in the European Union by patent holders, who sub-jectively perceive the level of success to be very high before Germancourts. Elmer and Lewis (2010) argue that, based on litigation datashowing that 63 % of the cases between 2006 and 2009 were won bypatent holders, Germany should be qualified as such56. The authorsalso qualify the courts of Netherlands and France as ‘patent-friendly’based on diverse patent litigation factors such as win rate and costsof litigation57. The United Kingdom, on the contrary, is subjectivelyperceived as the ‘most patent-unfriendly’ country in Europe becauseof low win rates, long time to trial and high costs of litigation58. Illus-tratively, when the alleged infringer brought the case to the PatentsCourts in London, patent holders only won in 14 % of the cases in2006 to 2009

59.

54. Harhoff, “Economic Cost-Benefit Analysis of a Unified and Integrated Euro-pean Patent Litigation System,” p. 43 and Robert H. Gertner, “Asymmetric Informa-tion, Uncertainty, and Selection Bias in Litigation,” University of Chicago Law SchoolRoundtable 1, no. 1 (1993).

55. Not being worth enforcing because they cover a worthless invention or thepatent itself has no economic value, e.g., the patent claims are drafted narrowly andinclude elements easy to circumvent, see John R. Allison et al., “Valuable Patents,”Georgetown Law Journal 92 (2004): p. 92.

56. Michael Elmer and Stacy Lewis, “Where to Win: Patent-Friendly Courts Revea-led,” Managing Intellectual Property, October 2010, p. 2.

57. For France, also see Isabelle Romet, Amandine Métier, and Dora Talvard, “Pa-tent Enforcement in France,” in Patent Enforcement Worldwide,: Writings in Honour ofDieter Stauder, ed. Christopher Health (BIOS Scientific Publishers, 2015), p. 92.

58. Only 35 % of the patents at trial were found valid in the UK from 2000 to 2010,see Graham and Van Zeebroeck, “Comparing Patent Litigation Across Europe: AFirst Look,” p. 695.

59. Elmer and Lewis, “Where to Win: Patent-Friendly Courts Revealed,” p. 4.

30 introductory remarks

2.4.2 Empirics on Litigation

A literature review of patent litigation studies in Europe is presented,followed by those analyzing biotechnology patent litigation.

2.4.2.1 Europe

In Europe, the first large-scale empirical study of patent litigation wasled by Graham and Van Zeebroeck (2014)60. Relying on a database ofnearly 9,000 patent suits in the seven largest countries in the Euro-pean Union, the authors showed that the incidence and base of ju-dicial outcomes widely vary by country and by patented technology.Typically, the likelihood litigation results in a final decision in theNetherlands is much higher than in the UK. This study presented aEuropean landscape characterized by fragmentation and uncertaintyfor innovators, patent holders and their competitors.

A second empirical project was then launched to overcome the mainshortcoming of the latter study: The lack of comprehensive data oncourt cases across jurisdictions, and, for Germany and the UK, theaddition of data on settlement. Cremers et al. (2017) furnished anexhaustive picture of the litigation landscape in Germany, France,the Netherlands and the United Kingdom by evaluating a datasetof hand-coded suits filed between 2000 and 2008. They found sub-stantial differences across countries in terms of case loads, outcomes,share of appealed cases and characteristics of litigants and patents61.In summary, German regional courts hear the highest number of ca-ses in Europe, and the German cases are more likely to settle than inthe UK. When it comes to the speed of the proceedings, claims aredecided fastest in the UK. Interestingly, the authors established thatthe vast majority of patents are subject to litigation in only one coun-try. Litigation in the UK is an exception here (again), with one outof four cases litigated in parallel in another European jurisdiction. InGermany, litigation is dominantly led by domestic entities and in themachinery and engines field, while in the UK, mostly foreign entitieslitigate over pharmaceutical, chemistry and electronics products.

60. Graham and Van Zeebroeck, “Comparing Patent Litigation Across Europe: AFirst Look.”

61. Cremers et al., “Patent Litigation in Europe.”

2.4 litigation as evidence of heterogeneity 31

In addition, a handful of studies focused on patent enforcement inone particular jurisdiction. For the UK, Helmers and McDonagh (2013)built a dataset containing the complete set of patent litigation filed atthe courts in England and Wales from 2000 to 2008, showing the distri-bution by technology and by sector of the litigating companies62. Theauthors demonstrated that litigated patents are considerably more va-luable than a set of matched sample of control patents – as measuredby standard patent value proxies. In Germany, Cremers (2004) docu-mented the determinants of patent infringement suits from 1993 to1995 before the three primary infringement courts. The article pre-sented evidence that more valuable patents and patents having sur-vived opposition are more likely to encounter subsequent litigationactions63. Last, in France, Véron (2010) evaluated the decisions ofthe Tribunal de Grande Instance de Paris and subsequent appealsbetween 2000 and 2009

64. Amongst a set of variables, this data dis-plays the distribution by technology and shows that litigated patentsmostly belong to the mechanical technology. Specific problematics im-pacting the patent system or litigation also were studied in Europe.Mainly, the presence of non-practicing entities in Germany and inthe UK was analyzed by Love et al. (2017) and Helmers, Love, andMcDonagh (2014)65. The studies found that roughly ten percent ofpatent suits were filed by patent assertion entities between 2000-2013

for the UK and 2000-2008 for Germany.

In the United States, there is a large stream of empirical literatureassessing various angles of patent prosecution and enforcement. Thefirst statistical report to our knowledge was published by Federico(1956) and investigated the outcome of adjudicated patents from 1925

62. Helmers and McDonagh, “Patent Litigation in the U.K.: An Empirical Survey2000–2008.”

63. Katrin Cremers, “Determinants of Patent Litigation in Germany,” ZEW Discus-sion Paper, nos. 04-72 (2004).

64. Pierre Véron, “Le Contentieux des Brevets d’Invention en France, Étude Statis-tique 2000-2010,” 2010, for French descriptive data.

65. Brian Love et al., “Patent Assertion Entites in Europe,” in Patent Assertion En-tities and Competition Policy, ed. Daniel D. Sokol (Cambridge University Press, 2017),104–129 and Christian Helmers, Brian J. Love, and Luke McDonagh, “Is There a Pa-tent Troll Problem in the U.K.?,” Media and Entertainment Law Journal 24 (2014): 509–553.

32 introductory remarks

to 195466. Further studies containing descriptive statistics mainly in-

clude Koenig (1974), who collected patent cases from 1953 to 1978, fin-ding wide disparities in validity rates across the regional circuits andthat obviousness (lack of invention) was the most frequent basis forinvalidation67. The first modern large-scale study on patent litigationthen was Allison and Lemley (1998), who reported descriptive statis-tics about patent validity decisions from 1989 to 1996

68. Amongst arange of interesting findings, they reported that novelty argumentsfare better in courts than arguments regarding obviousness, or thatthe validity rates in different fields of technology did not vary. Thesame authors then updated their study with patent cases filed in 2008

and 2009. In short, Allison, Lemley, and Schwartz (2014) found thatthe overall dynamics remain the same: Patentees win at trial, but noton summary judgment, and patentees win individual issues but notoverall69. A landmark study undertook to evaluate the question ofthe technology- and industry-specificity of litigation. This based onsingle observations that courts seemed to treat certain patented in-ventions differently. Allison, Lemley, and Schwartz (2015) analyzedthe outcome by technology and industry of all substantive decisi-ons rendered by courts in every patent suit filed in 2008 and 2009

70.They found large differences between them. Illustratively, chemistrypatents had a significantly higher likelihood to prevail than softwareor biotechnology counterparts.

‘Protection Heterogeneity in a Harmonized System’ builds on this scho-larly work. While studies in Europe have shown differences in thedistribution and outcome of patent litigation across countries, thequestion of the differences by technology and industry were not sub-ject to study to date. The paper seeks to provide for a comprehensiveunderstanding of whether judges treat different technologies and in-dustries differently. To do so, the paper bases on hand-coding of the

66. Pasquale J. Federico, “Adjudicated Patents, 1948-1954,” Journal of the Patent Of-fice Society 38 (1956): 233.

67. Gloria Koenig, Patent Invalidity: A Statistical and Substantive Analysis (Clark Bo-ardman, 1974).

68. John R. Allison and Mark A. Lemley, “Empirical Evidence on the Validity ofLitigated Patents,” IPLA Quarterly Journal 26 (1998): 185–226.

69. John R. Allison, Mark A. Lemley, and David L. Schwartz, “Understanding theRealities of Modern Patent Litigation,” Texas Law Review 92 (2014): 1769–1801.

70. Allison, Lemley, and Schwartz, “Our Divided Patent System.”

2.4 litigation as evidence of heterogeneity 33

technologies and industries, another novel feature for a European pa-tent litigation study. It is the first attempt to determine whether theforum of dispute or the characteristic of the cases are decisive in Eu-ropean patent enforcement. Finally, it reviews a new timespan, theyears 2008 to 2012.

2.4.2.2 Biotechnology

Studies targeting biotechnology have only occurred in the United Sta-tes so far. Mills and Tereskerz (2010) have analyzed 72 lawsuits in-volving 28 stem cell patents rendered between 2006 and 2010, andidentified the plaintiffs and litigants71. In particular, the article foundthat the rate of litigation to issued patents was extremely small, butwhen litigated, it was often asserted more than once. Furthermore, itappears that universities were proactive in protecting their intellec-tual property. Holman (2008) led a study of 31 human gene patentlitigations between 1987 and 2010, and analyzed the outcome of the 6

decisions reaching a judgment in the merits72. Depending on the sub-field, litigation varied consequently. For instance, lawsuits involvingdiagnostic testing mainly settled, while therapeutic proteins rarely. Inthe study by Allison, Lemley, and Schwartz (2015), the outcome of 50

biotechnology decisions (cases from 2000 and 2009) was assessed73.Surprisingly, only 2 of 37 biotechnology patentees prevailed at trial,compared to a roughly 25 % overall win rate across all technologies.As a response to this finding, Holman (2015) investigated whether bi-otechnology patents really overwhelmingly lose, and argued that theoutcome was favorable for patentees in 44% of the time by focusingon favorable litigation outcome74.

71. Ann E. Mills and Patti M. Tereskerz, “Empirical Analysis of Major Stem CellPatent Cases: The Role of Universities,” Nature Biotechnology 28 (2010): 325–328.

72. Christopher M. Holman, “Trends in Human Gene Patent Litigation,” Science322, no. 5899 (2008): 198–199.

73. Allison, Lemley, and Schwartz, “Our Divided Patent System.”74. Christopher M. Holman, “Do Biotech Patent Lawsuits Really "Overwhelmingly

Lose?": A Response to Our Divided Patent System,” Biotechnology Law Report 34

(2015): 59–67.

34 introductory remarks

Investigating biotechnology-related cases and case-law in Europe forthe first time is a crucial task75. Biotechnology has been at the cen-ter of several debates relating to patent law76. Obtaining a better un-derstanding of its enforcement before the courts is of value. ‘ParallelLitigation: The Case of Biotechnology’ collects and analyzes novel datato provide for a comprehensive picture of the enforcement of bio-technology patents before the European courts, focusing on the issueof harmonization.

75. Geertrui Van Overwalle, “Policy Levers Tailoring Patent Law to Biotechnology:Comparing U.S. and European Approaches,” UC Irvine Law Review, 2011, 435–517 ar-gues that ‘[a]nalyzing biotechnology-related case law of the courts in the thirty-eightUSD contracting states seems like a daunting task, but one that might be extremelyinformative on the workings of European patent law in practice’.

76. Questions on the scope and legal standards, on industrial applicability andsufficiency of disclosure, or exhaustion in the context of self-reproducing materialall challenge the traditional patent law regime, see Denis Barbosa and Kuntz Grau,Exclusions from Patentable Subject Matter and Exceptions and Limitations to the Rights:Biotechnology, (World Intellectual Property Organization, 2010) and Gerd Winter, “Pa-tent Law Policy in Biotechnology,” Journal of Environmental Law 4, no. 2 (1992) on thesteps intellectual property had to take to enable the protection of biotechnology.

3P R O T E C T I O N H E T E R O G E N E I T Y I N AH A R M O N I Z E D S Y S T E M

*This Chapter was co-authored with Erasmus Elsner, ETH Zurich, Centerfor Law & Economics.

3.1 introduction

With the goal of fostering innovation in Europe, central and declaredpolicy has aimed to harmonize patent systems over the last fifty ye-ars. Through the harmonization of the laws governing patent rights,the development of a single and unified market for innovation andtechnology is pursued. European countries have moved towards ho-mogeneity by adjusting their national regimes under the roof of theStrasbourg Convention of 1963 and the European Patent Conventionin 1973 (EPC)1. Creating common standards across the internal mar-ket seeks to establish a level playing field2. A harmonized set of ru-

1. Convention on the Unification of Certain Points of Substantive Law on Patentsfor Invention, signed November 27, 1963 and Convention on the Grant of EuropeanPatents of 5 October 1973. For a historical account, see Joseph Yarsky, “HasteningHarmonization in European Union Patent Law Through a Preliminary Reference Po-wer,” Boston College International and Comparative Law Review 40 (2017): 167–193 andKate H. Murashige, “Harmonization of Patent Laws,” Houston Journal of InternationalLaw 16 (1994): 591–614 naming it ‘[a] giant step toward harmonization’.

2. See Alberto Alesina, Angeloni Ignazio, and Schuknecht Ludger, “What Doesthe European Union Do?,” Public Choice 123 (2005): 275–319 discussing the trade-offbetween the benefits of centralization (economies of scale, positive externalities) andits costs (heterogeneity of preferences in a large union) and more generally, see JohnF. Duffy, “Harmony and Diversity in Global Patent Law,” Berkeley Technology LawJournal 17 (2002): 685–726 and Craig A. Nard and John F. Duffy, “Rethinking PatentLaw’s Uniformity Principle,” Northwestern University Law Review 101, no. 4 (2007):1619–1676 for the benefits of harmonization. Henk Kox, Arjan Lejour, and RaymondMontizaan, “The Free Movement of Services Within the EU,” CPB Document, no. 69

(2004) estimate that bilateral trade of services within the European Union could beboosted by 30 to 60 % by a uniform regulatory system.

35

36 protection heterogeneity in europe

les allow for an autonomous system of patent protection, hinging onsubstantive patentability requirements and the establishment of Euro-pean patents. This effort towards harmonization, however, is under-mined if the practices of the national courts systematically diverge3.Anecdotal evidence such as the Epilady litigation has raised doubtupon the success of said harmonization. Claiming that the companyRemington had violated its patented device through a device with thesame function but different mechanism, Epilady launched infringe-ment proceedings in several European countries. Whereas the courtsin Belgium, Germany, Italy and the Netherlands ruled in favor ofthe infringement, judges in Austria, France and the United Kingdomcame to the opposite conclusion4. Disconcertingly, the courts had tointerpret the same patent, operate under very similar facts and uti-lized a common standard of interpretation; yet they came to polardecisions5. Even under harmonized laws, national divergent interpre-tations remain possible6.

The European Patent Office (EPO) offers a centralized examinationprocedure for the 38 current member states of the European PatentConvention. Once granted, a European patent however requires to bevalidated at national level and moves in a country-by-country enfor-cement regime. Beyond the costs of validation of patents7, the currentsystem is characterized by the legal complexity for litigating parties

3. Hugh Laddie, “Kirin Amgen - The End of Equivalents in England?,” Internatio-nal Review of Intellectual Property and Competition Law, 2009, 3–38 on how the benefitssecured by the creation of a central granting authority are undermined if the grantedpatents vary in their effects in the different Member States of the EPC.

4. A summary of the case can be found with Malwina Mejer and Bruno van Pot-telsberghe de la Potterie, “Economic Incongruities in the European Patent System,”European Journal of Law and Economics 34, no. 1 (2012): 215–234.

5. See John P. Hatter, “The Doctrine of Equivalents in Patent Litigation: An Analy-sis of the Epilady,” Indiana International & Comparative Law Review, 1995, p. 486 foran analysis of the lawsuits and the underlying doctrine of equivalents.

6. See Getchen Bender, “Clash of the Titans: The Territoriality of Patent Law vs.the European Union,” IDEA: The Journal of Law and Technology 40 (2000): 49–82 on theinherent conflict between national patent law and the EPC and Vincenzo Di Cataldo,“From the European Patent to a Community Patent,” Columbia Journal of EuropeanLaw 8 (2002): 19–36 characterizing the situation as ‘deadlock’.

7. The costs of validation for European patents are at least five times more ex-pensive than their US counterparts, see Mejer and van Pottelsberghe de la Potterie,“Economic Incongruities in the European Patent System.”

3.1 introduction 37

to enforce their rights at the country level8. In the event of a dispute,the enforcement of patents will be subject to a highly variable institu-tional set of enforcement rules, whose harmonization is far from com-plete. At the institutional court level, the dual German enforcementsystem – so-called bifurcation – dissociates infringement and validityclaims which are heard by separate courts; in other countries bothissues are tried by the same court. The cost structure for enforcinga patent in the UK is substantially higher compared to other coun-tries9. Resultant of the difference in experience of judgment-makingin patent cases, quality and speed of the courts diverge. A range ofprocedural provisions such as provisional measures, language of pro-ceedings or rulings on evidence and the onus of proof characterizethe individual states10. In terms of damages, the EU Enforcement Di-rective has harmonized the calculation methods, but variances in theamount of the level of damages across countries exists11.

Important differences between European countries in terms of out-comes, share of appealed cases and characteristics of litigants andlitigated patents have been revealed in the last years. The first large-scale empirical study of patent litigation was led by Graham and VanZeebroeck (2014). Relying on a database of nearly 9,000 patent suitsin the seven largest countries in the European Union, the authors sho-wed that the incidence and base of judicial outcomes widely variedby country and by patented technology. A second empirical project

8. Joseph Straus, “Reversal of the Burden of Proof, the Principle of “Fair andEquitable Procedures” and Preliminary Injunctions under the TRIPS Agreement,”The Journal of World Intellectual Property 3, no. 6 (2005): 807–822.

9. With costs in the UK in an aggregate of between £1 million and £6 million, seeChristian Helmers and Luke McDonagh, “Patent Litigation in the U.K.: An EmpiricalSurvey 2000–2008,” Journal of Intellectual Property & Practice 8 (2013): 846–861. Thesefigures are to be compared to costs of EUR 50,000 to 200,000 in France and of EUR25,000 to EUR 91,456 for the court fees and of EUR 40,000 to EUR 100,000 for attorneyfees in Germany, see Katrin Cremers et al., “Patent Litigation in Europe,” EuropeanJournal of Law and Economics 44, no. 1 (2017): 1–44.

10. See the detailed analysis by Stefan Luginbuehl, European Patent Law - Towards aUniform Interpretation (Edward Elgar, 2011), p. 64.

11. See Stuart J. Graham and Nicolas Van Zeebroeck, “Comparing Patent LitigationAcross Europe: A First Look,” Stanford Technology Law Review 17 (2014): p. 667 esti-mating the awarded damages in Belgium, France and Italy as low, in Germany andthe Netherlands as average, and in the UK as high; see EU Directive 2004/48/ECand it’s Article 13 paragraph 1 that obliged France to award the infringer’s profits asa damage (Luginbuehl, European Patent Law - Towards a Uniform Interpretation, p. 70).

38 protection heterogeneity in europe

by Cremers et al. (2017) furnished an exhaustive picture of the litiga-tion landscape in Germany, France, the Netherlands and the UnitedKingdom by evaluating a dataset of hand-coded suits filed between2000 and 2008. They found substantial differences across countries interms of case loads and outcomes, and characteristics of both litigantsand patents. In addition, several recent studies focused on patent en-forcement in one particular jurisdiction, presenting evidence of a Eu-ropean landscape characterized by fragmentation and uncertainty forinnovators, patent holders and their competitors12.

In examining the behavior of courts of different jurisdictions, casecomparison is complicated by the differentiating subject-matter. Fo-rum shopping enables entities to conduct a gamification of the systemby acknowledging the odds of winning and competence of differentcourts. As such, simply contrasting win rates between two jurisdicti-ons might not be sufficient. Natures of the cases tried by courts differ.Under a harmonized system, systematic differences in the likelihoodof patents to be enforced with success should correspond to thosecharacteristics. Put otherwise, patents with similar characteristics andsimilar strength should, in the limit, reach similar outcomes. By con-trast, if the differences rely solely on the country of jurisdiction – andnot on the characteristics of the underlying invention – the harmo-nization policy goal is not being achieved14. This would imply thattraditions existing before the entry into force of the European Pa-tent Convention have survived to some extent or that current policies

12. For the UK, Helmers and McDonagh, “Patent Litigation in the U.K.: An Empiri-cal Survey 2000–2008” built a dataset containing the complete set of patent litigationfiled at the courts in England and Wales from 2000 to 2008, showing the distributionby technology and by sector of the litigating companies13. In Germany, Katrin Cre-mers, “Determinants of Patent Litigation in Germany,” ZEW Discussion Paper, nos.04-72 (2004) documented the determinants of patent infringement suits from 1993 to1995 before the three primary infringement courts. Last, in France, Pierre Véron, “LeContentieux des Brevets d’Invention en France, Étude Statistique 2000-2010,” 2010,evaluated the decisions of the Tribunal de Grande Instance de Paris and subsequentappeals between 2000 and 2009.

14. Hatter, “The Doctrine of Equivalents in Patent Litigation: An Analysis of theEpilady” and Matthew Parker, “Giving Teeth to European Patent Reform, Overco-ming Recent Legal Challenges,” Emory International Law Review 26 (2012): 1079–1110

arguing that ‘[i]n actuality, the national courts of Europe have construed patentsdifferently despite applying the so-called uniform requirements of the EPC’.

3.1 introduction 39

diverge15. To study the extent of harmonization in Europe, this pa-per investigates the characteristics of patent disputes brought beforecourts.

We propose a model where both litigation and settlement are drivenby patent quality. We assume that patent quality depends on bothbroadness and definiteness of the patent, whereby broadness relatesto the scope of protection as defined by the patent’s claims and de-finiteness relates to the precision of the claims. In our model, patentholders and patent infringers decide before proceeding to court whet-her to settle or litigate based on differences in the perception of thepatent’s quality, i.e., based on relative patent quality assessments. Weassume that technology-specific differences in patent definiteness ex-ist, leading to different settlement and litigation rates across techno-logies. Each agent makes his or her own estimation of the patent qua-lity, whereby the patent’s definiteness is the factor which introducesheterogeneity in their estimation. If there is sufficient agreement, asettlement will be reached, otherwise the parties will proceed to trial.At trial stage, the divergent expectations of the parties are no lon-ger material, it is rather the assessment of the absolute patent qualityby the judge, i.e., the composite of both patent broadness and defi-niteness, which decides the outcome of the case. Notably, unlike thePriest and Klein (1984) model, we do not assume that either plaintiffor defendant is forming their respective expectation with respect toa common judicial standard of fault. This can be rationalized, as ourinstitutional setting is the existing system of heterogeneous Europeanpatent courts, whereby patents can be litigated in multiple courts andheld to different judicial standards.

The model predicts that technologies where the patent definitenessattribute can be estimated with high accuracy will have higher settle-ment rates. Furthermore, according to the model, the absolute patentquality will be determinative of case outcome only at the litigationstage. Therefore, in contrast to Priest and Klein (1984), the model pre-dicts that patent litigation cases where high quality patents are under

15. For infringement disputes, there was initially a divergence between countriessuch as Switzerland and the UK applying ‘peripheral claiming’, with claims definingboth the invention and limiting it, whereas in Germany and the Netherlands thecourts relied on ‘central claiming’, which states that the claims define the inventiononly, being the starting point of the scope of protection.

40 protection heterogeneity in europe

dispute will have a higher probability of success and, thus, we expectto find higher win rates in our empirical dataset.

The paper utilizes empirical methods to investigate whether the cha-racteristics of the patents or the country of litigation predict the out-come of litigation. Examination of the patent characteristics is con-ducted in line with our model, in that the patent quality, and un-derlying technology and industry are tested. This paper unfolds asfollows: the theory and the hypotheses are introduced first, followedby the description of the data and coding strategy. The results of thestudy are presented; discussion and conclusion conclude.

3.2 theory

3.2.1 Litigated Patents: the Tip of the Iceberg

Most patents have little monetary value16 and, therefore, expire unin-fringed17. Patent litigation data can always only be a subsample ofboth patent disputes and the entire patent universe. It is, therefore,subject to a strong selection bias. Early law and economics scholars,in particular Priest and Klein (1984) with their seminal paper, havesuggested that the cases which are litigated are the hardest ones, na-mely those with a 50 % chance of winning and losing. The Priest-Klein hypothesis has since been extended and/or rejected by nume-rous authors18. As a result of this strong selection bias, it is suggestedthat no inferences can be made about legal standards from plaintiff

16. Kimberly A. Moore, “Worthless Patents,” Berkeley Technology Law Journal 20

(2005) and Mark A. Lemley, “Rational Ignorance at the Patent Office,” NorthwesternUniversity Law Review 95 (2001): 1495 both finding that the majority of patents lapsebecause of a failure of the patent holders to pay the renewal fee, indicating that thereis no economic value to the protection.

17. See Lemley, “Rational Ignorance at the Patent Office” estimating that of abouttwo million US patents in force, only about 2,000 different patents are subject tocourt disputes and, of that, about 100 cases per year make it to trial.

18. Kevin M. Clermont, “Litigation Realities Redux,” Notre Dame Law Review 84

(2009): 1951–1956, Kevin M. Clermont and Theodore Eisenberg, “Litigation Reali-ties,” Cornell Law Review 119 (2002): 119–154 and Daniel Kessler, Thomas Meites,and Geoffrey P. Miller, “Explaining Deviations from the Fifty-Percent Rule: A Multi-modal Approach to the Selection of Cases for Litigation,” The Journal of Legal Studies25, no. 1 (1996): 233–59.

3.2 theory 41

trial win rates. We introduce a divergent expectation model for patentinfringement disputes which uses building blocks of the Priest-Kleintheorem and integrates these into a subjective expected utility model.We build our model in three steps: Firstly, we model the trade-off ofthe patent holder between settlement and trial in an expected utilitysetting; secondly we introduce a divergent expectations framework;and lastly, we dive deeper into the dimensions of patent quality. Un-der our model, and in contrast to Priest-Klein, the cases at trial do notnecessarily have a 50 % chance of winning, but because of divergentexpectations of the parties, a population of patent cases of differentquality will end before the courts, and the latter quality will be deter-minative of the final outcome by the judge.

3.2.2 Microeconomic Model of Patent Disputes

Patent holders whose patents are infringed have a discrete choicebetween settling or litigating such claims. Rubinfeld and Scotchmer(1993) have modeled success and defeat in litigation as two mutuallyexclusive states of the world in an expected utility setting. Buildingon this framework, we represent the patent holder’s choice set in pa-tent infringement cases as a function of patent broadness and patentdefiniteness – with the technology-specific factor determining the re-lative weights. We define broadness as the scope of coverage of a pa-tent and definiteness as the precision of the claims specifications. Thedefiniteness defines the ease with which protection in a technologyfield can be substantiated (e.g., through the proof of exact chemicalcompositions for drug patents vs. the often rather vague descriptionof functional properties for mechanical patents).

The starting point of our model is the occurrence of a patent infrin-gement dispute. At the principal node, the patent holder can decidewhether to settle or litigate his respective patent claim. Through asettlement, he can avert the ambiguity of a judicial decision, sincehe obtains certainty over the payoff at settlement – not leaving it toan exogenous decision body. As such, the settlement option appearsdeterministic rather than probabilistic. However, this is not how wemodel this decision point. Rather, we assume that the decision agentdeploys backward induction across the entire decision tree depicted

42 protection heterogeneity in europe

in Figure 3.1. This means that in our model the uncertainty of trialis carried forward all the way to the initial occurrence of the patentdispute. Therefore, the settlement decision becomes a function of theuncertain trial outcome as set out below.

PatentDispute

Litigation

Patent Claim Wins

pL(θ)

Patent Claim Loses1− pL

(θ)pS (θ)

Settlement

1−pS

(θ)

Figure 3.1: Litigation vs. Settlement Decision Tree

We first assume that the patent holder will litigate with probabilitypS(θ) and settle with probability 1− pS(θ). If he decides to litigate,he enters the ‘litigation lottery’, where he can either win or lose. Wefurther assume that each patent is endowed with a patent qualityθ, which determines both the success probability in a patent infrin-gement trial, denoted as pL (θ), and the probability of an adverseoutcome, i.e., the probability that the claim is not upheld at trial, de-noted by 1 − pL (θ). Since the patent holder makes his decision bybackward induction, we assume that the success probability at trial isdeterminative of the patent holder’s initial settlement choice19.

We can, therefore, model the trade-off between settlement and trialusing a von Neumann-Morgenstern expected utility framework. Thepatent holder’s utility function is, thus, given by:

Uθ(S,L) = (1− pS(θ))u(S− cS) + pS(θ)u(L− cL) (1)

19. A microfoundation concerning these assumptions can be derived inspired byGeorge L. Priest and Benjamin Klein, “The Selection of Disputes for Litigation,” TheJournal of Legal Studies 13, no. 1 (1984): 1–56, but this goes beyond the scope of thispaper.

3.2 theory 43

Figure 3.2 depicts the claimant’s indifference curve: the vertical axis,labeled S, represents the patent holder’s expected payoff upon settle-ment, whereas the horizontal axis, labeled L, represents his expectedpayoff upon litigation, with the latter being the weighted sum of thewin/lose state payoff. The costs associated with negotiating the sett-lement are denoted as cS, while the costs of going to trial are deno-ted as cL. The settlement and litigation payoff, S and L, respectively,must exceed cS and cL, respectively, for there to be some utility forthe agent u(·) > 0. We further assume that the costs of going to trialexceed those of settlement, cL > cS. The indifference curves In furt-her to the northeast represent greater expected utility and are, thus,preferred by the patent holder, i.e., u(I3) � u(I2) � u(I1).

I1

I2

L

S

Figure 3.2: Litigation vs. Settlement Indifference Curves

The slope of the patent holder’s indifference curve reflects the patentquality and, thus, the probability of winning a potential infringementsuit pL (θ). The slope is the marginal rate of substitution:

MRS = −pS(θ)

(1− pS(θ))

u ′(S)

u ′(L)(2)

Therefore, the lower the patent quality, θ, the flatter the patent hol-der’s indifference curve and the larger the patent holder’s relativepreference for the settlement option, since for a given settlement amount,the holder of the lower quality patent requires more litigation payoffunits to render him indifferent.

44 protection heterogeneity in europe

3.2.2.1 Divergent Expectations

Both plaintiff and defendant hold different subjective expectations ofthe success probability at trial pL (θ). In a simplified discrete setup,patents can be classified by agents as either ‘high’ quality patents, H,or ‘low’ quality patents, L, so that θ ∈ {L,H} at this stage. If no sett-lement is agreed, the patent quality will be assessed by the judge,resulting in the assignment of a ‘judicial’ patent quality, J, whichthen determines the success probability at trial and defines the agentsexpected utility function at trial. Negotiation occurs in a bargainingzone, denoted by Z, determined by (i) the patent holder’s estimationof pL (θ)H, (ii) the patent infringer’s estimation of pL (θ)I

20 and (iii) arange of expected litigation outcomes between Lmin and Lmax21.

Ih

Ij

Il

Lmin Lmax

L

S

Figure 3.3: Patent Infringement Bargaining Area

Figure 3.3 illustrates a scenario where the patent holder perceives thepatent to be of low quality and, therefore, as having relatively lowsuccess probability in adjudication, resulting in a relatively flat indif-ference curve Il. On the other hand, the patent infringer’s assessment

20. This can be plotted as a second, quasi indifference curve of the patent holder.It can also be thought of as the patent infringer’s subjective assessment of what thepatent holder’s ‘fair’ indifference curve should look like given his estimation of thesuccess probability.

21. Whereby Lmin = min(Lpatent holder,Lpatent infringer) and Lmax =

max(Lpatent holder,Lpatent infringer)

3.2 theory 45

of the patent is that it is a patent of high quality, leading to a steepsettlement curve Ih. While only formed at the trial stage, Ij denotesthe indifference curve using the judicial success probability. In this set-up, it would be pareto-efficient for the patent holder and the patentinfringer to reach a settlement at all litigation payoff levels, Ln. This isbecause a bargaining zone exists in the north-east and the south-westof the patent holder and the patent infringer, respectively, i.e., in theutility-enhancing zone (represented by the gray area in Figure 3.3).

3.2.2.2 Patent Quality Parameters

Shifting from a discrete setup to a continuous setup, where patentquality can not only take two discrete states to θ ∈ {N}, we proposethat patent quality should be modeled as a function of the patent’sbroadness, denoted as B, and the patent’s definiteness, denoted as D,with the technology-specific factor, denoted as αj, determining therelative weights:

θ(B,D) = B(1−αj)Dαj

(3)

Figure 3.4 represents θ(B,D) = B(1−αj)

Dαj, i.e., the patent value as a

function of the patent’s definiteness, D, with the patent’s broadnessheld constant at different initial endowment levels. The technology-specific factor, αj, determines whether patent broadness or definite-ness has more influence on patent quality and adjudication successprobability.

θ = 1.8(1−0.1)D0.1

θ = 1.5(1−0.2)D0.2

D

θ

Figure 3.4: Patent Quality Function

46 protection heterogeneity in europe

To illustrate this by analogy, patent disputes can be compared to boun-dary disputes in land law – with the land owner and trespasser beingthe equivalent to the patent holder and infringer, respectively. Patentbroadness can in this analogy be thought of as the size of the landprotected, while patent definiteness is the equivalent of the land’sfencing or trespassing detection system. The larger the size of theland, the higher the likelihood of trespassing and the easier it is forthe land owner to prove in court that trespassing has occurred, thesmaller the land the more important his fencing/detection systemsand the evidential threshold established by courts become.

For the purpose of our model, we further assume that patent broad-ness and the technology factor are observable exogenous variablesand that differences in the parties’ estimation of patent quality θ aresolely due to patent definiteness. We assume that the accuracy ofestimating the patent’s true level of legal protectability, i.e., it’s defini-teness is technology-specific and that the standard error of estimateof the patent’s definiteness can be denoted as:

σjest =

√√√√∑Njk=1 (D

jk −D

j)2

Nj(4)

These differences in the standard errors of estimates across techno-logies lead to different respective settlement rates. Put formally, iftwo technologies, A and B, exist with agents having more trouble toascertain the boundaries of a patent in technology A (due to lowerprediction accuracy of the patent definiteness attribute), then the ini-tial bargaining zone is larger for this technology and the likelihoodof reaching a settlement decreases.

σAest > σBest ⇒ ZA > ZB (5)

Thus, we summarize our propositions as follows.

Proposition 1. The patent value is a function of the patent’s defini-teness and broadness with the technology-specific factor as a scalingfactor.

3.2 theory 47

Proposition 2. Only patent disputes where it is difficult for the par-ties to reach agreement over patent quality (and, by extension, on thesuccess chances at trial) defined with regard to patent definitenesswill proceed to trial. Where the parties can agree on the merits ofthe case, irrespective of whether this entails a high or a low absolutesuccess probability, settlement will be preferred. Thus, it is only re-lative success probability that determines the settlement vs. litigatedecision, i.e., the estimation of the success probability of the patentholder versus that of the patent infringer.

Proposition 3. If no settlement is reached, the divergent expectationsof the success probability are no longer determinative of the finaloutcome. Rather, the key factor is the judge’s assessment of patentquality and, thereby, of the adjudication success probability p(θ). Itis, thus, the ‘absolute’ value of pL(θ) that matters at the adjudicationstage.

Since the selection of cases proceeding to adjudication is made onrelative patent quality, this does not tell us anything about absolutecase quality and success likelihood at trial. Since patent quality, asdetermined by the judge, depends on patent broadness, definitenessand the technology factor, the litigation sample does not allow us tomake inferences as to the population of contiguous cases. In otherwords, heterogeneity in case outcomes among technologies would bein line with our model’s predictions, as success probability at trialdepends on a host of factors in patent disputes.

3.2.2.3 Relation with the Priest-Klein Hypothesis

In relating our model to Priest and Klein (1984), we rely on the Priest-Klein model formalization of Klerman and Yoon-Ho (2014)22. Likeour model, Priest and Klein (1984) is a divergent expectation model,where the litigation condition depends on the relative estimate ofthe plaintiff’s versus the defendant’s success probability. In the Priestand Klein (1984) model, divergent expectations result from an ‘error’which can result from a number of factors, such as errors in the inter-pretation of the law, errors in interpreting the respective case factual

22. Daniel Klerman and Alex Lee Yoon-Ho, “The Priest-Klein Hypotheses: Proofs,Generality and Extensions,” Legal Studies Research Paper Series, nos. 14-43 (2014).

48 protection heterogeneity in europe

merits, or incomplete information concerning either. The equivalentto the errors term under Priest and Klein (1984), resulting from noisysignals, in our model is the patent definiteness factor, which we con-sider to be the noise factor that captures residual uncertainty for bothparties.

Similar to our model with our assumptions, settlement in the Priestand Klein (1984) setting will occur where the plaintiff’s estimate ofthe success probability is below that of the defendant. However, un-like the Priest and Klein (1984) model, which presumes that both par-ties of a litigation base their estimates around a common judicial faultstandard, our model assumes that parties make case assessments withregards to patent quality, without taking into account a common judi-cial standard. Only in the second stage, when the parties have alreadydecided to proceed to court, does the judicial standard come into play,through the judge’s own assessment of patent quality. The reasoningbehind this is that in the specific institutional setting which formsthe subject of our theoretical and empirical analysis, namely the Eu-ropean patent court system, there is no common judicial standard, asparties can litigate before multiple courts.

Under Priest and Klein (1984), in the limit, where the standard devi-ation of the estimation error term approach zero, no cases will go totrial as the parties will have certainty over litigation outcome and willalways settle. In turn, if the standard deviation of the estimation errorterm rises there will be fewer settlements and more cases proceedingto trial. This is in line with our model of the parties estimate of patentquality, which diverge depending on the definiteness factor: the lessdefinite the patents get, the more litigation is to expected.

Under Priest and Klein (1984), the hypothesis states that in the limitproportion of victories will approach 50 percent exactly as the estima-tion error diminishes and the litigation rates decline. In contrast tothis, we only introduce the judicial standard in a second step, namelywhen both parties have already decided to proceed to trial. If we as-sume that the judge’s error term follows the same distribution as thatof the parties, a similar result as under Priest and Klein (1984) is to beexpected. However, if one assumes, that the judge follows a fully in-dependent distribution from that of plaintiff and claimant, the Priestand Klein (1984) hypothesis does not hold true and we are to expect

3.3 hypotheses 49

a divergence in outcomes. Given the heterogeneity in legal standardsacross the European patent system, we assume that the latter is thecase. With the advent of the Unified Patent Court, however, we canexpect a gradual shift towards a more common distribution.

3.3 hypotheses

We formulate the following hypotheses flowing from our microeco-nomic model. We predict that the outcome of court decisions overpatents is predicted by the quality and, therefore, of the technologyand industry of the patent at stake.

Hypothesis 1 The quality of the litigated patent is directional for the out-come of patent litigation.

Explanation. Hypothesis 1 follows directly from Proposition 3. In theabsence of settlement, the patent quality will be determined by thejudge, resulting in the assignment of a ‘judicial’ patent quality whichdetermines the success probability at trial. The higher the quality ofa patent, the more likely it is to win. In a completely harmonizedpatent system, the ‘judicial standard’ would be fully uniform acrossjurisdictions and judges. This would mean that given a specific patentquality level, the judgment on the case’s merits should be the same.As we are trying to identify country-level differences in judicial beha-vior, we formulate the second hypothesis as follows.

Hypothesis 2 The technology and industry of the litigated patent is directi-onal for the outcome of patent litigation.

Explanation. Our model predicts that the success probability at trialdepends on patent quality (see Proposition 3), a function of the pa-tent’s definiteness and broadness with the technology-specific factordetermining the relative weights (see Proposition 1), which imply va-riations in patent quality across technologies and industries. It shouldbe noted that technology is already endogenous to case quality in ourmicroeconomic model. In our empirical identification strategy, howe-ver, we have chosen to use technology and industry as separate fac-tors, instead of treating them endogenous to patent quality.

50 protection heterogeneity in europe

3.4 data

We have compiled a novel dataset comprising 934 individual patent li-tigation infringement and counterclaim cases rendered between 2008

and 2012 across Germany, France and the United Kingdom. The 934

cases represented a total of 1,407 individual court decisions. We expli-citly excluded revocation decisions and infringement counterclaims,or non-infringement declarations.

3.4.1 Germany

Because Germany operates a bifurcated system, i.e., patent infringe-ment and patent validity cases are dealt with by different courts, wecollected the patent infringement and nullity data separately. The 569

infringement cases were aggregated from Düsseldorf, Mannheim andMunich, the three largest patent litigation courts in Germany23. ForDüsseldorf, we used the official online North Rhine-Westphalia casedatabase24 by filtering all cases with the keyword ‘patent’ appearingin the judgment of the Landgericht. For Mannheim and Munich, wemade use of Darts-IP to collect decision from the respective Landge-richt. The existence of decisions at the appeal level (Oberlandesgerichtand Bundesgerichtshof) was then verified. In total, we collected 471

infringement cases from Düsseldorf (and 164 related appellate decisi-ons), 75 from Mannheim (and 16 related appellate decisions) and 23

from Munich (and 7 related appellate decisions). Thereafter, we ex-tracted the patent codes from the database of the German Patent andTrade Mark Office and from Darts-IP. Nullity actions are handled bythe German Federal Patent Court (Bundespatentgericht) in Munichwith appeal to the Federal Court of Justice of Germany (Bundesge-richtshof). All revocation lawsuits with a decision falling within 2.5years after an infringement decision were counted as counterclaims.In fact we can confirm for 99% of all patents that is was the poten-tial infringer that challenged the validity of the (arguably) infringed

23. In fact, the regional court in Düsseldorf hears the largest number of cases inEurope, according to data by Cremers et al., “Patent Litigation in Europe.”

24. Available under http://www.justiz.nrw.de.

3.4 data 51

patent25. Basing our search on the patent codes, we have gathered atotal of 142 invalidity cases from the official Federal Patent Court and23 from the Federal Court of Justice online databases26 and comparedour results with the nullification actions listed in the Patent Gazette(Patentblatt)27.

3.4.2 France

In the absence of an official register for patent suits in France, we havebuilt our dataset of 304 cases resp. 401 decisions from the IP data plat-form Darts-IP, the most exhaustive database in the field28. This is un-derlined by the fact that, apart from the French Patent Office, Véron& Associés is the main supplier of IP case data to the platform29.Véron & Associés has aggregated all decisions rendered by the Tribu-nal de Grande Instance (TGI) de Paris, the Cour d’Appel de Paris andthe Cour de Cassation from the 1 January 2000 – therefore, coveringall three appellate levels. Notably, the TGI provides the richest datasource for French case data, since, even prior to the centralization ofpatent litigation in 2009 and the exclusive first instance jurisdiction ofthe TGI, the Parisian Court was already the most prominent patentcourt in France – hearing more than 50 % of all cases30.

25. Inversely, if a revocation decision occured within the same timespan beforethe infringement decision, the case was excluded and considered as revocation withinfringement counterclaim. The latter cases are rare, Anne Hees and Sven-Erik Brait-mayer, Verfahrensrecht in Patentsachen (Carl Heymanns Verlag, 2010) and Alfred Keu-kenschrijver, Patentnichtigkeitsverfahren (Heymann, 2011), p. 73-74 estimate that over90% of all revocation actions are filed in response to an infringement action.

26. Available under https://www.bundespatentgericht.de.27. Available under https://register.dpma.de/DPMAregister/Uebersicht. We es-

timate that our dataset covers 50% to 60% of all German patent infringement andcounterclaim decisions rendered between 2008 and 2012, see Graham and Van Zee-broeck, “Comparing Patent Litigation Across Europe: A First Look” with indicationthat Darts-IP estimates its German coverage to 50% for the years 2000-2009.

28. Cremers et al., “Patent Litigation in Europe.”29. Cremers et al., “Patent Litigation in Europe.”30. Towards an Enhanced Patent Litigation System and a Community Patent – How to

Take Discussions Further, (The European Council, 2007). We estimate that our data-set covers between 75% to 90% of all French patent infringement and counterclaimdecisions rendered between 2008 and 2012, see Graham and Van Zeebroeck, “Com-paring Patent Litigation Across Europe: A First Look” with indication that Darts-IPestimates its French coverage to 90% for the years 2000-2009.

52 protection heterogeneity in europe

3.4.3 United Kingdom

The overwhelming majority of patent suits are heard in England andWales31, with a shared jurisdiction of the Patents High Court (PHC),part of the High Court of England and Wales, and the IntellectualProperty Enterprise Court (IPEC; formerly the Patent County Courts).While the Patent County Courts historically dealt with smaller claimsof less complex variety, with a market share of less than 10%32, theIPEC has become, after some restructuring in the court system, aneffective forum for IP disputes in England and Wales33. Our datasetis based on the PHC and the IPEC Diary, basically listing all casesscheduled for a hearing or an application.34 Thus, starting from theDiaries, we were able to collect 61 infringement cases (and 24 ap-pellate decisions) from the website of the British and Irish Legal In-formation Institute35, Thomson Reuter’s Westlaw36 database and onDarts-IP37.

31. Hence, we excluded the by far less important litigation in Scotland and Nort-hern Ireland.

32. See Helmers and McDonagh, “Patent Litigation in the U.K.: An Empirical Sur-vey 2000–2008” and Graham and Van Zeebroeck, “Comparing Patent LitigationAcross Europe: A First Look.”

33. Our evaluation indicates that, based on the 69 available court decisions listed onthe IPEC Diary as per July 2015, the main share of rulings are issued in patent relatedcases (32.4 %) dominated by infringement trials (19.1 %), followed by copyright (20.6%), trademark and design (19.1 % each), goodwill (2.9 %) and unavailable litigation(5.8 %).

34. Available for the PHC under https://www.justice.gov.uk/courts/court-li

sts/list-patents-court-diary and for the IPEC under https://www.justice.gov.uk/courts/court-lists/intellectual-property-enterprise-court-diary.

35. Available under https://www.bailii.org.36. Available under https://www.westlaw.co.uk.37. While the Diaries intend to be as accurate as possible, they do not furnish an

exhaustive overview of UK patent litigation. Some settled cases are not listed andparties may, in some cases, request to not be listed (information derived from phonecalls with the clerks in charge of keeping the IPEC and PHC Diaries). We estimatethat our dataset covers between 60% to 80% of all UK patent infringement and coun-terclaim decisions rendered between 2008 and 2012, see Graham and Van Zeebroeck,“Comparing Patent Litigation Across Europe: A First Look” with indication thatDarts-IP estimates its UK coverage to 60% to 80% for the years 2000-2009.

3.5 data coding 53

3.5 data coding

3.5.1 Decision Coding

We hand-coded all decisions, categorizing them across numerous di-mensions, mainly by technology and industry, but also by the levelof jurisdiction (first instance, intermediate appeal level and supremecourt) and the nature of the ruling rendered (infringement vs. inva-lidity). The ‘patent case’ was our unit of analysis: each outcome wascoded separately for each patent, even when they were assessed inthe same verdict. A ‘win’ was reported if the patent holder could en-force its infringement claim before the courts, i.e., at least one of theclaims was found to be infringed and that claim was, if challenged,upheld as valid.

3.5.2 Patent Quality

Empirically speaking, patent quality is not a directly observable vari-able. Innovation economics scholars have been relying on inferencesfrom patent metrics or survey methods38. For instance, it has beenshown that the private economic value of a patent, as estimated by in-ventors, correlated with the number of citations they yielded39. Ourmicroeconomic model suggests that patent quality depends on thebroadness of the patent claim and definiteness of the patent, with atechnology- or industry-specific scaling factor determining the rela-tive weights. For our empirical identification strategy, we have mat-ched both patent broadness and patent definiteness to empirical fac-tors commonly used in the literature. We relied on an econometricfactor model that is based on the work Lanjouw and Schankerman

38. E.g., Alfonso Gambardella, Dietmar Harhoff, and Bart Verspagen, “The Valueof European Patents,” European Management Review 5 (2008): 69–84.

39. Dietmar Harhoff, Francis M. Scherer, and Katrin Vogel, “Citation Frequencyand the Value of Patented Innovation,” Review of Economics and Statistics 81 (1999):511–515.

54 protection heterogeneity in europe

(2004)40. In particular, we develop a composite index for patent qua-lity using multiple characteristics of patents.

To arrive at the composite index, we formulate a factor model withseparate broadness indicators of a patent’s underlying, unobservabledefiniteness: the number of claims and the number of distinct techno-logical fields the invention is allocated to. The data was extractedfrom espacenet41; and the number of technology fields representedby the number of 4-digit International Patent Classification subclas-ses42. We use a multiple-indicator model with patent definiteness atthe latent common factor:

yki = λkdi +βXi + εki (6)

where yki is the observation of the k’th patent indicator for the i’thpatent, di is the patent’s definiteness with factor loading λk and Xiare the patent broadness variables. The variance of d is normalizedby setting its variance to one d ∼ N(0, 1). Any uncommon variationwhich is not related to the broadness indicators is captured by anidiosyncratic error εki, which is assumed to be independently drawnfrom a N = 0,σ2k. The common definiteness factor di is simply theunobserved characteristic of a patented innovation that influences alltwo of the broadness indicators Xi: the number of claims and thenumber of technology classes. Thinking about our land law compari-son above, where patent broadness is represented by the size of theland and patent definiteness by the fencing, we believe that the dis-

40. Jean O. Lanjouw and Mark Schankerman, “Patent Quality and Research Pro-ductivity: Measuring Innovation with Multiple Indicators,” The Economic Journal 112

(2004): 441–465; see also Bronwyn H. Hall, Grid Thoma, and Salvatore Torrisi, “TheMarket Value of Patents and R&D: Evidence from European Firms,” NBER WorkingPaper 13426 (2007), and Béatrice Dumont, “Does Patent Quality Drive Damages inPatent Lawsuits? Lessons from the French Judicial System,” Review of Law & Econo-mics 11, no. 2 (2015): 355–383 for a recent use of the factor model to capture patentquality and link it to the damages in French patent lawsuits and Bodo Knoll, MartinaBaumann, and Nadine Riedel, “The Global Effects of R&D Tax Incentives: Evidencefrom Micro-Data,” Beiträge zur Jahrestagung des Vereins für Socialpolitik 2014: Evidenz-basierte Wirtschaftspolitik - Session: Taxation II, 2014, for a quality-adjusted count ofpatent applications. In addition, see OECD, “A Framework for Biotechnology Statis-tics,” 2005, for a graphic view of the evolution of patent quality proxied by twocomposite indexes.

41. www.espacenet.com.42. OECD, “A Framework for Biotechnology Statistics.”

3.5 data coding 55

tinction makes intuitively sense. Since the patentable land of techno-logical innovation is an ever-expanding land, patent definiteness isneeded to fence off the different parts (patent claims) and parcels ofthe land (number of technologies).

To arrive at the Patent Quality Index we performed a principle compo-nents factor analysis, using varimax rotations. We then run a Horn’sparallel analysis, a method to determine the number of factors to re-tain from the factor analysis. We kept those factors whose average ei-genvalue were greater than the average eigenvalues (and higher than1, according to the Kaiser criterion). The two factors explained for56.6% of the variance. The factor loading matrix for this final solutionis presented as follows:

Table 3.1: Factor Loadings

Variable Factor Loading

Number of IPC 0.75

Number of Claims 0.75

It should be noted that, following both our microeconomic modeland Lanjouw and Schankerman (2004), the next logical step wouldhave been to re-calculate different weights for the specific technolo-gies and industries. However, we have decided on using the aboveuniform factors, which effectively treat technology and industry asexogenous factors with one common weighting, for all quality speci-fications. Instead, we decided to use both technology and industry asseparate factors, as this allowed us to identify heterogeneity on thatdimension more clearly in the main regression results.

3.5.3 Technology Coding

Due to the limitations of the patent classification (such as the Inter-national Patent Classification) in capturing patents at a conceptuallevel43, measuring the patent breadth requires hand-coding by techno-

43. The classifications are inadequate for delimiting technologies since they weredesigned with the purpose of identifying the function of the patent to facilitate prior

56 protection heterogeneity in europe

logy field44. We followed the classification developed by Allison, Lem-ley, and Schwartz (2015) that distinguishes between mechanical, elec-tronics, chemistry, biotechnology, software and optics patents (see Ap-pendix A for further details). Each patent was assigned to one singletechnology area.

3.5.4 Industry Coding

By determining the industries, we aim at assessing a further dimen-sion of patent litigation. The industry classification allows us to iso-late technologies that are used in several industries, or to aggregateindustries relying on inventions in several technologies. Accordingly,we also followed the industry classification by Allison, Lemley, andSchwartz (2015) and assigned each patent to one of the followingindustries: computer and other electronics, semiconductor, pharma-ceuticals, medical devices, biotechnology, communications, transpor-tation, construction, energy, goods and services for industry use, andgoods and services for business use (see Appendix B).

3.6 descriptive results

The readership recall that only infringement disputes are recordedin our sample. That is, only lawsuits brought by the patent holderagainst one or several alleged infringers are examined. To countersuch claims, infringers may either or both raise non-infringementdefenses or validity arguments. Therefore, even when the patent isfound to be infringed by the courts, the patent holder may overalllose at trial when his patent is declared invalid.

art searches, see John R. Allison et al., “Valuable Patents,” Georgetown Law Journal 92

(2004): p. 28-29.44. In the words of Allison et al., “Valuable Patents,” p. 28-29: ‘if economists want

to measure patent breadth, they will have to hand-code the patents by technologyarea or at least find a better measure than the ones that exist today’.

3.6 descriptive results 57

The picture painted by our data is complex. Patent litigation win ra-tes are highest in Germany, with an average of about 53%, followedby France, where an average of 37% of the claims have been success-ful. When patent holders brought infringement suits to the UK, theywere successful in less than one of four cases (see Table 3.2). When in-creasing granularity, we find that the variation between infringementrates across countries, varying from about 56% to 64%, was smallerthan in invalidity rates. When patent validity was examined, the va-riance increased from 37% to 54%. The low likelihood of success invalidity trials in the UK must be underlined.

Table 3.2: Win Rates by Country

Overall Win Rate by Country

Germany France UK

Frequency Win % Frequency Win % Frequency Win %

Overall 569 53.60 304 36.84 61 24.59

Infringement 569 60.81 205 55.61 61 63.93

Validity 142 49.3 225 54.22 60 36.66

Of note is the great disparity in the share of patents that undergovalidity examination. In the UK, validity counterclaims were raised inall but one dispute. In Germany, this share was of less than a quarter.This might be explained by institutional differences, since validityclaims need to be raised in front of a separate specialized court inGermany.

The division of litigated patent type is not congruous across juris-diction (see Figure 3.5). In France, upwards 70 percent of all casesinvolved mechanical inventions. In contrast, a comparatively dispro-portionate share of the patents were of the chemistry or softwaretechnology in the UK.

58 protection heterogeneity in europe

Mechanical - 57.64%Electronics - 15.82%Chemistry - 7.56%Software - 11.42%Optics - 5.98%Biotech - 1.58%

Mechanical - 71.38%Electronics - 15.46%Chemistry - 9.87%Software - 1.32%Optics - 1.64%Biotech - 0.33%

Mechanical - 52.46%Electronics - 16.39%Chemistry - 18.03%Software - 8.20%Optics - 1.64%Biotech - 3.28%

Figure 3.5: Technology Distribution in Germany, France and in the UK

From a descriptive perspective, we observe that the win rates bytechnology concealed remarkable variation, across and cross-country(see Tables 1, 2 and 3 in the Appendix C). There appears to be no ten-dency of one technology being particularly strong or weak across allthree jurisdictions. The only exception were software patents. Theyunderperformed consistently in validity trials across all jurisdictions.This mirrors the perception that software is less reliant on patentsand that the line between patent-eligible software patent claims fromineligible ones is still blurry45. Across industries, the only consistenttheme is that pharmaceutical patents underperformed consistentlyin terms of infringement and validity (see Table 4, 5 and 6 in theAppendix C). The relative weakness of pharmaceutical patents is sur-prising, especially considering the traditional view that the currentpatent system is most accommodating for this industry and that thedata presented by Allison, Lemley, and Schwartz (2015) shows thatpharmaceutical patents have fared well before US courts46. As to therest, the apparent commonality is the result of performing well in onecountry but not in the other.

45. See Bradford L. Smith and Susan O. Man, “Innovation and Intellectual PropertyProtection in the Software Industry: An Emerging Role for Patents,” University ofChicago Law Review 71 (2004) suggesting that the growth of the software industry islinked to the implementation of copyright and patent regimes for software in thelate 1970s and early 1980s; for the contrary view that patents are not serving thesoftware industry, see Mark H. Webbink, “A New Paradigm for Intellectual PropertyRights in Software,” Duke Law & Technology Review 4 (2005).

46. John R. Allison, Mark A. Lemley, and David L. Schwartz, “Our Divided Pa-tent System,” University of Chicago Law Review 82, no. 3 (2015): 1073–1154 statingthat these results fit with the classical representation that pharmaceutical patents arestrong and valuable, and are perceived as critical to protect R&D investments, seeamongst many Claus R. Gawel, “Patent Protection as a Key Driver for Pharmaceuti-cal Innovation,” Pharmaceuticals Policy and Law, 18, nos. 1-4 (2006): 45–53.

3.7 empirical results 59

As to the quality of the patents litigated before the three jurisdictions,we observe that the average quality was highest in the UK, followedby Germany and France (see Figure 3.6). The distribution of the pa-tent quality is skewed right, with an average q = 0.19 (see Figure3.7).

.15

.2

.25

.3

Aver

age

Pate

nt Q

ualit

y

FranceGermanyUK

Figure 3.6: Average Patent Qua-lity.

0

50

100

150

200

Freq

uenc

y

0 .2 .4 .6 .8 1

Patent Quality

Figure 3.7: Distribution Quality.

3.7 empirical results

We now turn to the empirical testing of the hypotheses developedpreviously. Firstly, we present the state of litigation across the threejurisdictions studied, Germany, France and the UK. Secondly, we ana-lyze the technology of the patented invention as a predictor for theoutcomes at trial, thirdly the industry of the patent. Finally, the asses-sment of the patent quality is discussed.

3.7.1 Econometric Specification

We model the plaintiffs (discrete) success in case i, denoted as Yi, asa function of technology T , industry I, country C and patent value Vto study the data on patent litigation. Our specification is

Yi,t,j,k = α+βjTi,t +βjIi,j +βkCi,k +βvVi + εi,t,j,k (7)

where the dependent variable is the case outcome, with Yi equal to1 if the claim is successful at trial and 0 if the claim is not upheld. A‘win’ was reported if the patent holder could enforce its infringement

60 protection heterogeneity in europe

claim before the courts, i.e., at least one of the claims was found tobe infringed and that claim was, if challenged, upheld as valid. Twofurther specifications use the infringement outcome and the invalidityoutcome as dependent variable, respectively.

T , I and C are indicators to control for level differences across thesefactors. V is the patent quality proxy we developed, a continuousvariable between ∼ N(0, 1), for the patent under dispute in case i.Furthermore, α is the constant, subscript t indicates technology t,subscript j indicates industry j, subscript k indicates country k andsubscript v indicates the coefficient for patent quality. We estimateβ coefficients for the different factors using logit regression. Finally,εi,t,j,k is an error term with the usual distributional assumptions. Weconsistently use ‘chemistry’ as the contrast dummy for technology,‘goods and services for consumer uses’ for industry and ‘Germany’for the country indicator.

3.7.2 Results

The logit regressions we conducted to test for the effect of the countryof litigation, technology, industry and quality of the patent on thelikelihood of success of patent holders led to the following results (seeTable 3.3 and Appendix Table A3.8 for the full regression table).

Country. Patents litigated before the courts in Germany were overallsignificantly more likely to prevail than patents in France (β = −.68,p < .01) and in the UK (β = −1.26, p < .01), when controlling fortechnologies, industries and patent quality. In other words, patentholders in Germany were more likely to enforce their patents withsuccess without regard to the underlying patent. We then distinguishbetween the decision on the infringement and on the validity of thepatent. We find that there was no significant effect of the country va-riable on infringement outcomes. By contrast, and in line with thedescriptive results, patents undergoing validity counterclaims in theUK were significantly more likely to be invalidated than their coun-terparts in Germany (β = 0.62, p < .1).

3.7 empirical results 61

Table 3.3: Summary Regression Results

(Definitive Win) (Infringement) (Invalidity)

Country

FRANCE -0.741∗∗∗ -0.178 0.157

(-4.74) (-1.01) (0.63)

UK -1.448∗∗∗

0.0968 0.623∗

(-4.38) (0.33) (1.79)

Technology

Software -0.677 -0.656 1.215∗

(-1.47) (-1.35) (1.65)

Industry

Biotechnology 1.097 2.191∗ -0.173

(1.17) (1.72) (-0.11)

Construction -0.975∗∗∗ -1.036

∗∗∗0.497

(-3.54) (-3.54) (1.27)

Goods & Services for Industrial Uses -0.464∗ -0.378 0.369

(-1.94) (-1.45) (1.07)

Medical Devices -0.558∗ -0.487 0.690

(-1.73) (-1.41) (1.52)

Pharmaceuticals -0.845∗ -1.236

∗∗1.491

∗

(-1.71) (-2.39) (1.93)

Patent Quality

Quality Proxy 2.369∗∗∗

1.935∗∗∗ -1.061

(3.92) (2.93) (-1.26)

Constant -0.0640 0.655 -0.213

(-0.17) (1.62) (-0.38)

chi2 84.40∗∗∗

47.95∗∗∗

35.36∗∗∗

Pseudo R2 0.0656 0.0426 0.0597

Observations 932 835 427

Standard errors in parentheses∗ p < .1, ∗∗ p < .05, ∗∗∗ p < .01

Technology. When controlling for differences across countries, thetechnology underlying patents was not a significant predictor for theoverall outcome at trial nor the infringement decisions. However, soft-ware patents were significantly more likely to be invalidated relativeto chemistry patents (β = 1.25, p < .1).

62 protection heterogeneity in europe

Industry. A number of industries were associated with a significantlyhigher or lower likelihood to win at trial. Patents in the pharmaceuti-cal industry were significantly less likely to prevail overall (β = −0.79,p < .05), but also on decisions over infringement (β = −1.236, p < .05)and validity (β = 1.49, p < .1). On the contrary, biotechnology pa-tents were significantly more likely to win infringement disputes(β = 2.19, p < .1); with a caveat due to the low sample size. Pa-tents deployed in construction were significantly less likely to prevailoverall (β = −0.94, p < .01) and in infringement trials (β = −1.04,p < .01). Two additional industries were significant predictors forthe overall outcome. Patents in the sphere of medical devices and ofgoods and services for industrial and business goods were less likelyto win overall (β = −0.56, p < .1 and β = −0.97, p < .01). This lowerperformance seems to be driven by the higher share of validity dis-putes in these specific industries rather than by any bias at invalidityor infringement trials. In fact, these two industries did not predictinfringement or invalidity with any level of significance.

Quality. The quality of the patent undergoing litigation was a signi-ficant predictor of both its overall success and the infringement out-come (β = 0.303, p < .01 and β = 1.2, p < .01). There was, however,no significant effect of the patent quality variable on the outcome atinvalidity trial. Patents of higher quality were more likely to prevailin litigation and to be found infringed, but not to be held valid. Evenwhen excluding the UK, the patent quality proxy did not correlatewith the likelihood of patent holders to prevail for validity challen-ges.

3.8 discussion

3.8.1 Differences Across Jurisdiction

Our data suggest that Germany should be the most attractive forumfor patent holders, as patents claims are most successfully litigatedthere. Overall, the choice of forum was decisive; although for diffe-rent reasons depending on the particular jurisdiction. The courts inthe UK appears to be somewhat biased against patent holders; that isthey were more likely to invalidate patents irrespective of the charac-

3.8 discussion 63

teristics of the underlying patents. On the other hand, the outcomevariation in France seems to result from higher share of patents un-dergoing validity trials and in the particular industries at stake.

Germany and the UK. The reputation of the UK courts is to be ‘anti-patents’, that is to be an unfriendly jurisdiction for patent holders be-cause of low win rates and high costs of litigation47. As such, the UKis thought to constitute a propitious jurisdiction for challenging thevalidity of patents or requesting declarations of non-infringement48.In line with this perception, we find that less than a quarter of allpatent infringement lawsuits ultimately resulted in a win. By distin-guishing between infringement and the validity challenge raised, amore refined picture emerges. The UK courts were awarding infrin-gement in over two third of all instances, the highest rate amongstthe three countries of study. The crux of the matter lies in the invali-dity counterclaims. Controlling for the patent at stake, its technology,industry and quality, patents were significantly more likely to be inva-lidated in the UK than in Germany. This implies that there may exista negative bias against patents in validity trials in the UK, comparedto Germany. It is unclear whether this negative bias against patentholders suggested by our data stems from differences in proceduralrules, such as more stringent disclosure and expert witness rules inthe UK, or from the judge-specific application of the substantive pa-tent law. Irrespective of the origin, such a stricter validity standardcontradicts the notion of a harmonized European patent system.

Germany and France. When comparing the success chances of patentholders in Germany and with those in France, we find that althoughthere is a significant difference in the overall success probability, thisis not the case for the individual infringement and invalidity outco-mes. This implies that the difference between the two jurisdictions

47. Illustratively, when the alleged infringer brought the case to the Patents Courtsin London, patent holders only won in 14 % of the cases in 2006 to 2009, see MichaelElmer and Stacy Lewis, “Where to Win: Patent-Friendly Courts Revealed,” Mana-ging Intellectual Property, October 2010, and further Helmers and McDonagh, “PatentLitigation in the U.K.: An Empirical Survey 2000–2008.”

48. Gary Moss, Matthew Jones, and Robert Lundie-Smith, “Just How ‘Anti-Patent’are the UK Courts?,” Journal of Intellectual Property Law & Practice 5 (2010): 148–157

reviewing this perception and contending that the courts became more friendly tothe patentees in 2008 and 2009.

64 protection heterogeneity in europe

does not lie in a bias of the courts, but rather in the share of patentsundergoing validity challenge and in the characteristics of the patentsat stake. Over two thirds of all French patents were subject to validitycounterclaims. By contrast, in Germany, this share was less than aquarter. While in both cases close to half of the patents were foundto be invalid, the difference in the share of patents undergoing suchexamination led to a substantial difference in the overall outcome. Asmentioned above, the most probable cause of this is the German in-stitutional setting, which has a separate trial procedure for invalidityclaims. Under the German patent litigation system, the alleged infrin-ger must file a lawsuit with a specialized patent court, rather thanbringing the validity defense within the infringement proceedings49.Furthermore, we find level differences in the industries of the pa-tents litigated. The share of patents from the construction industry,which we find to be significantly more likely to lose on infringementgrounds, was highest in France. There were also more pharmaceuticalpatents respectively and less biotechnology patents litigated in Francethan in Germany. Finally, patents in Germany were, on average, ofhigher quality. The higher the quality, the more likely the patent wasto prevail in infringement disputes50.

France and the UK. Patent holders in France were overall more likelyto prevail compared to the UK, but this did not result from diffe-rences in the appraisal of infringement or invalidity. There was nosignificant impact of the country variable on both of these outcomes,suggesting that differences in the technology, industry, or quality ofthe patents explained the overall difference. France, for instance, was

49. For a theoretical and empirical investigation of the costs and benefits of bifur-cated systems, see Katrin Cremers et al., “Invalid but Infringed? An Analysis of theBifurcated Patent Litigation System,” Journal of Economic Behavior & Organization 131

(2016): 218–242 demonstrating that infringers are less likely to challenge validity inbifurcated systems than in non-bifurcared ones.

50. Surveys have shown that most companies are unsatisfied with the amount ofthe reimbursement of procedural costs in French patent litigation, qualifying them as‘way too low’, as well as with the amount of damages awarded, while they were sa-tisfied with the reimbursement in Germany, see Gregoire Triet and Luc Santarelli,Propriété industrielle : le coût des litiges. Etude comparée entre la France, l’Allemagne,l’Angleterre, les Etats-Unis, l’Espagne et les Pays-Bas, (Ministère de l’Economie, desFinances et de l’Industrie, May 2000); see also Graham and Van Zeebroeck, “Compa-ring Patent Litigation Across Europe: A First Look.”

3.8 discussion 65

the ground-field for only very few software patent disputes, whichwere significantly more likely to be found invalid. A large part of thelitigated patents in France were in the mechanical technology cate-gory, which performed strongly in court51. When contrasting Franceto the UK, the selection effects in the UK and the level differencesin validity challenges, rather than the forum, appear to explain thedifferences in outcome.

3.8.2 Differences Across Patent Quality

Our empirical finding that patent quality significantly predicts patentlitigation outcome corresponds with the prediction of our model. Asdiscussed above, we hypothesize that the judge will make assessmentof patent quality at trial, resulting in the assignment of a ‘judicial’patent quality which determines the success probability at trial. Weacknowledge that our notion of patent quality focuses on the standa-lone patent quality only, not on its prosecution history or to the cha-racteristics of the applicant, inventor or of the judges52. This beingsaid, our patent quality measure was an significant predictor withrespect to overall outcomes and infringement outcome at trial.

Furthermore patent quality provides further insights on the fragmen-tation of the European enforcement system. According to our rese-arch, fragmentation in Europe stems largely from differences in theappraisal of validity counterclaims. In that regard, our empirical fin-ding that validity decisions are not associated with patent qualityprovides an interesting insight. One explanation for this might bethat patent holders have better access to information about the va-lidity of their patent and existing prior art, and could thus be in a

51. According to findings by Véron, “Le Contentieux des Brevets d’Invention enFrance, Étude Statistique 2000-2010,” the incidence of litigation was higher for me-chanical patents for the time span 2000 to 2009 than for any other technology.

52. In the United States, Ronald J. Mann and Marian Underweiser, “A New Lookat Patent Quality: Relating Patent Prosecution to Validity,” Journal of Empirical LegalStudies 9, no. 1, 1–32 found statistically significant relations between validity decisi-ons and ex ante aspects of the prosecution history such as the existence of internalpatent office appeals; for an overview of the studies relating to the effects of gender,religion or political preferences on case outcomes within the past two decades, seeStephen J. Choi and Mitu G. Gulati, “Bias in Judicial Citations: A Window into theBehavior of Judges?,” Journal of Legal Studies 37 (2008): 87–129.

66 protection heterogeneity in europe

stronger position to defend their case, making the case outcome lessdependent on absolute patent (or at least our empirical estimator ofthe latter).

The jurisdiction with the highest quality patents was the UK, wherethe overall win rate was much lower than in France and in Germany.This should be put in the context of patent litigation costs in theUK, which are substantially higher than in the other two jurisdicti-ons. This may crowd out lower quality inventions, where the partiesprefer to settle the dispute or even to renounce to the enforcement oftheir rights since the expected payoff from trial may be too low. Thisenforcement threshold may lead to a secondary effect of market dis-tortion by discouraging the enforcement of valid (but rationally unen-forceable) patents53. In other words, the UK system could be said toimpose a ‘tax’ on innovation, as UK innovators have lower incentivesto invest in lower quality patents that cannot be enforced purely onthe basis of litigation costs rather than substantive aspects54.

3.8.3 Future of Litigation in Europe

The Unified Patent Court (UPC) is a proposed supra-national com-mon court which is supposed to commence operations within thenear future. It has the goal of harmonizing patent enforcement by pro-viding a one-stop litigation forum for patent disputes. In an effort tobring greater harmony and cross-national predictability to patent liti-gation, a single court shall, in principle, deal with all civil litigation re-lating to European patents. This means that the UPC will have exclu-sive jurisdiction over all infringement and invalidation proceedingsfor European patents (including the new Unitary Patent) throughout

53. Thomas S. Ellis, “Distortion of Patent Economics by Litigation Costs,” CASRIPPublication Series: Streamlining International Intellectual Property, 1999, ‘[i]t is, simplyput, that the escalating, indeed skyrocketing litigation costs of the 1970’s and 1980’shave distorted patent markets and patent economics. Put another way, it is my ob-servation that the escalating costs associated with litigating patent infringement andvalidity issues discourage challenges to patents, thereby essentially equating the en-try barriers for presumptively valid, but weaker patents with those entry barriersassociated with strong or judicially tested patents.’.

54. See James Bessen and Michael J. Meurer, “The Private Costs of Patent Litiga-tion,” Boston University School of Law Working Paper, nos. 07-08 (2007) for costly litiga-tion as tax on innovation when it flows from the risk of unavoidable infringement.

3.8 discussion 67

the 25 European territories involved (all EU Member States exceptSpain, Poland and Croatia)55. The Court of First Instance will havea Central Division seated in Paris and sections in Munich and Lon-don. The working of the courts will be technology based. As such,the cases will be attributed based on subject matter with London forpatents falling into the International Patent Classification of WIPOsections (A) Human Necessities and (C) Chemistry and Metallurgy,Munich for patents in section (F) Mechanical Engineering, Lighting,Heating, Weapons and Blasting and Paris for all other sections. Furt-hermore, Member States may set up local division, in general oneper state with the exception for Germany with four local divisions,and regional divisions such as the ‘Nordic-Baltic’ regional division inStockholm56. Appeals on points of law and facts can be lodged withthe Court of Appeal, located in Luxembourg. Furthermore, questionson the interpretation in matters of EU law may be submitted to theEuropean Court of Justice. The Unitary Patent, which is a new typeof European patent with validity in all 25 states, will be granted oncethe Unified Patent Court enters into force57.

The court will enter into force when at least thirteen states, includingthe three largest patent granting states France, Germany and the UK,have ratified the Agreement on the Unified Patent Court. The UK ra-tified the UPC Agreement in April 2018, but the impact of Brexit andthe capacity of the UK to remain a member of the UPC after its poten-

55. See Kevin P. Mahne, “A Unitary Patent and Unified Patent Court for the Euro-pean Union: An Analysis of Europe’s Long Standing Attempt to Create a Supranati-onal Patent System,” Journal of the Patent and Trademark Office Society 94 (2012): 162–191 for a historical perspective and Dimitris Xenos, “The European Unified PatentCourt: Assessment and Implications of the Federalisation of the Patent System inEurope,” SCRIPTed: A Journal of Law, Technology and Society 10, no. 2 (2013): 246–277

for a criticism of the lack of democratic process-making as the UPC establishes amonopolistic source of legal power.

56. See Art. 7 para. 3 Unified Patent Court Agreement and Impact of the UnitaryPatent Protection and the Unified Patent Court in the Nordic-Baltic Region, (Ministry ofJustice of Sweden, October 2014).

57. Richard Pinckney, “Understanding the Transitional Provisions of theAgreement on the Unified Patent Court,” European Intellectual Property Review 37,no. 5 (2015): 268–277 and Alan Johnson and Luke Maunder, Challenging the UPCOpt-Out – How Exactly Will It Work?, (Bristows, June 2017).

68 protection heterogeneity in europe

tial withdrawal from the EU are unclear58. In Germany, a constituti-onal complaint pending in the German Federal Constitutional Courtregarding the implementation law passed by the German Parliamentis causing delay to the ratification. At the time of writing this article,the fate of the UPC therefore remains undetermined. Despite theseuncertainties, the UPC Preparatory Committee is currently procee-ding on planning, among others, for the recruitment and training ofUPC judges59. The latter, namely the training of patent judges on thesame cases will be key to achieving homogeneity beyond the same setof statutory requirements, as it will reduce intra-UPC heterogeneityin patent decisions.

The ultimate aim of the UPC is to provide for a one-stop litigation fo-rum for patent disputes, thus enabling litigation with European-wideeffect. The court shall, thus, avoid or reduce intra-community tradebarriers, the high costs of parallel litigation, and inconsistent decisi-ons or strategic litigation that may follow from a fragmented courtsystem60. Harhoff (2009) estimated that the total private savings fromhaving access to a unified patent court would range between EUR147 to 289 million annually. A variety of challenges will mark boththe transitional and long-run operation of the court, creating barriers

58. For the widely debated topic of the participation of the UK in the UPC, see, e.g.,Ansgar Ohly and Rudolf Streinz, “Can the UK Stay in the UPC System After Brexit?,”Journal of Intellectual Property Law & Practice 12, no. 3 (2017): 245–258; Matthias Lam-ping and Hanns Ullrich, “The Impact of Brexit on Unitary Patent Protection and itsCourt,” Max Planck Institute for Innovation & Competition Research Paper, nos. 18-20

(2018); and Winfried Tilmann, “The Harmonisation of Invalidity and Scope of Pro-tection Practice of the National Courts of EPC Member States,” International Reviewof Intellectual Property and Competition Law, 2006, 62–74.

59. At a conference, chairman Alexander Ramsay of the UPC Preparatory Com-mittee said work to prepare for the UP system is going on: ‘[w]e are trying touse the time that we have to make sure the system works when we’re starting.’,see http://patentblog.kluweriplaw.com/2018/02/15/plans-training-judges-un

ified-patent-court-ready.60. See, Dietmar Harhoff, “Economic Cost-Benefit Analysis of a Unified and Inte-

grated European Patent Litigation System,” Final Report, Tender No. Markt/2008/06/D,2009, Mejer and van Pottelsberghe de la Potterie, “Economic Incongruities in theEuropean Patent System.” Undesirable effects of parallel litigation include addedfinancial burden of proceeding in different courts, consumption of judicial resourcesin multiple jurisdictions and inconsistent decisions, see Yoshimasa Furata, “Interna-tional Parallel Litigation: Disposition of Duplicative Civil Proceedings in the UnitedStates and Japan,” Pacific Rim Law & Policy 1 (1995): 1–58.

3.8 discussion 69

to such goals61. Firstly, there will be a transitional period of sevento fourteen years, during which national courts and the UPC willhave parallel jurisdiction over actions concerning (non-unitary) Eu-ropean patents62. The potential for forum shopping and divergencesin decisions will not vanish for the foreseeable future and will alsobe part of the new system. In particular, infringement lawsuits canbe brought before local and regional divisions, which will have ju-risdiction to decide on potential counterclaims for revocation or re-ferrals to the Central Division63. Pan-European litigation will not bea concrete possibility as thirteen state parties to the European PatentConvention will not be part of the new system64. This will give rise tothe possibility of parallel, duplicate or inconsistent litigation and deci-sions65. Litigation with European-wide effects remains a target ratherthan a reality. However, with the Central Division in London of theCourt of First Instance having exclusive jurisdiction over revocationlawsuits, and the Court of Appeal leveling up eventual heterogeneous

61. See, e.g., Roberto Romandini and Alexander Klicznik, “The Territoriality Prin-ciple and Transnational Use of Patented Inventions – The Wider Reach of a UnitaryPatent and the Role of the CJEU,” International Review of Intellectual Property and Com-petition Law 44, no. 5 (2013): 524–540 and Matthias Brandi-Dohrn, “Some CriticalObservations on Competence and Procedure of the Unified Patent Court,” Internati-onal Review of Intellectual Property and Competition Law 43 (2012): 372–389.

62. Art. 83 Unified Patent Court Agreement. Parallel litigation during the tran-sitional period raise a number of questions of lis pendens, see an overview underhttp://eplaw.org/eu-division-of-jurisdiction-in-transitional-period.

63. See Art. 33 para. 4 lit. a-c Unified Patent Court Agreement. For a discussion ofthe extent to which the establishment of the UPC can contribute to enhance judicialcoherence, see Federica Baldan and Esther Van Zimmeren, “The Future Role of theUnified Patent Court in Safeguarding Coherence in the European Patent System,”Common Market Law Review 52, no. 6 (2015): 1529–1577.

64. The European Court of Justice held that an agreement creating a unified patentlitigation system including courts in countries outside the EU not compatible withthe provisions of the Treaty on European Union and the Treaty on the Functioningof the European Union (European Court of Justice, Opinion 1/09, March 8, 2011).Henceforth, the States parties to the European Patent Convention but not to the EUwere excluded. Three Member States refused to join the new system, namely Spain,Poland and Croatia.

65. A European patent validated in Spain, Germany and France would thereforebe subject in parallel to the jurisdiction of a national court for its Spanish part andof the unified court for its German and French parts. This will require coordinatinglitigation for pan-European enforcement of patent rights, but also to a certain extentthe continued potential for forum shopping.

70 protection heterogeneity in europe

infringement and invalidity counterclaims before the local or regionaldivisions of the Contracting Member States, the UPC will be able tostrengthen the harmonization of European patent litigation.

3.9 conclusion

This paper investigates country-, technology- and industry-specificityof the European patent system. We develop a general model of patentdisputes which predicts that both settlement and litigation are drivenby patent quality. While at the settlement stage, the relative patentquality estimations of patent holder and infringer respectively deter-mine whether a settlement is reached or not, at trial stage, it is thejudge’s estimation of patent quality that decides the case. As a re-sult, the model predicts that high quality patents will prevail in court,despite the selection effect from the settlement stage. We test the pre-diction of the model at the litigation stage using a dataset that coverspatent litigation in Germany, France and the United Kingdom for theperiod between 2008 and 2012. We find evidence that the Europeanpatent system is heterogeneous across jurisdictions. The data showsthat the forum in which a patent case is brought matters more thanthe individual characteristics of litigated patent.

Albeit its numerous challenges, the UPC could be a real paradigmshift when it comes to heterogeneity in the European patent system,setting it on the path of a unitary patent system, both nominally andin practice. As evidenced by recent developments surrounding theUPC, this first step towards homogeneity will be countered by natio-nal resistance and institutional challenges. However, it is to be hopedthat, with time, the idea of a single court for the enforcement of pa-tents in Europe will be the norm. The state of innovation will certainlydepend heavily on the capacity of Europe to form a united front.

3.10 appendix 71

3.10 appendix

a . technology coding

We hand-coded the patents at dispute by following the classificationdeveloped by Allison, Lemley and Schwartz (2015) that distinguishesbetween mechanical, electronics, chemistry, biotechnology, softwareand optics patents as follows66:

1. Mechanical: An invention in which the claims cover the use of mecha-nical parts, either solely or predominantly, sometimes combined withheat, hydraulics, pneumatics, or other power sources or power transfertechniques.

2. Electronics: An invention in which the claims cover the use of tra-ditional electronic circuitry or the storage or transmission of electricenergy.

3. Chemistry: An invention in which the claims cover chemical reacti-ons, chemical compounds with specific elements and proportions, andchemical processes specifying specific elements and amounts or pro-portions. Closely related inventions such as those on purportedly no-vel metal alloys and nonmetallic composites are also included whenthe claims cover the specific components and proportions of such amal-gams. This technology area includes “small-molecule” chemistry; DNA,antibodies, and other large molecules are included in the biotechnologycategory instead. Although many of the chemistry technology patentswere assigned to the pharmaceutical industry category, they are alsofound in other industry categories such as semiconductors.

4. Biotechnology: An invention in which the claims cover processesinvolving advanced genetic techniques intended to construct new mi-crobial, plant, or animal strains; a product created from such a pro-cess; or the way such a process or product is used in biotechnologyresearch. Although there are a number of different genetic-engineeringtechniques, for several reasons we decided not to disaggregate thesetechniques into separate technology areas.

66. Allison, Lemley, and Schwartz, “Our Divided Patent System.”

72 protection heterogeneity in europe

5. Software: An invention in which the claims cover data processing—theactual manipulation of data (and not merely transmission, receipt,or storage of data), regardless of whether the code carrying out suchdata processing is on a magnetic storage medium, embedded in a chip(“firmware”), or resident in flash memory.

6. Optics: An invention in which the claims cover the use of light wavesor light energy. We also assigned certain patents in the “primary” soft-ware classification to one of that technology’s subsets, namely, softwarebusiness methods. As we defined it, the software business method cate-gory includes software patents that cover models, methods, and techni-ques for conducting business transactions. Business-method patentsare notoriously difficult to define, with possible definitions varyinggreatly in scope.

b . industry coding

Similarily, we utilized classification developed by Allison, Lemley andSchwartz (2015) to distinguish patents by industry of use67:

1. Computer and Other Electronics: This industry encompasses in-ventions of all kinds that purport to advance the state of the art incomputing or computer device manufacturing, or to enhance users’experiences in employing computing technology. The category inclu-des software and computer hardware inventions that seek to serve theaforementioned purposes. Also included are inventions predominatedby the use of traditional electronic circuitry when those inventionspurport to advance the art in that technology or enhance users’ ex-periences in employing electronics technology. In contrast with ourprior studies, here we combine the computer and traditional electro-nics industries because we find fewer and fewer patents covering tra-ditional electronics without also including significant data processingelements. Traditional electronics inventions without data processingelements do continue to exist, but their frequency and importance israpidly declining—the industries clearly have been merging for quitesome time.

67. Allison, Lemley, and Schwartz, “Our Divided Patent System.”

3.10 appendix 73

2. Semiconductor: The semiconductor industry category includes in-ventions of any kind intended to advance the state of the art in rese-arching, designing, or fabricating semiconductor chips. Technologiesemployed in semiconductor industry inventions may include software,chemistry, optics, and mechanical.

3. Pharmaceutical: The pharmaceutical industry category includes pa-tents on drugs for treating diseases or other abnormal conditions inhumans or animals, as well as processes for producing or using suchdrugs. The technologies found in pharmaceutical industry inventionsare overwhelmingly chemistry or biotechnology.

4. Medical Devices, Methods, & Other Medical: This industry cate-gory includes , non-biotechnology inventions of any kind used for rese-arch on, or for the diagnosis or treatment of, diseases or other abnormalconditions in humans or animals. Patents on processes and productsfor pharmaceutical purposes are not included in this category. All ofthe different technology fields are represented in the medical industrycategory.

5. Biotechnology: This category includes those inventions that are inthe biotechnology technology category that do not relate to the pro-duction of pharmaceutical compositions or medical diagnostics or tre-atment, but that instead purport to advance the science of biotechno-logy itself.

6. Communications: The communications industry category includesinventions of all kinds intended to advance the state of the art in com-munications. Technologies represented in the communications indu-stry include software, electronics, optics, and mechanics.

7. Transportation: This category includes patents on any type of inven-tion related to the production of automobiles or vehicles of any otherkind intended for transporting people or cargo, and inventions relatedto the provision of transportation services. Several different technologyareas are represented in this industry category.

8. Construction: The construction industry category includes inventi-ons of all kinds related to the erection or maintenance of structures, orto excavation.

9. Energy: This category includes inventions of any kind associated withpower generation, transportation, or consumption.

74 protection heterogeneity in europe

10. Goods & Services for Industrial & Business Uses: This categoryincludes patents on products and services of all kinds intended for in-dustrial and business purposes, i.e. goods and services for wholesaleuses that are not in another, more specific category. Many software-implemented business method inventions are included in this cate-gory..

11. Goods & Services for Consumer Uses: This category includes pa-tents on products and services of all kinds intended for personal con-sumer purposes, i.e. goods and services for retail uses that are not inanother, more specific category. Many software-implemented businessmethod inventions are included in this category.

c . caseload and appeal

Table A3.1: Summary Caseload by Country

County First Instance Decision Appellate Decisions

Germany (All) 711 210

Germany (Infringement) 569 187

Germany (Revocation) 142 23

France 304 97

UK 61 24

3.10 appendix 75

d. litigation outcomes by technology and industry

Table A3.2: Overall Win Rate by Technology

Overall Win Rate by Technology

Technology Germany France UK

Frequency Win % Frequency Win % Frequency Win %

Mechanical 328 55.66 217 35.94 32 31.25

Electronics 90 51.11 47 53.19 11 10.0

Chemistry 43 45.51 30 23.33 11 36.36

Software 65 44.62 4 0 5 0.0

Optics 34 58.82 5 40.0 1 0.0

Biotechnology 9 77.8 1 0.0 2 0

Total 569 53.60 304 36.84 61 24.59

Table A3.3: Infringement by Technology

Infringement by Technology

Technology Germany France UK

Frequency Win % Frequency Win % Frequency Win %

Mechanical 328 59.15 152 54.11 32 59.38

Electronics 90 60.0 36 66.67 10 80.0

Chemistry 43 62.8 20 47.37 11 54.55

Software 65 60.0 - - 5 100.0

Optics 34 73.53 3 66.67 1 0.0

Biotechnology 9 77.8 1 0.0 2 0.0

Total 569 60.81 205 55.61 61 63.93

76 protection heterogeneity in europe

Table A3.4: Invalidity by Technology

Invalidity by Technology

Technology Germany France UK

Frequency Win % Frequency Win % Frequency Win %

Mechanical 64 57.81 165 55.15 32 56.25

Electronics 30 60.0 29 62.07 10 0.0

Chemistry 10 20.0 22 45.45 10 40.0

Software 22 31.82 4 0.0 5 0.0

Optics 13 30.77 5 60.0 1 0.0

Biotechnology 3 66.7 - - 2 0

Total 142 49.3 225 54.22 60 36.66

Table A3.5: Overall Win Rate by Industry

Overall Win Rate by Industry

Technology Germany France UK

FrequencyWin % FrequencyWin % FrequencyWin %

Goods & Services for Industrial Uses 163 55.21 81 28.4 17 29.41

Construction 74 40.54 58 22.41 2 100.0

Transportation 57 59.65 55 43.64 7 42.86

Goods & Services for Consumer Uses 56 57.14 48 54.2 3 33.3

Medical Devices, Methods & Medical 47 59.57 15 13.3 9 11.1

Computer and Other Electronics 54 53.7 14 42.86 2 0.0

Communications 61 45.9 13 92.31 12 8.3

Pharmaceuticals 16 37.5 15 33.3 9 22.2

Energy 20 70.0 4 25.0 2 0

Semiconductor 9 44.4 1 0.0 - -

Biotechnology 12 83.3 - - - -

Total 569 53.60 304 36.84 61 24.59

3.10 appendix 77

Table A3.6: Infringement by Industry

Overall Win Rate by Industry

Technology Germany France UK

FrequencyWin % FrequencyWin % FrequencyWin %

Goods & Services for Industrial Uses 163 60.12 51 47.06 17 88.24

Construction 74 43.24 37 37.83 2 100.0

Transportation 57 63.16 41 60.98 7 42.86

Goods & Services for Consumer Uses 56 64.29 36 66.67 3 100.0

Medical Devices, Methods & Medical 47 68.09 6 33.3 9 22.2

Computer and Other Electronics 54 64.81 9 66.67 2 50.0

Communications 61 62.3 12 100.0 12 91.67

Pharmaceuticals 16 50.0 10 60.0 9 22.2

Energy 20 70.0 3 33.3 2 0

Semiconductor 9 66.67 - - - -

Biotechnology 12 91.67 - - - -

Total 569 60.81 215 55.61 61 63.93

Table A3.7: Invalidity by Industry

Overall Win Rate by Industry

Technology Germany France UK

FrequencyWin % FrequencyWin % FrequencyWin %

Goods & Services for Industrial Uses 39 58.97 64 51.56 17 35.29

Construction 13 61.54 44 45.45 2 100.0

Transportation 8 62.5 39 61.54 7 71.43

Goods & Services for Consumer Uses 11 36.36 39 71.79 3 33.3

Medical Devices, Methods & Medical 17 41.18 11 27.27 9 66.67

Computer and Other Electronics 13 53.85 12 58.33 2 0.0

Communications 26 34.62 5 80.0 12 0.0

Pharmaceuticals 3 0.0 8 25.0 8 28.57

Energy 4 50.0 2 50.0 2 0

Semiconductor 5 60.0 1 0.0 - -

Biotechnology 3 66.67 - - - -

Total 142 49.3 225 54.22 60 36.67

3.10 appendix 79

e . coding book

This Section reports the procedure followed in the collection and co-ding of the patent decisions.

1. Type of cases. We collected all substantive decisions by courtson patent infringement and subsequent validity counterclaims.That is, revocation decisions were excluded, unless counted ascounterclaim (see ‘Bifurcation’ for Germany). When a patent in-fringement decision was raised as a counterclaim to a revocationaction, it was excluded. We excluded injunctions, procedural de-cisions, and non-infringement decisions.

2. Bifurcation. Because of the differences in patent system, infrin-gement and revocation had to be treated differently in Germany.For disputes before German courts, we joined patent infringe-ment disputes (before the regional courts) and invalidation dis-putes (before the Bundespatentgericht). A patent infringementdispute followed by an invalidity dispute over the same patentwas considered equivalent to a counterclaim in France or inthe UK. To be counted as such, the decision had to be rende-red between the same parties, over the same patent and overa timespan of 2 years after the (first or last) infringement de-cision. If the parties were anonymised, we assumed that theywere identical if the decision was rendered within a timespanof two years. In fact, we confirmed for 120 of 123 invalidity ca-ses in Düsseldorf that it was one of the alleged infringer thatchallenged the validity of the (arguably) infringing patent. In-versely, if an invalidation decision was rendered before the in-fringement decision, the case was excluded and considered asrevocation with infringement counterclaim. The latter cases arerare, with estimations that 90% of all revocation actions are filedin response to an infringement action68.

3. Timespan. The first instance decision had to be rendered bet-ween January 1, 2008 and December 31, 2012 to be counted asrelevant. The first substantive decision was determinant, that isprevious procedural decisions would not affect this count. The

68. Hees and Braitmayer, Verfahrensrecht in Patentsachen and Keukenschrijver, Pa-tentnichtigkeitsverfahren, p. 73-74.

80 protection heterogeneity in europe

search was first conducted in 2015 with a complete second co-ding run in 2017. Litigation cases in Mannheim and Munichwere coded in 2018.

4. Outcome. The outcomes are reported from the standpoint ofthe patent holder. An overall ‘win’ was reported if the patentholder could enforce its infringement claim before the courts,i.e., at least one of the claims was found to be infringed and thatclaim was, if challenged, upheld as valid. For infringement spe-cifically, a ‘win’ was reported if at least one claim was found tobe infringed, not matter its validity. For validity specifically, a‘win’ was reported if the claim in dispute, or in case of severalclaims at stake, one of the claims that was found to be infrin-ged was found valid. Assume the following example to clarifythe coding: At decision stage, patent claim 1 was found to beinfringed and valid, while claim 2 was declared not infringedand invalid; the overall outcome would be coded as a ‘win’, theinfringement outcome as a ‘win’ and the validity outcome as a‘win’.

A further difficulty in outcome coding is the presence of par-tial invalidity decisions in Germany (‘Teilnichtigkeit’). For suchdecisions, since infringement and validity proceedings are sepa-rated, it is unclear whether the infringing act falls within theamended claim. In our main specification, if the amended claimis upheld, we assumed that a violation was given and recordedan overall ‘win’. In a further specification, we excluded thoseinstances, which did not affect out results69.

69. With an overall win rate at 53.11% compared to 53.44% across all technologies,and within 1% difference by technology.

3.10 appendix 81

Table A3.8: Full Regression Results

(Overall Win) (Infringement) (Invalidity)

Country

FRANCE -0.741∗∗∗ -0.178 0.157

(-4.74) (-1.01) (0.63)

UK -1.448∗∗∗

0.0968 0.623∗

(-4.38) (0.33) (1.79)

Technology

Biotechnology -0.180 -1.334 0.184

(-0.22) (-1.48) (0.14)

Electronics 0.0437 -0.445 -0.139

(0.12) (-1.12) (-0.26)

Mechanical 0.319 -0.204 -0.345

(1.05) (-0.63) (-0.81)

Optics 0.192 0.202 0.696

(0.43) (0.41) (1.07)

Software -0.677 -0.656 1.215∗

(-1.47) (-1.35) (1.65)

Industry

Biotechnology 1.097 2.191∗ -0.173

(1.17) (1.72) (-0.11)

Communications -0.0816 0.373 0.734

(-0.21) (0.91) (1.30)

Computer and Electronics -0.0362 -0.00375 -0.212

(-0.09) (-0.01) (-0.35)

Construction -0.975∗∗∗ -1.036

∗∗∗0.497

(-3.54) (-3.54) (1.27)

Energy 0.264 0.0473 0.583

(0.54) (0.09) (0.67)

Goods & Services for Industrial Uses -0.464∗ -0.378 0.369

(-1.94) (-1.45) (1.07)

Medical Devices -0.558∗ -0.487 0.690

(-1.73) (-1.41) (1.52)

Pharmaceuticals -0.845∗ -1.236

∗∗1.491

∗

(-1.71) (-2.39) (1.93)

Semiconductor -0.959 -0.211 0.118

(-1.35) (-0.27) (0.13)

Transportation -0.0545 -0.202 -0.0669

(-0.20) (-0.67) (-0.17)

Patent Quality

Quality Proxy 2.369∗∗∗

1.935∗∗∗ -1.061

(3.92) (2.93) (-1.26)

Constant -0.0640 0.655 -0.213

(-0.17) (1.62) (-0.38)

chi2 84.40∗∗∗

47.95∗∗∗

35.36∗∗∗

Pseudo R2 0.0656 0.0426 0.0597

AIC 1240.06 1115.585 594.5897

BIC 1331.969 1205.407 671.6686

Observations 932 835 427

Standard errors in parentheses∗ p < .1, ∗∗ p < .05, ∗∗∗ p < .01

4PA R A L L E L L I T I G AT I O N : T H E C A S E O FB I O T E C H N O L O G Y

The first paper of this dissertation presented evidence of the failingharmonization of the patent systems in Europe. Demonstrated wasthe lack of importance of the characteristics of the litigated patent;rather, it was the forum to which the case was brought that was de-cisive. Accordingly, the creation of uniform substantial patent lawstandards in Europe appears to be countered by national judges anddifferences in application of harmonized patent law1. The dissertationtakes an additional approach to study the European harmonizationin greater detail by investigating the enforcement of identical patentsin several countries. The lack of consistent application of harmonizedpatent law may be illustrated best in cases where one patent is litiga-ted across a number of jurisdictions, but decided differently in each.Anecdotal evidence such as most famously the Epilady v. Remingtonlitigation (for a patent on epilators), and patent enforcement cases bySara Lee Corporation (for a patent on coffee pads), Document Secu-rity System Inc. (for a patent for stripes on banknotes) and Angiotech(for a patent on stents) resulted in contradictory judgments beforedifferent national courts2. The courts had to interpret the same Euro-pean patent, operate under very similar facts and utilize a commonstandard of interpretation; yet they came to polar decisions3.

1. Similarly Matthew Parker, “Giving Teeth to European Patent Reform, Overco-ming Recent Legal Challenges,” Emory International Law Review 26 (2012): 1079–1110

contending that ‘[i]n actuality, the national courts of Europe have construed patentsdifferently despite applying the so-called uniform requirements of the EPC’.

2. For Epilady versus Remington, see for Germany, Oberlandesgericht Düseldorf,2 U 27/89, November 21, 1991 and for the UK, Improver Corp v Remington Con-sumer Product Ltd, High Court of Justice, [1990] F.S.R. 181, May 16, 1989; for asummary of these disputes, see Malwina Mejer and Bruno van Pottelsberghe de laPotterie, “Economic Incongruities in the European Patent System,” European Journalof Law and Economics 34, no. 1 (2012): 215–234.

3. See John P. Hatter, “The Doctrine of Equivalents in Patent Litigation: An Analy-sis of the Epilady,” Indiana International & Comparative Law Review, 1995, p. 486.

82

parallel litigation : the case of biotechnology 83

Despite the harmonization efforts, enforcing a patent infringementlawsuit in Europe still occurs on a per-country basis. That is, procee-dings in a national court can only be brought in respect to damagesthat occurred within that forum4. Previous practices by courts – espe-cially in the Netherlands – to grant cross-border injunctions have beenstruck down or largely limited by the European Court of Justice5. Si-milarly, issues of patent validity, with the exception of the oppositionbefore the European Patent Office, are to be determined under the re-spective national law. Challenging the validity of a European patentin one country thus has effect for this specific national part of the Eu-ropean patent only. Despite scholars proposing to render mandatorycomparative interpretation for judges and require them to considereach other’s case law with regard to the extent of protection6, paten-tability standards and revocation grounds of patents, there is no legalobligation for courts to do so7. In recent years, judges have recogni-

4. The so-called ‘Shevill’ doctrine, see ECJ, Shevill (C-98/93); see also the BrusselsRegime with its EU Regulation 44/2001 replaced by EU Regulation 1215/2012.

5. Article 6 EU Regulation 44/2001 respectively Article 8 of EU Regulation1215/2012 allows parties to sue several defendants domiciled in different countriesfor a given proceeding in the courts for the place where any one of them is domi-ciled, provided the claims are so closely connected that it is expedient to hear anddetermine them together to avoid the risk of irreconcilable judgments resulting fromseparate proceedings. In GAT v Luk (C-4/03), July 13, 2006, the European Courtof Justice held that national courts have no jurisdiction over the infringement of fo-reign parts of an European patent when the validity of latter patent is challenged. InRoche v Primus (C-539/03), July 13, 2006, the European Court of Justice dismissedthe argument that not allowing multiple actions against foreign subsidiaries in a sin-gle jurisdiction would lead to ‘irreconcilable judgments’. Since European patents aregoverned by national laws only, the legal findings cannot be contradictory.

6. See Haertel Kurt, “Die Harmonisierungswirkung des Europäischen Patents-rechts,” Gewerblicher Rechtsschutz und Urheberrecht, Internationaler Teil, 1981, 479–490.

7. Schutz (UK) Limited v Werit (UK) Limited [2013] UKSC 16, paragraph 40 no-ted in that regard that ‘there can be no question of the courts in this jurisdictionfeeling obliged to follow the approach of the German courts, any more than the Ger-man courts could be expected to feel obliged to follow the approach of the Englishand Welsh courts’. See, however, Peter Walter, “Die Auslegung staatsvertraglichenund harmonisierten Rechts: Gewicht und Bedeutung von Entscheidungen auslän-discher Gerichte und der Beschwerdekammern des EPA,” Gewerblicher Rechtsschutzund Urheberrecht, 1998, 866–870, arguing that the interpretation of ‘European patentlaw’ or ‘international uniform law’ as stipulated by the EPC obliges courts to con-strue the laws in the spirit of international uniform law and aim at the creation ofcross-boarder legal uniformity.

84 parallel litigation : the case of biotechnology

zed the need to try to achieve consistency, and to depart from a pointauthoritatively decided in another country only if convinced that itsreasoning is erroneous8.

In its first Section (Section 4.1), this paper studies decisions overidentical patents by different courts to shed light on the extent ofharmonization in Europe. While a certain level of variation betweencourts might exist, systematic variance would be a further sign ofthe lack of success in the application of harmonized patent laws. TheEuropean effort towards harmonization is undermined if the practi-ces of the national courts systematically diverge9. In this study, bio-technology is being used as a larger example of such practice. Thisemerging technology exemplifies the problematic nature of litigationacross the European patent system. High cost R&D and internatio-nal go-to-market lead to biotechnological innovation requiring wide-scale patent validation, unlike mechanical patents which are occasi-onally sufficient to be validated at a single national level10. As anemerging technology with small players, biotechnology is dispropor-tionally affected by costs of litigation11. The qualitative assessment ofover a hundred biotechnology infringement and revocation decisionspresents the divide between the United Kingdom versus Germanyand the Netherlands. Patent claims were interpreted more narrowly,

8. Lord Justice Jacob in Grimme Maschinenfabrik GmbH & Co KG v Derek Scott(t/a Scotts Potato Machinery, [2010] EWCA Civ 1110: ‘[b]roadly we think the princi-ple in our courts – and indeed that in the courts of other member states – should beto try to follow the reasoning of an important decision in another country. Only ifthe court of one state is convinced that the reasoning of a court in another memberstate is erroneous should it depart from a point that has been authoritatively decidedthere. Increasingly that has become the practice in a number of countries, particu-larly in the important patent countries of France, Germany, Holland and Englandand Wales. Nowadays we refer to each other’s decisions with a frequency whichwould have been hardly imaginable even twenty years ago. And we do try to beconsistent where possible’.

9. Hugh Laddie, “Kirin Amgen - The End of Equivalents in England?,” Internatio-nal Review of Intellectual Property and Competition Law, 2009, 3–38 on how the benefitssecured by the creation of a central granting authority are undermined if the grantedpatents vary in their effects in the different Member States of the EPC.

10. For a set of approximately 92,000 European biotechnology patents filed from1978 to 2014, patent holders filed with an average of 8 different patent offices com-pared to 6 for the all patent applications (own computation based on OECD PatentQuality Indicators database, 2015, with biotechnology classified in tech field 15).

11. See Section 4.1 for details and references.

parallel litigation : the case of biotechnology 85

and definitions and thresholds in validity questions were more re-strictive when comparing the arguments by courts in both the UKand abroad.

In the single European market, infringement and revocation in mul-tiple countries of a centrally granted patent are therefore subject todivergent outcomes. European policy makers are not unaware of thecurrent heterogeneity prevailing in patent disputes and of its costs12.In response, a new supra-national court will enter into effect in co-ming years. The Unified Patent Court shall harmonize patent enfor-cement by providing for a one-stop litigation forum for patent dis-putes. In an effort to bring greater commonality to patent litigation,a single court shall, in principle, deal with civil litigation relating toEuropean patents. In its second Section (Section 4.2), the paper inves-tigates whether the Unified Patent Court has the potential to achieveits ambition of countering parallel and duplicate litigation13. Tempo-rary and persistent obstacles to a uniform patent litigation system areidentified and analyzed. The hope of resolving the issue of duplicatelitigation through unified patent enforcement can succeed only parti-ally; there will remain potential for dissident decisions over Europeanpatents.

In its final Section (Section 4.3), the paper introduces a game theo-retical model demonstrating that no matter the level of harmoniza-tion, parallel litigation will exist. The current intuition is that undera harmonized set of rules parties will litigate in one forum only, asidentical or similar rules are applied by the courts. The stronger theharmonization, the less likely it is for courts to deviate from prece-dents issued in third countries. The model demonstrates that, on thecontrary, it is often optimal for rational patent holders not to settleinitially, but rather to proceed to court decision, independently of the

12. Dietmar Harhoff, “Economic Cost-Benefit Analysis of a Unified and IntegratedEuropean Patent Litigation System,” Final Report, Tender No. Markt/2008/06/D, 2009,estimated that the total private savings from having access to a unified patent courtwould range between EUR 147 to 289 million annually.

13. See, e.g., Roberto Romandini and Alexander Klicznik, “The Territoriality Prin-ciple and Transnational Use of Patented Inventions – The Wider Reach of a UnitaryPatent and the Role of the CJEU,” International Review of Intellectual Property and Com-petition Law 44, no. 5 (2013): 524–540 and Matthias Brandi-Dohrn, “Some CriticalObservations on Competence and Procedure of the Unified Patent Court,” Internati-onal Review of Intellectual Property and Competition Law 43 (2012): 372–389.

86 parallel litigation : the case of biotechnology

level of harmonization of substantial laws. When litigation in twocountries or more is available, the expected utility calculation of thepatent holder is European, not national. The trial becomes sequentialin nature, in that the patent holder can grant licenses for the diffe-rent countries even after a first trial. Under our model, the patentholder will often prefer the uncertainty of a trial to the certainty oflow royalty rates. This means that under the current framework, theaim of homogeneity cannot be reached with mere harmonization ofsubstantial legal rules.

4.1 biotechnology before the courts

Let us take the example of a biotechnology firm operating within theEuropean market with mission to develop antibody drug conjugates.When patenting its technology, the company must decide to file itsapplication with the European Patent Office (EPO), the national pa-tent offices or the Patent Cooperation Treaty. When selecting the EPO,the European patent will be subject to a uniform examination system,but grant a bundle of national rights in a country-by-country enforce-ment regime. In parallel, during a nine-month period, the oppositionsystem before the EPO enables third parties to challenge the validityof the European patent with effect for all states. The company musttherefore decide whether or not it will request protection for a de-termined country or save the costs of the yearly fee – at the risk ofbeing exposed to a concurrent production in said state14. At a laterstage, the company faces significant challenges in determining notonly whether third parties violate its patent, but because of the Eu-ropean architecture, also once the violation is detected, in enforcingits rights. Since national court decisions have effect for the nationalpart of the patent only, the decision of suing must take the costs oflitigation in several states into account15. While the main objective ofthe framers of the European Patent Convention was the creation of

14. The costs of validation for European patents are at least five times more ex-pensive than their US counterparts, see Mejer and van Pottelsberghe de la Potterie,“Economic Incongruities in the European Patent System.”

15. Undesirable effects of parallel litigation include the added financial burdenof proceeding in different courts, the consumption of judicial resources in multiplejurisdictions and inconsistent decisions, see Yoshimasa Furata, “International Parallel

4.1 biotechnology before the courts 87

uniform substantial patent laws16, under national patent law harmo-nized by the Strasbourg Convention of 1963 and the European PatentConvention of 1973

17, national divergent interpretations remain pos-sible18. As a consequence of the lack of an integrated jurisdiction,differences in the judicial interpretation of the patent scope or of thepatentability threshold may arise, leading to uncertainty for the pa-tent holders, potential infringers and users, and innovators19.

Patents litigated before several courts in Europe demonstrate the ex-tent of harmonization in practice. By targeting disputes that resultedin judgments on the merits by courts of two or more countries, thesuccess of the harmonization policy aim can be assessed. The enfor-cement of biotechnology patents was selected to study the divergen-ces that may arise before European courts. Biotechnology was cho-sen for three reasons. Firstly, preference was given to a technologyfor which patents are of particular importance. Biotechnology is ischaracterized by very high development costs for new products and

Litigation: Disposition of Duplicative Civil Proceedings in the United States andJapan,” Pacific Rim Law & Policy 1 (1995): 1–58.

16. See Laddie, “Kirin Amgen - The End of Equivalents in England?”: ‘[t]he majorobjective of the European Patent Convention (EPC) was to put in place a substanti-ally uniform patent code for Europe’.

17. Mainly by the Strasbourg Convention on the Unification of Certain Points ofSubstantive Law of Patents for Invention; the EPC then almost literally incorporatedthe Strasbourg Convention, see Antonina B. Engelbrekt, “Dilemmas of Governancein Multilevel European Patent System,” in National Developments in the Intersection ofIPC and Competition Law, ed. Hans Henrik Lidgard (Hart Publishing, 2011), p. 42.

18. Getchen Bender, “Clash of the Titans: The Territoriality of Patent Law vs. theEuropean Union,” IDEA: The Journal of Law and Technology 40 (2000): 49–82 on the in-herent conflict between national patent law and the European Patent Convention andVincenzo Di Cataldo, “From the European Patent to a Community Patent,” ColumbiaJournal of European Law 8 (2002): 19–36 characterizing the situation as ‘deadlock’.

19. Because of the low likelihood of a patent to end up in litigation – between 1-3%in most patent systems – the signals arising from precedent decisions are important,see Harhoff, “Economic Cost-Benefit Analysis of a Unified and Integrated EuropeanPatent Litigation System,” p. 14; Mejer and van Pottelsberghe de la Potterie, “Eco-nomic Incongruities in the European Patent System” and Dongwook Chun, “PatentLaw Harmonization in the Age of Globalization: The Necessity and Strategy for aPragmatic Outcome,” Journal of Patent Trademark Office Society 93, no. 2 (2011): 127–166: ‘competitors cannot know where to focus their research and development in-vestments until they know precisely what a patent applicant has been able to claimas its inventive territory. Accordingly, companies in this situation may make fewerinvestments in innovation that are potentially misdirected and wasteful’.

88 parallel litigation : the case of biotechnology

processes while the costs of imitation by reverse engineering is rela-tively low20. Second, a ‘cross-border’ technology was sought, one forwhich protection and enforcement would be requested across multi-ples European countries, rather than in isolated nations only21. Thi-rdly, a technology was searched for which such a fragmented enforce-ment system is particularly detrimental for innovation22. Consideringthat an estimate of over 80% of biotechnology firms are small- andmedium-sized enterprises – and that small companies face substanti-ally higher marginal costs to protect their patents than large firms23 –one can hypothesize that this is the case for biotechnology.

4.1.1 Data Collection

To study patent litigation in the field of biotechnology, decisions weregathered from Darts-IP, a leading database of patent lawsuits24. A to-tal of 127 individual patent litigation decisions over 47 patents ren-dered between 2000 and 2013 were identified (see Figure 4.1). Thesearch was conducted by manually reviewing all patents from the

20. Genetic Inventions, Intellectual Property Rights and Licensing Practices, (OECD,2002) states that ‘[i]n no other fields is the relationship between patent protectionand the incentives to innovate so strong’. This goes with the exception of diagnostickits and laboratory-developed tests, where development costs are not substantial,see Health Report of the Secretary’s Advisory Committee on Genetics and Society,Gene Patents and Licensing Practices and Their Impact on Patient Access to Genetic Tests,(April 2010), p. 4.

21. As mentioned previously, biotechnology patents are filed with above averagenumber of patent offices. Further, of the biotechnology patents decisions we analyze,over half of the judgments were the fruit of duplicate litigation.

22. The Considerations of the Agreement on a Unified Patent Court state that ‘thefragmented market for patents and the significant variations between national courtsystems are detrimental for innovation, in particular for small and medium sizedenterprises which have difficulties to enforce their patents and to defend themselvesagainst unfounded claims and claims relating to patents which should be revoked’.

23. Editorial, Who Speaks for Small Biotech?, 25 (Nature Biotechnology, 2007), 693 andJean O. Lanjouw and Mark Schankerman, “Protecting Intellectual Property Rights:Are Small Firms Handicapped?,” Journal of Law and Economics 57 (2004): 45–74.

24. Stuart J. Graham and Nicolas Van Zeebroeck, “Comparing Patent LitigationAcross Europe: A First Look,” Stanford Technology Law Review 17 (2014): 655–708

reporting that Darts-IP estimates its coverage to 90% in France and the Netherlands;60-80% in Belgium, Italy, Spain and the UK and 50% in Germany (2000-2009).

4.1 biotechnology before the courts 89

IPC classes representing biotechnology25 and whether they were fal-ling under the biotechnology definition by the Organization for Eco-nomic Co-operation and Development (OECD) , i.e. ‘the applicationof science and technology to living organisms, as well as parts, pro-ducts and models thereof, to alter living or non-living materials forthe production of knowledge, goods and services’26.

Figure 4.1: Biotechnology Litigation Geographical Coverage.

Decisions from Germany, France, Belgium, Italy, Spain, the Netherlands and the UK

were identified.

4.1.2 Methodology

Biotechnology Litigation. The 127 patent infringement and invali-dation decisions were clustered in 74 cases according to their levelof jurisdiction (first instance, intermediate appeal level and supremecourt)27. Overall, the disputes concerned a set of 47 patents. We hand-coded all decisions over various variables, categorizing them acrossnumerous dimensions, first general data (e.g., country, citation, date)and nature of the ruling rendered (infringement vs. invalidity). A

25. IPC subclasses A01H1/00, A01H4/00, A61K38/00, A61K39/00, A61K48/00,C02F3/34, C07G (11/00, 13/00, 15/00), C07K (4/00, 14/00, 16/00, 17/00, 19/00),C12M, C12N, C12P, C12Q, C12 R, G01N27/327, G01N33/ (53*, 54*, 55*, 57*, 68, 74,76, 78, 88, 92). The search was lead twice, in October 2015 and 2016.

26. OECD, “A Framework for Biotechnology Statistics,” 2005, we derive the IPCcodes from the same source.

27. For comparison purpose, in the bifurcated system in Germany, infringementand revocation lawsuits are joined when they take place amongst the same partiesand in a timespan of 2 years.

90 parallel litigation : the case of biotechnology

‘definitive win’ was reported if the patent holder could enforce itsalleged infringement claim before the courts, i.e. at least one of theclaims was found to be infringed and that claim was, if challenged,upheld as valid, or in revocation proceedings, at least one claim wasfound valid.

Duplicate Litigation . About a third of the patents in our sample – 16

out of the 47 biotechnological inventions – were subject to decisionson the merits in two or more countries. We define duplicates as de-cisions between the same parties over the same patent before two ormore courts of different countries. These disputes represented a totalof 66 of the total 127 judgments. Hence, more than half of all rulingswere the fruit of duplicate proceedings. All decisions over paralleldisputes were analyzed in order to identify contradictions on pointsof law. In particular, we focused on divergences over the scope of pro-tection of patent claims, patentability standards and the definition ofthe person skilled in the art. By targeting duplicates, we try to over-come the selection effect limitation. In a perfectly harmonized patentsystem, the parties may litigate in one country and assume that courtsin other countries would likewise come to the same conclusion ratherthan modify the first judgment. A first court decision would reducethe divergence of the parties’ expectations or their information asym-metry, which both would lead to a higher rate of settlement. The factthat certain disputes are subject to multiple litigation seems to indi-cate that the expectations of the parties are such as the outcome mayvary across country.

Limitations. In the analysis of general – and patent – litigation, themost important bias results from the fact that the vast majority oflawsuits settle before any court ruling. Only lawsuits reaching the de-cision stage are captured in our sample, and those are not a randomsample of the mass of underlying disputes. It is therefore unclearwhether inferences can be drawn from the outcome of the trials28. Se-cond, by limiting the study to biotechnological patents, a second se-lection bias potentially impacts our findings. We cannot exclude that

28. See generally Richard A. Posner, Economic Analysis of Law (Wolters Kluwer,2007), p. 765; and regarding patent litigation John R. Allison, Mark A. Lemley, andDavid L. Schwartz, “Our Divided Patent System,” University of Chicago Law Review82, no. 3 (2015): p. 1125-1133.

4.1 biotechnology before the courts 91

the differences we identify reflect the entire universe of patents of alltechnologies, or are limited to biotechnology. Finally, even decisionsover identical patents bear differences. Since litigation is sequential,the litigants might allocate different level of resources to the cases;further, cognitive fallacies and biases might affect the two or morejudges in different ways. Despite these limitations, the study offersan understanding of what the direct consequences of the lack of har-monization can represent for legal disputes for a particular techno-logy. By targeting duplicate disputes, we are in the unique positionto compare different tried cases as close as they can get.

4.1.3 Current Divide

After presenting high-level patent litigation trends patent at the coun-try level, we illustrate how the lack of harmonization impacts a vari-ety of legal standards crucial for patent holders.

4.1.3.1 Overview

At the national level, we observe in Table 4.1 that the overall patent li-tigation win rate appears to be highest in Germany, where on average60% of the claims have been successful. Both on the individual infrin-gement and invalidity level, the patent holder had a very high levelof success of approximately two-third. In line with the the inventor-friendly vs. third-parties friendly discourse, as well as with prior em-pirical work by Cremers et al. (2017), patent protection appears to belowest in the United Kingdom, where less than one of three patentprevailed at trial. Even after separating infringement from invalidityoutcomes, patent holders under-perform when litigating in the UK.

92 parallel litigation : the case of biotechnology

Table 4.1: Litigation outcome across Europe, by Country

Litigation Outcome for Patent Holder

Country Overall Infringement Invalidity

Frequency Win % Frequency Win % Frequency Win %

Germany 30 60.0 23 60.9 14 71.4

Netherlands 16 37.5 7 57.1 14 50.0

UK 13 30.8 9 44.4 13 38.5

France 7 42.9 5 60.0 4 50.0

Italy 5 60.0 2 100.0 5 60.0

Spain 2 0.0 1 0.0 1 0.0

Belgium 1 0.0 1 0.0 1 100.0

Total 74 45.9 48 56.3 52 53.8

Description: Litigation outcome in the 74 litigation cases across Europe, by state. The summary of all

case outcomes can be found in Appendix Figure A4.1.

4.1.3.2 Duplicate Litigation and The Division of Europe

The qualitative analysis of biotechnological inventions litigation con-veys a picture of a divided Europe; with the United Kingdom on theone side and Germany and the Netherlands on the other. There weredifferences in the interpretation of patent claims, leading to a narro-wer scope of the protection of patents in the United Kingdom. In theassessment of arguments behind the validity of patents, judges in theUnited Kingdom were more restrictive in their definition of the per-son skilled in the art, and in the threshold of obviousness, novelty,and sufficient disclosure.

4.1 biotechnology before the courts 93

A. Differences in Scope of Protection of Patents

Principles. Patent claims establish the scope of protection granted toan issued patent29. Infringement is judged by comparing the activityof the alleged infringer to the scope of protection as defined by theclaim30. To do so, patent claims must be interpreted by the courts.Because of the discretion courts possess in their interpretation, thereis room for dissidence across countries. Before the entry into forceof the EPC, two opposite methodologies prevailed in Europe. Accor-ding to ‘peripheral claiming’, a more restrictive methodology appliedin the UK and in Switzerland, claims define both the invention andlimit it (‘what is not claimed is disclaimed’). ‘Central claiming’, howe-ver, that was in use in Germany and the Netherlands, states that theclaims define the invention only, being the starting point of the scopeof protection31. Article 69 EPC hoped to bridge the gap by providingthat patent claims determined the extent of the protection while thedescription and drawings shall be used to inteprete the claims, fol-lowing the peripheral claiming approach. But the Protocole on theInterpretation of Article 69 EPC adopted at the Munich DiplomaticConference in 1973 decreed that the claims were neither to be limitedto the literal meaning, neither as a guideline only, but that the trackshall be a position between the extremes32.

29. Art. 69 para. 1 EPC specifies more precisely that the extent of protection isdetermined by the claims, but that the description and drawings shall be used tointerpret the claims.

30. Christopher A. Cotropia, “Patent Claim Interpretation Methodologies andTheir Claim Scope Paradigm,” William & Mary Law Review 47, no. 1 (2005): 49–132

and Dan L. Burk and Mark A. Lemley, “Fence Posts or Sign Posts: Rethinking PatentClaim Construction,” University of Pennsylvania Law Review 157 (2009): 1743–1799.

31. For a summary of these differences, see scholarship by Toshiko Takenaka, Inter-preting Patent Claims: The United States, Germany and Japan (Wiley-VCH, 1995); DonaldS. Chisum, “Common Law and Civil Law Approaches to Patent Claim Interpreta-tion: "Fence Posts" and "Sign Posts",” in Intellectual Property in the New Millennium:Essays in Honour of William R. Cornish, ed. David Vaver and Lionel Bently (CambridgeUniversity Press, 2004); Burk and Lemley, “Fence Posts or Sign Posts: Rethinking Pa-tent Claim Construction”; Nicholas Pumfrey et al., “The Doctrine of Equivalents inVarious Patent Regimes – Does Anybody Have it Right?,” Yale Journal of Law andTechnology 11, no. 1 (2009): 262–308 and the judicial decision Kirin-Amgen v. Roche,Transkaryotic Therapies et al., House of Lords, [2004] UKHL 46, October 21, 2004.

32. As such, the article defines a position striving a balance between protectionfor the patentee, reflecting German and Dutch approaches, and certainty for thirdparties, the UK focus.

94 parallel litigation : the case of biotechnology

Caselaw. A dispute over a blockbuster biotechnology drug exempli-fies the differences in claims construction that still exist in Europe.MedImmune and Medical Research Council alleged that Novartishad infringed its two patents jointly owned by sales of the pharma-ceutical product ranibizumab sold under the trademark Lucentis. Lu-centis is a monoclonal antibody fragment used for the treatment ofmacular degeneration of the eye produced in the US by Genentechand sold by Novartis in Europe with reported sales of $ 2.4 billionin 2014

33. The two patents included claims directed to phagemid sy-stems respectively pure phages displaying at their surface a popula-tion of binding molecules having a range of binding specificities34.While the English courts found the patents not to have been infrin-ged35, the German courts concluded that one of the patents was36.When diving into the argumentation of the court, divergences in theconstruction of claims emerge. Novartis contended that the skilledreader would understood the words ‘a range of binding specificities’to encompass a range of specificities to different antigens, whereasMedImmune argued that a range of specificities to a single antigenshould be understand under this claim. The UK judges followed thenarrower claim construction of Novartis, while the German judgessupported the construction of MedImmune37. Similarly, the wording‘by fusion with a gene III protein‘ embraced only the complete geneIII protein, or at least substantially complete gene III protein, for theUK but also parts of the gene III protein for the German court. Gen-entech having used fragments of gene III proteins to produce fusionswith antibody fragments, ranibizumab was only decreed to be madeby a process according to the claims in Germany, not in the UK. Ac-

33. Novartis top 20 product sales can be retrieved under: https://www.novartis.com/investors/financial-data/product-sales.

34. European Patent EP0774511 and EP2055777, filed the July 10, 1991.35. MedImmune et Medical Research Council v. Novartis, High Court of Justice,

[2011] EWHC 1669 (Pat), July 5, 2011 and Court of Appeal, [2012] EWCA Civ 1234,December 10, 2012.

36. For patent EP0774511, see MedImmune v. Novartis, Landgericht Düsseldorf, 4aO 47/10, November 10, 2011 and for patent EP2055777, see Landgericht Düsseldorf,4a O 143/10 , November 10, 2011.

37. MedImmune et Medical Research Council v. Novartis, High Court of Justice,[2011] EWHC 1669 (Pat), July 5, 2011, the court based its construction on the argu-ment of the choice of the wording ‘specificity’ rather than ‘affinity’ in the claim andtheir consistently distinct use in the patent.

4.1 biotechnology before the courts 95

cordingly, due to differences in patent claim construction, the scope ofprotection of patent rights was significantly narrower in the UK, andlimited to the claims. This exemplifies that despite the harmonizationefforts, judges appear to construct the claims in line with previoustraditions38.

B. Differences in the Person Skilled in the Art

Principles. The definition of the person skilled in the art is centralthroughout the lifecycle of a patent. Not only does it determine thevalidity of a patent but it also impacts infringement analyses39. Theinventiveness of an invention is only given should it be non-obviousto a skilled person40. Similarly, the sufficiency of a disclosure is tobe assessed from the standpoint of such person41. Variances in thedefinition of the skilled skilled person across Europe imply nationaldifferences in thresholds of patentability. Additional specialization ofthe skilled person increases inventiveness thresholds and decreasesdisclosure requirements. Since the level of knowledge and skill inthe field of the addressee is higher, alleged inventions become moreobvious and easier to carry out as described in the disclosure. Interms of infringement analysis, the claims are often analyzed from

38. See Andrew Rudge, Guide to European Patents (Thomson West, 2012), §. 6.1stating that ‘[n]ot only is this parallel litigation expensive, but the outcome also oftenvaries from country to country, depending on the way the same claims and prior artare interpreted according to different legal traditions’; see Dominique Guellec andBruno van Pottelsberghe, The Economics of the European Patent System: IP Policy forInnovation and Competition (Oxford Scholarship Online, 2007), p. 221: ‘common rulesthat are applied in similar, but not necessarily identical ways in all countries’.

39. Jonathan Darrow, “The Neglected Dimension of Patent Law’s PHOSITA Stan-dard,” Harvard Journal of Law & Technology 23, no. 1 (2009): 227–258; Rebecca Eisen-berg, “Obvious to Whom? Evaluating Inventions from the Perspective of PHOSITA,”Berkeley Technology Law Journal 19 (3 2004): 885–906; and Gregory Mandel, “The Non-Obvious Problem: How the Indeterminate Nonobviousness Standard Produces Ex-cessive Patent Grants,” UC Davis Law Review 42 (2008): 57–128.

40. Article 56 EPC; Article 33 para. 3 Patent Cooperation Treaty : ’[..] a claimedinvention shall be considered to involve an inventive step if, having regard to theprior art as defined in the Regulations, it is not, at the prescribed relevant date,obvious to a person skilled in the art’.

41. See Article 83 EPC and Part F Chapter III Section 1 Guidelines for Examinationin the European Patent Office stating that ‘objection of lack of sufficient disclosurepresupposes that there are serious doubts, substantiated by verifiable facts’.

96 parallel litigation : the case of biotechnology

the viewpoint of the skilled person by national courts42. As statedby the honorable Lord Hoffman: ‘[t]he question is always what theperson skilled in the art would have understood the patentee to beusing the language of the claim to mean.’43.

Caselaw. In two disputes, the level of specialization of the skilledperson was higher in the UK than abroad. In particular, judges havestarted to identify the person skilled in the art as a team, reflecting therise of collaborative research in science. In a set of legal disputes bet-ween Amgen and Roche over patents covering anaemia drug erythro-poietin by recombinant DNA technology, judges in the UK identifiedthe skilled person to be a team of three people with PhD’s and se-veral years experience in gene technology, molecular biology and cellbiology44. In Germany, the addressee was one scientist with PhD andseveral years experience in gene technology45. In the second instance,the person skilled in the art in the UK was a team of five scientists,including a clinician, a toxicologist, a synthetic chemist and two ana-lytical chemist with different expertise, compared to a single scientistin the Netherlands46. The patent related to a Copolymer-1, which isa synthetic polypeptide which has been suggested as a potential ther-

42. According to Article 1 Protocol on the Interpretation of Article 69 EPC of Oct.5, 1973 as revised by the Act revising the EPC of Nov. 29, 2000 ‘[n]or should [Article69] be taken to mean that the claims serve only as a guideline and that the actualprotection conferred may extend to what, from a consideration of the descriptionand drawings by a person skilled in the art, the patent proprietor has contemplated.’

43. Kirin-Amgen, Inc. v. Hoechst Marion Roussel, Ltd., (2004) UKHL 46 34.44. Dispute over patent EP0411678, see Kirin-Amgen v. Roche, Transkaryotic Ther-

apies et al., High Court of Justice, [2001] EWHC Patents 433, April 11, 2001; Courtof Appeal, [2002] EWCA Civ 1096, July 31, 2002; House of Lords, [2004] UKHL 46,October 21, 2004; and Amgen v. Genetics Institute, 99/3046, January 17, 2001 andalso Nature Reviews Drug Discovery, 3, 990, 2004.

45. Anonymous v. Genetics Institute, Bundespatentgericht, 3 Ni 4/03 (EU), May 25,2004; see Dennis Crouch, “THOSITA: Obvious to a Team Having Ordinary Skill inthe Art,” Patently-O, October 2011, for a note on the construct of the skilled team.

46. Dispute over patent EP0762888, filed the May 23, 1995, see Mylan v. Yeda Rese-arch and Development and Teva Pharmaceutical Industries, Rechtbank den Haag,C/09/431155 /HA ZA 12-1339, October 2, 2013; Generics UK v. Yeda Researchand Development and Teva Pharmaceutical Industries, High Court of Justice, [2012]EWHC 1848 (Pat), July 11, 2012; and Court of Appeal, [2013] EWCA Civ 925, July 29,2013. The identity of the team, however, was restricted by the High Court of Justiceas being a neurologist with experience in the treatment of multiple sclerosis havingseveral years of experience in the pharmaceutical industry; while Yeda’s formulationincluded additional skills in e.g. neuroimmunology.

4.1 biotechnology before the courts 97

apeutic agent for multiple sclerosis. Teams have arguably an overallbroader depth of knowledge – leading to higher thresholds in inven-tiveness47. While these different definitions not necessarily lead todifferences in the outcome of disputes, they illustrate the more subtlevariations that might hinder harmonization efforts overall.

C. Differences in Inventive Step Threshold

Principles. An inventive step is required for an invention to be paten-table under European Patent Convention48. Different methodologiesare applied to assess inventive step. The Board of Appeal of the EPOutilizes the ‘problem-and-solution approach’, identifying the closestprior art item, establishing the objective technical problem to be sol-ved and considering whether or not the claimed invention, startingfrom the closest prior art and the objective technical problem, wouldhave been obvious to the skilled person49. In the UK, inventive stepfollows the ‘Windsurfing/Pozzoli’ test. Accordingly, the steps are tofirst identify the notional skilled person and its relevant common ge-neral knowledge; then identify or construe the inventive concept ofthe claim and identify what, if any, differences exist between priorart and the inventive concept of the claim50. Germany, on the otherhand, does not rely on a formalized approach51. Courts shall analyzewhether the skilled person can find the claimed subject matter usinghis knowledge and skill on the basis of the prior art52. Traditionally,

47. Dennis Crouch, “Person(s) Skilled in the Art: Should the Now Established Mo-del of Team-Based Inventing Impact the Obviousness Analysis?,” Patently-O, May2011, with a similar analysis.

48. Art. 56 EPC. The purpose of the requirement is to complement the novelty barand ensure that inventions represent sufficient advances in technology, see also AmyNelson, “Obviousness or Inventive Step as Applied to Nucleic Acid Molecules: AGlobal Perspective,” North Caroline Journal of Law & Technology 6, no. 1 (2004): 1–40.

49. Part G Chapter VII Section 5 Guidelines for Examination in the European Pa-tent Office; Li Xiang et al., “A Comparative Analysis of the Inventive Step Standardin the EPO, SIPO and USPTO,” Journal of Intellectual Property Law & Practice 8, no. 7

(2013): 539–545.50. Lord Justice Jacob in Pozzoli Spa v BDMO SA and Another [2007] BusLR D117,

[2007] EWCA Civ 588, [2007] FSR 37.51. AIPPI Summary Report on The Patentability Criteria of Inventive Step under

https://aippi.org/download/commitees/217/SR217English.pdf.52. See Rainer Moufang, “Kapitel § 4,” in Patentgesetz mit Europäischem Patentübere-

inkommen, ed. Rainer Schulte (Carl Heymanns, 2013).

98 parallel litigation : the case of biotechnology

obviousness has been the most frequent basis for invalidation, beingraised in about three-quarter of all invalidity actions in Europe53.

Caselaw. The lawsuits of study reveal one illustration where despiteidentical formulation of the skilled person, courts established set dif-ferent thresholds for obviousness in Germany and in the UK. In Me-dImmune v. Novartis, the courts diverged as to the prior art requiredto establish obviousness. The invention at stake was related to an-tibody phage display – an invention which revolutionized antibodydrug discovery. The High Court of Justice in the UK found that phagedisplay was not obvious towards antigen display in 1990 but that itwas after a talk by Prof. Smith displayed the reasonable chances ofsuccess of the technique54. The Federal Patent Court in Germany stop-ped at the first stage and found obviousness of phage towards antigendisplay55. In these two instances, the identically defined notional skil-led inventor would find antibody phage display obvious respectivelynon-obvious towards antigen display.

D. Differences in Sufficiency of Disclosure

Principles. A patent must disclose the invention in a manner suffi-ciently clear and complete for it to be carried out by a person skilledin the art56. The required disclosure is that any feature essential tocarry out the invention are given, and the skilled person shall use hiscommon general knowledge to perform the invention without undueburden and without needing inventive skill.57. US courts frequentlyrevoke broad claims relating to biotechnology, especially proteins, for

53. Graham and Van Zeebroeck, “Comparing Patent Litigation Across Europe: AFirst Look,” p. 695; see Gloria Koenig, Patent Invalidity: A Statistical and SubstantiveAnalysis (Clark Boardman, 1974) for the US.

54. MedImmune et Medical Research Council v. Novartis, High Court of Justice,[2011] EWHC 1669 (Pat), July 5, 2011 and Court of Appeal, [2012] EWCA Civ 1234,December 10, 2012.

55. Novartis v. MedImmune et Medical Research Council, Bundespatentgericht, 3

Ni 5/10 (EU), January 24, 2012.56. Art. 83 EPC; Alan Devlin, “The Misunderstood Function of Disclosure in Patent

Law,” Harvard Journal of Law & Technology 23, no. 2 (2010): 401–446 discussing the‘incentive to disclose’ rationale of the patent system as normative justification for apatent system.

57. EPO Board of Appeal T 0727/95 (Cellulose) of May 21, 1999.

4.1 biotechnology before the courts 99

lack of sufficient disclosure58. Illustratively, sufficient disclosure is ge-nerally fulfilled when the claims define an antibody by sequence orby target, whereas a definition in term of activity may lead to difficul-ties59.

Caselaw. Patents might be found sufficiently disclosed by the court ofone country but not of another. The blockbuster patent the EuropeanPatent Office had granted Amgen covering erythropoietin (EPO) byrecombinant DNA technology illustrates this60. The claims definedan extremely broad patent including product claims to purified EPOitself61. The validity of the patent was examined by the courts in theUK, Germany and the Netherlands. Britain’s highest court, the Houseof Lords, ruled that the product claims were invalid because theywere too broad62. Amgen had developed a groundbreaking processfor making EPO and its analogues, instead it sought to patent the pro-tein EPO itself. The true construction of the claim, argued the court,would mean that the invention would cover any method of makingEPO. This led, however, to a classic breadth of claim objection, sincethe specification would not cover the specific process itself. In Ger-many, the patent was also nullified for lack of sufficient disclosure63.Similarly, the courts held that the insufficient presence of essentialfeatures induced difficulties of reproducibility (due to the broadnessof the claims); and the invention was deemed insufficiently disclosed.In the Netherlands, however, the patent was found sufficient in dis-closure64. When the skilled person used its common knowledge, con-

58. Sarfaraz K. Niazi and Justin L. Brown, Fundamentals of Modern Bioprocessing(CRC Press, 2015), p. 688.

59. In the latter case, the antibody should be defined by its function and at least oneexample provided of an antibody with such function, see Louise Holliday, “PatentingAntibodies in Europe,” MAbs 1, no. 4 (2009): 385–386.

60. EP0148605, filed December 12, 1984.61. Luigi Palombi, Gene Cartels: Biotech Patents in the Age of Free Trade (Edward

Edgar, 2009).62. Kirin-Amgen v. Roche, Transkaryotic Therapies et al., High Court of Justice,

[2001] EWHC Patents 433, April 11, 2001; Court of Appeal, [2002] EWCA Civ 1096,July 31, 2002; and House of Lords, [2004] UKHL 46, October 21, 2004.

63. Anonymous v. Amgen, Bundespatentgericht, 3 Ni 34/99 (EU), December 14,2000; details on the patent can be gained with Palombi, Gene Cartels: Biotech Patentsin the Age of Free Trade.

64. Amgen v. Boehringer Mannheim, Gerechtshof te ’s-Gravenhage, 96/581, Janu-ary 27, 2000.

100 parallel litigation : the case of biotechnology

tended the Dutch court, reproducibility was given. In its arguments,the court held that it considered previous jurisprudence on the dis-closure of chemical compounds by the EPO Board of Appeal as falseas it led to incorrect outcomes65. The common general knowledge ofthe skilled person was extended to include information which canonly be obtained after a comprehensive search. Under this enlargedsetting, the claimed invention was sufficiently disclosed.

E. Differences in Novelty

Principles. To be patentable, an invention must satisfy the require-ment of novelty, that is it shall not form part of the state of the art66.When the invention is compared with individual items of prior art,the examiner or judge must establish the state of the art and its con-tent and determine whether or not the invention differs from it. Thestate of the art, in return, is everything made available to the pu-blic by means of a written or oral description, by use, or in any ot-her way, before the filing of the patent application67. The disclosureof subject-matter must be enabling, that is such that the skilled per-son can reproduce that subject-matter using common general know-ledge68. To anticipate a patent, the subject matter must represent anenabled disclosure of the technical teaching of the patent at issue69.Despite the harmonization, differences on the methods to determinethe claim scope for assessing novelty exist at a national level70. Forinstance, in the UK, the scope of claims is construed consistently for

65. EPO Board of Appeal, T 206/83 OJ EPO 1987, 5-13. The decisions holding apatent valid in opposition procedures have no binding effect for national courts, andnational courts may correct the ‘errors’ of the EPO. Lord Justice Jacob phrases thesituation as ‘[w]e follow any principle of law clearly laid down by [the TechnicalBoard of Appeal of the EPO], only reserving the right to differ if we are sure thatthe commodore is steering the fleet on to the rocks.’ but not ‘questions of degreeover which judges may legitimately differ’, see Eli Lilly v. Human Genome Sciences,[2010] EWCA Civ 33, February 9, 2010.

66. Article 54 para. 1 and para. 2 EPC.67. Part G, Chapter VI Section 1 Guidelines for Examination in the European Patent

Office.68. Part G, Chapter VI Section 2 Guidelines for Examination in the European Patent

Office and EPO Board of Appeal T 26/85, September 20, 1988.69. Paul England, “Novelty of Patents in Europe and the UPC,” Journal of Intellec-

tual Property Law & Practice 12, no. 9 (2017): 739–746.70. See England, “Novelty of Patents in Europe and the UPC” for a review of the

differences in the application of the novelty requirement across Europe.

4.1 biotechnology before the courts 101

the purpose of the infringement and validity. This can give rise to theapplication of the doctrine of equivalents to prior art documents71.Infringers may accordingly build a squeeze defense, contending thatif the patent infringes then it covers the prior art item and is thereforeanticipated72. In Germany, where infringement and novelty are triedseparately, squeeze argument cannot be deployed.

Caselaw. In a dispute over methods of production of erythropoie-tin (EPO) between Amgen and Roche, both parties sought to revokeeach other’s patent. The first case was fought over patent whose re-gistered proprietor was Amgen. The UK High Court of Justice, whenassessing the novelty of such claim, drew no difference between theEPO made through the claimed method of recombinant EPO andEPO produced through the previously disclosed method of urinaryEPO. As such, the claim was held been anticipated73. At earlier stage,the patent had in vain been opposed before the Board of Appeal ofthe EPO, where it was considered novel based on the same items ofprior art74. In his judgment, Lord Hoffman argued that the Board ofAppeal based its decision on different findings of facts, and wouldit had made the same, it would have rejected the claim. The seconddispute involved a Roche patent. Amgen contended it was anticipa-ted by its own patent. Roche submitted that its patent was novel tothe extent its teaching disclosed recombinant EPO that will always beglycosylated and fucosylated, and the teaching could be related to adeposited cell line. The High Court of Justice found that an Examplein the Amgen patent would, if carried out, contain clear directions forthe skilled man who would inevitably produce such rEPO that wasin part O-glycosylated. The patent was therefore found anticipated inthe UK75. The same claim was examined by the Federal Patent Court

71. Part C, Chapter IV Section 7 Guidelines for Examination in the European PatentOffice stating that ‘when considering novelty, it is not correct to interpret the teachingof a document as embracing well-known equivalents which are not disclosed in thedocument; this is a matter of obviousness’.

72. See Gillette Safety Razor v Anglo Trading Case (1913) 30 RPC 465.73. Kirin-Amgen Inc and others v. Hoechst Marion Roussel Limited and others,

House of Lords, [2004] UKHL 46, October 21, 2004.74. Amgen, Janssen et Hoechst v. Genetics Institute, Opposition Division EPO,

April 6, 1995 and Board of Appeal, T 0277/95, April 16, 1999.75. Nature Reviews Drug Discovery, 3, 990, 2004; Kirin-Amgen v. Roche, Transka-

ryotic Therapies et al. High Court of Justice, [2001] EWHC Patents 433, April 11,

102 parallel litigation : the case of biotechnology

in Germany76. There, the same claim was found novel. Roche prevai-led on the ground that the properties of the cell line used for EPOproduction differed to the one disclosed by the Amgen patent. WhileChinese hamster ovary cells were used in both instances, the additi-onal o-glycosylation taught by the patent was not disclosed. Whilecomparing the same claim with the same item of prior art, the twocourts therefore reached opposite conclusions.

4.2 the future

The Unified Patent Court, which will enter into force when at leastthirteen states – the number currently being sixteen – including thethree largest patent granting states France, Germany and the UK– willratify the Agreement on a Unified Patent Court, shall remedy to thesituation of heterogeneity in Europe77. The Court will have exclusivejurisdiction over all infringement and invalidation proceedings forEuropean patents (including the new Unitary Patent) throughout the25 European territories involved (all EU Member States except Spain,Poland and Croatia)78. The UPC will have a Court of First Instance

2001; Court of Appeal, [2002] EWCA Civ 1096, July 31, 2002; and House of Lords,[2004] UKHL 46, October 21, 2004.

76. Amgen v. Genetics Institute, 99/3046, January 17, 2001; Anonymous v. GeneticsInstitute, Bundespatentgericht, 3 Ni 4/03 (EU) , May 25, 2004.

77. The impact of Brexit and the capacity of the UK to remain a member ofthe UPC after its (potential) withdrawal from the EU are widely debated topics,see, e.g., Ansgar Ohly and Rudolf Streinz, “Can the UK Stay in the UPC SystemAfter Brexit?,” Journal of Intellectual Property Law & Practice 12, no. 3 (2017): 245–258; Matthias Lamping and Hanns Ullrich, “The Impact of Brexit on Unitary Pa-tent Protection and its Court,” Max Planck Institute for Innovation & Competition Re-search Paper, nos. 18-20 (2018); and Winfried Tilmann, “The Harmonisation of In-validity and Scope of Protection Practice of the National Courts of EPC MemberStates,” International Review of Intellectual Property and Competition Law, 2006, 62–74.The UK ratified the UPC Agreement on April 26, 2018. In Germany, a constitutio-nal complaint pending in the German Federal Constitutional Court regarding theimplementation law passed by the German Parliament is causing delay to the ra-tification. The plaintiff contends a breach of the limits to surrender sovereigntythat are derived from the right to democracy (Art. 38 (1), clause 1, Basic Law),see http://patentblog.kluweriplaw.com/2017/08/16/upc-finally-some-news-fr

om-the-german-federal-constitutional-court.78. See Kevin P. Mahne, “A Unitary Patent and Unified Patent Court for the Euro-

pean Union: An Analysis of Europe’s Long Standing Attempt to Create a Supranati-

4.2 the future 103

(divided into local, regional and central divisions) and a Court of Ap-peal in Luxembourg. The Court of First Instance will have a CentralDivision located in Paris and sections in Munich and London. Withinthe Divisions, the cases will be attributed based on subject matterwith London inter alia specializing on Chemistry, Munich on Mecha-nical Engineering and Paris on Physics and Fixed Constructions79.This allocation will base on the WIPO International Patent Classifica-tions. Member State may set up local divisions, expected in Austria,Belgium, Denmark, England & Wales, France, Italy, Ireland and theNetherlands. Germany will be entitled to four local divisions plan-ned in Düsseldord, Mannheim, Munich and Hamburg. Further, regi-onal divisions are foreseen, one for Bulgaria, Greece, Cyprus, Roma-nia and Slovenia, another for Sweden, Estonia, Latvia and Lithuania(the Nordic-Baltic regional division)80. The Court of Appeal will hearappeals on points of law and facts against final decisions, decisionsterminating proceedings for one of the parties and a number of or-ders such as freezing orders or injunctions81. The unified court willfully respect the primacy of EU law, and accordingly follow the caselaw of the European Court of Justice82. The Unitary Patent, whichwill be granted as from the date the Unified Patent Court enters intoforce, will be a new type of European patent with validity in all 25

states83.

onal Patent System,” Journal of the Patent and Trademark Office Society 94 (2012): 162–191 for a historical perspective and Dimitris Xenos, “The European Unified PatentCourt: Assessment and Implications of the Federalisation of the Patent System inEurope,” SCRIPTed: A Journal of Law, Technology and Society 10, no. 2 (2013): 246–277

for a criticism of the lack of democratic process-making as the UPC establishes amonopolistic source of legal power.

79. See Annex II of the Unified Patent Court Agreement.80. See Art. 7 para. 3 Unified Patent Court Agreement and Impact of the Unitary

Patent Protection and the Unified Patent Court in the Nordic-Baltic Region, (Ministry ofJustice of Sweden, October 2014).

81. See Art. 73 para. 3 Unified Patent Court Agreement and Rule 220.1 of the DraftRules of Procedure of the UPC.

82. Article 21 UPC Agreement in relation with Art. 267 TFEU on preliminary ru-lings: ‘[d]ecisions of the Court of Justice of the European Union shall be binding onthe Court’.

83. Richard Pinckney, “Understanding the Transitional Provisions of theAgreement on the Unified Patent Court,” European Intellectual Property Review 37,no. 5 (2015): 268–277 and Alan Johnson and Luke Maunder, Challenging the UPCOpt-Out – How Exactly Will It Work?, (Bristows, June 2017).

104 parallel litigation : the case of biotechnology

The overarching goal of the Unified Patent Court is to provide fora one-stop litigation forum for patent disputes and building up auniform jurisprudence on validity and infringement of patents, thusenabling litigation with European-wide effect. The following sectioninvestigates whether the Unified Patent Court has the potential toachieve its ambition of countering parallel and duplicate litigation.Temporary and persistent obstacles to a uniform patent litigation sy-stem are identified and analyzed.

4.2.1 Transitional Period

During a transitional period of seven to fourteen years, national courtsand the UPC will have parallel jurisdiction over actions concerning(non-Unitary) European patents84. During that same period, patentowners can file an opt-out application for patents85. Opting-out Euro-pean patents protects from central validity challenges in the CentralDivision of the UPC – which will have effect for all single geographi-cal designations of the patent. The opt-out will remove the Europeanpatent from the jurisdiction of the UPC for its entire life, and can benotified by the latest one month before expiry of the transitional pe-riod86. It is expected that the pharmaceutical industry makes use ofthe opt-out for its most important patents, and try the unified courtfor some but not all its patent disputes87.

Parallel litigation during the transitional period raise a number ofquestions of lis pendens. In the event of an infringement action broughtto a national court, the defendant can lodge a nullity action with theUPC. The subject-matter of the dispute in the matter is not identical,yet may impact the stay decision at the national level88. If the patent

84. Art. 83 para. 1 and para. 5 Unified Patent Court Agreement; the transitionalperiod is seven years but may be prolonged up to seven additional years by theAdministrative Committee on the basis of a consultation with the users of the patentsystem and an opinion of the Court.

85. Art. 83 para. 3 Unified Patent Court Agreement.86. The debate about the intepretation of the article was clarified by the Preparatory

Committee of the UPC: An opt-out removes a patent entirely from the jurisdictionof the UPC, see https://www.unified-patent-court.org/faq/opt-out.

87. See Leela Barham, “European Pharma and the Unified Patent Court,” Phar-mExec.com, July 2016, on expected patent litigation strategies in the field.

88. Art. 29 of Reg. (EU) No. 1215/2012 (Art. 27 of Regulation (EC) No. 44/2001).

4.2 the future 105

holder then counterclaims, alleging the infringement of the Europeanpatent before the UPC, the UPC would probably have to stay its in-fringement decision as it cannot exclude the territory of a MemberState – where the infringement action is pending89. As exemplifiedby this case, parallel litigation will be a challenge in the early years ofthe court. For the transitional period to be a success, clear and consis-tent decisions on procedural aspects will be key, so that litigants canstreamline their litigation strategies accordingly.

4.2.2 Bifurcation

In the event of an infringement, the action should be brought to localand regional divisions of the UPC, whereas revocations are of thecompetence of the Central Division90. When the defendants bringa validity counterclaim before local or regional courts, the judgeshave the option to refer both actions to the Central Division withthe agreement of the parties, refer the counterclaim for revocation tothe Central Division and suspend or proceed with the infringementaction or proceed with both actions91. While under this regime du-plicate litigation will be banned, the potential for forum shoppingand divergences in region-dependent decisions will not vanish. Inparticular, infringement lawsuits will be brought before local and re-gional divisions, which will at their discretion decide on eventualcounterclaims for revocation or refer them to the Central Division.The eventual heterogeneity resulting from bifurcation will, however,be tempered by the Central Divisions of the Court of First Instancehaving exclusive jurisdiction over revocation lawsuits, and the Courtof Appeal leveling-up eventual heterogeneous infringement and in-validity counterclaims before the local or regional divisions of theContracting Member States92. The multinational mixture and the trai-

89. See the overview by Konstantin Schallmoser, available under http://www.eplawpatentblog.com/2013/October/Competence_AUPC_Transitional_regime.pdf%20.

90. Art. 33 Unified Patent Court Agreement for the forum rules; further Alba A.Betancourt, “Cross-Border Patent Disputes: Unified Patent Court or InternationalCommercial Arbitration?,” Utrecht Journal of International and European Law 32, no. 82

(2016): 44–58.91. Art. 33 para. 4 lit. a-c Unified Patent Court Agreement.92. For a discussion of the extent to which the establishment of the UPC can con-

tribute to enhance judicial coherence, see Federica Baldan and Esther Van Zimmeren,

106 parallel litigation : the case of biotechnology

ning program of the judges further aims at strengthening common in-terpretations standards93. The over-representation of national legallyqualified judges in (active) local and regional divisions (two out ofthree) has, however, the potential to lead to national resistance poc-kets94.

4.2.3 Non-Participating States

The European Court of Justice held that an agreement creating aunified patent litigation system including courts in countries outsidethe EU was not compatible with the provisions of the Treaty on Eu-ropean Union and the Treaty on the Functioning of the EuropeanUnion95. Henceforth, the states parties to the European Patent Con-vention but not to the EU were excluded. Furthermore, three Mem-ber States refused to join the new system. Spain illustratively critici-zed the high costs of enforcement in the UPC compared to Spanishcourts, particularly for SME, and reasons of language96. A Europeanpatent validated in non-participating and participating states will besubject in parallel to the jurisdiction of (a) national court(s) for itsnon-participating state part(s) and of the unified court for its partici-pating state part(s). This will require coordinating litigation for pan-European enforcement of patent rights, but also provide for the con-tinued potential for forum shopping and inconsistencies in decision-making97.

“The Future Role of the Unified Patent Court in Safeguarding Coherence in the Eu-ropean Patent System,” Common Market Law Review 52, no. 6 (2015): 1529–1577; Art.75 Unified Patent Court Agreement: ‘[t]he Court of First Instance shall be bound bythe decision of the Court of Appeal on points of law.

93. Christopher J. Bayliss, “The Unitary Patent and Unified Patent Court: PotentialChanges and Implications,” International Review of Intellectual Property and CompetitionLaw 5, no. 2 (2014): p. 464.

94. At the local division level, this is only true if the court hears 50 or more patentcases per year over three years.

95. European Court of Justice, Opinion 1/09, March 8, 2011.96. See Rey-Alvite Villar, M. Spanish Government Provides its Reasons not to Join

the Unitary Patent and UPC System. Bristows UPC News, March 22, 2017.97. See Reto Hilty et al., The Unitary Patent Package: Twelve Reasons for Concern, (Max

Planck Institute for Intellectual Property & Competition Law Research Paper No. 12-12, 2012).

4.2 the future 107

4.2.4 Enforcement and Opposition

Another challenge for uniformity is the relationship between the Uni-fied Patent Court and the European Patent Office. In an oppositionprocedure before the EPO, there is no possibility to refer questionsto the Court of Justice of the European Union, illustratively on theinterpretation of Directive 98/44/EC on the legal protection of bio-technological inventions, Directive 2004/48/EC on the enforcementof intellectual property rights and Regulation (EC) 469/2009 concer-ning supplementary protection certificate for medicinal products. Onthe contrary, the Unified Patent Court can or must refer questions onthe interpretation of European law before the Court of Justice98. It isargued that the EPO would ‘consider’ the interpretation of the UPC.Nevertheless, since the EPO Board of Appeal is not part of the judicialframework of the European Union, there will be no legal obligationto do so. Hence, there is great potential for divergent interpretationsof patentability standards, especially for biotechnology patents99.

Since the Unified Patent Court will also result in the possibility tocentrally challenge a patent, it is likely that the number of oppositi-ons will decrease100. With regard to the expected costs of litigation,one can hypothesize that filing an opposition will remain a cheap op-tion against valuable patents. The potential driver of the UPC in thisfield is of procedural nature. When an action is brought before theUPC relating to a patent which is also subject to an opposition, theUPC may of its own motion or at the request of a party request thatopposition proceedings before the European Patent Office be accele-

98. Article 21 UPC Agreement in relation with Art. 267 TFEU; Rob J. Aerts, “TheUnitary Patent and the Biotechnology Directive: Is Uniform Protection of Biotechno-logical Inventions Ensured?,” European Intellectual Property Review 9 (2014): 584–587.

99. The bio-industry responded positively to the announcement that the UK hasratified the UPC. The UPC is seen as reinforcing the position of SME and theexperience of the UK judiciary for handling life science cases has been praised(since the Central Division of the Court responsible for Chemistry will be basedin London)(https://www.thepharmaletter.com/article/bia-welcomes-uk-ratification-of-the-upc-agreement).100. Workshops on the Unitary Patent and the Unified Patent Court, (EPO’s Economicand Scientific Advisory Board, 2013), p. 39.

108 parallel litigation : the case of biotechnology

rated101. The Court may also stay its proceedings102. Considering thehigh number of patents subject to an opposition procedure in parallelto national nullification actions, the practical impact of the UPC mightbe to reduce such occurrences by making use of its acceleration andstay instruments.

4.2.5 Double Protection

A further potential for clash results in proposed recent draft bills atthe national level. Germany and France plan to protect the same in-vention by a national patent and by a patent for which the UPC isexclusively competent103. This situation would arise where an appli-cant would file the identical national and European patent applica-tion on the same day with identical or overlapping claims or wherethe national application was the priority application for the Europeanapplication and the priority application is not abandoned104. Thisdouble protection undermines the one-shop aim of the UPC. Whiledouble enforcement is prohibited, and provisions foresee the discre-tion of courts to stay the national proceedings until completion ofthe UPC action, the patent holder can initially forum shop based onthe particulars of the case105. In instances where the European patentis nullified by the UPC, the proprietor of the national patent maystill enforce the patent nationally106. Further, inadmissibility and stay

101. Article 33 para. 10 of the UPC Agreement, Rule 298 of the Preliminary Set ofProvisions for the Rules of Procedures of the Unified Patent Court, 18th draft, 19

October 2015.102. Rule 118 and 295 of Preliminary Set of Provisions for the Rules of Proceduresof the Unified Patent Court, 18th draft, 19 October 2015.103. For Germany, Drucksache 18/8827 would introduce a new Art. II § 18 Act onInternational Patent Treaties; for France, see Ordonnance 2018-341 du 9 mai 2018

relative au brevet européen à effet unitaire et à la juridiction unifiée du brevet.104. Peter Höcherl, “Double Protection and Forum Shopping under Germany’sDraft UPC Legislation,” BristowsUPC, 2017, for additional details.105. Höcherl, “Double Protection and Forum Shopping under Germany’s Draft UPCLegislation” on forum shopping.106. Jan Krauss and David Kuttenkeuler, “Go German Patents! – Special Rules Allo-

wing Double Patenting and Parallel Enforcement for the Unitary Patent System andNational Patents in Germany,” Boehmert & Boehmert, 2017,

4.3 theory of parallel litigation 109

of proceedings are not possible for interim or protective measures107.Therefore, it appears that the patent holder can with regard to prelimi-nary injunctions enforce a national and a European patent in parallelwithout any restriction108.

4.3 theory of parallel litigation

When enforcing a patent in several countries to claim damages for pa-tent infringement, patent holders have to sue the alleged infringer(s)in multiple forums109. Yet, at any moment in time, the parties canreach an agreement and settle over the dispute. In exchange for thepayment of royalties, the competitor will be able to use the inventionat stake. To study the rationale of parties engaging in parallel litiga-tion in several countries, the following section formalizes litigation.Litigation with settlement in two countries is formulated as a gametheoretical model. The model considers a patent dispute where onlythe infringement is at issue and the patent holder seeks to enforceits patent in two different countries against one potential infringer110.In practice, such parallel disputes where a plaintiff and a defendantfaced each other in multiple jurisdictions represented 26% of all pa-tent disputes in the UK and 15% in the Netherlands from 2000 to2008

111.

Under the model, in the first stage of the game, the patent holderrealizes that his patent has been potentially infringed simultaneouslyby the infringer in Country A and in Country B and makes a universalsettlement offer, requesting a licensing fee to settle the case in the

107. Art. II § 18 (4) Gesetz über internationale Patentübereinkommen vom 21. Juni1976.108. Höcherl, “Double Protection and Forum Shopping under Germany’s Draft UPCLegislation” for injunctive relief.109. The patent holder can also sue in one country, and await trial or settle in thiscountry before suing in other countries.110. A range of definitions may be applied to the concept of parallel litigation, asit might be limited to lawsuits with exactly the same parties and same claims, orlawsuits arising in claim preclusion for some or all parties, or reactive, repetitiveor derivative litigation, see James P. George, “International Parallel Litigation,” TexasInternational Law Journal 37 (2002): 499–540.111. See Katrin Cremers et al., “Patent Litigation in Europe,” European Journal of Lawand Economics 44, no. 1 (2017): 1–44.

110 parallel litigation : the case of biotechnology

two countries112. In the second-offer subgame, that only occurs if theinfringer rejected the settlement offer, the patent case goes to trialin Country A. The patent holder, observing the first court outcomethen makes a second universal settlement offer to settle the two cases.Depending on the acceptance or rejection of the offer by the infringer,the game ends or a second trial in Country B is realized.

4.3.1 Model Setting

Consider two players, labeled P for the patent holder and I for the po-tential infringer. The patent holder has patented an invention in twocountries with market profits of 2π, π > 0113. At the beginning of thegame, there is a dispute between the patent holder and the infringerwhich entered the protected market114. If the patent holder sues thepotential infringer and the courts concludes that a patent claim wasinfringed, the infringer is forced to stop to make or sell the patentedinvention. In this case, the patent holder earns the monopoly profitsβ2π where β = 1 and the infringer earns zero115. If the infringementsuit is not successful or the parties agree to settle, the patent holder

112. Aggregation of lawsuits enables plaintiffs to rationally invest in litigation andspread expenses over the portfolio of cases, thereby increasing the plaintiff’s leveragein settlement negotiation; it also generates efficiencies and transaction costs for bothdefendants and plaintiffs, see Theodore D. Rave, “Governing the Anticommons inAggregate Litigation,” Vanderbilt Law Review 66 (2013): 1183–1258. Global settlementsand licenses are common practice in the pharmaceutical market, see illustratively arecent global settlement between Mylan and Genentech and Roche in relation to pa-tents for Herceptin (trastuzumab) for all countries except Japan, Brazil and Mexico(http://newsroom.mylan.com/2017-03-13-Mylan-Announces-Global-Settlement-and-License-Agreements-with-Genentech-and-Roche-on-Herceptin-R).113. Within one country, the profits therefore are equal and represented by π.114. We consider the existence of the dispute as given, i.e., we do not study moni-toring efforts to establish the presence of the entry, the identity of the infringer, thearguments of infringement and validity. For a model where the patent holder canchoose his level of monitoring efforts before observing entry, see Claude Crampesand Corinne Langinier, “Litigation and Settlement in Patent Infringement Cases,”The RAND Journal of Economics 33 (2002): 258–274.115. For the sake of simplicity, we assume that the market payoff 2π is identical

whether the patent holder settles or proceeds to trial, although evidence suggeststhat a trial reduces the profits of firms, see Josh Lerner, “Patenting in the Shadow ofCompetitors,” Journal of Law and Economics 38, no. 2 (1995): 463–495.

4.3 theory of parallel litigation 111

earns the duopoly profits β2π whereas the infringer earns(1− β)2π,where β ∈ [0, 1] and β 6= 1116.

Similarly as to Crampes and Langinier (2002), we limit our study tothe situation in which the property rights of the patent holder cannotbe challenged, i.e. the infringer cannot launch an invalidity suit. Thevalidity of the patent could be accounted for as a factor impacting β.Should a patent be invalid, this would reduce to a greater extent theshare of the (duopoly) market profits and increase the share of the(duopoly) market profits of the infringer117.

We assume that infringers have superior information about the streng-th or weaknesses of the patent118. On the contrary to validity trials,where the patent holder may have private information about the pa-tent itself119, we argue that the infringer has superior private infor-mation as to whether his infringing violated the (publicly disclosed)patent claim120. Under our setting, the infringer classifies the patentsas either ‘high’ quality patent infringement claim, H, or ‘low’ qua-lity patent infringement claim, L, so that j ∈ {L,H}121. We define theprobability of a patent claim of winning at trial as θj. Accordingly,θH patent claims have a higher likelihood to prevail at trial than θLclaims. The patent holder’s prior beliefs about the quality of its pa-

116. For more details on the monopoly versus duopoly situation as determined bythe outcome of the trial, see Carl Shapiro, “Antitrust Analysis of Patent SettlementsBetween Rivals,” Antitrust Summer, 2003, 70–77.117. See Crampes and Langinier, “Litigation and Settlement in Patent InfringementCases.”118. Asymmetric information is a leading explanation for the failure by litigantsto reach Pareto-superior settlements before trial, see Keith N. Hylton, “AsymmetricInformation and the Selection of Disputes for Litigation,” The Journal of Legal Studies22, no. 1 (1993): 187–210.119. Michael J. Meurer, “The Settlement of Patent Litigation,” The RAND Journal ofEconomics 20, no. 1 (1989): 77–91.120. The assumption is in line with the general litigation literature assuming thatdefendants have superior information about their culpability, see Florian Baumannand Tim Friehe, On Discovery, Restricting Lawyers, and the Settlement Rate, DICE Dis-cussion Paper 155 (2014). As to the legal uncertainty about the interpretation of thepatent claims and the question of whether the infringing acts fall within the patentclaims, we argue that none of the actors has superior information.121. As in most theoretical work, only one side (the infringer) is privately informed;however, see Urs Schweizer, “Litigation and Settlement under Two-Sided IncompleteInformation,” Review of Economic Studies 56, no. 2 (1989): 163–177 for a two-sidedasymmetric information (signaling) model.

112 parallel litigation : the case of biotechnology

tent claim is p for θH and 1−p for θL. We assume that trial outcomesare providing information to the patent holders which update theirbeliefs about their likelihood of enforcing a either θH or θL patentclaim. Underlying the model is a screening problem, where the unin-formed party proposes the contracts122.

We define the costs of litigation in the two countries as CT = 2C,and determine that those costs have to be borne by the party losingat trial according to the British rule123. The settlement offers Si aredefined as S1 for the first offer and S2 ∈ {High,Low}. For the secondoffer, the patent holder can make a high or low offer after a trial winrespectively loss after having updated his beliefs upon the strength ofits patent. A S2High settlement offer thus equals to a higher licensefee to be transferred to the patent holder compared to S2Low.

4.3.2 Timing

The timing of the game is presented hereafter.

1. P files two infringement suits against I in Country A and Coun-try B and makes a universal take-it-or-leave-it settlement offer.

2. I chooses whether to accept or reject the settlement offer. If theinfringer accepts the offer, the patent holder realizes his payoffand the game is over. If he rejects it, the case goes to trial inCountry A and the players enter into a signaling game.

122. As in classical sorting or screening games, the first-mover acts to screen thesecond-movers types into those who accept and those who refuse, see Andrew F.Daughety and Jennifer F. Reinganum, “Economic Theories of Settlement Bargai-ning,” Annual Review of Law and Social Science 1, no. 1 (2005): 35–59.123. While the American rule on attorney fees provides that each party must pay itsown legal fees, see Alyseka Pipeline Service Co. v. Wilderness Society, 421 U.S. 240

(1975); for a model adopting the British rule in the US, see John J. Donohue, “TheEffects of Fee Shifting on the Settlement Rate: Theoretical Observations on Costs,Conflicts, and Contingency Fees,” Law and Contemporary Problems 54 (1991): 195–222.

4.3 theory of parallel litigation 113

P D

Settlement, Payoff Delivery

accept

Litigation, Enter Signaling Game

reject

Figure 4.2: Timing of the Game

In the second-offer subgame that follows:

1. The Country A trial outcome is realized, and P wins the casewith probability θ.

2. After observing the outcome, P can make a either high or lowuniversal settlement offer after a win, respectively a high or alow settlement offer after a loss.

3. I chooses whether to accept or refuse the offer. An infringeraccepting the offer transfers the respective license fee, in addi-tion to the costs of trial if he lost at trial in Country A. In theeventuality the case goes to trial in Country B, the respectivepayoffs of the two trials are summed.

H

P

I

A R

S2High

I

A R

S2Low

win

P

I

A R

S2High

I

A R

S2Low

loss

L

P

I

A R

S2High

I

A R

S2Low

win

P

I

A R

S2High

I

A R

S2Low

loss

Figure 4.3: Timing of the Second-Offer Subgame

114 parallel litigation : the case of biotechnology

4.3.3 Outlining the Solution

We present the solving of the game by backward induction and ex-plain the optimal strategies for rational players124. The strategy spa-ces of the players are for the patent holder P{S1;S2(High,Low)} andfor the infringer I{A,R;A,R} respectively.

4.3.3.1 Second Settlement Offer

For now, we consider the second settlement offer only. After the firsttrial, the patent holder can choose to make a low settlement offer byrequesting a low licensing fee that both the θL and θH type infringerwill accept, or a high offer that only the infringer facing a θH wouldaccept. The probability of the patent holder to enforce a patent claimof quality θH is p respectively θL is 1−p. After observing the outcomeof its patent at trial in Country A, the patent holder updates his beliefsas to the probability of enforcing a θH and type θL in consistence withBaye’s rule125. The trial outcome provides the plaintiff with additionalinformation as to the strength of his patent; he therefore updates hispriors as a result of the decision.

124. Backward induction relies on Zermelo’s theorem that ‘[e]very finite gameof perfect information has a pure strategy Nash equilibrium that can be derivedthrough backward induction. Moreover, if no player has the same payoffs at any twoterminal nodes, then backward induction results in a unique Nash equilibrium’, seeUlrich Schwalbe and Paul Walker, “Zermelo and the Early History of Game Theory,”Games and Economic Behavior 34, no. 1 (2001): 123–137.125. Information update occurs at different stages in litigation at pretrial bargaining(Robert H. Gertner, “Asymmetric Information, Uncertainty, and Selection Bias in Liti-gation,” University of Chicago Law School Roundtable 1, no. 1 (1993) and Neil Rickman,“Contingent Fees and Litigation Settlement,” International Review of Law and Econo-mics 19, no. 3 (1999): 295–317), especially when the defendant accepts or rejects thesettlement offer (Alison Watts, “Bargaining through an Expert Attorney,” Journal ofLaw, Economics, & Organization 10, no. 1 (1994): 168–186). For sequential litigationin several countries, however, we argue that the major information update is givenwhen the court in the first country reaches a decision on the merits.

4.3 theory of parallel litigation 115

Updated Beliefs. The patent holder assigns the following probabilityto type θH and type θL which are the only beliefs consistent withBaye’s rule:

µ(θH|win) =P(win|θH) ∗ P(θH)

P(win|θH) ∗ P(θH) + P(win|θL) ∗ P(θL)=

pθHp(θH) + (1− p)θL

(8)

µ(θH|lose) =P(lose|θH) ∗ P(θH)

P(lose|θH) ∗ P(θH) + P(lose|θL) ∗ P(θL)=

p(1− θH)

p(1− θH) + (1− p)(1− θL)

(9)

µ(θL|win) =P(win|θL) ∗ P(θL)

P(win|θL) ∗ P(θL) + P(win|θH) ∗ P(θH)=

(1− p)θL)

(1− p)(θL) + pθH(10)

µ(θL)|lose) =P(lose|θL) ∗ P(θL)

P(lose|θL) ∗ P(θL) + P(lose|θH) ∗ P(θH)=

(1− p)(1− θL)

(1− p)(1− θL) + p(1− θH)

(11)

• Patent Holder Loss

The infringer will accept settlements offers that are higher or equalto the expected value of litigation. In other words, when EU(A) >EU(R) the infringer accepts the offer. If he is requested to compensatebeyond that extent, the infringer will prefer to litigate than settlingthe case126.

Settlement Payoff. Under our model, the patent holder requests thepayment of a license fee compensating for his loss of monopoly mar-ket and for his legal fees occurred at trial. Therefore, the settlementpayment is equal to S2 = θ(βπ+C). When settling under these terms,the patent holder is will obtain the following payoff:

2βπ+ S2 −C = βπ(2+ θ) +C(θ− 1) (12)

126. This is the standard assumption in the scholarship on legal decision making,see William M. Landes, “Sequential versus Unitary Trials: An Economic Analysis,”The Journal of Legal Studies 22, no. 1 (1993): 99–134, but also Ivan P. P’ng, “StrategicBehavior in Suit, Settlement, and Trial,” Bell Journal of Economics 14, no. 2 (1983): 539–550 and David Rosenberg and Steven Shavell, “A Model in Which Suits Are Broughtfor Their Nuisance Value,” International Review of Law and Economics 5, no. 1 (1985):3–13 for negative expected utility lawsuits.

116 parallel litigation : the case of biotechnology

Indifference. Depending on the strength of the patent infringementclaim, the alleged infringer will accept or not the requested licensefee. The settlement amount making the infringer indifferent can benoted as follows:

2(1−β)π− S2 > (1−β)π+ (1− θ)(1−β)π+ θ(−C) (13)

When solving, we find that the amount to reach indifference is:

S2 6 θ((1−β)π+C) (14)

Hence, when facing a θH patent infringement claim, the infringer willbe willing to pay a higher settlement fee than when facing a θL claim.Conversely, when the patent holder seeks for licensing fees that donot match his underlying legal claim, the alleged infringer will notagree to settle. Considering our two types of patent infringementclaims θH and θL, the patent holder can make either a high settle-ment offer S2HighLose = θH((1 − β)π + C) that would only makethe infringer facing a θH patent indifferent, or a low settlement offerS2LowLose = θL((1−β)π+C) that would make infringers facing bothtypes of patents claims indifferent. If he requests a high license fee, aninfringer facing a high quality patent claim will settle, whereas the in-fringer facing a low quality patent will litigate. After updating his be-liefs, the patent holder makes a high offer only if the expected payoffof such strategy is higher, that is if EU(S2HighLose) > EU(S2LowLose)(see Point A of the Appendix for the equations).

After a first litigation loss, a rational patent holder will only makea high offer if the market profits π are a multiple of the litigationcosts C. Figure 4.4 illustrates the sensitivity of the parameters andthe required ratio between profits and litigation costs for the patentholder to make a high rather than a low offer.

4.3 theory of parallel litigation 117

Figure 4.4: Market Profits vs. Litigation Costs for High Offer after Loss

0

5

10

15

Mul

tiplie

r For

Hig

h O

ffers

0 .2 .4 .6 .8 1

var1p (likelihood to enforce θH) β (patent market share)θH (likelihood to win with θH) θL (likelihood to win with θL)

To compute Figure 4.4, the parameters are set as follows. On the x-axis, the variable parameters flows from 0 to 1, the non-variable para-meter p is 0.6, β is 0.6, θH is 0.7 and θL is 0.4. The resulting multiplieron the y-axis represents the indifference point for the patent holderto make a high or low offer. Illustratively, the curve for p gives themultiplier with a variable p and non-variable β, θH and θL.

When the profits are worth over five times the litigation costs, a ra-tional patent holder will almost always request a high license fees,no matter the underlying parameters. The larger the profits at stakein comparison to the costs of litigation, the more likely it will be thepatent holder selects to request high royalties. If the ratio of profits tolitigation costs is smaller the patent holder will – depending on theparameters – prefer to make low offers, increasing the likelihood ofreaching settlement agreement.

With an increase in p, that is in the likelihood to enforce a θH patentinfringement claim, the patent holder has a stronger legal position.As reflected in Figure 4.4, the higher p, the stronger his preferencefor high settlement fees – even after his first loss. Similarly, when

118 parallel litigation : the case of biotechnology

the probability of θH patent claims to prevail at trial is high, he willchoose to request a high settlement fee. On the contrary, when lowquality claims are more likely to prevail at trial, the patent holderwill be more likely to seek a low license fee. This is due to the factthat with the Bayesian update, the patent holder updates his priorsand estimates his likelihood to enforce a low quality claim as higher.Finally, as to the share of market profits β, i.e., the share of the profitsthat would remain within the patent holder’s control and earning inthe case of a duopoloy, the higher it is, the more the patent holderseeks settlements. Since the patent holder is not as dependent on theoutcome of the trial, he will make low offers and settle the case, ex-cept if the profits at stake are disproportionally high. If one sought toinclude validity challenges in the model, β would be reduced. In theeventuality of patent nullification, third parties may enter the previ-ously protected market. According to Figure 4.4, this would increasethe preference of patent holders towards high settlement offers, thatis he would be less eager to settle than previously.

• Patent Holder Win

Indifference. After a first patent holder trial win, the settlement amountmaking the infringer indifferent is as follows:

2(1−β)π− S2 > −C+ (1− θ)(1−β)π+ θ(−C) (15)

When solving, we find that:

S2 6 (1+ θ)((1−β)π+C) (16)

Settlement Offers. The high offer by the patent holder is S2HighWin =

(1+ θH)((1− β)π+C), and the low settlement offer is S2LowWin =

(1+ θL)((1− β)π+ C). After updating his beliefs, the patent holdermakes a high offer if EU(S2Highwin) > EU(2LowWin) rather than alow offer only if the expected payoff of such strategy is higher (seePoint B Appendix for the equations). The latter will only be acceptedby the infringer facing a high quality patent, whereas the infringerfacing a low quality patent will litigate.

Using the identical non-variable parameters as above, we observe inFigure 4.5 that the market profits at stake must only be a fraction

4.3 theory of parallel litigation 119

of the litigation costs for the patent holder to proceed with a highoffer, when it was a multiplier after a first loss. This may be explai-ned by the fact that the patent holder already secured his monopolyin Country A, and updated his beliefs after his win. He estimateshis likelihood to enforce a high quality patent infringement claim ashigher than before. Therefore, the β share, his initial probability to en-force a high quality patent, or even the strength at trial of low qualitypatent infringement claims are almost irrelevant to his expected pa-yoff calculation. This is reflected in the flat slope of these parametersin Figure 4.5.

Figure 4.5: Market Profits vs. Litigation Costs for High Offer after Win

.05

.1

.15

.2

.25

.3

Mul

tiplie

r to

Mak

e H

igh

Offe

r

0 .2 .4 .6 .8 1

var1p (likelihood to enforce θH) β (Patent Share Market)θH (likelihood to win with θH) θL (likelihood to win with θL)

Only θH, the strength of a θH patent claim before the court will in-fluence his decision to make a low or high offer. However, even withlower chances of success, the patent holder will make high offers alt-hough the profits are a fraction only of the litigation costs. When inte-grating validity challenges, the lack of relevance of β shows that suchchallenges do not influence the patent holder’s decision anymore.

120 parallel litigation : the case of biotechnology

4.3.4 Conclusions Subgame

Whether or not the patent holder won or lost at the first trial willaffect his decision to make a high or low settlement offer. After a winat trial, the patent holder prefers to make a high offer with muchlower profits at stake (a fraction of the litigation costs) than after afirst loss (a multiplier of the litigation costs)127.

4.3.5 First Settlement Offer

Suppose parameter values are such that the infringer’s beliefs are thatthe patent holder will make a high settlement offer after a first win(requesting a high licensing fee after the win) and a low settlementoffer after a first lose. When facing θH, the infringer thus acceptsthe high offer after the win and the low offer after the loss. In other,words if EU (A) > EU (R).

Indifference. At the first offer, for the infringer facing θH:

2(1−β)π− S1 > θH((−C) + 2(1−β)π− S2HighWin)

+ (1− θH)(2(1−β)π− S2LowLoss) (17)

When solving,

C(θH(2+ θH − θL)) − (−1+ b)π(θH + θ2H + θL − θHθL) > S1 (18)

On the contrary, when facing a patent of quality θL, the infringerrefuses the high offer after a first win and accepts the low after a loss.Accordingly, he will take the offer should he be rendered indifferenton whether to settle or litigate, EU (A) > EU (R).

Indifference. At the first offer, for the infringer facing θL:

2(1−β)π− S1 > θL((−C) + θL(−C) + (1− θL)((1−β)π))

+ (1− θL)(2(1−β)π− S2LowLoss) (19)

127. It appears that at the pure national (appeal) level, the propensity to settle is notimpacted by the outcome of the first instance court, see Katrin Cremers, “SettlementDuring Patent Litigation Trials. An Empirical Analysis for Germany,” The Journal ofTechnology Transfer 34, no. 2 (2009): 182–195.

4.3 theory of parallel litigation 121

When solving,

C(1+ θ2L) + 2π(θL − θLβ > S1 (20)

The patent holder, who does not know the quality of its patent claim,will prefer to make a S1 settlement offer that will only be acceptedby the infringer facing a θH patent if EU(offer for θH) > EU (offer forθL) (see point C of the Appendix).

According to Figure 4.6, that bases on the identical non-variable va-lues as above, the parameters widely impact the patent holder’s deci-sion to make either a high or low offer. The higher his likelihood toenforce a θH patent infringement claim and the higher his probabilityθH to prevail at trial, the smaller the required ratio of profits to litiga-tion costs for him to make a high offer. If these parameters are high,the patent holder will make high offers even when the profits arecomparatively small and request high license fees to settle. If theseparameters are low, he will rather make low offers. The market shareβ impacts the outcome as well, but the slope is less steep than theother parameters. If the market share the patent holder retains is low,the patent holder will seek high offers. If the market share is high,the patent holder will favor to settle initially, and make low offerswith comparatively high market profits at dispute. Since the patentholder does not lose as much market shares, he will prefer to settle toavoid litigation costs. With the inclusion of patent validity challenges,the reduction of β will accentuate the preference of the patent holderfor high settlement offers; he will request higher royalties from thealleged infringer.

122 parallel litigation : the case of biotechnology

Figure 4.6: Market Profits vs. Litigation Costs for High Offer

0

20

40

60

80

100

Mul

tiplie

r to

Mak

e H

igh

Offe

r

0 .2 .4 .6 .8 1

var1p (likelihood to enforce θH) β (patent market share)θH (likelihood to win with θH) θL (likelihood to win with θL)

4.3.6 Predictions

1. In sequential litigation, P makes a first high settlement offer thatheavily depends on the underlying parameters; depending onthe situation, the profits must represent a tenfold of the litiga-tion costs, in other the double.

2. After a first loss, P makes a high settlement offer only if theprofits at stake are comparatively high (a multiplier), leading tolower rates of double decisions.

3. After a first win, P makes a high settlement offer even if theprofits at stake are comparatively small (a fraction), leading tohigher rates of double decisions.

All in all, this theoretical model helps to explain that there exist se-veral situations in which it might be optimal for patent holders tomake high settlements offers that not all infringers will accept. In fact,the patent holder may request high licensing fees at the initial stage,even if the profits under dispute are not substantially higher than the

4.4 conclusion 123

litigation costs. If the patent holder wins its case in the first country,he will even make a high offer if the profits are a fraction of the litiga-tion costs only. On the contrary, there is a shift after the patent holderloses its case in the first country. The patent holder then makes onlyhigh offers when comparatively high profits are at stake.

These predictions help to explain why for patent holders it is oftenoptimal not to settle at first, no matter the profits at stake. In a systemwhere several courts exist, parties might litigate individual nationalsubparts of the global claim, yet at any moment settle over the wholegroup of lawsuits. The expected utility calculation from the stand-point of the patent holder is European, not national128. The possibi-lity to grant a license for the different countries is such that the patentholder might prefer the uncertainty of a first trial to the certainty of alow license fee129. This is also due to the fact that if the patent holdergrants the license or is defeated before the courts, he switches from amonopoly to a duopoly situation. Royalties ought to compensate forthis.

4.4 conclusion

At first glance, the patent litigation landscape in biotechnology con-veys an impression of a divided Europe. In the United Kingdom,biotechnology patents fared poorly at trial when compared to theirperformance in other large jurisdictions such as Germany, France orthe Netherlands. Such high-level analysis is supported by the qualita-tive assessment of duplicate decisions. When identical patents wereexamined by a court in the United Kingdom on the one side and inGermany and the Netherlands on the other, a trend emerges. In theUK, the interpretation of claims was stricter, and patents would havea narrower scope of protection than abroad. Similarly, the validity ofpatents was harder to establish, as the definition of legal constructssuch as the person skilled in the art was more restrictive, but also ob-

128. This mirrors, although in a different context, the point by Landes that the se-paration of lawsuits (liability and damages) reduces the likelihood that parties settleout of court, see Landes, “Sequential versus Unitary Trials: An Economic Analysis.”129. This conclusion is in line with the argument by Rave, “Governing the Anti-commons in Aggregate Litigation” that plaintiffs have higher leverage in settlementnegotiations as a benefit of aggregation.

124 parallel litigation : the case of biotechnology

viousness and novelty threshold found to be higher. These contradicti-ons illustrate the direct consequences of the lack of harmonization forlitigants and the uncertainty currently prevailing in Europe.

When a patent is centrally granted by the European Patent Office, ithas different effects in the different Member States of the EuropeanPatent Convention. The benefits of a centralized grant system appearseverely undermined if one patent varies both in scope of protectionand in validity depending on the forum of litigation. The legal tra-ditions that prevailed before the harmonization eventually managedto survive and would explain part of these incongruities. A notableexample is given by the methodologies of claim interpretation prevai-ling in the UK before the entry into force of the EPC. Latter defineda narrower scope of protection for patent claims. While harmonized,to some extent, by the European Patent Convention, the applicationof stricter standards in claim interpretation in the UK than abroadultimately is in line with such previous practices.

The novel and to be introduced unified enforcement regime hopes toremedy to this situation of heterogeneity. A variety of challenges willmark the transition but also the long-run operation of the UnifiedPatent Court. During the transitional period, parallel litigation willprevail. Afterward, the bifurcation between infringement and nulli-fication has the potential to lead to forum shopping and divergen-ces between the different local and regional courts. Similarly, pan-European litigation will not be a reality as thirteen States parties tothe European Patent Convention will not be part of the new system.Again, choices of forum and coordination of parallel litigation proce-dures shall remain possible. Additionally, the potential for differencesbetween the European Patent Office and the Unified Patent Court isstrong. Since there is no legal obligation for the EPO Board of Appealto follow the interpretation of the unified court, differences may arise.Last but not least, the double protection that recent national draftsforesee threaten the one-shop aim of the new court, enabling compa-nies to hedge themselves nationally and seek parallel enforcement ofpreliminary injunctions.

The divisions of the Unified Patent Court, such as the London Sectionof the Central Division dealing with chemical, pharmaceutical andlife sciences disputes will comprise judges from different contracting

4.4 conclusion 125

states. This should level out eventual differences in legal traditionsthat may exist. Once in operation, the court will be at an inflectionpoint. Adoption a pro-patentee, or pro-third parties approach, willtake the court in divergent directions. As such, philosophies of jud-ges selected from patent heavy states will be in a strong position toinfluence the direction of the court from its inception.

The Unified Patent Court will be in a key position to rehabilitateEuropean patent litigation and set if on the path to further homo-geneity. Considering the disparate effects of European patents, theunified enforcement regime, even if incomplete, is a significantly im-pactful step towards a homogeneous patent system in Europe. Thereexist and will be national resistance and institutional challenges forthe court to begin and develop its operations. However, the UnifiedPatent Court is in an unique position to establish an authority en-hancing regional integration. If the court manages to position itselfas an attractive and efficient venue for patent dispute litigants, it candemonstrate the benefits for countries to have harmonized laws inter-preted by centralized courts. In that regard, a number of legal uncer-tainties, as illustrated in the previous section, will provide the courtwith the opportunity to set clear jurisdictional precedents in both ma-terial and procedural matters. Should the court mark its path, non-participating EU countries will be tempted to reconsider their posi-tion, and non-participating non-EU countries to follow the leadingcases of the court. While a perfectly harmonized legal system, in pa-tent law and in other legal fields, seems to remain rather a goal thana reality, the Unified Patent Court can contribute to and reinforce theidea that single courts can be the norm in Europe. The last fifty yearsshould serve to demonstrate that harmonizing laws cannot suffice ifthe agents of their implementation, in casu the judges, construct thesestandards in line with national traditions. The current national ge-opolitical trends are not supportive of regional integration, and thesuccess or failure of the Unified Patent Court will serve as a furthertest case for the European Union’s ability to navigate in troubled wa-ter.

Centralizing the patent enforcement system is, however, necessary toeliminate parallel litigation. From a game theoretical standpoint, thepaper demonstrates that even in case of uniform application of har-

126 parallel litigation : the case of biotechnology

monized patent laws, parallel litigation will exist. The model showsthat it is often optimal for patent holders to proceed to trial. In asystem where several courts exist, parties might litigate individualnational subparts, yet at any moment settle over the whole group oflawsuits. The expected utility calculation from the standpoint of thepatent holder is European, not national. Accordingly, it is argued thatthe elimination of parallel litigation cannot be reached through themere harmonization of substantial law. With its current architecture,the Unified Patent Court will most likely not reach such goal. It has,however, the potential to reduce the occurrence of parallel litigation,and eventually, over time, become the one-stop forum for patent en-forcement in Europe it aims to be.

4.5 appendix 127

4.5 appendix

Point A. Patent Holder Loss.

For simplification, we replace the patent holder updated beliefs asfollows: p(1−θH)

p(1−θH)+(1−p)(1−θL)by P1 and (1−p)(1−θL)

(1−p)(1−θL)+p(1−θH)by P2.

The high offer is made by the patent holder if his expected payoff ishigher than from making a low offer:

P1 (2βπ) + (θH((1−β)π+C))+P2 ((βπ−C) + θL(π)

+ (1− θL)(βπ−C)) > 2βπ+ θL((1−β)π+C) (21)

When solving,

C(θH − θL − 2P2 + θLP2)

2β− θH +βθH + θL −βθL − 2βP1 − 2βP2 − θLP2 +βθLP2> π

(22)

Point B. Patent Holder Win.

For simplification, we replace the patent holder updated beliefs asfollows: p(θH)

p(θH)(1−p)θLby P3 and (1−p)θL

(1−p)(θL)pθHby P4

The high offer is made by the patent holder if his expected payoff ishigher, therefore:

−C(P3 − 1+ P3θH − P4 + P4θL − θL)

β(−1+ θL + P3 − θHP3 + P4 − θLP4) − 1− θL + P3 + θHP3 + P4 + θLP4> π

(23)

When solving, we find as follows:

1+C−CθH + θL1−β+ θH −βθH + 2βP3 + P4 + θLP4

> π (24)

128 parallel litigation : the case of biotechnology

Point C. First Settlement Offer.

The settlement offer the patent holder makes that will only be accep-ted by the infringer facing a θH:

p(2βπ+C(θH(2+ θH − θL)) − (−1+ b)π(θH + θ2H + θL − θHθL))

+(1−p)(2(θL)(π)+2(1−θL)(βπ−C)) > 2βπ+C(1+θ2L)+2π(θL−θLβ)

(25)

When solving,

C(3− 2θL + θ2L − 2p− 2pθH − pθ2H + 2pθL + pθHθL)

p(θH −βθH + θ2H −βθ2H − θL +βθL − θHθL +βθHθL> π (26)

Point C. Tables and Figures.

Definitive Outcome Infringement Outcome Construction Claims Invalidity Outcome Novelty Outcome Inventiveness Outcome Skilled Person Sufficiency Outcome

EP2055777 Identical D >UK D >UK - - - - -

EP0774511 Identical Identical D >UK Identical - Identical Similar -

EP1466983 Identical - - Identical Identical D, not in FR

EP0153114 Identical Identical Similar Identical Identical Identical D, �NLEP1386617 Identical - - Identical Identical UK, �NLEP0733710 Identical Identical Similar Identical Identical Identical D similar BE, �NL Identical

EP0796912 Identical Identical Similar - - - - -

EP0733712 Identical Identical Similar

EP0546090 Identical Identical - - - - - -

EP0430402 Identical - - Identical Identical D, �ITEP0148605 NL >UK, D NL >UK - NL >UK, D UK, �D,NL UK,NL >D

EP0411678 D >UK, NL - - D >UK, NL D >UK, NL D >UK D >UK, in NL

EP0912898 D >IT - - D >IT D, �ITEP0762888 Identical Identical Identical NL >UK Identical

EP0322438 Identical Identical Identical Identical D, ?

EP0607156 Identical Identical Identical Identical

Table A4.1: Duplicate Litigation Summary

4.5 appendix 129

Figure A4.1: Biotechnology Litigation Summary

Patent number Opposition Country of LitigationDuplicates 16 patents 40 cases

EP0153114 - NL,DE 0 0 0 0 1EP0322438 - IT,DE 1 1 1 1 1EP0546090 valid# UK,ES 0 0 0 0 1EP0607156 - NL,UK 1 1 1 1EP0762888 - UK,NL 1 1 1 1 1EP0774511 withdrawn UK,DE 0 0 0 0 0 0EP1386617 valid# UK,NL 0 0 0 0 0EP1466983 - DE,FR 0 0 0 0 0EP2055777 revoked# UK, DE 0 0 0 0 0EP0148605 valid# NL,UK,DE 1 0 0 1 0 1 0 0EP0411678 revoked# UK,NL,DE 0 0 1 1 0 0 1EP0430402 valid# DE,DE,IT 1 1 1 1 1 1 1EP0733712 - NL,DE,DE 1 1 1 1 1 1 1 1EP0796912 valid# NL,DE,DE 1 1 1 1 1 1 1EP0912898 valid# DE,IT,IT 1 0 0 1 0 0EP0733710 - BE,NL,DE,DE 0 1 1 1 0 1 1 1 1 1

Non-Duplicates 31 patents 35 casesDE19756864 - DE 1 1EP0134048 valid DE 0 0EP0181150 valid DE 1 1EP0194212 valid NL 0 0EP0258017 valid DE 0 0EP0305337 withdrawn NL 0 0EP0379606 valid DE 1 1EP0383569 revoked# UK 0 0EP0403506 valid DE 1 1EP0420358 revoked# UK 0 0EP0472651 valid# DE 0 1 1EP0584715 - DE 0 0EP0608235 valid# FR 1 1 1EP0614984 valid# NL 0 0EP0746398 - DE 0 0EP0761817 - FR 0 0EP0939804 valid# UK 1 1EP0959132 - DE 1 0 1EP1033995 - NL 0 0EP1131416 revoked DE 0 0EP1151004 - DE 1 1EP1238986 - UK 1 1 1EP1281760 - NL 0 0EP1290938 - NL 0 0EP1301539 - DE 1 1EP1956886 revoked# NL 0 0EP1994937 revoked# UK 0 1 0IT1311929 valid# IT 1 1 1EP0587780 valid FR,ES 1 0 1 0EP0178978 valid FR,FR 0 0 0 0EP0690991 valid# FR,DE* 1 0 1 0 1

Total 47 patents 75 cases

Definitive Outcome

This table displays the whole sample of litigated cases we study. *These cases were not counted as duplicate since the patentholder enforced its patent against two different infringers in two countries. #These oppositions were lead before or in parallel to a national court decision on the validity of the patent. 1, win for the patentholder; 0, loss for the patentholder. BE, Belgium; DE, Germany; ES, Spain; IT, Italy; FR, France; NL, Netherlands; UK, United Kingdoms.

Overall Infringement Invalidity

*These cases were not counted as duplicates since the patent holder enforced its patent against two different infringers in two

countries. These oppositions were lead before or in parallel to a national court decision on the validity of the patent. 1, win for

the patent holder; 0, loss for the patent holder. BE, Belgium; DE, Germany; ES, Spain; IT, Italy; FR, France; NL, Netherlands;

UK, United Kingdom. .

Part III

F U N D I N G O F S C I E N C E I N T H E U N I T E DS TAT E S

5I N T R O D U C T O RY R E M A R K S

Economic growth is driven by investments in innovation1. Researchand development investments are made by firms with the aim ofcreating new products, or improving existing ones. Beginning withNelson (1959) and Arrow (1962), economists pinned down the poten-tial for market failure in investments in R&D2. Mainly, this resultsfrom the fact that the social returns to investments exceed the pri-vate returns to the individual firms3. Because of the ‘public good’4

characteristics of R&D, the benefits of innovations cannot be fully ap-propriated by the investing firm, as other firms may free ride5.

1. With theoretical models by Paul Romer, “Endogenous Technological Change,”Journal of Political Economy 98, no. 5 (1990): 71–102 resp. Philippe Aghion and PeterHowitt, “A Model of Growth through Creative Destruction,” Econometrica 60, no. 2

(1992): 323–351 where growth is driven by technological change resp. vertical inno-vations involving creative destruction; for an empirical standpoint, see Ufuk Akcigit,John Grigsby, and Tom Nicholas, “The Rise of American Ingenuity: Innovation andInventors of the Golden Age,” NBER Working Paper, no. 23047 (2017) finding a posi-tive correlation between innovation and drivers of regional performance for the USbetween 1880 and 1940.

2. Kenneth Arrow, “Economic Welfare and the Allocation of Resources for In-vention,” in The Rate and Direction of Inventive Activity: Economic and Social Factors(National Bureau of Economic Research, 1962), 609–626 identifies indivisibility, inap-propriability, and uncertainty as the three reasons for such market failure to achieveoptimality in resource allocation.

3. For empirical support documenting a social return of R&D higher that the pri-vate returns to firms, see Zvi Griliches, “The Search for R&D Spillovers,” ScandinavianJournal of Economics 94 (1992): 29–47 and Bronwyn H. Hall, “The Private and SocialReturns to Research and Development,” in Technology, R&D, and the Economy, ed.Bruce L. Smith and Claude E. Barfield (Brookings Institution, 1996), 140–183.

4. See, e.g., Joseph E. Stiglitz, “Knowledge as a Global Public Good,” Global PublicGoods 19 (1999): 308–326.

5. The public good – free rider – problem lies at the foundation of the current in-tellectual property laws, see, e.g., Suzanne Scotchmer and Green Jerry, “Novelty andDisclosure in Patent Law,” The RAND Journal of Economics 21, no. 1 (1990): 131–146.Certain factors, such as the cost of imitation, might mitigate the problem of under-investment, see Richard C. Levin et al., “Appropriating the Returns from Industrial

132

5.1 impact of public funding 133

Even if appropriability was complete, firms may disregard sociallyvaluable R&D projects6. Such market failure is generally consideredthe justification for public investments in science7, which represent alarge share of expenditures of developed nations8. Public research iswidely recognized as a key contributor to innovation and economicgrowth. However, little is known as to its mechanisms and returns.Several authors called for more attention on the subject by a specia-list scholarly community9. Despite the critical need for measurementof the effectiveness and efficiency of government funding decisions,there are few studies and little data10. The second pair of papers ofthis dissertation seeks to contribute to this body of work. This chapterreviews the literature on the impact of public funding and of univer-sity patenting. The two papers build upon this social context; a risingpatenting activity, partially driven by universities.

5.1 impact of public funding

The close linkage between scientific knowledge and the rise of techno-logy arguably enabled the market economies of Western nations toachieve increasing prosperity11.

Research and Development,” Brookings Papers on Economic Activity 18, no. 3 (1987):783–832 quantifying latter cost to up to 50-75% of the cost of the original invention.

6. Furthermore, the cost of external capital might accentuate the market failure,see Bronwyn H. Hall, “The Financing of Research and Development,” Oxford Reviewof Economic Policy 18, no. 1 (2002): 35–51.

7. José Ángel Zúñiga-Vicente et al., “Assessing the Effect of Public Subsidies onFirm R&D Investment: A Survey,” Journal of Economic Surveys 28, no. 1 (2014): 36–67.

8. Public funding represented 32.3% of the EU28 and 27.7% of the US total R&Dspendings in 2014, data available under http://ec.europa.eu/eurostat/statisti

cs-explained/index.php/R_%26_D_expenditure.9. See John H. Marburger, “Wanted: Better Benchmarks,” Science 308, no. 5725

(2005): 1087–1087 with a call for econometric models with a large variables in alarger number of countries; see Adam B. Jaffe, “Building Program Evaluation Intothe Design of Public Research Support Programs,” Oxford Review of Economic Policy18, no. 1 (2002): 22–34 for encouraging program design for evaluation that wouldproduce data less subject to selection bias.

10. Joshua D. Sarnoff, “The Likely Mismatch Between Federal Research Develop-ment Funding and Desired Innovation,” Vanderbilt Journal of Entertainment Techno-logy Law 18, no. 2 (2016): 369-371.

11. In the words of Mariana Mazzucato, The Entrepreneurial State: Debunking Publicvs. Private Sector Myths (Anthem Press, 2013), p. 6 ‘nearly all the technological revo-

134 introductory remarks

“physical science and industrialism may be conceivedof as a pair of dancers, both of whom know their steps,and have an ear for the rhythm of the music. If the part-ner who has been leading chooses to change parts and tofollow instead, there is perhaps no reason to expect thathe will dance less correctly than before” – Toynhee (1962)12

There are two main channels through which public investments inscience may impact technological development13. These two differentstrands may be described as the indirect and direct applied role ofgovernment funding. Firstly, indirectly, by way of its informationalvalue, the knowledge created may affect innovation by private actors.The ‘transport’ mechanism of these spillover effects from universi-ties to firms may be journal publications and reports, or any formof knowledge transfer such as public conferences, meetings and in-formal conversation between researchers or consulting activity14. Infact, about a third of the federally funded grants generate articles thatare cited by patents15. An additional investment of USD 10 million ingrants generates 2.3 additional private-sector patents16. Secondly, go-

lutions in the past – from the Internet to today’s green tech revolution – requireda massive push by the State.’ See also Vannevar Bush, Science, the Endless Frontier;A Report to the President on a Program for Postwar Scientific Research (National ScienceFoundation, 1960); Nathan Rosenberg, Exploring the Black Box (Cambridge UniversityPress, 1994); and Francis Narin and Dominic Olivastro, “Status Report: Linkage bet-ween Technology and Science,” Research Policy 21, no. 3 (1992): 237–249 for the linkbetween scientific knowledge and national prosperity.

12. Arnold J. Toynhee, A Study of History, vol. I (Oxford University Press, 1962).13. For a range of other government set innovation policy tools, such as prizes,

tax credits, or R&D subsidies, see Daniel J. Hemel and Lisa L. Ouellette, “Beyondthe Patents-Prizes Debate,” Texas Law Review 92, no. 2 (2006): 303–382; for prizes, seeMichael J. Burstein and Fiona Murray, “Innovation Prizes in Practice and Theory,”Harvard Journal of Law and Technology 29, no. 2 (2016): 401–452.

14. Adam B. Jaffe, “Real Effects of Academic Research,” American Economic Review79, no. 5 (1989): 957–70 finds geographically mediated ‘spillovers’ from universityresearch to commercial innovation particularly in areas such as drugs and medi-cal technology; but see Wesley M. Cohen, Richard R. Nelson, and John P. Walsh,“Links and Impacts: The Influence of Public Research on Industrial R&D,” Manage-ment Science 48, no. 1 (2002): 1–23 about the difficulty to measure knowledge and itslink to downstream outcomes, as well as with determining causality.

15. Danielle Li, Pierre Azoulay, and Bhaven Sampat, “The Applied Value of PublicInvestments in Biomedical Research,” Science 356 (2017): 78–81.

16. See Pierre Azoulay et al., “Public R&D Investments and Private-Sector Paten-ting: Evidence from NIH Funding Rules,” NBER Working Paper, no. 20889 (2015)

5.1 impact of public funding 135

vernmental spending directly increases innovative activity throughthe patenting of novel inventions by funded scientists. Across allfields, about one of ten federal grant generates a patent17. Similarly,a tenth of the drugs approved by the Food and Drug Administra-tion are covered by a public-sector patent18. According to these stu-dies, university research had a smaller direct effect on industrial R&Doutside a few technology fields such as drugs, while the indirect ef-fect was substantial. As such, science seems to be mostly employedas a set of tools and stock of knowledge to be tapped in problem-solving19.

The lower level of direct connections between scientific and technicaladvances is accentuated by the fact that interactions between publicand private players are complex and subtle, the lags are long, thefeedbacks intricate, and deriving causality in studies is complex20.Public investments are particularity important in fields where the pri-vate sector has inappropriate incentives to invest in21. Thus, researchareas characterized by scientific uncertainty, high R&D costs or diffi-culty of appropriation of basic knowledge are especially dependenton public funds, and may, in a second step, encourage crowding-inof private sector investments, i.e., encourage private firms to makecomplementary investments in R&D22.

who derive this rate by using the link between federally funded grants with thepublications they support and the patents citing those publications.

17. Li, Azoulay, and Sampat, “The Applied Value of Public Investments in Biome-dical Research.”

18. Bhaven N. Sampat and Frank Lichtenberg, “What are the Respective Rolesof the Public and Private Sectors in Pharmaceutical Innovation?,” Health Affairs 30

(2011): 332–339.19. This observation is shared by Michael Gibbons and Ron Johnston, “The Roles

of Science in Technological Innovation,” Research Policy 3, no. 3 (1974): 220–242 andAlvin K. Klevorick et al., “On the Sources and Significance of Interindustry Differen-ces in Technological Opportunities,” Research Policy 24, no. 2 (1995): 185–205.

20. The difficulties in capturing the link between basic research and private profitare discussed, e.g., by Richard R. Nelson, “Institutions Supporting Technical Advancein Industry,” American Economic Review 76, no. 2 (1986): 186–189 and Harold Varmus,The Art and Politics of Science (W. W. Norton & Company, 2009).

21. Nicholson W. Price, “Grants,” Berkeley Technology Law Journal forthcoming(2019).

22. See Li, Azoulay, and Sampat, “The Applied Value of Public Investments inBiomedical Research.”

136 introductory remarks

Grants typically rely on public officials to determine the goals to beachieved, the level of subsidization of a particular technology and theperson or team best at reaching this goal23. As such, when the costsand benefits of a potential invention are not observable and hardlyforeseeable by the sponsor, the rewards run the risk to lead to under-or over-investment24. Decentralized systems, such as patents and pri-zes, may on the contrary enable for outsiders to become innovators25.Agencies attributing grants have also been accused to suffer of beingsubject to political pressure26 and mismanagement27. And, once thedecision to fund has been made, the government pays whether or notthe R&D produces any returns28. As innovation mechanism, grantsappear most effective when the government has a comparative ad-

23. See Hemel and Ouellette, “Beyond the Patents-Prizes Debate,” p. 327 forgovernment-set vs. market-set transfers.

24. See Nancy Gallini and Suzanne Scotchmer, “Intellectual Property: When Is Itthe Best Incentive System?,” in Innovation Policy and the Economy, Volume 2, NBERChapters (National Bureau of Economic Research, 2002), 51–78 arguing that whenvalue and cost are not observable by the sponsor, the best mechanism is probably IP;see Kerry Grens, “An Economic Gamble: What Does Society Get for the Billions itSpends on Science?,” The Scientist, 2007, quoting Pierre Azoulay ‘we want to knowif we are underinvesting or overinvesting in life sciences research’.

25. For the virtue of decentralization, see Jonathan H. Adler, “Eyes on a ClimatePrize: Rewarding Energy Innovation to Achieve Climate Stabilization,” Harvard En-vironmental Law Review 35, no. 1 (2011): p. 13-14 citing a McKinsey & Companyreport: ‘[t]he history of science is replete with instances of outsiders proposing noveland ultimately revolutionary solutions to problems that had vexed insiders’.

26. See Josh Lerner and Colin Kegler, “Evaluating the Small Business InnovationResearch Program: A Literature Review,” in The Small Business Innovation ResearchProgram: An Assessment of the Department of Defense Fast Track Initiative, ed. Charles W.Wessner (National Research Council, 2000), p. 315-317 for a review of the literatureon distortions in the award process and on the theory of regulatory capture.

27. See Adler, “Eyes on a Climate Prize: Rewarding Energy Innovation to AchieveClimate Stabilization,” p. 29 and Benjamin K. Sovacool, “Replacing Tedium withTransformation: Why the US Department of Energy Needs to Change the Way it Con-ducts Long-Term R&D,” Energy Policy 36, no. 3 (2008): 923–928 on the risk-aversion offederal agencies; see Cohen, Nelson, and Walsh, “Links and Impacts: The Influenceof Public Research on Industrial R&D” on the distortion due to officials selectingfirms on their likelihood of success in order to claim credit for the firms, regardlessof the impact of the public fund.

28. John Alic, David Mowery, and Edward Rubin, U.S. Technology and InnovationPolicies: Lessons for Climate Change, (Report Prepared fo the Pew Center on GlobalClimate Change, 2003), p. 11 arguing that the government has commonly failed atdefining technical attributes and design features to pick winners in the marketplace.

5.2 rise of academic patenting 137

vantage in the cost and benefit analysis of potential projects29. Grantsbeing ex ante rewards financed by general revenues, they are to befavored where markets proxy poorly for the social benefits of newinventions, where such inventions would require large capital inves-tments, or where cross-subsidization by nonusers is desirable30. De-pending on the specific research environment, the optimal incentiveschemes vary. Grants thus act an alternative or a complementary toolto patents, prizes and tax credit mechanisms31.

5.2 rise of academic patenting

Over the last decades, a major policy aim has been to foster greaterinteraction between public research and industry in order to increasethe social and private returns from public funding. In parallel, a gene-ral strengthening of patents laws and legislations improving techno-logy transfer has accompanied the expansion of academic patenting32.In 1980, the United States adopted the Bayh-Dole Act, which grantedrecipients of federal funds the right to patent inventions and licensethem to firms.

29. See Hemel and Ouellette, “Beyond the Patents-Prizes Debate,” p. 375 and Bur-stein and Murray, “Innovation Prizes in Practice and Theory,” p. 449-450: when un-certainty and information asymmetry are low, grants are likely to be a good solution.

30. See Hemel and Ouellette, “Beyond the Patents-Prizes Debate,” p. 375 namingspace exploration an optimal field for grants, with a large NASA institutional know-ledge, capital-intensity and large social benefits in detecting and deflecting asteroidsthat would otherwise collide with Earth.

31. There is an extensive literature on the economics of invention incentives, see,e.g., Brian Wright, “The Economics of Invention Incentives: Patents, Prizes, and Rese-arch Contracts,” American Economic Review 73, no. 4 (1983): 691–707; Eric A. de Laat,“Patents or Prizes: Monopolistic R&D and Asymmetric Information,” InternationalJournal of Industrial Organization 15, no. 3 (1997): 369–390; Suzanne Scotchmer, “Onthe Optimality of the Patent Renewal System,” The RAND Journal of Economics 30, no.2 (1999): 181–196; Gallini and Scotchmer, “Intellectual Property: When Is It the BestIncentive System?”; and Hugo Hopenhayn, Gerard Llobet, and Matthew Mitchell,“Rewarding Sequential Innovators: Prizes, Patents, and Buyouts,” Journal of PoliticalEconomy 114, no. 6 (2006): 1041–1068 on patents and prizes.

32. For a thorough survey on theoretical and empirical evidence with regard to therelationship between the strength of patent protection and technology transfer, seeBronwyn H. Hall, “Does Patent Protection Help or Hinder Technology Transfer?,”chap. 2 in Intellectual Property for Economic Development, ed. Sanghoon Ahn, BronwynH. Hall, and Keun Lee, Chapters (Edward Elgar Publishing, 2014), 11–32.

138 introductory remarks

The Bayh-Dole Act was a reaction to the fact that of its nearly 30’000

patents, the US government commercialized less than 5% throughlicenses33. For the following twenty years, the patenting activity ofuniversities more than doubled every five year34. Other countries fol-lowed this trend and reformed funding regulations and employmentlaws to enable research institutions to file, own and license the in-tellectual property rights derived from government research funds.Austria, Germany, Denmark and Japan abolished the ‘professor’s pri-vilege’, under which academics owned the patent. A share of theroyalty is nowadays given to the academic inventors in exchange35.

Single patents, such as DNA cloning developed at Stanford Univer-sity and co-transformation of eukaryotic cells at Columbia University,or more trivial ones such as the Gatorade drink at the University ofFlorida provide for millions of dollars in licensing revenues36. Whileuniversity licensing revenues have risen starkly,37, most US universi-ties do not generate enough licensing revenues to cover the operating

33. See US Government Accounting Office, Report to Congressional Committees,Technology Transfer, Administration of the Bayh-Dole Act by Research Universities, (1978)for further historical and institutional details.

34. Patenting mapped by Richard R. Nelson, “Observations on the Post-Bayh-Dole Rise in University Patenting,” Journal of Technology Transfer 26 (2001): 13–19;Innovation’s Golden Goose, The Economist, December 2002: ’possibly the most inspi-red piece of legislation to be enacted in America over the past half-century’ as it‘helped to reverse America’s precipitous slide into industrial irrelevance’.

35. See Dov Greenbaum and Christopher Scott, “Hochschullehrerprivileg - A Mo-dern Incarnation of the Professor’s Privilege to Promote University to IndustryTechnology Transfer,” Science, Technology and Society 15, no. 1 (2010): 55–76 for a criticof this trend and Olof Ejermo and John Källström, “What is the Causal Effect of R&Don Patenting Activity in a “Professor’s Privilege” Country? Evidence from Sweden,”Small Business Economics 47, no. 3 (2016): 677–694 finding empirical support for theprofessor’s privilege in Sweden.

36. The gene splicing method of the Cohen-Boyer patents collected near USD $255

mio. in licensing fee, see Maryann P. Feldman, Alessandra Colaianni, and ConnieK. Liu, “Lessons from the Commercialization of the Cohen-Boyer Patents: The Stan-ford University Licensing Program,” in Intellectual Property Management in Health andAgricultural Innovation: A Handbook of Best Practices (2008), 1208–1797; the methods forinserting DNA into eukaryotic cells yielded a total of USD 790 million in licensingfee, see Alessandra Colaianni and Robert Cook-Deegan, “Columbia University’s AxelPatents: Technology Transfer and Implications for the Bayh-Dole Act,” The MilbankQuarterly 87, no. 3, 683–715.

37. University licensing revenues went from USD 221 mio. (1991) to USD 698 mio.(1997), Nelson, “Observations on the Post-Bayh-Dole Rise in University Patenting.”

5.3 technology transfer 139

costs of their technology transfer offices38. The trend line towards pa-tenting and commercialisation has not been without detractor39. Theavailability of the option to patent may alter the incentive to contri-bute to public goods, i.e., non-excludable advances to the scientificliterature. Academic patenting might shift the research agenda fromfundamental towards commercial, slow down the rate of dissemina-tion of research findings, or impact the trade-off between publishingand patenting40.

5.3 technology transfer

Against the backdrop of the rise of university patenting, the need forefficient technology transfer increased over the years. Under the linearmodel of innovation, the knowledge and technology in the hands ofresearch-based organizations, such as universities, must be transfer-red to organizations that produce goods for sale in the marketplace.Even considering more mixed models, under which upstream rese-arch spawns new research projects, but downstream technical advan-ces also shape upstream research41, the transfer of academic intellec-tual property is key. Universities have traditionally relied on contrac-tual licensing as predominant mode of diffusion42.

38. Walter D. Valdivia, University Start-Ups: Critical for Improving Technology Transfer,(Center for Technology Innovations at Brookings, 2013) finding that the top 10% (16

universities) make three-quarter of all academic licensing gross income.39. Editorial, Patents Pending, 11 (Nature Materials, 2012) making the case that aca-

demic patenting should shift from a simple licensing revenues model towards thegoal of making an impact.

40. See Jerry Thursby and Marie Thursby, “Who Is Selling the Ivory Tower? Sourcesof Growth in University Licensing,” Management Science 48, no. 1 (2002): 90–104 fora model examining whether the increased licensing results from the increase in thewillingness to license or from the change in academic research; see Pierre Azoulay,Waverly Ding, and Toby. Stuart, “The Impact of Academic Patenting on the Rate,Quality and Direction of (Public) Research Output,” Journal of Industrial Economics57, no. 4 (2009): 637–676 finding that patenting had a positive effect on both the rateand quality of publications.

41. See, e.g., Cohen, Nelson, and Walsh, “Links and Impacts: The Influence of PublicResearch on Industrial R&D” examining the manufacturing sector.

42. Rebecca L. Siegel, Kimberly D. Miller, and Ahmedin Jemal, “Cancer Statistics,2018,” CA: A Cancer Journal for Clinicians 68, no. 1, 7–30; other modes of transferinclude published publications and reports, or informal exchanges at conferences ormeetings.

140 introductory remarks

Alternatively, entities can collaborate directly. R&D collaborations areconsidered crucial for achieving product innovation because they pro-vide the firms with access to external knowledge needed to innovatetheir own products43. In high technologies especially, licensing, strate-gic alliances, and spinoffs have placed universities as driving force44.Major R&D collaborations were launched during World War II in thepharmaceutical, petrochemical, synthetic rubber and atomic weaponsindustry45. With the passage of the Cooperative Research and Deve-lopment Agreement in 1986 shortly after the Bayh-Dole Act, federallaboratories were encouraged to collaborate with non-federal partiesby providing for personnel, services, facilities and equipment – butnot funding – toward the conduct of a specified development effort.The non-federal partner can negotiate an exclusive or non-exclusivelicense to any invention resulting from the collaborative research. Thebenefits from such collaboration range from enabling parties to cap-ture knowledge spillovers, reduce duplication and support the exploi-tation of economies of scale46.

5.4 studies

The second set of papers of this dissertation have a common startingpoint: the governmental funding of science. The rise of academic fun-ding led to an increasing number of patents, that called for increasingtechnology transfer. Identified here are two key patterns to foster in-novation in emerging technologies, public funding and technologytransfer.

43. Annique C. Un and Kazuhiro Asakawa, “Types of R&D Collaborations and Pro-cess Innovation: The Benefit of Collaborating Upstream in the Knowledge Chain,”Journal of Product Innovation Management 32 (2015): 138–153 classifying four typesof R&D collaborations (with universities, suppliers, competitors, and customer) interms of their position in the knowledge chain and contextual knowledge distance.

44. Susan Rosegrant, David Ralph David Lampe, and Robert Lampe, Route 128:Lessons from Boston’s High-Tech Community (Basic Books, 1992) and Annalee Saxenian,Regional Advantage: Culture and Competition in Silicon Valley and Route 128 (HarvardUniversity Press, 1996).

45. David C. Mowery, “Collaborative R&D: How Effective Is It?,” Issues in Scienceand Technology 15, no. 1 (1998): 37–44.

46. Michael L. Katz, “An Analysis of Cooperative Research and Development,” TheRAND Journal of Economics 17, no. 4 (1986): 527–543.

5.4 studies 141

‘Patent Thickets in Nanotechnology’ considers nanotechnology as a casestudy for a field where patents are thought to block commercializa-tion advances. A web of overlapping intellectual property rights rai-ses entry costs; influencing firms in their decisions whether or not toenter the nano-space. Accordingly, the paper analyzes the presence ofpatent thickets in Europe and in the US, as well as ways firms mightmitigate the effects of thickets through collaboration.

‘Federal Funding in Cancer Research’ shifts its focus on the public fun-ding itself. The paper investigates the class of disease that benefitedfrom most attention and funding, cancer. By studying a field drivenby governmental investments, the investigation enables us to compre-hend the efficacy of funding of emerging technologies over time. Theexamination of nanotechnology and cancer research provides grea-ter understanding about the impact of public funding on emergingtechnologies.

6PAT E N T T H I C K E T S I N N A N O T E C H N O L O G Y

*This Chapter was co-authored with Marius Fischer, Max Planck Institutefor Innovation and Competition.

6.1 introduction

Nanotechnology is said to bring about the next technological revolu-tion. Concerned with the scale set by a billionth of a meter and, in par-ticular, control of matter at that scale, nanotechnology does not onlypromise fundamental scientific insights, but also novel applicationsranging from materials to medicine. In fact, as nanotechnology tou-ches upon – depending on the definition even includes – single mo-lecules and their manipulation, it impacts virtually every single dis-cipline. While the concept of nanotechnology is usually traced backto Feynman (1960) and his notoriously famous speech ‘There’s Plentyof Room at the Bottom’1, the term itself was introduced by Taniguchi(1974) only roughly 15 years later2. Today, the term nanotechnologypops up in discussions concerning dirt-repellent surfaces, sustainablewater supplies and elevators into space alike. Not least due to the in-fluence of Drexler (1986) and his apocalyptic and dystopian view of‘nanobots’ feeding on carbon, in turn extinguishing humankind3, dis-cussions about nanotechnology have long left laboratories, resultingin what might right now well be labeled a hype4.

1. Richard Feynman, “There’s Plenty of Room at the Bottom,” Engineering andScience 23, no. 5 (1960): 22–36.

2. Norio Taniguchi, “On the Basic Concept of ‘Nano-Technology’,” Proceedings ofthe International Conference on Production Engineering, Tokyo, 1974, ‘[n]ano-technologyis the production technology to get the extra high accuracy and ultra fine dimension,i.e., the preciseness and finesse in the order of 1 nm (nanometer), 10−9 in length’.

3. Eric Drexler, Engines of Creation: The Coming Era of Nanotechnology (AnchorBooks, 1986).

4. See Marius Fischer, Upstream-Patente in der Nanotechnologie, Ein Vergleich zwis-chen Europa und den USA (forthecoming) and David M. Berube, Nano-Hype: The TruthBehind the Nanotechnology Buzz (Prometheus Books, 2006).

142

6.2 nanotechnology and the patent system 143

However, despite a profound socio-economic interest in nanotechno-logy, and a market volume in the billions5, the expected revolutionfailed to materialize so far. While there are quite a few products that– not least due to the breadth of the definitions usually referred to –technically qualify as nanotechnology, the industry is arguably stillin its infancy. More than 55 years have passed, and yet the NobelPrize in Chemistry 2016 was awarded for pioneering work on nano-technology. Its documentation explicitly notices this by stating that‘we are at the dawn of a new industrial revolution of the twenty-firstcentury’6. As Morris (2016) puts it, nanotechnology seems to be stuckin the status of emerging science for decades, failing to make its wayfrom invention to innovation7.

6.2 nanotechnology and the patent system

There is a significant amount of literature – especially originatingfrom the US – blaming the patent system for the slow progressionof nanotechnology from conceptual invention to seizable innovation8.The hype that emerged around nanotechnology and its sheer infi-

5. Estimates for the global market for nanotechnology vary considerably, from fo-recasts of USD 64 billion by 2019 by Nanotechnology: A Realistic Market Assessment,(BCC Research, 2014) to USD 4.4 trillion in 2018 by Nanotechnology Update: Corpora-tions Up Their Spending as Revenues for Nano-Enabled Products Increase, (Lux ResearchInc., 2014).

6. Royal Swedish Academy of Sciences, Background on the Nobel Prize in Chemistry2016: Molecular Machines, (2016), p. 47.

7. Emily Morris, “The Irrelevance of Nanotechnology Patents,” Connecticut LawReview 49 (2016): 499–551.

8. See Ted Sabety, “Nanotechnology Innovation and the Patent Thicket: Which IPPolicies Promote Growth?,” Nanotechnology Law & Business Volume 1.3 (2004): 262–283; Raj Bawa and Stephen Maebius, “The Nanotechnology Patent "Gold" Rush,”Journal of Intellectual Property Rights 10 (2005): 426–433; Mark A. Lemley, “PatentingNanotechnology,” Stanford Law Review 58 (2005): p. 618; John Miller et al., The Hand-book of Nanotechnology (John Wiley / Sons, 2004), p. 224; Maurice Schellekens, “Pa-tenting Nanotechnology in Europe: Making a Good Start? An Analysis of Issuesin Law and Regulation,” Journal of World Intellectual Property 13, no. 1 (2010): p. 47;Amber Stiles, “Hacking through the Thicket: A Proposed Patent Pooling Solutionto the Nanotechnology Building Block Patent Thicket Problem,” Drexel Law Review 4

(2012): 555–592; and André Sabellek, Patente auf nanotechnologische Erfindungen (MohrSiebeck, 2014), p. 32.

144 patent thickets in nanotechnology

nite possibilities has sparked a patent ‘land-grab’9 or ‘gold-rush’10 inwhich market participants try to gain control over as much of thenew discipline as early as possible. Hence, a patent thicket, often de-fined with Shapiro (2001) as a ‘dense web of overlapping intellectualproperty rights that a company must hack its way through in orderto actually commercialize new technology’11, is born, tangling up in-novation instead of setting it free. Nanotechnology patents largelycover components of modular and complex technologies, whilst thefunctionalities of many of these components may often partially oreven completely overlap12. In addition, patents protecting these com-ponents may then overlap, too13. And when such overlapping rightsbelong to several firms, ultimately a particularly vicious patent thic-ket emerges14. The consequences of such a thicket can be severe, andnanotechnology is believed to already suffer them. Firstly, a web ofoverlapping intellectual property rights increases entry costs, keeping

9. Peter J. Paredes, “Written Description Requirement in Nanotechnology: Cle-aring a Patent Thicket?,” Journal of the Patent & Trademark Office Society 88 (2006):p. 497.

10. See Bawa and Maebius, “The Nanotechnology Patent "Gold" Rush”; Joel D’Silva,“Pools, Thickets and Open Source Nanotechnology,” European Intellectual PropertyReview 31, no. 6 (2009): p. 300; and Graham Reynolds, “Nanotechnology and theTragedy of the Anticommons: Towards a Strict Utility Requirement,” University ofOttawa Law & Technology Journal 6 (2009): p. 96.

11. Carl Shapiro, “Navigating the Patent Thicket: Cross Licenses, Patent Poools,and Standard Setting,” in Innovation Policy and the Economy, vol. 1 (MIT Press, 2001).

12. See Bronwyn H. Hall et al., A Study of Patent Thickets, (Intellectual PropertyOffice, 2013).

13. Edward J. Egan and David J. Teece, “Untangling the Patent Thicket Literature,”University of California Berkeley Tusher Center for Management Intellectual Capital Wor-king Paper, no. 7 (2015): the overlap can be horizontal when patents are adjacent toeach other or vertical when related through cumulative innovation.

14. The number of different owners does not have to be large in order for a patentthicket to emerge, see Dan L. Burk and Mark A. Lemley, “Policy Levers in PatentLaw,” Virginia Law Review 89, no. 7 (2003): p. 1627 and Dan L. Burk and Mark A.Lemley, The Patent Crisis and How the Courts Can Solve It (The University of ChicagoPress, 2009), p. 89. More than three firms owning overlapping rights might alreadybe enough; for the ‘triplets’ approach, see Georg von Graevenitz, Stefan Wagner,and Dietmar Harhoff, “Incidence and Growth of Patent Thickets: The Impact ofTechnological Opportunities and Complexity,” The Journal of Industrial Economics 61,no. 3 (2013): 521–563.

6.2 nanotechnology and the patent system 145

firms out of the market altogether15. Secondly, since they create sub-stantial transaction costs, they affect the decisions of actors to conti-nue their operations in the marketplace16.

The nanotechnology discipline is seen by many to be particularlyprone to patent thickets and its impedimentary effects. In contrastto biotechnology17, the patentability of nanotechnological inventionshas never been seriously challenged18. While ethical concerns largelyprevented privatization of the building blocks of biotechnology, thesebuilding blocks were increasingly patented in nanotechnology19. Thisincrease in patenting activity is also thought to result from the highfragmentation of the nanotechnology sector, with a large number ofactors seeking intellectual property protection20. The whole problemis exacerbated by the multi- rather than interdisciplinary nature ofnanotechnology that may spawn overlapping and low quality pa-tents21. Finally, since nanotechnology is a research-driven field, pa-

15. Bronwyn H. Hall, Christian Helmers, and Georg von Graevenitz, “TechnologyEntry in the Presence of Patent Thickets,” Institute for Fiscal Studies Working Papers,no. W16/02 (2016).

16. See Bronwyn H. Hall and Rosemarie H. Ziedonis, “The Patent Paradox Revi-sited: An Empirical Study of Patenting in the U.S. Semiconductor Industry, 1979-1995,” The RAND Journal of Economics 32, no. 1 (2001): 101–28 and Egan and Teece,“Untangling the Patent Thicket Literature” who study 164 definitions of patent thic-kets and finds that the four underlying economic issues most commonly referredto are saturated invention space, diversely-held complementary inputs, overlappingpatents and moral hazard due to low quality patents.

17. Although, in a way, biotechnology may be seen as nanotech’s predecessor interms of harsh tensions between modern, cumulative innovation and exclusionaryrights, as well as in terms of technological evolutionary patterns, see Corine Genet,Khalid Errabi, and Caroline Gauthier, “Which Model of Technology Transfer for Na-notechnology? A Comparison with Biotech and Microelectronics,” Nanotechnology:Introducing the Future, Technovation 32, no. 3 (2012): p. 224.

18. See, e.g., Sabety, “Nanotechnology Innovation and the Patent Thicket: Which IPPolicies Promote Growth?,” p. 506.

19. As far as we can tell, this was originally observed by Lemley, “Patenting Nano-technology,” p. 605, 606, 618.

20. See Vivek Koppikar, Stephen B. Maebius, and Steven Rutt, “Current Trendsin Nanotech Patents: A View from Inside the Patent Office Intellectual Property,”Nanotechnology Law & Business 1 (2004): p. 27 arguing that most patents are held bydifferent entities and cover similar inventions; the same observation is made by areport assessing c. 1000 US nanotechnology patents, see Nanotechnology Gold RushYields Crowded, Entangled Patents, (Lux Research and Foley & Lardner LLP, 2005).

21. Nanotechnologies are both multidisciplinary, see, e.g., Jyoti S. Bhat, “Concernsof New Technology Based Industries – The Case of Nanotechnology,” Technovation

146 patent thickets in nanotechnology

tents are frequently filed at an early stage, and with broad claims,which may result in overlapping and/or invalid patents22. As such,over-patenting practices by universities have been accused to contri-bute to the formation of the patent thicket in nanotechnology23.

However, commentators – rightly – point out that there is more toinnovation than patents. Alternative intellectual property protection,such as trade secrets, can complement patenting strategies. Maintai-ning key technical information such as research data, lab notebookentries, experimental techniques and failed experiments as trade se-crets can be an effective strategy to develop nanotechnology intellec-tual property assets24. Trade secret misappropriation lawsuits againstformer employees disputes include large settlements and damagesaward25. That being said, the absence of truly revolutionary, nanotech-enabled products is presumably rooted in a variety of factors thatmay or may not be connected to the patent system at all. The mostimportant and almost obvious of these are intrinsically long deve-lopment cycles, limited access to equipment and infrastructure, aswell as inefficient tech-transfer mechanisms that – assuming a simplebut paradigmatic model of technological innovation – get nanotechfrom the hands of universities and public research institutions intothose of commercial actors26. It seems virtually impossible, therefore,to conclusively settle the question whether or not patents impede

25, no. 5 (2005): 457–462 and cross-industrial in application, see, e.g., Jonathan D.Linton and Steven T. Walsh, “Integrating Innovation and Learning Curve Theory:An Enabler for Moving Nanotechnologies and Other Emerging Process Technologiesinto Production,” R&D Management 34, no. 5 (2004): 517–526.

22. See Fischer, Upstream-Patente in der Nanotechnologie, Ein Vergleich zwischen Eu-ropa und den USA for a more detailed investigation of the multidisciplinary andresearch-driven properties of nanotechnology and their possible implications for theemergence of a patent thicket.

23. Stiles, “Hacking through the Thicket: A Proposed Patent Pooling Solution to theNanotechnology Building Block Patent Thicket Problem” and Lemley, “PatentingNanotechnology.”

24. Scott W. Cummings, “The Role of Trade Secrets in Today’s NanotechnologyPatent Environment,” Nanotechnology Law & Business 5 (2008): 41–52.

25. See Lisa L. Ouellette, “Nanotechnology and Innovation Policy,” Harvard Journalof Law & Technology 29 (2015): 58-59 for an account of three lawsuits.

26. For the corresponding discussion, see Morris, “The Irrelevance of Nanotechno-logy Patents”; see Genet, Errabi, and Gauthier, “Which Model of Technology Transferfor Nanotechnology? A Comparison with Biotech and Microelectronics” finding thatlarge firms are central in nanotechnology co-patenting.

6.3 nanotechnology and collaborations 147

nanotechnology innovation without recourse to empiricism. But besi-des anecdotal evidence, such as the reported existence of 250 patentsissued on buckminsterfullerene nanoparticle technology27, no suchempirical evidence exists to the best of our knowledge.

6.3 nanotechnology and collaborations

Patent thickets are characterized in that they rise transaction and se-arch costs. In patent-intensive high-technology industries, it is com-mon that complex products require from a hundred to several thou-sand patented inputs28. Dealing with a thicket therefore requires coope-ration with rival innovators – which implies transaction costs29. Ty-pically, this is achieved trough cross-licensing. A cross-license is anagreement between two companies that grant each other the right topractice the patent of the other30. The exchange of complementarytechnologies in this form facilitates the product development and de-creases the risk of being held-up31. Similarly, patent pools have fre-quently been described as remedy to patent thicket. In a patent pool,

27. See Albert Halluin and P. Westin Lorelei, “Nanotechnology: The Importance ofIntellectual Property Rights in an Emerging Technology,” Journal of the Patent andTrademark Office Society 86 (2004): p. 229.

28. Egan and Teece, “Untangling the Patent Thicket Literature”; Iain M. Cockburnand Megan J. MacGarvie, “Patents, Thickets and the Financing of Early-Stage Firms:Evidence from the Software Industry,” Journal of Economics & Management Strategy18, no. 3, 729–773: ‘[o]ne salient feature of patent thickets is the potential for highercosts associated with negotiating with many parties. To the extent that there arefixed costs of conducting a negotiation, having to deal with more parties will driveup costs of obtaining licenses. There may also be transactions costs associated withbargaining and coordinating negotiations with multiple licensors’.

29. See Shapiro, “Navigating the Patent Thicket: Cross Licenses, Patent Poools, andStandard Setting,” 122 explaining that antitrust law even increases the transactioncosts with its hostility to cooperation among horizontal rivals when it is preciselycooperation that is required to ‘navigate the patent thicket’.

30. For cross-licensing as a mean to resolve the threat of hold-ups in patent thickets,see Peter C. Grindley and David J. Teece, “Managing Intellectual Capital: Licensingand Cross-Licensing in Semiconductors and Electronics,” California Management Re-view 39 (1997): 8–41; Shapiro, “Navigating the Patent Thicket: Cross Licenses, PatentPoools, and Standard Setting”; and Bharat Anand and Tarun Khanna, “The Structureof Licensing Contracts,” Journal of Industrial Economics 48, no. 1 (2000): 103–135.

31. Chaim Fershtman and Morton I. Kamien, “Cross Licensing of ComplementaryTechnologies,” International Journal of Industrial Organization 10, no. 3 (1992): 329–348.

148 patent thickets in nanotechnology

the patent rights are aggregated by patent owners and made availa-ble to member and non-member licensees32. Such ‘one-stop shopping’conserves on the cost of licensing from dispersed patent holders andreduces the likelihood a patent holder strategically hold out for exor-bitant licensing fee33. A further form of cooperation for innovators isto collaborate on R&D projects34. Ex-ante cooperation helps to pre-vent the formation of a hold-up by avoiding patent infringement liti-gation, reducing transaction costs and promoting innovation35. Suchjoint research and development enhances efficiency and innovationthrough the pooling of resources36.

Assuming the previously described patent thickets, innovation ac-tors might employ public and private collaboration as a mean of ci-rcumventing the thicket produced blockade. Through collaboration,

32. Patent Pools and Antitrust – A Comparative Analysis, (World Intellectual PropertyOrganization (WIPO), March 2014); Josh Lerner and Jean Tirole, “Efficient PatentPools,” American Economic Review 94, no. 3 (2004): 691–711; and Ryan Lampe and Pe-tra Moser, “Do Patent Pools Encourage Innovation? Evidence from 20 Industries inthe 1930s,” in Standards, Patents and Innovations, ed. Timothy Simcoe, Ajay K. Agra-wal, and Stuart J. Graham (NBER Books, 2014).

33. Robert P. Merges and Michael Mattioli, “Measuring the Costs and Benefits ofPatent Pools,” Ohio State Law Journal 78 (2017): 281–347.

34. Ralph Siebert and Georg von Graevenitz, “How Licensing Resolves Hold-Up:Evidence from a Dynamic Panel Data Model with Unobserved Heterogeneity,” CEPRDiscussion Papers, no. 5436 (2006).

35. See Siebert and von Graevenitz, “How Licensing Resolves Hold-Up: Evidencefrom a Dynamic Panel Data Model with Unobserved Heterogeneity,” Ralph Siebertand Georg von Graevenitz, “Are Licensing Agreements Appropriate Instrumentsto Cut Through the Patent Thicket?,” Working Paper, 2011, and Anna Kingsbury,“Patent Collaboration: Licensing, Patent Pools, Patent Commons, Open Source andCommunities of Innovation,” New Zealand Intellectual Property Journal, September2013, 3–9.

36. The literature strategic alliances suggests that firms might access, through colla-boration, to complementary assets (e.g., Shona L. Brown and Kathleen M. Eisenhardt,“Product Development: Past Research, Present Findings, and Future Directions,” Aca-demy of Management Review 20, no. 2 (1995): 343–378), know-how (e.g., Rene Belderboset al., “Co-Ownership of Intellectual Property: Exploring the Value-Appropriationand Value-Creation Implications of Co Patenting With Different Partners,” ResearchPolicy 43, no. 5 (2014): 841–852), technologies (e.g., Pierre Mohnen and Cathy Hoa-reau, “What Type of Enterprise Forges Close Links with Universities and Govern-ment Labs? Evidence from CIS 2,” Managerial and Decision Economics 24, nos. 2-3(2003): 133–145) and enhance learning capabilities (e.g., Wolfgang Becker and Jur-gen Dietz, “R&D Cooperation and Innovation Activities of Firms – Evidence for theGerman Manufacturing Industry,” Research Policy 33, no. 2 (2004): 209–223).

6.4 research questions 149

the technology is transferred from public to private hands. Theseagreements decrease transactions costs, as the jointly filed patent canbe developed directly by the commercial partner. Among a range ofdisciplines such as chemistry, mechanical engineering, and biotechno-logy, nanotechnology figures among the main beneficiaries of this in-flux of private capital. In the science-based field of nanotechnology,corporations partner with universities and public sector agencies, fre-quently on multiple projects37. To name but a few examples, IBMannounced a $90 million strategic partnership with the ETH Zurichfor a new nanotechnology laboratory ten years ago38. Another pro-minent case has been a $25 million research alliance between Philipsand the Massachusetts Institute of Technology (MIT) in the area of lig-hting solutions technology. Both partners jointly own a highly citedquantum dot patent portfolio, originally co-developed by MIT andHewlett Packard39. In Japan, Hitachi is a sponsor of several researchprograms at the University of Tokyo, which has led to the joint de-velopment of several patents for carbon nanotube synthesis40. Thatacademic institutions and technology companies should partner upis the ‘biggest lesson’ from Thomas Insel, former National Institutesof Health director who entered the private sector in Silicon Valley41.Establishing partnerships amongst different types of actors has beena primary objective of European and worldwide nanotechnology in-novation policies42.

6.4 research questions

This paper seeks to empirically test whether or not nanotechnologyis subject to a patent thicket. Following the leading definition byShapiro (2001), we define a patent thicket as a ‘dense web of over-

37. Allianz. Opportunities and Risks of Nanotechnology, Report in Co-Operationwith the OECD International Futures Programme. 2005, http://oecd.org/science/nanosafety/37770473.pdf.

38. For the press release, https://www.zurich.ibm.com/news/08/nanotech.html.39. Patent US6501091, filed on October 7, 1998 was cited by close to 500 US patents.40. For patents filed in Europe, see illustratively EP 2889268 and EP 2476648 filed

September 9, 2010 as well as EP 2865445 filed June 20, 2013.41. Thomas Insel, “Join the Disruptors of Health Science,” Nature 551 (2017): 23–26.42. Giuseppe Calignano, “Italian Organisations Within the European Nanotechno-

logy Network: Presence, Dynamics and Effects,” Die Erde 14, no. 4 (2014): 241–259.

150 patent thickets in nanotechnology

lapping rights’43. We use this concept to construct a novel measurevia network analysis. For the purpose of this model, clustering coeffi-cients refer to the tendency of patents within a technology to form ci-rcles of connected nodes. We describe and calculate patent clusteringfor nanotechnology in Europe and in the US and compare the resultsto clustering in a field notoriously prone to thickets, the semiconduc-tor industry. Specifically, the article hypothesizes that if nanotechno-logy is as a field as inter-connected as the semiconductor industry,chances are there is a thicket present in the nano-patent space. Thiswould then in turn support the view that the patent system is, parti-ally at least, responsible for the lack of progress in nanotechnology. If,conversely, nanotechnology appears substantially less clustered thanthe semiconductor industry, patents may well be less relevant for thefield’s slow development.

We introduce a measure of patent thicket quantifying the clustering ofboth backward and forward patent citation networks. Backward citati-ons disclose the prior knowledge upon which the patents have reliedat the time of the filing. The underlying intuition is that if the scopesof two patents overlap, they are likely to be influenced by the sameknowledge44. Forward citations, on the other hand, are the citationsa given patent receives after its publication. In this regard, highlycited patents may act as blocking patents, i.e., bottlenecks in the in-novative process45. The greater the clustering of a network, the morethe technology is inter-connected. Such inter-connectedness suggeststhat the technology is built upon overlapping rights and/or may havefuture innovation hampered by bottlenecks.

Subsequently, we also shed light on one specific technology transfermechanism, i.e., collaboration between private and public actors. Weask whether such collaborations might help the industry to mitigate

43. Shapiro, “Navigating the Patent Thicket: Cross Licenses, Patent Poools, andStandard Setting.”

44. For measures of the knowledge distance between two inventions, see Ryan Wha-len, “Citation Distance: Measuring Knowledge Translation, Integration, Diffusion,and Scope,” Proceedings of the Annual Conference of CAIS, 2016, and Jeff Alstott et al.,“Mapping Technology Space by Normalizing Patent Networks,” Scientometrics 110,no. 1 (2017): 443–479.

45. On defensive patents and forward citations, see David S. Abrams, Ufuk Akcigit,and Jillian Grennan, “Patent Value and Citations: Creative Destruction or StrategicDisruption?,” NBER Working Paper, no. 19647 (2013).

6.4 research questions 151

the impedimentary effects of thickets, if any. The research questionswe study can be formulated as follows.

Research Question 1 The number of nanotechnology patents is increasingstrongly.

Explanation. A large number of patents promotes the emergence of apatent thicket as patent landscapes grow inherently more complex thelarger the number of relevant rights is. Yet, a large number of patentsalone does not necessarily give rise to a thicket as those patents maybe independent from each other46.

Research Question 2 Nanotechnology patents form a dense web of priorart.

Explanation. Since patent thickets by definition consist of patents thatcover functionally similar components and/or partially or even com-pletely overlap each other, clusters of patents building upon commonprior art should emerge47.

Research Question 3 Nanotechnology patents form a web of blocking pa-tents.

Explanation. The hold-up potential, i.e., the possibility that new pro-ducts (inadvertently) infringe on (several) third party rights and arehence blocked/kept characterize patent thickets48.

Research Question 4 Nanotechnology patents resulting from private-publiccollaborations are increasing in number and quality over time.

Explanation. To mitigate economic issues arising from the presenceof patent thickets, it becomes more efficient for companies to investin research upfront rather than license ex post49.

46. In this sense Cockburn and MacGarvie, “Patents, Thickets and the Financingof Early-Stage Firms: Evidence from the Software Industry” ‘[h]owever, it may notbe just the absolute number of patents in an area that can deter entry, but also theextent to which those patents form a ‘thicket’ in the sense of generating transactionscosts above and beyond simple blocking power’.

47. See Egan and Teece, “Untangling the Patent Thicket Literature.”48. See Hall et al., A Study of Patent Thickets.49. See Ralph Siebert and Georg von Graevenitz, “Does Licensing Resolve Hold Up

in the Patent Thicket?,” Münchener Wirtschaftswissenschaftliche Beiträge, nos. 2008-01

(2008).

152 patent thickets in nanotechnology

6.5 related literature

So far, first empirical estimations of patent thickets have been deri-ved from social network analysis and algorithmic methods. Mainly,the measurement sought to determine the density of patent citationsnetworks, either clustered at the patent or at the company level.

Patent level. Clarkson (2005) and De Korte and Clarkson (2006) uti-lized network analysis methods to compare density measures of bac-kward citation networks50. The density is proposed to be higher for apatent thicket than for the surrounding patent universe. Such densitymetrics rely upon work in the field of graph theory51. Network den-sity is the ratio of the number of actual edges. As such, it determineshow close a network is to complete, with density equaling 1 if andonly if all edges are actually established. Mathematically, this may beexpressed as follows:

∆ =2|E|

|N|(|N|− 1)(27)

where |E| is the number of actual edges and |N| the number of nodes.Assessing the density, however, misses the element of connectedness.Two technological fields might share the same density coefficient, butyet one be forming a dense cluster, the other be in isolation.

Company level. von Graevenitz, Wagner, and Harhoff (2013) and vonGraevenitz, Wagner, and Harhoff (2011) introduced a density mea-sure counting constellations in which three firms each own patentsciting a patent or more of the two other in the same technology52.

50. Gavin Clarkson, “Patent Informatics for Patent Thicket Detection: A NetworkAnalytic Approach for Measuring the Density of Patent Space,” January 2005, andDavid De Korte and Gavin Clarkson, “The Problem of Patent Thickets in ConvergentTechnologies,” Annals of the New York Academy of Sciences 1093 (2006): 180–200.

51. Frank Harary, Robert Z. Norman, and Dorwin Cartwright, Structural Models:An Introduction to the Theory of Directed Graphs (John Wiley & Sons, 1965).

52. von Graevenitz, Wagner, and Harhoff, “Incidence and Growth of Patent Thic-kets: The Impact of Technological Opportunities and Complexity” and Georg vonGraevenitz, Stefan Wagner, and Dietmar Harhoff, “How to Measure Patent Thickets– A Novel Approach,” Economics Letters 111, no. 1 (2011): 6–9; see Timo Fischer andJoachim Henkel, “Patent Trolls on Markets for Technology – An Empirical Analysisof NPEs’ Patent Acquisitions,” Research Policy 41, no. 9 (2012): 1519–1533 presentingevidence that non-practicing entities are more active in fields with a higher share oftriples.

6.5 related literature 153

The advantage of such metric is that is integrates transaction costs inthat it looks at patent thickets at the company level. However, the cen-tral assumption that if a company owns a patent citing any patent ofanother company in the same technology there is overlap of rights canbe criticized. For a cross-disciplinary field such as nanotechnology ortechnologies with large patent aggregators, citing any patent of the ot-her company’s portfolio might be a condition too easy to satisfy. A re-cently developed methodology utilizing natural language processingby deGrazia, Frumkin, and Pairolero (2018) exploits claims similarityto capture the inventive overlap; therefore remedying to such issue.By focusing on vertical horizontal claims, it offers a clean identifica-tion of one specific channel of patent thickets. Altogether, analyses atthe company level suffer from the limitation that they do not capturethe indirect links between patents as established by prior art.Our method focusing on clustering has several advantages for our

study. Firstly, this provides us with an indication of not only the den-sity but also the connectedness of a technology. Secondly, the met-hod enables us to compare nanotechnology to other fields, such asthe semiconductor industry, where the presence of a patent thickethas been established53. Thirdly, we can use the method to track de-velopments of the technology over time, and see how the technologyevolved. The methodology bears, however, two main limitations. Theanalysis is carried out at a patent-level, while the strategic players arethe companies. Nanotechnology being a fragmented field in terms ofownership, the need to differentiate at the company level seems lesspressing54. Furthermore, it relies on the assumption that cited priorart limits the use of the citing patent55. Since we assume prior art tobe only more likely to share an overlapping scope of protection withthe citing patent than non-cited prior art would, we are of the opinionthat this limitation is minor; and that the advantages of the methodoverweight its drawbacks.

53. See, e.g., Hall and Ziedonis, “The Patent Paradox Revisited: An Empirical Studyof Patenting in the U.S. Semiconductor Industry, 1979-1995” and Rosemarie H. Zie-donis, “Don’t Fence Me In: Fragmented Markets for Technology and the PatentAcquisition Strategies of Firms,” Management Science 50, no. 6 (2004): 804–820.

54. According to our data in Europe for instance, the top ten patent applicants holdless than ten percent of all patents.

55. Hall et al., A Study of Patent Thickets.

154 patent thickets in nanotechnology

6.6 data collection

We aggregated data on nanotechnology patents by identifying all pa-tent documents from the B82Y sub-class of the International PatentClassification. This sub-class comprises all patent documents relatingto nanostructures56. For Europe, we extracted the patents from theOECD, REGPAT database, February 2016. This search yielded 2,396

patent applications filed from 1991 to 2012. For the US, we used thedata provided by PatentsView and collected 48,212 applications forgranted patent from 1971 to 2017. In this paper, we use the term ‘pa-tent’ for the data on patent applications and patent grants57.

For the network analysis, citation data was extracted from Patents-View and the OECD Citations database, September 2016. We furt-her gathered data on semiconductor patents from the H01L sub-class.Since we sought to map networks of comparable dimensions to ana-lyze the development of the respective technology, we generated net-works comprising only (the first) 2,396 patent documents throug-hout58. Although the main analysis does not include all available USdocuments then, it nevertheless allows us to trace the beginnings ofnanotechnology’s evolution in each region – arguably the most inte-resting time frame. For the sake of robustness, we further conduct anempirical analysis over a longer period of time in the US.

For the section on collaboration, we merged nanotechnology and se-miconductors patents with additional information on the applicantsusing PATSTAT, Spring 2016, and the OECD, HAN database, Septem-ber 2016. PATSTAT divides applicants into a set of categories inclu-ding company and individual inventors, which we defined as privatesector organizations, and government non-profit, hospital, and uni-

56. See Nanotechnology and Patents, (European Patent Office, 2013); in fact all pa-tent offices worldwide started to classify nanotechnology uniformly under the IPCsystem with the new symbol B82Y introduced on 1 January 2011.

57. With data on both in Europe and on granted patents only in the US.58. For US nanotechnology, the first 2,396 granted patents cover the timespan 1974

to 1991; for EP semiconductors, the first 2,396 patent applications cover the timespan1978 to 1983; for US semiconductors, the first 2,396 applications for granted patentscover the timespan 1965 to 1971.

6.7 results 155

versity which we considered to form the public sector59. For approx-imately 10% of the applicants this information was not available andwas thus hand-coded. We aggregated data on the patents themsel-ves using the OECD Patent Quality Indicators database, Spring 2016,primarily to extract the family size of patents. The latter indicate thenumber of patent offices at which a given invention has been pro-tected, and is used as a proxy for the value of patents60.

6.7 results

We present our results relating to the growth in patenting, the clus-tering of the field and private-public collaborations in nanotechno-logy.

6.7.1 Increase in Patenting

Indeed, the number of patents in nanotechnology is ever-increasing.Moreover, the total number, but also – and in particular – the shareof public patents, i.e., patents filed by public entities and universities,has been steadily increasing, as well, the latter even constantly sur-passing 40% in Europe and 30% in the US, respectively, since 2011

(Figure 6.1 and 6.2)61.

59. Using the same distinction between public and private sector, see Gregory D.Graff et al., “The Public–Private Structure of Intellectual Property Ownership inAgricultural Biotechnology,” Nature Biotechnology 21 (2003): 989–995.

60. See OECD, “A Framework for Biotechnology Statistics,” 2005, and Jean O. Lan-jouw, Ariel Pakes, and Jonathan Putnam, “How to Count Patents and Value Intellec-tual Property: The Uses of Patent Renewal and Application Data,” Journal of IndustrialEconomics 46, no. 4 (1998): 405–432.

61. In this graphic, public patents include all patents involving at least one publicactor, and in turn also the private-public patents.

156 patent thickets in nanotechnology

Growth in Nanotechnology Patents

Share Public

0

10

20

30

40

Private

Public

1995 2000 2005 20100

50

100

150

200

Figure 6.1: Europe.

Share Public (%)

0

5

10

15

20

25

30

35

Private (n)

Public (n)

1985 1990 1995 2000 2005 20100

500

1000

1500

2000

Figure 6.2: United States.

Two readings of these figures are possible. On the one hand, this maybe seen as an indicator for the still infantile status of nanotechnology,with a field still in the batch of universities. On the other hand, thismay be seen as an indicator for a further exacerbation of the thicketproblem in the sense that even more of nanotechnology is ‘stuck’ inthe hands of non-commercial actors that lack sufficient drive to turnnanotechnology inventions into innovations – an observation that fu-els a substantial part of the thicket argument in the first place.

6.7.2 Clustered Web of Patents

To assess the overlap of rights that characterizes patent thickets, wecompare sets of citation networks. Mainly, we seek to analyze net-works of backward and forward citations of patents.

6.7.2.1 Descriptive View

For a qualitative, descriptive view, we compare plots of US and Eu-ropean citation networks for nanotechnology with those for the semi-conductor industry. The latter is well known to suffer from thicketeffects62, a finding the plots in Figures 6.3b, 6.3d, 6.4b and 6.4dintuitively convey and underline: Semiconductor patents are heavily

62. Hall and Ziedonis, “The Patent Paradox Revisited: An Empirical Study of Pa-tenting in the U.S. Semiconductor Industry, 1979-1995”; Ziedonis, “Don’t Fence MeIn: Fragmented Markets for Technology and the Patent Acquisition Strategies ofFirms.”

6.7 results 157

interconnected, in turn leading to the emergence of a large cluster atthe center of each network. The effect appears strongest for forwardcitations in the US (Figure 6.4c) and weakest for forward citations inEurope (Figure 6.4a).

In contrast, the situation appears to be quite different and differen-tiated in nanotechnology. No clustering is visible in Europe, neitherwhen considering backward citations (Figure 6.3a) nor when conside-ring forward citations (Figure 6.4a). Not only does nanotechnologyhence seem to build upon independent, largely disjoint streams rat-her than relying on a specific, coherent set of prior art. There alsodo not seem to exist any fundamental blocking patents representingbottlenecks for the innovative process, irrespective of whether theyqualify as nanotechnology or not.

Yet, nanotechnology does indeed resemble the semiconductor indu-stry in terms of interrelation by citation in the US. Although not aspronounced, one can observe a clustering at the center of the net-works for backward (Figure 6.3c) as well as forward citations (Figure6.4c) . Translating this graphical observation, one is led to assume thatthere do in fact exist overlapping patent rights as well as fundamentalblocking patents in the US.

158 patent thickets in nanotechnology

Figure 6.3: Backward Citation Networks

6.7 results 159

Figure 6.4: Forward Citation Networks

160 patent thickets in nanotechnology

6.7.2.2 Empirical View

We find confirmation for these qualitative findings in a quantitativeapproach drawing on complex network analysis. The clustering coef-ficient of a single vertex (of at least degree 2) is the probability thatany two randomly chosen neighbors of said vertex are linked toget-her themselves. The clustering coefficient of the whole graph maythen be defined as the average of this value across all vertices (of atleast degree 2)63. Clustering coefficients as summarized in Table 6.1mirror that, in Europe, nanotech citation networks are less cluste-red than their semiconductor counterparts. Somewhat surprisingly,though, US nanotechnology patent documents seem to cluster evenmore strongly than those pertaining to semiconductors, underliningonce again that the concerns raised by the US literature in particularmight not be easily discarded after all.

Table 6.1: Clustering Coefficients

Backward Citations

Nanotech Semiconductor

EU 0.009 0.03

US 0.102 0.018

Forward Citations

Nanotech Semiconductor

0 0.042

0.138 0.02

Density measures, albeit their limitation in that they do not capturethe connectiveness of patents and their potential overlap in scope,can also be computed for comparison. Network citation densities innanotechnology and semiconductor are compared in Table 6.2. Accor-dingly, in Europe nanotechnology patents form a less dense networkthan semiconductor patents. In the US, the two technological fieldsreach an equal density coefficient.

63. Matthieu Latapy, “Main-Memory Triangle Computations for Very Large(Sparse (Power-Law)) Graphs,” Theoretical Computer Science 407, nos. 1-3 (2008): 458–473.

6.7 results 161

Table 6.2: Density Coefficients

Backward Citations

Nanotech Semiconductor

EU 0.0002 0.0024

US 0.00025 0.00025

Forward Citations

Nanotech Semiconductor

0.00012 0.00034

0.00007 0.00007

Finally, we find confirmation for the presence of a thicket in nano-technology in the US by comparing the average clustering and densitycoefficients over a longer timespan. We observe that nanotechnologyclusters more intensely than semiconductor patents not only for theearliest period, but also when comparing up to the first 6,000 patentsfiled in the field. Density coefficients are higher for semiconductors, afurther suggestion that higher density of a field does not imply higherclustering (see Figure 6.5). This supports our findings of a presenceof a patent thicket in the US.

Figure 6.5: Clustering and Density of Prior Art

.00005

.0001

.00015

.0002D

ensi

ty A

vera

ge

.01

.015

.02

.025

.03

Clu

ster

ing

Aver

age

3000 4000 5000 6000 7000

Number of Patents

Nanotechnology Clustering Semiconductor ClusteringNanotechnology Density Nanotechnology Density

162 patent thickets in nanotechnology

6.7.2.3 The Multidisciplinarity of Nanotechnology

Yet, it is nanotechnology’s multidisciplinary structure that representsa caveat with a view to the presented findings: It may well be thatonly certain subparts of nanotechnology suffer from thicket effectssuch that corresponding networks do in fact only exhibit partial andspotty, albeit then heavy, clustering not captured by this analysis. Themost promising silo of nanotechnology, nanomedicine, can be takenas an example to illustrate the relationship between clustering andinnovation. Ever since the Food and Drug Administration approveda nanotherapeutic product for the first time back in 1995 – Doxil, ananticancer drug using PEGylated nano-liposomes – the dawn of nano-medicine has been proclaimed. In fact, however, despite a number ofclinical successes, nanotechnology has not (yet) transformed the diag-nosis and treatment of diseases. For pharmaceuticals, in vivo imagingand in vitro diagnostics, nanotechnology is poised to contribute signi-ficantly, but the technology has yet to bear its fruits64.

When examining the networks of patent citations in nanomedicine(classified IPC B82Y5), we can observe a trend towards consolidationresp. clustering (see Figure 6.6). In the early stages from 1974 to 1990,three small ‘islands’ of prior art had formed in isolation. Then theybridged in 1995. From then on, nanomedicine grew increasingly clo-ser. Nowadays, despite spanning a variety of areas, almost all patentsare linked to each other either directly or indirectly through commonprior art. Furthermore, although the vast majority of cited patents arenon-nanomedicinal (88%), a majority of patents refers to at least onenanomedicine patent (up to 80% in 2014). Rather than branching outin specialized subfields, nanomedicine has been translated into a fieldincreasingly consolidated. The criticism levied against patent thicketsfor nanotechnology on the whole appear to be replicated in kind atthe more refined technology silo level.

64. See Subbu Venkatraman, “Has Nanomedicine Lived up to its Promise?,” Nano-technology 25, no. 37 (2014): 1–4; Wim De Jongand and Paula Borm, “Drug Deliveryand Nanoparticles: Applications and Hazards,” International Journal of Nanomedicine3, no. 2 (2008): 133–149 and Lisa Bregoli et al., “Nanomedicine Applied to Transla-tional Oncology: A Future Perspective on Cancer Treatment,” Nanomedicine: Nano-technology, Biology and Medicine 12, no. 1 (2016): 81–103.

6.7 results 163

Figure 6.6: Nanomedicine Prior Art Networks

164 patent thickets in nanotechnology

6.7.3 Private-Public Collaboration in Nanotechnology

A distinct feature of nanotechnology innovation is that patents are lar-gely held by public entities. While the public sector generally holdsabout one percent of all patents65, this share surpassed an average30% in nanotechnology as demonstrated in the dataset spanning 1991

to 2014 (Figure 6.1, 6.2 and Appendix Figure A6.1). The importanceof university patenting calls for reliable technology transfer mecha-nisms, so that the patented subject matter may be turned into com-mercialized products by private actors. As indicated above, universitypatenting does not, by itself, establish impedimentary effects on inno-vation. On the contrary, patents created in collaboration with privateactors might contribute to the mitigation of thicket effects by reducingtransaction costs. Patent data can indicate whether there actually ex-ists some form of collaboration between public and private actors66.Simply put: The number and/or share of patents filed jointly by aprivate and a public organization, i.e., so-called ‘mixed patents’, maybe seen as an indicator for the intensity of private-public collabora-tion.

First of all, the share of patents filed jointly by private and publicentities is above average in nanotechnology with 6.18% in Europeand 1.91% in the US (see Figure 6.7, 6.8 and Appendix FigureThat being said, an increase in mixed patenting may only have a me-aningful impact on the efficiency of the innovative process if the cor-responding patents are of value. We thus seek to determine whethermixed patents are of higher quality than purely private or public pa-tents. Therefore, we make use of two measures commonly employedin innovation science: forward citations, a proxy for the technologicalimportance of the patented invention67; and family size, which has

65. Lemley, “Patenting Nanotechnology.”66. As such, it appears that ex ante licensing – the sharing of knowledge before its

full development – is more common in industries with strong intellectual propertyrights, see Grindley and Teece, “Managing Intellectual Capital: Licensing and Cross-Licensing in Semiconductors and Electronics.”

67. As well as, to a certain extent, the value of said patent, see, e.g., Manuel Trajten-berg, “A Penny for Your Quotes: Patent Citations and the Value of Innovations,” TheRAND Journal of Economics 21, no. 1 (1990): 172–187; Bronwyn H. Hall, Adam Jaffe,and Manuel Tratjenberg, “Market Value and Patent Citations,” The RAND Journal ofEconomics 36 (2005): 16–38; and Dietmar Harhoff, Frederic M. Scherer, and Katrin

6.7 results 165

Growth in Private-Public Patents

Share Mixed (%)

0

2

4

6

8Mixed (n)

2000 2002 2004 2006 2008 2010 20120

5

10

15

20

25

30

Figure 6.7: Europe.

Share Mixed (%)

0

1

2

3

4

5

6Mixed (n)

1985 1990 1995 2000 2005 20100

50

100

150

Figure 6.8: United State.

repeatedly been found to be associated with the economic value ofan invention68.

Somewhat surprisingly, private-public patents (almost) consistentlyreceived the lowest number of citations (see Table 6.3 and 6.4 and Fi-gure 6.9 and 6.10.), although there admittedly exist some outliers. InEurope, a high-impact patent co-filed by MIT and Hewlett Packardis the only one69. In the US, the citation averages of mixed patentspeaked in the period 1995-1999, even surpassing those of their fullyprivate counterparts. Highly cited patents from that period includeinventions in the field of nanotubes70, nanomedicine71, nanomagne-tism72 or skin care73.When running a one-way ANOVA test for US data, we find thatthe number of forward citations of private-public patents was signi-

Vopel, “Citations, Family Size, Opposition and the Value of Patent Rights,” ResearchPolicy 32, no. 8 (2003): 1343–1363.

68. Lanjouw, Pakes, and Putnam, “How to Count Patents and Value IntellectualProperty: The Uses of Patent Renewal and Application Data” and Harhoff, Scherer,and Vopel, “Citations, Family Size, Opposition and the Value of Patent Rights.”

69. EP 2325897 filed Apr. 1, 1999, a quantum dot light emitting diodes.70. See, e.g., US 6277318 filed Aug. 18, 1999, a University of North Carolina and

Agere Systems method for fabrication of patterned carbon nanotube films.71. See, e.g., a Wisconsin Alumni Research Foundation and Vical delivery method.72. See, e.g., US5648885 filed Aug. 31, 1995, a University of Alabama, Hitachi and

Victor Co of Japan magnoresistive spin-valve sensor.73. See, e.g., US6013683 filed Dec. 17, 1998, a University of Delaware and Dow

Corning microemulsion.

166 patent thickets in nanotechnology

Average Forward Citations Five Year Over Time

●

●●

●●

●

■

■

■

■

■

■

◆ ◆

◆

◆

◆ ◆

● Private Patents

■ Public Patents

◆ Mixed Patents

1996 1998 2000 2002 2004 2006 2008 2010

1

2

3

4

5

6

Figure 6.9: Europe.

●

●

●

●

●

● ●■

■

■

■

■

■■

◆

◆

◆

◆

◆

◆ ◆

● Private Patents

■ Public Patents

◆ Mixed Patents

1985 1990 1995 2000 2005 2010

2

4

6

8

10

12

Figure 6.10: United States.

ficantly lower as compared to private and public patents74. This con-trasts with previous research finding that co-ownership of patents sig-nificantly correlates with higher forward citation rates75. Accordingto our data, patents filed by public institutions obtained the largestnumber of citations. This is consistent with other work arguing thatuniversity patents are more basic and hence have greater technologi-cal impact76. Also, and in line with the fact that the value distributionof patents is highly skewed, with a long tail into the high value side77,two thirds of all European and one third of all US nanotechnology pa-tents did not receive a single citation78.Lastly, we turn to our second value proxy of choice, family size. Firstand foremost, one would expect private actors to seek for patent pro-tection in more jurisdictions since they have the financial strength

74. There was a statistically significant difference between the three groups as de-termined by one-way ANOVA (F(2, 31958) = 22.187,p = .000), and we use a Tukeypost hoc test to assess the differences between groups. In Europe in contrast, wherethe sample size was very limited, there was no significant difference.

75. Belderbos et al., “Co-Ownership of Intellectual Property: Exploring the Value-Appropriation and Value-Creation Implications of Co Patenting With Different Part-ners.”

76. Adam B. Jaffe and Gaétan de Rassenfosse, “Patent Citation Data in SocialScience Research: Overview and Best Practices,” Journal of the Association for Infor-mation Science and Technology 68, no. 6 (2017): 1360–1374.

77. See John R. Allison et al., “Valuable Patents,” Georgetown Law Journal 92 (2004)and Kimberly A. Moore, “Worthless Patents,” Berkeley Technology Law Journal 20

(2005) finding that more than half of the patent holders let their patents expire bynot paying the maintenance fees.

78. This number is similar to the level observed in our data for all 2.9 millionEuropean Patents filed from 1978 to 2015.

6.8 conclusion 167

Table 6.3: EU Forward Citations 5-Y(Up to 2008)

Type Number Mean 95%CI

Private Patents 697 1.62 1.37-1.86

Public Patents 226 1.76 1.26-2.25

Mixed Patents 34 1.20 0.46-1.95

Table 6.4: US Forward Citations 5-Y(Up to 2008)

Type Number Mean 95%CI

Private Patents 25,439 5.15 5.03-5.27

Public Patents 5,868 5.16 4.9-5.42

Mixed Patents 654 3.69 3.12-4.26

Table 6.5: EU Family Size(Up to 2012)

Type Number Mean 95%CI

Private Patents 1549 5.34 5.2-5.48

Public Patents 699 4.07 3.93-4.22

Mixed Patents 148 5.03 4.61-5.45

Table 6.6: US Family Size(Up to 2012)

Type Number Mean 95%CI

Private Patents 10,318 3.65 3.58-3.72

Public Patents 2,107 3.87 3.69-4.04

Mixed Patents 222 3.22 2.86-3.57

to bear the additional costs. This is indeed true for Europe, whereprivate patents attain the largest geographical scope (see Table 6.5),somewhat closely followed by mixed patents. In the US, on the otherhand, the geographical scope of protection of mixed patents is thelowest and that of public patents the largest (see Table 6.6).

Taken together, these results suggests that private-public collabora-tion does not – at least at this point in time – lead to innovation withthe highest technological impact. Compared to purely private or pu-rely academic ones, the hybrid collaborative model underperformedin its research output as measured in (proxied) patent quality.

6.8 conclusion

The notion that society is awaiting the nanotechnology revolution invain due to relentless patenting activity in the field has proven quiteresilient thus far. That being said, voices blaming other factors andquestioning the existence of a patent thicket in the nanotechnologyspace in the first place have grown ever louder recently. In terms ofempirical citation data, both sides of the discussion seem to be right,though, after all: While Europe was spared the formation of a thicketso far, the nanotech patent landscape appears to be rather intricateand convoluted in the US. When examining nanomedicine, the criti-cism levied against patent thickets for nanotechnology on the whole

168 patent thickets in nanotechnology

appear to be replicated in kind at the more refined technology silo le-vel. Ultimately, these findings connect neatly to the fact that most ofthe critical discussion concerning patents on nanotechnology occursin that very region, although they are hard to square with each othergiven that the development of nanotechnology can surely be expectedto proceed in an international fashion. Unfortunately, the chances thatprivate-public collaboration, albeit increasingly popular, can mitigatethicket effects appear to be rather limited. Such collaborations simplydo not yield innovation with the highest impact. This analysis must,however, not divert from the fact that the pooling of assets, know-how, and technologies may provide innovators with better ways toaddress a variety of challenges of innovation in nanotechnology. Fi-nancing and supporting collaborative environments may be key fornations to foster growth in nanotechnology, but the fruits in terms ofquality are yet to bear.

Hence, the perceived lack of progress in nanotechnology may havedifferent reasons on the two continents after all. In the US, in linewith the more general debates about the more permissive approachof the US Patent and Trademark Office, a large number of nanotechno-logy patents was granted, presumably with larger overlap in scope.Reading these signs to imply the presence of a thicket suggests thatpatents might indeed hinder innovation in nanotechnology in the US.It should be noted, however, that there arguably still exists a vari-ety of other factors that may be at least equally as important for the(lacking) development of the field.

Policies seeking to foster the development of the field in the US couldfocus, amongst other things, on implementing and assuring more re-strictive grant practices by the USPTO and the application of strictnovelty and obviousness thresholds by courts. In Europe on the otherhand, patents probably played a smaller role than other factors suchas intrinsically long development cycles and limited access to equip-ment, funding, and infrastructure. Therefore, European nanotechno-logy programs could rather aim at developing large public grant pro-grams or encouraging R&D subsidies for private actors as well asat facilitating collaboration between public and private innovators.Developing reliable frameworks and close institutional networks bet-

6.8 conclusion 169

ween entities active in nanotechnology may be key for European coun-tries seeking to develop or strengthen their presence worldwide.

170 patent thickets in nanotechnology

6.9 appendix

Figure A6.1: EU Public and Private Patent Applications.

0

100

200

300

400

91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14

Public Sector PatentsPrivate Sector Patents

Figure A6.1 represents the yearly number of patent applications filed by privatesector applicants versus applications filed by at least one public sector applicant inEurope.

Figure A6.2: EU Percentage Patent Applications, by Type.

0

20

40

60

80

100

1990 1995 2000 2005 2010

Public PartnershipsPrivate PartnershipsPrivate-Public PatentsPublic PatentsPrivate Patents

Figure A6.2 depicts the yearly percentage of patent applications filed (i) indepen-dently by private sector entities, (ii) independently by public sector entities, (iii)jointly by private-public collaborations, (iv) jointly by private sector collaborationsand (v) jointly by public sector collaborations in Europe.

7F E D E R A L F U N D I N G I N C A N C E R R E S E A R C H

7.1 introduction

Cancer research is a field that has been especially dependent on pu-blic funding. The National Cancer Institute (NCI), an independentinstitute of the National Institutes of Health (NIH), was establishedin 1937 to provide for, foster and coordinate research relating to can-cer1. It has since then been at the forefront of the war against can-cer and, as the largest funder of cancer research in the world, hasspent an estimate of more than USD 100 billion on research and treat-ment2. New screening, diagnosis, and treatment methods3 in conso-nance with groundbreaking pathogenetic discoveries in the oncologicfield, mainly in the form of oncogene4 and cancer metabolism rese-arch5, have led to an overall decrease of the cancer death rate by a

1. National Cancer Act, Senate Bill 2067 - Enacted August 5, 1937 (Public Law244) and for a historical account, see Michael B. Shimkin, “As Memory Serves–AnInformal History of the National Cancer Institute, 1937-57,” Journal of the NationalCancer Institute 59, no. 2 (1977): 559–600.

2. See Gina Kolata, “Grant System Leads Cancer Researchers to Play It Safe,” NewYork Times, July 2009, citing the figure of USD 105 billion and Margaret Cuomo,A World Without Cancer: The Making of a New Cure and the Real Promise of Prevention(Rodale, 2012) with the number of USD 90 billion. In 1999 illustratively, the NCIprovided for close to half of the total cancer research funding in the US, see MichaelMcGeary and Michael Burstein, Sources of Cancer Research Funding in the United States,(Report on Sources of Cancer Research Funding in the United States, June 1999).

3. With screening and diagnosis progress in ultrasound (sonography), computedtomography (CT scans), magnetic resonance imaging (MRI scans) and positron emis-sion tomography (PET scans) to name but a few since the 1970s; and treatment ad-vances in chemotherapy, immunotherapy, and targeted therapies.

4. A gene with the potential to cause cancer, such as as the TP53 gene, or BRCA2

and HER2 for breast cancer, see Klaus Bister, “Discovery of Oncogenes: The Adventof Molecular Cancer Research,” Proceedings of the National Academy of Sciences 112, no.50 (2015): 15259–15260.

5. With a deeper biochemical understanding of the metabolic reprogramming intumor cells, see Natalya N. Pavlova, “The Emerging Hallmarks of Cancer Metabo-lism,” Cell Metabolism 23, no. 1 (2016): 27–47.

171

172 nih and cancer-related patenting

quarter since 19756. To pursue its aims, the NIH supports research

enterprise both intramural in its facilities and extramural in univer-sities, medical schools, hospitals or institutes. The NIH has awardedsupport to 156 recipients of 92 Nobel Prizes, and 37 of them receivedfunding by the NCI in particular7.

The mission of the NIH is to both seek fundamental knowledge andto apply that knowledge to enhance human life, extend longevity, andreduce illness and disability8. This distinction follows a linear concep-tion of innovation and differentiates between the aim of developingtheoretical knowledge (basic research) and the practical problem sol-ving (applied research). The assumption that public investments inscience have practical returns in addition to the pure knowledge theygenerate underlines modern science policy9. Recent scholarship hasstarted to develop measures of those returns on investment. Typically,this strand of the literature estimates the number of public grants re-quired to generate one patent, or the amount of funding necessary togenerate one additional private-sector patent10.

6. From 50% in 1975 to 66% in 2012 for the most common types of cancer, see An-nual Report to the Nation 2017 available under https://seer.cancer.gov/report_to_nation/survival.html and Rebecca L. Siegel, Kimberly D. Miller, and AhmedinJemal, “Cancer Statistics, 2018,” CA: A Cancer Journal for Clinicians 68, no. 1, 7–30.

7. See the listing under https://www.nih.gov/about-nih/what-we-do/nih-almanac/nobel-laureates; in 2018, the recipients of the Nobel Prizes in Chemistry and inPhysiology or Medicine had received NCI funding in their career.

8. The goal of the funding program is to maximize the summed impact of thescientific community, see Jean-Michel Fortin and David J. Currie, “Big Science vs.Little Science: How Scientific Impact Scales with Funding,” PLoS ONE 8, no. 6 (2013):1–9 for a model on the optimal allocation of research funds.

9. Vannevar Bush, Science, the Endless Frontier; A Report to the President on a Pro-gram for Postwar Scientific Research (National Science Foundation, 1960) ‘[n]ew pro-ducts, new industries, and more jobs require continuous additions to knowledgeof the laws of nature, and the application of that knowledge to practical purposes’and Mariana Mazzucato, The Entrepreneurial State: Debunking Public vs. Private SectorMyths (Anthem Press, 2013), p. 6: ‘nearly all the technological revolutions in the past– from the Internet to today’s green tech revolution – required a massive push by theState’.

10. See, e.g., Danielle Li, Pierre Azoulay, and Bhaven Sampat, “The Applied Value ofPublic Investments in Biomedical Research,” Science 356 (2017): 78–81; Pierre Azou-lay et al., “Public R&D Investments and Private-Sector Patenting: Evidence fromNIH Funding Rules,” NBER Working Paper, no. 20889 (2015) and Bhaven N. Sampatand Frank Lichtenberg, “What are the Respective Roles of the Public and PrivateSectors in Pharmaceutical Innovation?,” Health Affairs 30 (2011): 332–339.

7.1 introduction 173

Over the last decade, academics and scientists raised concerns aboutthe current funding system of the NIH11. While certain experts see theselection of research for grant funding as solely based on merits12, ot-her openly criticized it. The former director of the NIH, Raynard King-ton, for instance recognized that the system provides disincentives tofunding really transformative research13. Several scholars criticizedthe NIH grant mechanism, arguing it lacked of predictive ability14

and was focusing on ‘conform’ research15. Further criticism point outto a mechanism subject to gender and racial bias16 and leading toduplication of grants by its different agencies17. This criticism is exa-mined for the field of cancer research.

11. For a review of the NIH grant mechanism in practice, see Nicholson W. Price,“Grants,” Berkeley Technology Law Journal forthcoming (2019).

12. See Jeffrey M. Drazen and Julie R. Ingelfinger, “Grants, Politics, and the NIH,”New England Journal of Medicine 349, no. 23 (2007): 2259–2261 arguing that the ‘thesystem works remarkably well [..] is rigorous and comprehensive, identifying andfunding the most meritorious work’; see also Danielle Li and Leila Agha, “Big Namesor Big Ideas: Do Peer-Review Panels Select the Best Science Proposals?,” Science 348,no. 6233 (2015): 434–438 with empirical support for the current peer review system.

13. Kolata, “Grant System Leads Cancer Researchers to Play It Safe.”14. See Ferric Fang and Arturo Casadevall, “Research Funding: the Case for a Mo-

dified Lottery,” mBio, no. 2 (2016) proposing that peer review is used to identify themost meritorious proposals, from which funded applications are selected by lottery.

15. See Joshua M. Nicholson and John P. Ioannidis, “Research Grants: Conformand Be Funded,” Nature 492 (2012): 34–36 with a study suggesting that the highestcited scientists do not have current NIH funding; for a critic of the latter metho-dology, see Steven L. Salzberg, “NIH Funding: It Does Support Innovators,” Nature493 (2013): 26 and David Gorski, “The NIH Funding Process: Conformity and Me-diocrity?,” Science-Based Medicine, December 2012, for a call towards a bigger com-mitment to supporting high-risk, high-return, interdisciplinary research, see JeffreyMervis, “White House Panel Urges Agencies to Take More Risks,” Science 338 (2012):1274; President’s Council of Advisors on Science and Technology, Report Transfor-mation and Opportunity: The Future of the U.S. Research Enterprise, (2012) and Kolata,“Grant System Leads Cancer Researchers to Play It Safe.”

16. See Jennifer R. Pohlhaus et al., “Sex Differences in Application, Success, andFunding Rates for NIH Extramural Programs,” Academic Medicine 86, no. 6 (2011):759–767 showing that men with previous experience as NIH grantees had higherapplication and funding rates than women at similar career points; see Donna K.Ginther et al., “Race, Ethnicity, and NIH Research Awards,” Science 333 (2011): 1015–1019 finding that Asian or African-American applicants are less likely to receive NIHinvestigator-initiated research funding compared with whites.

17. See Harold R. Garner, Lauren J. McIver, and Michael B. Waitzkin, “Same Work,Twice the Money?,” Nature 493 (2013): 599–601 finding that between 2007 and 2011

only, there were 39 concerningly similar grant pairs, involving over USD $20 mio.

174 nih and cancer-related patenting

7.2 research questions

This study estimates the applied value of federal funding for cancerresearch over time. The focal point is to analyze whether grants arestill, under the current funding framework, the optimal policy toolto foster technological progress in the cancer field. This is executedusing the applied returns, i.e., patents, as measure. The patenting acti-vity resulting directly from funded scientists and indirectly from pri-vate entities are quantified over time. For the current grant system tobe the optimal innovation policy instrument under this setting, thispaper studies the two following research questions:

Research Question 1 The overall direct and indirect applied returns toinvestments increase as overall public funding increases.

Explanation. The direct applied value of public funding is reflectedin the patents filed by publicly funded researchers. With an increasein public investment in research, a higher innovation output in termsof quantity or quality is expected18. The indirect value of public fun-ding is reflected in the number of publications resulting from federalresearch subsequently cited as prior art by a patent19.

Research Question 2 The direct and indirect applied returns to investmentby grant are constant or increase.

Explanation. The direct applied value of a grant is reflected in the pro-bability it results in a patent. The indirect value of a grant is reflectedin the probability it results in a publication cited by a subsequent pa-tent. While we expect certain changes over time due to e.g. increasingpatenting and publishing standards, we do not expect the trends in

18. National Academies of Sciences, Engineering, and Medicine, Beyond Patents: As-sessing the Value and Impact of Research Investments: Proceedings of a Workshop - in Brief,(Policy and Global Affairs; Government University Industry Research Roundtable,2017): ‘[t]he number of patents is one metric for measuring the value of research’ andGeorge Symeonidis, “Comparing Cournot and Bertrand Equilibria in a Differentia-ted Duopoly with Product R&D",” International Journal of Industrial Organization 21,no. 1 (2003): 39–55: ‘product R&D raises product quality’.

19. See Li, Azoulay, and Sampat, “The Applied Value of Public Investments inBiomedical Research” for details on the use of these proxies, in particular on patent-publication reference being a measure of knowledge diffusion in that it identifiespatents building on funded research.

7.2 research questions 175

the cancer field to be different from benchmarks such as the returnsto the funding of overall biomedical research.

Explanation Factors. Should these two research questions be negated,we argue that the most likely explanations are that the decreasingreturns are caused by one or some combination of (i) saturation ofthe cancer innovation space, (ii) inefficiencies in the grant selectionprocedure, or (iii) inefficiency of grants as policy tool.

Limitations. The outcome measures we use, patents as proxies andpublications subsequently cited by patents, bear two main limitati-ons. Firstly, this approach assumes that the applied value of publicfunding is a measure for the total value of funding. This can be coun-tered with the argument that the mission of the public funder is lar-ger and includes the development of purely basic research with nopractical implementation. However, the NIH has always had a practi-cal approach20. The total value of public funding is reflected, at le-ast partially, by the practical output21. The second assumption thispaper makes is that measuring patents is a valid proxy for quanti-fying the applied value of funding. This measure is imperfect andnoisy: knowledge spillovers such as conferences, meetings, technicalreports, hiring of previously funded scientists also contribute directly

20. During World War II, it worked almost entirely on war-related issues such asthe development of yellow fever and typhus vaccines, oral saline therapy as first-aidmeasure on the battlefield, or apparatus to supply oxygen for high altitute flying,see https://history.nih.gov/exhibits/history/docs/page_06.html; the NIH alsostates in its mission that it contributes to society by ‘driving economic growth andproductivity’, see https://www.nih.gov/about-nih/what-we-do/impact-nih-rese

arch; Michael S. Lauer and Richard Nakamura, “Reviewing Peer Review at theNIH,” New England Journal of Medicine 373, no. 20 (2015): 1893–1895 with the ar-gument that NIH should measure the impact of grant projects to investigate thesuccess of the NIH peer review process.

21. Norman Sharpless, Director of the NCI: ‘the Bayh-Dole [..] has led to manyimportant successes that have occurred to cancer patients’ at the Policy, Politics andLaw of Cancer Conference at Yale Law School (https://law.yale.edu/solomon-center/events/policy-politics-and-law-cancer/conference-videos); EkaterinaGalkina Cleary et al., “Contribution of NIH Funding to New Drug Approvals2010–2016,” Proceedings of the National Academy of Sciences, 2018, finding that NIHfunding is directly or indirectly associated with every drug approved from 2010 to2016; Francis S. Collins, “Reengineering Translational Science: The Time Is Right,”Science Translational Medicine 3, no. 90 (2011): 1–6 arguing that ‘dramatic technologi-cal advances and associated basic insights into disease mechanisms—research thathas been supported heavily by NIH and other funding agencies’.

176 nih and cancer-related patenting

to practical problem solving22. Private firms might use other form ofprotection such as trade secrets to protect knowledge partially deri-ved from public source. Patents, however, are a proxy that have theadvantage of being not only measurable23, but also to represent anessential protection mechanism in the biomedical industry24.

7.3 data collection

This paper makes use of patent data recently aggregated by the USPatent and Trademark Office within the frame of the Cancer Moons-hot Task Force25. The curated dataset comprises of nearly 270’000

cancer-related patent documents spanning the 1976 to 2016 period26.In addition, data was computed on the patents themselves using theOECD Patent Quality Indicators database, Spring 2016, to extract thenumber of citations a patent received (forward citations) over a periodof five years after the publication date, as well as the patent scope,the family size, the backward citations, the non-patent literature ci-

22. Adam B. Jaffe, “Real Effects of Academic Research,” American Economic Review79, no. 5 (1989): 957–70.

23. Although patent metrics are an imperfect measure of innovative outputs (ArielPakes and Zvi Griliches, “Patents and R&D at the Firm Level: A First Report,” Eco-nomics Letters 5, no. 4 (1980): 377–381), they may provide for a rich innovation lands-cape to inform, support and strengthen the factual basis for policy debates (AntonyTaubman, “Shedding Light on the Life Sciences: Patent Landscaping for Public Poli-cymakers,” WIPO Magazine 4 (2008)).

24. For surveys, see, e.g., Edwin Mansfield, Mark Schwartz, and Samuel Wagner,“Imitation Costs and Patents: An Empirical Study,” Economic Journal 91, no. 364

(1981): 907–918 contending that 90% of the pharmaceutical innovations would nothave been introduced without patents.

25. In January 2016, former President Obama established the Task Force in pursuitof the mission to dramatically accelerate efforts to prevent, diagnose, and treat cancer.Under the lead of the former Vice President Biden, federal agencies and privatesectors collaborations committed to transform cancer research. As part of this effort,the USTPO established the "Horizon Scanning Tool", focusing on leveraging patentdata, see http://bit.ly/WhiteHouseMemoMoonshot.

26. The classification methodology results from a combination of patent classifica-tion and keyword searches, see Jesse Frumkin and Amanda F. Myers, USPTO Moons-hot Patent Data, (U.S. Patent and Trademark Office, August 2016). We focus on the267’643 patent applications filed from 1976 to 2015, thus excluding the smaller datapre-1976 and the (incomplete) data of 2016. Data on all published patent applicati-ons is only available since 2001, while for the earlier period, only data on grantedpatents is available.

7.4 national institutes of health 177

tations and the number of claims27. In addition, this paper uses thereplication data from Li, Azoulay, and Sampat (2017)28 for the dataon 365,380 NIH grants and matches to indirect (citation) linked pa-tents and to direct (Bayh-Dole) linked patents from 1985 to 2007

29. Inthis dataset, we identified 48,111 grants listing the National CancerInstitute and its divisions as administering agency in RePORTER, aNIH publicly available database30. The data on the average size ofR01 equivalent grants bases upon calculation from NIH data31.

7.4 national institutes of health

The National Institutes of Health (NIH) is the largest public funder ofbiomedical research worldwide. The NIH invests its nearly USD 37.3billion in medical research (in 2018) by awarding grants to more than300,000 researchers at thousands of universities and medical schoolsand by supporting nearly 6,000 scientists in its own laboratories. TheNIH is part of the US Department of Health and Human Services,and organized into 27 independent Institutes and Centers specializedby disease (e.g., National Cancer Institute), field of science (e.g. Natio-nal Institute of General Meducal Sciences), or by human developmentstage (e.g., National Instiute on Aging)32. The main share of its budgetgoes to funding research at universities and research institutions. Thecornerstone or the NIH review process for extramural research is theso-called ‘dual peer review’ procedure to analyze proposals by resear-chers. Firstly, the grant applications are assessed by panels of externalscientists from the field itself. Applications are graded according totheir significance, technical merit, innovativness and the qualificati-ons of the investigators. Secondly, for those that pass this threshold,

27. OECD, “A Framework for Biotechnology Statistics,” 2005, for more details onthis methodology.

28. Available under doi:10.7910/DVN/H2QA1R.29. This includes nearly all NIH grants over this period, with half of these being

R01-equivalent grants.30. Available under https://exporter.nih.gov/ExPORTER_Catalog.aspx.31. https://report.nih.gov/catalog_results.aspx?refUrl=index&sS=filter&s

I=&sP=&sM=12&sA=63&sD=7&sV=&sY=&fI=&fP=2&fM=12&fA=63&fD=7&fV=&fY=1996.32. Deepak Hedge and Sampat N. Bhaven, “Can Private Money Buy Public

Science? Disease Group Lobbying and Federal Funding for Biomedical Research,”Management Science 61, no. 10 (2015): 2281–2298.

178 nih and cancer-related patenting

a National Advisory Council – composed of scientists and represen-tatives of the public – makes a recommendation to the Director of theInstitute or Center. The latter ultimately takes the final funding de-cision33. For all Research Project Grants for instance, reviewers shallassess the overall impact, significance, investigators, innovation, ap-proach, and environment with additional review criterias dependingon the research at hand (such as protection of human subjects whenthose are involved)34. Per year, the NIH handles approximately 80,000

applications and engages approximately 20,000 reviewers35. Typically,to assess the significance of a project, the NIH guidance guidelinesask the reviewers to consider how the completion of the grant wouldchange the concepts, methods, technologies, treatments, services, orpreventative interventions that drive this field. For the assessment ofthe innovation, reviewers should determine whether the applicationseeks to refine, improve or apply new theoretical concepts, instru-mentation or interventions36. The practical impact of the project istherefore, at least partially, weighed in.

7.5 background on cancer patenting

The last decades have been characterized by a tremendous amountof technological progress in the field of cancer research. Because ofthe extensive R&D costs for the conception of drugs, their decade-lasting research, design, and clinical trial process over a regulatory la-byrinthine, the pharmaceutical industry has traditionally relied on pa-tents37. When looking at cancer-related patenting, Figure 7.3 shows

33. Hedge and Bhaven, “Can Private Money Buy Public Science? Disease GroupLobbying and Federal Funding for Biomedical Research.”

34. See Definitions of Criteria and Considerations for Research Pro-ject Grant (RPG/X01/R01/R03/R21/R33/R34) Critique, available underhttps://grants.nih.gov/grants/peer/critiques/rpgD.htm

35. NIH Peer Review: Grants and Cooperative Agreements (2013), available underhttps://grants.nih.gov/grants/peerreview22713webv2.pdf.

36. Definitions of Criteria and Considerations for Research Project Grant (RP-G/R01/R03/R15/R21/R34) Critique.

37. See, e.g., Dan L. Burk and Mark A. Lemley, “Policy Levers in Patent Law,” Vir-ginia Law Review 89, no. 7 (2003): 1575–1696 and Brian Kahin, “Through the Lensof Intangibles: What Patents on Software and Services Reveal about the System,” inPatents, Innovation and Economic Performance, OECD Conference Proceedings, ed. Adam

7.5 background on cancer patenting 179

the stark increase in patent application since 1976, particularly from1990 to 2004. In 2002, a record number of more than 10’000 patentapplications were filed within one year for the first time38. Not onebut all cancer technology as defined by the United States Patent andTrademark Office (USPTO)39 have been subject to rising patentingactivity40.

The strong growth of patents around the year 2000 has its root in va-rious explanations. The race fostered by the Human Genome Projectand Celera Genomics to sequence human DNA lead to breakthroughsin large-scale sequencing41. New technological prospects in cancer im-

B. Jaffe, Josh Lerner, and Scott Stern (OECD Publishing, 2004). PhRMA claim costsof $1.3 billion per new drug, with studies in the same range ($802 million, JosephA. DiMasi, Ronald W. Hansen, and Henry G. Grabowski, “The Price of Innovation:New Estimates of Drug Development Costs,” Journal of Health Economics 22 (2003):151–185) but also substantially lower ($43 million, Donald W. Light and Rebecca War-burton, “Demythologizing the High Costs of Pharmaceutical Research,” ManagementScience 50 (2004): 804–820).

38. Note that the technology fields are not exclusive, so that one patent may beclassified in two or more fields.

39. The USPTO classified the patents in one ore more of the seven following techno-logy fields: Drugs & Chemistry (including small molecule and large protein-basedtherapeutics), Diagnostic & Surgical Devices (including in vitro diagnostics andmedical devices), Radiation Measurement (including a subset of radiation therapyinventions), Data Science (including computer-related and imaging technologies),Food & Nutrition (including cancer prevention foods), Model Systems & Animals(including genetically modified cell lines and animals for testing of therapeutics andother experimentation), Cells & Enzymes (emphasizing cellular and molecular bi-ology), and Other & Pre-classification (interdisciplinary inventions and documentswhich have not yet been classified).

40. Jannigje G. Kers et al., “Trends in Genetic Patent Applications: The Commerci-alization of Academic Intellectual Property,” European Journal of Human Genetics 22

(2014): 1155–1159 show a similar increase in genetic patent applications in the year2000, followed by a decline of almost 50%; Shyh-Jen Wang, “The Stem Cell PatentLandscape as Relevant to Cancer Vaccines,” Human Vaccines 7, no. 10 (2011): 1100–1108 finds the same trend for US stem cell patents.

41. Matthew Herper, “Illumina Promises To Sequence Human Genome For $100 –But Not Quite Yet,” Forbes, January 2017, with the costs for sequencing the humangenome falling from $300,000 in 2006 to $1000 in 2017.

180 nih and cancer-related patenting

0

5000

10000

15000

1960 1970 1980 1990 2000 2010

Drugs & Chemistry

Cells & Enzymes

DNA, RNA, or Protein Sequence

Diagnostic & Surgical Devices

Radiation Measurement

Data Science

Food & Nutrition

Model Systems & Animals

Other & Pre-Classification

Figure 7.3: Growth in Cancer Patents Filings.

munotherapy42 or in stem cells43 paved the way for promising com-mercial applications44. This year also marked a change in patent re-gulation according to which patent applications for DNA sequenceshave to explain the function of the DNA sequence that was to be pa-tented45. This is eventually reflected in the increase in the numberof patents that disclose nucleic acid or amino acid sequences (DNA,RNA and Protein Sequences in the graphic), which may be used as

42. The year 1995 marks a shift in attitudes to cancer immunotherapy with evi-dence highlighting that the adaptive and innate immune systems can recognize andeliminate tumors, see Christopher Parish, “Cancer Immunotherapy: The Past, thePresent and the Future,” Immunology And Cell Biology 81 (2003): 106–113.

43. Gretchen Vogel, “Stem Cells Named Breakthrough of the Year,” Science, Decem-ber 1999, on stem cells.

44. Humira, an immunosuppressive medication manufactured by AbbVie was thetop-selling drug in the US in 2016 with $13.6 billion in sales, see https://www.meds

cape.com/viewarticle/886404.45. See United States Patent and Trademark Office, Manual of Patent Examining

Procedure, Ninth Edition, Revision 07.2015, Section 2421.

7.5 background on cancer patenting 181

a proxy for a rise of patents in personalized medicine or genomics46.One potential technical explanation to the surge of patent applica-tions is the implementation of electronic filing at the USPTO. Since2003, all newly filed patent applications are converted to electronicapplications, and in 2006 a web-based filing system for patent appli-cations was launched47. This might have increased the ability of theUPSTO to locate keywords on patent applications. The sudden jumpwould reflect the better delimitation of cancer-related patents, whilea number of false negatives was affecting the data previously.

The later decline – or stagnation – since the early 2000s may even-tually be explained by an increasing public concern about geneticpatenting48. Commentators also suggested that the completion of ad-vances in gene sequencing as well as the biotech market bubble in2000 profoundly impacted biotechnology and helps to explain thisdrop49. It is also possible that the peak in 2000 reflects that the in-dividual technologies have reached maturation50. Finally, changes inpatent regulations and practices by the USPTO may have contribu-ted to this decline as well. The courts and the USPTO have arguablybecome more demanding, so that applicants shall disclose fully theeffects of their inventions51. Recent Supreme Court cases, Associationfor Molecular Pathology v. Myriad Genetics, Inc., and Mayo Colla-borative Services v. Prometheus Laboratories, Inc. also impacted thepatenting of biology-based innovation52. After Myriad, the isolationof genomic DNA sequences identical to those found in nature are nolonger patent eligible subject matter. Mayo in return made unclearto which extent personalized medicine, more precisely patent claims

46. See Frumkin and Myers, USPTO Moonshot Patent Data.47. https://www.uspto.gov/about-us/news-updates/electronic-\patent-appl

ication-records-replace-paper-files-uspto.48. For instance, in 2001 the New York Times wrote about patents on genetically

engineered seeds and initiated a public discussion on genetic patenting, see SabraChartrand, “Patents in Supreme Court: Patents, Part of Fierce Battle over GeneticEngineering,” New York Times, March 2001,

49. See Stacy Lawrence, “Patent Drop Reveals Pressures on Industry,” Nature Bio-technology 22, no. 8 (2011): 930–931.

50. See 2nd Report of Session 2008-09, Genomic Medicine, (House of Lords Scienceand Technology Committee, 2009), p. 610 with this argument genetic screening.

51. See Lawrence, “Patent Drop Reveals Pressures on Industry”; the USPTO publis-hed new Utility Examination Guidelines becoming effective as of 5 January 2001.

52. 133 Supreme Court 2107 (2013) and 132 Supreme Court 1289 (2012).

182 nih and cancer-related patenting

to diagnostic methods involving natural correlations are patentable53.These new developments have impacted the filing strategy of innova-tors from 2012 onwards54.

7.6 results

7.6.1 Decrease in Overall Returns

This Section analyzes first the direct applied returns from total fun-ding, i.e., looking at the patents that were filed by publicly fundedresearchers, and second its indirect returns, i.e, assessing the publica-tions resulting from federal research subsequently cited as prior artby a patent.

Direct value of federal funding. The direct impact of the NIH oncancer-related inventions shifted over the years. Until the mid-1990,the share of federally funded patents rose, and up to one out of sixcancer patents benefited from NIH support (Figure 7.4).

1980 1990 2000 20100

2

4

6

8

10

12

14

Time

ShareNIHPatents

0

5000

10000

15000

NumberTotalPatents

Figure 7.4: Growth in Cancer Patents Filings & NIH.

These numbers show the major role of federally supported researchon innovation in cancer research’s earlier stages. As the patentingactivity continued to raise strongly, this percentage has been falling

53. Jennifer Gordon, “The Impact of Myriad and Mayo: Will Advancements in theBiological Sciences Be Spurred or Disincentivized? (Or Was Biotech Patenting NotComplicated Enough?),” Cold Spring Harbor Perspectives in Medicine 5, no. 5 (2015):a020917.

54. Mateo Aboy et al., “Myriad’s Impact on Gene Patents,” Nature Biotechnology 34

(2017): 1119–1123 finding that despite a drop in gene-related patents, the effect ofthe Myriad ruling have, however, been less profound than predicted.

7.6 results 183

over time to less than 1% in 2015. In total, 7,498 or 2.8% of all cancer-related patents received NIH funding.

When correcting for inflation55, the funding budget of the NCI hasbeen raising until 2004, but has since been declining by about onethird (Figure 7.5.)56. Interestingly, the patenting activity of researchsupported by the NIH first followed this trend, but since the mid-1990, not only the relative but also the absolute numbers of patentshas been falling (Figure 7.4 and 7.5). The number of patents ge-nerated by federally supported scientists declined in the followingyears. This suggests that the grants awarded at that time were not, atleast in terms of number of generated patents, particularly success-ful57. Potentially, the federal funds were raised with the expectationto reiterate their past successes.

1980 1990 2000 20102

3

4

5

6

7

8

Time

BudgetNCI($B)

0

100

200

300

400

NumberFederally-FundedPatents

Figure 7.5: Budget of NCI and NIH Patents.

The decline in terms of numbers and share of federally funded pa-tents is surprising considering the increase in funding, but not quitetelling in itself. Eventually, the NIH funded a lower number of pro-jects resulting in patents, but ultimately these projects had the greaterimpact. In fact, Kalutkiewicz and Ehman (2014) have suggested thatNIH-funded patents yield superior citation counts, which are com-

55. Using the GDP deflator from the World Development Indicators for the UnitedStates, with reference to the year 2015.

56. Although the NCI is not the sole NIH funder of cancer research, it representsby far the largest funding agency in cancer research, for instance 86.5% in FY 1997,see McGeary and Burstein, Sources of Cancer Research Funding in the United States.

57. It must be kept in mind that there exist a time lag between the year of theallocation of the funding and the resulting patent filing.

184 nih and cancer-related patenting

monly used as proxy of both the private and the social value of thepatented invention58. For cancer-related patents, NIH-supported pa-tents were, according to this proxy, of higher economic value thantheir counterparts in the early period from 1976 to 1990, but the con-trary is true since then (see Table 7.1).

Table 7.1: Forward Citations 5-Years Drugs & Chemistry

NIH-Funded Patents Non-Funded Patents Difference tests

Variables N Mean SD N Mean SD t-test DF

1976-1990 487 4.63 6.1 7,880 3.91 6.38 2.49** 553

1991-2013 5,221 11.32 32.31 175,438 12.88 38.47 -3.4*** 5669

* p < .1, ** p < .05, ***

Levene tests of equal variances revealed that the variances in standard deviation are not homo-

genous (p < 0.05), so that two-samples t-test of unequal variances were performed using the

Satterthwaite Approximation. The Satterthwaite degree’s of freedom (DF) are reported.

This tendency is further accentuated by an assessment of the distri-bution of patents: until 1990, the share of federally funded patentswas highest in the top 20% most cited patents, but since then thisshare culminates in middle-class patents, in terms of citations (Figure7.6.). The likelihood for a patent of generating no single citation washigher for non-funded, than for funded patents59. These various pie-ces of the puzzle mirror a conservative funding strategy, rather than

58. Forward citations are commonly used as a proxy for the technological impor-tance of the patented invention as well as, to a certain extent, of the value of saidpatent, see, e.g., Manuel Trajtenberg, “A Penny for Your Quotes: Patent Citations andthe Value of Innovations,” The RAND Journal of Economics 21, no. 1 (1990): 172–187;Bronwyn H. Hall, Adam Jaffe, and Manuel Tratjenberg, “Market Value and PatentCitations,” The RAND Journal of Economics 36 (2005): 16–38; and Dietmar Harhoff, Fre-deric M. Scherer, and Katrin Vopel, “Citations, Family Size, Opposition and the Valueof Patent Rights,” Research Policy 32, no. 8 (2003): 1343–1363; see David S. Abrams andBhaven N. Sampat, “Pharmaceutical Patent Citations and Real Value,” University ofPennsylvania Working Paper, 2017, for the field of pharmaceutical patents.

59. Approximately one out of six patents of federally funded patent were nevercited, versus slightly about one out of four non-funded patent.

7.6 results 185

a high-risks high-rewards one60. Over the last years, federal fundinghas resulted in a low amount of both low and high value patents61.

1.8

2.0

2.2

2.4

2.6

2.8

3.0

ShareofFederally-FundedPatents

(1990-2016

)

Bottom 20% 20-80% Most Cited Patents Top 20%

4.5

5.0

5.5

6.0

6.5

7.0

7.5

Bottom 20% 20 80% Most Cited PatentsShareofFederally-FundedPatents

(1976-1990

)

Figure 7.6: Trapeze Distribution of Patent Quality.

As to the inventions making it to market phase, another proxy for thevalue of the patents, the data shows that the number of NIH-fundeddrugs and chemistry cancer patents resulting in an FDA approvalremained constant over the years, and was lower by 12% than theirnon-funded counterparts (Figure 7.5)62. Additional proxies of patentquality revealed that some differences persisted over the two peri-ods. For instance, non-funded patents were filed in in more countries,while funded patents were broader in their technological scope (seeAppendix Table A7.1 and A7.2).

60. Shane Crotty, “The New NIH "Rule of 21" Threatens to Give Up on AmericanPreeminence in Biomedical Research Based on a Flawed Concept and Flawed Analy-sis,” Medium, 2017, ‘[a]merican science has always been "going for the gold", aimingto be the best of the best, the greatest in the world, like many American endeavors’.

61. See the Appendix for another set of proxies composed of patent scope, familysize, backward citations, non-patent literature and number of claims.

62. Overall, only c. 0.5% of all cancer-related patents in the drugs and chemistrytechnology reached an FDA approval, more specifically 0.55% versus 0.45%.

186 nih and cancer-related patenting

Indirect value of federal funding. Despite the overall strong increasein funding, the number of grants awarded by the National Cancer In-stitute decreased after 1985, and then, from 1990 onwards remainedstable until 2007 (Figure 7.7). The number of grants resulting in pu-blications that would later be cited by subsequent patents followeda parallel trend. What is surprising is that while there was a boomin non-federally funded cancer patents from the year 2000 onwards,this is not reflected in the present data. The stock of prior art built bypublicly funded scientists seem not to have increasingly been used bythe fast-growing number of patent applications.

●

●●

●●

● ●●

● ● ● ● ● ● ● ● ● ● ● ●● ● ●

■

■

■

■

■

■ ■

■

■ ■■ ■ ■ ■

■■

■ ■ ■■

■ ■ ■

● Linked Grants

■ Total Grants

1990 1995 2000 2005

1000

2000

3000

4000

5000

Figure 7.7: Grants with Linked Publication.

The average size of R01 equivalent grants, which form close to halfof all NIH grants of the Li, Azoulay, and Sampat (2017) dataset, in-creased over the year even when accounting for inflation (AppendixFigure A7.1)63. It appears that part of the increasing budget of the Na-tional Cancer Institute was allocated to increase the size of the grantsrather than their number. This seem to reflect that research is beco-ming more costly in terms of materials and regulations, especiallyfor following ethical protocols. Nevertheless, the fact that the increa-singly larger grants did not impact cancer innovation to a larger ex-tent, despite the doubling or tripling of patents that could potentiallycite the funded research is troubling.

63. Using the GDP deflator from the World Development Indicators for the UnitedStates, with reference to the year 2007.

7.6 results 187

7.6.2 Decrease in Grant Returns

This Section now first investigates the direct applied returns fromgrants awarded by the National Cancer Institute, i.e., it studies thelikelihood a grant leads to a patent by a federally funded scientist,and second the indirect applied returns from such grants, i.e., thenumber of grants it takes to yield a publication that is later cited bya patent. Combined, these two measures enable us to analyze thereturns of federal funding at the refined grant-level, rather than ontotal spending.

Direct value of federal funding. On total, c. 8% of all grants by theNational Cancer Institute yielded Bayh-Dole patents – that is patentsprimarily filed by universities, academic medical centers, and nonpro-fit research institutes64. This is higher than the returns for the NIHglobally, that resulted in c. 5% of its grants leading to Bayh-Dole pa-tents65. This overperformance is constant over time (Figure 7.8).

●

●

●●

●

● ●● ●

●●

●●

● ●● ●

● ●● ●

● ●

■

■

■■

■

■

■

■

■ ■

■

■ ■

■

■

■■

■

■ ■■

■ ■

● All NIH Grants

■ All NCI Grants

1990 1995 2000 2005

5

10

15

20

Figure 7.8: Probability of Grants to Patent

64. With the enactment of the Bayh-Dole Act, universities were allowed to elect toretain the title (i.e., obtain patents) of any subject inventions created under federalfunds, see Rebecca Eisenberg, “Patents: Help or Hindrance to Technology Transfer?,”in Biotechnology: Science, Engineering, and Ethical Challenges for the Twenty-First Century,ed. Frederick B. Rudolph and Larry McIntire (Joseph Henry Press, 1996), 161–174 fora landmark analysis.

65. See Li, Azoulay, and Sampat, “The Applied Value of Public Investments inBiomedical Research.”

188 nih and cancer-related patenting

The share of grants that lead to Bayh-Dole patents went up to over20% in 1987-1988, and is falling since then. Despite the increase of thesize of the grants (Appendix Figure A7.1), the returns by grant neverreached the levels of these earlier years.

Indirect value of federal funding. The returns by federal grant dra-matically shifted in 1995. Until that date, grants by the National Can-cer Institute yielded far superior results than the NIH grants (seeFigure 7.9). Depending on the year, over 40% to 60% of the grantswould yield a linked publication, i.e., a publication later cited by apatent. For NIH grants generally, this share was of a third. The trendsthen inversed in 1995, and since then, NIH patents yield higher re-turns in that regard than their NCI counterparts.

●● ● ●

●● ● ●

●

●

●●

●

●● ●

●

●

●● ●

●

●

■

■ ■

■■

■ ■

■

■ ■ ■■

■■

■■

■■

■■

■

■■

● All NIH Grants

■ All NCI Grants

1990 1995 2000 2005

10

20

30

40

50

60

Figure 7.9: Probability of Grants to Linked Publication.

Again, this is surprising considering the rise in the budget of theNational Cancer Institute, in the average size of the grants and the tri-pling of total patent applications from 2000 to 2004-2014. The relevantprior art on which these new patents built is not expanding the cita-tion rate of publicly funded research as one could expect. This seemsto indicate a decreased importance of the research by the NationalCancer Institute overall. The NIH has, during the last decades, facedcriticism over the insufficient returns of publications, especially after

7.7 discussion 189

raising its budget in 200066. Nevertheless, the NIH could maintain

and even increase the share of grants leading to linked patents untilthen, while this was not the case for the NCI.

7.7 discussion

In short, this paper presents evidence that there was a shift in theapplied returns to publicly funded cancer research around the year1995. Firstly, there is a decrease in overall returns: a lower absoluteand relative number of cancer-related patents received funding bythe NIH, and those were of lower technological impact. Second, thereis a decrease in the returns by grant: A lower likelihood for a grantby the NCI to yield a Bayh-Dole patent, and a lower likelihood for agrant to generate a publication subsequently cited by a patent. Thepresent Section seeks to discuss potential explanation factors that inisolation or combined might have lead to this shift.

Grant as policy instrument. Several commentators have presentedevidence of decreasing marginal returns of grants. Mainly, the scien-tific output by dollar spent on a scientist, as measured in terms ofpublications and their impact factor, increases first, peaks and then de-creases67. Also, larger grants seem not to lead to larger discoveries68.

66. See Michael S. Lauer et al., “Marginal Returns and Levels of Research GrantSupport Among Scientists Supported by the National Institutes of Health,” bioRxiv,2017, showing a decrease in the share of highly-cited papers per year.

67. With Jeremy Berg, Measuring the Scientific Output and Impact of NIGMS Grants,(NIGMS Feedback Loop Blog, September 2010) for a study of 2,938 investigator be-nefiting from grants from the National Institute of General Medical Sciences and fin-ding a plateau with grants larger than $700,000; see Philippe Mongeon et al., “Con-centration of Research Funding Leads to Decreasing Marginal Returns,” ResearchEvaluation 25, no. 4 (2016): 396–404 for a similar study in Quebec; see also Laueret al., “Marginal Returns and Levels of Research Grant Support Among ScientistsSupported by the National Institutes of Health” and Jon Lorsch, A Shared Respon-sibility, (NIGMS Feedback Loop Blog, January 2015) arguing that ‘funding smaller,more efficient research groups will increase the net impact of fundamental biome-dical research’. For a criticism, see Jeremy Berg, “Research Output as a Function ofGrant Support: the Scatter Matters,” Sciencemagazine, 2017, 2925–2935 and Crotty,“The New NIH "Rule of 21" Threatens to Give Up on American Preeminence inBiomedical Research Based on a Flawed Concept and Flawed Analysis.”

68. Fortin and Currie, “Big Science vs. Little Science: How Scientific Impact Scaleswith Funding” showing that greater funding is not strongly associated with greater

190 nih and cancer-related patenting

Grants scale only to a certain extent69. In our data, we find signs ofsuch a trend. Despite an increase in the average size of grants by theNCI, their returns in terms of patents and linked publications are de-creasing. This is nevertheless not true for the NIH. Despite criticismof the lack of diffusion of NIH research70, the publications resultingfrom its funding were more likely to be used in future patents thanthe NCI counterparts. If the choice of grants as policy instrument wascausal for this decline, we would arguably expect to find similar ef-fects in the general biomedical research, and not only in the field ofcancer71. In line with this, we argue that the choice of the policy in-strument of grants per se does not seem to be the principal drivingforce of the shift.

Saturation of cancer research. The productivity slowdown by publi-cly funded researchers may be caused by the exhaustion of inven-tive and technological opportunities72. Eventually, decreasing margi-nal returns to investment would converge towards a saturation pointwhere it becomes more and more costly to innovate. In fact, the out-put of funded research peaked in 1995 in terms of patents. Since then,despite an increase of funding, the number of patents is falling con-stantly. This hypothesis can be put into context with the raise in thecosts of conducting cancer research, mainly due to clinical trial73 and

productivity using funding data by the Natural Sciences and Engineering ResearchCouncil of Canada and Michael S. Lauer, “Applying the Relative Citation Ratio as aMeasure of Grant Productivity,” Extramural Nexus, 2016, for diminishing returns interms of citations for higher budgets and longer duration of funding for R01 grants.

69. See a recent study by Wayne P. Wahls, “High Cost of Bias: Diminishing MarginalReturns on NIH Grant Funding to Institutions,” bioRxiv, 2018, demonstrating aninverse correlations between funding and scientific output for NIH grants at theinstitution level.

70. David Gordon et al., “Publication of Trials Funded by the National Heart, Lung,and Blood Institute,” The New England Journal of Medicine 369, no. 20 (2013): 1926–1934 for the lack of publication of the results of publicly funded clinical trials.

71. In parallel, see Joseph E. Stiglitz and Arjun Jayadev, “Medicine for Tomorrow:Some Alternative Proposals to Promote Socially Beneficial Research and Develop-ment in Pharmaceuticals,” Journal of Generic Medicines 7, no. 3 (2010): 217–226 for acall for promoting prizes over patents in the field of pharmaceuticals.

72. See Zvi Griliches, “Patents: Recent Trends and Puzzles,” NBER Working Paper,no. 2922 (1989) for a study of historical patent data and the productivity slowdownin the 1970s.

73. See, e.g., DiMasi, Hansen, and Grabowski, “The Price of Innovation: New Esti-mates of Drug Development Costs” finding that the average costs to develop a drug

7.7 discussion 191

regulatory compliance74. The ‘easy’ problems have been demystified,what is left to solve in cancer research are larger and more costlyunknowns. The increase of the costs of research could explain part orall of the decline. It is nevertheless striking that the overall fundingof the National Cancer Institute or the average size of R01 grantsincreased over the years, accounting, in part at least, for the risingcost of research. It could, however, be the case that cancer researchbecame exponentially expensive so that the increase of the fundingbudget was not sufficient to keep up with the incurred costs of rese-arch. The idea of a saturated cancer research market does only poorlyfit the explosion of (non-funded) patent applications since 2000. Onealternative factor helping to explain the slowdown might be overfun-ding. Data suggests that cancer research might per se be overfundedwith regard to its societal burden (in terms of disability-adjusted life-years) compared to heart disease, stroke, depression, injuries or alco-hol abuse75. More prominently, amongst the different types of cancer,certain are overfunded relative to other less publicly visible, yet moredeadly cancers76. Because of public awareness, breast cancer has thehighest funding per-case and per-death77. The lack of optimization of

rose of annually 7.4% between 1987 and 2000, mainly due to the costs of clinical tri-als; Roger Collier, “Rapidly Rising Clinical Trial Costs Worry Researchers,” CanadianMedical Association Journal 180, no. 3 (2009): 277–278: ‘the cost of conducting a clinicaltrial for a drug is rising like mercury on a summer afternoon’.

74. See Paul M. Stewart et al., “Regulation - the Real Threat to Clinical Research,”The BMJ 337 (2008) and David J. Stewart, Simon N. Whitney, and Razelle Kurzrock,“Equipoise Lost: Ethics, Costs, and the Regulation of Cancer Clinical Research,” Jour-nal of Clinical Oncology 28, no. 17 (2010): 2925–2935; more specificity on the ethicalregulation in biomedical research, see Charles Warlow, “Over-Regulation of ClinicalResearch: a Threat to Public Health,” Clinical Medicine 5, no. 2 (2005): 33–38.

75. For a correlation graphic and an analysis as to the NIH diseases funding levelsand burden of disease, see the study by Leslie Gillum et al., “NIH Disease FundingLevels and Burden of Disease,” PLoS ONE 6 (2012).

76. Compare spendings of $1,630 for each lung cancer death and of $13,452 for bre-ast cancer by the NCI in 2006, see Matthew Parker, “Giving Teeth to European PatentReform, Overcoming Recent Legal Challenges,” Emory International Law Review 26

(2012): 1079–1110.77. See, e.g., Irwin Martin and Sowmya Mallela, “Funding of Cancer Research: Do

Levels Match Indidence and Mortality Rates?,” Therapeutic Innovation & RegulatoryScience 41, no. 1 (2014): 33–35 for the US and Ashley J. Carter, Beverly Delarosa,and Hannah Hur, “An Analysis of Discrepancies between United Kingdom CancerResearch Funding and Societal Burden and a Comparison to Previous and UnitedStates Values,” Health Research Policy and Systems 13, no. 1 (2015): 62–72 for the UK.

192 nih and cancer-related patenting

research spending with respect to the burden of the disease does notoptimize social welfare78. With more diversity, the returns on invest-ments might be higher79.

Shift in grant subject-matter. An alternative explanation is that grantswere awarded for more basic research and less translational and app-lied research since 1995

80. The decrease in the absolute number of pa-tents would simply reflect a decrease in science policy priority81. Onewould eventually expect less patents in this case. However, this goesagainst the current trend of the NIH, that has been cutting its budgeton basic research, which has fallen from two-thirds in the 1980s and1990s to a half in 2013

82. On the contrary, science policy has focusedon increasing economic incentives for research to be commercial and

78. Cary Gross, Gerard F. Anderson, and Neil R. Powe, “The Relation betweenFunding by the National Institutes of Health and the Burden of Disease,” The NewEngland Journal of Medicine 340, no. 24 (1999): 1881–1887 and Ashley J. Carter andCecine Nguyen, “A Comparison of Cancer Burden and Research Spending RevealsDiscrepancies in the Distribution of Research Funding,” BioMedCentral Public Health12 (2012): 526–537.

79. With the same argument, see Mark Peifer, “The Argument for Diversifying theNIH Grant Portfolio,” Molecular Biology of the Cell 28, no. 22 (2017): 2935–2940.

80. See Richard R. Nelson, “The Simple Economics of Basic Scientific Research,”Journal of Political Economy 67 (1959): 297–306 and Desiree Schauz, “What is BasicResearch? Insights from Historical Semantics,” Minerva 52, no. 3 (2014): 273–328 forthe distinction between basic and applied research; Ammon J. Salter and Ben Mar-tin, “The Economic Benefits of Publicly Funded Basic Research: a Critical Review,”Research Policy 30, no. 3 (2001): 509–532 for the economic benefits of publicly basicresearch; Fang and Casadevall, “Research Funding: the Case for a Modified Lottery”and Kimberly A. Moore, “Populism and Patents,” New York Law Review 82 (2007):69–111 for details on translational research.

81. Norman Sharpless, Director of the NCI describes this view as follows: ‘[w]hat Iencounter much more commonly is this notion that biology is an engineer problem.If we changed how we did things, we would start to make more progress. [..] All weneed to do is to fund applied research with an engineering mentality’ at the Policy,Politics and Law of Cancer Conference at Yale Law School (https://law.yale.edu/solomon-center/events/policy-politics-and-law-cancer/conference-videos).

82. See Eric Hand et al., “A Back Seat for Basic Science,” Nature 496, no. 7445

(2013): 277–279; also Stefano Bertuzzi and Don Cleveland, “The Curious Incidentof the Translational Dog that Did Not Bark,” Trends in Cell Biology 25, no. 4 (2015):187–189 arguing that the main determinant in the shift for the National Instituteof Neurological Disorders and Stroke was the decreased number of applications inbasic science.

7.7 discussion 193

with patenting potential83 Even if such claim was correct for cancerresearch, and the subject-matter shifted to more basic research, wewould expect an increase of publications and thus a larger stock ofknowledge. This is not the case in our data: since 1995 there is a lowerprobability a NCI grant yields a publication subsequently cited by apatent compared to a NIH grant84. Also, it has in fact been shownby Li, Azoulay, and Sampat (2017) that basic and applied researchwere equally likely to be cited by future patents. Another interes-ting alternative is the shift in cancer research towards data-drivenanalytics. The aggregation of large cancer disease knowledge and pa-tient database85 is predicted to have a huge impact on cancer researchand cancer care, argues the current director of the NCI86. It is, howe-ver, doubtful that this shift in focus has already materialized in ourdata.

Grant allocation system. It seems that the decrease in the appliedreturns to publicly funded research is not most likely to be explainedby the choice of grants as policy instrument, saturation of cancer rese-arch, and the shift of grant subject-matter. On the contrary, it appearsthat the allocation of funds for cancer research might be subject toflaws in the grant system – lack of predictive ability, gender and racialbias, focus on conform rather than transformative research, and du-plication of grants. Risk aversion by the NIH encourages researchersto present incremental research rather than novel, but risky ideas pus-hing the boundaries of discovery87. The NIH funding is considered

83. Ann Johnson, “The End of Pure Science? Science Policy from Bayh-Dole to theNNI,” in Discovering the Nanoscale, ed. Davis. Baird, Alfred Nordmann, and JoachimSchummer (IOS Press, 2003), 20–33 and Daniel L. Kleinman, “Untangling Context:Understanding a University Laboratory in the Commercial World,” Science, Techno-logy, & Human Values 23, no. 3 (1998): 285–314.

84. Galkina Cleary et al., “Contribution of NIH Funding to New Drug Approvals2010–2016” showing that NIH funding contributed to published research associatedwith all new drugs approved by the FDA from 2010 to 2016.

85. Such as canSAR by The Institute of Cancer Research or the American Associa-tion for Cancer Research’s Genomics, Evidence, Neoplasia, Information, Exchange.

86. Jocelyn Kaiser, “New NCI Director Expects Big Data to Revolutionize CancerResearch Care,” Science, December 2017, on this point.

87. https://www.aaas.org/news/scientists-frustrated-system-often-funds-incremental-work-over-risk-taking-historian-says.

194 nih and cancer-related patenting

to incentivize safe research rather than truly exploratory one88. Thiswould be in line with our results. Scientists generally suspect peerreviewers to be reluctant to select high-risk research89. Peer reviewis thought be biased against speculative, unorthodox and multidis-ciplinary research90. The incentives to lead lower-risk research arealso accentuated by the publication system, as while an innovativepublication is more likely to achieve high impact than a conserva-tive one, the reward might not compensate for the risk of failure topublish91. A plausible explanation for the slowdown in publicly fun-ded cancer research is that increasingly conservative projects werepresented by researchers, or that such projects were more likely to re-ceive federal funding (or perceived to be so). In the tension betweenplausibility and scientific value on the one side, and originality andcreativity on the other, the balance might have shifted92. In fact, it hasbeen proposed that the surge of scientists seeking grants in the 1990s

88. Pierre Azoulay, Joshua Graff Zivin, and Gustavo Manso, “Incentives and Crea-tivity: Evidence from the Academic Life Sciences,” The RAND Journal of Economics 42,no. 3 (2011): 527–554; but see Hyunwoo Park, Jeongsik Lee, and Byung-Cheol Kim,“Project Selection in NIH: A Natural Experiment from ARRA,” Research Policy 44, no.6 (2015): 1145–1159 for a study finding that NIH selects and funds risky projects.

89. See Daryl E. Chubin and Edward J. Hackett, Feerless Science: Peer Review and U.S.Science Policy (State University of New York Press, 1990).

90. This is suggested by studies by Chubin and Hackett, Feerless Science: Peer Reviewand U.S. Science Policy arguing that scientific feuds influence the outcomes of peerreview; Garett D. Travis and Harold M. Collins, “New Light on Old Boys: Cognitiveand Institutional Particularism in the Peer Review System,” Science, Technology, &Human Values 16, no. 3 (1991): 322–341 positing that frontier science is more likelyto suffer from cognitive cronyism (preference for cognitive similarities) than main-stream research; and Donald W. Braben, Pioneering Research: A Risk Worth Taking(Wiley-Interscience, 2004), p. 70 even claiming that ‘the natural inclination to opposemajor challenges to the status quo has become institutionalized’.

91. See Jacob G. Foster, Andrey Rzhetsky, and James A. Evans, “Tradition and Inno-vation in Scientists’ Research Strategies,” American Sociological Review 80, no. 5 (2015):875–908 with a network study of 6.4 million biochemistry papers finding that rese-arch is six times more likely to confirm prior results than carving out new territories;see Andrey Rzhetsky et al., “Choosing Experiments to Accelerate Collective Disco-very,” Proceedings of the National Academy of Sciences, 2015, inferring the typical rese-arch strategy in biomedicine and finding that ‘[t]his strategy generates conservativeresearch choices focused on building up knowledge around important molecules’.

92. See Thomas Heinze, “How to Sponsor Ground-Breaking Research: a Compari-son of Funding Schemes,” Science & Public Policy 35 (2008): p. 302 and Gerald Pollack,“Revitalizing Science in a Risk-Averse Culture: Reflections in the Syndrome and Pres-criptions for its Cure,” Cellular and Molecular Biology 51 (2005): 815–820 arguing that

7.8 conclusion 195

lead to more caution in grant proposals, which coincides with ourresults93.

The lack of patents as success. All explanation factors have so far con-sidered the decrease in the number of patents as a negative event. Analternative story is that this is on the contrary to be viewed as positivefor the development of technological progress in cancer. When uni-versities focus on upstream research, that is research removed fromthe commercial end product, patents might inhibit future innovationand preempt downstream applications94. While some patent theoristshave claimed that such broad upstream monopolies are necessary fortechnological progress by allowing the owner to coordinate down-stream research and avoid duplication efforts95, other consider com-petition crucial and therefore these patents as impediments96. Themain response to counter this hypothesis is that the decrease of pa-tents was not only in numbers, with lesser and lesser patents filed,but but also in their technological impact. Furthermore, the under-performance of grants in resulting in scientific literature later cited bypatents by the private sector is not to be explained by this story.

7.8 conclusion

Funding agencies seek to allocate academic research funding for thegeneration of high-quality and high-impact output. Agencies oftencall for transformative research, with an impact ‘much greater in mag-nitude’ than what would normally be expected from investments inscience97. To that end, as much as 20,000 reviewers seek to determinewhich of the 80,000 grant proposals should be awarded funding by

‘[m]ost grant systems were not designed to deal with proposals aimed at promulga-ting paradigm shifts’.

93. See Kolata, “Grant System Leads Cancer Researchers to Play It Safe.”94. Michael A. Heller and Rebecca S. Eisenberg, “Can Patents Deter Innovation?

The Anticommons in Biomedical Research,” Science 280, no. 5364 (1998): 698–701.95. Edmund W. Kitch, “The Nature and Function of the Patent System,” Journal of

Law and Economics 20, no. 2 (1977): 265–290.96. Robert P. Merges and Richard R. Nelson, “On the Complex Economics of Patent

Scope,” Columbia Law Review 90 (1990): 839–916.97. Engineering National Academies of Science and Medicine, Fostering Transforma-

tive Research in the Geographical Sciences, (2015).

196 nih and cancer-related patenting

the National Institutes of Health, and which not. The system seeks toidentify proposals based on their scientific merit and the excellenceof the applicant98. This study investigates grants as innovation policymechanism over a period of forty years in the field of cancer rese-arch.

The study finds that there was a shift around the year 1995, whenthe returns to public investments started to decrease. Since then, boththe overall returns and the returns by grants are on the fall. Espe-cially alarming is that there is not only a slowdown in the absolutenumber of patents but also that the technological impact of federallyfunded patents is decreasing. This is reinforced by an assessment ofthe indirect returns to funding. Grants are less likely to generate pu-blications subsequently cited by patents. In that regard, the viewpointthat research by the NIH should develop a set of tools and stock ofknowledge that the private sector can rely upon in problem-solving,and not necessarily patents per se, fails to convince. Amongst a rangeof explanation factors, the simplest and most likely that matches ourdatapoints is that the grant mechanism in cancer research might notbe innovative enough. In line with previous scholarly work depictingthe federal funding system as risk-adverse, the findings of this studysupport the view that the funding strategy sets its focus on incre-mental rather than on uncertain – but potentially game-changing –projects. The output of such funding practice was average, and didnot contribute to the innovative transformations that occurred sincethe early 2000s as much as it could potentially have.

Configuring review criteria might be a first step to strengthen thepractical impact of funded research. While the current review crite-ria by the NIH include, partly, the practical impact of a project, wepropose to refine the definitions or to add a new criteria with particu-lar emphasis on patenting. In particular, it is foreseen that reviewersrank the significance of a project – how the completion of the grantwould change concepts, methods, technologies, treatments and servi-ces driving the field – and its innovation – whether the applicationseeks to refine, improve or apply new theoretical concepts, instrumen-

98. See the mission statement of the ERC: ‘[t]he ERC’s mission is to encouragethe highest quality research in Europe through competitive funding and to supportinvestigator-driven frontier research across all fields, on the basis of scientific excel-lence’ under https://erc.europa.eu/abouterc/mission.

7.8 conclusion 197

tation or interventions99. These two criteria contribute to the overallimpact score. The NIH could include a sixth criteria, illustrativelytitled ‘patenting’, that would request the reviewer to detail whetherand how the proposed project can be turned into a patent directlyor fuel patentable inventions indirectly. Alternatively, and more con-servatively, the criteria ‘significance’ or ‘innovation’ could specificallybe defined as having a patent-related component. Accordingly, thedirect and indirect patenting potential could be subject to review; ormore generally, emphasize the importance of public-health orientedinventions100. Implementing review and valuation mechanisms thatalign with the goals of the NIH can help both applicants and revie-wers in their proposals and assessment. The previous literature sug-gests that the bias lies with the reviewers, who tend to be risk-adverseand favor conservative projects. Adjusting the review criteria mighttherefore be a first step towards more risk-friendly funding policies,that would in return incentivize applicants to present such projects.Should the NIH lower its barriers for risky projects by specifying thatis seeks to encourage research with potential practical impact in itscriteria, reviewers might factor in the higher likelihood of failure ofsuch ventures.

Translating biological discoveries into clinical applications can furt-her be strengthened by reenforcing targeted initiatives at the Natio-nal Cancer Institute and National Institutes of Health level. In orderto re-engineer translational science, a flagship initiative of the NIHwas to launch Clinical and Translational Science Awards in 2005; thiswas followed by the establishment of the National Center for Ad-vancing Translational Sciences in 2013 with a yearly budget of USD500 mio. In addition, mission-specific centers, such as the Translatio-nal Conte Centers (National Institute of Mental Health), and General

99. Definitions of Criteria and Considerations for Research Project Grant (RP-G/R01/R03/R15/R21/R34) Critique; innovation is defined as follows: ‘[d]oes theapplication challenge and seek to shift current research or clinical practice paradigmsby utilizing novel theoretical concepts, approaches or methodologies, instrumenta-tion, or interventions? Are the concepts, approaches or methodologies, instrumenta-tion, or interventions novel to one field of research or novel in a broad sense? Is arefinement, improvement, or new application of theoretical concepts, approaches ormethodologies, instrumentation, or interventions proposed?’.100. See also Amy Kapczynski and Talha Syed, “The Continuum of Excludabilityand the Limits of Patents,” Yale Law Journal 122 (2013): p. 1952.

198 nih and cancer-related patenting

Clinical Research Centers supported by the National Center for Re-search Resources provide for substantial resources for translationalresearch. Latter provide for large-scale research project funding fora diversity of fields and universities: illustratively a project on bio-engineering at Stanford with 1725 grants or on neurology at UCLAwith 776 grants101. Cancer research can benefit from the shift towardstranslational research, encouraged by both present and former NIHDirectors102. Establishing translational cancer research as a field at-tractive for private investments can also be a way to complementpublic funds, and to help to close the gap from bench-to-bedside.

101. RePORTER database, retrieved under http://projectreporter.nih.gov.102. Collins, “Reengineering Translational Science: The Time Is Right” shaping avision to pursue opportunities for ‘disruptive translational innovation’ and EliasA. Zerhouni, “Translational and Clinical Science — Time for a New Vision,” NewEngland Journal of Medicine 353, no. 15 (2005): 1621–1623 to ‘truly transform[..] humanhealth.

7.9 appendix 199

7.9 appendix

Table A7.1: Quality Proxies Drugs & Chemistry 1976-1990

NIH-Funded Patents Non-Funded Patents Difference tests

Variables N Mean SD N Mean SD t-test DF

Patent Scope 487 2.94 1.81 7,880 2.65 1.6 3.42*** 533

Family Size 487 6.29 6.04 7,880 8.7 8.54 -8.41*** 579

BW Citations 487 4.16 6.05 7,880 5.27 6.23 -3.89*** 551

NPL Citations 487 14.37 16.79 7,880 5.72 10.48 11*** 509

Claims 487 14.33 14.11 7,877 12.23 11.5 3.22*** 526

* p < .1, ** p < .05, ***

Levene tests of equal variances revealed that the variances in standard deviation are not homo-

genous (p < 0.05), so that two-samples t-test of unequal variances were performed using the

Satterthwaite Approximation. The Satterthwaite degree’s of freedom (DF) are reported.

Table A7.2: Quality Proxies Drugs & Chemistry 1991-2013

NIH-Funded Patents Non-Funded Patents Difference tests

Variables N Mean SD N Mean SD t-test DF

Patent Scope 5,221 3.57 1.81 175,438 3.22 1.66 13.87*** 5486

Family Size 5,221 6.56 6.03 175,438 9.83 8.14 -38.07*** 5841

BW Citations 5,221 14.39 23.57 175,438 18.42 27.48 -12.1*** 5650

NPL Citations 5,221 36.05 32.19 175,438 19.81 29.39 36*** 5482

Claims 5,221 16.56 15.47 118,856 1.59 15.44 -0.1 5651

* p < .1, ** p < .05, ***

Levene tests of equal variances revealed that the variances in standard deviation are not homo-

genous (p < 0.05), so that two-samples t-test of unequal variances were performed using the

Satterthwaite Approximation. The Satterthwaite degree’s of freedom (DF) are reported.

200 nih and cancer-related patenting

● ●

● ●●

●●

●● ●

●●

●●

●

●

●

●

● ● ●

●●

● Average Size R01 Grant

1990 1995 2000 2005

250000

300000

350000

Figure A7.1: Average Size of R01 Grants.

Part IV

E P I L O G U E

8C O N C L U D I N G R E M A R K S

This dissertation sought to answer a range of questions about theimpact of legal institutions on innovation in emerging technologies.A set of approaches utilizing interdisciplinary methods were develo-ped to investigate how institutions such as courts, funding agencies,and patent offices related to innovative activity. The practical reali-zation of innovative policies relies heavily on factors outside of thescope of the legislation, such as the implementation by institutionsand their agents or the cooperation of innovators. When introducingmeasurements of high-level policies, the dissertation finds that oftenthe policy aim does not align with the reality.

Take Europe and its effort to create a centralized market for innova-tion. Although claimed and foreseen by a variety of legal instruments,the European-wide enforcement of patent rights has not been realizedbecause of the lack of homogeneity in the application of harmonizedstandards . The choice of the forum of litigation, rather than the cha-racteristics of the particular rights at hand, are decisive in ensuringthe enforcement of patents (Chapter 3). The current but also futurelitigation system bear inherent flaws that will lead to the parallel en-forcement of patents across Europe. The uncertainties resulting the-reof for innovators are translated at the infringement and invaliditylevel, as illustrated by the study of biotechnology (Chapter 4). Thedissertation demonstrates how the harmonization of laws governingpatents or other rights cannot suffice by itself as implementing agentsconstruct interpretation standards in line with their prior practices ortraditions.

At the science policy level, large funding policies by the agencies andtolerant patent grant practices by patent offices can lead to unwantedeffects. For nanotechnology, the dissertation presents evidence thatthe patent system ultimately led to the formation of an intricate andconvulated patent thicket in the US. The chances that public-privatecollaboration, albeit increasingly popular, can mitigate thicket effects

203

204 concluding remarks

appear to be rather limited (Chapter 6). For cancer-related research,it appears that despite increasing the level of public investments overtime, funding agencies adopted an incremental, rather than a high-risk high-reward strategy, that resulted in a productivity slowdown(Chapter 7).

When we recall the set of patterns instrumental for emerging techno-logies to grow (governmental funding, transfer of knowledge andcorporate innovation), our findings suggest they are, for individualtechnologies at least and under a specific studied angle, not fulfilled.Public funding has taken a conservative approach, hindering inno-vation. The transfer of knowledge is staunched by the rise of over-lapping patents and firms fail to remedy to the lack of transfer byengaging in collaborations with universities. Corporate innovation ischallenged by the practical heterogeneity of a theoretically harmoni-zed patent enforcement system.

These findings have a number of policy implications for the design oflegal institutions and policies. The dissertation shows that innovationmechanisms that were adapted to foster emerging technologies halfa century ago might need to be refined and adjusted over time. Thepolicymaker, the funding agencies and the courts seem to be – tosome extent – hindering innovation rather than setting it free.

Regional integration and harmonization of laws must be accompa-nied by the harmonization of standards governing implementing agen-ts, such as the judges serving in courts. As illustrated by patent rig-hts, the social benefits of centrally granted rights can be underminedif courts inteprete harmonized laws according to their own previoustraditions. In that regard, establishing a common court of appeal fornational courts, or rather as proposed with the Unified Patent Court,a supranational enforcement system appears necessary to counter na-tional resistance and legal uncertainty. Such court can level-up nati-onal differences by creating a common European standard of inter-pretation. Latter can be strengthened by a multinational compositionof judge panels, as well as by judicial training. However, it must bestressed that that under the Unified Patent Court as foreseen, therewill exist a set of option for litigants to conduct a gamification of thesystem, and seek to exploit differences at the national and supranatio-nal level. To that end, it is essential that the unified court is consistent

concluding remarks 205

in its precedents, and follows the jurisprudence of the Court of Ap-peal. In doing so, the Unified Patent Court can present itself as anattractive and efficient venue for patent disputes, indirectly pressu-ring non-participating countries to seek consistency and abstain fromdeparting from points of law authoritatively decided at the supra-national level. The court will be able to demonstrate such benefitsif inventors, patent holders, and their competitors make use of thecourts.

Science policies, to be optimal in their scope, must start to evolveand be reassessed to account for changes in technology. When pa-tent offices grant patents with great ease, they fuel the potential foroverlapping rights to emerge. The application of restrictive noveltyand obviousness standards by the patent offices and strict standardsby the courts can carve out patents of lower quality, which altoget-her might hinder innovation rather than foster it. Similarly, fundingpractices should factor their practical outcome. When latter is toolow, a change is needed. Illustratively, implementing more risk-pronefunding practices through the modification of funding rules or thedevelopment of specific high-risk high-rewards projects can help toboost output overall. The more conservative a grant system becomes,the less effective it is compared to alternative innovation policy me-chanisms. If the grant system fails at identifying the highest-impactprojects, its efficiency is in doubt. In that regard, the regular asses-sment by quantitative models can help the policy maker to estimatewhether or not it reaches its policy goals. The public release and easyaccess to detailed datasets by funding agencies, patent offices andcourts can contribute to such knowledge, as it provides scientists ofall fields with a datasource of study. Measuring the research and in-novative output of such institutions might help them to decide in fullknowledge of the facts. It is therefore the conviction of the author thatfurther research in science policy and in the interdisciplinary studyof the law and the legal institutions should be encouraged.

B I B L I O G R A P H Y

2nd Report of Session 2008-09, Genomic Medicine. House of Lords Scienceand Technology Committee, 2009.

Aarnio, Aulis. Essays on the Doctrinal Study of Law. Springer, 2011.

Aboy, Mateo, Johnathon Liddicoat, Kathleen Liddell, Matthew Jordan,and Cristina Crespo. “Myriad’s Impact on Gene Patents.” NatureBiotechnology 34 (2017): 1119–1123.

Abrams, David S., Ufuk Akcigit, and Jillian Grennan. “Patent Va-lue and Citations: Creative Destruction or Strategic Disruption?”NBER Working Paper, no. 19647 (2013).

Abrams, David S., and Bhaven N. Sampat. “Pharmaceutical Patent Ci-tations and Real Value.” University of Pennsylvania Working Paper,2017.

Adams, Charles W. “The Court of Appeals for the Federal Circuit:More Than a National Patent Court.” Missouri Law Review 49

(1984): 43–84.

Adelman, Martin J., and Sonia Baldia. “Prospects and Limits of thePatent Provision in the TRIPS Agreement: The Case of India.”Vanderbilt Journal of Transnational Law 20 (1996): 507–534.

Adler, Jonathan H. “Eyes on a Climate Prize: Rewarding Energy Inno-vation to Achieve Climate Stabilization.” Harvard EnvironmentalLaw Review 35, no. 1 (2011): 1–45.

Aerts, Rob J. “The Unitary Patent and the Biotechnology Directive:Is Uniform Protection of Biotechnological Inventions Ensured?”European Intellectual Property Review 9 (2014): 584–587.

Aghion, Philippe, and Peter Howitt. “A Model of Growth throughCreative Destruction.” Econometrica 60, no. 2 (1992): 323–351.

206

Bibliography 207

Akcigit, Ufuk, John Grigsby, and Tom Nicholas. “The Rise of Ame-rican Ingenuity: Innovation and Inventors of the Golden Age.”NBER Working Paper, no. 23047 (2017).

Alesina, Alberto, Ignazio Angeloni, and Federico Etro. “InternationalUnions.” American Economic Review 95 (2005): 602–615.

Alesina, Alberto, Angeloni Ignazio, and Schuknecht Ludger. “WhatDoes the European Union Do?” Public Choice 123 (2005): 275–319.

Alic, John, David Mowery, and Edward Rubin. U.S. Technology andInnovation Policies: Lessons for Climate Change. Report Prepared fothe Pew Center on Global Climate Change, 2003.

Allison, John R., and Mark A. Lemley. “Empirical Evidence on theValidity of Litigated Patents.” IPLA Quarterly Journal 26 (1998):185–226.

. “Who’s Patenting What? An Empirical Exploration of PatentProsecution.” Vanderbilt Law Review 6 (2000): 2099–2148.

Allison, John R., Mark A. Lemley, Kimberly A. Moore, and R. DerekTrunkey. “Valuable Patents.” Georgetown Law Journal 92 (2004).

Allison, John R., Mark A. Lemley, and David L. Schwartz. “Under-standing the Realities of Modern Patent Litigation.” Texas LawReview 92 (2014): 1769–1801.

. “Our Divided Patent System.” University of Chicago Law Re-view 82, no. 3 (2015): 1073–1154.

Alstott, Jeff, Giorgio Triulzi, Bowen Yan, and Jianxi Luo. “MappingTechnology Space by Normalizing Patent Networks.” Scientome-trics 110, no. 1 (2017): 443–479.

Anand, Bharat, and Tarun Khanna. “The Structure of Licensing Con-tracts.” Journal of Industrial Economics 48, no. 1 (2000): 103–135.

Anderson, Jonas J. “Court Competition for Patent Cases.” Universityof Pennsylvania Law Review 163, no. 3 (1997): 631–698.

208 Bibliography

Arrow, Kenneth. “Economic Welfare and the Allocation of Resourcesfor Invention.” In The Rate and Direction of Inventive Activity: Eco-nomic and Social Factors, 609–626. National Bureau of EconomicResearch, 1962.

Azoulay, Pierre, Waverly Ding, and Toby. Stuart. “The Impact of Aca-demic Patenting on the Rate, Quality and Direction of (Public)Research Output.” Journal of Industrial Economics 57, no. 4 (2009):637–676.

Azoulay, Pierre, Joshua Graff Zivin, Danielle Li, and Bhaven N. Sam-pat. “Public R&D Investments and Private-Sector Patenting: Evi-dence from NIH Funding Rules.” NBER Working Paper, no. 20889

(2015).

Azoulay, Pierre, Joshua Graff Zivin, and Gustavo Manso. “Incentivesand Creativity: Evidence from the Academic Life Sciences.” TheRAND Journal of Economics 42, no. 3 (2011): 527–554.

Balconi, Margherita, Stefano Brusoni, and Luigi Orsenigo. “In De-fence of the Linear Model: An Essay.” Research Policy 39, no. 1

(2010): 1–13.

Baldan, Federica, and Esther Van Zimmeren. “The Future Role of theUnified Patent Court in Safeguarding Coherence in the EuropeanPatent System.” Common Market Law Review 52, no. 6 (2015): 1529–1577.

Barbosa, Denis, and Kuntz Grau. Exclusions from Patentable SubjectMatter and Exceptions and Limitations to the Rights: Biotechnology.World Intellectual Property Organization, 2010.

Barham, Leela. “European Pharma and the Unified Patent Court.”PharmExec.com, July 2016.

Basmann, Robert L., Michael McAleer, and Daniel Slottje. “PatentActivity and Technical Change.” Journal of Econometrics 139 (2007):355–375.

Baumann, Florian, and Tim Friehe. On Discovery, Restricting Lawyers,and the Settlement Rate. DICE Discussion Paper 155. 2014.

Bibliography 209

Baumol, William J. “Contestable Markets: An Uprising in the Theoryof Industry Structure.” American Economic Review, 72, no. 1 (1975):1–15.

Bawa, Raj, and Stephen Maebius. “The Nanotechnology Patent "Gold"Rush.” Journal of Intellectual Property Rights 10 (2005): 426–433.

Bayliss, Christopher J. “The Unitary Patent and Unified Patent Court:Potential Changes and Implications.” International Review of Intel-lectual Property and Competition Law 5, no. 2 (2014): 433–475.

Becker, Wolfgang, and Jurgen Dietz. “R&D Cooperation and Innova-tion Activities of Firms – Evidence for the German Manufactu-ring Industry.” Research Policy 33, no. 2 (2004): 209–223.

Belderbos, Rene, Bruno Cassiman, Dries Faems, Bart Leten, and BartVan Looy. “Co-Ownership of Intellectual Property: Exploring theValue-Appropriation and Value-Creation Implications of Co Pa-tenting With Different Partners.” Research Policy 43, no. 5 (2014):841–852.

Bender, Getchen. “Clash of the Titans: The Territoriality of Patent Lawvs. the European Union.” IDEA: The Journal of Law and Technology40 (2000): 49–82.

Berg, Jeremy. Measuring the Scientific Output and Impact of NIGMSGrants. NIGMS Feedback Loop Blog, September 2010.

. “Research Output as a Function of Grant Support: the ScatterMatters.” Sciencemagazine, 2017, 2925–2935.

Berg, Sanford V., and John Tschirhart. Natural Monopoly Regulation:Principles and Practice. Cambridge University Press, 1988.

Bertuzzi, Stefano, and Don Cleveland. “The Curious Incident of theTranslational Dog that Did Not Bark.” Trends in Cell Biology 25,no. 4 (2015): 187–189.

Berube, David M. Nano-Hype: The Truth Behind the Nanotechnology Buzz.Prometheus Books, 2006.

Bessen, James, and Michael J. Meurer. “Lessons for Patent Policy fromEmpirical Research on Patent Litigation.” Lewis and Clark Review9, no. 1 (2005): 1–27.

210 Bibliography

Bessen, James, and Michael J. Meurer. “The Private Costs of PatentLitigation.” Boston University School of Law Working Paper, nos. 07-08 (2007).

. Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innova-tion at Risk. Princeton University Press, 2008.

Betancourt, Alba A. “Cross-Border Patent Disputes: Unified PatentCourt or International Commercial Arbitration?” Utrecht Journalof International and European Law 32, no. 82 (2016): 44–58.

Bhat, Jyoti S. “Concerns of New Technology Based Industries – TheCase of Nanotechnology.” Technovation 25, no. 5 (2005): 457–462.

Bister, Klaus. “Discovery of Oncogenes: The Advent of MolecularCancer Research.” Proceedings of the National Academy of Sciences112, no. 50 (2015): 15259–15260.

Boldrin, Michele, and David K. Levine. Against Intellectual Monopoly.Cambridge University Press, 2008.

. “The Case Against Patents.” The Journal of Economic Perspecti-ves 27, no. 1 (2013): 3–22.

Braben, Donald W. Pioneering Research: A Risk Worth Taking. Wiley-Interscience, 2004.

Brandi-Dohrn, Matthias. “Some Critical Observations on Competenceand Procedure of the Unified Patent Court.” International Reviewof Intellectual Property and Competition Law 43 (2012): 372–389.

Branstetter, Lee G. “Do Stronger Patents Induce More Local Innova-tion?” Journal of International Economic Law 7, no. 2 (2004): 359–370.

Bregoli, Lisa, Dania Movia, James Gavigan-Imedio, Joanne Lysaght,John Reynolds, and Adriele Prina-Mello. “Nanomedicine App-lied to Translational Oncology: A Future Perspective on CancerTreatment.” Nanomedicine: Nanotechnology, Biology and Medicine12, no. 1 (2016): 81–103.

Brown, Shona L., and Kathleen M. Eisenhardt. “Product Develop-ment: Past Research, Present Findings, and Future Directions.”Academy of Management Review 20, no. 2 (1995): 343–378.

Bibliography 211

Burk, Dan L., and Mark A. Lemley. “Is Patent Law Technology-Specific?”Berkeley Technology Law Journal 17 (2002): 1155–1206.

. “Policy Levers in Patent Law.” Virginia Law Review 89, no. 7

(2003): 1575–1696.

. “Fence Posts or Sign Posts: Rethinking Patent Claim Con-struction.” University of Pennsylvania Law Review 157 (2009): 1743–1799.

. The Patent Crisis and How the Courts Can Solve It. The Univer-sity of Chicago Press, 2009.

Burstein, Michael J., and Fiona Murray. “Innovation Prizes in Practiceand Theory.” Harvard Journal of Law and Technology 29, no. 2 (2016):401–452.

Bush, Vannevar. Science, the Endless Frontier; A Report to the Presidenton a Program for Postwar Scientific Research. National Science Foun-dation, 1960.

Calignano, Giuseppe. “Italian Organisations Within the European Na-notechnology Network: Presence, Dynamics and Effects.” Die Erde14, no. 4 (2014): 241–259.

Campbell, Randy L. “Global Patent Law Harmonization: Benefits andImplementation.” Indiana International & Comparative Law Review605, no. 13 (2003): 605–638.

Cane, Peter, and Herbert Kritzer. The Oxford Handbook of EmpiricalLegal Research. Oxford University Press, 2010.

Carroll, Michael W. “One for All: The Problem of Uniformity Costin Intellectual Property Law.” American University Law Review 55,no. 4 (2009): 845–900.

. “One Size Does Not Fit All: A Framework for Tailoring Intel-lectual Property Rights.” Ohio State Law Journal 70, no. 6 (2009):1361–1434.

212 Bibliography

Carter, Ashley J., Beverly Delarosa, and Hannah Hur. “An Analysisof Discrepancies between United Kingdom Cancer Research Fun-ding and Societal Burden and a Comparison to Previous andUnited States Values.” Health Research Policy and Systems 13, no. 1

(2015): 62–72.

Carter, Ashley J., and Cecine Nguyen. “A Comparison of Cancer Bur-den and Research Spending Reveals Discrepancies in the Dis-tribution of Research Funding.” BioMedCentral Public Health 12

(2012): 526–537.

Chartrand, Sabra. “Patents in Supreme Court: Patents, Part of FierceBattle over Genetic Engineering.” New York Times, March 2001.

Chisum, Donald S. “Common Law and Civil Law Approaches to Pa-tent Claim Interpretation: "Fence Posts" and "Sign Posts".” In In-tellectual Property in the New Millennium: Essays in Honour of Wil-liam R. Cornish, edited by David Vaver and Lionel Bently. Cam-bridge University Press, 2004.

Choi, Stephen J., and Mitu G. Gulati. “Bias in Judicial Citations: AWindow into the Behavior of Judges?” Journal of Legal Studies 37

(2008): 87–129.

Chubin, Daryl E., and Edward J. Hackett. Feerless Science: Peer Reviewand U.S. Science Policy. State University of New York Press, 1990.

Chun, Dongwook. “Patent Law Harmonization in the Age of Globa-lization: The Necessity and Strategy for a Pragmatic Outcome.”Journal of Patent Trademark Office Society 93, no. 2 (2011): 127–166.

Clarkson, Gavin. “Patent Informatics for Patent Thicket Detection: ANetwork Analytic Approach for Measuring the Density of PatentSpace,” January 2005.

Clermont, Kevin M. “Litigation Realities Redux.” Notre Dame Law Re-view 84 (2009): 1951–1956.

Clermont, Kevin M., and Theodore Eisenberg. “Litigation Realities.”Cornell Law Review 119 (2002): 119–154.

Bibliography 213

Cockburn, Iain M., and Megan J. MacGarvie. “Patents, Thickets andthe Financing of Early-Stage Firms: Evidence from the SoftwareIndustry.” Journal of Economics & Management Strategy 18, no. 3,729–773.

Cohen, Julie E., and Mark A. Lemley. “Patent Scope and Innovationin the Software Industry.” California Law Review 89 (2001): 1–57.

Cohen, Wesley M., Richard R. Nelson, and John P. Walsh. “ProtectingTheir Intellectual Assets: Appropriability Conditions and WhyU.S. Manufacturing Firms Patent (or Not).” NBER Working Paper,no. 7552 (2000).

. “Links and Impacts: The Influence of Public Research on In-dustrial R&D.” Management Science 48, no. 1 (2002): 1–23.

Colaianni, Alessandra, and Robert Cook-Deegan. “Columbia Univer-sity’s Axel Patents: Technology Transfer and Implications for theBayh-Dole Act.” The Milbank Quarterly 87, no. 3, 683–715.

Collier, Roger. “Rapidly Rising Clinical Trial Costs Worry Researchers.”Canadian Medical Association Journal 180, no. 3 (2009): 277–278.

Collins, Francis S. “Reengineering Translational Science: The Time IsRight.” Science Translational Medicine 3, no. 90 (2011): 1–6.

Cooter, Robert. “Bargaining with the State: Offsets and Mitigationin Developing Land.” In Internationalisierung des Rechts und seineökonomische Analyse; Internaionalization of the Law and its EconomicAnalysis, edited by Thomas Eger, Claus Ott, and Jochen Bigus,423–429. Gabler Edition Wissenschaft, 2008.

Cooter, Robert, and Thomas Ulen. Law & Economics. Berkeley LawBooks, 2016.

Cotropia, Christopher A. “Patent Claim Interpretation Methodologiesand Their Claim Scope Paradigm.” William & Mary Law Review47, no. 1 (2005): 49–132.

Cotropia, Christopher A., Jay P. Kesan, and David L. Schwartz. “He-terogeneity Among Patent Plaintiffs: An Empirical Analysis ofPatent Case Progression, Settlement, and Adjudication.” Journalof Empirical Legal Studies 15, no. 1, 80–125.

214 Bibliography

Crampes, Claude, and Corinne Langinier. “Litigation and Settlementin Patent Infringement Cases.” The RAND Journal of Economics 33

(2002): 258–274.

Cremers, Katrin. “Determinants of Patent Litigation in Germany.”ZEW Discussion Paper, nos. 04-72 (2004).

. “Settlement During Patent Litigation Trials. An EmpiricalAnalysis for Germany.” The Journal of Technology Transfer 34, no. 2

(2009): 182–195.

Cremers, Katrin, Max Ernicke, Fabian Gaessler, Dietmar Harhoff, Chris-tian Helmers, Luke McDonagh, Paula Schliessler, and Nicolasvan Zeebroeck. “Patent Litigation in Europe.” European Journal ofLaw and Economics 44, no. 1 (2017): 1–44.

Cremers, Katrin, Fabian Gaessler, Dietmar Harhoff, Christian Hel-mers, and Yassine Lefouili. “Invalid but Infringed? An Analysisof the Bifurcated Patent Litigation System.” Journal of EconomicBehavior & Organization 131 (2016): 218–242.

Crotty, Shane. “The New NIH "Rule of 21" Threatens to Give Up onAmerican Preeminence in Biomedical Research Based on a Fla-wed Concept and Flawed Analysis.” Medium, 2017.

Crouch, Dennis. “Person(s) Skilled in the Art: Should the Now Esta-blished Model of Team-Based Inventing Impact the ObviousnessAnalysis?” Patently-O, May 2011.

. “THOSITA: Obvious to a Team Having Ordinary Skill in theArt.” Patently-O, October 2011.

Cummings, Scott W. “The Role of Trade Secrets in Today’s Nano-technology Patent Environment.” Nanotechnology Law & Business5 (2008): 41–52.

Cuomo, Margaret. A World Without Cancer: The Making of a New Cureand the Real Promise of Prevention. Rodale, 2012.

D’Silva, Joel. “Pools, Thickets and Open Source Nanotechnology.”European Intellectual Property Review 31, no. 6 (2009): 300–306.

Dam, Kenneth W. “The Economic Underpinnings of Patent Law.” TheJournal of Legal Studies 23, no. 1 (1994): 247–271.

Bibliography 215

Darrow, Jonathan. “The Neglected Dimension of Patent Law’s PHO-SITA Standard.” Harvard Journal of Law & Technology 23, no. 1

(2009): 227–258.

Daughety, Andrew F., and Jennifer F. Reinganum. “Economic Theo-ries of Settlement Bargaining.” Annual Review of Law and SocialScience 1, no. 1 (2005): 35–59.

De Jongand, Wim, and Paula Borm. “Drug Delivery and Nanoparti-cles: Applications and Hazards.” International Journal of Nanome-dicine 3, no. 2 (2008): 133–149.

De Korte, David, and Gavin Clarkson. “The Problem of Patent Thic-kets in Convergent Technologies.” Annals of the New York Academyof Sciences 1093 (2006): 180–200.

De Laat, Eric A. “Patents or Prizes: Monopolistic R&D and Asymme-tric Information.” International Journal of Industrial Organization15, no. 3 (1997): 369–390.

Deakin, Simon. “Legal Diversity and Regulatory Competition: WhichModel for Europe?” European Law Journal 12, no. 4 (2006): 440–454.

deGrazia, Charles, Jesse Frumkin, and Nicholas A. Pairolero. “Em-bracing Invention Similarity for the Measurement of VerticallyOverlapping Claims.” USPTO Economic Working Paper, nos. 2018-01 (2018).

Devlin, Alan. “The Misunderstood Function of Disclosure in PatentLaw.” Harvard Journal of Law & Technology 23, no. 2 (2010): 401–446.

Di Cataldo, Vincenzo. “From the European Patent to a CommunityPatent.” Columbia Journal of European Law 8 (2002): 19–36.

DiMasi, Joseph A., Ronald W. Hansen, and Henry G. Grabowski.“The Price of Innovation: New Estimates of Drug DevelopmentCosts.” Journal of Health Economics 22 (2003): 151–185.

Donohue, John J. “The Effects of Fee Shifting on the Settlement Rate:Theoretical Observations on Costs, Conflicts, and ContingencyFees.” Law and Contemporary Problems 54 (1991): 195–222.

216 Bibliography

Drazen, Jeffrey M., and Julie R. Ingelfinger. “Grants, Politics, and theNIH.” New England Journal of Medicine 349, no. 23 (2007): 2259–2261.

Drexler, Eric. Engines of Creation: The Coming Era of Nanotechnology.Anchor Books, 1986.

Dreyfuss, Rochelle C., and Andreas F. Lowenfeld. “Two Achievementsof the Uruguay Round: Putting TRIPS and Dispute SettlementsTogether.” Virginia Journal of International Law Association 37 (1997):275–334.

Duffy, John F. “Harmony and Diversity in Global Patent Law.” Berke-ley Technology Law Journal 17 (2002): 685–726.

Dumont, Béatrice. “Does Patent Quality Drive Damages in PatentLawsuits? Lessons from the French Judicial System.” Review ofLaw & Economics 11, no. 2 (2015): 355–383.

Editorial. Who Speaks for Small Biotech? 25. Nature Biotechnology, 2007.

. Patents Pending. 11. Nature Materials, 2012.

Egan, Edward J., and David J. Teece. “Untangling the Patent Thic-ket Literature.” University of California Berkeley Tusher Center forManagement Intellectual Capital Working Paper, no. 7 (2015).

Eisenberg, Rebecca. “Patents and the Progress of Science: ExclusiveRights and Experimental Use.” University of Chicago Law Review56 (1989): 1025–1028.

. “Patents: Help or Hindrance to Technology Transfer?” In Bio-technology: Science, Engineering, and Ethical Challenges for the Twenty-First Century, edited by Frederick B. Rudolph and Larry McIntire,161–174. Joseph Henry Press, 1996.

. “Obvious to Whom? Evaluating Inventions from the Perspectiveof PHOSITA.” Berkeley Technology Law Journal 19 (3 2004): 885–906.

Eisenberg, Theodore. “Why Do Empirical Legal Scholarship?” SanDiego Law Review 41, no. 4 (2004): 1741–1746.

Bibliography 217

. “The Origins, Nature, and Promise of Empirical Legal Studiesand a Response to Concerns.” University of Illinois Law Review 5

(2011): 1713–1738.

Eisenberg, Theodore, Paula L. Hannaford-Agor, Valerie P. Hans, Ni-cole L. Waters, G. Thomas Munsterman, Stewart J. Schwab, andMartin T. Wells. “Judge-Jury Agreement in Criminal Cases: APartial Replication of Kalven and Zeisel’s The American Jury.”Journal of Empirical Legal Studies 2, no. 1, 171–207.

Ejermo, Olof, and John Källström. “What is the Causal Effect of R&Don Patenting Activity in a “Professor’s Privilege” Country? Evi-dence from Sweden.” Small Business Economics 47, no. 3 (2016):677–694.

Ellis, Thomas S. “Distortion of Patent Economics by Litigation Costs.”CASRIP Publication Series: Streamlining International Intellectual Pro-perty, 1999.

Elmer, Michael, and Stacy Lewis. “Where to Win: Patent-FriendlyCourts Revealed.” Managing Intellectual Property, October 2010.

Engelbrekt, Antonina B. “Dilemmas of Governance in Multilevel Eu-ropean Patent System.” In National Developments in the Intersectionof IPC and Competition Law, edited by Hans Henrik Lidgard. HartPublishing, 2011.

England, Paul. “Novelty of Patents in Europe and the UPC.” Journalof Intellectual Property Law & Practice 12, no. 9 (2017): 739–746.

Fang, Ferric, and Arturo Casadevall. “Research Funding: the Case fora Modified Lottery.” mBio, no. 2 (2016).

Federico, Pasquale J. “Adjudicated Patents, 1948-1954.” Journal of thePatent Office Society 38 (1956): 233.

Feldman, Maryann P., Alessandra Colaianni, and Connie K. Liu. “Les-sons from the Commercialization of the Cohen-Boyer Patents:The Stanford University Licensing Program.” In Intellectual Pro-perty Management in Health and Agricultural Innovation: A Hand-book of Best Practices, 1208–1797. 2008.

218 Bibliography

Fershtman, Chaim, and Morton I. Kamien. “Cross Licensing of Com-plementary Technologies.” International Journal of Industrial Orga-nization 10, no. 3 (1992): 329–348.

Feynman, Richard. “There’s Plenty of Room at the Bottom.” Engineer-ing and Science 23, no. 5 (1960): 22–36.

Fischer, Marius. Upstream-Patente in der Nanotechnologie, Ein Vergleichzwischen Europa und den USA. forthecoming.

Fischer, Timo, and Joachim Henkel. “Patent Trolls on Markets forTechnology – An Empirical Analysis of NPEs’ Patent Acquisi-tions.” Research Policy 41, no. 9 (2012): 1519–1533.

Fortin, Jean-Michel, and David J. Currie. “Big Science vs. Little Science:How Scientific Impact Scales with Funding.” PLoS ONE 8, no. 6

(2013): 1–9.

Foster, Jacob G., Andrey Rzhetsky, and James A. Evans. “Traditionand Innovation in Scientists’ Research Strategies.” American Soci-ological Review 80, no. 5 (2015): 875–908.

Fowler, Michael R., and Julie M. Bunck. Law, Power, and the SovereignState: The Evolution and Application of the Concept of Sovereignty.Penn State University Press, 1995.

Freeman, Christopher. Technology, Policy, and Economic Performance:Lessons from Japan. Pinter, 1987.

Fromer, Jeanne C. “Patentography.” New York University Law Review85 (2010): 1444–1520.

Frumkin, Jesse, and Amanda F. Myers. USPTO Moonshot Patent Data.U.S. Patent and Trademark Office, August 2016.

Fuller, Lon L. “American Legal Realism.” University of PennsylvaniaLaw Review 82 (1934): 429–462.

Furata, Yoshimasa. “International Parallel Litigation: Disposition ofDuplicative Civil Proceedings in the United States and Japan.”Pacific Rim Law & Policy 1 (1995): 1–58.

Bibliography 219

Galkina Cleary, Ekaterina, Jennifer M. Beierlein, Navleen Surjit Kha-nuja, Laura M. McNamee, and Fred D. Ledley. “Contribution ofNIH Funding to New Drug Approvals 2010–2016.” Proceedings ofthe National Academy of Sciences, 2018.

Gallini, Nancy, and Suzanne Scotchmer. “Intellectual Property: WhenIs It the Best Incentive System?” In Innovation Policy and the Eco-nomy, Volume 2, 51–78. NBER Chapters. National Bureau of Eco-nomic Research, 2002.

Gambardella, Alfonso, Dietmar Harhoff, and Bart Verspagen. “TheValue of European Patents.” European Management Review 5 (2008):69–84.

Garner, Harold R., Lauren J. McIver, and Michael B. Waitzkin. “SameWork, Twice the Money?” Nature 493 (2013): 599–601.

Gawel, Claus R. “Patent Protection as a Key Driver for Pharmaceuti-cal Innovation.” Pharmaceuticals Policy and Law, 18, nos. 1-4 (2006):45–53.

Genet, Corine, Khalid Errabi, and Caroline Gauthier. “Which Modelof Technology Transfer for Nanotechnology? A Comparison withBiotech and Microelectronics.” Nanotechnology: Introducing theFuture, Technovation 32, no. 3 (2012): 205–215.

Genetic Inventions, Intellectual Property Rights and Licensing Practices.OECD, 2002.

George, James P. “International Parallel Litigation.” Texas InternationalLaw Journal 37 (2002): 499–540.

Gertner, Robert H. “Asymmetric Information, Uncertainty, and Se-lection Bias in Litigation.” University of Chicago Law School Round-table 1, no. 1 (1993).

Gibbons, Michael, and Ron Johnston. “The Roles of Science in Techno-logical Innovation.” Research Policy 3, no. 3 (1974): 220–242.

Gillum, Leslie, Christopher Gouveia, Ray Dorsey, Mark Pletcher, Co-lin Mathers, Charles McCulluoch, and Claiborne Johnston. “NIHDisease Funding Levels and Burden of Disease.” PLoS ONE 6

(2012).

220 Bibliography

Ginther, Donna K., Walter T. Schaffer, Joshua Schnell, Beth Masimore,Faye Liu, Laurel L. Haak, and Raynard Kington. “Race, Ethnicity,and NIH Research Awards.” Science 333 (2011): 1015–1019.

Godin, Benoît. “The Linear Model of Innovation: The Historical Con-struction of an Analytical Framework.” Science, Technology, & Hu-man Values 31, no. 6 (2006): 639–667.

Gordon, David, Wendy Taddei-Peters, Alice Mascette, Melissa Ant-man, Peter G. Kaufmann, and Michael S. Lauer. “Publication ofTrials Funded by the National Heart, Lung, and Blood Institute.”The New England Journal of Medicine 369, no. 20 (2013): 1926–1934.

Gordon, Jennifer. “The Impact of Myriad and Mayo: Will Advance-ments in the Biological Sciences Be Spurred or Disincentivized?(Or Was Biotech Patenting Not Complicated Enough?)” Cold SpringHarbor Perspectives in Medicine 5, no. 5 (2015): a020917.

Gorski, David. “The NIH Funding Process: Conformity and Medio-crity?” Science-Based Medicine, December 2012.

Graff, Gregory D., Susan E. Cullen, Kent Bradford, David Zilberman,and Alan B. Bennett. “The Public–Private Structure of Intellec-tual Property Ownership in Agricultural Biotechnology.” NatureBiotechnology 21 (2003): 989–995.

Graham, Stuart J., and Matthew Higgins. “The Impact of Patentingon New Product Introductions in the Pharmaceutical Industry.”MPRA Paper, no. 4574 (2007).

Graham, Stuart J., and Nicolas Van Zeebroeck. “Comparing PatentLitigation Across Europe: A First Look.” Stanford Technology LawReview 17 (2014): 655–708.

Greenbaum, Dov, and Christopher Scott. “Hochschullehrerprivileg -A Modern Incarnation of the Professor’s Privilege to PromoteUniversity to Industry Technology Transfer.” Science, Technologyand Society 15, no. 1 (2010): 55–76.

Greenberg, Brad A. “Rethinking Technology Neutrality.” MinnesotaLaw Review 100 (2016): 1495–1562.

Bibliography 221

Grens, Kerry. “An Economic Gamble: What Does Society Get for theBillions it Spends on Science?” The Scientist, 2007.

Griliches, Zvi. “Patents: Recent Trends and Puzzles.” NBER WorkingPaper, no. 2922 (1989).

. “The Search for R&D Spillovers.” Scandinavian Journal of Eco-nomics 94 (1992): 29–47.

Grindley, Peter C., and David J. Teece. “Managing Intellectual Capi-tal: Licensing and Cross-Licensing in Semiconductors and Elec-tronics.” California Management Review 39 (1997): 8–41.

Gross, Cary, Gerard F. Anderson, and Neil R. Powe. “The Relationbetween Funding by the National Institutes of Health and theBurden of Disease.” The New England Journal of Medicine 340, no.24 (1999): 1881–1887.

Guellec, Dominique, and Bruno van Pottelsberghe. The Economics ofthe European Patent System: IP Policy for Innovation and Competition.Oxford Scholarship Online, 2007.

Hagedoorn, John. “Sharing Intellectual Property Rights — An Explo-ratory Study of Joint Patenting Amongst Companies.” Industrialand Corporate Change 12, no. 5 (2003): 1035–1050.

Hall, Bronwyn H. “The Private and Social Returns to Research andDevelopment.” In Technology, R&D, and the Economy, edited byBruce L. Smith and Claude E. Barfield, 140–183. Brookings Insti-tution, 1996.

. “The Financing of Research and Development.” Oxford Reviewof Economic Policy 18, no. 1 (2002): 35–51.

. “Open Innovation & Intellectual Property Rights: The Two-Edged Sword.” Japan Spotlight Jan/Feb (2010): 18–19.

. “Does Patent Protection Help or Hinder Technology Trans-fer?” Chap. 2 in Intellectual Property for Economic Development, edi-ted by Sanghoon Ahn, Bronwyn H. Hall, and Keun Lee, 11–32.Chapters. Edward Elgar Publishing, 2014.

222 Bibliography

Hall, Bronwyn H., Christian Helmers, and Georg von Graevenitz.“Technology Entry in the Presence of Patent Thickets.” Institutefor Fiscal Studies Working Papers, no. W16/02 (2016).

Hall, Bronwyn H., Christian Helmers, Georg von Graevenitz, andChiara Rosazza-Bondibene. A Study of Patent Thickets. IntellectualProperty Office, 2013.

Hall, Bronwyn H., Adam Jaffe, and Manuel Tratjenberg. “MarketValue and Patent Citations.” The RAND Journal of Economics 36

(2005): 16–38.

Hall, Bronwyn H., Grid Thoma, and Salvatore Torrisi. “The Mar-ket Value of Patents and R&D: Evidence from European Firms.”NBER Working Paper 13426 (2007).

Hall, Bronwyn H., and Rosemarie H. Ziedonis. “The Patent ParadoxRevisited: An Empirical Study of Patenting in the U.S. Semicon-ductor Industry, 1979-1995.” The RAND Journal of Economics 32,no. 1 (2001): 101–28.

Halluin, Albert, and P. Westin Lorelei. “Nanotechnology: The Im-portance of Intellectual Property Rights in an Emerging Techno-logy.” Journal of the Patent and Trademark Office Society 86 (2004):220–236.

Hand, Eric, Beth Mole, Lauren Morello, Jeff Tollefson, Meredith Wad-man, and Alexandra Witze. “A Back Seat for Basic Science.” Na-ture 496, no. 7445 (2013): 277–279.

Harary, Frank, Robert Z. Norman, and Dorwin Cartwright. StructuralModels: An Introduction to the Theory of Directed Graphs. John Wiley& Sons, 1965.

Harhoff, Dietmar. “Economic Cost-Benefit Analysis of a Unified andIntegrated European Patent Litigation System.” Final Report, Ten-der No. Markt/2008/06/D, 2009.

Harhoff, Dietmar, Francis M. Scherer, and Katrin Vogel. “Citation Fre-quency and the Value of Patented Innovation.” Review of Econo-mics and Statistics 81 (1999): 511–515.

Bibliography 223

Harhoff, Dietmar, Frederic M. Scherer, and Katrin Vopel. “Citations,Family Size, Opposition and the Value of Patent Rights.” ResearchPolicy 32, no. 8 (2003): 1343–1363.

Harhoff, Dietmar, and Stefan Wanger. “The Duration of Patent Exa-mination at the European Patent Office.” Management Science 55,no. 12 (2009): 1969–1984.

Hatter, John P. “The Doctrine of Equivalents in Patent Litigation: AnAnalysis of the Epilady.” Indiana International & Comparative LawReview, 1995, 461–494.

Heck, Philip. “Was ist diejenige Begriffsjurisprudenz, die wir bekämp-fen ?” Deutsche Juristen-Zeitung, 1909, 1457–1461.

Hedge, Deepak, and Sampat N. Bhaven. “Can Private Money BuyPublic Science? Disease Group Lobbying and Federal Fundingfor Biomedical Research.” Management Science 61, no. 10 (2015):2281–2298.

Hees, Anne, and Sven-Erik Braitmayer. Verfahrensrecht in Patentsachen.Carl Heymanns Verlag, 2010.

Heinemann, Andreas. “Recht, Ökonomie und Realität.” In Law & Eco-nomics – Festschrift für Peter Nobel zum 70. Geburtstag, edited byRobert Waldburger, Peter Sester, Christoph Peter, and CharlotteM. Baer, 21–41. 2015.

Heinze, Thomas. “How to Sponsor Ground-Breaking Research: a Com-parison of Funding Schemes.” Science & Public Policy 35 (2008):302–318.

Heise, Michael. “An Empirical Analysis of Empirical Legal Scholars-hip Production, 1990-200.” University of Illinois Law Review 9 (2011):1739–1752.

Helfer, Laurence R. “Regime Shifting: The TRIPs Agreement and NewDynamics of International Intellectual Property Lawmaking.” YaleJournal of International Law 29 (2004): 1–83.

Heller, Michael A., and Rebecca S. Eisenberg. “Can Patents DeterInnovation? The Anticommons in Biomedical Research.” Science280, no. 5364 (1998): 698–701.

224 Bibliography

Helmers, Christian, Brian J. Love, and Luke McDonagh. “Is There aPatent Troll Problem in the U.K.?” Media and Entertainment LawJournal 24 (2014): 509–553.

Helmers, Christian, and Luke McDonagh. “Patent Litigation in theU.K.: An Empirical Survey 2000–2008.” Journal of Intellectual Pro-perty & Practice 8 (2013): 846–861.

Hemel, Daniel J., and Lisa L. Ouellette. “Beyond the Patents-PrizesDebate.” Texas Law Review 92, no. 2 (2006): 303–382.

Herper, Matthew. “Illumina Promises To Sequence Human GenomeFor $100 – But Not Quite Yet.” Forbes, January 2017.

Hilty, Reto, Thomas Jaeger, Matthias Lamping, and Hanns Ullrich.The Unitary Patent Package: Twelve Reasons for Concern. Max PlanckInstitute for Intellectual Property & Competition Law ResearchPaper No. 12-12, 2012.

Höcherl, Peter. “Double Protection and Forum Shopping under Ger-many’s Draft UPC Legislation.” BristowsUPC, 2017.

Holliday, Louise. “Patenting Antibodies in Europe.” MAbs 1, no. 4

(2009): 385–386.

Holman, Christopher M. “Trends in Human Gene Patent Litigation.”Science 322, no. 5899 (2008): 198–199.

. “Do Biotech Patent Lawsuits Really "Overwhelmingly Lose?":A Response to Our Divided Patent System.” Biotechnology LawReport 34 (2015): 59–67.

Holmes, Oliver W. The Common Law. Barns & Noble, 2004.

Hopenhayn, Hugo, Gerard Llobet, and Matthew Mitchell. “Rewar-ding Sequential Innovators: Prizes, Patents, and Buyouts.” Jour-nal of Political Economy 114, no. 6 (2006): 1041–1068.

Hylton, Keith N. “Asymmetric Information and the Selection of Dis-putes for Litigation.” The Journal of Legal Studies 22, no. 1 (1993):187–210.

Impact of the Unitary Patent Protection and the Unified Patent Court in theNordic-Baltic Region. Ministry of Justice of Sweden, October 2014.

Bibliography 225

Innovation’s Golden Goose. The Economist, December 2002.

Insel, Thomas. “Join the Disruptors of Health Science.” Nature 551

(2017): 23–26.

Jackson, John H. “Sovereignty-Modern: A New Approach to an Out-dated Concept.” American Journal of International Law 97, no. 4

(2003): 782–802.

Jaffe, Adam B. “Real Effects of Academic Research.” American Econo-mic Review 79, no. 5 (1989): 957–70.

. “Building Program Evaluation Into the Design of Public Rese-arch Support Programs.” Oxford Review of Economic Policy 18, no.1 (2002): 22–34.

Jaffe, Adam B., and Gaétan de Rassenfosse. “Patent Citation Data inSocial Science Research: Overview and Best Practices.” Journalof the Association for Information Science and Technology 68, no. 6

(2017): 1360–1374.

Jaffe, Adam B., and Josh Lerner. Innovation and Its Discontents: HowOur Broken Patent System Is Endangering Innovation and Progress,and What to Do About It. Princeton University Press, 2004.

Jankowski, John E., Albert N. Link, and Nicholas S. Vonortas. StrategicResearch Partnerships: Proceedings from an National Science Founda-tion Workshop. National Science Foundation, July 2001.

Johnson, Alan, and Luke Maunder. Challenging the UPC Opt-Out –How Exactly Will It Work? Bristows, June 2017.

Johnson, Ann. “The End of Pure Science? Science Policy from Bayh-Dole to the NNI.” In Discovering the Nanoscale, edited by Davis.Baird, Alfred Nordmann, and Joachim Schummer, 20–33. IOSPress, 2003.

Kahin, Brian. “Through the Lens of Intangibles: What Patents on Soft-ware and Services Reveal about the System.” In Patents, Innova-tion and Economic Performance, OECD Conference Proceedings, edi-ted by Adam B. Jaffe, Josh Lerner, and Scott Stern. OECD Publis-hing, 2004.

226 Bibliography

Kaiser, Jocelyn. “New NCI Director Expects Big Data to RevolutionizeCancer Research Care.” Science, December 2017.

Kalutkiewicz, Michael J., and Richard L. Ehman. “Patents as Proxies:NIH Hubs of Innovation.” Nature Biotechnology 32 (2014): 536–537.

Kalven, Harry Jr., and Hans Zeisel. The American Jury. Little, Brown /Company, 1966.

Kapczynski, Amy, and Talha Syed. “The Continuum of Excludabilityand the Limits of Patents.” Yale Law Journal 122 (2013): 1900–1963.

Katz, Michael L. “An Analysis of Cooperative Research and Develop-ment.” The RAND Journal of Economics 17, no. 4 (1986): 527–543.

Kerkmeester, Heico. “Methodology: General.” In Encyclopedia of Lawand Economics. Volume I. The History and Methodology of Law andEconomics, 383–401. Edward Elgar, 2000.

Kers, Jannigje G., Elco Van Burg, Tom Stoop, and Martina C. Cornel.“Trends in Genetic Patent Applications: The Commercializationof Academic Intellectual Property.” European Journal of HumanGenetics 22 (2014): 1155–1159.

Kessler, Daniel, Thomas Meites, and Geoffrey P. Miller. “ExplainingDeviations from the Fifty-Percent Rule: A Multimodal Approachto the Selection of Cases for Litigation.” The Journal of Legal Stu-dies 25, no. 1 (1996): 233–59.

Keukenschrijver, Alfred. Patentnichtigkeitsverfahren. Heymann, 2011.

Kingsbury, Anna. “Patent Collaboration: Licensing, Patent Pools, Pa-tent Commons, Open Source and Communities of Innovation.”New Zealand Intellectual Property Journal, September 2013, 3–9.

Kitch, Edmund W. “The Nature and Function of the Patent System.”Journal of Law and Economics 20, no. 2 (1977): 265–290.

Kleinman, Daniel L. “Untangling Context: Understanding a Univer-sity Laboratory in the Commercial World.” Science, Technology, &Human Values 23, no. 3 (1998): 285–314.

Bibliography 227

Klerman, Daniel, and Alex Lee Yoon-Ho. “The Priest-Klein Hypot-heses: Proofs, Generality and Extensions.” Legal Studies ResearchPaper Series, nos. 14-43 (2014).

Klevorick, Alvin K., Richard C. Levin, Richard Nelson, and SidneyWinter. “On the Sources and Significance of Interindustry Diffe-rences in Technological Opportunities.” Research Policy 24, no. 2

(1995): 185–205.

Kline, Stephen J. “Innovation Is Not a Linear Process.” Research Mana-gement 28, no. 4 (1985): 36–45.

Kline, Stephen J., and Nathan Rosenberg. “An Overview of Innova-tion.” In The Positive Sum Strategy: Harnessing Technology for Eco-nomic Growth, edited by Landau Ralph and Nathan Rosenberg,275–306. National Academy of Sciences, 1986.

Knoll, Bodo, Martina Baumann, and Nadine Riedel. “The Global Ef-fects of R&D Tax Incentives: Evidence from Micro-Data.” Beiträgezur Jahrestagung des Vereins für Socialpolitik 2014: EvidenzbasierteWirtschaftspolitik - Session: Taxation II, 2014.

Koch, Norbert, and Franz Froschmaier. “The Doctrine Of Territoria-lity In Patent Law and The European Common Market.” Idea: ThePatent, Trademark, and Copyright Journal of Research and Education,1965, 343–360.

Koenig, Gloria. Patent Invalidity: A Statistical and Substantive Analysis.Clark Boardman, 1974.

Kolata, Gina. “Grant System Leads Cancer Researchers to Play ItSafe.” New York Times, July 2009.

Koppikar, Vivek, Stephen B. Maebius, and Steven Rutt. “Current Trendsin Nanotech Patents: A View from Inside the Patent Office Intel-lectual Property.” Nanotechnology Law & Business 1 (2004): 24–30.

Kox, Henk, Arjan Lejour, and Raymond Montizaan. “The Free Mo-vement of Services Within the EU.” CPB Document, no. 69 (2004).

228 Bibliography

Krauss, Jan, and David Kuttenkeuler. “Go German Patents! – SpecialRules Allowing Double Patenting and Parallel Enforcement forthe Unitary Patent System and National Patents in Germany.”Boehmert & Boehmert, 2017.

Kurt, Haertel. “Die Harmonisierungswirkung des Europäischen Pa-tentsrechts.” Gewerblicher Rechtsschutz und Urheberrecht, Internati-onaler Teil, 1981, 479–490.

Laddie, Hugh. “Kirin Amgen - The End of Equivalents in England?”International Review of Intellectual Property and Competition Law,2009, 3–38.

Lampe, Ryan, and Petra Moser. “Do Patent Pools Encourage Inno-vation? Evidence from 20 Industries in the 1930s.” In Standards,Patents and Innovations, edited by Timothy Simcoe, Ajay K. Agra-wal, and Stuart J. Graham. NBER Books, 2014.

Lamping, Matthias, and Hanns Ullrich. “The Impact of Brexit on Uni-tary Patent Protection and its Court.” Max Planck Institute for In-novation & Competition Research Paper, nos. 18-20 (2018).

Landes, William M. “Sequential versus Unitary Trials: An EconomicAnalysis.” The Journal of Legal Studies 22, no. 1 (1993): 99–134.

Landes, William M., and Richard A. Posner. The Economic Structure ofIntellectual Property Law. Belknap Press, 2003.

Lanjouw, Jean O., Ariel Pakes, and Jonathan Putnam. “How to CountPatents and Value Intellectual Property: The Uses of Patent Rene-wal and Application Data.” Journal of Industrial Economics 46, no.4 (1998): 405–432.

Lanjouw, Jean O., and Mark Schankerman. “Patent Quality and Re-search Productivity: Measuring Innovation with Multiple Indica-tors.” The Economic Journal 112 (2004): 441–465.

. “Protecting Intellectual Property Rights: Are Small Firms Han-dicapped?” Journal of Law and Economics 57 (2004): 45–74.

Bibliography 229

Larouche, Pierre. “Legal Emulation Between Regulatory Competitionand Comparative Law.” In National Legal Systems and Globaliza-tion: New Role, Continuing Relevance, edited by Pierre Laroucheand Péter Cserne. Springer, 2013.

Latapy, Matthieu. “Main-Memory Triangle Computations for VeryLarge (Sparse (Power-Law)) Graphs.” Theoretical Computer Science407, nos. 1-3 (2008): 458–473.

Lauer, Michael S. “Applying the Relative Citation Ratio as a Measureof Grant Productivity.” Extramural Nexus, 2016.

Lauer, Michael S., and Richard Nakamura. “Reviewing Peer Reviewat the NIH.” New England Journal of Medicine 373, no. 20 (2015):1893–1895.

Lauer, Michael S., Deepshikha Roychowdhury, Katie Patel, RachaelWalsh, and Katrina Pearson. “Marginal Returns and Levels ofResearch Grant Support Among Scientists Supported by the Na-tional Institutes of Health.” bioRxiv, 2017.

Lawrence, Stacy. “Patent Drop Reveals Pressures on Industry.” NatureBiotechnology 22, no. 8 (2011): 930–931.

Lemley, Mark A. “Rational Ignorance at the Patent Office.” Northwes-tern University Law Review 95 (2001): 1495.

. “Patenting Nanotechnology.” Stanford Law Review 58 (2005):601–630.

Lerner, Josh. “Patenting in the Shadow of Competitors.” Journal ofLaw and Economics 38, no. 2 (1995): 463–495.

. “Where Does State Street Lead? A First Look at Finance Pa-tents, 1971-2000.” Journal of Finance 57, no. 2 (2002): 901–930.

Lerner, Josh, and Colin Kegler. “Evaluating the Small Business In-novation Research Program: A Literature Review.” In The SmallBusiness Innovation Research Program: An Assessment of the Depart-ment of Defense Fast Track Initiative, edited by Charles W. Wessner.National Research Council, 2000.

Lerner, Josh, and Jean Tirole. “Efficient Patent Pools.” American Eco-nomic Review 94, no. 3 (2004): 691–711.

230 Bibliography

Levin, Richard C. “A New Look at the Patent System.” The AmericanEconomic Review 76, no. 2 (1986): 199–202.

Levin, Richard C., Alvin K. Klevorick, Richard Nelson, and SidneyWinter. “Appropriating the Returns from Industrial Research andDevelopment.” Brookings Papers on Economic Activity 18, no. 3

(1987): 783–832.

Li, Danielle, and Leila Agha. “Big Names or Big Ideas: Do Peer-Review Panels Select the Best Science Proposals?” Science 348,no. 6233 (2015): 434–438.

Li, Danielle, Pierre Azoulay, and Bhaven Sampat. “The Applied Valueof Public Investments in Biomedical Research.” Science 356 (2017):78–81.

Light, Donald W., and Rebecca Warburton. “Demythologizing theHigh Costs of Pharmaceutical Research.” Management Science 50

(2004): 804–820.

Linton, Jonathan D., and Steven T. Walsh. “Integrating Innovation andLearning Curve Theory: An Enabler for Moving Nanotechnolo-gies and Other Emerging Process Technologies into Production.”R&D Management 34, no. 5 (2004): 517–526.

Lorsch, Jon. A Shared Responsibility. NIGMS Feedback Loop Blog, Ja-nuary 2015.

Love, Brian, Christian Helmers, Fabian Gaessler, and Max Ernicke.“Patent Assertion Entites in Europe.” In Patent Assertion Entitiesand Competition Policy, edited by Daniel D. Sokol, 104–129. Cam-bridge University Press, 2017.

Luginbuehl, Stefan. European Patent Law - Towards a Uniform Interpre-tation. Edward Elgar, 2011.

Lundvall, Bengt-Ake. National Systems of Innovation: Towards a Theoryof Innovation and Interactive Learning. Pinter, 1992.

Mahne, Kevin P. “A Unitary Patent and Unified Patent Court for theEuropean Union: An Analysis of Europe’s Long Standing At-tempt to Create a Supranational Patent System.” Journal of thePatent and Trademark Office Society 94 (2012): 162–191.

Bibliography 231

Mandel, Gregory. “The Non-Obvious Problem: How the Indetermi-nate Nonobviousness Standard Produces Excessive Patent Grants.”UC Davis Law Review 42 (2008): 57–128.

Mann, Ronald J., and Marian Underweiser. “A New Look at PatentQuality: Relating Patent Prosecution to Validity.” Journal of Empi-rical Legal Studies 9, no. 1, 1–32.

Mansfield, Edwin, Mark Schwartz, and Samuel Wagner. “ImitationCosts and Patents: An Empirical Study.” Economic Journal 91, no.364 (1981): 907–918.

Marburger, John H. “Wanted: Better Benchmarks.” Science 308, no.5725 (2005): 1087–1087.

Marshall, Alfred. Principles of Economics. MacMillan, 1920.

Martin, Irwin, and Sowmya Mallela. “Funding of Cancer Research:Do Levels Match Indidence and Mortality Rates?” Therapeutic In-novation & Regulatory Science 41, no. 1 (2014): 33–35.

Mathis, Klaus. Effizienz statt Gerechtigkeit? Auf der Suche nach den phi-losophischen Grundlagen der Ökonomischen Analyse des Rechts. Dunc-ker & Humblot, 2009.

Mazzucato, Mariana. The Entrepreneurial State: Debunking Public vs.Private Sector Myths. Anthem Press, 2013.

McAleer, Michael, and Daniel Slottje. “A New Measure of Innovation:The Patent Success Ratio.” Scientometrics 63, no. 3 (2005): 421–429.

McGeary, Michael, and Michael Burstein. Sources of Cancer ResearchFunding in the United States. Report on Sources of Cancer Rese-arch Funding in the United States, June 1999.

McGinnis, John O., and Mark L. Movsesian. “The World Trade Con-stitution.” Harvard Law Review, 2000, 511–605.

Mejer, Malwina, and Bruno van Pottelsberghe de la Potterie. “Eco-nomic Incongruities in the European Patent System.” EuropeanJournal of Law and Economics 34, no. 1 (2012): 215–234.

232 Bibliography

Menell, Peter S., and Suzanne Scotchmer. “Intellectual Property Law.”In Handbook of Law and Economics, edited by Mitchell A. Polinskyand Steven Shavell, 1473–1570. Elsevier, 2007.

Merges, Robert P., and Michael Mattioli. “Measuring the Costs andBenefits of Patent Pools.” Ohio State Law Journal 78 (2017): 281–347.

Merges, Robert P., and Richard R. Nelson. “On the Complex Econo-mics of Patent Scope.” Columbia Law Review 90 (1990): 839–916.

Mervis, Jeffrey. “White House Panel Urges Agencies to Take MoreRisks.” Science 338 (2012): 1274.

Meurer, Michael J. “The Settlement of Patent Litigation.” The RANDJournal of Economics 20, no. 1 (1989): 77–91.

Miller, John, Ruben Serrato, Jose Represas-Cardenas, and Griffith Kun-dahl. The Handbook of Nanotechnology. John Wiley / Sons, 2004.

Mills, Ann E., and Patti M. Tereskerz. “Empirical Analysis of MajorStem Cell Patent Cases: The Role of Universities.” Nature Bio-technology 28 (2010): 325–328.

Mohnen, Pierre, and Cathy Hoareau. “What Type of Enterprise For-ges Close Links with Universities and Government Labs? Evi-dence from CIS 2.” Managerial and Decision Economics 24, nos. 2-3(2003): 133–145.

Mongeon, Philippe, Christine Brodeur, Catherine Beaudry, and Vin-cent Larivière. “Concentration of Research Funding Leads to De-creasing Marginal Returns.” Research Evaluation 25, no. 4 (2016):396–404.

Moore, Kimberly A. “Worthless Patents.” Berkeley Technology Law Jour-nal 20 (2005).

. “Populism and Patents.” New York Law Review 82 (2007): 69–111.

Moore, Underhill, and Charles C. Callahan. “Law and Learning The-ory: A Study in Legal Control.” Yale Law Journal 53 (1943): 1–136.

Bibliography 233

Morris, Emily. “The Irrelevance of Nanotechnology Patents.” Connec-ticut Law Review 49 (2016): 499–551.

Moser, Petra. “How Do Patent Laws Influence Innovation? Evidencefrom Nineteenth-Century World’s Fairs.” American Economic Re-view 95, no. 4 (2005): 1214–1236.

Moss, Gary, Matthew Jones, and Robert Lundie-Smith. “Just How‘Anti-Patent’ are the UK Courts?” Journal of Intellectual PropertyLaw & Practice 5 (2010): 148–157.

Moufang, Rainer. “Kapitel § 4.” In Patentgesetz mit Europäischem Paten-tübereinkommen, edited by Rainer Schulte. Carl Heymanns, 2013.

Mowery, David C. “Collaborative R&D: How Effective Is It?” Issues inScience and Technology 15, no. 1 (1998): 37–44.

Moy, Carl. “Patent Harmonization, Protectionism and Legislation.”Journal of the Patent and Trademark Office Society 74 (1992): 777–810.

Mueller, Janice M. “The Tiger Awakens: The Tumultuous Transfor-mation of India’s Patent System and the Rise of Indian Pharma-ceutical Innovation.” University of Pittsburgh Law Review 68, no. 3

(2006): 491–641.

Murashige, Kate H. “Harmonization of Patent Laws.” Houston Journalof International Law 16 (1994): 591–614.

Nanotechnology and Patents. European Patent Office, 2013.

Nanotechnology Gold Rush Yields Crowded, Entangled Patents. Lux Rese-arch and Foley & Lardner LLP, 2005.

Nanotechnology Update: Corporations Up Their Spending as Revenues forNano-Enabled Products Increase. Lux Research Inc., 2014.

Nanotechnology: A Realistic Market Assessment. BCC Research, 2014.

Nard, Craig A., and John F. Duffy. “Rethinking Patent Law’s Uni-formity Principle.” Northwestern University Law Review 101, no. 4

(2007): 1619–1676.

234 Bibliography

Nard, Craig A., and Andrew P. Morris. “Constitutionalizing Patents:From Venice to Philadelphia.” Review of Law & Economics 2, no. 2

(2006): 223–321.

Narin, Francis, and Dominic Olivastro. “Status Report: Linkage bet-ween Technology and Science.” Research Policy 21, no. 3 (1992):237–249.

National Academies of Science, Engineering, and Medicine. FosteringTransformative Research in the Geographical Sciences. 2015.

National Academies of Sciences, Engineering, and Medicine. BeyondPatents: Assessing the Value and Impact of Research Investments: Pro-ceedings of a Workshop - in Brief. Policy and Global Affairs; Govern-ment University Industry Research Roundtable, 2017.

Nelson, Amy. “Obviousness or Inventive Step as Applied to NucleicAcid Molecules: A Global Perspective.” North Caroline Journal ofLaw & Technology 6, no. 1 (2004): 1–40.

Nelson, Richard R. “The Simple Economics of Basic Scientific Rese-arch.” Journal of Political Economy 67 (1959): 297–306.

. “Institutions Supporting Technical Advance in Industry.” Ame-rican Economic Review 76, no. 2 (1986): 186–189.

. “Observations on the Post-Bayh-Dole Rise in University Pa-tenting.” Journal of Technology Transfer 26 (2001): 13–19.

Niazi, Sarfaraz K., and Justin L. Brown. Fundamentals of Modern Biop-rocessing. CRC Press, 2015.

Nicholson, Joshua M., and John P. Ioannidis. “Research Grants: Con-form and Be Funded.” Nature 492 (2012): 34–36.

Nobel, Peter. “Zur Koexistenz von Recht und Ökonomie: Ökonomis-che Analyse des Rechts als Aufforderung zum Dialog.” In DerKanton St. Gallen und seine Hochschule – Beiträge zur Eröffnung desBibliotheksbaus, edited by Rolf Dubs, Yvo Hangartner, and AlfredNydegger, 454–462. Hochschule St. Gallen für Wirtschaftsrechts-und Sozialwissenschaften, 1989.

. “Die wirtschaftliche Betrachtungsweise im Recht.” Schweize-rische Juristen-Zeitung 113 (2017): 457–469.

Bibliography 235

. “Einführung – Die Aktiengesellschaft als Kapitalgesellschaft.”In Berner Kommentar zum Aktienrecht, edited by Peter Nobel, 42–64. Stämpfli, 2017.

Nordhaus, William. Invention, Growth, and Welfare: A Theoretical Treat-ment of Technological Change. MIT Press, 1969.

OECD. “A Framework for Biotechnology Statistics,” 2005.

Ohlhausen, Maureen K. “Patent Rights in a Climate of IntellectualProperty Rights Skepticism.” Harvard Journal of Law & Technology30, no. 1 (2016): 1–49.

Ohly, Ansgar, and Rudolf Streinz. “Can the UK Stay in the UPC Sy-stem After Brexit?” Journal of Intellectual Property Law & Practice12, no. 3 (2017): 245–258.

Orsenigo, Luigi, Fabio Pammolli, and Massimo Riccaboni. “Techno-logical Change and the Dynamics of Networks of CollaborativeRelations: The Case of the Bio-Pharmaceutical Industry.” ResearchPolicy 30 (2001): 485–508.

Ouellette, Lisa L. “Nanotechnology and Innovation Policy.” HarvardJournal of Law & Technology 29 (2015): 33–75.

. “Patent Experimentalism.” Virginia Law Review 101 (2015): 65–128.

P’ng, Ivan P. “Strategic Behavior in Suit, Settlement, and Trial.” BellJournal of Economics 14, no. 2 (1983): 539–550.

Pakes, Ariel, and Zvi Griliches. “Patents and R&D at the Firm Level:A First Report.” Economics Letters 5, no. 4 (1980): 377–381.

Palombi, Luigi. Gene Cartels: Biotech Patents in the Age of Free Trade.Edward Edgar, 2009.

Paredes, Peter J. “Written Description Requirement in Nanotechno-logy: Clearing a Patent Thicket?” Journal of the Patent & TrademarkOffice Society 88 (2006): 489–512.

Parish, Christopher. “Cancer Immunotherapy: The Past, the Presentand the Future.” Immunology And Cell Biology 81 (2003): 106–113.

236 Bibliography

Park, Hyunwoo, Jeongsik Lee, and Byung-Cheol Kim. “Project Se-lection in NIH: A Natural Experiment from ARRA.” Research Po-licy 44, no. 6 (2015): 1145–1159.

Parker, Matthew. “Giving Teeth to European Patent Reform, Overco-ming Recent Legal Challenges.” Emory International Law Review26 (2012): 1079–1110.

Patent Pools and Antitrust – A Comparative Analysis. World IntellectualProperty Organization (WIPO), March 2014.

Pavlova, Natalya N. “The Emerging Hallmarks of Cancer Metabo-lism.” Cell Metabolism 23, no. 1 (2016): 27–47.

Peifer, Mark. “The Argument for Diversifying the NIH Grant Portfo-lio.” Molecular Biology of the Cell 28, no. 22 (2017): 2935–2940.

Pinckney, Richard. “Understanding the Transitional Provisions of theAgreement on the Unified Patent Court.” European IntellectualProperty Review 37, no. 5 (2015): 268–277.

Pohlhaus, Jennifer R., Hong Jiang, Robin M. Wagner, Walter T. Schaf-fer, and Vivian W. Pinn. “Sex Differences in Application, Success,and Funding Rates for NIH Extramural Programs.” Academic Me-dicine 86, no. 6 (2011): 759–767.

Pollack, Gerald. “Revitalizing Science in a Risk-Averse Culture: Re-flections in the Syndrome and Prescriptions for its Cure.” Cellularand Molecular Biology 51 (2005): 815–820.

Posner, Richard A. “Natural Monopoly and Its Regulation.” StanfordLaw Review 21, no. 3 (1969): 548–643.

. “Legal Scholarship Today.” Harvard Law Review 115, no. 5

(2002): 1314–1326.

. Economic Analysis of Law. Wolters Kluwer, 2007.

President’s Council of Advisors on Science and Technology. ReportTransformation and Opportunity: The Future of the U.S. Research En-terprise. 2012.

Price, Nicholson W. “Grants.” Berkeley Technology Law Journal forthco-ming (2019).

Bibliography 237

Price, William J., and Lawrence W. Bass. “Scientific Research and theInnovative Process.” Science 164, no. 3881 (1969): 802–806.

Priest, George L., and Benjamin Klein. “The Selection of Disputes forLitigation.” The Journal of Legal Studies 13, no. 1 (1984): 1–56.

Pumfrey, Nicholas, Martin J. Adelman, Shamnad Basheer, Raj S. Dave,Peter Meier-Beck, Yukio Nagasawa, Maximilian Rospatt, and Mar-tin Sulsky. “The Doctrine of Equivalents in Various Patent Re-gimes – Does Anybody Have it Right?” Yale Journal of Law andTechnology 11, no. 1 (2009): 262–308.

Rave, Theodore D. “Governing the Anticommons in Aggregate Liti-gation.” Vanderbilt Law Review 66 (2013): 1183–1258.

Report of the Secretary’s Advisory Committee on Genetics, Health,and Society. Gene Patents and Licensing Practices and Their Impacton Patient Access to Genetic Tests. April 2010.

Reynolds, Graham. “Nanotechnology and the Tragedy of the Anti-commons: Towards a Strict Utility Requirement.” University ofOttawa Law & Technology Journal 6 (2009): 79–114.

Rickman, Neil. “Contingent Fees and Litigation Settlement.” Interna-tional Review of Law and Economics 19, no. 3 (1999): 295–317.

Romandini, Roberto, and Alexander Klicznik. “The Territoriality Prin-ciple and Transnational Use of Patented Inventions – The WiderReach of a Unitary Patent and the Role of the CJEU.” Internati-onal Review of Intellectual Property and Competition Law 44, no. 5

(2013): 524–540.

Romer, Paul. “Endogenous Technological Change.” Journal of PoliticalEconomy 98, no. 5 (1990): 71–102.

Romet, Isabelle, Amandine Métier, and Dora Talvard. “Patent Enfor-cement in France.” In Patent Enforcement Worldwide,: Writings inHonour of Dieter Stauder, edited by Christopher Health, 145–180.BIOS Scientific Publishers, 2015.

Rosegrant, Susan, David Ralph David Lampe, and Robert Lampe.Route 128: Lessons from Boston’s High-Tech Community. Basic Books,1992.

238 Bibliography

Rosenberg, David, and Steven Shavell. “A Model in Which Suits AreBrought for Their Nuisance Value.” International Review of Lawand Economics 5, no. 1 (1985): 3–13.

Rosenberg, Nathan. Exploring the Black Box. Cambridge UniversityPress, 1994.

Rotolo, Daniele, Diana Hicks, and Ben Martin. “What Is an EmergingTechnology?” Research Policy 44, no. 10 (2015): 1–40.

Rowley, Charles K. “Public Choice and the Economic Analysis ofLaw,” 123–173. Kluwer Academic Publishers, 1989.

Royal Swedish Academy of Sciences. Background on the Nobel Prize inChemistry 2016: Molecular Machines. 2016.

Rubinfeld, Daniel L., and Suzanne Scotchmer. “Contingent Fees forAttorneys: An Economic Analysis.” The RAND Journal of Econo-mics 24, no. 3 (1993): 343–356.

Rudge, Andrew. Guide to European Patents. Thomson West, 2012.

Ruffner, Markus. “Reformbedarf des Aktienrechts – eine "Law andEconomics" – Perspektive.” Schweizerische Juristen-Zeitung 101 (2005):253–261.

Rzhetsky, Andrey, Jacob G. Foster, Ian T. Foster, and James A. Evans.“Choosing Experiments to Accelerate Collective Discovery.” Pro-ceedings of the National Academy of Sciences, 2015.

Sabellek, André. Patente auf nanotechnologische Erfindungen. Mohr Sie-beck, 2014.

Sabety, Ted. “Nanotechnology Innovation and the Patent Thicket: WhichIP Policies Promote Growth?” Nanotechnology Law & Business Vo-lume 1.3 (2004): 262–283.

Salter, Ammon J., and Ben Martin. “The Economic Benefits of PubliclyFunded Basic Research: a Critical Review.” Research Policy 30, no.3 (2001): 509–532.

Salzberg, Steven L. “NIH Funding: It Does Support Innovators.” Na-ture 493 (2013): 26.

Bibliography 239

Sampat, Bhaven N., and Frank Lichtenberg. “What are the RespectiveRoles of the Public and Private Sectors in Pharmaceutical Inno-vation?” Health Affairs 30 (2011): 332–339.

Samuelson, Paul A. “The Pure Theory of Public Expenditure.” Reviewof Economics and Statistics 36, no. 4 (1954): 387–389.

Sarnoff, Joshua D. “The Likely Mismatch Between Federal ResearchDevelopment Funding and Desired Innovation.” Vanderbilt Jour-nal of Entertainment Technology Law 18, no. 2 (2016): 363–418.

Saxenian, Annalee. Regional Advantage: Culture and Competition in Sili-con Valley and Route 128. Harvard University Press, 1996.

Schankerman, Mark. “How Valuable is Patent Protection? Estimatesby Technology Field.” The RAND Journal of Economics 29, no. 1

(1998): 77–107.

Schauz, Desiree. “What is Basic Research? Insights from HistoricalSemantics.” Minerva 52, no. 3 (2014): 273–328.

Schellekens, Maurice. “Patenting Nanotechnology in Europe: Makinga Good Start? An Analysis of Issues in Law and Regulation.”Journal of World Intellectual Property 13, no. 1 (2010): 47–76.

Scherer, Frederic M. “First Mover Advantages and Optimal PatentProtection.” The Journal of Technology Transfer 40, no. 4 (2015): 559–580.

Schwalbe, Ulrich, and Paul Walker. “Zermelo and the Early Historyof Game Theory.” Games and Economic Behavior 34, no. 1 (2001):123–137.

Schweizer, Urs. “Litigation and Settlement under Two-Sided Incom-plete Information.” Review of Economic Studies 56, no. 2 (1989):163–177.

Scotchmer, Suzanne. “On the Optimality of the Patent Renewal Sy-stem.” The RAND Journal of Economics 30, no. 2 (1999): 181–196.

. Innovation and Incentives. MIT Press, 2004.

240 Bibliography

Scotchmer, Suzanne, and Green Jerry. “Novelty and Disclosure in Pa-tent Law.” The RAND Journal of Economics 21, no. 1 (1990): 131–146.

Seawright, Jason, and John Gerring. “Case Selection Techniques inCase Study Research: A Menu of Qualitative and QuantitativeOptions.” Political Research Quarterly 61, no. 2 (2008): 294–308.

Shannon, John. “Federalism’s "Invisible Regulator" – Interjurisdictio-nal Competition.” In Competition Among States and Local Govern-ments: Efficiency and Equity in American Federalism, edited by DaphneKenyon and John Kincaid, 117–126. Urban Institute Press, 1991.

Shapiro, Carl. “Navigating the Patent Thicket: Cross Licenses, PatentPoools, and Standard Setting.” In Innovation Policy and the Eco-nomy, vol. 1. MIT Press, 2001.

. “Antitrust Analysis of Patent Settlements Between Rivals.”Antitrust Summer, 2003, 70–77.

Shimkin, Michael B. “As Memory Serves–An Informal History of theNational Cancer Institute, 1937-57.” Journal of the National CancerInstitute 59, no. 2 (1977): 559–600.

Siebert, Ralph, and Georg von Graevenitz. “How Licensing Resol-ves Hold-Up: Evidence from a Dynamic Panel Data Model withUnobserved Heterogeneity.” CEPR Discussion Papers, no. 5436

(2006).

. “Does Licensing Resolve Hold Up in the Patent Thicket?”Münchener Wirtschaftswissenschaftliche Beiträge, nos. 2008-01 (2008).

. “Are Licensing Agreements Appropriate Instruments to CutThrough the Patent Thicket?” Working Paper, 2011.

Siegel, Rebecca L., Kimberly D. Miller, and Ahmedin Jemal. “CancerStatistics, 2018.” CA: A Cancer Journal for Clinicians 68, no. 1, 7–30.

Smeets, Roger. “Does Patent Litigation Reduce Corporate R&D? AnAnalysis of US Public Firms.” Working Paper Rutger University,2014.

Bibliography 241

Smith, Bradford L., and Susan O. Man. “Innovation and IntellectualProperty Protection in the Software Industry: An Emerging Rolefor Patents.” University of Chicago Law Review 71 (2004).

Solow, Robert M. “Technical Change and the Aggregate ProductionFunction.” Review of Economics and Statistics 39, no. 3 (1957): 312–320.

Sovacool, Benjamin K. “Replacing Tedium with Transformation: Whythe US Department of Energy Needs to Change the Way it Con-ducts Long-Term R&D.” Energy Policy 36, no. 3 (2008): 923–928.

Stack, Alexander. International Patent Law: Cooperation, Harmonizationand an Institutional Analysis of WIPO and the WTO. Edward Elgar,2011.

Stallman, Richard. “Patent Law is, at Best, Not Worth Keeping.” LoyolaUniversity Chicago Law Journal 45, no. 2 (2013): 389–399.

Stewart, David J., Simon N. Whitney, and Razelle Kurzrock. “EquipoiseLost: Ethics, Costs, and the Regulation of Cancer Clinical Rese-arch.” Journal of Clinical Oncology 28, no. 17 (2010): 2925–2935.

Stewart, Paul M., Anna Stears, Jeremy W. Tomlinson, and Morris J.Brown. “Regulation - the Real Threat to Clinical Research.” TheBMJ 337 (2008).

Stiglitz, Joseph E. “Knowledge as a Global Public Good.” Global PublicGoods 19 (1999): 308–326.

Stiglitz, Joseph E., and Arjun Jayadev. “Medicine for Tomorrow: SomeAlternative Proposals to Promote Socially Beneficial Research andDevelopment in Pharmaceuticals.” Journal of Generic Medicines 7,no. 3 (2010): 217–226.

Stiles, Amber. “Hacking through the Thicket: A Proposed Patent Pool-ing Solution to the Nanotechnology Building Block Patent Thic-ket Problem.” Drexel Law Review 4 (2012): 555–592.

Straus, Joseph. “Reversal of the Burden of Proof, the Principle of “Fairand Equitable Procedures” and Preliminary Injunctions underthe TRIPS Agreement.” The Journal of World Intellectual Property 3,no. 6 (2005): 807–822.

242 Bibliography

Swan, Trevor W. “Economic Growth and Capital Accumulation.” TheEconomic Record 32, no. 2 (1956): 334–361.

Symeonidis, George. “Comparing Cournot and Bertrand Equilibriain a Differentiated Duopoly with Product R&D".” InternationalJournal of Industrial Organization 21, no. 1 (2003): 39–55.

Takenaka, Toshiko. Interpreting Patent Claims: The United States, Ger-many and Japan. Wiley-VCH, 1995.

Taniguchi, Norio. “On the Basic Concept of ‘Nano-Technology’.” Pro-ceedings of the International Conference on Production Engineering,Tokyo, 1974.

Taubman, Antony. “Shedding Light on the Life Sciences: Patent Lands-caping for Public Policymakers.” WIPO Magazine 4 (2008).

Thumm, Nikolaus. “Introduction.” In Intellectual Property Rights: Na-tional Systems and Harmonization in Europe, edited by NikolausThumm, 1–4. Springer, 2000.

Thursby, Jerry, and Marie Thursby. “Who Is Selling the Ivory Tower?Sources of Growth in University Licensing.” Management Science48, no. 1 (2002): 90–104.

Tilmann, Winfried. “The Harmonisation of Invalidity and Scope ofProtection Practice of the National Courts of EPC Member Sta-tes.” International Review of Intellectual Property and CompetitionLaw, 2006, 62–74.

Towards an Enhanced Patent Litigation System and a Community Patent –How to Take Discussions Further. The European Council, 2007.

Toynhee, Arnold J. A Study of History. Vol. I. Oxford University Press,1962.

Trajtenberg, Manuel. “A Penny for Your Quotes: Patent Citations andthe Value of Innovations.” The RAND Journal of Economics 21, no.1 (1990): 172–187.

Travis, Garett D., and Harold M. Collins. “New Light on Old Boys:Cognitive and Institutional Particularism in the Peer Review Sy-stem.” Science, Technology, & Human Values 16, no. 3 (1991): 322–341.

Bibliography 243

Triet, Gregoire, and Luc Santarelli. Propriété industrielle : le coût deslitiges. Etude comparée entre la France, l’Allemagne, l’Angleterre, lesEtats-Unis, l’Espagne et les Pays-Bas. Ministère de l’Economie, desFinances et de l’Industrie, May 2000.

Tushnet, Mark, Peter Cane, John Baldwin, and Gwynn Davis. “Empi-rical Research in Law.” In Oxford Handbook of Legal Studies, 173–203. Oxford University Press, 2005.

Ullrich, Hanns. “TRIPS: Adequate Protection, Inadequate Trade, Ade-quate Competition Policy.” Pacific Rim Law & Policy Association 4,no. 1 (1995): 153–210.

Un, Annique C., and Kazuhiro Asakawa. “Types of R&D Collabora-tions and Process Innovation: The Benefit of Collaborating Up-stream in the Knowledge Chain.” Journal of Product InnovationManagement 32 (2015): 138–153.

US Government Accounting Office. Report to Congressional Committees,Technology Transfer, Administration of the Bayh-Dole Act by ResearchUniversities. 1978.

Valdivia, Walter D. University Start-Ups: Critical for Improving Techno-logy Transfer. Center for Technology Innovations at Brookings,2013.

Van Overwalle, Geertrui. “Policy Levers Tailoring Patent Law to Bi-otechnology: Comparing U.S. and European Approaches.” UCIrvine Law Review, 2011, 435–517.

Van Aaken, Anne. "Rational Choice" in der Rechtswissenschaft: zum Stel-lenwert der ökonomischen Theorie im Recht. Nomos, 2003.

Varmus, Harold. The Art and Politics of Science. W. W. Norton & Com-pany, 2009.

Venkatraman, Subbu. “Has Nanomedicine Lived up to its Promise?”Nanotechnology 25, no. 37 (2014): 1–4.

Véron, Pierre. “Le Contentieux des Brevets d’Invention en France,Étude Statistique 2000-2010,” 2010.

Vogel, Gretchen. “Stem Cells Named Breakthrough of the Year.” Science,December 1999.

244 Bibliography

Von Graevenitz, Georg, Stefan Wagner, and Dietmar Harhoff. “Howto Measure Patent Thickets – A Novel Approach.” Economics Let-ters 111, no. 1 (2011): 6–9.

. “Incidence and Growth of Patent Thickets: The Impact ofTechnological Opportunities and Complexity.” The Journal of In-dustrial Economics 61, no. 3 (2013): 521–563.

Wahls, Wayne P. “High Cost of Bias: Diminishing Marginal Returnson NIH Grant Funding to Institutions.” bioRxiv, 2018.

Walter, Peter. “Die Auslegung staatsvertraglichen und harmonisiertenRechts: Gewicht und Bedeutung von Entscheidungen ausländis-cher Gerichte und der Beschwerdekammern des EPA.” Gewerbli-cher Rechtsschutz und Urheberrecht, 1998, 866–870.

Wang, Shyh-Jen. “The Stem Cell Patent Landscape as Relevant to Can-cer Vaccines.” Human Vaccines 7, no. 10 (2011): 1100–1108.

Warlow, Charles. “Over-Regulation of Clinical Research: a Threat toPublic Health.” Clinical Medicine 5, no. 2 (2005): 33–38.

Watts, Alison. “Bargaining through an Expert Attorney.” Journal ofLaw, Economics, & Organization 10, no. 1 (1994): 168–186.

Webbink, Mark H. “A New Paradigm for Intellectual Property Rightsin Software.” Duke Law & Technology Review 4 (2005).

Whalen, Ryan. “Citation Distance: Measuring Knowledge Translation,Integration, Diffusion, and Scope.” Proceedings of the Annual Con-ference of CAIS, 2016.

Whytock, Christopher A. “The Evolving Forum Shopping System.”Cornell Law Review 96, no. 3 (2011): 481–534.

Williams, Heidi L. “Intellectual Property Rights and Innovation: Evi-dence from Health Care Markets.” Innovation Policy and the Eco-nomy 16 (2016): 53–87.

Winter, Gerd. “Patent Law Policy in Biotechnology.” Journal of Envi-ronmental Law 4, no. 2 (1992).

Workshops on the Unitary Patent and the Unified Patent Court. EPO’sEconomic and Scientific Advisory Board, 2013.

Bibliography 245

World Intellectual Property Report, Breakthrough Innovation and EconomicGrowth. World Intellectual Property Organization, 2015.

Wright, Brian. “The Economics of Invention Incentives: Patents, Pri-zes, and Research Contracts.” American Economic Review 73, no. 4

(1983): 691–707.

Xenos, Dimitris. “The European Unified Patent Court: Assessmentand Implications of the Federalisation of the Patent System inEurope.” SCRIPTed: A Journal of Law, Technology and Society 10, no.2 (2013): 246–277.

Xiang, Li, Dai Lingli, Ouyang Shiwen, and Zhu Ning. “A Compara-tive Analysis of the Inventive Step Standard in the EPO, SIPOand USPTO.” Journal of Intellectual Property Law & Practice 8, no.7 (2013): 539–545.

Yarsky, Joseph. “Hastening Harmonization in European Union PatentLaw Through a Preliminary Reference Power.” Boston College In-ternational and Comparative Law Review 40 (2017): 167–193.

Yu, Peter K. “Five Disharmonizing Trends in the International Intel-lectual Property Regime.” In Intellectual Property and InformationWealth: Issues and Practices in the Digital Age, edited by Peter K.Yu, 73–112. Cambridge University Press, 2004.

Zerhouni, Elias A. “Translational and Clinical Science — Time for aNew Vision.” New England Journal of Medicine 353, no. 15 (2005):1621–1623.

Ziedonis, Rosemarie H. “Don’t Fence Me In: Fragmented Marketsfor Technology and the Patent Acquisition Strategies of Firms.”Management Science 50, no. 6 (2004): 804–820.

Zúñiga-Vicente, José Ángel, César Alonso-Borrego, Francisco J. Forca-dell, and José I. Galán. “Assessing the Effect of Public Subsidieson Firm R&D Investment: A Survey.” Journal of Economic Surveys28, no. 1 (2014): 36–67.

C U R R I C U L U M V I TA E R A P H A E L Z I N G G

Address: 123 Carl Street Born May 23, 1990

San Francisco, CA-94117 Swiss Nationality

ACADEMIC POSITION

Waseda University, Institute for Advanced Study Toyko

Assistant Professor since 2018

Singapore Management University, School of Law Singapore

Visiting Professor April 2019

University of Hong Kong, Law and Technology Centre Hong Kong

Visiting Scholar March 2019

University of California, Berkeley, School of Law Berkeley

Visiting Scholar 2017-2018

Max Planck Institute for Innovation and Competition Munich

Visiting Scholar 2016-2017

EDUCATION

ETH Zurich, Center for Law & Economics Zurich

(Dr. sc. ETH Zurich) Defense 22/03/2019

Fostering Innovation in Emerging Technologies

Committee: Stefan Bechtold, Stuart Graham, David Schwartz

University of California, Berkeley Extension Berkeley

(Advanced Biosciences Program) 2017

ETH Zurich, Dept. Chemistry and Applied Biosciences Zurich

(Studies in Pharmaceutical Sciences) 2015-2016

University of Michigan, Institute for Social Research Michigan

(Program in Quantitative Methods) Summer 2016

University of Zurich Zurich

(Master of Law, magna cum laude) 2013-2014

Universite Pantheon-Assas, Paris II Paris

(Licence en Droit prive, avec mention) 2009-2012

University of Fribourg Fribourg

(Bachelor of Law, magna cum laude) 2009-2012