patents innovation and economic performance oecd conference pptx

Intdusion PATENTS AND ECONOMIC PERFORMANCE: ESTABLISHING THE LINKS {Intemational Pateming and the European Patent Office ‘A Quantitative Assessment naan Eaton, Semuel Kort ane Josh Lemer

Patents, Innovation and Economic Performance

OECD CONFERENCE PROCEEDINGS

Patents, Innovation

and Economic Performance

orc (@

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

ORGANISATION FOR ECONOMIC CO-OPERATION

AND DEVELOPMENT Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came Into fore on 3th September 161, the Organisation for Economic Co-operation and Development (OECD) shall promote policies designed:

~ to achieve the highest sustainable economic growth and employment and a rising standard of living in member counties, while maintaining financial stability, and thus to contribute to the development ofthe world economy;

~ to contribute to sound economic expansion in member as well as non-member countries in the

‘process of economic development; and

= to contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance with international obligations

‘The original member countries of the OECD are Austria, Belgium, Cansda, Denmark, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain,

‘Sweden, Switzerland, Turkey, the United Kingdom and the United States The following countries became members subsequently through accession at the dates indicated hereafter: Japan {28th April 1964), Finland (28th January 1968), Australia (7th june 1973), New Zealand (25th May 1973),

‘Mexico (18th May 1954), the C2ech Republic (21st December 1985), Hungary (?Un May 1986), Poland {@and November 1996), Korea (12th December 1996) and the Slovak Republic (14th December 2000) The (Commission of the European Communities takes part in the work of the OECD (Article 13 of the OECD Convention),

sas» pn a cam ops Geom se a El nh te Ce

toi od hou Cpe Car

CCSTP work to date has tacked various aspects of intellectual property rights and patent policy notably bioteehvology and public research organisations Ic has resulted in several conferences and publications: Generic Inventions, IPR and Licensing Practices (OECD, 2002) and Turning Science fino Business: Patening and Licensing at Public Research Organisations (OECD, 2003), The August

2008 conference was also part a this work and addressed the broader ise ofthe relationship of the conomy and innovation with the patent system Although there is ample evidence of the greater importance of patents in pater of ianovation and economic grow, and although the patent sem has experienced in-depth changes over the past years to date there has bee litle systematic evalu tion to beter inform policy eoiees, Kis the aim of the CSTP peojoct to contribute to such evaluation

‘This publication is organised in five parts, which broadly correspond to the thematie sessions of the conference:

‘+ Patents and economie performance: establishing the links

‘+ Changes in patent regimes

‘+ Patents, enueprencusship and diffusion of technology

4+ IPR for software and services,

+ Camentanlliu policy challenges

These proceedings were prepared by Catalina Martinez, Jerry Sheehan and Dominique Guelec, with the editorial assistance of Joseph Lous They are published on the responsibility of the Seretary~ (General ofthe OECD

Intdusion

PATENTS AND ECONOMIC PERFORMANCE:

ESTABLISHING THE LINKS

{Intemational Pateming and the European Patent Office

‘A Quantitative Assessment naan Eaton, Semuel Kort ane Josh Lemer

Japan's Patent System and Business Innovation

Reassessing Pro-patent Policies

Kacayuki Mowohashi

Sop and Natur of the Patent Surge: A View from Germany

Knut Blind, Job Edler, Rainer Enesch and Ulrich Sclanoch

Understanding Business Patnting and Licensing: Resulis of a Survey

Jerry Sheehan, Cotaling Martine: ad Domingue Guellec

proving Patent Quality: Connecting Bzonomie Research an Policy Siepho A Merril

How Do Thied-Pary Patents Foster Innovation?

Thierry Sueur

(CHANGES IN PATENT REGIMES

Overview of Recent Changes and Comparison of Patent Regimes in the

United! Stats, Japan and Europe

(Catalina Marine: and Dominique Guellec

Recent Changes and Expected! Developments in Patent Regimes

‘A European Perspective ch Sa:

PATENTS, ENTREPRENEURSHIP AND

‘TECHNOLOGY DIFFUSION

[Enabling Imovation, Encouraging Invest:

Inlleetusl Proper’ and Genomic Enepreneusip Thane Kreiner

‘Technology Licensing

iets $ Vonortas an Young Kim

Comments on Patents, Enteprencurship and Technology Dilasion

In

Hồ

113

us ist

INTELLECTUAL PROPERTY RIGHTS

FOR SOFTWARE AND SERVICES,

“Through the Lens of Intangibles

‘What Patents on Sofware asl Services Reveal about the System)

Briar Kain

ram Innovation Surveys?

Jacques Mairesse and Pierre Molen

“The Software Patent Experiment

James Bessen and Robert M Hunt

‘The Use of Patents, Copyrights and Trademars in Software

Evidence from Litigation

‘Stuart Graham and Deepak Somaya

Why are Software Patents So Trivial?

‘The Pacem System: Curent and Future Poly Challenges

‘A View from the European Patent Office Manuel Dezantes

‘The Patent System: Current andl Future Policy Challenges

‘A View from the United States Patent and Trademark Office Lois E Roland

309 ais 3a

333 Conference Agenda

Lis of Panieipants

EXECUTIVE SUMMARY Patented inventions are inreasingly present throughout the economy and ther inluence on inno vation and economic pedformance i pervasive, Over the past vo decades, the mumber of patent applications fled cach yeae in major patent offices has grown ala rapid pace, espocally in new areas such as information and communications technologies (ICT) and biotechnology Increased inve ress and growing investment in esearch explain pat ofthe growth in patenting, hut changes in patent fegimnes that have expanded the realm of patent protection and sttengthened rights of patent oer, together sith a more staegie bohaviour of patents, have also played a role Ensuring thatthe patent system continues to serve its dual role of providing incentives to invention and facilitating diffusion of Will require increased vigilance by poliey makers and robust

‘empirical evidence for measuring the eects of patents on innovation and economic performance

Ministers in charge of science and technology policy from all OECD countries concluded atthe January 2004 meeting ofthe OECD Comite for Scientific and Technological Policy at Ministerial level that “patent regimes play am increasingly complex role in encouraging innovation, ding Seionific and technical knowledge, and enhancing market entry and fiom ereation AS suck, they

‘shoul be subject to closer seratiny by science, technology and innovation poliey makers."

'D Conference on Intcllecual Property Rights (IPR), Innovation and Eeonomic Per held in Pars on 28-29 August 2003, foreshadowed this need by providing policy makers with Factual evidence and analysis that shed ight onthe poley debate about patents and by sting out implications forthe development of IPR regimes that contribute more efficiently to innovation and economic performance, Organised al the inttive of the OECD Directorate of Science, Technology and Industry, as par of a broader project on IPR, the conference was designed as a forum for

‘iscusion among researchers, sakeholers and policy makers A number of poiey.riewed empifcal studies undertaken by economists and Tegal expen, in most par especially prepared for the project,

‘were presented and discussed at the conference Results and conclusions from those studies Were tested agtins the views of policy makers and practitioners from the business community and patent ices Discussions were organised around a numer of thems, including the lnk between patents and economic performance, recent changes in patent regimes, the impact of patents on entrepreneur

‘hip and technology diffusion, andthe protection of intellectual propery in software an seie+,

‘The presentations and discussions fed to the conclusion that the stengthening and extension of the patent system correspon to broaler changes in the economy, notably the transition to inewasingly slobal knowledge-based economies characterised by growing innovation and heightened dependence

fn intellectual assets a8 Key source of economic value and competitive advantage, Broad general

‘ations about the effects of patenting on innovation and economic performance ate difficult make, a8 the elfetivenes of patents seems to vary considerably by industry sector and technological Fld For example, whereas most panicipans agreed that patents provide incentives to innovate in the phar-

‘maceutialssetor, opinions were divided as epands software Nevertheless, discussions reflected that the expansion of patent poteetion as certainly influenced industrial structure by, fr example, Ti

tating the breakdown of vercally integrate industries e.g, semiconductors and pharmaccutieals) and

«eating opportunities for sal firms that by virtue of tei intellectual property ean atact capital and imtegrte themselves into slobal value chins (eg biotechnology) AL the same time, participants iden- tified several areas for which aditional attention is needed in onlr o ensure that patents continae to both prtcct inventions and encourage disclosure: 1) enhancing the dtfusion of patented technology: 2)ensuringthotoueh examination and high-quality of fsued patents; and 3) improving imernational syondinalin,

Participants mised concems shou the possible effets of patenting on difasion of knowledge and

fo access to patented knowledge for follow-on research, especialy in new technological Fields Two tain afcas of inlerest were identified regarding policy directions to improve the diffusion of know= Tedge and follow-on innovation

‘+ Exemptions for research use of patented inventions, Paricipans indicated a need for beter ‘monitoring the evolution of exceptions for research use of patented inventions Research

‘exemptions allow research institutions, sveh as universities, to use freely patented research resuls from other institutions when the purpose is non-commercial Recent out desisions in the United States have nartowed the seope of application of the exception and the definition and status of esearch exempons in other regions is heterogencous and sometimes uncertain

‘+ Markets for tecinology, notably patent icensing agreements, play an ineressingly important role in the economy, especially 38 innovation becomes mote cooperative, They conte (0 the diffusion of technology in an eru characterised by greater patent protection and favour the

‘reation of seience-based SMES On the bass of improved understanding of such markets, jovernments might consider poliey measures to remove obstacles wo their development

Key to an effective patent system is ensuring the quality of patens Low-quaity patents include those that are overly broad, of for which inventiveness fs weak, Such patents combate to congestion

in the patent system and give patent holdets more protection than might be warranted, reducing the nel benefit of patents to society However ensuring high-quality standards for patents can be costly Paricpant identified two ateas in nc of atetion,

1+ Now areas of patent protection, notably biotechnology software and business methods, have ‘ised new ists tha the patent system have had ficulties in addressing There is need foe

‘developing the capability to rapidly buildup expertise in nes areas and leạm how best apply basic patenting principles and ensute the granting of high quality patents, More pot

‘oriented studies based on empirical evidence have to be undertaken to face future challenges ts further new areas emerge

‘© Patent administration Concerns regarding the quality of patents are not Kimited to new areas, ‘Growing workloads a patent offices make it more dificult to maintain the quality of al issued Patents Participants noted that post-grant measures, such as opposition systems, can help off Set such problems They also identified a need fora better assessment ofthe issue of patent

‘quality including definitions and measurement, and for improving existing solutions,

International isues were also high onthe conference pariipants’ agenda The question of patent

‘uninistation and enforcement in developing countries wis hotly debated, Even among OECD eoun- tres, business representatives highlighted the challenges of protecting inventions atoss multiple juris ditions, and patent officials identified «nee for greater co-operation,

‘+ Developing counries are curently strengthening their patent systems, mainly under pressure from developed countries but also with the view of encouraging indigenous inventions The nel of development of « county, notably its innovative capability, determines its ability to tse efficiemly a patent system As a result it might not bein the interest ofall developing

‘counties to adopt patent systems as strong as in developed counties in all aspects, More

‘economic analysis distinguishing between the poorest and middle income among the less

‘develope countries is need in this domain

‘+ Reinforced imernavional co-operation among patent authorities was seen as a privity by ‘many panicipants, especially as relales to patenability itera and prior-am sears This

‘could not only reduce the administrative burden on pateat offices, hut provide a more

‘consistent IPR fame ork for Firms and other inventors that exploit thet IP globally

These proceedings summarise the presentations and discussions held at the conference and include a compilation of writen consibutions prepared by a numberof pariipants The publication is

‘organised into five parts that roughly follow the conference siete, The fst part explores the links between patenting, imovation an economic perfomance, The secord describes recent changes in Patent regimes The third analyses the impact of patents on entrepreneurship and iffusion of tech ology The fourth part looks a the protection of imtelleetual property in sofware and services andthe Impact of patents on diffusion of Knowledge in this area The last part concludes with the views of patent officials, policy makers and experts on current and Tature challenges for patent policy including fssues related 1 adaping patent systems to developing countries,

INTRODUCTION

“The number of patent applications submited to major patent offices has grown rapidly inthe past

to decades Patented inventions are inereasngly present throughout the economy an thee influence

‘on innovation and economic performance i becoming more important and pervasive, The growth in patenting refleets increased investments in research, development and innovation, more stnlegie haviour of patentees and changes in pateat regimes that have led 19 an expanding realm of patent roection and stronger rights for patent holers Patents provide incentives innovate and contibute

to technology dfusion, ia panicular by equing dislosure ofthe funetioning of anew invention and facilitating exchanges via licensing agreements Nevertheless, 2 number of questions must bee assessed regarding the determinants of patenting and the impact of patenting on innovation and eco fomie performance In particular, as innovation processes have changed and new types of inventions have been introduced, there isa need to ensure thatthe patent sfstem continues to ake the right balance between the appropriation ofthe fruits of imnovation by patent holders and the diffusion of technology forthe society a8 a whole Policy oriented research based on empirical evidence to address these questions is seare and an open dialogue among poli

‘The Conference on IPR, Innovation and Beonomie Performance organised atthe end of August

3008 hy the OECD Directorate for Science, Technology and Tadusey was designed as a Forum for experts, policy makers and stakeholders to exchange views on these issues The conference showed that the relationship helwsen patents and innovation is complex and varies aeross industry sectors, technological fields and types of firms Inteased patenting seems to he a mix of grovving innovation, inernationalisation and satepic patenting by firms ax new ways of organising research and innovation, Teasing out the relative contiutons of cach of these factors is a dificult task PPanicipants identified a certain number of areas deserving attention fom policy makers to ensure that patens continue to both proicet inventions and encourage diffusion These included enhancing the secess To patented technology, ensuring the quality of issued patens, especially in emerging areas, improving intemational co-ordination and finding the best vay of implementing IP protection in developing counties

‘This introduction summarises the conference presentations and discussions ts followed by a compilation of writen contibutions prepared by a number of participant, organised into five pans thar roughly follow the conference suture The first part explores the link between patenting, inno- vation and economic performance The seco describes recent changes in patent regimes in major OECD regions The third analyses the impact of patents on entrepreneurship and ditfusion of tech nology The fourth pat looks athe protection of intellectual propery in software and services, with & pamicular focus on the impact of patents on diffusion of knowledge inthis aca The last pant con-

‘ides with the Views of patent officials, poiey makers and expens on curtent and future challenges for patent policy including issues relate to the implementation of patent systems in developing

Patents and economic performance: Establishing the inks

“The focus ofthe first session was on tends, detemninans and obstacles to patenting Participants tried 10 answer questions sch as whether the siege in patenting has been a generalised phenomenon e108 Sectors anid counties and atempted to identify the determinants and impacts ofthe inteasing Jmportance of patents,

Many interpretations were offered fr the factors driving growth in patenting, Pasiipants agrod

‘on the need todo futher ressarch om the deteninants ad impacts of patents on the economy in order

to be better prepared cies to dra eatsaiis to face future challenges in his il and implement policy changes, but recognised the

‘The econometric analysis of EPO patents presented by Samuel Kortum from the Univesity of

‘Minnesota suggests that half the growth in filings at the European Patent Office (EPO) during the 1990s cam be atrbuted to a decrease in fees (see Chapter I, Intemational Patonting atthe EPO: A Quantitative Assessment) According to the model developed in his work, whichis restricted to the carly stages of EPO filing, about 60% of the nearly 6% annual increase in European patents between 198 and 2000 can he attrbuted to the decline in EPO application Fes He atefbues the remaining row to an inereas in innovative ouput of around 2.3%, However, better data on patent counts and patemting costs (eg translation costs, national agent fees renewal costs) discussants noted, would be needed to further elaborate an economctri model on the determinants of patenting in Europe

Statistical analysis and information from industry surveys in differen counties offered other imerpretations tothe determinants of patenting For example, the work presented by Jakob Elder, from the Fraunhofer Insitute for Systems and Innovation Research, is hase on a study of EPO patents filed by German firms and a survey of more than S00 German patentees, Eller fads that German patent applications at the Buropean Patent Office more than doubled during the 199Ds, whereas R&D expenliures grew ata much lower rate (ee Chapter 3, Scope and Nature ofthe Patent Surge: A View Jom Gennany) Respondents 10 Ealer's survey reported that pressure 19 compete and co-operate, rather than decreasing filing ost, or ineveased inventiveness, had heen the main divers of increased Patening in Germany during the 199s, Edler suggests that this factor led to-a breaking up ofthe tioct relationship between R&,D expenditures and patent applications during the period

tectng their inventions in recent years, They noted tht a increase in the numberof inventions had heen a very important driver oftheir inereased patenting, especially inthe ICT and pharmaceutical sectors More than half of responding fims recognised that they are now patenting inventions they

‘would not have sought to patent ten yeas ago, even if they were patentable thea, Ineed, as inated

no have hecomne more important i the ast Tew yea Tor very large Finns In contrast,

to Kons reslis, however, the survey indicated that changes in the cost or administrative proce dures for patenting had only played a moderately important role in the patenting strategies of res- pondants, which were essentially large fins

Opinions were divided as to wheter pro-patent policies have had any’ visible effects on business innovation, Arguments in favour of pro-patent policies had largely relied on the conventional wisdom that patents have a positive effect on innovation and economie performance However, participants noted that concerns about the effectiveness of the teat system have started wo be raised and empirical Studies available to date show that impacts tend to vary across sectors and counties,

Recent changes inthe US patent system have coincided with rapid technological advance, revival

of productivity growth and restoraion of US compettivensss, and prompted other countries to em Jate the US patent poliey model, noted Stephen Merrill, of the National Academies (NAS), (see Chapter 5, Inproving Patent Quality: Connecting Economic Research and Policy) However, the US patent model is increasingly under scrutiny, and concerns have been raised about patent quality neg patenting oF incursions of patent protection ino the domain of ideas, Proposals for reform have been expressed ina number of US fora during 2003." Although progress has been mein unde: Maning the patent system, Merl concluded, thors sil noed to beter communicae the results of economic research 10 policy makers The design of a framework for evaluating the patent system Should take ino account factors such as patent quality, inclines of dcisions ad transaction costs Kazuyuki Motohash, from Hitolsubashi University, insisted on the difficulty of discerning the effets of propatent policies on business inovation and diffusion of technology, especialy inthe presence of increased technological opportunities, Patent strategies vary across firms and Sectors and

‘ver time, Based on statistical analysis a interviews with large companies of the TT and biophann ceutical sectors, Motohashi showed! that large part of the growth in patent applications in the late 1080 was due toa sunge in IT and pharmaceutial patents, which coincided with inerased RAD and expanded patentability in those areas, He also noted that dhe growing importance of patents have Ted

TP sections within fms to play a more active role in R&D and product development sales, pat ularly inthe pharmaceuticals sector The role of patents for technology diffusion is also reflected in his study where he notes that FT and biotechnology firms tend to increasingly rely on external know ledge via licensing (see Chapter 2, Japan's Paten stem and Business Innovation: Reassessing Pro- patent Policies) He concluded that patent policy making should take a balanced approach and ‘onside impacts on both technology diffusion and expansion of protection, Iniatives such as those Jncudd in the Japanese Framework for Intellectual Propet Strateay lainhed in June 2002, be sug ested, should take ino account characterises of different sectors

The important ole of patents forthe diffusion of technology was stressed by Thierry Sueur, Air Liquide Vice President of Intllecwal Prope Sueur noted that a systematic analysis of patents from third panies represen very imponant source of techaical information for fans and provides inspiration for new inventions (see Chapter 6, How Do Third Parts Patents Faster Inmavation?), Third party patents facilitate information on competitors’ strategies, and motivate in-house inventors by Challenging them to find atemative solutions Among the necessary conditions for the net effect of patents on economic performance to he positive, Suen mentioned predictability and high quality san

‘ards fr patents, along te same lines of previous speakers

“Thẻ need to monitor and improve patent quality’ inorder to be better prepared for new challenges inthe future was again stressed during the open discussion atthe end ofthe sesion Pateipants also recognised the ulilficultes involved in extracting meaningful conclusions from patent statistics and surveys and recommended caution inthe interpretation of results in his complex area

1 NAS Commie on IR inthe Knowledge-Based anon a wie: atonalacademiesorg

——

‘Changes in patent regimes

Speakers and discussant inthis session observed that most changes to patent regimes exper

‘need inthe past rwo decades have gone in te diteetion of expanding and strengthening patent eo-

‘The United States, Europe and Japan have experienced similar pattems of expanding and strengthening patent protection in cent yeas, as rellected in a study of recent changes in patent regimes presented by Catalina Martinez of the OECD (see Chapter 7, Overview of Recent Changes

‘and Comparison of Patent Regimes in the United Sates, Japan and Europe) Changes have inchased the expansion of subject mate, a roadening of patent scope, a substantial increase of damage claims

in patent tigations and growing flexibility and simplicity of patent procedures International harmon sation has heen an important factor in the recent developments of patent regimes across counties,

‘marked inthe frst place by the TRIPs agreement setting minimum standards for IP protection within WTO counties Nevertheless, county differences ant unseted questions remain along a number of dimensions For example, as regards the patentability of software (curently under disession in Europe), the patentability of business methods and the status of research exempsions As note inthe proviogs session, the rationale behind most rforms seems to have been that strong pact regimes bring high economie henefits to nations in terms RAD and innovate, However, pro-patent polices have been putin ple without much regard to thei of Foreign direct investment, incentives wo invest in effets on competition oF the lifusion of knowledge, which are important questions and deserve futher research

‘Along the same fines, Akira Goto, from the Univesity of Tokyo, noted that dhere has not been sich discussion to date onthe impact of rent changes inthe Japanese patent regime Japan's ro- Patent policies have hoen diven mainly by foreign pressure due to intemational ade disputes and US pro-paten policies and domestic facors sch as rising levels of high technology ‘Technological progress had enhanced the compiivenessof the Japanese industy, which in tum triggered foreign and domestic pressure to strengthen the patent system Pro-patent policies initiated in the late 1980s in Japan included the expansion of subject matter, an ipeease in the number of claims patent, Timits on compulsory Heensing, higher incentives for university patenting and licensing nowt in damages awarded in tigation eases, Following these changes, R&D and patent applications Slowed down, although the number of claims per application and the number of grants grew cons

‘erably in the 1990s Funher research, Goto concluded, will he nected to assess the impact of stonger patent protection on innovation in Japan hase om this evidence,

to avoid or correct Based on his empirical work on the matter, Harhoff concluded that post-grat

‘opposition proceedings, enabling third aries to oppose patent grants, appear tobe an efficient remedy

10 contol the quality of patents issued at EPO,” He also noted that although re-examination pro ceedings a! USPTO could in principle serve similar purposes, in practice, they do not Between 1981 and 1998, 86% of EPO grants were opposed, with one-third of them resulting in revocation and nother thin! in amendment, In contnit only 0.35 of USPTO grants were re-examined between 198] fand 1998, and only 8 minor fration of them was revoked and proceedings were initiated by patent

lowers in as many as 40% of (Bose cases Hatoff rscommerdel USPTO to adopt an opposition system similar (© the one in place at EPO, for which, he suggested, there is also some room for improvement

As noted by Ulrich Sehatz, former Principal Director for Inerational Affairs at EPO (see Chapter 8, Recent Changes and Expected Developments in Patent Regimes: A European Perspective) the priority i the political agenda for patent reform in Europe scemt o he the integration of diferent rational systems and, ultimately, seting up & European patent system Ongoing projects comprise intergovernmental initiatives sue as the 2000 revision of the European Patent Convention, the 2000 London Agreement on translation requirements and the daft European Patent Liigalion Agreement (EPLA) aiming to reduce the uncerainty caused by multiplicity of juisprudences within EPC eoantiee, tll under discussion Negotiations on the Commanity Patent pursued by the European Commission have put on hold discussions on these lst two inergovernmental iiiatives, However, the implementation ofthe Community Patents ot sieaighlforwant either and manly due to diferent positions of Member countries on tunsaton requirements and the implementation of a centralised patent ott in Europe

‘operation Treaty (PCT); negsiations onthe Substantive Patent Law Treaty (SPIT), whose objective is

1 harmonise substantive isues such as the definition of prior aft, novelty, inventive stephan toiousness induseal appleaBiiQ/udy, the drafting and iterprotation of claims al the r-

‘quirement of sufficient disclosure of the invention; and the protection of folklore and traditional Knowledge Gurry drew attention «9 a changing multilateral environment for the development of TP protection where the policy dialogue is becoming more difiul © manage as becomes more closely connected to broader issues such as health polly, economic development and teritrialty of Fits, Some actors have responded by pursuing bilateral epoiations outside the mulilateral arena he noted,

To conelude, Gury insisted on the need for more empirical data to feed into the imterational

he noted that examining sn enforeing patents is very expensive business, aa developing counties cannot always afford it They need financial assistance to upgrade their systems in onder to protect their indigenous inventions and atract foreign onc It is neither desirable nor possible for devcloping countries to develop large patent offices He suggested that they should instead have improved access

to the expertise and capacity ofthe patent offices of developed counties,

A large part of the open discussion at the end of the session centred on the need to increase co-

‘operation among OECD countries and the estion of whether enforcing patent rights in developing

‘huntecs is beneficial to their own development, Some directions for policy wore suggested to help countries improve the efficiency oftheir patent spstems, notably fostering greater co-operation among,

‘major patent offices in a number of areas, including prior art search,

15

Patents and entrepreneurship

Participants in this session noted thatthe impact of patents on competition, and entrepreneurship largely depends on the likelihood of patent holders to stecesfully enforce the exelisve rights eon ferred by patents, Which i in turn affeted hy market conditions and differs across technology ate,

‘The question of whether the patent system inereases innovative activity is very old and has not yet been completely answered, as pointed out by Bronwyn Hal, fom the University of California a Berkeley Hall noted tht the growth in patent applications observed in the United States since the eration ofthe Court of Appeals for the Federal Circuit in 1982 has been largely due to increased patenting by US corporations from electronics, computing and communieation industries Empirical evidence shows that although strengthening patent protection does not generally result in an inrease

in innovative activity (except probably in pharmaceutical and chemicals), it does increase the strategic use of patents I also affects the structure of industries by allowing trad in disembodied knowledge, faciitaing the venieal disintegration of knowledge-based industries, and enabling the entry of nev firms tha only posess intangible assets Changes in patent regimes also seem rediret innovation towards patentable inventions, when patent proteetion is relatively more effective than trade secret Jaw.” Hall stressed the importance of granting high-quality patents asthe benefits of patents tend to inrease with quay (more novel, less obvious, beter disclosure) But the ost remain about the sae, Congestion costs, however, intease propotionaly tothe numberof patents, she coneluded, so that more snot bote?

er patent has not risen Empirical evidence on US patent suits indicates that there is substantial heterogencity of litigation rates across patents and patent holder Patents are mote likely toe li fated the more valuable they are, in ater words, the larger their scope, in tems of number of claims, and the higher the numberof inventions building on them, in terms of future citations hy claim Also, Schankerman showed that patent holders tnd to have a higher probability of filing sits he le con" entated tei technology area andthe smaller hee patent porto

Panicipants suggested that the cost of patent enforcement is higher for small firms than fr large firms, notably cause sinall fens do not have aeeess fo the same amount of mechanisins re firs tse 10 resolve ligation threats (risk-pooling aetoss large patent portfolios, access to legal expen, se) They also noted that although some countries have implemented ä number of measures to roduce the burden of small firms in patent enforcement, none of them is frequently used These measures include patent tigation insurance schemes, patent pools and licensing agents, allematve dispute rso- ution mechanisms and competition policy remedies such as compulsory Ticening Some of these solutions are being considered or have been already implemented by the European Commission, as noted by David Ellard, of the Directorate General on Interal Market.” In particular he mentioned the project ta implemen a European system of patent ligation insurance and discussed some issues a the

3 P Moser 2003) “How Do Patel Lvs Inlaece laovaion? Evidence from Nineteenth Fa NBER Working Paper No 9909 wen: nbenoncpapersty 9300, Centary Wold

4, 4 Lanjouw and M, Schakerman (2003), "An Empirical Analysis te Enforsemen of Pant Rights nthe United Stats” in Paton he Knowledge: Based Economy, Wesley Caen and Steven Meri es (Nato rreprenciship sealable st hapeurypaeninvcommlinte ral matredindprpinde

"6

Jmerface between competition and IP la, such asthe potential effets of compulsory licensing on

‘market competition,

“Thẻ risk of litigation also affets the likelihood to obtain suppor fom a venture capitalist inthe aly stages of fim cretion, Venture capitalists spend a latge amount of time evaluating the intel lectual propery of start-ups and their tigation risks They do not want to ad an akitional risk lo those already inherent in cresting a new firm, The higher is the risk of ligation the lower the probability of raising venture capital Jean-Bernard Schmidt, Chairman of the European Private Equity and Ventre Capital Assocation, stressed the Tact hat one of the Key ingredients for a sary

to be successful sa srong differentiating factor that may'laer sve it a competitive advantage inthe

‘market (apart from good management and target market) This factor may be related 10 technical Enow-how, specific kowledge ofan industry, marketing skills, patents and other forms of intellectual propery The effetivenss of patents asa differentiating factor Vates seross markets a technology fares, and appears to be higher, Tor instance, in fe sciences, microclectonics of hardware than i Soltware or healthcare services As regu life sciences, Thane Kreiner made clear in his resent tion that patents had heen critical othe success of his company Atiymetix (see Chapter 9, Enabling Innovation, Encouraging vesiment:Imellectual Property and Genomic Enepreneutship)

Discussions at the end ofthe session conted around the impact of patent quality and enforcement

‘mechanisms on market entry and competition Discussans identified the need to devote more resources 40 the analysis of patent quality and is improvement, and identified a number of factors contributing to an increase of patent quality: strict patentability equements, sufficient information tiselosure and legal eeraimy about validity

Patents and technology diffusion

‘The focus of the fourth session of the conference was on the impact of patents on technology Aitfusion Participants were asked whether patents facilitate the diffusion of technology and whether transaction costs were higher in somt sectors o for some types of fim than fo others

Patent licensing agreements are one of the mechanisms whereby inventions are diffused int the economy However it difficult to analyse recent tends in licensing as agreements are based on Private contacts among firms Nevenholess, information from most available databases, whieh are

‘mainly based on public announcements, indicates tha licensing agroements signiicaly increased from the mid-1980s to the mid-1990s and reached peak around 1995, Nicholas Vonortas, from George Washington University finds that thee factors increase companies’ likelihood o engage more ffi in Tcensing agreements: similarity in technology and market profile; existence of previous licensing agreements (o each other of independently); and importance of IP protection to thet industrial sector (see Chapter 10, Technology Licensing A key finding of his say s that reducing transaction cat is most important when licensing across stor, whereas strategic and compettion-

‘elated factors are most importa when licensing within an indus

Innovation’ Pp 285-340 nS Merl, R, Levin and M Meyers, es, Patents inthe Ẩmonledgediord.

đdown©Ín ieet ing negotiations for commercially worthubile projects They revealed that firms and Universes tend to adopt “working solutions” when they face IP aeess problems, These solutions include licensing inventing around patens, going offshore, court challenge infingement (use ofthe patented technology without a license often invoking research exemptions, and allemative solutions Such asthe creation of public or quasi-public databases to make research tots and other foundational inventions widely available (e GenBank, SNPs Consortium, Metek Gene Index) Nevertheless, Cohen noted, new technologies, new entrans and egal decisions nay yet upset the delicate balance between protection and diflusion in the biotechnology sector, and policy makers should remain vis- lạnh

In contrast to the practical approach taken by Cohen, Sandra Thomas from the Nuftild Counell

fn Bioethics, and Thane Kreiner, from Affymetrix, questioned the overall legitimacy of patenting genetic inventions, and recommended reforms to dramatically Timi the number and scape of gene patents

“Thomas presented the conclusions and recommendations ofthe Nuffield Couns recent repo

‘on The Ethics of Potenting DNA, which ate based on the principle thatthe patent system should promote the public interest, providing incentives for research, diffusing new technologies and making available medicines and diagnostics.” She identified two tends characterising the eurtent situation of the patent system: protetion of investments and proeetion of information, noting that protection of mvenions should be the nue instead, Recommendations from the Nalield Coundi o alesate the tension created by incresing patent protection in biomlicine included: 1) re-examination of the srounds fr eligibility of patents fr genetic inventions: 2) rigorous aplication of patenting criteria by Patent examiners, in particular the inventive step requirement: 3) rection ofthe scope of elms t0 Specific defined uses and elimination of per se claims; 4) clarification of research exemptions; and 5) compulsory lense legislation when public interest sa sake

In acltion, Kreiner noted that gene patents create incentives to patent base information instead

of products and may lead to less follow-on innovation, create patent thickets and limit access to clinical tests, To avoid these problems he proposed a framework where only biotechnology inventions With proven funetionality would be patented, and simple catalogues of information, such as genes, proteins or expressed sequence tags would not He stressed the need to codify research exemptions and treate pools of existing zene patens, especialy those derived from publicly funded research, in order

to guarantee sufficient access to fundamental gentic discoveries by researchers working inthe fel, both inthe public and pevate sector

Richard Johnson, senior partner of the US law firm Arnold & Ponte, etumed tothe broader topie of how 10 get the policy framework ght in order to respeet the social eonract inherent inthe rant of inellectual property rights (see Chapter II, Comments on Patents, Enreprenewrship anal

‘Technology Difiusion Private and public sectors are both keenly inlerestd in promoting technology Aitfusion: however, he noted tht thre recent tends are making the balance between creating incen- tives to rescarch and facilitating diffusion increasingly dificul First, new produets and markets draw

‘na variety of technologies and bundles of TP rights: second, actors involved in RAD have no longer clearly delineated roles, universities may do development work and firms basic research, bluring the line betwoen basic and applied research; and thir, the IP landscape is increasingly complex In some Scotors, the intersection of patens, copyright, and database protectin rights is likely to create new aceess problems and high transetion costs to which there ate no simple solutions One interesting challenge to the economics of P willbe the development of ew insituional and market mechanisns

mẽ ap

1%

1o facilitate innovation and diffusion 1s essen, he also warned, o prepare poliey makers for new challenges

‘The discussion at the end ofthe session futher elaborated the issue of research exemptions and access to foundational discoveries, Speakers and dscussants were largely in favour of defining a space

in which haste research could be fee of overly burdensome IP restrictions, Participants worted thai the preset patchwork of naional research exemptions may he breaking down, However it was cog ised that defining a research exemption that Would allow those who create research fols adequate protection isa dificult challenge and recommended Tuner research and discussions on the issue, Panicipants also noted that researchers in biotechnology and software are increasingly using the public domain as an atemative means of protection, which not only involves a high degre of cifusion but also invalidates the novelty requirement preventing others from patenting This tend rises new issues fad represents another challenge fr policymakers that pateipasthoweh neded tobe adresse

Patents for sotware and services

Te fifth session ofthe conference focused on issues related to patents for software and sevice-

‘elated inventions Parcpants addressed questions such as whether patenting plays a ferent role in protecting IP or in enhancing competitive advantage in services and manufacturing sectors They aso looked at whether altemative means of protetion for software and services are move effective than patens As regards software-elated inventions in panicular, they discussed the effect of patents on Aitfusion and fuer innovation and how the emergence of open source software has affected the role of patents inthe sector

‘The distinction between service and manufacturing sectors does not appear to he appropriate noe

‘meaningful as regards the analysis of IPR The large heterogeneity within diffewent types of services| and manufacturing sectors as regal innovation and use of IPRS calls for analysis based om mone etl industry characteristics, as show in the work presented by Pierre Mohnen from MERIT (sce Chapter 13, fnelectual Property in Services: What Do We Lear From Inmovation Surveys?) Based

‘on data from the Third Community Innovation Survey in France, Mohnen observed that innosative firms in high-tech manufacturing use all types of IP mechanisms (except design registrations) mone than innovating firms in low-tech manufacturing, and also more than innovating firms in services In tura, innovative firms in services use lead-time, copyrights, complexity and trademarks more than innovative firms in lou-lech manufacturing and nonsianovating Firms in services tend 10 use TP

‘mechanisms (except patents) more than non-innovaing fims in manufaetuing, The split between innovating and non-innovating lies seems to be much more striking than the spit between services

‘nd manufacturing Morcover, within the manufastring/srvces spit low-tech manufacturing ssetors appear closer to services than to highfech manufacturing sectors

‘munication Studies, University of Michigan Parens on Software and Serices Reveal about she System) PaleMing sollware inetaes the com (see Chapter 12, Through the Lens of Intangible: Whot plexity of knowledge management ad ases transition costs as each software prod may comprise

‘many’ patentable components Moreover, Kahin claimed, the software industry has experienced Very rapid growth inthe pas in the absence of patent protection On the other hand, patenting business methods may confer exclusive rights over abstract ideas, as patent that simply articulates a general business problem might cover al its possible solutions As a result, Kahin proposed « exclude husiness methods an software ‘as Such” from patentabiity, as well as algorithms, abstrt ideas and hase sience, and questioned the one-size fis-all model of patent protection,

»

‘The open souree community is also against the patentability of abstract ideas and algorithms, 35 pointed out by Hartmut Pile of the Foundation for a Free Information Infastucture an Eurolinx ane He argued that ideas should be inthe public domain as there is no intervention of laws of nature inthe production of ideas and no investment needs to be recovered, in contrast 0 inventions {see Chapter 16, Why Are Sofware Patents So Trivial?) Pilch also claimed that most software patents ranted so Far tend to be tvial and broad and insisted on the ned vo do more economic analssis on the impact of software patents on innovation, This research would be especially important to assess the proposal for auirective on the patentability of software currently under discussion in Europe, He con- clulsd by wishing thatthe final dnetve, following amendments suggested by his group, ensures tht patents ate granted fr inventions, ao for abstract ideas,

Participants agreed that even ifthe conventional wisdom is that more and stronger patents stim- Jate innovation in all fields, inching software, very few empirical studies have investigated the impact of patents on software innovation and lle evidence in citer diretion has been found o date

‘ates, software patents ae mainly granted in the United Sates to large, established manufacturing firms, which also rank hish as patentces in other technology felis Recent legal changes in the United States might have encouraged strategic patenting and led to lower R&D intensity by these fms However as pointed oot in the dseussion, a correlation does not nacessaily imply causality and ai- tional research may be needed

“The recent sein software patenting has been accompanied by significant growth in patent ligation

in software patents that has ourpace grow in patent Iitigtion gencrally Firms within the soltware sector have been involved in a sina fation of suits involving software patents, and most of the ligation related to software patents has involved either Firms producing hardware or Firms frm a wide ange of industries that crploy software i their products of process, Moreover, Graham finds

‘har patent tigation suis involving fims in the software industry tend 1 he inated by snall firms, individuals and indus’ outsiders, with incuibents offen being on the defensive side

20

Uni the 1980s, trade secret protetion and contrat law were the main means of protetion for software, with copyright gxining popularity since then, Patents are now becoming increasingly im: portant, IPR protection for software has evolved aver time, logether with the evolution ofthe software industry as noted by Marie-Theérése Huppert from Microsoft (ce Chapter 11, The Role of IPR for the Software Industry; A Changing Landscape?) The pressure from users to make source cade aail- able and the success of open soure sofware have imposed new challenges to software developers

‘sho now nee stonger protection mcans, sich as pens Patent law requires information disclose, but protects the invention sll, not just a specific implementation or its expression which s protected

by copyfiaht Moreover, patent prosecution ensures that enforcement applies 1o innovative features of produits, rather than to Simply ineremental features, Patents may not affet the Fundamentals of inno

‘ation and competition in the software inst, Huppert argued, but affect the behaviour of users and developers, as shown by the US experience, and will lead lowards the adoption of more pen and Jteroperble systems in the software industry

{A final point was also made on he role of patent offices to increase patent quality Patent offices

‘may provide training and funding to ensure the quality of examinations, improte databases of poe and get beter documentation concerning patents of new technologies In addition, they should fuirantce thatthe inventive ep of non-abviousness requirement is fulille A recommendation to Patent offices, based on theoretical results, would be to mata a high standard of inventiveness, especially Tor software and servicetelated invention, dscussants noted, They ako suggested that industries should pve feedback on the quality of patents granted by patent offices and agreed on the importance to undertake empirical research on patent quality

(Current and future policy challenges

‘The last session of the conference was intended as s forum of discussion among high-level officals, policy makers and experts on the main challenges currently faced by the patent system and Aicetons for future reform I was chaired by Francis Gurry, WIPO Depaty Director General, and speakers included Ichiro Nakayama, Incelletal Property Straepy Headyuarters, Japan; Manuel Desantes, EPO Vice President, International ATfis; Lois Boland, USPTO Acting Ditector Office of International Relations; and John Barton, George E Osborne Profesor of Law at Stanford Law School and chair of the UK Commission on IPR (sce their written contbutions in Chapters 18-22)

2Ð

book

2001 with the objective o find beter ways to integrate IPR and development policy" The final report feleased by the Commission in 2002 recognised that although the patent system is needed in the developed word, especially for pharmaceuticals, it does not work forthe poorest, as they lack the esoures ad incentives to fully pariipate inthe system,

‘The case of transition economies with a significant scientific and technological capability is itferent, Barton argued, as they are now entering global markets and are increasingly confront the Stronger patent systems of developed countries However, although strength he helpful oatret foreign ditet investment to some sectors in those counties, song patents ate not

as necessity fo aract investors as high educational levels and a good busines climate Siengthening their patent systems wil likely lock-in the competitive advantage of global leaders atthe expense of Indigenous firms, who will not be able to imitate, may not Renefit from licensing and might be acquired by mukinationals investing in thei counties, Carel qualitative and quantitative analysis read to be produced i assess whether strengthening IP regimes in developing counres is relatively

‘more beneficial to leading multinationals thant indigenous firms

‘One discussant suggested that ational technical assistance could help developing countries take better advantage ofthe developed world's patent system and indicated that interational organi Sations could provide financial support in that direction, However it was noted that funds available for

TP technica assistance at international organisations, WIPO in particular, are too sal or that mis son, Moreover, Barton atgued, its nt clear to whal extent developing agencies should spend mney

on strengthening the I systems of developing counties, ather than on providing them with medicines and food There is need to think avoid a duplication ofthe prfi-center model for patent offices in developing couites One solution, of beter IP systems for developing countries, Barton followed, ato

he suggested, could be to ereate regional offices and explore a better application of measures such a5 compulsory licenses and research exemptions,

‘The discussion concluded with & recommendation to ake a balanced approach for Finding the best implementation of IP protection in developing counties, considering the interests of both the

‘developed an! he developing world

Conclusion

‘Most panieipants agreed that patents play an increasingly important role in innovation and eco omic performance, They recognised thatthe relationship between patents and innovation is complex and aes across industry sector techpologieal field and ype of firm Although there is general

‘agreement on the effectiveness of patents 38a means to provide incentives to innovate in some Sectors (Gg pharmaceuticals), eonchisive evidence is scarce in others (e.g software) Nevertheless, partic pants noted the expansion of patent protection has affected firm behaviour and indusry structure in rnost sectors Ina context of increase technological opportunities, patents have ereated market oppor- tunities for seience-based firms and falitated acess 0 external pools of technical Knowledge via licensing, Ensuring the high quality of issued patents is necessary 0 enforce the dul role of patents

in fostering innovation and echaology diffusion, Policy makers should remain Vigilant and ens thai the patent system has the tols it needs to avoid the grant of low quality patents, characterised by extremely broad elsims, weak inventiveness and insufficient diselosure, which may Tead to congestion inthe patent system and induce anticompetitive strategie behavious,

‘A unanimous call was raised by participants t0 improve existing empirical evidence related 10 patents and increase poieyoriented research on the relationship between patents, innovation atl eo nomic performance Some directions for esearch were identified based on a number of unanswered

‘questions raised atthe confernee, These inluded issues such a the impact of patents onthe develop-

‘ent of technology markets, the role of research exemypsions on he ifieion of technology for follow

‘on innovation, the impct of patents on sofware innovation, the deiiton and contol of patent quality

sn IP protection in developing counties,

With the objective to address these questions and serve a frum for futher discussion on the

‘economics of patent policy, participants strongly encouraged the OECD to continue its work on IP

Fa

Chapter 1 INTERNATIONAL PATENTING AND THE EUROPEAN PATENT OFFICE: A QUANTITATIVE ASSESSMENT

Jonathan Eaton [New York University and NBER

‘Samet Kortum University of Minnesota, Federal Reserve Bank of Minneapolis, NBER

Josh Lerner Harvard University and NBER

‘Siandard source of data om intematonal parenting suggest a sarge in international potent activity inthe let deve ofthe twentieth century, The World Inellectaa! Property Organtation (OVIPO) reports a five fold nereane in non-resident patent applications in Denmark and nearly a rpling in Germans, or example During thx period the European Patent Office (EDO) evolved

to provide asinpler and cheaper means of seeking patent precion in most European counties We find that some of the apparent explosion in patenting arses from a ovestaement in the tusiber of countries designaved in European patents Correcting Jor this mismeasurement, however the data nevertheless ply substantial growih: 1) published European patent appl cations grew by nearly 6% annually benveen 1991 and 2000: 2):his growth was not at the expense of patents sought directly through matinal patent offices, which were small in number

‘lative to Baropeon patent throaghowt the period: dnd 3) the eumber of destinations designated for protection na typical European patent ha grown substantially, tthe pont there most not ‘designate all EPO members A simple sractaral model of an insentor's panting decision

‘utributes about 60% of the growth in European patnts and al ofthe growth in the breadth of Uesignation wo he decline tn EPO application and designation fees Asrbating the remaining [rowthin patenting to an increase in Innovative outpatinples tha lnovadve duy ket at ở

‘more modest bus nevertheless impressive anual rate of 2.3% over the decade

mr

‘The teratonal patent body that has moved furthest inthis dreton isthe European Patent (Office (EPO) irbegan withthe ternational Patent lastiute, established by a group of European ‘countries in ‘The Hague in 1047 as a common resource for patent sarching and archiving,

"ushering in a mow interational approach to patemting Two year later, the Coun of Europe aavocated the creation ofthe European Patent Office The EPO however, was not established

‘ntl 1973 by dhe Convention onthe Grant of European Patents Despite what its name might suggest, the EPO does not issve Earopean Union patents or even Lurope-wige pens The European Patent Convention, under which the EPO was founded, provides single patent grant procedre, bul not yet a single patent (3s envisioned in proposals fora “Europetn Patent Neverhcless, we will use the term “European patent” co refer to patents processed though the EPO Afr gran, the European patent becomes set of raion patents, We refer to this xe sa European patent aly

‘An applicant scking to protect an invention with a European Patent must pay a foe for the

“European Patent itelf aod an additonal fee for each county Belonging tothe EPO designed for protection The eost of designating a county for protection in a European patent i typical Tower than the cost of filing fora patent dred trough the individual national paca offices Hence the European patent provides an attractive istrment for applicants seeking protcton in several destinations among members ofthe EPO Since the European patent itself is relatively expensive, however, applicants seking protection in only afew members will id it cheaper Tile drei through the relevant atonal patent oc,

%

We consider these 16 counties os destinations for patents We include in our analysis patent pplications fom any ofthese 16 EPO members as wel as those from Australia, Canada, Japan, [Nonvay, New Zealand, and the United State Thus, we consider 22 counties as sources of patents The total number of national patents (in any of the 16 destinations} duc to emerge from "European patent applications (rom any ofthe 22 sources) ple, trom under 430 000 in 199T t

‘over 1.3 millon in 2000, The increase was due about equally to growth fn the number of published EPO patent apliations (fom just over 5% 00 to just over 98 000) an an increase i he average fail size pr apliaton (fom about 7.310 aod 13.3 counties)

‘The cont foreach country designate for national protection fell more deamaticlly.K ws Lite

‘over USD 200 forthe fist si yess falling to under USD 100 in 1997 Maeover, after 30 June 1090, the EPO offered fe designation in any EPO member afer payment for seven, Hene the cost of either patenting a all or of patenting widely tach the EPO fll dramtically

‘Our analysis oF the EPO comprises thee steps We fst examine the various sources of data

‘on patent application, and pot to serious pitfalls in ntspreting Worl! Inelectual Property Organization (WIPOy data on eros-country patenting daring this decade The WIPO data, published in Industrial Property Staines, suggest tht patenting over the petiod grew much

‘more than it actually did The reasons are twofold First the WIPO data appear to double count applications through the Patent Cooperation Treaty (PCT! that designate the EPO Second they aan EPO application s an application to all EPO members even when not all members are designate for protetion at the Gime of poblieaton, The OPCD te to produee a refined version fof the WIPO data in ils Posi Science and Technology Satistcs that appeared to comet he First problem but ot the second We ales these reporting ssi

With our revised dat in hand the second step of our analysis provides an overview of the

‘basie pater in the evolution of EPO st naonat patenting ever the decade, Our approach is sided by the “gravity” Iiteature exaniningerosenaional pater in patenting In particular we Follow Eaton and Koren (1996) and Kortum and Lerner (1998) in distinguishing "soweo™ and

“destination” effects in cross-national patenting This approach allows us to assess how inno vation varies across counzis and evolves over time, and the probability that innovators will

‘choose to file fr patent protection ina given county We incorporate data on whether lings

‘were made det inthe national patent offices of EPO members or dough the EPO wo analyze the impact ofthe EPO

‘Ringo, USD 0 in Spi, USD 815 n France, and USD 882 te Nethrand

4 Slama (981) san eal contbuton

1 The European patent has slmostentely replaced dvet applications to national patent offices, For almost all EPO members in most yeas, patents that do not originate ith the EPO constitute fewer than 10% of patent applications arriving athe national patent ies,

2 Im comast with the eater “gravity” Iterture, which suggested that the proximity of ‘he source and destination mattered a great deal to patent application, we find that the pattern of counties designated for protection varies only slightly actoss sources The roleof gevgraphy ionly apparent ina fe eases and diminishes over the decade,

3 By 2000 most European patents designate all EPO members for protection Hence by ‘then applicants were exploiting the European putent to protect their inventions across the Continent, more or less ignoring the national route This siverslity contrasts anificaily with the past Pinan (1993), having assembled da set of patent Families for 1974 shows thatthe mean patent family size was two counties, and 705

‘of inventions that were patente a al were patented in only one count

‘The third step of our analysis is estimating the extent which the growth inthe numberof [EPO patent applications crn the 1990s is det ewer Tes rather than an increase i inventive ctv, Todo so we develop an estimate a srl srvtural model ofan aplian’s decision to patent in siffeont counties The decision is governed by a comparison of the cost of Hlng + patent application ina destination with the Value of patent thot, which we teat as the product

ff randoms and systematic components, Applicants may ether fle in individual destinations

‘rough their national patent offices or apply to a portfolio of EPO members though a European patent An invention may turn out not to Be Worth pateating anywhere, in whieh ease no Pro- fection is sug

‘We estate the parameters ofthe model to match: I foreach current EPO member county,

‘the number of EPO patent applications designating that country 2) for each cutent EPO member country the numberof patent applications sought directly though the coresponding national patent office, and 3 the distribution of faily sizes of EPO patent applications al for 1998 We {hen see hows well the estimated model ca explain developments thraighout the decade in these

‘We find that we can explain the increase in the breath of patenting among EPO members totally in terms ofthe declining cost of designating ational members, in particular the bulk Aiscount offered afler 1999 Moreover, the decline in the overall cost of secking protetion

"wough the European patent can explain wel over haf ofthe TF increase inthe amber of EPO patents, We atuibute the residal increase in EPO piteting to an increase in inventive ouput

‘ver the decade

‘ue pager is relate to a variety of other work on the operation ofthe global patent system,

In particular a numb of other papers have looked at isituonal changes and thie ip 08

‘he propensity to patent (nd, in some eases, to inven) Kortum and Lerner (1998) undertake a varictyof statistical dacompostions and modsl-fting exercises 10 explote the impact of the establishment of the Court of Appeals for the Federal Circuit onthe umber of filings in the United States MeCalman (2001) quantifies the impact of international patent harmonization elfsted by the TRIPs agrcement of the 1993 General Agreement of Taif and Trade by estimating and simulating a model of international patenting Lerner (2002) explores the overall response of patenting by domestic and foreign ets to patent policy changes over the past 180

»

‘year, comparing applications fled in the nation undertaking the poley change with those led in

‘Groat Britain, a nation witha relatively constant patemt polie Eaton and Kortum (1999) and Eaton, Guterez, and Kortum (1998) model the effect of ger, and more gsographiclly extensive, patent protection on innovative activity and growth,

‘Mowe generally, our work is related to dhe lieratue op pateatstatses as indieators of scientific progress and technological innovation Griliches (1990) provides an insight, and rather cautionary, survey of the lisrture, which includes such classic contributions as

‘Schmooklers Invention and Economic Growth (1966), Scherer (1965), and, ofcourse, the many contributions of Griliches himself and his students, aoably various papers in Griliches (1984), Schankerman and Pakes (1986), Giches, Paks, and Hall (1987), and Griliches (1989), This Bienture hạt Tong recognized how ehanges i patent istttions pose challenges to the technometic use of patent dla, For instance, when discussing the seeming slowdown in patenting during the 1970s i his 1991 review atele, Cliches wrged researchers to he eautions

‘out the possiblity of "a saisticl mirage, caused by a bureaucratic nghe than an economic or teehnoloiealcyele”

‘We proceed as follows, We next describe the sources of data, The thin! section then dis- cusses specific ses in reconciling WIPO and EPO patent data Having sorted out these ise as Satisfacoiy as we can, the Fourth section provides a presentation of some basic features ofthese

‘uta, The fh section presents our model ofthe decision to patent among EPO members, ouF estimation results, anda diseussion of what hey imply The Final section concludes

Patent families in the EPO publications data

‘The standard source of data on patent applications and grants sero counties is WIPO, which reports the tral numberof patch applications to each tional patent offic by country of residence of the applicant Until eealy the OECD has compiled these numbers, making some adjustments in an atempe to avoid double counting ofthe same patemt document (as described below) and published them in is Basic Science and Tecnology Seats volumes (e.g OECD, 19095), These patent statistics have served among other thing, as indicators of the intensity of inventive ativiyin source couniev an the dffsion of eas across borders

‘A dahack ofthese vources is that neither allows usta observe the extent to which patents sought in diferent national offices represent the same invention Since aplicants often seek protection in multiple destinations, these dala donot provide unambiguous insight ino pplicans’ decisions about he range of eountries i Which to seck protection An increase in the hummer of patent applications worldwide, for example, could reflet either moce ideas being patented or else the same number of ideas being patented more widely We ied dats on the pottlio of destinations in Which single inventions ae patented to resolve this ambiguity

6 The dạn igingnhh bervcen the incor” he advil or inviduals who came op with the id, slip hy skip ron whch ee ent he dc

‘oun Coatce an van Potelserghe de ls Potae(2001) examin the isbn dt ab ase inthe EPO dat

a

‘This modtng procedure requires data on the anilyof destinations reported in BPO patents

‘and on ttionl patents applied foe dieetly tough the national patent ofees of EPO members Before tuning oan analysis ofthese data we must Fist confont serious measurement ies in

ow these data ae epoted

Reconciling EPO and OECD/WIPO patent data

‘To construct complete accounts of patent applications in EPO counties we merged the data from the OECD on applications (hy destination county, country of applicant, and year of application) with the EPO publications data agregated (om the level of individual EPO Publications) by designated country, country of applicant, ad year of aplication Each EPO publication with designation in Fan se, for example, counts 38

‘ne patent application in France and one pat ium Our plan was fo measure the numberof patents applied for dreey though the national pateat office in each destination from cach source by the total number indicted in the OECD data minus the number of EPO publications fom that source designating that destination,

This procedure yielded a measure of national patents hat, very fequendy, implied thatthe

‘majority of patent applications did not go through the EPO Moreover, the measure indicates explosive growth inthe number of non-EPO patents to EPO destinations Pusuing the anomalies inthis measure, we tured to the anna IPO Industrial Property Stasis (Table It Az “Patent ‘Application Fie by Non-residents Broken Down econding tothe Country of Residence of the

‘Applicat whieh ae the Basis forthe OECD numbers

Conceptually applications for pent protection in an EPO member country’ ean follow any fone of four mutually exclusive routes: 1) applications to the national office ively 2) appli eitons va the PCT, 3) applications vathe EPO, snd 4 applications vi the EPO via the PCT [7 Th data al publ EPO pen ppicions we povided ous in avery convient fons by whe OECD,

x

In each year, starting in 1985, WIPO preseas up 0 four separate numbers fr each EPO:

‘member destination (what WIPO refers to the "reporting county of organization”) WIPO defines these four measures as follows

("Patent applications filed directly with he office concerned

+ (Py “Designations in international applications filed under the PCT" (not reported for some EPO members)

‘+ (REY: “Designations in European patent applications filed under the European Patent Convention

‘+ (RPE): “Designations in applictions filed under the PCT with view to obtaining @ European patent

1K might at first scem that these four cutgorics in the WIPO data correspond to the four mutually exclusive routes to apply fora patent Indeed, WIPO reports the sum of these four numbers in boldface The OECD data on patent applications however, report applications by Foreign residents asthe sum of only (N}, (RE) and (RPL) When the country of apliant i the sme asthe destination county the OECD makes alditionalajstments, The OECD procedure Suggest hat the (P) applications are contained within the (RPE) applications Evidence infor

ff the OECD's proces that WIPO iself reports that curently over 98% of PCT applications

‘designate the EPO each year (while ifthe (P) applications were distinct fom the (RPE) this percentage would be much sills)

‘he EPO, appears to move rowshlyin proportion tothe EPO measure, asi sould, On the other hhand, the series Tabeled WIPO (RPE), representing WIPO's count of German applications ariving atthe EPO via the PCT, takes on essentially the same vale in each LPOsmember, ‘Apparently WIPO treats 4 PCT application designating the LPO œ ifthe resulting European patent application would designate all EPO mentor countries The WIPO (RPE) measure ths

‘verstaes the st of countries which svc applicants wil actully seek protection Figure 1} Sows thatthe same isue ase, hut even move powerfully, when examined fom the perspective (oF US applicants

s

‘nn

Figuees Ie and Ltd show thatthe over counting in WIPO (RPE) makos a huge diference for the patern over the past decade of pateming in EPO-member counties We illustrate for the

‘cases of German applications in Denmark, and for US applications in Germany For the WIPO Series we plot the sum of (N).(P) (RE), and (RPE), the number that WIPO reports atta patent applications As mentioned above the OECD series exeludes WIPO (P, yielding a slighly Tower crv, Buta more dramatic aljstment comes abot if ve ignore the WIPO statistic a simply

‘ely on counts of published EPO patent applications, This last adjustment goes too fa, however,

if we want a measure of the total numberof patent applications fom one county seeking proeton ia anather The EPO data miss pate applications fed deely with the tational patet offices, as measured by WIPO (N) Addiag WIPO (N) wo the EPO count, however, i 4 Fatty minor adjustment For the ease of German applications in Denmark, (he adjusimen 100 small even fo be apparent in the figure

Table 1.1 illyseates the problem numerically Ie reports data on patent applications fom

‘twee sources (Germany Japan, and the United States) to de destinations (Belgium, Germany, and Denmark) as wol 351 the EPO ise, in the year 1996 It sepors, est, the number of EPO published patent applications designating the destination as fepote by the EPO fllowed by the

‘OECD/WIPO figure on total applications It then reports the four pes of applications reported

£0 Pa sa nanan Boigu 160828 nea 2888 Germany 110953 ssis_— ẽnn"

GamA 217478 30B 18G 0884 HB] tr Denmark — 33619 - 20/38 eae 1788887

£0 T82 wa m HH 930310980 Fist, note thatthe OECDAWIPO figure corresponds to the sum of WIPO (N), WIPO (RE), and WIPO (RPE) Thus, as reported above, WIPO (P) is not incl in the OECD/WIPO Figures Second, note thatthe nbmher of WIPO (RPE) patent applications foreach ofthe tree sources is vitally the same aeross EPO destinations This i the sane phebomenon we saw in Figures [a and [1b I suggest hat WIPO (RPE; is automatically counting each application to

‘the EPO though the PCT as it would be designated in every EPO member

{The EPO publications dat eeasinly oper single patent sought by join apticant om more than ‘is county We use the OBCD's rocodre of viding thse patent publcos se hượt sơ te

‘Siren sources, Hence the ptr for on-inpr wus nthe EPO clu

6

‘The number of WIPO (RE) patents varies much more across destinations than WIPO (RPE),

‘but ot enough to explain she much greater variation i the EPO figures across destinations Not, Finally, that when Japan isthe source, the sum of WIPO (RE) and WIPO (RPE) for EPO as a destination corresponds gute closely withthe EPO column for EPO total published applications

‘When the United States the sours, this sum i somewhat higher The discrepancy hetwecn the [EPO and WIPO figures for overall EPO patenting can be explsined by the distinction between patent applications (WIPO) and published patent appeatons (EPO) Give that applications are

‘ceasionaly withdrawn before publication, we eapect the WIPO figure to exceed the cores: ponding figure fom the EPO Our explanation forthe discrepancy at the level of vational {estinatins is tha WIPO is abating EPO applications as destined fr all EPO members wh

in fact only a much smaller subset of destintions were typically designated in the published application,

ta from other years tll he same story By 2000, however, the problem is lagely over as

by den published EPO pateut aplicaons typically designate actly ll members (a phetomenon

‘tha our analysis Below Seeks to expla)

1 the analysis that follows we calculate total EPO-member applications asthe sum of (N) applications fom WIPO and the counts constructed fm EPO publications Since there awe

‘elaively few (N) applications even as far back as 1986, the overall uends in our totals will Targslyeppear just slighly higher than those we construct fom the EPO publications data For example, from US applicants in 19N6 there were only 156 (N) applications in Austria, 329 in Belgium and 2034 in Germany (representing at mest 20% of application trom the United States in those counties)

Am overview ofthe data

‘The cros-county patent data have many dimensions to them: most importantly source, destination, and time An applicant from some country can apply for patent protection in any number of countries either directly or by designating members ofthe EPO As just discussed, even by 1985 the European patent was by far the most popular means of seeking protection Among EPO member It grew in popularity tothe extent that by 2000 the ation of patents Sought directly through the national patent offices rarely’ coasted more than 2% ofthe ttl Germany and Great Britain are the only EPO members that hy 20) il received sizeable percentage of apliatons foe patent protection dvetly Hough ther national patent fee, this avoiding the EPO."

Since the EPO patents tll most ofthe story, aaa fom a single consistent souree, we Aleck most of our stbsequent analysis toward the European pate, We fst lok attends in

‘who uses the EPO, and then Took atthe pattens of designated counties in EPO patent pubs Tictions, We then lok atthe interaction of source and destination, and th sizeof fami In this analysis the destination counties ate the major EPO members: Austa, Belgium, Switer-

9 Almas 39% of Austian, 23% of Swiss, 174 of Japanese and 65% of US applications for patent po ston in Germany skip the EPO and go arty 4 the German pute oie In he ease of those

‘ching pan protection in re a hi perenape cacti 10% fo appcatone rm Nast, Belper, Cana, Denmark, Toland Isp, New Zasot, apd the United San Sce we have invent seeking psfecdon through the Geman and UK pen afc wete applying or pens

‘ter eome nse or outside the EPO membership

mr

land, Germany, Denmark, Spain, Finland, France, Great Britain, Greece, olan, aly, Laxem-

‘our, Netherlands, Portugal, and Sweden The source counties include these 16 plus Austali, (Canada, Japan, Norway, New Zealand, andthe United Stats

Sources of European patents

Figue 12a depicts the evolution of European patents during 1991-2000 perio, aecordin tà the country of applicant, what we refer to asthe sour.” For clay we have dropped some of the smaller source countries Over the decade the total numberof patent publications grew fom

58 210 0 98 50S The exeat 0 which this growth reflects increased innovative utvty rather ‘than changes inthe patent system isthe lopic of Section 5

“Looking at individual sources, US applicants were the most fequent users of the European patent, accounting for around 30% of the total throughout the period The Japanese shar fell from 22 to 188 while Germany replaced Japan as te second-most frequent usr with its share rising from 18 t0 21% France and Great Britain follow with initial shares of 88 and S4%, respectively, falling by about one percentage point ove the decade During the psiod the Finnish find Swedish shces grew from 0'7 to 1:2 an! fom 1.6 to 2.5% respectively, Overall, however, the picture is ome of relative stability in terms of whose applicants sere making se of the European patent, Notably, ney half of the total EPO publications involved applicants from non member counties, mostly from the United Stats and Japan,

Destinations designated in European patents

Figure 1.2b depicts the evolution of European patents during the same decade according to [BPO members desienited for patent protection in the published EPO applications The Toal umber of EPO publication is he top curve, Throughout the decide around 98% a the European patents designated German, the most popuae destination, France and Britain follow closely at round 90%, rising t0 98% in the last 160 years In the fst half of the decade other EPO members sited significa Ul, dhe fourth most popular destination, was dsighatd in 73%

‘oF applications in 1990, a share which remainc|tale unt the las wo years, when its requency jumped to $7% Switzerland and Sweden were designated in just under half the patents, gain nll he lst to years when they started Yo be named in nearly 80% The smallest members were designated in about 25% ofthe European patents for most ofthe decade

10, Ax mecioned, wpa may lit igpleanl rỏm naBõHde coinbdek Ta thọ xt re allowed the Staton of el apps se om tit oun

‘A remarlable development i the movement toward university of designation in the lst two years of the decade, By 2000 even the sallest EPO members were designated in vsary 80%

‘of European patents An abvioas explanation isthe substantial drop in the size ofthe fee pet designe that occurred late inthe decade

Sourcesdestination interaction

To what extent do source countries differ in thir propensity to designate particular destination countries? AS mentioned by 2000 applicants were designating most EPO members Jmespective of where they were fom, while during 1991-1998 larger destinations were dsig- nated! much more often,

An cxamination of designated states by source during the period reveals a coupe of patterns

‘Most strikingly Japanese applicants tended to omit smaller EPO members much more than average A number of smaller counties (New Zealand, Denmark, ad reland, for example) tended to designate smaller members including those oer than themselves, somewhat more tan average, Otherwise designation patterns fom the olor source ours (eg Australis, Canad, France, Germany, Great Brttn, the Netherlands, and the United States) conform quite closely {o the overall average Thre i also some evidence of gengraphy playing role: appliants ae more likely to designate countries nearby

In summary, however, we find differences in designation propensities across sources stint slight and iienyrate so s6 otto warrant spermatic investigation a his point The rain sty in these data ober the pera is the general movement toward universality in desig hating EPO members For this feason We ignoce diferences across sources in designation ropenstes i what Follows

‘To the techpometrician, however this enhanced ability comes at a cost By lowering the har

to patenting in a large numberof counties, the European patent removes the need fr applicants

to spend much time thinking about where their inventions will ulimaely prove useful Hence

‘where they apply for patents may reveal less about the flow of ideas across rer than was the case a decade air

A simple structural model of patenting choice

‘As discussed above, the 1990's saw major growth inthe amount of patenting taking place mong EPO members, with much ofthis increase reflecting growth inthe sizeof patent fais But that leaves a substantial increase inthe number of inventions secking pant The lowering fof the fees both for taking out a European pateat and for designating an EPO member for protetion may explain some of the increase both in the numberof LPO patents and in the umber of eoxnres designated por patent, But how much? To get handle onthe answer we Iodel the applicants decision abot how and where among EPO members to abtain patent

0

protection for an invention We estimate this model using a cross-section of data on EPO and

‘ational patening for 1995 We then see what the model predicts forthe ether years of the cade, and how much ofthe increase in patenting it attributes to the Towering of Fes

‘The applicant faces a trade-off between the value of obtaining patent protection in a estination and the cost of applying for «patent there The European patent complicates the Picture since, by paying the fixed fee cost forthe European patents the applicant can apply for Prodeeion in any EPO membse at a mich lower cost than f she fied drelly through the national patent oti

We rece all thee complications ino a simple ypecitiaton in whieh the value of patent rotcton in countey’ nfo invention j depends onthe size of 1's market, Yon he size OF the

‘work! market 7, andon the quality of invention jas applies o county al aime

Vy (i= ally +8¥ ig lhe

Whore and b ate parameters, The tem mu (J) representing the quality of invention j as it applies to destination wat ime, has evo components

‘wherever itis patented while (7 latest the value of invention j specie to destination vt

fs important to remember that V4) isnot the value ofthe invention in markt itll but just

‘he incremental vale of having it proected bya patent there

‘costs cf lo apply fora European patent in year rand, having paid this cost, an additonal cost! foreach country designated by the European patent For an applicant to fle deel in country through the national patent office costs eat ie

We simulate an invention as a parila draw of (J) anda 4) We simulate 100 000 inventions by drawing (4#1) + 100,00 standard normal random variables, weighting the first by

1 tn pact our model des no insade the payment of renewal fxs spt of be ost of pateig

a