Pre-emptive patents - Enseignement supérieur et Recherche
Transcription
Pre-emptive patents - Enseignement supérieur et Recherche
Pre-emptive patents: what they are and what they do Dominique Guellec1, Catalina Martinez2, Pluvia Zuniga3 1,3 Directorate for Science, Technology and Industry, OECD 2 Institute of Public Goods and Policies, IPP-CSIC EAS Internal Seminar 1 Introduction ` ` ` Rise of patenting worldwide Concern about quality of patents Perception of increasing strategic use of patenting (anecdotal evidence, surveys): ◦ Bloking others to patent : pre-emption ◦ Defense itself and secure freedom to operate How important is the blocking phenomenon aiming at pre-empting others to patent? Can we characterise pre-emptive patents with the aid of patent data? 2 Survey Evidence – Strategic Patenting – Cohen et al (2000) “The Carnegie Mellon Survey” for US Reasons (nonexclusive) (% of companies declaring) Products Processes – Prevent copying 96 78 – Patent blocking 82 64 – Prevent suits (defensive) 59 47 – Use in negotiations 48 37 – Enhance reputation 48 34 – Licensing revenue 28 23 – Measure performance 6 5 Carnegie Mellon definition: “blocking patents’” are meant to prevent other firms’ attempts to patent a related invention 3 What are blocking patents? Block or “kill” other patent applications Entry barriers into markets and technologies : pre-emption ¾ if granted, market power to holder - Useful to attack others and gain licensing profits; to exchange technology; as protection from being accused of “infringing” others ¾ If not granted (withdrawn or refused), free patent market place - Useful to deter entry into markets: create uncertainty; ensure freedom to operate - Withdrawals can be seen as very defensive strategies : one entry barriers generated through prior art others can have their patents refused 4 Two main orientations of blocking patenting Offensive blocking: conceived to threaten, attack, and weaken others • Surrounding high value patents with a “picket screen” of minor patents (covering incremental innovations on its periphery) • Restraining freedom to operate of competitors by filing patents at the margin of their areas of activity to improve bargaining positions in cross-licensing • To achieve opportunistically licensing or advantageous settlement conditions in case of litigation (e.g. patent trolls, ingenious drafting of patent claims) Defensive blocking: to ensure freedom to operate • “building fences” to counter-suit in case of attacks • To keep others off the market (with no intention of commercialisation) • To prevent attacks by designing overlapping inventions across several patents (e.g. patent thickets) 5 Offensive patenting: x Patent sharks (“throlls”) companies or individuals (Henkel and Reitzig, 2007): Forgent Networks Inc (JPEG image compression standard); by early 2005 received more than US$ 100 million in licensing fees from 35 companies, and is suing 44 companies x NTP suing RIM (blackberry) threatening to shut down RIM's operations in the U.S. and issued a settlement for US$ 612.5 million in March 2006 Defensive/offensive patenting: x semiconductors since the 1990s (Hall and Ziedonis, 2001); software industries, e.g. IBM ; Oracle, etc. Mutually blocking (overlapping rights): x Plant biotech patents (Marco and Reusser, 2007) x Comestics, i.e. L’oreal and Henkel KGaA 6 ` When entry into an industry can only take place through the invention and patenting of a substitute for a monopolist´s product, they show that it is rational for the incumbent, under certain conditions, to preempt entry of potential competitors by patenting the substitute himself and letting the patent sleep ◦ the pre-emption threat needs to be credible (R&D commitment) ◦ Patenting a substitute technology would always be valuable to the incumbent because it enables to preserve monopoly profits, which typically exceed the duopoly profits with entry. ◦ The more drastic the underlying innovation on which the R&D competition is based, the lower the incentives for and the profits with pre-emptive patenting Îpre-emptive patenting associated more to incremental innovation ` Ceccagnoli (2008) : pre-emption tends to improve remarkably the appropriability of returns to R&D (especially for incumbents with strong market power and facing high threats of entry), thereby increasing the value from their intangible assets 7 Can we identify blocking behaviour based on patent data? • Prior art is used in most patent laws to asses novelty and inventive step (non-obviousness, US) : patent and non patent literature public at the date of the filing Citation categories (EPO and PCT search reports) Category X Meaning Particularly relevant documents when taken alone (a claimed invention cannot be considered novel or cannot be considered to involve an inventive step) Y Particularly relevant documents if combined with one or more other documents of the same category such a combination being obvious to a person skilled in the art A E Documents defining the general state of the art (but not belonging to X or Y) Potentially conflicting documents – any patent document bearing a filing or priority date earlier than the filing date of the application searched but published later than that date, and the content of which would constitute prior art X, Y and E are critical documents for validating patentability A documents define the general state of the art 8 How does blocking work? Model: Multinomial logit on the fate of the citing patent application (1: grant, 2: withdrawal, 3: refusal), related to: Conflicting backward citations (X, Y, E) State of the art backward citations (A) Forward citation profile of patents compromising patentability (“pure X”, “X and A”) procedural status of the cited patent (granted, withdrawn, refused, pending) Data: PATSTAT April 2007: 396 707 EP [atent applications and their corresponding EP cited patents (either directly or through an equivalent) • They correspond to all EPO patent applications granted, refused or withdrawn with priority years 1990-2000 citing other EPO applications, directly or indirectly, under X, Y, E or A citation categories 9 Expectations Are citations compromising patentability of other inventions blocking the patenting of other inventions because : ¾ they represent a relevant contribution to the art and thus “naturally” block further filings,? Îcitation category A ¾ Or, because they were “strategically” designed to limit patentability in a specific field (entry deterrence)? Înever cited as category A (dubious quality; rather incremental inventions) ¾Further, are un-granted patent applications (cf withdrawn or pending, refused) equally effective in enacting barriers to entry into technology? Patent data can help us answer these questions. 10 ` ` Blind, Cremers and Mueller (2007) Traditional protection motive is related to more forward citations than the blocking or exchange motives. Within blocking, offensive blocking is related to a significantly lower number of citations than defensive blocking Grimpe and Hussinger (2007) Deal value of acquisition target firm in M&A increases with the “blocking potential” of its patent portfolio, measured in terms of number of X and Y forward citations All EP filings Filings 859706 EP filings citing Pure X 97467 EP filings citing X and A 328178 EP filings citing Pure A 170622 Grants 475016 45985 189596 106241 Withdrawals 254792 33784 93258 45672 Refusals 24030 2595 9512 5030 Grants over filings 59% 47% 58% 62% Withdrawals over filings 30% 35% 28% 27% Refusals over filings 3% 3% 3% 3% Technology field Electronics-electricity Instruments Chemicals-materials Pharma-biotech Industrial processes Mechanics-transport Consumers-construction Total Non cited Cited as pure X Cited as pure X withdrawn Cited as pure X granted Cited as X and A Cited as pure A All filings 65% 62% 63% 66% 65% 66% 70% 11% 12% 14% 17% 10% 10% 9% 3% 3% 4% 6% 3% 2% 3% 5% 7% 8% 7% 7% 7% 5% 10% 12% 11% 9% 9% 9% 7% 13% 13% 12% 7% 14% 14% 13% 100% 100% 100% 100% 100% 100% 100% 65% 557 889 12% 104 716 3% 27 260 7% 56 705 10% 86 423 13% 113 660 100% 859 706 The MNL model Citation category X (“X”) Y (“X”) E (“X”) A (“A”) “X” from “pure X” “X” from “X and A” “A” from “X and A” “A” from “pure A” “X” from “pure X” granted “X” from “pure X” withdrawn “X” from “pure X” refused “X” from “pure X” pending “X” from “X and A” granted “X” from “X and A” withdrawn “X” from “X and A” refused “X” from “X and A” pending “A” from “X and A” granted Applicant countries Priority years Technology classes Observations Wald Chi2 Log pseudolikelihood Refusal 0.035 0.003 -0.25 -0.11 Cited profile Cited status Withdrawa Refusal Withdrawal Refusal Withdrawal 6.45 l 4.01 8.04 -3.77 0.09 8.03 -0.02 3.95 -0.13 -3.47 -0.01 -1.56 0.07 4.14 -0.12 12.90 1.55 8.05 -0.07 8.59 -0.13 1.23 YES YES YES YES YES YES 396 707 22 411.85 -283 868.42 YES YES YES YES YES YES 396 707 22 628.82 -283 738.13 0.18 8.64 0.87 6.38 -0.08 4.80 -0.24 -5.13 YES YES YES YES YES YES 396 707 24 623.35 -282 568.69 14 Self citations Priority lag Refusal Withdrawal Refusal Withdrawal 0.109 7.80 “X” from “pure X” no self -0.005 8.89 “X” from “pure X” self -0.025 4.38 “X” from “X and A” no self “X” from “X and A” self 0.003 1.79 “A” from “X and A” no self -0.099 -2.89 “A” from “X and A” self -0.280 -5.82 “A” from “pure A” no self -0.012 -1.15 “A” from “pure A” -0.006 -3.27 “X” X” self lag <=2 years 0.11 9.42 2 years 0.07 7.58 “X” from “pure “X andX”A”lag lag> <= 2 years -0.003 3.87 “X” from “X and A” lag > 2 years -0.01 3.79 “A” from “X and A” lag <= 2 years -0.21 -2.35 “A” from “X and A” lag > 2 years -0.12 -3.75 “A” from “pure A” <= 2 years 0.18 -1.18 “A” from “pure A” > 2 years -0.04 -1.69 Applicant countries YES YES YES YES Priority years YES YES YES YES Technology classes YES YES YES YES Observations 396 707 396 707 Wald Chi2 22 854.90 22 668.00 Log pseudolikelihood -283 607.75 -283 716.52 Cited withdrawals being more powerful when close invention date 15 Main findings Pure X have a stronger blocking impact than other cited patent applications Pure X withdrawn have the strongest blocking power among cited patent applications Further, withdrawn patent applications, of any kind, have the strongest blocking power within their groups 16 Scope and cost Claims (ln) IPC classes (ln) inventors (ln) Backward patent citations (ln) Science linkage Applicant country United States Japan Germany France United Kingdom Australia Switzerland China Korea Other EU15 countries All other countries Technology class Electronics Instruments Chemicals-materials Pharmaceuticals-biotech Industrial processes Machines, mechanics, transport purex v. all purex v. cited 1 2 pure x withdrawn v. pure x granted 3 0.6** 0.4** 0.1 0.5** 0.2 -0.2 -1.3** -0.6** 0.2 3.8** -0.3 0.4 -1.6** 2.4** 10.0** 0.3 0.5 -1.0** -1.3** -0.4 -0.7 -1.4** 0.8 -0.4 -1.3** -0.9** 1.5 5.1** 0.9 -0.4 2.7* 5.4** -0.3 10.3 4.3* 0.7 2.6* 12.3** -4.8* -1.2 -4.4 7.8** 11.6** 3.4 1.9 0.5 3.3 6.7* -0.2 0.5** 2.4** 3.6** 0.5** 0.0 -2.9** -1.1* 5.5** 14.3** -1.3** -2.6** 1.2 -0.5 0.1 6.2** -5.1** -5.8** Narrow inventions less number of inventors; cumulative inventions More likely to be filed by AngloSaxon applicants More frequently filed in pharma biotech 17 Conclusions Evidence of pre-emptive patenting o compromising patentability but not contributing to the art o not necessarily ending into a grant Evidence on the impact of withdrawn patents o (very) defensive patenting: applicants playing on uncertainty and asymmetric information to pre-empt entry into markets and technologies Refused patent applications have a profile “in between” granted and withdrawn applications Blocking strategies with pre-emptive orientation are more often found in biotech-pharma 18