Patents: Their Effectiveness and Role Wesley M. Cohen Fuqua School of Business Duke University November 14, 2003 Presentation in the Information Ecology Series of the Center for the Study of the Public Domain Collaborators: Ashish Arora, Marco Ceccagnoli, Akira Goto, Akiya Nagata, Richard R. Nelson, John P. Walsh Sources of support: Alfred P. Sloan Foundation, National Science Foundation, Center for Global Partnership

W.M. Cohen, "Patents: Their Effectiveness and Role" Background A strengthening and broadening of patent protection over the past 20 years 1982 Creation of CAFC and pro-patent trends in the courts Expansion of what is patentable and who can patent Dramatic growth in corporate patenting over past two decades, and growing emphasis on exploitation of IP among top management. W.M. Cohen, "Patents: Their Effectiveness and Role"

Cause for questioning both public and private policies Forty year empirical legacy: patents not central to protection in most U.S. industries (except drugs) Growing concern over impact of increasing “strategic” uses of patents, as in semiconductors, on innovation and competition Concern over patenting of research inputs on subsequent discovery Where innovation is cumulative, a foundational patent may block access to key building block to subsequent research (e.g., human embryonic stem cells) Anticommons: As number of patents reading on a product innovation grow, negotiations across rights-holders become more difficult, threatening development of worthwhile therapies. W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Overview Survey-based evidence on effectiveness of patents in protecting inventions Uses of patents across industries The “quid pro quo”- a cross-national study of patent disclosures and their importance for innovation Impact of patenting on R&D incentives in U.S. manufacturing W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Data Carnegie Mellon Survey (CMS) administered to R&D lab managers in the U.S. manufacturing sector in 1994. 1478 of 3,240 labs responded; 46% response rate (54% adjusted rate) U.S. sample broadly representative of firm size distribution Also reporting on comparable Japanese survey (643 of 1,219 labs responded; 53%) W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Patent effectiveness Context: A variety of mechanisms are employed by firms to protect their inventions Secrecy Lead time Complementary capabilities (e.g., sales and service, manufacturing) Patents Our measure of effectiveness of each of these % of firm’s innovations for which a mechanism was effective in protecting competitive advantage from that innovation W.M. Cohen, "Patents: Their Effectiveness and Role"

Relative effectiveness of mechanisms For product innovations Top mechanisms overall: Secrecy and lead time Patents least effective overall Relatively effective in drugs and med. equipment. Less effective in semiconductors & communications equipment Larger firms report patents to be more effective (r = .22 for product patents) Since mechanisms not mutually exclusive, “effectiveness” reflects centrality to strategy Do not conclude from our findings that patents do not stimulate R&D, even where less effective W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Effectiveness of Appropriability Mechanisms for Product and Process Innovations Mean % of Product/Process Innovations for Which Mechanism Considered “Effective” W.M. Cohen, "Patents: Their Effectiveness and Role"

Effectiveness of mechanisms in four industries W.M. Cohen, "Patents: Their Effectiveness and Role"

Limits on patent effectiveness: Why firms do not patent Most important reasons for not applying Demonstration of novelty (32%) Information disclosure (24%) Ease of inventing around (25%) Negative partial correlation between firm size and defense cost (r = -.18) as reason not to patent suggests why larger firms report (product) patents to be more effective. W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Need to consider how patents are used across industries to help understand how they may affect innovation. W.M. Cohen, "Patents: Their Effectiveness and Role"

Reasons to patent for aggregate sample Reasons (nonexclusive) Products Processes Prevent copying 96% 78% Patent blocking 82 64 Prevent suits 59 47 Use in negotiations 48 37 Enhance reputation 48 34 Licensing revenue 28 23 Measure performance 6 5 W.M. Cohen, "Patents: Their Effectiveness and Role"

A source of industry differences Our analysis of differences across industries in uses of patents builds on the following observation: When number of patents per commercializable innovation are great, unlikely that any one firm holds all necessary rights, fostering mutual dependence and associated behaviors W.M. Cohen, "Patents: Their Effectiveness and Role"

Distinguish between “complex” and “discrete” product industries Complex product industries: Where a product protected by numerous (e.g., hundreds) patents (e.g., computers, communications equipment), that is, more patents read on a product. Discrete product industries: Where a product protected by relatively few patents (e.g., drugs, chemicals) W.M. Cohen, "Patents: Their Effectiveness and Role"

Distinction implies different uses of patents by industry type Complex product industries: Patents used to block rival use of complements to Assure inclusion or “player” status in cross-licensing negotiations in order to gain access to rivals’ technologies Gain freedom to operate and design freedom via “mutually assured destruction” Discrete product industries: Patents used to block substitutes by creating patent “fences;” not to compel cross-licensing. W.M. Cohen, "Patents: Their Effectiveness and Role"

What uses did we find in discrete versus complex product industries?

Product patent uses across industry types Patent Uses Discrete Complex (patent applic. wtd.) Negotiations 33% 81% Player: Block and negots. 29 61 Fences: Block but not negs./lics. 45 11 There is, however, still substantial heterogeneity of uses even within industries. W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Social costs of “strategic” uses of patents in complex product industries Mutual dependence and associated “player strategy” may spawn patent non-cooperative portfolio races, resembling “arms races,” (cf. Hall and Ziedonis [2001] for semiconductors), raising cost of innovation Such strategic uses may increase litigation May yield “patent harvesting” where firms patent inventions that they would have generated anyway, suggesting patents may have little R&D incentive effect. Portfolio races may also deter entry and associated innovation The costs that strategic uses add to innovation may be troublesome if patents play little role in stimulating innovation to begin with. W.M. Cohen, "Patents: Their Effectiveness and Role"

But--Possible benefits Extensive cross-licensing among incumbents may also: Promote information sharing and associated efficiencies and complementarities across rival R&D efforts Avert license stacking and possible breakdowns in negotiations over rights due to large number of claimants (predicted by the anticommons argument). Because strategic uses yield profits, such uses may provide incentive to conduct R&D W.M. Cohen, "Patents: Their Effectiveness and Role"

Next study: The quid pro quo of disclosure Patents supposed to promote innovation in two ways: Appropriability Disclosure Disclosure often overlooked in U.S. Can patent disclosures importantly affect innovation? Compared patenting and related information flows in U.S. and Japan W.M. Cohen, "Patents: Their Effectiveness and Role"

Importance of R&D-related information flows Saves on duplicative R&D Complementarity effects, improving R&D productivity and incentives May promote entry But can diminish appropriability and associated incentive to invest in R&D R&D information flows (spillovers) contribute importantly to productivity growth W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Intraindustry R&D Information Flows and Appropriability in Japan and U.S. Measure #1: Whether information from rivals: Suggests new R&D projects: J > U Contributes to completion of existing projects: J >> U Measure #2: When aware of rival’s major R&D project? Japan: 44% of respondents aware prior to development U.S.: Only 16% aware prior to development Per more information flow, appropriability is less For unpatented and patented process and product innovations, imitation lags longer in U.S. by 40%-80%. W.M. Cohen, "Patents: Their Effectiveness and Role"

Why more information flow and less appropriability in Japan? Patents may be key Policy differences (at time of survey) Priority to first-to-file in Japan versus first-to-invent in U.S. In 1994, disclosure 18 months after application in Japan versus upon grant in U.S. Pre-grant (and pre-examination) opposition period in Japan prior to 1996. W.M. Cohen, "Patents: Their Effectiveness and Role"

Another reason for more disclosure due to patents in Japan Compared to U.S., in Japan patent claims interpreted more narrowly and there are fewer claims per patent (15 average in U.S. versus 5 in Japan)=> more patents per product. Implies mutual dependence across firms’ patent holdings more pervasive in Japan Uses of patents across all Japanese industries resemble strategic uses in U.S. complex product industries W.M. Cohen, "Patents: Their Effectiveness and Role"

Product patent uses across industry types in Japan Patent Uses Discrete Complex (patent applic. wtd.) Negotiations 84% 86% Player: Block and negotiations 83 81 Fences: Block but not negs./lics. 11 7 W.M. Cohen, "Patents: Their Effectiveness and Role"

Therefore, would expect patents to contribute much more to information flows across rivals in Japan

W.M. Cohen, "Patents: Their Effectiveness and Role" Finding: Patents are a much more important information source for R&D in Japan Compared channels of R&D information flow across rivals Top five (of 10) in both nations: patents, informal information exchange, products (e.g., reverse engineering), publications, public meetings/confs. One of top five channels stands out as much more important in Japan: patents First place in Japan, scoring 30% higher than next most highly ranked channel, while in middle of pack in U.S., and absolute score 70% greater than in U.S. W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Observations Not only technology, but policy and norms associated with number of patents per product affect uses of patents These uses, such as cross-licensing, affect information flows across rivals W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Implications Japanese experience suggests patent policy may significantly increase R&D spillovers. Also, such policies do not necessarily diminish R&D incentives In U.S. study, found positive effect of R&D info. flows across rivals on industry R&D Also Japan’s R&D intensity greater than that of U.S. on average—despite lower appropriability Patent reform efforts in U.S. should give at least equal time to disclosure W.M. Cohen, "Patents: Their Effectiveness and Role"

But do patents “promote the progress of science and the useful arts?” In light of finding that R&D relatively “unimportant” in protecting inventions across most U.S. industries, does patenting stimulate R&D, that is, do they increase the returns to R&D, even in such industries? W.M. Cohen, "Patents: Their Effectiveness and Role"

Impact of patents on R&D Direct appropriability incentive Indirect impact (via increasing effectiveness of rival patents) May decrease average payoff to R&D Cumulative technologies: may reduce incentives for follow-on R&D Spillovers: Patent disclosures may enhance R&D productivity. W.M. Cohen, "Patents: Their Effectiveness and Role"

Does patenting stimulate R&D? Conducted two-stage analysis to examine effect of patenting on product R&D Estimated average patent premium--the proportional increment to the value of innovations realized by patenting For “typical” innovations and for patented innovations (conditional premium) Analyzed impact of increasing patent premium on incumbents’ product R&D and patenting Ignored: Patents may reduce entry  reducing innovation Patents enhance efficiency of markets for technology which can increase R&D efficiency Clarify that the distribution is across innovations of a representative firm W.M. Cohen, "Patents: Their Effectiveness and Role"

Considerations in analysis Account for effect of both appropriability incentive and patent disclosure impact on R&D information flows Recognize R&D and patenting driven by many of same variables, and the one affects the other. W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Novel data CMU Survey provides data on key variables: Patent propensity-- % of (product) innovations that firms patent Patent applications – self-reported number of patent applications Patent effectiveness -- % of firm’s (product) innovations for which a patent was effective in protecting competitive advantage from that innovation’s R&D expenditure (for product innovation) W.M. Cohen, "Patents: Their Effectiveness and Role"

Patent Premium, selected industries Average Conditional Drugs and medicines 0.9 1.65 Biotech 0.97 1.66 Computers and other office equipment 0.7 1.58 Semiconductors 0.4 1.50 Medical instruments 1.12 1.72 Aggregate sample 0.49 1.53 Premium >1 => positive expected return to patenting Large variation across industries in patent premia Moreover, industries vary substantially in the variance of the p. premium distribution, as suggested by the within industry estimates and complementary analysis. That means for example that in drugs and biotech, conditional on patenting, the expected premium is very high, but you have a substantially higher probability of getting either very high premiums or very low premiums (this is consistent with the blockbuster drugs idea....) W.M. Cohen, "Patents: Their Effectiveness and Role"

Average and conditional patent premia, by industry W.M. Cohen, "Patents: Their Effectiveness and Role"

Elasticity w.r.t patent premium R&D Patent Appli-cations Patent apps/ R&D Drugs and medicines 0.96 1.4 0.45 Biotech 1.02 1.46 0.44 Computers and other office equip 0.74 1.29 0.55 Semiconductors 0.46 1.07 0.61 Medical instruments 1.18 1.48 0.30 Aggregate sample 0.57 1.16 0.59 W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Implications Positive effect of patenting on R&D overall Even in industries such as semiconductors where patents much less effective than other mechanisms and strategic uses pervasive. We find some “harvesting”--the patenting of inventions that would have been generated anyway--in all industries, but especially where patent premium is lowest Our study did not consider: Negative effects of patents on entry Positive effects due to markets for technology, W.M. Cohen, "Patents: Their Effectiveness and Role"

W.M. Cohen, "Patents: Their Effectiveness and Role" Conclusions Many ways to protect inventions While patents not as featured as other mechanisms, they stimulate R&D broadly, though more in some industries than others Patent disclosures can contribute importantly to R&D information flows, but apparently little such effect in U.S. Pervasive “player” strategy raises concerns (i.e., costs and barrier to entry) for both managers and policymakers To understand impact of patents on innovation, need more work on effects of patents on industry entry and markets for technology W.M. Cohen, "Patents: Their Effectiveness and Role"