the tragedy of under-innovation: intellectual property...

26
Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix The Tragedy of Under-Innovation: Intellectual Property Rights and the Anticommons in Biofuel Technologies Annabelle Berklund Colorado State University [email protected] October 31, 2014 1 / 26

Upload: others

Post on 03-Aug-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

The Tragedy of Under-Innovation:Intellectual Property Rights and the Anticommons

in Biofuel Technologies

Annabelle Berklund

Colorado State University

[email protected]

October 31, 2014

1 / 26

Page 2: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Roadmap

1 Motivation

2 BackgroundCommons and AnticommonsExamplesDetour

Innovation and Patents

3 Modeling the AnticommonsExisting LiteratureModel SpecificationTheoretical Conclusions

4 Empirical TestBiofuelsBiofuel Patent Data

5 Conclusions

6 Appendix

2 / 26

Page 3: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Motivations

With large R&D incentives for Biofuels production, why haven’tmore made it to market?

RFS, MTBE ban, International PoliciesLarge amounts of Public and Private R&D funding

Patents, a form of intellectual property rights (IPR), have beenscrutinized in the media recently, are they part of biofuel’s problem?

Research Question:Tragedy of the Anticommons in IPR

Have fragmented patents rights, which increase transaction costs,inhibited biofuels innovation?

3 / 26

Page 4: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Commons and Anticommons

Commons and Anticommons

The Tragedy of the Commons

Lack of property rights →overuse

The Tragedy of the Anticommons

Overlapping/fragmented propertyrights → suboptimal use

Why is this a problem?Negative externalities

1 Reduces the use value of allother owners rights

2 Under-use today → futuregrowth consequences

Prisoner’s dilemma

4 / 26

Page 5: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Examples

Examples of the Anticommons

Complementary oligopoly

Two independent firms hold monopolies on complementaryinputs (Cournot, 1963)

Quaker Oats Klondike Big Inch Land PR Ploy

Eastern European Store Fronts post 1991 (Heller, 1998)

River Regulation

Overlapping water regulation agencies in the U.S. →suboptimal use of river-basin (Kosnik, 2012)

Fisheries (Filipe, 2011)

Intellectual Property Rights (Heller and Eisenberg, 1998;Bessen and Maskin, 2009)

5 / 26

Page 6: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Detour

The U.S. Patent System

Patent term in U.S. is 20years

The average application topublication lag is 18 monthswith great variation.

Complex application process→ need for intermediary

U.S. Patent examiners haveperverse incentives →increased patenting in U.S.(Shapiro, 2001; Lei &Wright, 2000).

Bayh Dole Act, 1989 →increased universitypatenting

6 / 26

Page 7: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Detour

Innovation and Patents

Patents create excitability over the use of claimed ideas

Allows inventors to earn temporary monopoly profits to recoup R&Dexpenses

Positive Impact

Patent allows firm to act as a gatekeeper and supports follow-oninnovations(Kitch, 1977; Arora, 1995).

Negative Impact

Asymmetric information between upstream and downstreaminnovators (Bessen, 2004; Bessen & Maskin, 2009).

Divergent expectations about payoffs from patents → inefficientlicense bargaining (Galasso, 2012; Lacetera & Zirulia, 2012; Priest& Kliein, 1984).

Increased transaction costs associated with acquiring licenses toexisting patent rights slows down innovation (Heller, 1998).

7 / 26

Page 8: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Existing Literature

Existing Literature

Anticommons in IPR

Largely addresses bargaining and transaction costs for patent rights

post invention discovery, by private firms who are assumed to

appropriate all patent royalties (Buchanan & Yoon, 2000; Bessen &

Maskin, 2009; Comino, et. al, 2011; D’Agata, 2012; Schulz, et. al, 2002).

Principal-Agent Models of Researcher Incentives

Private: Researchers of various types exert effort based on theirfirm’s incentive schemes and reservation utility (Dasgupta &Stiglitz,1980; Aghion, et. al, 2000; Stern, 2006; Lissoni, et. al, 2013)

Public: Researchers allocate efforts towards basic or applied researchprojects, constrained by university incentive structure (Lach andSchankerman, 2008; Huffman & Just, 2000)

8 / 26

Page 9: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Existing Literature

Drawbacks and Solutions

Drawbacks of Existing Literature

Ex post bargaining doesn’t address the question of whether ornot a researcher embarks on a research trajectory ex ante,given their knowledge of the existing patents in place, andhow institutional settings change these outcomes.

To my knowledge no comprehensive model of public andprivate research institutional constraints exists

Proposed solutions to the Anticommons in IPR include:

Patent pooling, licensing, acquisitions, partnerships andcollaboration, and legal intervention (Shapiro, 2001;Audretsch & Feldman, 2002)

All of these require some monetary or effort costs→transaction cost

9 / 26

Page 10: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Model Specification

Goals for Modeling Researcher Incentives

Researchers constrained by their institutional setting exerteffort on R&D.

Projects have an underlying probability of success

Conditional on market conditionsResearchers knowledge of this probability is incompleteThe IPR surrounding the existing knowledge stock changes theprobability and costs of commercialization.

Researchers gain new knowledge about the profitability of anidea each period as more information is acquired → Bayesianupdating.

10 / 26

Page 11: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Model Specification

Incentives and Invention in Universities

Lach and Schankerman’s 2008 (RAND) paper addresses several ofthe researcher incentives I aim to model, but in the context ofuniversity Rresearchers, their model assumes:

Researchers allocate time to basic, applied, or commercialresearch to maximize expected utilityShadow prices for effort levels are determined by the universitythrough their incentive scheme.University constraints, such as royalty shares, changeresearcher incentivesEffectiveness of the technology licensing office determinesfaculty royalty rates.

My model generalizes Lach and Schankerman’s model toinclude: the possibility of both public and private research

principals, an IPR fragmentation cost function, and Bayesianupdating.

11 / 26

Page 12: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Model Specification

Effort

e: effort towards basic researchz : effort towards new applied researchq: effort towards commercialization of applied research

12 / 26

Page 13: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Model Specification

Modeling Researcher Incentives Mathematically

Researchers who are constrained by their institutional settingschoose their level of effort towards a vector of research efforts(e, z , q) to maximize their expected utility:

E (U(e, z , q)|s, θ, γ)

Where:s = share of licensing revenue to inventor, ε{0, 1}θ = effectiveness of the IP intermediary ε{0, 1}γ = fragmentation index ε{0, 1}

U(e, z , q) = V (s ∗ r(z , q), p(e, z , q))− C (e, q, z)

C (e, z , q) is a convex effort cost functionp(e, z , q) is a concave publications production functionr(z , q) is revenue from individual’s research efforts

13 / 26

Page 14: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Model Specification

Expected Revenue per Researcher

Ideas are patented if: θυ > υ + c(γ)c(γ) is a convex cost function, with 0 ≤ γ ≤ 1γ = 1 represents sole ownership of IPRγ = 0 represents complete fragmentation of IPRn(z) is an increasing and concave function for the number of newapplied research projectsEffort q produces an invention with potential commercial valueυ(q) = ψ(q)ε

ψ(q) is increasing and concaveε is an independent stochastic shock observed after effortchoices are made, with distribution function G (ε)

Expected patent profits per researcher are then:

r(z , q) = θn(z)ψ(q)∞∫

υ+c(γ)θψ(q)

εdG (ε).

s = share of licensing revenue to inventorθ = Effectiveness of the IP intermediaryγ = fragmentation index

14 / 26

Page 15: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Theoretical Conclusions

Model Conclusions

Given our First Order Conditions:

V2pe − Ce = 0sV1rz + V2pz − Cz = 0sV1rq + V2pq − Cq = 0

it can be shown that increased fragmentation of IPR decreasesresearcher’s applied and commercial efforts:

∂z∂γ>0, ∂q

∂γ>0

Note: the specific distribution function G (ε) determines the magnitudeof these changes.Next Step: Add Bayesian Updating following Weiler, et. al (2008) suchthat:

E [U(ri,j |θi , γj)]

where researcher’s knowledge of γj is updated each period, contingentupon the effectiveness of the IP intermediary at IDing the existing IPR.

15 / 26

Page 16: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuels

Application to Biofuels

Strong innovation incentives for biofuels arose with the increasingthreat of climate change, national security, and rising gas prices.

Gas Price Instability: Oil Embargo, Increased NG Production, Carbon Tax andTradingU.S. Policies: MTBE Ban, Clean Air Act, Renewable Fuel Standard

Despite these incentives we haven’t reached 2nd genfuel goals:

16 / 26

Page 17: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuel Patent Data

Data: Biofuels Patents

Our constructed data set contains:

28,776 Patent Families Internationally over 60 years7,342 U.S. Patents

We then break the data into 6 technology spaces:

17 / 26

Page 18: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuel Patent Data

Summary Statistics and Trends

18 / 26

Page 19: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuel Patent Data

International Patent Family Trends

19 / 26

Page 20: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuel Patent Data

Technology Trends by Fuel Type

20 / 26

Page 21: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Biofuel Patent Data

Empirically Testing the Anticommons

Empirical Question:

Given an existing invention, what is the probability a follow-oninvention exists given the fragmentation of IPR and other inventioncharacteristics?

Pr(Yi ) = β0 + β1~x1,i + β2~x2,i + β3~x3,i + β4~x4,i + β5~x5,i

where:

yi =

{1 if a follow-in invention exists;0 otherwise.

~x1: density/fragmentation of technology space~x2: inventor characteristics~x3: assignee characteristics~x4: backward citations and number of claims fixed effects~x5: includes controls for yearly variation

21 / 26

Page 22: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Next Steps

Bayesian Updating

Fragmentation Index

Forward and backward missing citation identification (Lei,2014)Publication data along a technology trajectoryPatent Ownership dynamics/concentration ratio along apathway

Researcher and Institution Characteristics

Geography and time fixed effects

22 / 26

Page 23: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Questions/Comments?

Happy Halloween!

23 / 26

Page 24: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

References

Aghion, P., Dewatripont, M., and Stein, J.C., (2008). Academic freedom, private sector focus, and the process ofinnovation. The RAND Journal of Economics, 39(3), 617-635Audretsch, D.B. & Feldman, M.P. (May 2003). Knowledge Spillovers and the Geography of Innovation. Handbookof Urban and Regional Economics, 4, 1-41.Bessen, J., & Maskin, E. (Jan 2009). Sequential Innovation, Patents, and Imitation. RAND Journal of Economics,40 (4), 611-635.Buchanan, James, and Yong Yoon. Symmetric Tragedies: Commons and Anticommons, Apr 2000. Journal of Lawand Economics: 43(1): 1-13.Comino, S., Manenti F.M., Nicol, A. (Nov 2011). Ex-ante licensing in sequential innovations, Games and EconomicBehavior,73(2), 388-4.Comino, S., F. Manenti, & A. Nicol. (Nov 2011). Ex-ante licensing in sequential innovation. Games and EconomicBehavior, 73(2), 388-401.DAgata, Antonio. Geometry of Cournot-Nash Equilibrium with Application to Commons and Anticommons, 2010.The Journal of Economic Education: 41(2): 169-176.Filipe, J.A., Ferreira, M.A., Coelho, M. & Pedro, I. (Jul 2011). Modeling Anti-Commons. The Case of Fisheries.International Journal of Academic Research, 3(4), 456-460.Fuglie, K.O., Heisey, P.W., King, J.L., Pray, C.E., Day-Rubenstein, K., Schimmelpfennig, D., Wang, S.L., and R.Karmarkar-Deshmukh. Research Investments and Market Structure in the Food Processing, Agricultural Input, andBiofuel Industries Worldwide, Dec, 2011. United States Department of Agriculture, Economic Research Service,Report number 130.Griliches, Z. (1990). Patent Statistics as Economic Indicators: A Survey. Journal of Economic Literature, 28 (4),1661-1707

24 / 26

Page 25: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

References Cont.

Heller, Michael. The Tragedy of the Anticommons: property in the Transition from Marx to Markets, Jan 1998.Harvard Law Review: 111 (3): 621-688.Heller, M.A. & Eisenberg, R.S. (1998). Can Patents Deter Innovation? The Anticommons in Biomedical Research.Science, 280, 698-701.Huang, K. & Murray, F. (2009). Does Patent Strategy Shape the Long-Run Supply of Public Knowledge?Evidence from Human Genetics. Academy of Management Journal, 52 (6), 1193-1221.Huffman, W.E. & Just, R.E. (2000). Setting Efficient Incentives for Agricultural Research: Lessons fromPrincipal-Agent Theory. American Journal of Agricultural Ecoonomcis, 82(4), 828-841.Kosnik, L. (2012). The anticommons and the environment. Journal of Environmental Management 101, 206217.Lach, S. & Schankerman, M., (2008). Incentives and Invention in Universities. The RAND Journal of Economics,39(2), 403-433.// Lacetera, N. & Zirulia, L. (2012). Individual preferences, organization, and competition in amodel of R&D incentive provision. Journal of Economic Behavior & Organization, 84, 550-570. Lissoni, F, Montobbio, F, Zirulia, L. (2013). Inventorship and authorship as attribution rights: An enquiry intothe economics of science credit. Journal of Economic Behavior & Organization, 95, 49-69.Lei, Zhen and Brian Wright. Why Weak Patents? Rational Ignorance or Pro-Consumer Tilt?. Agricultural andApplied Economics Association in its series 2009 Annual Meeting, July 26-28, 2009, Milwaukee, Wisconsin withnumber 49279. Shapiro, C. (Jan 2001). Navigating the Patent Thicket: Cross Licenses, Patent Pools, andStandard Setting. Innovation Policy and the Economy, 1, 119-150.Stern, S. (2004). DO Scientists Pay to Be Scientists? Management Science, 50(6), 835-853.

25 / 26

Page 26: The Tragedy of Under-Innovation: Intellectual Property ...annabelle.berklund.com/uploads/3/7/8/1/37816803/... · Prisoner’s dilemma 4/26. Motivation Background Modeling the Anticommons

Motivation Background Modeling the Anticommons Empirical Test Conclusions Appendix

Patent Family Composition

A patent family groups allapplications andpublications for the sameinvention.

Domestic (SingleJurisdiction) PatentFamily:

all filings made inthe same country

International(Multi-Jurisdiction)Patent Family:

filings made inmultiple countries

26 / 26