PatentSim Study Revisited: Patent Expertise Equalizes Innovation in Patent and Non-Patent Systems

    By James DeGiulio --

Earlier this month, we reported on a study published last year in the The Columbia Science and Technology Law Review, which suggested a commons systems can generate significantly greater amounts of innovation, productivity, and social utility than currently predominating patent systems that combine both patent and open source protection for inventions ("Patents and the Regress of Useful Arts").  The authors of this study, Dr. Andrew W. Torrance (at right), Associate Professor at the University of Kansas School of Law and a Research Associate at the Biodiversity Institute at the University of Kansas, and Dr. Bill Tomlinson (below left), Assistant Professor in the Informatics Department of the Bren School of Information and Computer Sciences at the University of California-Irvine, have now tested their PatentSim software with individuals having patent "expertise," leading to some differing results.  Their newest study, published in the Southern Illinois University Law Journal, ("Patent Expertise and the Regress of Useful Arts") shows that there is no statistically significant difference in levels of innovation generated by expert users across the pure patent, patent/open source, and commons systems.

The study utilized a computer program, PatentSim, to simulate the behavior of inventors and competitors under conditions approximating patent and non-patent systems.  A more detailed description of the PatentSim software can be found in our previous report.

In this iteration of the study, the PatentSim subjects were second and third-year law students at the University of Kansas School of Law who had formally studied patent law and open innovation (including open source).  These subjects represented "expert" users of patent and open innovation systems.  This is in contrast to the original published study, which tested first-year law students with no knowledge of patent protection, open innovation, or any other intellectual property concepts.  Though the productivity and social utility levels (as defined by the authors) were consistent with the first study with uneducated subjects, the rates of innovation were noticeably different with expert users.

Innovation was defined as the mean number of unique inventions.  Graph 1 (below) illustrates the relative amounts of innovation generated in the Pure Patent, Patent/Open Source, and Pure Commons systems.  The data reveals that there is no significant difference in innovation (p=0.441) between the Pure Patent and Patent/Open Source systems, no significant difference in innovation (p=0.197) between the Pure Patent and Pure Commons systems, and no significant difference in innovation (p=0.802) between the Patent/Open Source and Pure Commons systems.

Productivity was defined as the mean number of total inventions created.  Graph 2 (below) illustrates the relative amounts of productivity generated in the Pure Patent, Patent/Open Source, and Pure Commons systems.  The data reveals that there is a significant difference in innovation (p=0.017) between the Pure Patent and Patent/Open Source systems, a highly significant difference in productivity (p=0.0006) between the Pure Patent and Pure Commons systems, and a highly significant difference in productivity (p=0.0032) between the Patent/Open Source and Pure Commons systems.

Social Utility was defined as the mean amount of money with which each subject ended each trial.  Graph 3 (below) illustrates the relative amounts of social utility generated in the Pure Patent, Patent/Open Source, and Pure Commons systems.  The data reveals that there is a significant difference in social utility (p=0.020) between the Pure Patent and Patent/Open Source systems, a highly significant difference in innovation (p=0.0000051) between the Pure Patent and Pure Commons systems, and a significant difference in innovation (p=0.0021) between the Patent/Open Source and Pure Commons systems.

In contrast to the authors' original highly-publicized study, where the rates of innovation for the commons system were considerably higher than the patent and patent/open source systems, the current study with educated subjects revealed no statistical difference in innovation rates.  This suggests, perhaps obviously, that some knowledge of intellectual property is required to effectively capture the innovation incentives that patent protection provides.  If the results of the simulation are to be believed, the current patent system encourages innovation just as much as other non-patent systems, provided subjects are aware of patents and how they function.

Interestingly, and not without some irony, the authors of the study have chosen patent protection for their PatentSim game rather than open sourcing.  As of the date of this post, their patent application (U.S. Application No. 12/189,770; U.S. Patent Application Publication No. 2009/0043552) is still pending and has yet to be examined.  The single independent claim reads as follows:

1.  A system for simulating or operating an intellectual property environment, said system comprising:
    an administrator module;
    a rules module; and
    a user interface;
    wherein said administrator module sets rules for said rules module either before and/or during the operation of the environment;
    wherein a user interfaces with said system through said user interface;
    wherein said user puts together an invention object, using two or more invention elements from an invention repository or some other source;
    wherein said user has an option to manufacture said invention object in said intellectual property environment; and
    wherein said user has an option to apply for a patent or other protection for said invention object in said intellectual property environment.

James DeGiulio has a doctorate in molecular biology and genetics from Northwestern University and is a third-year law student at the Northwestern University School of Law.  Dr. DeGiulio was a member of MBHB's 2009 class of summer associates, and he can be contacted at [email protected].

Comments

[Skeptical]

All the comments previously mentioned about the faultiness of the program still apply.

Changing the user set won't change the fact that the program cannot and does not reflect Real Life.

Nonetheless, it would be interesting to see how drastic a change in results might occur if the user pool was not students at all, but rather patent professionals in the appropriate field of art.

[Dale B. Halling]

The study is still flawed. If you did the same study with real property, you would get the same results. But there is no real world evidence that the world is better with real property law. In addition, the study’s definition of commons give the participants more control over their technology than any real world contract or other law allows.

[AK]

The study needs to be repeated with experienced executives who are responsible for decisions, not students with no accountability to shareholders or investors.

[6]

They should have added at the end of claim 1: wherein the current patent system is revealed as largely being a bunch of bunk.

or perhaps:
wherein said user realizes that currently patents have a negligable effect on innovation and starve the economy of profits; and wherein the user becomes concerned about this.

Oh and someone should tell them that "system" claims have to fit within one of the stat classes and cannot overlap two of them.

"an administrator module;"

is presumptively a structural limitation.

" wherein said user puts together an invention object, using two or more invention elements from an invention repository or some other source;"

Is obviously a method step.

"Nonetheless, it would be interesting to see how drastic a change in results might occur if the user pool was not students at all, but rather patent professionals in the appropriate field of art."

Innovations would go up by .00000000000001 and profits would dip by 500%+.

[Noise above Law]

Ouch Kev,

Trainwreck free from me - sorry that you will still have to put up with utter nonsense from 6.

All the other posts are correct - no matter who uses a flawed model, the results still are tainted. As it is, the simulation is but a game and there is no useful signal to be gleaned from the results.

[Bob]

The authors' definition of "innovation" is flawed. In the study, "innovation" is defined to be nearly synonymous with "invention." Few credible economists would define innovation in this limited way, as it fails to account for the expense of developing the invention into a marketable product. For example, there's a big difference between discovering that compound X inhibits enzyme Y and developing an oral dosage form for safely delivering an effective amount of compound X to a human being.

As Posner and Landes point out, the goal of patent protection is not merely to generate more inventions. Rather, it is to permit the inventor (or her assignee) to have a capitalizable property interest, which can be used to attract the funding necessary to transform the invention into an innovative product.

This study should be retitled: "If we pretend that patent law operates like copyright law, then we find that 'patent' protection does not generate significantly higher levels of invention." Doh!