Section 2 Symposium: Herbert Hovenkamp on Patents and Exclusionary Practices

Herbert Hovenkamp —  6 May 2009

hovenkampHerbert Hovenkamp is Professor of Law at The University of Iowa College of Law.

One interesting aspect of the DOJ Report on Section 2 is the scant, episodic treatment of IP issues. The Report rejects the presumption of market power for patent ties (p. 81); has a very brief discussion of refusal to license patented parts in which it properly rejects the reasoning of the Ninth Circuit’s Kodak decision and aligns itself with the Federal Circuit’s Xerox decision (p. 121-122). The Walker Process case, which held that an infringement action based on an improperly acquired and unenforceable patent could violate §2, is cited in a footnote, and only for the proposition that market power is required in a §2 case (p. 25 n. 53). Finally, the Report contains a brief discussion of the presence of intellectual property in measuring incremental cost for purposes of analyzing predatory pricing (p. 63).

I suggest that the Antitrust Division and the case law develop a theory about the unreasonably exclusionary use of patents that generally divides the territory between pre-issuance and post-issuance conduct. This division has much less to do with the exclusionary power of patents than with the presence or absence of a relevant regulatory agency. The patenting process is characterized by very intensive agency regulation up to the time that a patent issues, but almost no regulation thereafter. This suggests a rather sharp line between pre-issuance and post-issuance conduct. The one exception is for pre-issuance conduct that the agency did not supervise adequately, mainly because it was never presented to the agency in the first place. This is consistent with the general theory of “implied immunity,” under which regulated conduct is immune from the antitrust laws only to the extent that it is within the jurisdiction of a federal agency, was actually made known to the agency, and assessed under criteria that took competitive conseqeunces into account.For example, applying this distinction would have led to a better analysis in the Federal Circuit’s decision in Dippin’ Dots, Inc. v. Mosey, 476 F.3d 1337, 1346-1347 (Fed. Cir. 2007), cert. denied, 128 S.Ct. 375 (2007) which rejected on conduct grounds an antitrust challenge to a patent that was unenforceable because the patent applicant had lied about substantial sales that would have prevented patentability under the “on sale” bar. 35 U.S.C. §102b. The Federal Circuit held that while a blatant lie about undisclosed prior sales rendered the patent unenforceable, permitting an antitrust action to proceed would require evidence of some kind of fraudulent or perhaps anticompetitive conduct in addition to the sworn falsehood itself. What the court neglected, however, was that there was more. The fraud before the PTO was certainly enough to invalidate the patent, but in this case the patentee had also filed infringement suits a decade after the patent issued, when any information about pre-application sales would be even more difficult to discover. Policing the integrity of the patent issuing process befalls the PTO and the Federal Circuit as supervisor of its conduct. But policing post-issuance practices such as the filing of improper infringement actions befalls antitrust. Of course, one might doubt that a plaintiff could make out the structural prerequisites for monopolization in a case about an ice cream formula, but that is a different issue.

One irony of competition policy in innovation intensive markets is that patent law’s own problems of poorly defined boundaries and inadequate notice not only make the patent system less workable, they also undermine effective use of competition law to pursue anticompetitive innovation restraints. For example, the Walker Process doctrine condemns improperly brought patent infringement suits as monopolistic, and excessive use of infringement claims is an important mechanism by which dominant firms restrain the innovations of rivals. But effective use of the doctrine requires a court to determine that no reasonable, well informed person would have brought the infringement action in question. Patent boundaries are so poorly defined that one can rarely state that conclusion with confidence except in cases of clearly improper conduct. As a result the great majority of such claims are rejected.

The holdup patent continuation cases provide another good example of the difference between immune and non-immune conduct. As a general proposition it is not an antitrust violation for a patentee to surprise an innocent infringer with a patent, even if the patent claims in question were drafted and approved after the infringer’s technology was developed. Kingsdown Medical Consultants v. Hollister, 863 F.2d 867 (Fed. Cir. 1988), cert. denied, 490 U.S. 1067 (1989) Existing patent law permits patent continuations, or “late claiming,” with retroactive application of the claim to the date that the original patent application was filed. This law makes it possible for an applicant to file a patent application in January, 2004, observe a rival’s technology on the market in March of 2004, and then write a claim on that technology for submission in a continuation application. I would greatly prefer that the system were different and that patent enforceability be more dependent on priority in the common law property sense. But it is not antitrust’s purpose to interfere with the prerogatives of federal reglatory agencies that are enforcing regimes established by Congress.

Deception of others in order to induce them to adopt technology that conflicts with the patent is a different matter. Even if the deception occurs prior to patent issuance, as in Qualcomm, Inc. v. Broadcom Corp., 548 F.3d 1004 (Fed.Cir. 2008), it is generally not brought to the attention of the regulatory agency. That is, the PTO may well know that the applicant is writing continuation claims, but it does not know whether the applicant is lying to others, such as an SSO, about what it is doing. In that case an antitrust inquiry is appropriate. Once again, this is not the same thing as saying the firm has violated the antitrust laws. The conduct and structural requirements for a Section 2 violation must still be made out; however, the immunity for activity adequately regulated by a federal agency has been lost.

6 responses to Section 2 Symposium: Herbert Hovenkamp on Patents and Exclusionary Practices


    The antitrust community seems to have a severe case of “To a man with a hammer everything looks like a nail”.


    The specification is not suppose to define the boundaries of an invention. That is the purpose of the claims. The PTO rules attempting to limit continuation are not laudable, they are an attempt to further weaken our patent laws. These rules will punish inventor’s of ground breaking technology by limiting their access to the patent system.

    The concerns about Rambus should properly fall within contract law not antitrust law. There are plenty of remedies for improperly obtained patents, including rule 11 sanctions. Antitrust law has a sad history of being more about politics than law. Infusing more antitrust law into patent law disputes will decrease the incentives for innovating by making validity and enforcement of patents more uncertain.

    Herb Hovenkamp 8 May 2009 at 7:55 am

    The comment is factually correct, but the fact is that if the original specification really did specify the precise boundaries of a patent and gave clear notice of what infringed and what did not, then claims would not be necessary. The reason we have disputes such as Rambus is because a specification does not necessarily give adequate notice of an invention’s boundaries. I have not argued, although others have, that patent continuations be abolished. But the fact is that they can be abused and limitations need to be imposed on the abuses. For example, late claims approved after a prescribed delay could be given prospective force only. The PTO rules regarding continuations that the Federal Circuit rejected a month ago in Tafas v. Doll, 559 F.3d 1345 (Fed. Cir. 2009), were a laudable attempt in this direction. Hopefully Congress will address the issue.


    “This law makes it possible for an applicant to file a patent application in January, 2004, observe a rival’s technology on the market in March of 2004, and then write a claim on that technology for submission in a continuation application.”

    Mr. Hovencamp left out an important condition at the end of the above sentence – “but only if supported by the patent’s original specification”. Suprisingly there is no mention at all of a patent’s specification which one would expect when discussing patents, continuation applications and priority dates.

    There are perfectly valid reasons why continuation applications exist and Mr. Hovencamp does nothing to acknowledge this. Indeed the blog post does much to confuse the issue by implying patent owners can arbitrarily add claims to cover competitor’s products. If one is going to suggest the removal of continuation applications one should consider the lost benefits plus the secondary implications of such a change (ie asking “and then what”).

    Herb Hovenkamp 7 May 2009 at 2:24 am

    Mr. Halling suggests that claim construction lies at the heart of most patent/antitrust disputes and that the problem is minimal because claims clearly define patent boundaries. While we might disagree about the latter proposition, the fact is that very few patent/antitrust disputes involve claim construction. A Walker Process case could conceivably arise on the basis of an unreasonably broad construction of a claim, but in that case the infringement plaintiff and defendant would have access to the same information. The much more serious ambiguity problems arise from defects that are not on the record, such as prior use or sales that are typically known only to the patent applicant. The problem in the patent holdup cases is frequently that the claims have not yet been approved at the time the deception occurs, because they are subject to ongoing continuation applications.


    Mr. Hovencamp is poorly informed about the boundaries of patents. Patent boundaries are clearly defined by the claims. The courts have consistently narrowed the scope of the claims, which define the invention. As a result, literal interpretations of the claims are the only interpretations that survive.

    To suggest antitrust law has well defined boundaries and patents do not is the height of absurdity.