INTRODUCTION 838 I. BACKGROUND 842 A. Why the Value of Litigation Depends on Prior Art 842 B. Prior Work 848 II. METHODOLOGY 851 A. Collection and Coding 851 B. Reproducibility and Selection Effects 855 III. PRIOR ART IN THE DISTRICT COURT 858 A. Anticipation vs. Obviousness 858 B. Categories of Prior Art 860 C. Printed Publications 862 D. US Patents as Prior Art 865 E. Activity 869 F. Foreign Patents 873 IV. POTENTIAL BENEFITS OF REPLACING LITIGATION WITH EXAMINATION 876 A. Prior Aft and Examination Time 878 B. Examining for Anticipation 881 C. Examining for Obviousness 883 V. REASONS FOR INFERENTIAL CAUTION AND FUTURE WORK 888 A. Invalidity Basis is Changing Over Time 888 B. IPR Selection Effects 890 C. Changes in Prior Art Rules 891 CONCLUSION 892 INTRODUCTION
Patent litigation is complicated, fact intensive, and expensive. Nonetheless, for at least fifty years, the Supreme Court has pursued an express policy of welcoming patent disputes into court and discouraging patent settlement. (1) Citing the public's "paramount interest" in patent adjudication, (2) the Court has abrogated traditional doctrines like licensee estoppel, (3) applied nonmutual claim preclusion to patentees, (4) changed the rules of appellate court procedure, (5) and subjected settling patent litigants to potential antitrust liability. (6) All of this to drum up more patent litigation and guide it toward final judgment.
Standing alone, this policy is a bit puzzling, occurring as it does within a court system that typically encourages settlement. (7) But the puzzle is only compounded when one realizes just how exceptional this prolitigation posture makes patent law. Among the many kinds of civil cases one might think serve a public interest--First Amendment litigation, civil rights suits, challenges to agency actions, and so on--only patent litigation has been anointed for judicial encouragement. (8) What is it that makes patent cases so special, that they should be encouraged while almost every other type of civil suit is quietly counseled out of the courtroom?
While the Supreme Court's answers to that question have been terse, (9) scholars have explored the public benefits of patent litigation in some detail. (10) At a high level, there are two distinct theories of how patent litigation might serve the public interest. The first is that judicial scrutiny of an issued patent may end a "patent monopoly," thereby increasing competition to the benefit of consumers. (11) The second is that revoking an undeserved patent may send a signal to prospective inventors about the need to comply with the substantive demands of patent law, thereby increasing the incentivizing power of the patent system in the future. (12) Though it is possible for patent litigation to yield other benefits, these are the two patent-specific theories that have been most extensively developed and that are consistently invoked to support policies of encouraging patent litigation. (13)
But both of these are only theories. To actually justify our longstanding policy of encouraging patent litigation, we need to know quite a bit about the particulars of how patent cases play out in practice. Under either theory, the public benefits of patent litigation can turn quite dramatically depending on the circumstances of a given case. (14) For example, it is possible that most patent cases vindicate the public domain based on prior art that the purported inventor should have known about. But it is also possible that many cases actually involve faultless inventors embroiled in fact-bound disputes with primarily private significance. To fully assess the public interest in patent litigation, we need to know something about the evidence underlying judicial determinations of invalidity. (15)
To this point, we have known very little. While prior studies have reported the raw number of patents invalidated under various statutory provisions, (16) no study has examined the evidence supporting those conclusions. We know, for example, that among all patent cases filed in 2008 and 2009, 154 summary judgment motions were filed on the basis of anticipation, (17) and that 31 of those motions ultimately succeeded. (18) But we have no idea whether those invalidations involved a strong public interest, or whether they would be better characterized as private fights over the technicalities of patent law. (19) So while the value of patent litigation has been vigorously theorized, the data necessary to test those theories has so far been lacking.
This Article provides the first comprehensive look at the evidence district courts rely upon when invalidating patents. To construct our dataset, we collected every district court ruling, verdict form, and opinion (whether reported or unreported) invalidating a patent claim over a recent six-and-a-half-year period. We then coded individual invalidation events on a reference-by-reference, claim-by-claim basis. In the end, we observed 3320 invalidation events based on 817 distinct prior art references.
This data sheds critical light on the public value of patent litigation. And it turns out to reveal starkly different answers depending on the legal issue in a case. Invalidations based on anticipation bear many indicia of publicly beneficial error correction. Encouragingly, invalidity based on obscure prior art appears to be rare. This suggests that the bulk of anticipation invalidations have the potential to incentivize future inventors.
But the story is not so rosy for cases of obviousness. A significant number of obviousness invalidations rely on prior art that would have been difficult or impossible to find at the time of invention. This suggests that many obviousness invalidations lack the potential for incentivization found in cases of anticipation and may even be disincentivizing for inventors who lose patents on the basis of facts they simply could not have known. This does not mean litigation of these cases is necessarily undesirable--these cases may yet be justified under some other theory of how patent litigation benefits the public. (20) But, at a minimum, the data fails to show that the prevailing policy of encouraging obviousness litigation has a sound empirical basis.
Aside from the public value of patent litigation, this data is relevant to another set of policy questions as well: whether it would be cost effective to allocate more time and money to the initial process of patent examination. For more than a decade, the consensus view--forged by Mark Lemley's highly influential article, Rational Ignorance at the Patent Office (21)--has been that increased investment in patent examination would cost more than it would save. (22) In recent years, scholars have questioned this premise, suggesting that better patent examination could justify its costs by significantly reducing the need for patent litigation. (23) We offer a new perspective on this debate by assessing the feasibility of patent examiners actually finding the prior art that later proves fatal in court. We conclude that Lemley's original predictions were implausibly low, but that the latest scholarship rebutting Lemley's work appears quite optimistic. While increased investment in examination may still turn out to be cost justified, our findings suggest that litigation savings will not be as large as recent scholarship has predicted.
To be clear from the outset, this is a study of district court litigation as an error-correction tool, not a study of patent quality. All of our observations are drawn from patents that were litigated to a judicial decision, and these may not be representative of all patents in general or even litigated patents in general. Because only a tiny fraction of issued patents are ever asserted, and only a tiny fraction of those are litigated to a decision, the potential for selection effects is significant. (24) This data sheds light on the nature of the error correction afforded by district courts and the feasibility of providing that same error correction through alternative mechanisms, but should not be used to make any inferences about the nature of patent invalidity overall.
This Article proceeds in five parts. Part I explains why the public benefits of patent litigation depend on the prior art at issue in a case and briefly recounts prior studies of district court patent litigation. Part II introduces the sources and methodology used in this study. Part III presents our findings related to the public value of patent litigation. Part IV explores the likelihood that increased investment in examination could avoid the need for patent adjudication. Part V discusses several reasons for inferential caution and suggests avenues for future work. A brief conclusion follows.
Before turning to the methodology of this study, two pieces of background are in order. Section A explains why the value of patent litigation can depend on the prior art at issue in a case. Section B then briefly discusses several prior studies of district court patent litigation and explains why they are insufficient to address the questions of interest here.
Why the Value of Litigation Depends on Prior Art
Courts and scholars have long heralded the public benefits of patent litigation. For nearly a century, the public's "paramount interest" in ensuring that patents are "free from fraud" and "kept within their legitimate scope" was accepted as a self-evident proposition that required neither proof nor elaboration. (25) Today, courts, commentators, and litigants often cite the public benefits of patent litigation as if they were simply an accepted fact. (26)
It is one thing to tout the benefits of patent litigation; it is another to propose a defensible theory of how adjudication of patent cases might yield public benefits not found in other forms of civil litigation. Prior scholarship has primarily focused on two distinct theories of how patent cases might uniquely serve...