Extreme value or trolls on top? The characteristics of the most-litigated patents.

AuthorAllison, John R.

INTRODUCTION I. STUDY DESIGN A. Technology Areas B. Industry Areas C. Entity Status D. Statistical Analysis II. RESULTS A. The Characteristics of the Most-Litigated Patents B. The Technologies and Industries of the Most-Litigated Patents C. The Owners of the Most-Litigated Patents D. The Katz Effect E. Logistic Regression III. IMPLICATIONS A. Extreme Value? B. Trolls on Top? APPENDIX: THE MOST-LITIGATED PATENTS INTRODUCTION

Patent reform has become, perhaps improbably, one of the most contentious issues facing Congress and the courts over the past six years. The fights range across a number of major issues, which not only separate patent owners from patent defendants and those who believe in innovation incentives from those who believe in market competition, but also divide patent owners themselves along both industry and technology lines. Advocates on both sides paint seemingly irreconcilable pictures of the patent system, either as a stable system with clearly defined legal rights essential to innovation or as a system rampant with litigation abuse by "patent trolls" who use the legal system to divert money from innovative companies. (1)

Far too much of this debate is based on anecdote and assumption, not real data. Pharmaceutical patent owners assume that most of the world works the way their industry does; so, too, do information technology (IT) companies. Patent trolls are variously portrayed as responsible for the majority of all patent lawsuits, for no more than two percent, or as mythical creatures that do not actually exist. (2)

The opening of the Stanford IP Litigation Clearinghouse in December 2008 (3) allows us to collect data that give a unique perspective on many of these debates. Using this data, we identify the patents litigated most frequently between 2000 and 2007 and compare those patents to a control set of patents that have been litigated only once in that period. The results are startling. The most-litigated patents are far more likely to be software and telecommunications patents, not mechanical or other types of patents. They are significantly different from once-litigated patents in ways that signal their value up front. And they are disproportionately owned by nonpracticing entities (i.e., "trolls"). The results do not answer all of the policy questions; we offer only one important piece of a larger mosaic. But our findings have significant implications for debates over patent reform, since we show both that the most-litigated patents are the most valuable ones (4) and that they are most commonly in the hands of companies other than the ones building new products.

In Part I we describe our study. In Part II we report our results. Part III discusses the implications of our findings.

  1. STUDY DESIGN

    Only about 1.5% of all patents are ever litigated in court. (5) The majority of patents are worth no more than a few thousand dollars; (6) litigated patents are almost by definition extreme outliers, since the parties are willing to spend millions of dollars per side in legal fees in order to litigate them. (7) In prior work, two of the authors demonstrated that litigated patents have significantly different characteristics than other patents. (8) They include more claims, cite more prior art, are cited more often by later patents, file more continuation applications, and come from larger "families" of patents. (9) They are also concentrated in certain industries. For example, semiconductor patents are particularly unlikely to be litigated because the industry is concentrated and cross-licenses are common. (10) Many of these characteristics are within the control of the patent applicant, and most are known by the time the patent issues. Allison et al. suggest that these characteristics are evidence of the private value of patents. (11)

    That prior work depended significantly on a randomly selected sample of cases actually litigated. The development of the Stanford IP Litigation Clearinghouse in December 2008 opened up a second alternative. The Clearinghouse collects every patent-infringement lawsuit filed since January 1, 2000, in searchable format, and links those suits to the patents in suit.(12) Using that database, we identified every patent that has been litigated eight or more times between 2000 and 2007 (including cases still pending). We identified 106 such patents. (13) For purposes of our study, we also identified a randomly selected control set of 106 patents that have been litigated only once during this time period. This allows us to extend the work that Allison et al. did in 2004, comparing the "ordinary" litigated patents (already outliers, as we have seen) to the most-litigated patents. If Allison et al. are correct, we would expect those most-litigated patents to exhibit even more evidence of private value, and perhaps even more of an industry skew.

    To test these hypotheses, we collected a variety of data about both the patents and the patent lawsuits. (14) For each litigated patent, we collected information about small-entity status (i.e., whether the patent owner at issue is an individual, university, or small business, as opposed to a large business); whether the patent is assigned before litigation; the number of continuation applications filed leading to issuance of the patent; the raw and adjusted number of "forward citations" (citations to the patent by later patents); the number of "prior art references" the patent makes to U.S. patents, foreign patents, and nonpatent prior art; and the number of claims in each patent.

    We also categorized each patent into both an industry and a technology in order to ascertain whether significant differences existed in the technology and industry areas. (15) We did not use the Patent and Trademark Office (PTO) classification system because our prior work has found that it is badly flawed. (16) Instead, we did our own categorization. In our description of technology and industry areas for inventions that we actually encountered in our data sets, we attempted to define the areas in a comprehensive way, and our definitions are thus broad enough to include specific inventions not actually found in our data sets. (17)

    1. Technology Areas

      (1) Software: An invention in which data processing is a sufficiently critical element that at least one claim element in the patent consists of data processing--the actual manipulation of data--regardless of whether the code carrying out that data processing is on a magnetic storage medium or embedded in a chip. The latter is often called "firmware." (18) This category includes the two software patent subsets described below.

      (2) Pure software: An invention consisting only of data processing; all claim elements in the patent consist of data processing. However, we include in this definition a patent on data processing in which there is a trivial nondata processing element such as a generic input, output, or storage element clearly not intended to represent any novel technical advance. This category is a subset of software.

      (3) Software business method: Also a subset of software patent, this category includes software patents that cover methods for conducting business transactions. Business-method patents are notoriously difficult to define, with possible definitions varying greatly in scope. For this study, we used a narrow definition limited to those patents the claims of which obviously covered only such things as automated generation of customer proposals, advertising, and the use of online catalogs.

      (4) Mechanical: A process, machine, or product that consists solely of the use of mechanical parts, sometimes combined with heat, hydraulics, pneumatics, or other power sources; or an invention in which the above is a critical part. Some inventions classified as mechanical will also be in one or more other classifications, such as electronics. While many different types of inventions fit into this category, it is not a catchall "other" category.

      (5) Electronics: A process, machine, or product in which the invention or a critical part thereof makes use of traditional electronic circuitry or involves electric-energy storage. An invention in this classification may also be included in other classifications, such as chemistry, mechanics, or optics.

      (6) Optics (other than imaging): A process, machine, or product in which the invention or a critical part thereof employs light waves. Optics technology will sometimes also be classified in one or more other areas, such as electronics or chemistry.

      (7) Imaging: A process, machine, or product in which the invention or a critical part thereof involves the creation or processing of images for various purposes. The imaging may be analog or digital. The majority of imaging patents have medical uses, but some serve other purposes such as security or meteorology.

      (8) Biotechnology: A process involving advanced genetic techniques intended to construct new microbial, plant, or animal strains; a product created from such a process; or the way such a process or product is used in biotechnology research. Although there are a number of different genetic-engineering techniques, for several reasons we decided not to disaggregate these techniques into separate technology areas.

      (9) Chemistry: A process consisting of chemical reactions, a product resulting from such a process, an invention of which a chemical process or product is a critical part, or an invention consisting of a purportedly novel use of chemical substances. Closely related inventions such as those on novel metal alloys and nonmetallic amalgams are also included. An invention in the field of chemistry may be included in one or more other classifications, such as electronics or optics.

    2. Industry Areas

      (1) Computer: This industry encompasses both software and computer-hardware inventions, including not only hardware products but also machines and processes for making computer hardware. As discussed...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT