San Diego Law Review


Michael Risch

Library of Congress Authority File


Document Type



This Article builds on the very best parts of current cross-sectional work by adding a longitudinal component that finds every case that asserted a set of patents and then separately tracing the outcome of every patent asserted in each of those cases. Part II discusses the debate about patent troll litigation. It focuses on how other studies have measured patent quality through patent litigation data. Part III discusses this study’s methodology: the collection of data about highly litigious NPEs and a control group of randomly selected nonNPEs. It describes how the cases were selected and the data collected in different phases of the study. This includes collection of previously untapped data: reexamination outcomes. Part IV presents the data in a variety of formats, and follows each subpart with a discussion of potential policy implications associated with the data. Subpart IV.A shows the growth in case complexity, including defendant counts, transfers and consolidations, and selection of district court venue. This subpart includes a linear regression showing the effect of NPEs on case duration. Subpart IV.B shows how much more often the NPEs settled and all of the different ways that cases ended. For example, NonNPEs ended a large portion of their cases with consent judgments, though the NPEs used consent judgments as well. Subpart IV.C presents data about invalidity in a number of ways: in terms of adjudicated patents, all patents, and cases. This subpart presents a novel regression estimating the likelihood that an asserted patent will be invalidated and finds that NPE status is not among the factors. Subpart IV.D examines infringement findings and shows that the primary concern with NPEs may be noninfringement rather than invalidity. Both subparts C and D show that decisions on the merits of cases are so rare that it is difficult to base policies on them. This Article concludes with some thoughts about how the results might guide policy.