Gaming the Patent System: An Empirical Analysis of Litigation Economics and Possible Solutions

Abstract

This paper attempts to form an economic model of how an entity could game the patent litigation system by suing multiple parties and offering settlements below a median amount. Empirical data is taken from district court cases involving patent infringement suits to determine mean and median settlement values. From there, suggested solutions are presented that may have varying effect on the gaming of patent litigation and the patent system itself. Patent litigation and settlement are based on the underlying economics and rules imposed by the litigation system itself. Inherent in this analysis is the assumption that accused infringers will take whatever option is expected to cost them the least. Therefore, when the cost of taking a license from the patentee is less than the expected cost of litigation, the accused infringer will take a license - without regard to the merit of the underlying suit. Once patent litigation has been initiated, this license offer comes in the form of settlement. The terms of settlement agreements would be very helpful to use in determining at what prices accused infringers have settled in previous cases. Settlement agreements are rarely made public, however, because of the impact they may have on other accused infringers. This allows only speculation on how risk-adverse patentees and accused infringers are gaming the patent litigation system and on what their litigation expected values are. There are two factors that determine the overall cost of litigation to an accused infringer: the money paid to litigate (actual cost) and the possible judgment damages from losing the suit. Each possible outcome has a certain probability of occurring in patent infringement which can be approximated based on historical data. All outcomes should be seen as future cash outflows. Then, a weighted average based on their probability of occurrence gives the true cost of litigation. The first part of Section I will focus on the actual cost of litigation and the judgment award amounts found based on the District Court docket research. The second part of Section I will then focus on ascertaining the probabilities of outcomes during patent litigation. The final part of Section I will formulate the expected value of a patent litigation settlement based on the calculated outcomes and probabilities.This paper will eventually discuss possible solutions to gaming the patent system based on the economic models and criteria above.