Significance
Patents are costly to enforce. Patent trials typically cost each side over $1 million, and filing a patent suit may decrease the market value of a firm by 2 or 3 percent. Yet patent holders are going to court in record numbers. Why don’t they just settle?

This paper investigates why computer and research medicine (i.e. pharmaceutical and biotechnology) firms fail to settle patent lawsuits despite the high cost of taking these cases to trial. The author argues that once suits are filed, disputants cannot settle claims where the plaintiff stands to gain more in winning a lawsuit than the defendant stands to lose. These cases of so-called asymmetric stakes occur when the patent confers unique advantages on its holder.

To test this theory, the author examines two forms of asymmetric stakes: strategic stakes and mutual hold-up. In the case of strategic stakes, a firm with a patent monopoly can leverage specific strategic assets. For example, the main producer of insulin can leverage its existing investments with a patent on a new technique for making the hormone. The greater the value of a patent, the more difficult it becomes to craft a settlement suitable to both parties. Self-citations (references to patents held by the patentee) can serve as a measure of strategic stakes, and a patentee with a high strategic stake in a litigated patent should be less likely to settle. Prior research suggests that strategic stakes are probably important in the research medicines industry.

Mutual hold-up occurs when firms depend upon one another for reciprocal access to patents in order to manufacture their products. This is common for companies that specialize in computer operating systems, where individual products consist of several different patented inventions, often owned by different firms. When faced with a patent lawsuit, an operating system manufacturer may protect itself by threatening its rival with a countersuit over another patent. The degree of hold-up is measured here by the number of in-citations—references that a patentee makes to the patents in its rival’s portfolio. The author proposes that the greater the number of in-citations (as well as the number of countersuits), the more likely the parties are to settle. Of course, both self-citations and in-citations are, at best, imperfect measures of the attributes the author wishes to capture.

Results
To test these hypotheses, the author looks at patent and litigation data collected by the Federal Judicial Center and the US Patent and Trademark Office for suits filed between 1983-1993 in the computer and research medicines industries. Results from a statistical analysis of the data support the strategic stakes hypothesis. A high strategic stake increases the probability that a dispute will be decided in court. However, contrary to prior research, this occurred primarily in the computer industry, not research medicines. The effect may not appear significant for research medicines because of the differences between the pharmaceutical and biotechnology segments of the industry included in the data set.

Empirical evidence for mutual hold-up proved inconclusive. However, the higher percentage of countersuits in the computer industry and the fact that sets of related countersuits are all settled on the same date do provide indirect evidence for the hypothesis.