A Better Patent System through Better Data

If you are a sports fan with an analytical bent, finding statistics describing the performance of your favorite team, player, or league is almost trivial. Hypothetically speaking, if you have fond memories of escaping the grind of law school to go watch minor league baseball, and, still hypothetically speaking, if you particularly remember a couple of good performances by a center fielder with an otherwise unremarkable career, you can find that guy’s stats with just a few keystrokes. One of the unsurprising ironies of the world is that statistical analysis of what is ostensibly only entertainment (sports) is readily available, but statistical analysis of what is asserted to be a critical element of the economy (the patent system) is limited to proprietary studies or what the sports statisticians discount as counting statistics such as how many patents issued or how many infringement suits were filed.

Seemingly anyone with any interest at all in patents has an opinion - often a strong opinion - about what should be done to “fix” the patent system. Precisely what aspect of the patent system needs to be fixed is not always a point of agreement. Everyone seems to have an anecdote that supports their own view, but open and verifiable patent data beyond basic counts has been hard to come by, much less any agreed upon ways of analyzing that data statistically. Without actual data describing the fundamentals of the patent system to consider, reaching a consensus on what needs to be changed in that system is unlikely, and identifying the optimal changes to the system is nearly impossible.

Into this data desert enters Professors Christopher A. Cotropia, Jay P. Kesan, and David L. Schwartz. The preliminary draft of their paper, Patent Assertion Entities (PAEs) Under the Microscope: An Empirical Investigation of Patent Holders as Litigants, will not ultimately answer questions as to what, if any, changes should be made to patent law, but it provides a start on the often skipped step of rigorously understanding our current patent system to shed light on what changes may or may not be in order. The research described required the individual analysis and classification of over 7,500 patent lawsuits filed in two calendar years in order to permit the classification of the parties to those suits. This classification of parties to patent suits provides a snapshot of what kinds of entities are actually asserting patents and what kinds of entities are being subjected to patent suits. The preliminary draft of the paper is available for download here, and is well worth a read. A brief summary of their findings is that while the total number of suits filed by non-operating companies (aka, “trolls”) increased from 2010 to 2012, the increase seems to be entirely due to procedural changes to the joinder rules that were made in 2011. Essentially, more patent infringement suits were filed in 2012 than in 2010 because multiple defendants could no longer be so easily brought into the same suit. The total number of patent infringement defendants in 2012 was roughly the same as in 2010. Critically, though, no one has to take the authors word for it - their methodology is clearly described and their data set is available.

As the authors themselves readily acknowledge, much more work remains to provide a properly rigorous analytical framework for understanding the patent system and its impact on the economy. Some of the relevant pieces of data, such as cease and desist letters, are not readily available. However, if statistical analysis of data has been used to build a better baseball team, surely we should take a lesson in building a better patent system.