Chicago-Kent Professor David Schwartz and University of Illinois Jay Kesan have written a thought-provoking article looking at what empirical data tells us about trolls (click here for their Patently-O summary of the article).  Blog readers will remember that Schwartz has recently written an interesting article analyzing the empirical characteristics of trolls and their counsel — my analysis is here and here.  Schwartz and Kesan point to a lack of empirical support for troll criticisms, and ultimately conclude that we need more efficient avenues for determining the outcomes of patent infringement suits, a conclusion that is hard to dispute.  Here is what they say about standard troll criticisms:

  1. Trolls assert patents of “dubious” quality.  Schwartz and Kesan argue that there is little evidence supporting the standard concern that trolls assert “dubious” patents.  While I would not suggest that everyone troll patent is low quality, I suspect that every retailer that has been sued by a few trolls knows from experience that many are either questionable patents or are read far too broadly in order to make a claim on the asserted technology.  The problem here is that the system acts as camouflage for these patents.  Because cases based upon weak patents or tortured analysis tend to take the same path through the district courts, as well as the Patent Office on reexam, there is no way to distinguish the cases.  Even time to settlement is a misleading statistic because the real drivers of settlement are complex — quality of patents, exposure of the accused product, reasonableness of counsel, etc.  Schwartz and Kesan suggest that we need other, less-expensive ways to test a patent’s validity.  That would be great, but unless those mechanisms carry the experience and weight of the federal courts, it is hard to imagine they will be any more effective than reexamination proceedings are today.
  2. Trolls pocket too much of the damages/settlements.   Looking at the standard law and economics analysis, Schwartz and Kesan argue that most trolls return a larger percentage of any winnings to the inventors than they do to contingent fee counsel.  But that avoids the reality that this si not a simple shifting of dollars from one innovator to another.  In fact, in the case of retailers targeted by trolls, you are shifting money and time from the retailer who generally did not develop the technology and may only understand its inputs and outputs to the troll, who also did not develop the technology.  The fact that a percentage of those amounts may be returned to an inventor cannot erase the costs that are improperly foisted upon a retailer and, to some degree, its customers.  That transaction cannot be a net-positive transaction for society.  It would be a better argument if the trolls were limited to seeking payment from the technology developers.
  3. Trolls are not innovators.  Schwartz and Kesan argue that the fact that trolls are not innovators can be addressed by the patent laws.  In essence, non-innovative technologies can be invalidated and tortured infringement analysis can be stopped by non-infringement analysis.  The problem here is the obvious and large cost of proving invalidity or noninfringement.  While the system will eventually weed out invalid patents and those that are not infringed, it costs individual retailers a significant amount, in both fees and internal resources, to push a case to that point.

While I question whether the data can tell the entire story in the troll world, Schwartz’s and Kesan’s ultimate point is a good one:

In sum, the debate about NPEs underscores the fundamental point that patent litigation is about whether a valid and enforceable patent claim has been infringed.  The more efficiently we can determine this answer in a lawsuit, the better off we will be.