The system for resolving domain-name disputes is unique in that it gives the complainant the unilateral ability to choose the arbitration provider.  As a result, providers, whether motivated by profit or prestige, have incentives to favor the complainant, a trademark owner who claims that a domain name violates its mark.  The domain name-dispute resolution system should be reformed to allow both complainant (trademark owner) and respondent (domain-name registrant) to strike an equal number of arbitration providers. This reform would give providers an incentive to be unbiased.

The dynamics of the domain-name dispute-resolution system provide an example of forum selling.   The idea of forum shopping is well known: plaintiffs choose the court that is most favorable to them.  Forum selling is the idea that courts and judges are not passive participants in forum selection.  Sometimes they actively seek more cases, and to do so they favor the party with the power to select the forum, which is usually the plaintiff.  While most courts and judges do not want to hear more cases, some seek the power, prestige, and benefits to their localities that higher caseloads can bring.  When plaintiffs have broad jurisdictional choice, those courts and judges tilt the law in favor of the plaintiff, because it is the plaintiff who ordinarily has the power to choose the court.   The result is that the judges and courts that are most pro-plaintiff have a disproportionate effect on the law, and the law takes a pro-plaintiff tilt.

The prime example of forum selling is patent litigation in the Eastern District of Texas.  For the last decade, judges in that district have openly sought more patent cases.  They have publicly stated they find such cases more interesting.  It is also undisputed that the large number of patent cases in their district has benefited the local bar and economy.  Although judges in the Eastern District claim that they attract cases by being more efficient and more expert, their decisions favor patent plaintiffs in a number of subtle, but important ways.  In particular, they allow plaintiffs to de facto choose the judge, almost never grant summary judgment, and almost never stay cases pending reexamination.  As a result, more than a quarter of patent cases are filed in the Eastern District of Texas, even though it is home to no major technology companies and no large cities.

Other examples of forum selling include state-court class-action litigation, pre-modern common law adjudication in England, and bankruptcy.  Domain-name dispute-resolution is different, because it involves arbitration. Nevertheless, the incentive to hear more cases is clearer for arbitration, because arbitration providers and arbitrators have direct financial incentives to hear more cases.

There are two dominant views about how trademark owners choose arbitration providers. According to one view, trademark owners choose the arbitration provider which is most likely to rule in its favor. This view is consistent with the idea of forum selling, because competition among arbitration providers gives them incentives to be biased in favor of the complainant.  Another view, which would not be consistent with forum selling, is that trademark owners choose the fastest, most efficient arbitration provider.

Data analysis

 Since December 1999, trademark disputes relating to many of the most important domain names have been governed by the Uniform Domain-Name Dispute-Resolution Policy (UDRP).  -That policy allows the complainant (trademark owner) to choose unilaterally from among approved dispute resolution providers.  By early 2000, there were four approved providers: the World Intellectual Property Organization (WIPO), the National Arbitration Forum (NAF), eResolution, and a fourth that never received a significant number of cases.  This article focuses on the first 18 months, before eResolution went out of business.

Table 1. Summary Statistics (December 1999 – June 2001)

Table 1 shows that WIPO and NAF had the largest market shares and together received over 90 percent of the cases.  eResolution heard less than 7 percent.  WIPO and NAF arbitrators also ruled for the complainant most often (80 percent or more).  These raw statistics are consistent with the idea that complainants chose dispute-resolution providers based on how favorably they ruled for the complainant.  The fact that even eResolution decided for the complainant more than half of the time reflects the fact that most disputes were simple cyber-squatter cases, in which the respondent clearly had no right to the domain name.

NAF resolved cases the fastest.  WIPO was the slowest, taking 50 percent more time.   These statistics are inconsistent with the idea that complainants chose providers based on speed.Analysis of trends over time does not significantly change these conclusions.

Table 2. Change over Time (December 1999 – June 2001)

Table 2 is flatly inconsistent with the idea that speed was the key factor.  WIPO became noticeably slower, yet it was the only provider whose market share increased.  eResolution became significantly less complainant friendly – its percentage of decisions for the complainant fell 16.2 percent — and its market share fell 3.1 percent.  Although 3.1 percent may not seem like a large decline, since eResolution started with only 8.6 percent of the cases, a 3.1 percent decline is actually a loss of more than one third of its market share.

Solution

A better system would borrow from procedures already used routinely for the selection of arbitrators and jurors.  Both the complainant and the respondent could be given the list of approved arbitration providers.  Each party could then strike an equal number of providers until only one or two providers was left.  If only one provider remained, then that provider would resolve the dispute.  If two providers remained, then the dispute would be assigned randomly to one of those two providers.

The advantage of this solution is that it would encourage arbitration providers to be unbiased.  Each party would use its strikes to eliminate the most biased providers.  Thus, in contrast to both the present system, this solution would reward neutrality rather than bias.

This post comes to us from Professor Daniel Klerman of the USC Gould School of Law. It is based on his recent articles, “Forum Selling and Domain-Name Disputes,” available here, and “Forum Selling,” (with Greg Reilly) available here. .