Good question, as I would like to know the answer to this myself. Panelists seem to have differing views on the subject, making it very difficult to know 'what the law is'. Accordingly, some cases result in RDNH and some don't, with seemingly little apparent consistency.
I think if there is one general conclusion that one could possibly make as to when a Panel tends to declare Reverse Domain Name Hijacking, is that RDNH is declared when they feel that "the circumstances unquestionably warrant it". Of course, this is highly subjective, and that may explain the apparent lack of consistency in cases where Reverse Domain Name Hijacking was requested.
For example, in Zija International, Inc. v. Deep Frontier, WIPO Case No. D2013-0107, the respected three-member Panel said the following:
"Contrary to the Respondent’s view, the Panel finds that, under the specific facts of this case, the Complainant’s actions in filing its Complaint do not rise to the requisite level of bad faith to warrant a finding of reverse domain name hijacking."
This is a fairly typical Panel reason for generally denying a request of Reverse Domain Name Hijacking, and is consistent with my aforementioned view that Panels simply exercises their discretion if they feel that overall, it is "appropriate". Nevertheless, the Panel in this case, interestingly, did provide a further explanation of why RDNH was not appropriate in this particular case:
"Specifically, the Panel believes that the Complainant had a reasonable basis, though misguided and ultimately incorrect, to believe that it might prevail in this proceeding and thus did not institute the proceeding merely to harass the Respondent and hence as an abuse of the administrative procedure."
As can be seen from the above excerpt, the Panel in this case gave some further insight into the reason why they exercised their discretion to deny RDNH. Essentially, the Panel explained that a "reasonable basis" for bringing a Complaint, even if "misguided" and "ultimately incorrect", can be sufficient to overcome an allegation of RDNH. In other words, a Complainant is entitled to bring a lousy Complaint and to lose as a result, without necessarily being considered a domain name "hijacker".
But, when is a complaint so lousy, that it ceases to provide a "reasonable basis for a UDRP complaint"? In the Zija case, which was a dispute over the four-letter domain name, ZIJA.COM, the domain was registered in 2002, whereas the complainant's trademark did not exist until two years later. Accordingly, the Panel correctly stated that:
"It is rather axiomatic that if a complainant was not in existence at the time a domain name was registered, then the domain name could not have been registered in bad faith – as the respondent could not have had any prior knowledge of the complainant."
So, even though it was "impossible" for the domain name to have been registered in bad faith, since the complainant did not bring its complaint merely to "harass" the respondent, it was excused from having brought a complaint that was entirely futile from the beginning. Shouldn't the complainant have been aware of this futility from the beginning however? And since it was or ought to have been aware of the futility, and proceeded nonetheless, is that not recklessness that rises to the level of "bad faith", and even arguably, "harassment"?
For example, in the case of Proto Software, Inc. v. Vertical Axis, Inc/PROTO.COM (WIPO Case No. D2006-0905), the Panel ruled that, "a finding of Reverse Domain Name Hijacking may be made if the Complainant knew or should have known at the time that it filed the Complaint that it could not prove that the domain name was registered in bad faith".
The Panel in the Proto case also ruled that:
"Complainant is represented by Counsel who even on a rudimentary examination of the Policy and its’ application in this area should have appreciated that the Complaint could not succeed where the Respondent’s domain name had been registered three years prior to filing a trademark application or actual use of the mark", and accordingly made a finding of RDNH. The Panel also made the important observation that;
"Initiating domain name dispute resolution proceedings necessarily involves putting the parties to a considerable expenditure of time and in many cases cost and the Policy must not be used unless the complainant has a reasonable and credible belief it is entitled to succeed. In particular, proceedings must not be commenced in a brash and totally unjustifiable attempt to pressure a domain name owner into releasing a legitimately held domain name that considerably pre-dates any trademark rights held by the complainant, see Sustainable Forestry Management Limited v. SFM.com and James M. van Johns “Infa dot Net” Web Services, WIPO Case No. D2002-0535."
Accordingly, it appears difficult to distinguish the Zija case from the Proto case, as both involved complaints about domain names which pre-existed the trademark rights on which the case was based, yet only the Panel in the Proto case decided that RDNH was warranted in the circumstances. Clearly in the one case, the Panel felt that the complainant's futile case and waste of everyone's time and money was less serious than in the other similar circumstance. Perhaps this distinction without an apparent difference boils down to the make-up of the Panel, although it is of course difficult to determine.
Perhaps the answer to this issue is to have every complainant expressly certify prior to commencing the UDRP Complaint, that the trademark on which the Complaint is based, pre-existed the domain name registration. This would at least put complainants on express notice of this fundamental requirement of a UDRP, and thereby possibly avoid needless UDRP cases. Such a solution would also ostensibly make it easier to make a finding of RDNH where this certification turned out to be false.