By David R. Johnson and Susan P. Crawford*
The ICANN reform committee is heading towards a proposal that would eliminate any requirement that policies binding on registries and registrars be supported by documented consensus among affected parties. In order to ensure that ICANN is "effective," the committee will recommend that policy-making authority should be vested in a carefully chosen Board. We have previously commented that this approach conflicts with the contracts that give ICANN its only powers and creates regrettable seat-claiming controversies. We have also pointed out that ICANN's legitimacy rests right now on the basic, voluntary bargain reached by those who have to implement the policies that ICANN makes: all of the entities with contracts with ICANN have agreed not to be irrational holdouts against a widely-supported solution to a problem that affects the global interoperability of the domain name system. We have been worried about the strong potential for mission creep, over-regulation and abuse of ICANN's monopoly power that the current plan creates.
But we have also been thinking about why consensus matters and whether it might be possible to create an effective top-down policy-making regime that serves the same basic goals that the consensus process supports. We are also sympathetic to those who have grown to hate the word "consensus" because of the abuses that have been inflicted in its name. We think we have found a solution to this top-down versus bottom-up conundrum. The answer is: create several competing policy-development fora and allow registries to choose among them. As long as we ensure that entries in the root do not conflict, and that all accredited fora require minimum policy undertakings of their members, we should allow competition for policy-making.
We take as a given the proposition that there can only be one authoritative root zone file. But it is not at all a given that the Department of Commerce can listen to only one source of recommendations as to what should go into or come out of that file. Above certain mandatory policy minimums, we do not see why there should be only one organization attempting to develop policies to guide the operations of DNS registries and registrars.
What minimum policies should we require each ICANN to mandate? Very few -- and only those that are necessary to prevent wrongdoers from imposing costs on others without any justification.
Each policy-development forum may find it desirable to develop policies in additional areas. It may even be desirable to develop those policies by means of voting by an "effective" Board that seeks input from various sources and decides what it thinks is best for its global internet community. What is decidedly not desirable (or necessary) is to have only one such policy development Board.
- Because we need to be able to track down wrongdoers, any policy-development forum recognized by DoC should be required to establish a policy requiring some mechanism for appropriate access to accurate whois data by those with justification for seeking it.
- Because virtually everyone agrees that cybersquatting imposes unjustified costs, any policy-development forum should be required to have some cheap, fair, and globally efficient dispute resolution mechanism focused on that problem.
- Because exposing registrants to loss of their domain name in the event of registrar or registry failure is unwarranted and unnecessary, any policy-development body recognized by DoC should require that critical data be escrowed.
- Because registrants may invest heavily in developing a domain name and will be vulnerable to abuse if locked in to a particular name, any policy development body recognized by DoC should be required to establish some form of price cap (e.g., renewals may not be priced higher than new registrations) to address this "lock in" problem.
Registries should be able to choose among at least three different fora (and to move from one to the other freely) in order for true policy competition to exist. We already have the seeds of this competitive regime -- with ccTLDs forming their own policy-development body and some who have not been admitted to the ICANN regime talking of developing a competitive (but non-conflicting) alternative. Many of the gTLDs will choose the existing ICANN. ICANN would surely become accountable to those affected by its rules if those affected could choose a different forum. Someone would of course need to keep the single authoritative IANA database, but they could take inputs from all policy-making bodies relating to the entities that have signed up for their respective regimes.
No matter how open and transparent the policy-making process of a top-down Board, there will always be questions about whether the Board really speaks on behalf of the global Internet community. One good way to help assure that is to allow the members of that community who will be bound by any resulting rules to choose among different "representatives." If the ccTLD organization could itself welcome or introduce gTLDs, and those who cannot obtain ICANN's recommendation for entry into the root zone file could seek that recommendation from another group with different ideas and different decision-making processes, capture by a small crowd of insiders would be rendered much more difficult. Such a system is not quite global democracy. It is certainly not "consensus." But competition among (responsible) policy-making fora would be the next best thing -- a solution DoC could accept on behalf of the entire global internet community with assurance that it was not empowering a new quasi-governmental body to abuse its monopoly position.
Consensus is a nice ideal. If we have only one forum within which to decide which DNS operational issues are "global" in character and must be the subject of rules binding on all registries and registrars and registrants, then seeking a documented consensus is still the best idea. But we have listened carefully to the desire of some ICANN participants to establish a more effective top-down approach to decision-making. Top-down decisions can be legitimate if they reflect a robust, voluntary political process for selection of the decision-making board and vetting of proposed policies. But such a system can only really be voluntary if there are alternatives and if those bound by the rules are free to choose among them.
We do not want competition among roots, or a form of policy-making competition that could produce a "race to the bottom" by creating havens for cybersquatting or anonymous law violation. We think it is essential to have protections of price caps on registration fees and escrow for critical data in place. But we believe the world should want competition for policy-development in other areas where the answers are not so clear: business models and alternatives among lawful practices of registries, decisions regarding how to implement internationalized names, and decisions on how to open new TLDs are all examples of policy areas where diversity would be a good thing. It seems to us that all of these policy fora will want to have good relationships with governments.
In short, it seems to us that the DoC (whose White Paper and MoU requires "bottom up" policy development) and the registries (that have signed contracts that require them to comply with ICANN policies only when based on a documented "consensus") should both be willing to go along with the proposed ICANN reform on one key condition: that there are at least three competing policy-development fora among which registries (and, indirectly, registrars and registrants) may freely choose in deciding whether and how to develop policies over and above the minimum necessary to prevent clearly unjustified wrongdoing.
DoC would accordingly require all contracts between registries and ICANN-like bodies to allow the registries concerned to retain their "delegations" while shifting their commitment to follow community-developed policies to another qualified policy-development body. DoC can still insist that there be only one authoritative, non-conflicting root zone file. But it should endorse true competition in the area of developing DNS policies. If the reformed ICANN proves to have the most attractive policy-development processes, inducing most registries and registrars to sign up for its form of "governance," more power to it.
*David R. Johnson is the founder of Graphical Groupware. He can be reached at email@example.com. Susan P. Crawford is a partner with Wilmer, Cutler & Pickering, an international law firm based in Washington, D.C. She can be reached at firstname.lastname@example.org. This paper was written in our personal capacities, and not on behalf of any client of Wilmer, Cutler & Pickering. It does not necessarily represent the views of any Wilmer, Cutler & Pickering client.