How ICANN accountability reforms dodged a bullet (barely)

This is the story of how ICANN accountability reforms dodged a bullet (barely), and how a greedy GAC got its wrist slapped. Actually the story is not quite over, and we do not know for sure yet whether the ending is a happy one. But that is why the story must be told now.

ICANN’s enhanced accountability process was supposed to make ICANN’s all-powerful board and staff more accountable to the community. By creating new forms of oversight to replace the role of the U.S. NTIA, however, the reform process also made it necessary to realign decision making power – and that in turn made it possible to redistribute power among ICANN’s Supporting Organizations and Advisory Committees.

For most of the involved community, the goal of the reform process was to keep the relative power of SOs and ACs the same, while making ICANN the corporation more accountable to the stakeholders involved. For ICANN’s Governmental Advisory Committee (GAC), however, board accountability was not the primary goal. Pressure from the GAC, after all, is responsible for some of ICANN’s most egregious accountability failures and its worst instances of mission creep. The GAC’s main concern always seems to have been to increase its own influence over the policy process.

During the CCWG deliberations, the push for a larger or more powerful GAC role had three components: 1) participation as a voting member in the community mechanism that would be used to exercise oversight over ICANN; 2) a demand that the board would need a 2/3 majority, rather than a simple majority, to overturn GAC advice; and 3) an assertion of the right of the GAC to unilaterally change the rules governing what constitutes formal advice to the board.

The GAC succeeded in achieving the first of those goals easily. The second and third demands, which deal with the role of GAC advice in the policy process, ran into serious trouble. Fortunately, the role of GAC advice was earmarked for special consideration by one of the “stress tests” used by the CCWG to assess the resiliency of their proposed reforms. Stress Test 18 posited a scenario in which the GAC changes its operating procedures (which it can do with a simple majority) to allow formal Advice to the board to be based on a majority vote or some other threshold short of full consensus. Consensus in the GAC currently means the absence of a formal objection from any government. This means that any individual government has a blocking veto. If that bar is lowered significantly, a dominant faction of governments could subordinate the entire ICANN policy process to its wishes by tying up the board in the consultations required by ICANN’s bylaws. The only way for the reforms to pass Stress test 18, the CCWG concluded, would be to make ICANN’s fundamental bylaws require full consensus before something would be considered ‘GAC advice.’

Lowering the bar for the provision of advice would strengthen the hand of governments in ICANN immeasurably. Many observers do not appreciate how significant such a change would be. They think that GAC’s advisory role means that the board can just overrule the GAC when it pleases. In fact, as this writer has argued in more detail here, the ICANN bylaws give the GAC an incredible amount of strategic power over the policy making process. GAC advice can come at practically any time in the policy development process, and can ignore or even contradict the painstaking compromises, bargains and issue analyses developed by the GNSO. Even if it is incoherent or poorly thought out, GAC advice cannot simply be dismissed; if the board disagrees the bylaws require it to negotiate with the GAC to reconcile any differences. That requirement can hold up the conclusion of the policy process. The GAC always gets the last word in the policy process and in order to expedite the conclusion of the process, the board has a strong incentive to bend or alter whatever the GNSO has done in order to keep the GAC happy.

By requiring advice to be full consensus advice, the current practice not only raises the bar for the GAC to produce advice, it also limits the amount of advice that the GAC is able to give. Not surprisingly, then, most vocal participants in CCWG from the GAC hated stress test 18.  Concerned as it always is with its own parochial interests rather than with the efficiency and fairness of the policy process as a whole, the GAC spokespersons claimed that enshrining full consensus in the fundamental bylaws was an attack on its ability to define its own operating procedures.

In response to the first and second CCWG draft proposals, the GAC said that it would accept the consensus requirement but reserved to itself the right to redefine consensus. So if GAC operating procedures (which can be changed by a simple majority of voting members) defined “consensus” as 51%, then consensus advice would require only 51% of the votes.

After further pushback from the community, the GAC tried to paper over the problem with ambiguous wording. It proposed to define consensus as “the absence of any significant objection,” leaving “significant” undefined. GAC’s claims were, predictably, supported by ALAC, another advisory committee seeking a more influential role for itself in ICANN. The CCWG co-chairs, trying desperately to achieve agreement within the group, tried to push this unclear compromise through, but ran into strong opposition from GNSO participants. There was a standoff with time running out.

The standoff was broken when the U.S. government intervened. Acting on behalf of Assistant Secretary Strickling, the NTIA’s Suzanne Radell sent a message to the CCWG list saying:

As the CCWG finalizes its proposals for enhancing ICANN’s accountability, we feel we should reiterate our view, as we stated last July, that ICANN preserve and clarify the current practice of the Board  in responding to advice it receives from the Governmental Advisory Committee (GAC).  Specifically, ICANN should amend its Bylaws to clarify that the Board is required to enter into a formal consultation process with the GAC only where it receives GAC advice that is consensus advice based on the current definition within the GAC’s Operating Principles, that is, advice to which no GAC member has raised a formal objection.

The statement went on to say:

Anything less than consensus places the Board in the awkward, if not impossible, position of trying to choose between governments with conflicting opinions.  NTIA sees any deviation from the current standard of consensus as introducing instability into the system while also inadvertently diminishing the important role of governments.  … Asking the Board to interpret any other threshold of support seems counter to the spirit of the CCWG’s efforts to empower the community in a clear and consistent manner.  It also undermines the work done to implement the relevant recommendations of ATRT1 to fix what the community diagnosed as a dysfunctional Board-GAC relationship.

The NTIA intervention finally forced GAC members to back down. In order to solidify the GAC’s willingness to go along with a requirement of full consensus for advice, other stakeholders expressed a willingness to accept the GAC’s wish that its advice could only be overturned by a 2/3 majority of the board. The final language, which appeared to be acceptable to all, was as follows:

The advice of the Governmental Advisory Committee on public policy matters shall be duly taken into account, both in the formulation and adoption of policies. In the event that the ICANN Board determines to take an action that is not consistent with the Governmental Advisory Committee advice, it shall so inform the Committee and state the reasons why it decided not to follow that advice.   Any GAC advice approved by a full GAC consensus, understood to mean the practice of adopting decisions by general agreement in the absence of any formal objection, may only be rejected by a vote of two-thirds (2/3) of the Board, and the Governmental Advisory Committee and the ICANN Board will then try, in good faith and in a timely and efficient manner, to find a mutually acceptable solution.

The last paragraph of the ‘acceptable language,’ however, seems to once again offer the GAC the ability to redefine consensus in a way that lowers the bar:

The GAC has the autonomy to refine its Operating Procedures to specify how objections are raised and considered (for example, disallowing a single country to continue an objection on the same issue if no other countries will join in an objection).  When transmitting consensus advice to the Board for which the GAC seeks to receive special consideration, the GAC has the obligation to confirm the lack of any formal objection.

The unresolved question now is this: currently, the definition of “a full GAC consensus” means the absence of any formal objection. Presumably that means that current practice goes into the fundamental bylaws. But what if the GAC unilaterally changes its definition of an ‘objection’? Would the bylaws definition of ‘advice’ change with it? Or would the fundamental bylaws have to be modified by the community mechanism before that change would go into effect? We strongly favor the latter alternative. Before the nature of GAC consensus is changed, the entire ICANN community should provide strong support for it. Indeed, if that is not what the CCWG agreed to, we are back to a situation in which the GAC can unilaterally change, by mere majority rule, a fundamental bylaw that affects the entire community.

Stay tuned.

This blog has been warning for nearly a decade that the threat of a governmental takeover of ICANN’s little corner of Internet governance does not come from the sluggish International Telecommunication Union (ITU), but from the ITU-within-ICANN, the GAC. If anyone was not convinced of that before, the drama over stress test 18 should convince them now.