By: Dr. Milton Mueller, Georgia Institute of Technology, School of Public Policy

September 15, 2015 – Full Article as .PDF

1. The IANA TransitionMilton-Mueller-Artwork-4

On March 14, 2014, the United States government announced its intention to end its direct role in overseeing the Internet’s Domain Name System (DNS) and IP address allocation. Those aspects of global Internet coordination, known as the IANA functions

[1], are one of the few centralized points of control on the Internet. In the wrong hands, they could be used to interfere with Internet services in ways that could limit the freedom of both users and providers.

The IANA transition, as it is called, means that the U.S. government will no longer be the contracting authority for the IANA functions. It would end 17 years of government oversight of ICANN, and turn the DNS and IP addressing over to self-governance by the technical community, civil society, business and Internet users. This is a praiseworthy attempt by the U.S. to be consistent about supporting a truly nongovernmental, multistakeholder model for Internet governance.

In ushering in the transition, the responsible U.S. government agency, the National Telecommunications and Information Administration (NTIA), has stipulated that the transition proposal must be guided by five principles:

  1. It must support and enhance the multistakeholder model;
  2. It must maintain the security, stability, and resiliency of the Internet DNS;
  3. It must meet the needs and expectations of the global customers and partners of the IANA services;
  4. It must maintain the openness of the Internet;
  5. It must not replace the NTIA role with a government-led or an inter-governmental organization solution.

The last principle, which addresses the role of governments in the regime, has become a major concern in the U.S., especially among some critics of the transition. Opponents of the transition have claimed that U.S. control of ICANN was a bulwark against a government-controlled Internet, and should be kept in place. While that claim is false – indeed, it is practically the opposite of the truth – some of the reforms proposed by ICANN’s Cross-Community Working Group on Enhanced Accountability do have the potential to strengthen the role of governments at the expense of private sector and civil society. Because these changes are embedded in obscure process and bylaw changes, it may not be evident to casual observers how dangerous they can be. While ICANN would still be an institution based in the private sector and civil society, a combination of intended and unintended consequences of changes in the role of the GAC could subvert the 5th principle of the NTIA over the longer term. The role of governments in the accountability reforms, therefore, needs to be carefully analyzed.

The dangers of enhanced governmental influence do not come from the favorite bogeymen of the conservative nationalists (the International Telecommunication Union, Russia, China, or the UN). Ironically, the real threats from governments in the IANA transition are coming from within ICANN itself. Specifically, the problems stem from the expanding role of ICANN’s Governmental Advisory Committee (GAC), its claim to have a special status over the formation of public policy, and from the misguided tendency to see national governments as “just another stakeholder” in the multistakeholder system.

This paper looks at the IANA transition from the perspective of how it affects the role of governments in ICANN. The paper begins by taking a careful look at how governments are involved within ICANN currently, and how their role has evolved over the years. Next, the paper asks whether the accountability reforms proposed as part of the transition make the threat of governmental involvement better or worse. Finally, the paper takes up the more fundamental issue: what, if anything, is wrong with governmental involvement in ICANN and Internet governance? Why not think of governments as just another stakeholder? Why shouldn’t we give them equal footing? All too often, discussions of the role of governments in the regime fail to confront those questions carefully and thoroughly. This paper will do that. Unless we understand the purpose of and rationale for the limitation on states’ involvement, we will not be able to understand whether the transition is making the problem better or worse.

2. The Role of Governments in ICANN

This section reviews the involvement of governments in ICANN. It begins with an assessment of the type of governmental involvement that most Americans prefer to overlook: the involvement of the U.S. government itself.

2.1 The Special Role of the USG

The special U.S. role as contractor for the IANA functions dates back to 1998. There were earlier National Science Foundation Cooperative Agreements to operate aspects of what were, at the time, research and education networks, but not until 1998 did a U.S. federal government interagency process assert control over the contents of the DNS root zone file and make the NTIA responsible for contracting out the IANA functions.

Contrary to the lurid rhetoric of American conservative nationalists, U.S. oversight of ICANN was never intended to safeguard Internet freedom. On the contrary, the Commerce Department explicitly refused to incorporate freedom of expression protections into its arrangements with ICANN (despite calls from many digital rights advocates for it to do so). According to an influential 1996 Clinton administration Framework for Global Electronic Commerce, the real policy objective of the U.S. was to institute a DNS governance regime based on private contracting. A 1998 Commerce Department White Paper referred to this goal as “privatizing” the technical management of the DNS. It is true that such privatization was meant to sidestep intergovernmental treaties so as to create a globalized jurisdiction for Internet governance – but the privatized regime was originally intended to exclude all governments, including the U.S. The NTIA role as contractor for IANA functions was meant to end after two years.

That didn’t happen, however. The U.S. government got deeply involved because of the monopoly power of Network Solutions, Inc. Back then, NSI (now known as Verisign) was the only commercial registry for generic top level domains. There was no vertical separation of registries and registrars to facilitate competition at the retail level. NSI held this monopoly because of an old 1991 contract with the National Science Foundation that was drawn up prior to the commercialization of domain name registration. When NSI was allowed to start charging fees for registrations, domain names suddenly became a profitable business. From that point on, decisions about the IANA functions had significant implications for competition policy, trademark protection and freedom of expression. Yet there was no mechanism for the National Science Foundation or the techies it supported to make those kinds of policy decisions.

The creation of ICANN was supposed to provide a stable, private sector-based multistakeholder institution for developing public policy for the DNS. Its most immediate goals were to introduce competition to the industry and to resolve domain name-trademark disputes. The U.S. role was perceived as a necessary but temporary means of getting the fledgling ICANN off the ground and counterbalancing the power of a relatively large and wealthy corporation with strong Washington lobbying connections (NSI). It was also a way to exempt NSI from antitrust liability for managing and publishing the DNS root zone.[2]

The retention of U.S. government oversight for 15 more years than originally intended undermined the legitimacy of the multistakeholder approach. On the one hand, the U.S. government and business claimed that ICANN was a private sector-led, nongovernmental regime, and warned about the problems and dangers associated with subjecting the Internet to governmental power. On the other hand, ICANN was under the direct authority of a government. Aside from the double standard, the U.S. role encouraged other governments to see the IANA functions as a source of control, and to seek the same level of influence over them as the U.S. It was not unreasonable for foreign observers to ask: if the United States had the power to approve all changes to the DNS root zone, why shouldn’t all other governments be involved in that decision too?

2.2 The Creation of GAC

In order to boost support for the regime among other governments, the founders of ICANN created a Governmental Advisory Committee (GAC). The GAC was to be composed of “representatives of national governments, multinational governmental organizations and treaty organizations, each of which may appoint one representative to the Committee.”[3] The GAC, they thought, would give governments a role in the ICANN regime, but it would only be a limited, purely advisory one. It would, as one board member said at the time, keep them in, but also keep them out. The role of the GAC was defined in ICANN’s first set of bylaws as follows:

The Governmental Advisory Committee should consider and provide advice on the activities of the Corporation as they relate to concerns of governments, particularly matters where there may be an interaction between the Corporation’s policies and various laws, and international agreements. The Board will notify the chairman of the Governmental Advisory Committee of any proposal for which it seeks comments under Article III, Section 3(b) and will consider any response to that notification prior to taking action.

Note that in this original model, the GAC only provided advice when the board asked for it. The creation of the GAC meant that governmental representatives did not participate directly in ICANN’s policy development organs. Nor could they appoint board members. Instead, governmental representatives were sequestered from all other stakeholders, and deliberated and reached consensus primarily among themselves. Indeed, for the first 6 or 7 years of GAC’s existence, it did not even allow nongovernmental stakeholders to observe, much less speak at, their meetings (unless invited for a short interventions). They were, in short, not “just another stakeholder” but resembled an intergovernmental organization more than an open, multistakeholder process.

2.3 Progressive Expansion of GAC Influence

Following its first meeting in 1999, the GAC immediately took steps to expand its influence over ICANN policy making. The policies it recommended invariably served governmental interests. In 2000, for example, it developed a set of principles asserting that national governments should be in control of the delegation and regulation of country code top level domains.[4] Although this policy contradicted RFC 1591, which had guided ccTLD delegations since 1994, ICANN began to use the GAC principles as “best practice” in ccTLD re-delegations.

Late in 2001, ICANN made an even more consequential concession to governments: it altered its bylaws to give GAC advice a privileged status. In Article XI 2 (1), j and k the bylaws said that GAC advice “shall be duly taken into account, both in the formulation and adoption of policies.” If the board deviated from a GAC recommendation, the board was required to “inform the [GAC] and state the reasons why it decided not to follow that advice.” In such cases, the bylaws now required GAC and the ICANN Board to negotiate to find a mutually acceptable solution. Only if no such solution could be found could the board proceed with a policy that did not follow GAC advice.

This seemingly innocuous requirement to take GAC advice “into account” had, and continues to have, an overwhelming impact on ICANN’s policy development process. The definition of what constituted “advice” and when it could be offered was vague and open-ended. The GAC could proffer advice demanding the acceptance of a policy or provision that had already been extensively considered, and rejected, by the nongovernmental parts of ICANN’s bottom up process. The GAC could advise that consensus policies developed in the bottom up process be modified to its liking. Special interest groups who did not get what they wanted from the bottom up process could lobby GAC members to get a second bite at the apple by procuring GAC advice favoring their view. In effect the GAC controlled the end game of the process, and acquired the ability to elevate specific viewpoints and policies in the policy development process, and to hold up the ratification of any policy until those views were either implemented or negotiated away.

While it appeared as if the ICANN board had complete freedom to reject advice that it considered ill-founded, such rejection became increasingly difficult during and after the World Summit on the Information Society (WSIS). WSIS was a United Nations intergovernmental process (2002 – 2005) that subjected ICANN, U.S. unilateral control of ICANN, and the private sector-based governance model to concerted attack from the world’s governments. During WSIS, most governments in the developing world were just waking up to the economic and political significance of the Internet and claimed that they, not some private California corporation, should be responsible for public policy related to the Internet. As one might expect, the WSIS critics of the ICANN regime seized on the unilateral U.S. role in overseeing ICANN to attack the legitimacy of the ICANN regime and the multistakeholder model.

In order to appease this governmental assault on the regime, the U.S. (under the Republican Bush administration!) made two more fateful deviations from the original concept of ICANN. First, it accepted a politically binding statement from WSIS that the multistakeholder model assigned different “roles” to different stakeholders, with the role of governments being to make public policy.[5] Second, it encouraged governments to get more involved in ICANN via the GAC. ICANN does not exclude governments, they claimed, it welcomes them! As a consequence of both concessions, the U.S. began to support a more aggressive role for the GAC in influencing policy development. In such an environment, the board had to think twice about alienating governmental representatives by rejecting its advice.

In the five years following WSIS, one sees a clear change in the status of the GAC. It was no longer a committee that provided advice to the board upon request, nor was it even commenting or advising on the policies developed by ICANN’s multistakeholder Supporting Organizations. Instead, the GAC became a competing and parallel policy development organ within ICANN. For any given policy development process, such as those related to new top level domains, Whois, or the registry and registrar contracts, the GAC would develop its own policy recommendations independently, and largely without reference to what was happening in the Generic Names Supporting Organization. When it came time to finalize a policy, the board was presented with two conflicting, uncoordinated and often competing proposals: the GAC and the Generic Names Supporting Organization (GNSO). But because of the privileged status of its advice, the GAC always had the last word. The GAC members have openly stated that their so-called advisory status actually gives them more power over the policy making process than any other part of ICANN. The GAC is, as several of its members have stated in their meetings, “first among equals in the ICANN ecosystem.”[6]

2.4 The GAC and the New gTLD Program

The GAC’s privileged role is evident in a quick review of the development of ICANN’s new generic TLD program, which took more than 7 years to complete. Formal authority for new TLD policy supposedly rested with ICANN’s GNSO, a multi-stakeholder body with balanced representation of different stakeholder groups. In 2006, the GNSO started developing policy for new TLDs. In 2007 the GAC developed a set of “principles” that it wanted to guide the GNSO’s TLD policy.[7]

The Principles call for blocking, in advance, any ideas or proposals for new TLDs that might be controversial or offensive. It said new TLD policies should respect “sensitivities regarding terms with national, cultural, geographic and religious significance.” It proposed that no one should be able to use country, territory or place names or “regional language or people descriptions” without the approval of governments or public authorities. (Note that there is no international law that gives states such an authority.) Naturally, the GAC was highly protective of the acronyms of intergovernmental organizations (members of GAC), even if the acronyms were shared with dozens of other organizations or were common words such as WHO. But GAC was not satisfied with asserting its right to regulate the content of names at the top level of the DNS. The 2007 Principles asked all new registries to create procedures to block second-level names with “national or geographic significance” as well. This should be done, GAC demanded, “at no cost and upon demand of governments, public authorities, or international organizations.”

Clearly, GAC was trying to set the parameters for new TLD policy, not provide advice about it. As bad as the GAC Principles were, however, at least they were relatively general principles. Regardless of their merit as policy guidelines, the GAC principles could be referenced by the ICANN policy development process and forged into clear, specific rules that would unambiguously tell applicants for new TLDs what was permitted and what was not.

But as the new TLD program wore on, GAC abandoned its role as provider of ex ante policy guidance and started expressing approval or opposition to specific new TLD applications. In many cases, it even started to rewrite the rules. The GAC advised applicants for criticism-oriented top level domains (.FAIL, .GRIPE, .SUCKS, .WTF) that they “should develop clear policies and processes to minimize the risk of cyber bullying/harassment.” Disregarding the entire concept of freedom of expression, some governments objected to the very existence of these domains. In the famous .AMAZON case, the GAC was used to veto the application of a company for a TLD corresponding to its internationally-recognized trademark. The veto came even though the Amazon Company’s application complied with all the rules, procedures and requirements of the ICANN new TLD program and broke no domestic or international law. It was vetoed simply because a lot of Latin American governments didn’t like it.

The best example of the GAC’s overreach is provided by the 2013 Beijing communique.[8] It not only expressed objections to specific applications, but concocted an entirely new regulatory framework for more than 200 new TLD registries. After nearly two thousand applications had been made according to a published set of rules, the GAC decided that hundreds of the proposed new TLDS had to meet detailed new regulations called “safeguards” because the words they used (e.g., .GAMES, .HEALTH, .CARINSURANCE) were associated with “regulated industries” and would be “likely to invoke a level of implied trust from consumers, and carry higher levels of risk associated with consumer harm.”

The GAC seemed to be suffering from the delusion that it could eliminate deception and fraud on the Internet by dictating the policies of domain name registries ex ante, rather than by applying existing consumer protection laws to the way domains were actually used ex post. In effect, GAC was attempting to use ICANN to impose rigid, global regulations on certain Internet content and services. Though doomed to be ineffective in this case, the GAC had succumbed to the temptation to leverage centralized authority over domain names to regulate content and services on the Internet. This problem became an important theme in ICANN’s accountability reform process, and we will return to it later.

Regardless of how misguided the advice was, ICANN’s board was required by its bylaws to try to accommodate it, and it worked. After 11 months of additional negotiations with applicants and the GAC, the ICANN board and staff came up with a complicated “Implementation Framework” for the demands made in the Beijing GAC communique. What matters here is that these last-minute interventions from the GAC ended up having more impact on the policies and restrictions applied to new TLD applicants than the actual policy produced by the multistakeholder process.

Notably, this push to extend regulation from domains to services and content was not led by authoritarian countries. It was countries such as the United States, Australia, the European Commission or Brazil who pushed for the safeguards and the extensive regulation of geographic names.

2.5 The GAC as Intergovernmental Organization

The GAC’s involvement in new TLD policy provides a good idea of what an Internet governed by national governments and intergovernmental treaties would be like. The Internet’s idea of permissionless innovation would be thrown out the window. There is a concerted attempt to use the centralized control over the domain name system as leverage to regulate Internet content and services. The interests of the public are subordinated to the interests of national governments.

Despite all the talk about the ICANN regime as “multistakeholder” and how this is preferable to a UN-based or intergovernmental system, we must face the fact that there is already an intergovernmental organization inside ICANN – the GAC. Although it interacts with the board and the other SOs, the GAC is largely a silo organized around states, with one representative per country just like an intergovernmental organization. The GAC has moved beyond a passive, ex post advisory role and is now the site of a parallel policy making process that competes with and often modifies and overrules the multi-stakeholder policy development organs within ICANN. Many of the members of GAC are openly advocating a stronger role for governments in the policy making process. To put the issue as bluntly as possible: if ICANN were to be transformed into a government-led or intergovernmental organization, it would not happen via the ITU or a UN agency, or a new international treaty. The most rational and efficient course of action would be to try to take over ICANN from within, by continuing to empower the GAC. GAC operating rules now require full consensus among all states before it can issue ‘advice,’ which means that, in theory at least, any individual government can block the GAC from doing something. But currently, nothing prevents those rules from changing.

3. Governments and the IANA Transition

Despite its flaws, U.S. government oversight provided a rudimentary, last resort accountability mechanism for ICANN. If ICANN went completely off the rails it was always possible for NTIA to take away the IANA functions contract and give it to someone else. At the very least, periodic renewal of the IANA functions contract provided the community and the U.S. government an opportunity to hold ICANN’s feet to the fire. After the announcement of a US withdrawal, it was reasonable to ask: to whom will ICANN be accountable without the tether to the U.S. government? The private actors involved in global Internet governance realized that they needed to develop new, more robust accountability mechanisms for ICANN. Indeed, stronger accountability became a quid pro quo for allowing the transition to happen.

But as important and essential as the IANA transition and enhanced accountability process are, they create risks. Empowering the Internet community to make ICANN accountable means, in effect, a redesign of ICANN, one which can reconfigure the power relationships among stakeholders. Such a redesign cannot help but affect the role of governments in ICANN. So this part of the paper examines the changes proposed by the Cross Community Working Group on Enhanced Accountability (CCWG) in that light. Given the 5th requirement of the NTIA (the transition must not replace the NTIA role with a government-led or an inter-governmental organization solution), the impact of the transition on the role of governments is a critical issue.

The CCWG’s charter, adopted November 3, 2014, defined its objective as: “to deliver proposals that would enhance ICANN’s accountability towards all stakeholders,” and “to identify those [accountability] mechanisms that must be in place or committed to before the IANA Stewardship Transition.” In determining the composition of the CCWG, the GAC was treated as a “Chartering Organization” alongside the three Supporting Organizations (one for generic names, country code names and numbers), and all the other ICANN Advisory Committees (the At Large Advisory Committee, the Security and Stability Advisory Committee, and the Root Server System Advisory Committee). A chartering organization is assured of a certain level of representation on the CCWG, and it is presumed that every Chartering Organization would have to support the final proposal before it could be considered a ‘full consensus proposal’ and passed on to the NTIA for approval and implementation. Thus, in the accountability reform process governments, via the GAC, were given same status as nonstate actors. They were no longer in a separate, advisory role. This approach reflected the casual and increasingly accepted assumption that governments are “just another stakeholder” and that all ICANN-based entities, whether they are SOs or ACs, should be treated the same.

3.1 The Proposals of the Accountability CCWG

After nearly a year of contentious deliberations, the CCWG released its proposed reforms for public comment in August 2015. The plan contained four basic elements:

  1. A revised, narrower definition of ICANN’s mission, core values and fundamental commitments. The revisions are intended to clarify that ICANN’s mission is limited to “coordinating the development and implementation of policies …to ensure the stable and secure operation of the DNS…” The revisions are also intended to ensure that “ICANN’s Mission does not include the regulation of services that use the DNS or the regulation of the content these services carry or provide,” and that ICANN’s powers are ‘enumerated;’ i.e., anything not articulated in the Bylaws are outside the scope of ICANN’s authority. The reforms also obligate ICANN to conduct its activities in accordance with fundamental principles.
  2. In order to enforce the limitations and commitments in the revised mission, core values and commitments, the reforms strengthen and make more accessible the Independent Review Process (IRP). In effect, the CCWG created an ‘independent judiciary’ for ICANN’s regime of global governance. This, in combination with the limited mission, is perhaps the most critical aspect of the accountability reforms.
  3. In order to prevent ICANN’s board from simply amending its bylaws to undo the accountability reforms coming out of this process, the CCWG proposed to create a category of bylaws known as ‘fundamental bylaws’ that could only be changed with supermajorities and the approval of the community. The reforms also gave the community the authority to approve normal bylaw changes.
  4. The reforms created new ‘community empowerment’ mechanisms that use the membership aspect of California Nonprofit Public Benefit Corporation law to approve or disapprove board actions. Using a complicated process known as the Sole Member Community Mechanism (SMCM), community empowerment gives the Supporting Organizations and some Advisory Committees (including the GAC), as a collective unit, the power to prevent ICANN’s board from changing fundamental bylaws, the power to approve the budget, and the power to remove board members or the entire board.

The GAC and governments figured prominently in the discussion of these reforms. Although the issue was never posed quite so explicitly, the CCWG had to face a critical question: was the GAC part of the community that needed to be empowered to make ICANN’s board more accountable? Or was GAC, with its special relationship to the board and the coercive power of states behind it, a big part of the accountability problem that also needed to be carefully checked and balanced?

There were three main battlefields over the role of governments and the GAC. One major point of controversy was a Core Value that dealt with ICANN’s relationship to governments. Another was Stress Test 18, which considered how the reforms would work if governments attempted to make decisions by majority rule rather than by consensus. The third was a debate over the vote allocations in the Sole Member Community Mechanism. Each of these issues will be taken up in turn, below.

Core Value 7

Core value 7 articulates ICANN’s commitment to take into account advice from governments. In its initial proposal, the CCWG proposed to formulate this Core Value as follows:

While remaining rooted in the private sector, including business stakeholders, civil society, the technical community, and academia, recognizing that governments and public authorities are responsible for public policy and duly taking into account the public policy advice of governments and public authorities, in accordance with the Bylaws and to the extent consistent with these Fundamental Commitments and Core Values.

By echoing the WSIS Tunis Agenda’s claim that stakeholders have different ‘roles’ and that the role of governments is to be exclusively responsible for public policy, this core value already constitutes a major concession to governments. But the GAC representatives in the CCWG objected to the wording of this Core Value. Speaking in the GAC meeting, the government of Portugal expressed outrage that governmental advice should be limited in any way.

Our position is that the public-policy issues obviously have to be dealt with by the government. Obviously governments [should] not be limited in [any] respect. So any wording in the proposal [providing for] a limitation to the role of governments, I think that this would be unacceptable, because you cannot limit the role of governments in these kind of issues.

It is important to note that the last sentence of Core Value 7 is not a limitation on what advice governments can offer, but a limitation on what the ICANN board can do. The Core Values are supposed to guide ICANN’s corporate behavior, not the behavior of governments, and Core Value 7’s main intent was to restrict the board to following GAC advice only if it was “in accordance with the Bylaws” and “consistent with [ICANN’s] Fundamental Commitments and Core Values.” But Portugal, and most other governments, interpreted this as a limitation on the kind of advice they could offer, and argued strongly against it in the CCWG. Does that mean that governments think ICANN should follow their advice even if doing so would violate its mission and core values? No one seems to have raised that question in the CCWG. In the end, the GAC objections succeeded in deleting those limiting phrases from the last sentence. The core value now reads:

While remaining rooted in the private sector, including business stakeholders, civil society, the technical community, and academia, recognizing that governments and public authorities are responsible for public policy and duly taking into account the public policy advice of governments and public authorities.

This implies that GAC can offer advice advocating any action it likes in the name of “public policy,” and the ICANN board must ‘duly take it into account.’ This could defeat the entire purpose of the reform process related to keeping ICANN within a narrow and well-defined mission. If “advice from public authorities” constitutes a get-out-of-jail-free card for the ICANN board, then ICANN can do anything, including abuse its powers and break its own bylaws, so long as governments tell it to. Note that the current language does not even specify that the advice ICANN will ‘duly take into account’ must be formal consensus advice from the GAC. Apparently, any public policy advice from any public authority or government will do. This is not an enhancement of ICANN accountability but an abdication of it.

The CCWG responded to these concerns by saying, “we propose to clarify that the Independent Review Process applies to all violations of the ICANN Bylaws, including violations resulting from ICANN’s action or inaction based on input from advisory committees or supporting organizations.” In other words, the IRP could be used to challenge and overturn any board action based on GAC advice that violated the bylaws. But if this is true, why would the governments object to having that limitation stated in the core value? If the advice governments provide can only be implemented if it is consistent with ICANN’s mission, core values and bylaws, why not say so right there in Core Value 7, which defines ICANN’s relationship to public authorities? The elimination of this language is thus cause for concern. At the very least it sends a confusing signal to governments about their proper role; at worst, it might provide a loophole.

Stress Test 18: Consensus or Majority Votes in the GAC?

Recognizing the danger of outsize governmental influence on the ICANN process, the CCWG proposed an amendment to the bylaws that would prevent the GAC from offering formal advice unless it had a full consensus of its members. In other words, the ICANN’s board’s obligation to take GAC advice into account, and to find a mutually acceptable solution would only be triggered if the GAC had full consensus. A majority vote would not have the same bylaw status as consensus advice. In this case, the CCWG did the right thing. It tried to impose a permanent requirement that all GAC advice be consensus advice. This would give any individual government in the GAC veto power over the substantive advice the committee could forward to the ICANN board.

Yet many members of the GAC viewed this, too, as an unacceptable restraint on the GAC’s freedom of action. The government of Spain, for example, said in a GAC meeting,

“We have to maintain the status quo of GAC as first among equals in the ICANN ecosystem. This is key. Therefore, the GAC should not suffer any reduction in terms of its legitimacy or capacity to provide advice to the board. So stress test 18 … should not be accepted, because in no case can a government accept a limitation to its capacity to provide advice.”

The CCWG-Accountability responded to these concerns by retaining the bylaw amendment specifying that only consensus advice would trigger the obligation on the ICANN board to try to find a mutually acceptable solution. This would allow the GAC to change its Operating Principle 47 to use majority voting for formal GAC advice, and allow GAC to give ICANN advice at any time, with or without consensus. But ICANN bylaws would require the board to attempt to find a mutually acceptable solution only on advice that had GAC consensus.

GAC and the SMCM

As noted above, the reforms attempt to strengthen bottom-up governance by ICANN stakeholders by creating a Sole Member Community Mechanism (SMCM). The SMCM is designed to give the community around ICANN – the ‘bottom’ of the so-called ‘bottom-up process’ – real authority over the ICANN board. These powers include the power to prevent ICANN’s board from changing fundamental bylaws, the power to approve the budget, and the power to remove board members or the entire board. This is a long-overdue reform that would truly enhance ICANN’s accountability.

The devil, of course, is in the details. Who speaks for the community and what is the process by which its will can be expressed? The CCWG answered this question by relying on ICANN’s existing Supporting Organizations and Advisory Committees to make up the Community Mechanism. (This was not surprising, because the CCWG itself was composed of these SOs and ACs.) A specific number of votes in the SMCM was allocated to each AC or SO, and specific supermajorities was required for each type of action. For example, any attempt by the board to change a fundamental bylaw would have to be approved by a 75% vote of the SMCM.

But the CCWG’s proposed voting allocations became a serious point of controversy. Reflecting the balance among its own chartering organizations (rather than a more rational calculation regarding accountability) the CCWG proposed the following vote allocation:

  • 5 – Generic Names SO
  • 5 – Address SO
  • 5 – Country Code Names SO
  • 5 – At Large AC
  • 5 – Governmental AC
  • 2 – Root Server System AC
  • 2 – Security and Stability AC

This allocation of voting power, if implemented, would dramatically change ICANN’s nature. Inexplicably, it gives almost half (48%) of the votes to Advisory Committees (including the GAC). This is deeply problematic for a number of reasons. First, Supporting Organizations, not ACs, are intended to be the primary representational and policy development organs of ICANN. Second, SOs are the entities that elect board members. With the partial exception of the At Large Advisory Committee (ALAC), which appoints a board member and plays a major role in the board Nominating Committee, all other Advisory Committees have no role in board selection. They are there to provide expert or specialized advice to the board.

Third, and most relevant to this paper, it is troublesome for the GAC, with its special advisory role, to also be part of the SMCM. Given that ICANN’s articles and bylaws have always barred governments from appointing members of the ICANN board, it is inappropriate for the GAC to play a role in voting on the removal of board members. Overall, the CCWG proposal alters the role of the GAC. Suddenly, GAC becomes a stakeholder group of the same type as the ‘individual internet users’ who supposedly comprise the ALAC, the domain name users and suppliers in the GNSO, the managers of country code top level domains in the CCNSO, and the Regional Internet Registries in the ASO. Suddenly GAC is standing alongside all these other interest groups voting directly on bylaw changes, on whether policy decisions should be upheld or challenged by the IRP, approving or disapproving ICANN’s budget, and so on.

But if governments just want to participate in ICANN as ‘normal’ stakeholders like business, civil society and technical people, does it still make sense for their policy advice to have a special, privileged status? Aren’t they double dipping if they do so, claiming a distinct and privileged status on the one hand while demanding ‘equal footing’ in the SMCM on the other? Two analysts at the Heritage Foundation said:

In our view, this dual role for governments is unacceptable and would provide them enhanced authority within the post-transition ICANN. We strongly believe that an acceptable proposal must require the GAC to either choose to participate in the Sole Member or retain their privileged advisory role, but not have both opportunities to influence ICANN.

The Advisory Committee on Security and Stability (SSAC) expressed concern about the way participating in the SMCM would undermine its mission. In a statement before the GAC, the Chair of the SSAC said:

We have drawn the conclusion that we don’t have any special standing for the advice we give other than it be evaluated on its merits… We are concerned about the way in which the proposed new SO/AC membership model might affect the way SSAC operates, considering its narrow focus on security and stability matters and its reluctance to become involved in issues outside of that remit.

Even ICANN itself, in its comments on the CCWG report, expressed concern about the “still unresolved issue of changing the role of advisory committees and considering the extent to which groups may be gaining new powers either disproportionate to their participation within ICANN today, or gaining powers disproportionately to other SOs or ACs.”

To summarize this section, the accountability reforms proposed by the CCWG have many positive features, but they also could alter the role of governments in ICANN in negative ways. The GAC demanded modification of Core Value 7 in a way that shows that governments do not respect the limited mission that ICANN reformers are aiming for. Indeed, there are many within the GAC who want governments to be more than just advisory. The debate over Stress test 18 shows that many GAC members want it to be possible for GAC advice to emerge from a majority vote in a one government, one-vote system like the UN General Assembly. And the proposed voting allocations in the new Sole Member Community Mechanism further enhance the powers of the GAC.

5. Why is it Important for IG to be Independent of Governments?

With all the discussion about the role of governments in ICANN, it is useful to back up and reconsider some first principles. The NTIA requires that the post-transition ICANN not be “government-led” or turned over to an intergovernmental institution. This principle has broad support. But why is it important? What is wrong with being “government-led?” The NTIA also requires that the transition “support and enhance the multistakeholder model.” But what is “the multistakeholder model” and what makes it better than the alternatives? Presumably, the MS model includes governments, but by itself the term does not specify what role governments should have in it. Defending the “multistakeholder model,” therefore, does not really provide a very good reason why intergovernmental institutions or government-led solutions are to be avoided.

If we don’t understand what the purpose of limiting governments’ role is, we could easily lose sight of that purpose while instituting complicated changes in ICANN. The results may worsen the problems associated with governmental participation in Internet governance. This section of the paper explains why ICANN should not be government-led or dominated.

The first point is that governments are not “just another stakeholder” in a multi-stakeholder system. Governments represent an alternate governance regime founded on a completely different logic of representation and decision making. Every national government claims to be a sovereign power, which means that it possesses ultimate authority over everyone in their territory; they claim to be the legitimate representative of every single person in their jurisdiction. It is that claim, which is fundamental to what national governments are, that led to the WSIS Tunis Agenda’s statement that the role of national governments is to hold supreme authority over the development of “public policy” for the Internet. Other stakeholder groups, such as civil society and business, are (according to the WSIS Tunis Agenda) supposed to fulfill different roles. Policy is for states.

The ICANN model was based on a completely different idea. ICANN was based on the idea that the Internet community would make its own public policies. Public policy would be developed directly by all stakeholders, not exclusively by state-based representatives. ICANN is an institution for the development of global public policy. It is not surprising, then, that the CCWG proposal correctly portrays ICANN as paralleling traditional governmental structures, with a constitution (the Articles of Incorporation and corporate bylaws) a legislative branch (the Supporting Organizations), an executive branch (the board and staff), and a judicial branch (the Independent Review Process). ICANN is an alternate government for the global DNS. In some respects governments are competing with it for power. At any rate there is a contradiction between the ability of stakeholders in the ICANN regime to represent themselves directly in the multistakeholder policy development process, and the claim of governments that they represent everyone in their territory and hold a monopoly over legitimate policy making. Combining these two principles in one institution, as ICANN does with the GAC, is inherently unstable.

If one interrogates the meaning of the term “public policy” one finds that there is no inherent limitation on the term. Public policy refers to whatever social activity a public authority wants to influence, for any reason deemed to be in the state’s or its citizens’ interest. Ergo, everything ICANN does could be considered public policy or relevant to public policy. If governments truly do have a special role that gives them exclusive authority over public policy related to the domain name system, then ICANN should be run by governments; it should be part of an intergovernmental institution. That, of course, is not what most participants want, but it explains why consistent advocates of sovereignty, such as Brazil or China, want to change the role of governments in ICANN to make them more than mere advisors.

The entire purpose of ICANN was to provide a global governance regime that was not based on national governments and their intergovernmental treaty organizations. Why was an alternate governance regime necessary?

The reasons are deep and fundamental. First, governmental authority is based on territorial jurisdictions. State jurisdiction fragments the world into two hundred separate, autonomous islands. The internet, on the other hand, is global in scope, and governance of Internet identifiers such as domain names and IP addresses requires globalized coordination and globalized policies to support it. The main argument for so-called multistakeholder model, then, is that it facilitates globalized policy making. It engages a transnational political community focused more on the functioning and freedom of the internet than on the geopolitical interests and security of the state. Basing global governance on non-state venues, actors and institutions, interacting through private contracts rather than state-based treaties or laws, avoids fragmenting Internet Governance into 192 different jurisdictions, which would degrade the global nature of information flows and services on the internet.

A nonstate actor-based governance regime is also important because governments – including the U.S. – have powerful incentives to restrict open and transnational nature of the Internet in the interests of the power or security of their own state. A multistakeholder regime based on a transnational community of civil society, the private sector, and the technical community is more likely to protect and preserve Internet freedom and innovation. Situating Internet governance in nonstate actors who are directly involved with Internet usage and services means that the social and economic value of communication takes priority over the military and political uses of information. Basing internet governance on states, in contrast, makes the Internet a plaything of geopolitical and military rivalries among powerful states. Worst case, it could lead to one country, one vote coalitions in which democratic and liberal values get outvoted by authoritarian regimes.

It should be added that in some respects governments are stakeholders, too. Government agencies are users of Internet technology and services, and sometimes suppliers of network services. This kind of stake, however, takes place not at the national level but at the agency level. Within a single government, different agencies often have very different interests and ideas about appropriate policy. For example, the U.S. Library of Congress might have a very different view of information policy issues than the U.S. Department of Defense; and a European country’s Ministry of Justice might have a different view from its Data Protection authority. In a truly multistakeholder system, government agencies should participate pluralistically, in their capacity as users or suppliers of Internet services on a par with other users and suppliers, rather than as “supreme” policy making authority for every agency and every citizen.

6. Looking Ahead

The elimination of the special U.S. Government role as overseer and contractor of ICANN is necessary and long overdue. The reform of ICANN’s accountability deficit is a necessary corollary of that process, and is also long overdue. The overall goal of the transition is to complete the experiment in governance by nonstate actors started in 1997, while ensuring that ICANN, once free of state control, is accountable to Internet users and suppliers.

The transition and accountability reforms have made explicit and visible an ongoing clash between two distinct principles and methods of Internet governance, one based on state sovereignty and the other based on nonstate actors in a globalized space. This contradiction must be resolved properly if the ICANN regime is not to be made an instrument of enhanced state control over the internet. There is some danger that the GAC’s role in ICANN will be enlarged in a way that will start a long term erosion of ICANN’s unique status as a nongovernmental policy making institution. In order to avoid this outcome, we need to have a clear idea why multistakeholder institutions must be protected from domination by governments. Governments are not “just another stakeholder” and it is a confused version of multistakeholder governance that tries to put states on “equal footing” with other stakeholders.

This paper has shown that governments represent an entirely different type of governance regime that cannot easily coexist with a private sector-based, bottom up governance model based on open participation. Consequently, the GAC’s role in the ICANN process must be carefully contained, if not rolled back. It should not participate in the Sole Member Community Mechanism, its advice should be restricted to full consensus advice, and its advice, even if based on a consensus of states, should never be allowed to override the limits imposed upon ICANN by its mission, core values and commitments.


Milton-Mueller-2ABOUT THE AUTHOR: Milton Mueller is an internationally prominent scholar specializing in the political economy of information and communication. The author of seven books and scores of journal articles, his work informs not only public policy but also science and technology studies, law, economics, communications, and international studies. His books Networks and States: The global politics of Internet governance (MIT Press, 2010) and Ruling the Root: Internet Governance and the Taming of Cyberspace (MIT Press, 2002) are acclaimed scholarly accounts of the global governance regime emerging around the Internet. Mueller’s research employs the theoretical tools of institutional economics, STS and political economy, as well as historical, qualitative and quantitative methods.

Dr. Mueller’s prominence in scholarship is matched by his prominence in policy practice. He is the co-founder and co-director of the Internet Governance Project (IGP), a policy analysis center for global Internet governance. Since its founding in 2004, IGP has played a prominent role in shaping global Internet policies and institutions such as ICANN and the Internet Governance Forum. He has participated in proceedings and policy development activities of ICANN, the International Telecommunications Union (ITU), the National Telecommunications and Information Administration (NTIA) and regulatory proceedings in the European Commission, China, Hong Kong and New Zealand. He has served as an expert witness in prominent legal cases related to domain names and telecommunication policy. He was recently elected to the advisory committee of the American Registry for Internet Numbers (ARIN), and appointed in 2014 to the IANA Stewardship Coordination Group. Dr. Mueller has also been a practical institution-builder in the scholarly world, where he led the creation of the Global Internet Governance Academic Network (GigaNet), an international association of scholars.


[1] IANA stands for Internet Assigned Numbers Authority. The function of the IANA is defined in RFC 2434 (1998) as follows: “Many protocols make use of identifiers consisting of constants and other well-known values. Even after a protocol has been defined and deployment has begun, new values may need to be assigned … To insure that such quantities have consistent values and interpretations in different implementations, their assignment must be administered by a central authority. For IETF protocols, that role is provided by the Internet Assigned Numbers Authority (IANA).”

[2] Amendment 11 of the NSI Cooperative Agreement, made October 7, 1998, mandated competing registrars and also required the U.S. government to approve any root zone changes. https://archive.icann.org/en/nsi/coopagmt-amend11-07oct98.htm

[3] From the November 23 1998 bylaws of ICANN.

[4] GAC, Principles for the Delegation and Redelegation of ccTLDs, 23 February 2000. https://archive.icann.org/en/committees/gac/gac-cctldprinciples-23feb00.htm

[5] See Paragraphs 35 and 35 a), Tunis Agenda for the Information Society, WSIS-05/TUNIS/DOC/6(Rev.1)-E, 18 November 2005 http://www.itu.int/wsis/docs2/tunis/off/6rev1.pdf

[6] Comments of Spain, transcript of the July 24, 2015 GAC meeting in Buenos Aires. The claim that GAC was “first among equals” was also made by the UK, United States, Portugal and other states in the same meeting.

[7] GAC Principles regarding new TLDs. March 28, 2007. https://archive.icann.org/en/topics/new-gtlds/gac-principles-regarding-new-gtlds-28mar07-en.pdf

[8] See Annex 1 of the Beijing GAC communique, April 11, 2013. https://gacweb.icann.org/download/attachments/27132037/Beijing%20Communique%20april2013_Final.pdf?version=1&modificationDate=1365666376000&api=v2