Summary

  • GAC advice is not a legal command. The Board retains decisional authority, and the Bylaws expressly provide an exit when no mutually acceptable solution can be found. Consensus advice nevertheless requires at least 60 percent of the Board to reject it, written reasons, and good-faith consultation.
  • The real cost of rejection is cumulative: a higher voting threshold, delay, repeated explanation, negotiation, diplomatic friction, exposure to accountability review and the risk that governments pursue the same policy through national or regional law. That burden can influence a decision before any formal rejection vote occurs.
  • The .XXX decision in 2011 proves that refusal is possible; the .AMAZON dispute proves that deference without independent merits analysis can also be costly; and the 2023-2025 auction advice shows how consultation can replace the requested measure with a negotiated alternative while the Board formally rejects the advice as first presented.

A veto can exist as pressure without existing as law

The easiest answer to the title is no. Governments do not possess a veto over ICANN Board decisions merely because they act through the Governmental Advisory Committee. The current ICANN Bylaws do not require government consent to every policy, contract, budget or delegation. The GAC's own Operating Principles say that the committee has no legal authority to act for ICANN. Its Board liaison does not vote. Advice remains advice.

That answer is legally correct and institutionally incomplete. A veto is sometimes used as a strict noun: a power whose exercise prevents an act. It is also used more loosely to describe a position so costly to oppose that the decision-maker rarely chooses direct conflict. The second meaning is the one that matters here. ICANN has built a sequence around government advice that raises the price of refusal without transferring the final corporate decision to governments.

The sequence has four parts. The Board must duly take GAC advice into account during policy formulation and adoption. If it proposes an inconsistent action, it must notify the GAC and explain why. If the advice has been approved by full GAC consensus, the Board needs at least 60 percent support to reject it. The Board and GAC must then try in good faith, promptly and efficiently, to find a mutually acceptable solution. If they fail, the Board may proceed, but its final decision must again state why the advice was not followed.

Each element is weaker than a veto. Together they create an escalation track that no other advisory committee enjoys in the same form. A Board majority may be ready to disagree but unable to assemble 60 percent. Directors may dislike the delay and diplomatic collision. Staff may redesign an option to avoid triggering inconsistency. A supporting organization may modify a recommendation after learning that governments will oppose it. None of those accommodations appears in a register as an exercised veto, yet each can reflect the shadow of the rule.

The distinction is essential because both exaggerations damage legitimacy. Calling the GAC powerless ignores the special legal weight attached to its consensus advice. Calling it a sovereign chamber mistakes influence for authority and obscures the Board's fiduciary responsibility. A defensible account must identify who can decide, what refusal costs, and whether the public can see the bargaining that occurs between advice and final action.

The Bylaws create deference, not obedience

Article 12 gives the GAC a broad public-policy remit. Governments, distinct economies, multinational governmental organizations and treaty organizations participate in a forum concerned with the intersection between ICANN's work, law, international agreements and public policy objectives. The GAC may raise issues directly with the Board, comment, give prior advice, and recommend action or policy development.

The critical words are “duly taken into account.” They require serious consideration, but they do not define a predetermined result. Due consideration is a duty of attention and reason. It should be visible in the material reviewed, questions asked, competing commitments identified and rationale issued. It is not a duty to accept every proposed measure.

The text confirms this by describing what happens after disagreement. The Board must inform the GAC and give reasons. Consensus advice may be rejected by a 60 percent vote. The parties then seek a mutually acceptable solution. If none can be found, the Board issues a final explanation. A provision that specifies the lawful route after failed negotiation cannot simultaneously make GAC assent indispensable.

The final sentence also preserves the rights and obligations of GAC members on public-policy matters within their responsibilities. That reservation matters. A government that loses inside ICANN does not surrender its legislative, regulatory, diplomatic or litigating powers. It may act under national law, coordinate with other states, participate in regional rulemaking or challenge conduct through another available avenue. The Board therefore faces actors whose influence does not end at the edge of the corporation.

This is one reason ordinary corporate vocabulary understates the encounter. The Board is not negotiating with a group of shareholders, and the GAC is not a ministry with hierarchical authority over ICANN. A California nonprofit is deciding whether to depart from advice supported across a governmental forum. Its legal permission to depart may be clear while the political consequences remain uncertain.

The Board also cannot use government advice as a release from its own duties. It must remain within ICANN's Mission, Articles and Bylaws. It must evaluate relevant facts and explain material public-interest, security, stability, financial and other effects when required. Directors serve as individuals and exercise their own judgment. “The governments asked us to do it” is not a sufficient substitute for a reasoned Board decision.

That independence protects the multistakeholder settlement. It prevents consensus among governments from automatically displacing policy work by registries, registrars, technical bodies, civil society, businesses and individual users. But independence has to be exercised rather than merely asserted. If the Board consistently accepts advice without testing its rationale, the formal absence of a veto becomes less meaningful.

Consensus is both the source of weight and a brake on it

Not every GAC statement receives the strongest treatment. The Bylaws define GAC Consensus Advice as advice approved by full consensus, understood as general agreement in the absence of any formal objection. The GAC must say whether advice sent to the Board has that status. Principle 47 of the GAC Operating Principles uses the same conception and provides that, where consensus is not possible, the Chair conveys the full range of views to the Board.

This threshold performs two contrary functions. First, it gives consensus advice exceptional legitimacy. A position that survives deliberation among many governments and intergovernmental entities deserves attention as evidence of converging public-policy concern. The 60 percent Board rejection threshold and mandatory consultation translate that convergence into institutional weight.

Second, the threshold limits governmental power. A large coalition cannot automatically trigger the strongest Board obligation if another GAC member enters a formal objection. The special status depends on the absence of such an objection, not on a simple majority, population weighting, financial contribution or geopolitical rank. One government's formal opposition can therefore prevent the GAC from speaking with the legally privileged voice of consensus.

The 2016 accountability settlement made this link deliberate. Recommendation 11 of the CCWG-Accountability final report addressed a stress test in which the GAC might move from consensus to majority voting while the Board remained obliged to negotiate a mutually acceptable solution. The settlement restricted the enhanced treatment to consensus advice, raised the Board rejection threshold to 60 percent, and paired those changes with a GAC carve-out in specified Empowered Community challenges.

The designers were balancing two capture risks. One was excessive governmental influence over a private multistakeholder body. The other was a single government using a formal objection to block broadly supported public-policy advice. The final text did not eliminate either risk. It made them visible and distributed them across two institutions: the GAC controls whether it can reach consensus, and the Board controls whether it can assemble the votes and reasons to depart.

Consensus should not be romanticized as unanimity of peoples. GAC representatives speak through governmental and intergovernmental institutions. Their positions may be well consulted or narrowly formed; the GAC record alone cannot establish which. Absence of a formal objection does not prove equal enthusiasm. Resource disparities affect who can attend, study technical material and sustain negotiations. Some governments may remain silent for reasons unrelated to agreement.

Nor should a formal objection automatically be praised as a safeguard. It can protect minority interests, prevent premature closure and expose conflict. It can also become leverage for one state to dilute or block advice. The right question is not whether consensus is good in the abstract. It is whether the record identifies the advice, rationale, objections, qualifications and affected interests clearly enough for others to judge the coalition behind it.

A communiqué is not a single block of governmental will

GAC communiqués are central evidence, but they are easy to misread. A communiqué may contain formal advice, follow-up on earlier advice, issues of importance, factual updates and accounts of discussion. These categories do not all impose the same Board duties. Treating every paragraph as consensus advice inflates governmental authority; treating the whole document as commentary understates it.

The GAC Advice Register and the Board's scorecards were created in part to solve this classification problem. The current GAC advice status report identifies advice items, dates, topics, Board action and later status across many years. A reader can follow an item through clarification, deferral, acceptance, implementation, consultation or closure.

That record is valuable because advice often arrives at a higher level of abstraction than implementation requires. “Protect intergovernmental organization identifiers,” “avoid auctions,” or “improve access to registration data” may express a clear public-policy objective while leaving many operational choices open. The Board must determine what the advice requires, whether the proposed means are feasible and how the objective interacts with existing consensus policy, contracts, law and prior commitments.

Clarifying questions are therefore not necessarily evasion. They can expose whether the GAC is asking for a result, a method, a pause or a study. They can reveal whether an apparent conflict can be removed through a narrower measure. But repeated requests for clarification can also postpone a politically difficult decision. A status label such as “pending” or “in progress” says little unless the record shows what remains unresolved and who owns the next action.

Board scorecards create another ambiguity. An item may be “accepted” in principle while implementation is conditional on feasibility, policy work or further discussion. Advice may be “closed” because the Board responded, not because the requested outcome occurred. The public needs a distinction among consideration, agreement, execution and observed effect.

For this reason, the strongest evidence is a linked chain: communiqué language, consensus designation, Board understanding, clarification exchange, resolution, vote, reasons, consultation record, final action and later implementation. A slogan about governmental influence cannot replace that chain. Neither can a large spreadsheet whose statuses are not explained.

The cost of rejection starts before the vote

The first cost is agenda cost. Once directors believe a proposed action may be inconsistent with consensus advice, the matter requires a dedicated decision on consultation. Notices, briefing materials, calls and correspondence must be prepared. Other dependent decisions may wait. Staff and volunteer attention moves from policy design to conflict management.

The second cost is coalition cost. A simple majority is not enough to reject consensus advice. At least 60 percent of the Board must support refusal. Directors who agree on the ultimate outcome may disagree on the rationale, timing or diplomatic handling. Recusals, absences and vacancies can complicate vote planning. The need for a supermajority makes each uncertain director more consequential.

The third cost is reason-giving. The Board must state why it has chosen not to follow the advice, and if consultation fails, explain again in the final decision. Reasons must survive scrutiny from governments, the rest of the ICANN community and possible accountability proceedings. A weak explanation can be more damaging than delay because it suggests the Board has either ignored public policy or surrendered its independent judgment.

The fourth cost is negotiation. Good-faith consultation is not satisfied by sending a refusal letter and waiting. The parties must try, promptly and efficiently, to find a mutually acceptable solution. That may require changing the proposed measure, gathering evidence, testing legal options and consulting affected groups. Even when the Board ultimately prevails, the final policy may differ materially from the option that first triggered conflict.

The fifth cost is legitimacy. ICANN relies on governments to accept a private, transnational coordination model rather than seeking intergovernmental control of the naming system. A visible rupture can revive arguments that public-policy authority has been delegated too far from states. The Board may lawfully reject advice and still weaken confidence among the officials whose tolerance helps sustain the model.

The sixth cost lies outside ICANN. Governments can regulate registries, registrars, privacy, consumer protection, competition, cybersecurity and access to data within their jurisdictions. They can coordinate positions in other forums. A Board victory over advice can be followed by fragmented legal obligations for contracted parties. That possibility does not make the GAC a legislature, but it gives its members bargaining power.

The seventh cost is review. A disappointed applicant, community body or other eligible claimant may argue that the Board failed to exercise independent judgment, disregarded material information or departed from the Bylaws. Review does not exist to second-guess every policy choice, yet a poor record increases exposure. The .AMAZON dispute shows that mechanically following GAC advice can create review risk just as surely as rejecting it.

These costs are not all improper. Deliberation, reasons and consultation are often the point. They force the Board to confront public-policy concerns rather than dismissing them as external politics. The problem arises when cost becomes disguised control: when directors avoid a defensible option because the route to refusal is too burdensome, or when negotiation occurs without enough visibility to show whose interests changed the result.

The 60 percent rule changes bargaining even when never tested

A voting threshold influences conduct before a motion reaches the Board. The chair and directors will want to know whether the necessary coalition exists. Staff will draft options that can hold it. GAC representatives will understand that broad consensus places dissenting directors in a harder position. Other community bodies will anticipate the same arithmetic.

This is the core of veto-like pressure. A formal veto normally works after a proposal has been adopted or presented: the veto-holder blocks it. Here the pressure can work earlier. An option may be narrowed, delayed or abandoned because leaders expect fewer than 60 percent of directors would vote to reject consensus advice. The public sees no failed vote because none occurs.

The threshold is nevertheless not extraordinary when compared with other ICANN supermajorities. The Bylaws use elevated votes for significant governance acts and for departure from some supporting-organization recommendations. Supermajorities can protect stability, require wider agreement and limit a transient majority. The issue is not that 60 percent is inherently illegitimate. It is that the protected position comes from governments and affects a body designed to avoid governmental control.

The GAC carve-out partly answers that concern. When a Board resolution is based materially on GAC Consensus Advice, the GAC cannot count as a decision-maker in an Empowered Community challenge to the Board's implementation of that advice. This prevents governments from supplying the advice and then using community power to defend the Board's implementation of it. The adjusted thresholds preserve a route for other decisional entities.

That safeguard does not erase agenda power. The GAC may still shape the Board's option set, and other community bodies may find it difficult to challenge a policy framed as globally supported public policy. Conversely, the carve-out can leave the GAC unable to defend an implementation that it believes accurately reflects its advice. Institutional balance is achieved through imperfect counterweights, not through a single clean separation.

Empirical assessment is difficult because deterrence and agreement look alike. If the Board never proposes an inconsistent action, that could mean the advice was compelling, the policy was redesigned early, conflict was avoided for diplomatic reasons or directors expected to lose the supermajority vote. A count of formal rejections cannot distinguish these possibilities.

Meeting records, draft changes, written exchanges and interviews can help. The Board should identify when GAC advice was a material factor, what alternatives were considered, and whether the 60 percent requirement affected timing or substance. Without such evidence, claims that the rule either dominates the Board or never matters remain speculative.

Consultation is an obligation to seek agreement, not to obtain it

The 2013 Board-GAC consultation process gives practical shape to the Bylaw duty. Written notice identifies the advice at issue and the reasons the Board may not follow it. The parties arrange consultation, exchange views and try to narrow differences. The procedure is designed to make disagreement specific enough to resolve.

The phrase “mutually acceptable solution” is powerful because it directs both institutions toward accommodation. It is limited because the duty is to try. The Bylaws anticipate failure and authorize a final Board decision with reasons. Good faith is reviewable as conduct, but it cannot guarantee substantive agreement.

This distinction affects how delay should be judged. Time spent clarifying a genuine ambiguity or testing a compromise can improve policy. Time spent repeating fixed positions can conceal an inability to decide. “Timely and efficient” should therefore be treated as an independent requirement, not as decoration around good faith.

A useful consultation record would show the date of notice, the exact conflict, entities, materials exchanged, alternatives considered, concessions made, unresolved points, deadlines and final assessment. Confidential diplomatic exchanges may sometimes be justified, but a public institution cannot ask observers to infer good faith solely from the fact that meetings occurred.

Consultation should also remain connected to the wider community. A bilateral Board-GAC bargain may resolve governmental concerns while displacing a policy recommendation developed elsewhere. Registries, registrars, applicants, rights holders, users and technical bodies may bear the resulting costs. A solution is not legitimate merely because two institutional actors find it acceptable.

That is why community-wide discussion mattered in the recent auction dispute. The Board and GAC did not resolve the issue only through a private exchange. They used broader discussion to explore alternatives before settling on replacement strings alongside continued use of the established last-resort auction method. The compromise can still be criticized, but the route recognized that public-policy advice intersected with a policy framework built by another part of ICANN.

.XXX proves that the Board can pay the price and still refuse

The .XXX controversy remains the clearest historical answer to the claim that GAC advice is an absolute veto. It also shows how expensive rejection can become.

ICM Registry's application had a long and contested history. ICANN rejected it in 2007. ICM sought independent review, and the first major declaration under that mechanism found serious inconsistency in the way the Board revisited its earlier conclusion that the application met sponsorship criteria. In June 2010 the Board accepted and acted on parts of the panel's findings, performed further due diligence and returned to the proposed registry agreement.

Governmental opposition remained. The Board concluded in December 2010 that entering the agreement would conflict with three items of GAC advice and invoked the special consultation procedure. It sent a detailed position, met the GAC in Brussels, continued discussion at the San Francisco meeting and received a clarifying GAC letter.

The 18 March 2011 Board minutes record the end of that effort. The Board authorized execution of the .XXX registry agreement, adopted a detailed rationale, declared that good-faith consultation had been completed, and stated that no mutually acceptable solution had been found. It then incorporated reasons for not following the GAC advice.

The vote was politically revealing: nine in favor, three opposed and four abstentions. The current 60 percent rule was not yet in force, so the case cannot be used as a direct test of today's arithmetic. It can be used to show that the consultation clause has always contained an exit and that directors were willing, after extensive engagement, to use it.

The cost was measured in years, review proceedings, public comments, due diligence, contract revision, correspondence, an intersessional meeting, Board time, GAC time and a divided vote. Some of that cost arose from the Board's earlier treatment of the application rather than from GAC advice alone. It would be wrong to attribute the whole dispute to governments. Yet the final stage demonstrates the burden attached to a decision that openly departed from governmental advice.

The GAC's March 2011 advice also preserved nuance. Some members neither endorsed nor opposed the top-level domain; others were emphatically opposed and warned that governments might block access. This was not a single global legislature issuing a prohibition. It was a governmental forum conveying public-policy concern strong enough to require a formal response.

The Board's decision did not settle the moral or policy debate. It settled who held final authority under ICANN's governing arrangements. That is the defining difference between veto pressure and a veto: the Board could complete consultation, publish reasons and proceed.

.AMAZON shows the opposite danger: deference without judgment

The .AMAZON dispute is sometimes presented as proof that governments controlled the outcome because the GAC advised that the applications should not proceed and the Board accepted that advice in 2014. The later record tells a more complicated story.

Amazon challenged the decision through independent review. The panel recommended that the Board promptly reevaluate the applications and make an objective, independent judgment about whether well-founded, merits-based public-policy reasons justified denial. The Board's October 2017 response recognized that recommendation and asked the GAC for additional information about the merits.

The lesson is not that GAC advice lacked legitimacy or that geographic and cultural concerns were unreal. It is that the Board could not outsource its decisional responsibility. Deference becomes legally vulnerable when the record does not show an independent assessment of reasons, affected interests and governing criteria.

The Board then spent further years seeking a mutually acceptable solution among GAC members, the Amazon Cooperation Treaty Organization states and the applicant. In May 2019 it concluded that Amazon's proposed commitments were acceptable and directed continued processing, while explaining the elapsed time, exchanges and balancing it had undertaken. Governmental concern continued, and the GAC later clarified that a follow-up request was not new advice.

This case reveals a two-sided accountability trap. Reject consensus advice too casually and the Board appears dismissive of public authority. Accept it too mechanically and the Board appears to have surrendered its judgment. The safe route is not automatic compromise. It is a record showing that advice was taken seriously, reasons were tested, affected parties were heard and the Board itself decided.

The case also qualifies claims about governmental veto power. The original advice delayed and initially stopped the applications. It imposed substantial cost on the applicant and ICANN. Yet review reopened the decision, and the Board eventually moved forward despite continued governmental opposition. Influence was formidable but not final.

There is a broader institutional point. Review bodies should not decide the underlying public policy merely because the Board's reasoning was deficient. Their role is to enforce governing commitments and fair decision-making. A remand can require the Board to exercise judgment without dictating which policy it must select. That separation preserves both accountability and decisional responsibility.

The auction case shows how rejection can become negotiated substitution

The most useful modern case began with the GAC's June 2023 advice on contention-set auctions in the next new generic top-level-domain round. The GAC advised the Board to take steps to avoid auctions of last resort between commercial and noncommercial applications and suggested that alternatives such as drawing lots might be explored.

The Board deferred action while related policy recommendations were considered. On 8 June 2024, it formally determined that its intended action was or might be inconsistent with the consensus advice. Its rationale identified practical problems in classifying applications by business-plan characteristics, possible legal constraints on drawing lots, and inconsistency with adopted policy recommendations that retained last-resort auctions. It initiated the Bylaws consultation.

This resolution is valuable because it does not hide disagreement behind an indefinite status. The Board identified the exact advice, stated why it might refuse, and opened the special route before final action. The GAC then issued further advice seeking urgent community-wide discussion of alternatives.

Discussion produced a narrower option: applicants could submit a replacement string to reduce the likelihood of contention, while the established ascending-clock auction remained available. Board and GAC correspondence recorded this as a mutually acceptable solution. The 26 January 2025 scorecard then used an apparently paradoxical formulation: the Board rejected the advice as originally presented and adopted the mutually acceptable solution.

That outcome shows why binary statistics mislead. The GAC did not obtain its stated suggestion in full. The Board did not simply preserve its first design. Consultation altered the option set. The eventual substitute addressed part of the concern without displacing the adopted auction framework.

Was this a governmental veto? Formally, no. The Board rejected the original advice. Functionally, the advice prevented the Board from proceeding on its initial path without a year of notice, dialogue, broader discussion and design change. The pressure was veto-like because the cost of unmodified refusal produced accommodation. The result was not veto-like because the GAC did not control the final instrument and accepted a substitute.

The case also shows the constructive side of the burden. A higher cost of disagreement can force institutions to discover a better option than either opening position. Replacement strings may reduce contention for less-resourced applicants while avoiding a legally questionable lottery. Whether they work as intended will require evidence from the next round, but the design emerged from conflict rather than from capitulation.

Political cost is real even when legal cost is bounded

Governments bring forms of authority that other ICANN entities do not. They legislate, regulate, investigate, prosecute and represent states. They can affect the operating environment of registries, registrars and users. Their shared advice therefore carries political meaning beyond its Bylaw status.

But the GAC is not a treaty organization directing ICANN. Its members differ in constitutional form, rights commitments, market interests and Internet policy. Consensus may be difficult precisely because these differences are real. A communiqué is a product of committee diplomacy, not a universal public-interest oracle.

The Board's political cost depends on the subject. Refusing advice about a narrow meeting procedure is not equivalent to refusing advice about geographic names, access to registration data, child protection or intergovernmental identifiers. Salience, legal exposure, public mobilization and the unity of governmental support all matter. The 60 percent rule is constant; the political price is not.

The identity of dissent also matters. Opposition from one government may block GAC consensus but carry little wider weight, or it may signal a geopolitical division that makes Board action dangerous. Conversely, consensus among attending representatives may attract little attention in capitals. Formal status is a poor proxy for political intensity.

ICANN should therefore resist two temptations. It should not describe every departure as a brave defense of multistakeholderism. Sometimes government advice identifies a genuine legal or public-interest defect. It should also not describe every accommodation as proof of consensus. Sometimes the Board changes course because the institutional price of refusal is high.

Candor is more stabilizing than mythology. The Board can say that government views materially changed its risk assessment while still defending its independent decision. The GAC can acknowledge compromises and minority concerns without weakening its role. Other community bodies can explain where their recommendations yielded and why.

Government advice competes with other forms of legitimacy

ICANN's design deliberately distributes authority. The GNSO develops policy for generic names through a structured, bottom-up process. Advisory committees contribute security, user and governmental perspectives. The Board evaluates recommendations under specified standards and holds ultimate corporate responsibility. The Empowered Community can exercise defined powers. Review mechanisms test compliance.

GAC advice enters this design as a privileged input, not as the only source of legitimacy. A policy recommendation may reflect years of open work and carefully balanced interests. Consensus advice may identify public law or public-interest consequences that the policy body underweighted. The Board may be unable to satisfy both without revision.

When that happens, the Board should map the conflict. Which part of the policy recommendation would be displaced? Which governmental concern is supported by evidence? Can the objective be achieved through implementation rather than policy change? Who bears the cost? Does the proposed accommodation remain within the Board's authority? Is further policy work required?

A bilateral answer that ignores these questions can turn the GAC into a late-stage amendment chamber. That would weaken the predictability of bottom-up policy development. The opposite approach—treating consensus advice as a comment after the decisive work is complete—would make the GAC's special status hollow.

Early engagement is the usual answer, but it is not a cure. If governments participate early and fail to persuade the policy group, later consensus advice can look like an appeal from an unfavorable result. If they wait until the Board stage, other entities may say the concern arrived too late. Different governmental resources and decision cycles make sustained participation difficult.

The Board must therefore examine both substance and timing without inventing a forfeiture rule that the Bylaws do not contain. Late advice may still expose a serious legal problem. Repeated relitigation of a settled disagreement may still undermine procedure. Reasoned classification is more credible than a categorical preference for either governments or the policy body.

The hidden veto is strongest when evidence is weakest

Formal rejection records capture only visible conflict. They do not show proposals never advanced, language softened before publication, advice negotiated before a communiqué or Board options discarded because support seemed limited public evidence. These are precisely the places where veto-like pressure is most likely to operate.

The absence of records invites motivated stories. Critics can claim that governments dictated the result behind closed doors. Officials can claim that consensus advice merely confirmed what the evidence already required. The Board can describe a compromise as independent judgment. Without a traceable chronology, none of these accounts can be tested.

Confidentiality sometimes has a legitimate role. Diplomacy can fail if every exploratory concession is immediately public. Legal advice and security details may need protection. But confidentiality should protect defined material, not erase the existence, entities, subject and outcome of engagement.

ICANN could publish a GAC influence statement for every consequential Board decision involving advice. It would identify the advice, consensus status, Board vote threshold, alternatives considered, consultations held, material changes, unresolved objections and implementation owner. Where details are withheld, it would state the category and justification.

The GAC could complement this record by describing how consensus was tested. It need not publish every negotiating position. It should make clear whether formal objections were sought, whether members recorded reservations, and whether the advice represents a new position or a continuation of earlier text.

Other bodies should be able to show how their work was affected. If a GNSO recommendation changes because of GAC advice, the public should see whether the change went back through the policy body, occurred at implementation, or was imposed by the Board. Institutional location determines accountability.

A rejection ledger would measure cost without pretending to measure motive

The GAC Advice Register should evolve from a status archive into a decision ledger. For each consensus item, it could record the date received, requested action, material rationale, Board clarification questions, response date, initial Board position, whether inconsistency was identified, consultation dates, final vote, final disposition and implementation evidence.

Time should be measured in segments. How long did the Board take to acknowledge advice? How long before it adopted a scorecard? How long was an item deferred? How long did consultation last? How long from compromise to implementation? One total duration cannot reveal where accountability failed.

Disposition needs a controlled vocabulary. “Accepted” should distinguish acceptance of objective from acceptance of method. “Rejected” should distinguish direct refusal from refusal with a negotiated substitute. “Closed” should state whether action was completed, superseded, withdrawn or merely answered. “Pending” should identify the next responsible institution.

Votes should be visible. The record should show whether the 60 percent rule was engaged, the number eligible to vote, recusals, abstentions and the exact result. A formal threshold without voting evidence cannot be evaluated.

Reasons should be linked to claims. If legal feasibility drives refusal, publish the legal issue at a level compatible with privilege. If adopted consensus policy constrains the Board, identify the recommendation. If cost or security matters, state the evidence and uncertainty. Generic appreciation for the GAC is not a rationale.

Consultation effects should be explicit. What changed between notice and final action? Which change addressed which concern? Did the GAC confirm acceptability? Were affected community bodies consulted? The auction case provides a useful model because the substitute and correspondence can be identified.

The ledger should also record challenges. Reconsideration requests, independent reviews and other formal disputes can reveal whether the Board exercised its own judgment and complied with procedure. A later remand or reversal should update the original advice entry rather than appearing in a disconnected archive.

No ledger can establish motive. A director may vote for compromise because it is substantively better, politically safer or both. The objective is narrower: make the cost-bearing steps and decision changes observable. That evidence will support more disciplined inferences about influence.

Reform should preserve friction while reducing obscurity

Eliminating the special treatment of GAC advice would make Board autonomy clearer but could push governmental conflict into less transparent and more fragmented venues. Turning advice into a formal consent right would weaken the multistakeholder design and allow public officials to displace other affected groups. The better course is to preserve structured friction and improve its visibility.

First, the Board should make the trigger decision early. If an intended action may be inconsistent with consensus advice, it should say so rather than holding the item in an undefined state. Early notice gives the GAC and other bodies a fair chance to assess options.

Second, every advice item should include a rationale sufficient for the Board to evaluate mission, law, proportionality, affected interests and implementation. Consensus around a conclusion is not a substitute for reasons. Advice without an adequate rationale should still receive respectful consideration, but the Board should identify the gap.

Third, consultation should have a public timetable. Extensions may be necessary, but they should be explained. A duty to act promptly loses force if no one can see when the clock started or why it stopped.

Fourth, compromises that alter a supporting-organization recommendation should return to that body when the change is material and the Bylaws require further engagement. The Board-GAC relationship should not become an alternate policy route.

Fifth, the GAC should publish reservations and the treatment of formal objections in a consistent format. This would reduce the temptation to read consensus as uniform enthusiasm and help the Board assess the breadth of support.

Sixth, independent reviewers should sample accepted as well as rejected advice. Excessive governmental influence is more likely to hide in unquestioned acceptance than in a well-documented refusal. Review should ask whether the Board identified merits-based reasons and exercised independent judgment.

Seventh, ICANN should report implementation effects. Did the compromise solve the public-policy problem? Did it impose unanticipated cost? Did governments legislate anyway? Did a supposedly closed item return? Advice quality can be judged only through outcomes, not respectful correspondence.

Finally, institutional language should remain exact. “GAC Consensus Advice” is a defined category. “Mutually acceptable solution” describes an accommodation, not proof that every government or stakeholder consented. “Rejection” can coexist with a substitute. Precision prevents political weight from being mistaken for legal command.

Conclusion

The GAC does not possess a formal veto over the ICANN Board. Its own rules deny authority to act for ICANN, its liaison does not vote, and the Bylaws explicitly permit the Board to proceed after good-faith consultation fails. The .XXX decision demonstrates that the exit is real.

Yet the exit is deliberately expensive. Consensus advice raises the rejection vote to 60 percent, compels reasons, initiates negotiation and exposes the Board to political, reputational and review risk. Governments retain powers outside ICANN. The cumulative burden can alter proposals before a rejection vote is ever scheduled.

That is why “veto pressure” is a more accurate description than either veto or ordinary advice. It captures influence over the option set without transferring final authority. The .AMAZON dispute shows that the Board cannot escape responsibility by deferring automatically. The auction dispute shows that refusal may end in a negotiated substitute rather than a binary victory.

The arrangement can be legitimate. Public-policy concerns deserve a route that is harder to ignore than an ordinary comment, while a private multistakeholder coordinator must preserve independent judgment. Legitimacy depends on showing the route from advice to decision: consensus, rationale, vote, consultation, compromise, final reasons and effect.

The decisive question is therefore not whether governments can issue a command. They cannot. It is whether the institutional price of saying no is visible, proportionate and compatible with the authority the Board is still required to exercise. Where that price shapes policy in silence, the unnamed veto is strongest. Where the record shows disagreement, negotiation and a reasoned final act, pressure remains accountable without becoming rule.

Sources