The Birth of a Treaty

ILPI Publications > Background Papers

A brief overview of mandates and models for negotiating multilateral agreements in the field of arms and disarmament

December 2016


In December 2016, a majority of the UN Member States voted to start a process of negotiation to prohibit nuclear weapons. By adopting resolution L.41, the United Nations General Assembly decided to “convene in 2017 a United Nations conference to negotiate a legally binding instrument to prohibit nuclear weapons, leading towards their total elimination”.[1]

The purpose of this background paper is to provide a brief overview of some of the procedural rules and models that are typically applied for treaty negotiations, starting with adoption of the mandate to negotiate a treaty. It also discusses key features of the procedural regime for treaty negotiations including membership, participation and decision-making. It looks at practice both from under and outside the auspices of the UN.


Historically, the negotiation of multilateral treaties was done through diplomatic conferences on an ad-hoc basis, on the invitation of one or more states, and conducted as autonomous meetings of plenipotentiaries. This practice continues to date—with the conventions on anti-personnel mines (APMBC) and cluster munitions (CCM) as two recent examples—but over the past decades, treaty negotiations have increasingly also taken place within the framework of the United Nations as well as other international organisations. Mandates have usually been adopted by the member states, for example through resolutions, and negotiations have typically taken place on UN (or other IGO) premises, with the support of the relevant organisation’s secretariat.

Treaty negotiations in the United Nations framework tend to bring together a larger group of states than do ad-hoc conferences, and while the latter type is usually (though not always) presided over by the host state, UN treaty negotiations are conducted in much the same way as other UN meetings, with an elected chair nominated by and from the participating states. Despite some common features, however, UN treaty negotiations do not appear to have a fixed formula, either in terms of venue, mandate, participation or decision-making. The draft rules of procedure for treaty negotiations, both in the UN and other forums, are often inspired by the UNGA rules of procedure, but usually with decision-making rules more in line with customary practice for treaty negotiations, as codified in the Vienna Convention on the Law of Treaties (VCLT).


Adopted by whom and how

The mandate for negotiation of a treaty under the auspices of the UN is usually provided by the General Assembly itself, and this would normally be done in the form of a resolution—sometimes adopted by majority voting, and sometimes by acclamation or consensus.[2] A mandate for negotiation of a treaty (in the form of a resolution) does not qualify as an ‘important question’ under the UNGA rules of procedure and therefore requires only a simple majority to be adopted.[3]  If, however, a (simple) majority decides that a negotiation mandate should be categorized as an ‘important question’, the adoption would require a two-thirds majority of the member states present and voting.[4]

A mandate for negotiation of a treaty could also be adopted by an ad-hoc conference of interested states outside of the UN. This was the case for the APMBC and the CCM. A negotiation mandate adopted outside of the UN could in theory be endorsed by the UNGA, although this is not common practice. Sometimes, however, an endorsement by the UNGA takes place after a treaty has been negotiated.

There is no requirement that multilateral treaties are adopted or endorsed by the UNGA. In case of the AMBC and the CCM, both of these treaties were mandated and negotiated outside of the UN, but because the UN Secretary-General was given the task of depositary for both treaties, UNGA resolutions for the financing of this function were necessary.

In short, any group of states can decide to adopt a mandate to negotiate a treaty. A mandate adopted in the UNGA gives weight to a negotiation process, but it is not a requirement in order to achieve treaties with wide memberships. The Geneva Conventions is one example. The APMBC is another.

Content of a mandate

A negotiation mandate can range from being very specific to being relatively vague. The mandate normally deals with both substance and process. For example, in addition to specifying (to various degrees) the aim and content of the relevant treaty, mandates typically set a time frame and sometimes venue(s) for the negotiations. The UNGA resolution mandating the ATT negotiations [5] even went as far as specifying the decision-making rule for the diplomatic conference´s rules of procedure. It stated that the ATT diplomatic conference should be undertaken “on the basis of consensus”.[6] This is unusual; normally a conference adopts its own rules of procedure, and even if key provisions are often “pre-cooked”, the mandate itself does not normally go into this level of detail.


The question of venue is closely linked to that of membership in the negotiations. If the process is under UN auspices, negotiations will normally take place on UN premises. The two likely candidates are UN Headquarters in New York and Geneva, both of which are thought of as “disarmament” negotiation venues; the CD and the CCW takes place in Geneva, as does many meetings in connection to the APMBC and the CCM. The ATT was negotiated in New York, but inter-sessional meetings have to a large extent taken place in Geneva. So will its second Conference of States Parties (CSP2).

Treaties that have been negotiated outside of the UN have normally been negotiated at venues decided on and hosted by the interested states.[7] Some treaty processes under UN auspices have also had venues other than the UN headquarters, for example the UN Convention on the Law of the Seas (UNCLOS), which held one session in Caracas and the final meeting in Jamaica.[8]

Rules of Procedure (RoPs)


The RoPs will normally indicate that the treaty negotiations are open to either “all states” or “all UN member states”, the former alternative potentially indicating that states such as the Holy See, Taiwan or Palestine can join as full participants, as opposed to the second alternative, which limits participation to UN member states.

Treaty negotiations under the auspices of the UN are typically open to all UN member states. Observer States (non-member states) do not have the right to participate in decision-making, but will normally have the right to be present in the conferences, make statements and submit papers. It is often left to a credentials committee to determine which states can be full participants in the conference.

In some cases, negotiations are left to a specific body with a more limited membership. This is the case for all the treaties that have been negotiated in the CD and its predecessors.[9]

Sometimes, a standing committee or ad-hoc group of experts will be given mandates to draft treaty text (for example the ILC[10] or the ad-hoc expert group drafting the Pelindaba Treaty), and then approved by UNGA.

Another model for determining membership in a negotiation process is to establish some kind of threshold requirement. The APMBC and CCM negotiation processes were for example open to all states that committed to a specific goal (banning anti personnel mines and cluster munitions respectively). This would mean that states that were opposed to such prohibitions could not participate in the negotiations. Such states would probably not join the treaties later anyway, and with a more likeminded group of states, the treaty in question would arguably become stronger.

In a UN context, however, all member states will by default be able to participate, which suggests that an explicit criterion in a negotiation mandate might not be very well received by the membership. Aspirational ‘general agreement’ among the negotiating states would thus need to be achieved in a different way and on a basis of self-selection—meaning that, without formulating a criterion for participation, the mandate would have to be formulated in a way that makes ‘interested states’ decide to participate, while at the same time minimizing the potential for derailing the process by negative states. This suggests that both the end goal and the means of achieving that result may have to be spelled out in relative detail in the mandate itself.


The question of participation overlaps with that of membership but extends to identifying which entities (other than observer states) can be observers at the diplomatic conference.

RoPs often make special and explicit provision for non-state entities such as IGOs. Various UN agencies as well as the ICRC and other relevant regional organisations are often mentioned as observers in RoPs.  NGOs are likewise sometimes also mentioned specifically, or there is a rule specifying that the Conference may invite NGOs. Such entities are normally not allowed in “private” meetings (actual negotiations), but are often allowed to take part in “public” meetings, to make statements and to circulate papers. The spectre of rules on admission of NGOs varies—from the negotiations of the Tobacco Convention, where only NGOs sharing the goals of the convention were allowed, to the ATT, where the list of NGOs generated by the ECOSOC ensured very wide participation.[11]


Historically, treaty texts were adopted by ‘unanimity’,[12] understood as an expressed agreement from all the participating states (though abstentions would not normally be seen to block the adoption of the text). The principled explanation for requiring unanimity was that no group of states could impose obligations on other sovereign states. During the past decades, conferences aiming for full/general agreement on a treaty text have tended to pursue ‘consensus’ rather than ‘unanimity’; the difference being that consensus does not require a vote, and states may openly disagree with the draft text without choosing to block its adoption.[13]  The negotiation of the Arms Trade Treaty is a recent example of how the rule on consensus for adopting a treaty has been used in the UN treaty negotiation context.[14]

Article 9 (2) of the 1969 VCLT specifies that when a treaty is negotiated at an international conference, it should be adopted by a two-thirds majority.[15] The rule on a two-thirds majority for treaty adoption applies unless the same majority decides to change this requirement.[16]

There are thus two basic models for making decisions on matters of substance in treaty-negotiations: 1) two-thirds majority voting and 2) consensus/unanimity. Particularly when it comes to the rule on two-thirds majority voting, there are further variations, a number of them including obligations of the participating states to make efforts to achieve consensus before proceeding to a vote. One example is from the negotiation of UNCLOS, where the RoPs stipulated that decisions on matters of substance would be taken by a two-thirds majority, but before such a decision could be put to the vote, a two-thirds majority had to agree that ‘all efforts at reaching general agreement have been exhausted’.[17]

In the area of arms control and disarmament, it would seem that the most frequently used model for decision-making in treaty negotiations over the past decades has been consensus.[18] One reason for this could be that issues relating to weapons have been considered to be key for national security and therefore so important that it should command consent from all states. Another reason may be that consensus is believed to be important in order to achieve universality.[19]

Several disarmament treaties have been negotiated by committees that did not include all interested states. The NPT was negotiated by 17 states in the Eighteen-Nation Disarmament Committee (ENDC), the Biological Weapons Convention was negotiated mainly in the Conference of the Committee on Disarmament (at the time with 26 members), and the Chemical Weapons Convention (CWC) was concluded in the Conference on Disarmament (at the time with 38 members). All these treaties were submitted to the UNGA for endorsement or approval after agreement had been reached in the committee, and the initial mandates for negotiations came from the UN.

Two of the exceptions to the consensus-model with regard to arms treaties came with the above-mentioned APMBC and CCM. Both of these treaties were negotiated with a decision-making rule based on the VCLT rule on two-thirds majority voting, with a provision for making all efforts to reach consensus before going to a vote.[20] It should also be noted that, since the adoption of the CWC, three attempts made at negotiating a weapons-related treaty by consensus (CTBT, CCW (Protocol VI) and ATT)[21] have failed at reaching agreement. The CTBT[22] and the ATT were nevertheless presented to the UNGA and adopted through resolutions.

One key argument for avoiding the consensus rule when negotiating the prohibitions on anti-personnel mines and cluster munitions was that these were not primarily security related treaties, but based on humanitarian concerns with strong elements of international humanitarian law. The very same argument is true for the CCW,[23] but here, the consensus rule has prevailed. It should be noted, however, that the VCLT makes no distinction between matters of national security and other matters when it comes to decision-making.


All treaty negotiations incur costs and therefore require some form of funding in order to be completed. Treaty negotiations under the auspices of the United Nations would normally draw on the resources of the UN secretariat, which in turn requires approval (by consensus) of the Fifth Committee. If a main committee adopts a resolution deemed to involve expenditures, the Secretary-General submits a statement on programme budget implications (PBIs) to the Fifth Committee, via the Advisory Committee on Administrative and Budgetary Questions (ACABQ). The Fifth Committee then informs the General Assembly of the cost implications of the draft resolution.[24]

In addition to assessed contributions by member states over the regular budget (including the Contingency Fund), the General Assembly can of course also encourage member states (and others) to contribute specifically towards the implementation of a given decision. This is often an issue when it comes to sponsorship programmes and other expenses not normally covered by the UN budget.

For negotiation processes that do not take place under the auspices of the United Nations, funding must be found separately. In most cases, the state offering to host a conference would cover at least part of the expenses related to that meeting, but usually a full negotiation process would also require one or more dedicated states to put additional resources on the table in order to ensure both participation and proper implementation of the process. While such a model does not benefit from drawing on the assessed contributions of all UN member states, the meetings themselves need not be more expensive than if they took place in New York or Geneva.[25]

Concluding observations

It seems that in the field of arms and disarmament, States tend to yield to demands for consensus-based decision-making in international treaty negotiations when these demands are put forward by states thought to be of particular importance in order for the treaty regime to be effective (e.g. ATT). It would appear that participation of key states in a given treaty negotiation is considered more important than to ensure the highest possible standards in the resulting instrument: universality before the strength of the treaty.

One the one hand, this inclination to allow ‘important states’ to dictate the terms of the negotiation is understandable. If one wanted to negotiate for example international standards for the treatment of kangaroos and wallabies, it would seem reasonable if Australia were entitled to a veto. Yet on the other hand, this inclination tends to underestimate the normative effect that an international treaty can generate and the political impact it can have over time on non-adhering states—even when special-interest states refrain from becoming parties (CCM, APMBC and UNCLOS are examples).

A UNGA First Committee mandate could in theory establish a negotiation mandate with a threshold requirement for participation. It could also establish that the RoPs for such negotiations should be based on the VCLT, in other words, not based on consensus. It could, moreover, even establish a venue for negotiations and determine which entity would carry out secretariat functions. One must, however, be aware that any budgetary implications of a resolution adopted in the First Committee would have to pass through the Fifth Committee and the ACABQ, which could pose difficulties. It may be noted that if a UNGA resolution adopted on this issue by the First Committee were to be stopped in a forum such as the ACABQ, this could provide an argument for taking the process out of the UN.

[1] UN General Assembly resolution A/C.1/71/L.41, available at

[2] Consensus is a term that seems to have different meanings in different contexts, and understood differently by different actors. It is used here with a meaning synonymous to ’adopted without a vote’.

[3] See UNGA RoP Rule 83.

[4] See UNGA RoP Rule 85.

[5] The Arms Trade Treaty (2009) — A/RES/64/48

[6] A/RES/64/48, para 5.

[7] The CCM, for example, was negotiated in Oslo, Lima, Vienna, Wellington and finally Dublin for the actual diplomatic conference.

[8] See

[9] Examples include the BWC, CWC, CTBT, NPT.

[10] The International Law Commission.

[11] The pro gun lobby were as vocal and visible as the NGOs wishing to restrict arms export.

[12] See Robbe Sabel (2006), Procedure at International Conferences, Cambridge, pp. 312-315.

[13] Typically states make their disagreement known in national statements following adoption of a text.

[14] It may be noted that consensus was not achieved at either of the two diplomatic conferences for the negotiating the ATT.

[15] Article 9 (1), stating that treaties should be adopted by consent from all the states participating, is typically meant for bilateral or other treaties with a limited membership, and where the refusal to accede to or ratify the treaty after the adoption would be politically difficult.

[16] Vienna Convention on the Law of Treaties, Article 9.

[17] Rule 37. (Normally, the Chair or president of a Conference decides when consensus has been achieved.)

[18] Examples include CTBT, CWC, BWC, NPT, ATT, and CCW.

[19] This argument does not hold water if one looks at the CCW and the ATT for example.

[20] It may be noted that in neither of the diplomatic conferences on negotiating these two prohibitions were voting actually conducted, but the ”common ground” to be reflected in the treaty text was based on careful ”headcounts” during negotiations. It was thus not possible for a small number of states to get their wishes imposed – any objection to a provision had to command more than a third of the negotiating states.

[21] CCW Amended Protocol II on the use of mines, booby-traps and other devices was adopted by consensus in 1996, but was seen as weak and triggered the negotiations of a prohibition on anti-personnel mines that was adopted a year later.

[22] Although the CTBT is not yet in force.

[23] The title of the Convention being Convention on prohibitions or restrictions on the use of certain conventional weapons which may be deemed to be excessively injurious or to have indiscriminate effects.

[24] The PBI statement (A/C.1/67/L.60) relating to draft UNGA resolution A/C.1/67/L.11 on the Arms Trade Treaty stipulated that if the UNGA were to adopt the resolution (which contained the decision to convene the final negotiation conference from 18-28 March 2013) ’additional resources in the total amount of $563,800 would arise’.

[25] As an example, it has been claimed that the whole negotiation process of the CCM came to about the same price as a two-day NATO summit.