Safeguards Provisions in the Treaty on the Prohibition of Nuclear WeaponsPosted: April 11, 2018
I’m very pleased to host a guest post by Dr. Eirini Giorgou. I think she does a really excellent job in the piece of explaining the safeguards provisions in the TPNW, and the debates that led to their adoption, and then responding to a number of criticisms that have been leveled against how these provisions were crafted. I think this is really exemplary legal work. I also don’t find anything in it with which I disagree. So I highly recommend it to readers.
Dr. Giorgou is a legal adviser on weapons law in the International Committee of the Red Cross (ICRC), with experience in multilateral disarmament diplomacy. She was previously Disarmament adviser in the Department of Foreign Affairs and Trade of Ireland. In that capacity she participated in the negotiations on the Treaty on the Prohibition of Nuclear Weapons, including by facilitating negotiations on the Treaty’s Safeguards provisions. Dr Giorgou is a licensed lawyer and holds a PhD in International Law from the University of Geneva, Switzerland.
The views expressed herein are of the author and do not necessarily reflect the position and views of the ICRC or of the Department of Foreign Affairs and Trade of Ireland
Safeguards Provisions in the Treaty on the Prohibition of Nuclear Weapons
By: Eirini Giorgou
Since its adoption by 122 states on 7 July 2017, the Treaty on the Prohibition of Nuclear Weapons (TPNW) has frequently been made the target of criticism as regards its Safeguards provisions. The following provides an analysis of these clauses as contained in the Treaty’s Articles 3 and 4, outlines the main arguments that have been leveled against the Treaty in this regard, and responds to these.
What does (and what doesn’t) the TPNW have to say on Safeguards – and why.
To Safeguard or not?
It did not take long for negotiators to agree that Safeguards provisions were necessary and could not be omitted without sending a dangerous political message: that the new Treaty undermined the existing Safeguards regime and that, by creating a loophole, it encouraged states to withdraw from the NPT and from their existing Safeguards obligations. Once the decision to include such provisions was made, negotiating parties were faced with two major difficulties: agreeing on how strict a Safeguards standard they should adopt; and ensuring that this standard would neither undermine the existing IAEA Safeguards regime, nor create an unwanted legal status quo by subordinating the new Treaty to other instruments or international organizations.
Throughout the negotiations, the issue of Safeguards was closely linked to that of the ratification of or accession to the Treaty of states having possessed (at any time after the date set for its adoption, namely 7 July 2017) nuclear weapons. Very soon there was consensus around the fact that the Safeguards standard to be imposed by the TPNW, whatever that was, should be different – and higher – for this category of states, compared to those that had never possessed nuclear weapons, or had eliminated them prior to the cut-off date.
It was therefore agreed that a distinction be drawn clearly between those states who did not possess nuclear weapons after 7 July 2017 and those who did, and that the respective Safeguards rules and obligations be contained in different articles. Hence Article 3 only applies to the first category of states, while Safeguards provisions applicable to the second category are contained in Article 4.
What changes for non-nuclear possessor states?
Article 3 was the result of long and strenuous negotiations, and a compromise between three negotiating factions: those supporting an explicit reference to the IAEA model Additional Protocol (AP) as the mandatory Safeguards standard required of TPNW States Parties; those (few) objecting to such a provision and willing to accept only a reference to the IAEA model Comprehensive Safeguards Agreement (CSA) – an instrument less intrusive in terms of inspections than the AP – as the mandatory standard for Safeguards; and a minority according to which the TPNW need not contain Safeguards provisions at all, but that it should be a short, simple, straightforward framework-setting instrument focused on the core prohibitions, to be complemented by existing and future instruments (including with regard to Safeguards).
Paragraph 1 provides that ‘‘Each State Party to which Article 4, paragraph 1 or 2, does not apply shall, at a minimum, maintain its International Atomic Energy Agency safeguards obligations in force at the time of entry into force of this Treaty, without prejudice to any additional relevant instruments that it may adopt in the future.’’
This provision stipulates that states bound at the time of entry into force of the Treaty by Safeguards obligations undertaken in the context of the IAEA (be they a CSA only, or an AP) are obliged to maintain them. This is a direct legal obligation imposed upon TPNW States Parties. The objective is to ensure that states would not take the absence of an explicit mandatory Safeguards standard in the TPNW as an invitation to opt out from their existing Safeguards obligations.
At the same time, states are not obliged to upgrade to the stricter standard (AP), although they are implicitly encouraged to do so (‘‘at a minimum’’, ‘‘without prejudice to any additional relevant instruments that it may adopt in the future’’). Proposals were made for a reference to ‘‘stricter standards’’ and for language ‘‘encouraging’’ states to upgrade. The fact that both met with strong opposition is indicative of the level of sensitivity among states of this issue. Nevertheless, there was tacit agreement that by ‘‘additional relevant instruments’’ the provision refers to the AP, as well as to potential stricter standards to be developed in the future.
Paragraph 2 provides that ‘’Each State Party to which Article 4, paragraph 1 or 2, does not apply that has not yet done so shall conclude with the International Atomic Energy Agency and bring into force a comprehensive safeguards agreement […] Each State Party shall thereafter maintain such obligations, without prejudice to any additional relevant instruments that it may adopt in the future.’’
This provision stipulates that a minimum mandatory Safeguards standard is required for States Parties to the TPNW, and that this standard is the CSA. While the existence of a CSA is not a prerequisite for adherence to the Treaty, the State Party that is not bound by a CSA at the time of entry into force of the Treaty is under a direct obligation to negotiate and bring such an agreement into force, in accordance with set deadlines. Thereafter the obligation to maintain the CSA in the future applies, coupled with the implicit encouragement to adopt higher standards (e.g. the AP).
…and what changes for nuclear possessors?
Article 4 purports to cut a path into uncharted territory by providing avenues for accession by states having possessed nuclear weapons after 7 July 2017. It further stipulates the Safeguards obligations applicable to such states once they become party to the TPNW.
Paragraph 1 provides that ‘‘Each State Party that after 7 July 2017 owned, possessed or controlled nuclear weapons or other nuclear explosive devices and eliminated its nuclear-weapon programme, including the elimination or irreversible conversion of all nuclear weapons- related facilities, prior to the entry into force of this Treaty for it […] shall conclude a safeguards agreement with the International Atomic Energy Agency sufficient to provide credible assurance of the non-diversion of declared nuclear material from peaceful nuclear activities and of the absence of undeclared nuclear material or activities in that State Party as a whole. […] That State Party shall thereafter, at a minimum, maintain these safeguards obligations, without prejudice to any additional relevant instruments that it may adopt in the future.’’
This provision applies to nuclear possessor states that have taken the ‘‘destroy, then join’’ approach. For these states the Safeguards standard required is higher than that of Article 3(2). However, providing for an abstract, ‘‘catch-all’’ standard would have been neither appropriate nor effective, given the vast differences in the arsenals of nuclear possessor states. This is why the standard is described rather than defined (‘‘sufficient to provide credible assurance’’). The language used in part reflects the general purpose of the model CSA, as stated in INFCIRC/153 (Corrected), i.e. to verify that nuclear material is not diverted from peaceful uses to nuclear weapons, and in part corresponds to language used in the model AP (INFCIRC/540), i.e. ‘‘credible assurance of the absence of undeclared nuclear material and activities’’. The existence of such a Safeguards agreement is not a prerequisite for ratification or accession, but the State Party is under a direct obligation to conclude and bring into force such an agreement within set deadlines, and to maintain this standard in the future (again, states are implicitly encouraged to adopt stricter standards).
Paragraphs 2 and 3 provide that ‘‘each State Party that owns, possesses or controls nuclear weapons or other nuclear explosive devices shall […] conclude a safeguards agreement with the International Atomic Energy Agency sufficient to provide credible assurance of the non-diversion of declared nuclear material from peaceful nuclear activities and of the absence of undeclared nuclear material or activities in the State as a whole. […] That State Party shall thereafter, at a minimum, maintain these safeguards obligations, without prejudice to any additional relevant instruments that it may adopt in the future.’’
This provision applies to nuclear possessor states that have taken the ‘‘join, then destroy’’ approach. The provision is identical to that of paragraph 1, and as such implies an agreement equivalent in scope and purpose to the CSA plus AP formula, although the actual Safeguards required may be different, because what is ‘‘sufficient to provide credible assurance’’ will be determined on an ad hoc basis.
Constructive (and de-constructive) criticism
The TPNW’s Safeguards provisions are undoubtedly imperfect. Limited negotiation time, lack of expertise and at times a strong divergence of views resulted in what has been called a conservative, weak, vague and even ambiguous text. The main arguments of the TPNW critics will be addressed below in an attempt to evaluate to what extent they hold true and, where possible, to identify the reasons that led to apparent or actual shortcomings in the Treaty text.
‘‘The TPNW weakens the existing Safeguards regime.’’
The TPNW has been accused of weakening the existing Safeguards regime by (a) not imposing as a prerequisite for ratification or accession the existence of a Safeguards agreement and (b) by not using the AP as the minimum (‘golden’) standard in this respect.
Under the NPT the conclusion of a CSA is mandatory for non-nuclear-weapon States Parties (as per Article III), whereas the conclusion of an AP is optional. Valid arguments were thus made by states that establishing the AP as the minimum acceptable standard in the TPNW would mean changing the former’s nature from optional to mandatory, something that would exceed the mandate of the Negotiating Conference.
In reality, the TPNW takes the same approach as the NPT, imposing upon (non-nuclear-weapon) States Parties the direct obligation to negotiate and agree a Safeguards agreement. Moreover, the Treaty goes beyond the NPT, by obliging States Parties to maintain, as a minimum, their existing Safeguards standards, thus making the AP mandatory for states that are bound by it when the TPNW enters into force.
The TPNW goes beyond the NPT also with regard to nuclear possessor States Parties, by imposing on them a direct obligation to negotiate, conclude and maintain an adequate Safeguards agreement. Currently nuclear-weapon States are under no such obligation under the NPT, although they do have voluntary arrangements in place regarding safeguards for some of their nuclear material.
‘‘The TPNW’s Safeguards provisions are vague and lack specificity.’’
As regards states not possessing nuclear weapons, the Safeguards standard is clear, insofar as it refers to INFCIRC/153 (Corrected), i.e. the model CSA. This argument therefore refers mainly to the Safeguards clauses contained in Article 4, which are applicable to nuclear possessor States Parties.
In this respect the text is indeed vague – and legitimately so. Indeed, it would have been neither possible nor appropriate to set one single standard a priori and applicable to all nuclear possessor states, irrespective of differences among nuclear arsenals and of possible future developments in such arsenals prior to the entry into force of the TPNW for the state(s) in question. Moreover, creating a new Safeguards standard would have clearly exceeded the mandate of the Negotiating Conference, in addition to being perceived as infringing upon the authority of the IAEA.
‘‘The Treaty sets two different Safeguards standards. This discrimination is counterproductive…’’
…the argument goes, as it is not conducive to confidence-building and thus to complete, verifiable and irreversible disarmament. The loophole identified regards states not possessing nuclear weapons after 7 July 2017 that do not have an AP in force at the time of entry into force of the Treaty – these are not obliged by the TPNW to adopt one, whereas states having possessed nuclear weapons after the cut-off date are required to adopt Safeguards equivalent in scope and purpose to those of a CSA plus AP.
It is also argued that this differentiated treatment contradicts the NPT, in particular Action 30 (agreed in the Final Document of the 2000 Review Conference and reaffirmed at the 2010 Review Conference), which provides that when nuclear weapons have been eliminated the highest standard of Safeguards applied by the IAEA – currently CSA plus AP – should be applied universally to all states.
The first argument is a political and not a legal one. The political counterargument would be that what is currently stalling progress on nuclear disarmament is not the reluctance of a small number of states to adopt APs (for which they have no obligation), but the failure of nuclear-weapon states to implement their commitments under the NPT. It is also logical to argue that differentiated Safeguards standards are warranted considering how much easier it would be for a state having possessed nuclear weapons to conceal or re-acquire nuclear weapon-grade material and relevant technology, or to divert material to non-peaceful uses and/or to convert nuclear facilities, compared to a state that was not previously in possession of such weapons.
The second argument merits closer consideration. It uses an interpretation of Action 30 which implies that nuclear-weapon states should be the last ones to adopt an AP, following the elimination of nuclear weapons, and that non-nuclear weapon states have thus undertaken to adopt APs prior to the moment when ‘‘global zero’’ is reached. However, this interpretation ignores Action 28, which ‘‘encourages all States parties which have not yet done so to conclude and to bring into force additional protocols as soon as possible and to implement them provisionally pending their entry into force’’ – contrary to Actions 24 and 25, which call on and urge, respectively, states to adopt a CSA. The difference in language shows clearly that the adoption of an AP is, until the elimination of nuclear weapons is achieved, optional. The TPNW does not therefore contradict the NPT in this respect either.
Even if the above interpretation of Action 30 were accepted, Article 18 of the TPNW does not preclude states from complying with obligations stemming from other instruments, insofar as these are consistent with the TPNW. It is hard to argue that adopting an AP would not be consistent with the Treaty, simply because the minimum compulsory standard set by the TPNW is lower.
‘‘A state that joins the treaty while still possessing nuclear weapons is not required to accept any Safeguards until after it has eliminated its weapons, as the deadline for the entry into force of the Safeguards agreement is 18 months after the completion of implementation of the elimination plan.’’
There is thus a Safeguards gap while the elimination programme has not been agreed or is still being implemented. Since the elimination of a state’s nuclear weapons programme could take years, it is argued that, at the very least, the Treaty should have required such states to apply a CSA plus AP to civil and non-sensitive nuclear materials and facilities immediately upon becoming a party to the TPNW, and, ideally, to apply appropriate monitoring and verification arrangements to their sensitive materials and facilities during the disarmament process until the required CSA and AP (or equivalent standards) enter into force.
There is indeed a temporal gap between the entry into force of the TPNW for a nuclear possessor state that has taken the ‘‘join, then destroy’’ approach, and the elimination of its nuclear arsenal, insofar as Safeguards are concerned. However, this temporal gap is not unique to this category of states, but is equally the case for non-nuclear weapon states (under Article 3, and, indeed, under NPT Article III) and for nuclear possessor states having adopted the ‘‘destroy, then join’’ approach – although the deadlines imposed on the latter are different.
The negotiators recognised that the implementation of CSA plus AP while a nuclear weapons programme was still active would be impracticable, as demonstrated by the experience of South Africa. Further, establishing a new Safeguards standard appropriate for a (still) nuclear possessor state would have exceeded the mandate of the Negotiating Conference, as the issue could not have been resolved with a simple reference to existing model instruments forming part of the IAEA framework in force. The need for monitoring and ongoing verification of the dismantlement was acknowledged, but could not be addressed without either vesting the future competent international authority with this mandate (and thus bypassing the IAEA – a politically highly unsound and practically ineffective move), or providing explicitly that the IAEA would conduct monitoring and ongoing verification. The latter did not prove possible due to a deep divergence of views on the interpretation of the IAEA mandate, and would in any case have been avoided in light of the difficulties such a mandate was expected to encounter at the IAEA Board of Governors, whose membership is in large part unequivocally ‘‘hostile’’ to the TPNW.
However, for states taking the ‘‘join, then destroy’’ approach, ‘‘interim’’ Safeguards applicable during the elimination process and period can be defined and agreed in the context of the elimination plan which will have to be negotiated with and verified by the competent international authority (Article 4(2)). Thus, while the TPNW does not establish an obligation for ‘‘interim’’ Safeguards per se, such an obligation might very well be included as part of the mutually agreed elimination plan.
‘‘Article 3(2) rigidly applies a specific Safeguards agreement that will likely become outdated and thereby prejudges what new safeguards standards the IAEA may decide upon.’’
The TPNW sets the CSA as the minimum standard, explicitly stating that this is without prejudice to additional relevant instruments a state may adopt in the future. This caveat was added not only to implicitly encourage states to upgrade their Safeguards standards by adopting an AP, but also to accommodate any new, higher standards that might be elaborated in the future in the context of the IAEA and beyond. This argument is thus clearly contradicted by the text of the Treaty.
‘‘The timing requirement in Article 3(2) conflicts with the timing requirement in Article III (4) of the NPT and could be interpreted as extending the period allowed under the NPT.’’
Article III(4) of the NPT stipulates that ‘‘Non-nuclear-weapon States Party to the Treaty shall conclude agreements with the International Atomic Energy Agency […] Negotiation of such agreements shall commence within 180 days from the original entry into force of this Treaty. For States depositing their instruments of ratification or accession after the 180-day period, negotiation of such agreements shall commence not later than the date of such deposit. Such agreements shall enter into force not later than eighteen months after the date of initiation of negotiations.’’
It is true that the starting point for calculating the deadline as regards the obligation to negotiate, conclude and bring into force a Safeguards agreement is different in the two instruments. The NPT provides for a more complicated system, where starting points differ between the deadline for commencing negotiations and for bringing into force the Safeguards agreement, and depending on when a state deposited its instrument of ratification or accession. On the other hand, TPNW negotiators opted for a much simpler (and unified) system, where the starting point for all relevant deadlines is the entry into force of the Treaty for the state in question. What does this practically mean?
For states that deposited their instrument of ratification prior to the entry into force of the treaty, the deadline for commencing negotiations on a Safeguards agreement is 180 days under the NPT and 180 days under the TPNW (as for this category of states ‘‘original’’ entry into force of the treaty and entry into force for those states will coincide). The deadline for entry into force of the Safeguards agreement in 18 months from initiation of the negotiations under the NPT and 18 months from the entry into force of the treaty for those states (which, as explained, equals the ‘‘original’’ entry into force of the treaty) under the TPNW. It is clear that, between the two instruments, the one providing for a more extended deadline in this respect is the NPT, and not the TPNW.
For states that acceded to the treaty after its ‘‘original’’ entry into force, the deadline for commencement of negotiations on a Safeguards agreement under the NPT is 180 days from the original entry into force of the treaty, or at the very latest the date of accession, if this occurs beyond the 180-day period. This essentially requires prospective States Parties to take preparatory measures (i.e. commence negotiations) before joining the treaty. Under the TPNW, the deadline for commencing negotiations is 180 days from the entry into force of the treaty for the state in question, i.e. 270 days after the deposit of the instrument of ratification or accession. The deadline for commencing negotiations in this case is therefore considerably longer under the TPNW than under the NPT. As regards entry into force of the Safeguards agreement, the deadline under the NPT is 18 months after the initiation of negotiations, i.e. at the latest 18 months after the date of deposit of the instrument of ratification or accession. Under the TPNW, this date is 18 months after the entry into force of the treaty for the state in question, i.e. 21 months after the deposit of the instrument of ratification or accession (slightly longer than under the NPT).
Thus, the main difference under the two instruments is that the TPNW does not impose an indirect obligation on prospective States Parties to take preparatory measures (i.e. to commence negotiations on a Safeguards agreement) before joining the Treaty. And yes, the ‘‘grace period’’ under the TPNW is indeed in most cases longer than the one under the NPT.
In terms of ‘‘conflicting timing requirements’’, it does not seem reasonable to invoke a conflict here, as the deadlines refer to obligations stemming from different instruments whose violation is a violation of the respective instrument. In other words, an NPT State Party not having concluded a CSA within the deadline imposed by that instrument will not cease to violate the NPT after it becomes a State Party to the TPNW, simply because the deadline for the fulfillment of the obligation imposed by the latter is longer.
‘‘A [last] loophole refers to the Safeguards standard for states that ran a nuclear weapons programme or even possessed nuclear weapons prior to the adoption of the TPNW, but (claim to have) eliminated these prior to the entry into force of the Treaty for them.’’
It is argued that the Safeguards standard for such states should be higher than a simple CSA (which is now the case according to Article 3), in light of their experience, knowledge and potential access to technology. Alternatively, it is argued that the cut-off date (date of adoption of the TPNW) is flawed and should not have been included; a higher standard of Safeguards should thus apply to any state that ever possessed nuclear weapons, irrespective of whether the elimination was completed before the adoption of the TPNW or not.
Both options were indeed examined during the negotiations, but were strongly objected to by one state and were ultimately dismissed. In reality, the known cases of states that would fall within this category would appear to be sufficiently covered by existing IAEA instruments and arrangements, and do not, in principle, raise concerns that would justify the need to include them in the same category as those possessing nuclear weapons at the time of adoption of the TPNW.
Obviously, concerns remain with regard to suspected cases of states running clandestine nuclear weapons programmes. However, this is an issue that exceeds the scope of the TPNW and, insofar as it touches upon broader questions of enforcement and verification, could not have been addressed in the context of that instrument.
Let’s end with some universally acknowledged truths.
One – the TPNW is not, and was never meant to be, a Comprehensive Nuclear Weapons Convention. What principally distinguishes the former from the latter model is the existence of provisions, including obligations, related to the verification of the complete and irreversible elimination of nuclear weapons. These are indispensable for achieving and maintaining a world free of nuclear weapons – however, whether one likes it or not, this was not the role envisaged for the TPNW by its drafters.
Two – equally indispensable for a nuclear-free status quo are Safeguards capable of ensuring that any material and activity resulting in or involving nuclear energy is put to exclusively peaceful uses. Safeguards should be applied universally. The NPT sets the minimum standard for Safeguards applicable today, but states not party to the NPT are practically under no Safeguards obligations.
And three – when it comes to nuclear disarmament, there is no magic bullet. The road to global zero requires consistency, determination, good faith, and above all political will. The TPNW is not the end of the road, but a significant (and unprecedented) step in the right direction. As such, further steps, measures, building blocks or whatever else we may wish to call them will be needed to reach the destination.
The TPNW is now a reality. Despite its shortcomings, it shook the stagnating waters of nuclear disarmament. Instead of engaging in futile debates in favour or against it, states should join forces to maintain, and strengthen, this momentum. The fact remains, a solid majority of non-nuclear weapon states have made one additional effort towards the implementation of the NPT’s Article VI. It’s about time the nuclear-weapon states did the same – they have dozens of unimplemented undertakings and commitments to choose from.
The problem is, we may be running out of time.
 Broadly speaking, a CSA facilitates verification relating to the non-diversion of declared nuclear material and activities for nuclear weapons activities, and an AP facilitates verification concerning the absence of undeclared nuclear material and activities.
 The issue of Safeguards should not be confused with that of verification of the complete and irreversible elimination or conversion of a nuclear weapons programme. The latter will, for states to which Article 4 applies, be carried out by the future international competent authority. The absence in the TPNW of specific verification provisions is fully compatible with the nature of the instrument and with the mandate of the Negotiating Conference.