"Though we have acheived progress, our work is not over. That is why I support the mission of the Arms Control Association. It is, quite simply, the most effective and important organization working in the field today." 

– Larry Weiler
Former U.S.-Russian arms control negotiator
August 7, 2018
Staying Credible: How Precedents Can Help the IAEA Get Noncompliance Calls Right
Share this

Latest ACA Resources

Peter Jenkins

In recent years, the International Atomic Energy Agency (IAEA) has come under fire for its handling of safeguards noncompliance cases. A prominent example of such criticism is the one from the International Commission on Nuclear Non-proliferation and Disarmament (ICNND) which produced a comprehensive report and recommendations on every aspect of the nuclear Nonproliferation Treaty (NPT) in December 2009.[1]

In its report, the commission suggests that a delay “of nearly three years” in the case of Iran put at risk the credibility of IAEA noncompliance determinations and that the IAEA should confine itself to technical judgments, leaving all political judgments to the UN Security Council.[2]

This article seeks to show that this criticism is unwarranted and that the IAEA can be trusted to fulfill its statutory obligations to the Security Council. In particular, it argues that the IAEA statute requires the Board of Governors to make judgments that extend beyond the purely technical and that the board’s record to date in this respect has been good. It points out that the seven noncompliance cases that have come before the board constitute a body of precedents on which the board can draw in the future and suggests that using past cases as a guide can be considered akin to a case-law approach.

The issue may seem abstruse, but in reality, it goes to the heart of the nuclear nonproliferation regime. The violation of an IAEA safeguards agreement may be the first overt indication that a state is intent on acquiring nuclear weapons in contravention of the NPT. In the last decade, five safeguards agreement violations or multiple violations have come before the board for it to determine whether they constitute the noncompliance to which Article XII.C of the IAEA statute refers.[3] It is quite possible that the board will face similar challenges in the future.

The statute requires the secretariat and the board to cooperate in detecting and reporting to the Security Council instances of noncompliance with NPT safeguards arrangements. The secretariat is the technical arm of the agency, responsible for detecting any violation of the provisions of safeguards agreements.

Following the detection of a violation, the Department of Safeguards has to distinguish between those violations that should be reported to the board in a specific report as noncompliance, in accordance with the provisions of Article XII.C, and the far larger number of violations that need only be reported, if at all, in the annual Safeguards Implementation Report.

In the absence of any statutory guidance, the safeguards department has had to develop criteria for drawing this distinction. To judge from past practice, one or more of the following factors have been decisive:

•  whether the violation has made it impracticable to implement the safeguards agreement (e.g., if the state concerned has refused to cooperate with the agency);

•  whether the nature of the violation has suggested that diversion has occurred;[4]

•  whether the nature of the violation has suggested that diversion may have occurred, including to “purposes unknown”; and

•  whether there has been any indication of possible weapons intent.

The safeguards department is widely held to have performed its task of detecting safeguards agreement violations and distinguishing among them in terms of gravity with competence and impartiality. The ICNND in effect acknowledges this by concentrating its criticism and its recommendation on the board.

Noncompliance Reports

Article XII.C requires the director-general to transmit to the board all specific noncompliance reports made by the safeguards department. It is for the board, which is a political body, to determine whether the failure to comply reported by the secretariat does indeed constitute noncompliance under Article XII.C and, if so, when this needs to be reported to the Security Council. The wording of Article XII.C implies that there may be grounds to distinguish some of the reports transmitted by the director-general from others; the board will not necessarily find that every report merits being reported to the Security Council and the UN General Assembly.

The statute does not define what constitutes noncompliance or spell out the criteria that the board should use to arrive at a finding. Some believe that the sole criterion should be whether the secretariat has been “unable to verify that there has been no diversion of nuclear material required to be safeguarded”—the language of paragraph 19 of the standard NPT safeguards agreement.[5]

In practice, board members, noting that a clear definition of “unable to verify” is absent[6] and that the Model Additional Protocol has clarified that the agency’s mandate includes looking for indicators of any undeclared nuclear activities, have adopted a more policy-oriented approach. Whether the Security Council needs to be informed of a case has been the prime consideration. This has required a judgment as to whether a failure to comply has implications for international peace and security.[7] This is logical, because Article III.B.4 contains a passage that suggests that its framers wished the Security Council to be kept informed of matters pertaining to the maintenance of peace and security: “[I]f in connection with the activities of the Agency there should arise questions that are within the competence of the Security Council, the Agency shall notify the Security Council as the organ bearing the main responsibility for the maintenance of international peace and security, and may also take the measures open to it under this Statute, including those provided in paragraph C of article XII.”

In making a formal noncompliance finding in the Iranian case in September 2005, the board implicitly drew attention to this link between Article XII.C, which requires the board to find whether noncompliance has occurred, and Article III.B.4, which requires the board to notify the Security Council of questions that are within its competence. In Resolution GOV/2005/77, the board found not only that “Iran’s many failures and breaches of its obligation to comply with its NPT safeguards agreement” constituted “non-compliance in the context of article XII.C of the Agency Statute,” but also that a number of listed factors had “given rise to questions that are within the competence of the Security Council as the organ bearing the main responsibility for the maintenance of international peace and security.”

In order to determine whether specific cases have implications for peace and security, board members have looked at the specific characteristics and circumstances of each case. No two cases have been the same. Yet, the five cases in which the board has found noncompliance needing to be reported to the Security Council—Iraq (1991), Romania (1992), North Korea (1993, 1994, and 2003), Libya (2004) and Iran (2005)—have had in common the following characteristics:

•  undeclared activities detected and reported by the inspectorate have involved or been related to the production of fissile material;

•  the inspectors have reported evidence that nuclear material required to be safeguarded has not been placed under safeguards or, having been under safeguards, has been diverted;

•  the inspectors have reported evidence of the state concerned pursuing a policy of concealment prior to the breach or breaches coming to their notice;

•  the inspectors have reported evidence that the nondeclaration or diversion of material required to be safeguarded has been deliberate; and

•  the inspectors have reported grounds to suspect or evidence of an underlying intent to gain the capacity to produce nuclear weapons.

In three of the cases—Iraq, North Korea, and Iran—the inspectors also have reported denial of access, obstruction, or a significant absence of proactive cooperation in the period following the detection of a breach or breaches.

A key point that emerges from the five cases is that board members have not required proof; they have been content with evidence, or even just grounds for suspicion. As John Carlson has pointed out,[8] board members have recognized that requiring the sort of proof necessary for conviction in U.S. criminal trials, i.e., beyond a reasonable doubt, would set an unrealistic threshold for the inspectorate. Such a requirement would risk delaying a noncompliance finding and subsequent report to the Security Council to a point beyond that at which the council may be able to determine and avert a threat to peace and security. Rather, a balance of probabilities has been seen as sufficient to justify providing the Security Council with the opportunity to consider the matter.

The five cases make clear that board members have not required evidence of intent to manufacture or acquire nuclear weapons. Requiring evidence of the intent to gain the capacity to produce nuclear weapons or even the grounds to suspect it falls well short of that.

It may be instructive to note the characteristics of two cases reported by the secretariat that were not found by the board to constitute noncompliance.[9] The first occurred in 2004, when the inspectors reported that, over an extended period of time, there had been undeclared South Korean experiments involving the production of small quantities of fissile material using undeclared nuclear material. The inspectors also reported that certain South Korean authorities had initially taken some actions that could be interpreted as attempts to conceal some of this noncompliance.

Board members took note of these failures with concern, but accepted the assurance of the South Korean government that it had been unaware of these activities at the time. The board also gave weight to the absence of evidence to support suspicion of a governmental intent to gain the capacity to produce nuclear weapons and to evidence that the undeclared activities had been terminated. Calculating, in essence, that the noncompliance reported by the secretariat had no implications for peace and security, board members refrained from making a formal noncompliance finding.

Whether this judgment was correct has been questioned then and since.[10] It would have been systemically helpful if the board had made a formal noncompliance finding and had reported the case to the Security Council. Such a course of action would have set the bar for findings and reports low enough to include cases in which there is reasonable room for doubt as to whether the case presents implications for peace and security. Nevertheless, the course the board chose did not do any systemic damage; subsequent developments support the view that the case had no implications for peace and security.

In the second case, in 2005, the inspectors reported comparable failures to declare experiments resulting in the production of small quantities of fissile material using undeclared material over an extended period, this time in Egypt. As in the South Korean case, there was no evidence of an underlying intent to gain the capacity to produce nuclear weapons, and the quantities of material involved were small. Also, there was no evidence that the failures to report had been deliberate or that the relevant authorities had attempted to conceal these failures once they had come to their notice; and neither denial of access nor obstruction had occurred. The board again refrained from making a formal noncompliance finding—less questionably than in the South Korean case, although it will be important to note whether recently renewed investigations by the safeguards department into aspects of Egypt’s nuclear program suggest that the board should revisit the 2005 conclusion.

The statute is silent on the timing of the reporting of noncompliance to the Security Council. In practice, in every case but one, the report has followed within days of the finding. In the Iranian case, the board decided to address the timing of the report at a subsequent, unspecified session. It did so for political reasons; by making use of the flexibility afforded by the absence of a specific statutory provision, it sought to put pressure on Iran to resume suspension of uranium enrichment.

Defining Noncompliance

In addition to the ICNND recommendation that the IAEA should confine itself to technical judgments, there have been calls for a common understanding on the definition of noncompliance.[11] A precise definition of noncompliance is unnecessary and would be risky. It also is improbable that IAEA members would agree to it.

There would be a strong argument for some kind of definitional understanding if the board had failed to make a noncompliance finding in relation to a case that could reasonably be held to have had implications for peace and security. Leaving aside the special case of Iran, there is no evidence that this has happened. Even those who believe that a noncompliance finding could have been useful in the South Korean case would find it difficult to demonstrate that the case presented implications for peace and security.

The board’s failure to find Iran noncompliant in November 2003 was the result of a conscious political decision, with which almost all board members were content, to endorse the outcome of ministerial talks in Tehran on October 21, 2003. During those talks, the Iranian government offered to suspend all uranium-enrichment and reprocessing activities and to apply the Model Additional Protocol provisionally, in return for France, Germany, and the United Kingdom committing to the view that “the immediate situation” should be “resolved by the IAEA Board.”[12] Had this not been the case, the board undoubtedly would have found Iran noncompliant, despite the unpopularity at that time of the prime advocate of a noncompliance finding, the U.S. government.

Within weeks of Iran’s August 2005 breach of its 2003 commitment to suspend enrichment activity,[13] the board found Iran to have been noncompliant. The vote on September 24, 2005, was 22-1, with 12 abstentions. It is likely, although this cannot be proved, that the number of votes in favor would have been larger had not Iran, since 2003, taken steps to remedy its noncompliance, as required by Article XII.C, and perhaps had not members of the Nonaligned Movement (NAM) been susceptible to the influence of IAEA Director-General Mohamed ElBaradei, who argued privately against a noncompliance finding.

A precise definition would be risky because it might give rise to a majority of board members declining to report to the Security Council violations having implications for peace and security, arguing that they fell outside the definition. Excluding this risk by anticipating all possible circumstances under which noncompliance might occur at the stage of drafting a definition would be an almost impossible task for a single expert, let alone for a committee comprising the 35 members of the board or the full membership of the agency.

A precise definition also would deprive governments of most of the discretion that they currently enjoy as board members. They can use that discretion profitably to allow space for diplomacy or to give the benefit of the doubt to a state that has breached its safeguards undertakings. Board members and Security Council members receive instructions from the same source: governments. Being able to exercise political judgment in the IAEA board as well as in the Security Council is a plus for governments. Flexibility enhances the global community’s capacity to manage the nuclear nonproliferation regime; it does not impair that capacity.

Irrespective of whether a precise definition of noncompliance is necessary or desirable, it is unlikely that the board would be able to reach the necessary consensus on it. During the 2003-2005 period, all the leading members of the board indicated in some way that they were attached to the flexibility currently enjoyed by the board. It enabled France, Germany, Russia, the United Kingdom, and others to agree to defer finding Iran noncompliant as long as Iran suspended work on uranium enrichment and cooperated with IAEA inspectors in return. It left the United States and many others free to give South Korea the benefit of the doubt in relation to its safeguards agreement violations.

Moreover, NAM members would probably see any attempt to adopt a precise definition of noncompliance as intended to make referral to the Security Council more likely and as strengthening the nonproliferation pillar of the NPT without any compensating reinforcement of the nuclear disarmament or peaceful-use pillars. For those reasons, they would be likely to resist such an effort.

A case could be made for delegating the board’s responsibility for noncompliance findings to the director-general, on the grounds that he could be expected to be more neutral than the board. In practice, delegation to the director-general could turn out to be counterproductive. In the weeks leading up to the September 2005 vote on Iran, ElBaradei made no secret of his self-assigned mission to save the member states, his employers, from taking a step to which he felt opposed. Had the board’s responsibility rested with him, it is most unlikely that Iran would have been found noncompliant, despite the obvious implications for peace and security of what the inspectorate had reported since May 2003. Moreover, delegating the responsibility would be unacceptable to all who are attached to the political discretion that the board currently enjoys.


It seems possible that much of the criticism of the IAEA’s handling of noncompliance, leading to calls for the IAEA to confine itself to technical judgments or for a common understanding on the definition of noncompliance, originates in circles that would have wished Iran to be reported to the Security Council in the fall of 2003. In reality, it is wrong to infer from those events that the IAEA is incapable of dealing adequately with noncompliance. The board’s handling of the Iranian case that fall was atypical. Four of the five states that are permanent members of the Security Council and almost all other members of the board at that time believed that recourse to the political discretion afforded by the statute was justified, to leave the way open for a suspension of all uranium-enrichment work in Iran and to secure for the IAEA inspectors the access provided by provisional application of the additional protocol, not least because the Security Council had been divided earlier in the year over Iraq and had failed to respond adequately to North Korean noncompliance. In other noncompliance cases, the board has made findings promptly, and those findings have passed the test of time.

In the event of any future noncompliance reports from the safeguards department, board members can refer to past cases for criteria to arrive at findings that are consistent with the board’s duty to inform the Security Council of questions falling within the council’s competence. This body of precedents can be considered as having some of the characteristics of case law. The attraction of a case-law approach is that it offers guidance to those of whom a judgment is required, but leaves them free to take account of factors that framers of codified legal systems may be unable to foresee.

In this instance, no more than an approximation to a case-law approach can be possible. The board does not articulate explicitly its reasons for arriving at or refraining from noncompliance findings, and this could lead to disagreement as to the interpretation of past cases and the guidance to be derived from them. The preambular paragraphs of the board resolutions that frame noncompliance findings and board discussion records can go some way toward addressing that deficiency.

A case-law approach or, more precisely, something akin to it, can help fill the vacuum created by the absence of a definition of noncompliance in the IAEA statute while leaving intact the freedom of governments to give a state that has breached its undertakings the benefit of the doubt or to delay a finding to create space for diplomacy. It allows governments the possibility of exercising political judgment in the IAEA board as well as the Security Council, thereby preserving options for wise management of the nuclear nonproliferation regime. By giving some guidance but not too much, it can strike a good balance.

Peter Jenkins was permanent representative of the United Kingdom to the International Atomic Energy Agency from 2001 to 2006. He is currently an associate fellow of the Geneva Centre for Security Policy and a partner in ADRg Ambassadors. He wishes to acknowledge the advice and specific textual suggestions provided by John Carlson and Pierre Goldschmidt during preparation of this article. The views expressed in this article are the author’s own.


1. International Commission on Nuclear Non-proliferation and Disarmament (ICNND), “Eliminating Nuclear Threats: A Practical Agenda for Global Policymakers,” December 15, 2009, www.icnnd.org/reference/reports/ent/index.html.

2. The report reads as follows:

Under the IAEA’s Statute, safeguards inspectors have the responsibility of determining in the first instance whether a state is in compliance with its safeguards agreement. They are required to report any non-compliance to the IAEA’s Board of Governors (consisting of 35 of its 150 member states, meeting five times a year), and if the Board finds that non-compliance has occurred, it is required to report the non-compliance to the Security Council. Confidence in the security guarantees afforded by the NPT depends to a large extent on how well compliance problems are addressed by this system. A basic problem is that a finding of non-compliance almost inevitably involves both technical and political dimensions: it appears for example that concern about the possible adverse consequences of a non-compliance finding led in the case of Iran to the finding being delayed by three years, with significant risk to the integrity and credibility of the IAEA processes.

It is important, if that credibility is to be maintained, that the IAEA confine itself essentially to technical criteria, applying them with consistency and credibility, and leaving the political consequences to the Security Council to determine. Issues of standard of proof become relevant here, and the IAEA has not helped itself by in practice setting the bar higher than its own standard safeguards agreements, which provide, for example, that a state may be found in non-compliance if the agency is not able to verify that there have been no diversions.

Ibid., paras. 9.14-9.15.

3. The relevant section of the article states, “The inspectors shall report any non-compliance to the Director General who shall thereupon transmit the report to the Board of Governors. The Board shall call upon the recipient State or States to remedy forthwith any non-compliance which it finds to have occurred. The Board shall report the non-compliance to all members and to the Security Council and General Assembly of the United Nations.”

4. Diversion can involve declared nuclear material, i.e., nuclear material placed under safeguards, or undeclared nuclear material, i.e., nuclear material that should have been placed under safeguards. Diversion constitutes using such material for nuclear weapons or nuclear explosive devices or for “purposes unknown.” This last provision covers cases where the IAEA cannot establish the intended use of the nuclear material in question, in circumstances in which a nuclear weapon or explosive purpose is plausible.

5. IAEA, “The Structure and Content of Agreements Between the Agency and States Required in Connection With the Treaty on the Non-proliferation of Nuclear Weapons,” INFCIRC/153.CORR, June 1972, www.iaea.org/Publications/Documents/Infcircs/Others/infcirc153.pdf.

6. The absence of a clear definition is one reason that it would be a mistake for the board to confine itself to this criterion. The words “unable to verify” are open to two interpretations. They can imply that verification has proved impossible because a state under safeguards has withheld cooperation, or that IAEA inspectors have insufficient grounds to state that diversion has not taken place. The former interpretation is narrow; withholding cooperation is only one of several possible indications that a situation has arisen that ought to be cause for concern. Withholding cooperation has been a feature of only three of the five noncompliance cases that have been reported to the Security Council.

7. Under Article 39 of the UN Charter, the Security Council is responsible for “maintain[ing] or restor[ing] international peace and security.”

8. John Carlson, “Defining Noncompliance: NPT Safeguards Agreements,” Arms Control Today, May 2009, pp. 22-27.

9. At the time, the director-general was insisting that the secretariat use synonyms for noncompliance such as “breach of the obligation to comply.”

10. See, for example, Pierre Goldschmidt, “Exposing Nuclear Non-Compliance,” Survival, Vol. 51, No. 1 (February-March 2009), pp. 143-164.

11. For example, a British Cabinet Office document states,

We must continue to work with international partners and the Agency to find and report promptly to the UN Security Council any instance of non-compliance. This would be part of a wider strategy to build international consensus for ensuring that any non-compliance with IAEA safeguards agreements…triggers strong [UN Security Council] action. The UK would like to discuss with the Agency and Member States at the earliest possible opportunity past practices in exposing and addressing non-compliance by the Agency, with the aim of reaching a common understanding on the definition of non-compliance.

British Cabinet Office, “The Road to 2010: Addressing the Nuclear Question in the Twenty-First Century,” Cm 7675, July 2009, para. 6.3, www.cabinetoffice.gov.uk/media/224864/roadto2010.pdf.

12. British Secretary of State for Foreign and Commonwealth Affairs, “Iran’s Nuclear Programme: A Collection of Documents,” Cm 6443, January 2005, p. 41, www.fas.org/nuke/guide/iran/uk2005.pdf (“Joint Statement at the End of a Visit to the Islamic Republic of Iran by the Foreign Ministers of Britain, France and Germany,” October 21, 2003, para. 3c).

13. The board took action less than two years after its November 2003 session, not the “nearly three years” mistakenly alleged by the ICNND. ICNND, “Eliminating Nuclear Threats,” para. 9.14.