A regime to verify the abolition of nuclear weapons must involve three distinct components1 : Verifying the non-diversion of declared civilian nuclear materials; verifying the dismantlement of declared warheads, delivery systems, launchers and nuclear-weapon related facilities, and verifying the absence of undeclared nuclear warheads, their components, nuclear materials, activities and facilities.

The world has extensive experience of the first of these tasks—verifying that nuclear material from civilian programmes is not diverted. This is, of course, a core purpose of the International Atomic Energy Agency (IAEA) safeguards that all nonnuclear- weapon state parties to the Nuclear Non-proliferation Treaty (NPT) are required to accept. Certainly, there is much room for improvement. To my mind, a significantly more effective safeguards system for declared nuclear material would be needed to facilitate abolition. But, the IAEA has been safeguarding declared facilities for over five decades. The experience it has acquired clearly provides a huge head start in constructing a regime to ensure that, in a world without nuclear weapons, the diversion of declared nuclear material would be rapidly detected.

There is also much knowledge, particularly in Russia and the United States, about how to verify the dismantlement of nuclear weapons—at least when it comes to missiles, bombers and launchers. Warheads, by contrast, have never been subject to verified dismantlement. This is a difficult problem. But, it is also one that Russia and the United States have made progress—individually and cooperatively—in tackling. They have been joined in their efforts recently by Norway and the United Kingdom. I think almost all the participants in these endeavours would agree that there is much more work to be done. However, I also believe that with sufficient effort a ‘good enough’ solution is attainable.

The third and final task is verifying the absence of undeclared nuclear warheads, materials, activities and facilities. This task is crucially important. Indeed, given my basically optimistic conclusion about the feasibility of verifying the dismantlement of declared weapons, it is natural to ask why a state would go to the trouble of trying to defeat such a verification system when it had other, much more attractive options for cheating at its disposal.

One way of cheating would be to produce fissile materials for nuclear weapons in an undeclared facility. Detecting undeclared facilities—especially small gas centrifuge enrichment plants—is extremely difficult. The IAEA is already assigned this task in all non-nuclear-weapon states—but it has not always been successful.2 Clearly significant improvement would be needed here.

But, nuclear-armed states would have an even easier cheating option at their disposal. They could simply fail to declare some warheads and keep them hidden in a secret stash. Warheads are small and easily moveable. Although radioactive, they can easily be shielded. Moreover, the problem is not limited to warheads; states could—instead of or in addition to retaining clandestine warheads—retain an illicit stockpile of fissile material.

Detecting clandestine stocks of warheads or fissile material presents the single biggest challenge to verifying disarmament. Yet it is also the problem that has attracted the least attention. I want to focus the remainder of my remarks on what states can do—in the short term—to start to solve this problem.

One part of the solution is to try to verify how much fissile material today’s nucleararmed states have produced. By comparing past production to current holdings, nuclear-armed states could—in theory at least—demonstrate that they had declared all their fissile material. Unfortunately, in practice, for reasons I won’t discuss now, some discrepancy would be inevitable even in states that were faithfully complying with their obligations. Thus, although this technique certainly has value and would surely be part—an important part—of the regime for verifying an abolition agreement, it would also not be enough by itself.

On-site inspections would be a crucial supplement. If the international body charged with verifying disarmament—or, perhaps, depending on the terms of the prohibition agreement, a state party to it—had reason to suspect that another had retained proscribed items or materials, it could demand a ‘nuclear challenge’ inspection to investigate further. 

While nuclear challenge inspections would be likely to form an important part of any verification regime, a number of problems must be solved for them to become a truly useful tool.

Intrusive inspections risk compromising secrets that states have legitimate reasons to keep. These could be nuclear-related—for example, there may be commercially sensitive information, such as advanced centrifuge designs, that states and companies will not want to disclose. Classified information about conventional weapons programmes is, however, more likely to cause problems. When, for example, evidence of a possible link between Iran’s nuclear programme and its military complex at Parchin, near Tehran, came to light, the IAEA’s request for inspection access became highly contentious for this reason.

Realistically, for states to accept a regime permitting intrusive nuclear challenge inspections, some limits will have to be placed on the rights of inspectors. The most intrusive form of challenge inspection yet negotiated was that developed for the Chemical Weapons Convention. Under the convention, if a state obtains evidence that another may be secretly manufacturing chemical weapons, it may request that the Organisation for the Prohibition of Chemical Weapons conduct a challenge inspection at the suspect location. In theory, a challenge inspection could take place anywhere, including at a nuclear-weapon production facility. However, the host state is still permitted to manage access, for example by removing ‘sensitive papers from office spaces’ or shrouding ‘sensitive displays, stores and equipment’. Furthermore, the host state is only required to give such access as is consistent with national security concerns.

In the context of abolishing nuclear weapons, a key question is whether it is possible to negotiate an inspection regime that is capable of uncovering clandestine fissile materials but also allows states to keep legitimate secrets secret. Or, in other words, can nuclear challenge inspections be both effective and acceptable?

This is a question that all states—nuclear-weapon states and non-nuclear-weapon states—could start examining today by trying to develop appropriate managed access protocols. After all, in a world without nuclear weapons, such inspections could take place on the territory of any state, whether or not it used to have nuclear weapons.

A different problem is whether nuclear challenge inspections would be usable. There has never been a challenge inspection pursuant to the Chemical Weapons Convention. There have been accusations—particularly by the United States—that some of the convention’s signatories retain banned capabilities but not once has the state making such an accusation requested a challenge inspection. The failure to request a single inspection in the 14 years that the Chemical Weapons Convention has been in force seems to have increased reluctance to use them.

In the nuclear context, the IAEA has the similarly powerful right to conduct special inspections. However, it has only ever requested one, in North Korea in 1993. North Korea refused access, and this failure appears to have deterred the IAEA’s secretariat from calling any more.

Notably, the IAEA has not requested a special inspection in Syria in spite of extremely strong evidence that Syria had built a clandestine nuclear reactor; evidence that was—quite rightly in my view—deemed strong enough to find Syria in non-compliance with its safeguards agreement. This inevitably raises an important question: If the IAEA did not request a special inspection in Syria, under what circumstances would it ever ask for one?

Perhaps, however, it is not surprising that the IAEA and states party to the Chemical Weapons Convention have made so little use of special and challenge inspections.

Some states party to the Chemical Weapons Convention worry that if they request a challenge inspection on another state they may open themselves up to revenge requests on their own territory. The IAEA worries about the consequences of requesting a special inspection and being met by a refusal. Moreover, such inspections are only as useful as the information available to guide them. However, states that have intelligence that could be useful in this regard are often unwilling to share it, while some other states oppose the use of intelligence information in international verification entirely! On top of that, everyone worries about what would happen if a challenge or special inspection were instigated but nothing was discovered.

The key challenge to challenge inspections is making them usable. For an abolition agreement to be verifiable, it would be critical for nuclear challenge inspections to exist in practice as well as on paper.

Fortunately, there is a solution—and a relatively simple one. States and the IAEA need to start using challenge and special inspections. How much lower would the political barriers to their use become if they were requested routinely? How much more verifiable would an abolition agreement appear if we could point to the regular use of challenge and special inspections in today’s world?

So, let me advance some conclusions. Verifying abolition is not a solved problem; but, neither is it necessarily an insoluble one. Indeed, there is much that states—all states—nuclear-weapon states and non-nuclear-weapon states—could do today to help surmount the real challenges that do exist.

That said, as we all recognize, verification is never going to be perfect. The real question, of course, is whether it will be good enough. This is ultimately a political judgement, not a technical one.

In the final analysis, whether verification is deemed sufficient should depend upon the enforcement system as a whole. Verification is, after all, not an end in itself but a means to ensuring treaty compliance. If the response to treaty violations is swift, certain and effective then even an imperfect verification system may be good enough to deter violations and hence prevent rearmament.

This may appear a very pessimistic note on which to end; after all the international community’s response to treaty violations today could hardly be called swift, certain or effective. However, my conclusion is actually a profoundly optimistic one. After all, if there were no way to compensate for the inevitable deficiencies in verification, then nuclear-armed states and their allies would, in all probability, never agree to give up their nuclear weapons. It is precisely because politics can, perhaps, solve the problems that technology cannot, that a world without nuclear weapons could be an attainable goal.

1. This paper draws heavily from George Perkovich and James M. Acton, Abolishing Nuclear Weapons, Adelphi Paper 396 (Abingdon: Routledge for the IISS, 2008), chapter 2. This is reprinted in George Perkovich and James M. Acton (eds.), Abolishing Nuclear Weapons: A Debate, (Washington, DC: Carnegie Endowment for International Peace, 2009), http://www.carnegieendowment.org/files/abolishing_nuclear_weapons_debate.pdf.

2. The IAEA’s responsibility to verify the absence of undeclared nuclear material stems from paragraph 2 of the Comprehensive Safeguards Agreement (INFCIRC/153), which all non-nuclearweapon states party to the NPT are required to accept. The Additional Protocol gives the IAEA the tools it needs to draw credible conclusions in this regard.