LOOKING BACK: The Additional Protocol

Trevor Findlay

The theory of punctuated equilibrium posits that life on Earth has evolved not in constant, linear fashion but through long periods of stasis, interrupted by catastrophic events, such as meteor impacts, which suddenly push it in new, adaptive directions. Without stretching the metaphor too far, the evolution of nuclear safeguards can be viewed in this way. It has been characterized by long periods of continuity, interrupted by extraordinary events that have changed its nature and direction.

Notable among these events, welcome and disturbing alike, have been the advent of the 1968 nuclear Nonproliferation Treaty (NPT), the 1974 Indian nuclear test, and the 1991 discovery of Iraq’s violation of the NPT, along with the later cases of North Korea, Libya, and Iran. A significant adaptive outcome of the Iraq case was the strengthened safeguards system, of which the Model Additional Protocol, launched in 1997, is a major component. On the tenth anniversary of the Model Additional Protocol, this article reviews the evolution of nuclear safeguards that led up to it, assesses its achievements and failings, and speculates about the next likely punctuation in the evolution of nuclear safeguards.


The origins of the nuclear safeguards concept and perhaps the term itself may be found in the 1946 Acheson-Lilienthal report, which concluded that “a system of inspection imposed on an otherwise uncontrolled exploitation of atomic energy by national governments will not be an adequate safeguard” against the production of fissionable material for nuclear weapons.[1] After the Soviet Union rejected the 1946 Baruch Plan for the internationalization of all nuclear capabilities, the safeguards about which Acheson and Lilienthal had been so skeptical became the default concept. They were embodied in the statute of the International Atomic Energy Agency (IAEA), founded in 1957 as both facilitator of the spread of nuclear science and technology and its chief “safeguarder.”

The IAEA was mandated to “establish and administer safeguards designed to ensure that special fissionable and other materials, services, equipment, facilities and information…are not used in such a way as to further any military purpose.”[2] The agency could review the design of equipment and facilities, including reactors; require the maintenance of operating records for the use of nuclear material; and call for reports from states. Most extraordinarily, in terms of intrusiveness, the IAEA could conduct on-site inspections, giving it “access at all times to all places and data and to any person…as necessary to account for…materials” in order to determine compliance.[3]

In the early years, most IAEA safeguards resulted from the transfer of existing bilateral arrangements to the agency. Based on IAEA document INFCIRC/26, the safeguards applied only to materials and facilities transferred from one state to another, notably small reactors, and have been described as “technically amateurish.”[4] After the Soviet Union became convinced of the security benefits of containing nuclear proliferation, a more elaborate and intrusive model became possible, based on agency document INFCIRC/66/REV.2 of 1965. Some of these early agreements survive, notably those applied to select facilities in the three states still outside the global nonproliferation regime: India, Israel, and Pakistan.[5]

Enter the NPT: Comprehensive or Full-Scope Safeguards

The most dramatic punctuation of this early period was the negotiation and entry into force of the NPT. The treaty formalized a division of the world into nuclear-weapon states and non-nuclear-weapon states, along with the concept of a verifiable division between peaceful and nonpeaceful nuclear programs. The NPT imposed a legal obligation on its non-nuclear-weapon states-parties to place all of their nuclear activities under IAEA safeguards, hence the terms “full-scope” and “comprehensive.” Each state-party is obligated to negotiate a bilateral safeguards agreement with the IAEA based on frameworks and models encapsulated in IAEA document INFCIRC/153.

INFCIRC/153 remains the foundation of modern IAEA safeguards, used not just for the NPT but for nuclear-weapon-free zones, the first of which, established by the 1967 Treaty of Tlatelolco, preceded the NPT by a year. It envisaged the same methods and practices as INFCIRC/66,[6] notably nuclear materials accountancy and inspections, but introduced new concepts to improve effectiveness: subsidiary arrangements to tailor safeguards to each state and protect confidentiality; the focusing of safeguards on strategic points where verification might be most revealing; the use of instrumentation and nonhuman inspection techniques (today, increasingly, continuous real-time remote monitoring using video cameras); surveillance and containment as important complements to material accountancy; and a requirement that states establish state systems of accountancy and control. INFCIRC/153 also placed some limitations on the agency, regulating the designation and right of rejection of agency inspectors by states and setting out dispute resolution arrangements.

The detonation of a nuclear device in May 1974 by non-NPT party India was an unexpected punctuation for this new safeguards regime only four years after entry into force of the NPT. Although India had not violated its IAEA safeguards agreement because it did not have one, the test spotlighted the “peaceful nuclear explosion” loophole in safeguards language relating to “weapons purposes” and led to the establishment in 1975 of the Standing Advisory Group on Safeguards Implementation. That group has been charged with recommending technical improvements to safeguards ever since.

Thereafter, the safeguards system entered a period of equilibrium. The number of states-parties to the NPT increased to near universality, and apart from some relatively minor infractions discovered in the early 1980s, the IAEA was able to declare annually that it had no evidence of the diversion, by any non-nuclear-weapon state-party, of peaceful nuclear material or facilities to nonpeaceful purposes.

The Iraq Debacle and Response

The sharpest of punctuations to this equilibrium came with the revelation, after the 1991 Persian Gulf War, that Iraq had been clandestinely conducting a nuclear weapons program in parallel with its IAEA-inspected peaceful program. The IAEA’s failure to detect such activities, located in some cases just over the berm from where inspectors visited, produced widespread criticism of safeguards, both justified and unfair. It also provided a political window of opportunity for significant strengthening of the system.

The fundamental problem was that the IAEA could only monitor and inspect materials and facilities declared by states-parties. Determined proliferators could develop substantial undeclared nuclear capabilities undetected, either co-located with declared facilities or apart from them. Although a so-called special inspection, the equivalent of a challenge inspection in other disarmament regimes, could be requested, these were politically and practically difficult to launch. After the IAEA board, post-Iraq, reiterated its right to seek special inspections, it found itself peremptorily refused on its first attempt, in North Korea in 1993.

A further difficulty was the reliance on nuclear accountancy as the principal tool for detecting diversion and, in turn, the reliance on safeguards as the key tool in detecting noncompliance with the NPT. The agency lacked the full range of modern tools for such a verification challenge, although even states with satellite imagery and sophisticated intelligence services had missed Iraq’s illicit activities. Nonetheless, the agency demonstrated the continuing utility of safeguards in 1993 by detecting North Korea’s noncompliance with its new safeguards agreement by calculating, using materials accountancy, that declarations of its plutonium production were improbably low.

Together, the Iraq and North Korea crises, along with South Africa’s 1989 revelation that it had secretly produced a small nuclear arsenal,[7] although as a nonparty to the NPT, proved sufficiently punctuating to trigger relatively rapid, adaptive evolution of safeguards in a new direction.

Strengthened Safeguards, Including the Additional Protocol

Jolted by such wanton noncompliance cases, the IAEA board in 1993 mandated that the secretariat propose legal, technical, and financial means of strengthening safeguards. The recommendations resulted in a two-pronged program. Part one comprised measures the agency concluded it already had the legal authority to undertake and that could begin immediately. These included requests for additional information from states on their former and future nuclear facilities; increased use of unattended monitoring devices transmitting data direct to IAEA headquarters in Vienna; expanded use of short-notice and unannounced inspections at declared facilities; and the introduction of environmental sampling. In addition, the agency revolutionized its use of open-source information, including satellite imagery, which is increasingly cheaply available commercially, as well as accepting intelligence information from member states.

Part two involved states providing the agency with the legal authority for further measures by individually concluding a supplement to their comprehensive nuclear safeguards agreements. In May 1997, the board agreed on a Model Additional Protocol,[8] which expanded the verification responsibilities of the agency and each state-party. It increased transparency by extending states’ declaration, reporting, and site access obligations to encompass the range of nuclear fuel cycle activities from mining to the storage of nuclear waste. The protocol also requires states to report on nuclear equipment production, imports and exports, fuel cycle research and development, and future plans for facilities. Parties are required to provide an expanded declaration of their nuclear activities within 180 days of entry into force of their additional protocol.

The IAEA would now aim at a holistic, as opposed to a materials- and facilities-based, view of states’ nuclear activities. It would for the first time seek “credible assurance not only about declared nuclear material in a state but also about the absence of undeclared nuclear material and activities.”[9] Complementary access could be sought by inspectors to resolve ambiguities discerned at declared and undeclared sites.

The power of strengthened safeguards has been demonstrated by the revelation of previously undisclosed nuclear activities in Egypt, South Korea, and, more significantly, Iran. In the case of Iran, the extra information requirements and agency powers, including those under the additional protocol, have proved potent. (Iran has signed but not ratified an additional protocol, and has sent mixed signals about its willingness to comply.) Although Iranian opposition groups first revealed the deception, new safeguards measures have helped uncover the details of 20 years of dishonesty and provided a constant stream of leads for the IAEA to pursue through requests for further information and follow-up inspections. Environmental sampling has proved illuminating.

The new IAEA safeguards capabilities provide increased reassurance that such noncompliance in the future will be detected much earlier. Yet, limitations remain. A major challenge is that concluding an additional protocol is voluntary, making it likely that only states intent on complying will join without pressure, which has been applied in the cases of Libya and Iran. Although there have been calls for the IAEA board to make the protocol compulsory, there is little stomach for venturing so far at this stage. It will probably take another punctuation in the safeguards equilibrium for this to happen.

In the meantime, after a slow start the number of states with additional protocols is growing to the point where it is becoming the norm. As of September 2007, 84 states had such protocols in force. The real surprise is the number of NPT states-parties that have failed to comply with their legally binding obligation to have a comprehensive safeguards agreement in force. As of September 2007, they numbered 30, mostly African states. This partly explains why the adoption of the Model Additional Protocol has been so slow. Without a comprehensive agreement, there is nothing to which the protocol can be attached. Other countries, such as the United States, have a policy of not bringing the protocol into force until their domestic legislative and other mechanisms are in place. Indeed, states without such mechanisms in place risk missing their 180-day declaration deadline under the protocol, which some have. Political and legislative lethargy and a lack of awareness by states not represented in Vienna are added reasons for the slow adherence to the protocol, despite the IAEA’s valiant attempts through regional meetings and other outreach activities.

A further difficulty is that some of the additional powers that the IAEA acquires under the Model Additional Protocol are neutered in states that also have a Small Quantities Protocol (SQP). Introduced in 1971, such protocols hold in abeyance significant safeguards obligations, including declarations and inspections, when nuclear activities remain under a certain low threshold. Controversy over the SQPs arose when Saudi Arabia, a state with evident nuclear energy ambitions, sought one. In September 2005, the board directed the agency to begin renegotiating with SQP states to restore at least some of the IAEA’s powers based on a revised model. States with existing SQPs were invited to exchange letters with the IAEA to trigger implementation of the new model, while all future SQPs will be based on the new one. This would oblige states to submit a declaration on their nuclear holdings, however small, which in turn forces them to institute a national system of nuclear materials accounting and tracking. This should be especially useful in strengthening national measures to avoid theft and illicit transshipments of nuclear material.[10] The initiative is, however, again dependent on the goodwill of the states concerned and is proceeding slowly. Currently, 79 states have old SQPs in force; 13 have the new version in force with five others in process; and one has been rescinded.[11]

Finally, the Model Additional Protocol still leaves the IAEA a long way from the anytime, anywhere verification envisaged in its statute. Complementary access requires at least 24 hours’ notice. If inspectors are already at the site in question, they must give two hours’ advance notice. A demand for a special inspection remains an extraordinary, highly politicized option. There is still a possibility that undeclared facilities could go undetected. A state bent on noncompliance will take active measures to conceal its activities, including disinformation and delaying tactics of the type that Iran has deployed.

Nonetheless, the strengthened safeguards system increases the costs and risks for a potential proliferator. It has also to some extent liberated the IAEA from its past timidity, both mandated and self-imposed, and certainly emboldened current Director-General Mohamed ElBaradei in examining the entire range of signals of a proliferator’s intentions. It is notable, for instance, that the agency has concerned itself with evidence of the links between Iran’s military and its alleged peaceful nuclear program, something it previously would have felt was beyond its official verification remit.

The agency meanwhile has also moved to rationalize layers of safeguards imposed on selected states over the years, thereby increasing efficiency and, it is hoped, effectiveness through a program of integrated safeguards. To date, only a handful of countries, notably Australia, Canada and Japan, have been admitted to this select group. This is partly a reward for punctilious compliance with all aspects of safeguards, including the Model Additional Protocol, as candidate states must undergo rigorous examination and cross-examination before qualifying. An unspoken benefit for the IAEA is that verification resources can in theory be devoted to more problematic cases.

The existing system remains as underfinanced and under-resourced as ever. In addition to a decade of zero budgetary growth that ended in 2003, the agency has been obliged to assume increasingly onerous and numerous verification and safeguards obligations, notably in Iran, Iraq, Libya, North Korea, South Africa, and the non-nuclear successor states of the Soviet Union and Yugoslavia. It is also cooperating with the United States and Russia in repatriating highly enriched uranium from research facilities in vulnerable locations around the world. If the U.S.-Indian nuclear cooperation agreement, which envisages India designating a whole raft of nuclear facilities as peaceful and therefore subject to safeguards, ever enters into force, the cost could be as much as $10 million annually. The strengthened safeguards system is itself more costly and labor intensive, while integrated safeguards have yet to produce significant, if any, savings.

In June 2007, ElBaradei decried the board’s refusal to approve a requested increase of 4.6 percent in the IAEA annual budget of around $275.5 million, revealing that the agency’s safeguards function was being “eroded over time.”[12] He noted that the organization was forced to use an unreliable 28-year-old instrument for environmental sampling and rely on external laboratories for analysis, “which puts into question the whole independence of the agency’s verification system.” Moreover, there has been no general implementation of wide-area environmental sampling due to the projected cost. The threat to safeguards is compounded by what the U.S. Government Accountability Office has described as “a looming human capital crisis caused by the large number of inspectors and safeguards management personnel expected to retire in the next 5 years.”[13]

The recent attempt by the United States to engage IAEA board members in further improving safeguards ended in failure when the Advisory Committee on Safeguards and Verification folded last June after two unproductive years. Although the secretariat proposed at least 18 improvements for consideration, the committee was unable to adopt a work plan. With its membership open to all IAEA members, Iran was able to pursue a wrecking strategy. Even the United States seemed unwilling to devote the necessary political and financial capital into making the committee succeed. Perpetual issues over the relative attention devoted to nuclear safeguards and assistance in the peaceful uses of nuclear energy and nuclear disarmament also contributed to the paralysis.

The Next Punctuation: Future Challenges

Nuclear safeguards remain a work in progress. If the punctuated equilibrium theory is correct, it will require another crisis for significant new improvements to be made. This may occur if the much-heralded nuclear energy revival ever comes to fruition. Increased numbers of research and power reactors, additional nuclear trade and transport, and moves by more states to acquire the full nuclear fuel cycle will require increased IAEA safeguards capacity and spending. It may also awaken a sleeper issue that has long exercised the sharpest critics of safeguards: the fact that the current system cannot provide sufficient assurance of nondiversion of fissionable material from bulk-handling facilities, such as those involved in uranium enrichment, plutonium reprocessing, and fuel fabrication. These facilities handle such large volumes of nuclear material that significant amounts, in terms of the quantities required for an illicit nuclear device, will be unaccounted for, lodged in pipes or other equipment or subject to accounting and measurement errors. The system is also currently unable to verify rapid adaptation of enrichment and reprocessing plants from declared peaceful purposes to production of weapons-useable materials. Moreover, some critics claim that the IAEA’s 30-year-old criteria, suggested by the nuclear-weapon states, for how much nuclear material is needed to make a nuclear weapon (“significant quantity”) and how much time is required to convert such materials into a bomb (“conversion time”) need significant revision downward.[14]

If a nuclear energy revival permits increasing numbers of non-nuclear-weapon states to acquire such facilities, the safeguards system risks losing credibility. Proposals for fuel banks and regional or multilateral enrichment facilities and the phaseout of the use of plutonium for civilian purposes are widely deemed to be appropriate means for dealing with the proliferation implications of these developments, but all of them imply more powerful nuclear safeguards tools beyond even today’s strengthened system. In the meantime, it would be useful for the IAEA to tell its member states frankly where it is unable to achieve verifiability. This will not only help relieve the agency of the perennial burden of overblown expectations but should catalyze radical new improvements in areas where they are feasible.

Faced with such challenges, the future evolution of nuclear safeguards lies in the realization by the international community that this form of verification is a security bargain that deserves openness, hard-headed scrutiny, commitment, finances, and resources commensurate with its significance for international security.

Trevor Findlay is director of the Canadian Centre for Treaty Compliance, Carleton University, Ottawa, and of the Nuclear Futures Project at the Centre for International Governance Innovation, Waterloo, Ontario.


1. “Nuclear Safeguards: A Reader,” Congressional Research Service, Washington, DC, December 1983, pp. 46-51 (excerpts from “A Report on the International Control of Atomic Energy,” [Acheson-Lilienthal report]).

2. Statute of the IAEA, art. III.A.5.

3. Ibid., art. XII.A.6.

4. Carlos L. Büchler, “Safeguards: The Beginnings,” International Atomic Energy Agency: Personal Reflections (Vienna: IAEA, 1997), p. 48.

5. In NPT states-parties, they are “suspended” but would be reactivated automatically should NPT-based safeguards disappear.

6. Lawrence Scheinman, The International Atomic Energy Agency and World Nuclear Order (Washington D.C: Resources for the Future, 1987), pp. 153-154.

7. See Jack Boureston and Jennifer Lacey, “Shoring Up a Crucial Bridge: South Africa’s Pressing Nuclear Choices,” Arms Control Today, January/February 2007, pp. 18-21.

8. “Model Protocol Additional to the Agreement(s) between State(s) and the International Atomic Energy Agency for the Application of Safeguards,” INFCIRC/540 (Corrected), Vienna, September 1997.

9. Jan Lodding, “Non-Proliferation of Nuclear Weapons and Nuclear Security: IAEA Safeguards Agreements and Additional Protocols,” IAEA, 2002, p. 2 (quoting IAEA Director-General Mohamed ElBaradei).

10. See Jan Lodding and Bernado Ribeiro, “Strengthening Safeguards in States With Limited Nuclear Activities,” Trust & Verify, No. 123 (March 2006-March 2007), pp. 1-4.

11. Safeguards current status as of September 27, 2007. See www.iaea.org.

12. Julian Borger, “Nuclear Watchdog Might Not Cope in Atomic Crisis,” The Guardian, June 22, 2007, www.guardian.co.uk/korea/.

13. U.S. Government Accountability Office, “Nuclear Proliferation: IAEA Has Strengthened Its Safeguards and Nuclear Security Programs, but Weaknesses Need to Be Addressed,” GAO-06-93, Washington, DC, October 2005.

14. See Henry Sokolski, “Falling Behind: International Scrutiny of the Peaceful Atom,” Nonproliferation Policy Education Center, September 2007.