News Fighting Proliferation

Chapter 2

The Nuclear Nonproliferation Treaty:
Strengths and Gaps

Leonard Weiss

The evolution of a strong nonproliferation ethic in the world is, ultimately, the best stable, long-term tool to prevent the spread of nuclear weapons. Such an ethic can stimulate and is, in turn, stimulated by the creation of international institutions incorporating the notion of nonproliferation at their core. The Nuclear Nonproliferation Treaty (NPT),1 despite the confused philosophy of its provenance, has become such an institution and has demonstrated its value—especially during the past few years. It remains, however, a flawed institution that requires considerable tending to, including constant efforts to obtain a consensus of its parties concerning evolving interpretations of its provisions in order to maintain its effectiveness as a nonproliferation tool—if not its survival altogether.

One should not be surprised to learn that the treaty is an imperfect nonproliferation instrument. It was created in response to nonproliferation concerns arising from burgeoning nuclear trade accelerated by a misguided Atoms for Peace policy. This trade was promoted aggressively by nuclear policymakers, technocrats, and diplomats whose visions of prosperity, generated by nuclear technology, obscured the very real national and international security problems being created. Those problems, when they emerged, seem to have been viewed as much in terms of the threat to future nuclear commerce as they were in terms of the threat to life. Accordingly, the treaty was designed to endorse and encourage the spread of nuclear technology for peaceful purposes at the same time it was to constrain—indeed prevent—the development and manufacture of nuclear weapons.

The incompatibility of these aims became apparent after the treaty went into effect in 1970 as some nuclear suppliers— particularly Germany and France (one an NPT party and the other pledged at the time to act as an NPT party)—prepared to export technology and equipment for production of fissionable material, albeit under safeguards administered by the International Atomic Energy Agency (IAEA), to certain countries. These nations either were not NPT parties and were embarked on secret military programs to develop nuclear weapons (Pakistan and Brazil) or were NPT parties whose nonproliferation credentials were suspect at the time (South Korea).

What followed over the next few years (and continues today) was a development of other institutions outside the NPT—a development designed to patch the omissions, ambiguities, ill-conceived constraints, and other flaws in the treaty. Thus, we now have nuclear-supplier agreements, bilateral agreements, national and multinational export controls, national technical means of surveillance and international intelligence links, and positive and negative security assurances to assist us in keeping the nuclear genie in the bottle. These tools, along with the NPT and the associated IAEA safeguards system, are referred to collectively as the nuclear nonproliferation regime. This regime is still evolving in the direction of greater effectiveness but is not yet at the point where any of the states with nuclear weapons would be prepared to put their nuclear arsenals aside with confidence.

Why is this so? And why has it been necessary to create all these auxiliary tools to combat proliferation? What have we learned over the past 25 years that, had we known it in the 1960s, would have enabled us to construct a better NPT and a better safeguards system? In the end, does it matter? That is, would a stronger NPT enable us to rely on it for our security? The following review of the major elements of the treaty seeks to answer these questions.

Articles 1 and 2

Article 1 mandates that each state with nuclear weapons that is party to the treaty may not transfer nuclear weapons or explosive devices or control over such weapons or devices to any recipient, directly or indirectly, and may not in any way assist, encourage, or induce any state without nuclear weapons to manufacture or otherwise acquire these weapons or other nuclear explosive devices or to obtain control over such weapons or devices. Article 2 prohibits states without nuclear weapons from receiving those things which weapon states are prohibited in Article 1 from giving and specifically prohibits nonweapon states from manufacturing or otherwise acquiring nuclear explosive devices.

The first problem with Articles 1 and 2 is that it is unclear what constitutes “assistance,” “encouragement,” or “inducement” to a nonweapon state; the second problem is that it is unclear what constitutes “manufacture” of a device; the third problem is that it is unclear what constitutes a nuclear explosive device because no consensus exists on the definition of a nuclear explosion; and the fourth problem is that no prohibition exists regarding a nonweapon state’s assisting another nonweapon state with acquiring nuclear weapons.

George Bunn and Roland Timerbaev, two of the negotiators of the text of the NPT, have written on the question of what constitutes “manufacture” and quote the testimony of William C. Foster, chief of the American delegation, before the Senate Foreign Relations Committee: “The construction of an experimental or prototype nuclear explosive device would be covered by the term ‘manufacture’ as would be the production of components which could only have relevance to a nuclear explosive device.”2 He also makes reference to “activities” by a nonweapon state that would “tend” to put that state in noncompliance with Article 2 if the purpose of those activities was the acquisition of a nuclear explosive device.3

In order to allay concerns about how one would determine the purpose of certain fuel-cycle activities that could be either peaceful or weapons-related, Foster added that “neither Uranium enrichment nor the stockpiling of fissionable material in connection with a peaceful program would violate Article II so long as those activities were safeguarded.” The reference to safeguards in his statement is immaterial because if a program is indeed peaceful, then there is no violation of Article 2—even if the activity is unsafeguarded. (In that case, the party would be in noncompliance with Article 3, but that is another matter.) This points up a problem that runs throughout the NPT: lack of definitive interpretation. Bunn and Timerbaev write that the Foster criteria for manufacture have generally been accepted as authoritative interpretations by historians of the NPT negotiations, but the issue of whether all the current parties to the NPT would agree with those interpretations is unclear. It is important to note that until the Iraq situation arose, there was no indication that many of the parties to the NPT viewed the IAEA as an appropriate verification instrument to ensure that nonnuclear weaponization activities weren’t being carried out. Indeed, in the past, there were debates over whether IAEA inspectors were obligated to report any untoward activities they observed that were unrelated to the negotiated safeguards agreement (e.g., noting the presence of bomb components such as machined hemispherical metal shells somewhere on the premises).

However, the Iraq situation and South Africa’s decision to abandon its nuclear weapons program have allowed the IAEA to put its toe in the water on nonnuclear weaponization activities. In the case of Iraq, the United Nations Special Commission (UNSCOM) provided the agency information regarding the Iraqi program, and in the case of South Africa, the IAEA was invited to examine with full transparency the scope, nature, and facilities of that country’s weapon program after dismantlement, including some nonnuclear weapon components. These developments, coupled with the acceptance by NPT members of the IAEA’s ability to do “special inspections” in the wake of the Gulf War, constitute a start toward significant reform.

By contrast, one may also note that the US/North Korea Framework Agreement makes no mention of any nonnuclear weaponization activities or the disposition of any weapon components that North Korea may have manufactured, and the IAEA considers North Korea not in compliance with its safeguards obligations because of its failure to allow inspection of two nuclear waste sites. Ostensibly, if North Korea were to allow these inspections and the result were to show that all the plutonium in that country can be accounted for, then IAEA would consider North Korea an NPT party in good standing since there are no other allegations officially pending regarding its NPT commitments.

Since the existence of a North Korean nuclear weapons program is an assumption shared by most observers of the scene, it is hard to believe that some weapon components have not been manufactured by North Korea. However, the IAEA apparently will ignore this possible violation of the NPT, at least for the time being, until it can account for all the nuclear material in North Korea.

Another issue concerning manufacture is that of research and development (R&D), particularly design information. In 1975 Japan submitted a paper to the Geneva Disarmament Conference arguing that the NPT does not explicitly prohibit weapons-oriented R&D short of actual production of nuclear explosive devices.4 In rebuttal, much has been made of a statement made by the drafters during the NPT negotiations that receipt by a nonweapon state of “information on design” of nuclear explosives is barred by virtue of the prohibition on assistance in the “manufacture” of such explosives;5 however, it is unclear whether this restraint can be extended to prohibit a nonweapon state from doing its own design without external assistance.

It stretches credulity to argue that the Foster criteria bar such activity, based on an assumption that the only purpose of design is to acquire a nuclear explosive device. Some years ago, Los Alamos National Laboratory asked some recently hired young physicists with no weapons background to design a weapon based on the open literature to see if it could be done and thereby to gauge the possible extent of proliferation by this route. The purpose of the activity was not to manufacture nuclear weapons. The treaty’s vague language on “manufacture,” unless appropriately interpreted, would appear to allow anyone to design weapons, using the Los Alamos experiment and rationale, without violating the treaty. Once again, however, even if the treaty were airtight on this issue, verification of compliance would be virtually impossible.

Evidently, the Foster criteria do not settle the question of what constitutes “manufacturing.” Neither do they settle some other important questions that arise from consideration of the safeguards regime. Such consideration will also reflect on the question of what constitutes direct or indirect assistance or encouragement to manufacture or otherwise acquire nuclear weapons—discussed in a later section.

Article 3

Article 3 has four parts. Article 3.1 begins by requiring non-weapon states to accept safeguards, “as set forth in an agreement to be negotiated and concluded” with the IAEA in accordance with that agency’s statute and safeguards system “for the exclusive purpose of verification of the Parties’ NPT obligations with a view to preventing diversion of nuclear energy from peaceful uses to nuclear weapons.” The remainder of Article 3.1 states that safeguards procedures shall be followed with respect to all source or special fissionable material in all peaceful nuclear activities within the territory of the state, under its jurisdiction, or under its control anywhere.

Although there is nothing in this language explicitly referring to the effectiveness of safeguards, one may infer effectiveness from the context, because the treaty cannot be an effective nonproliferation instrument if it allows equipment, material, and technology that could be used for nuclear explosive purposes to be transferred with ineffective safeguards attached. Unfortunately, this point was not explicitly addressed by the drafters, and the question of the relationship of trade to effectiveness of safeguards (as opposed to the mere attachment of safeguards) has accordingly become a contentious issue.

In their deconstruction of the language of Article 3.1, Bunn and Timerbaev argue that 3.l authorizes the IAEA to verify that nonnuclear components for nuclear weapons are not being manufactured. Their case would not be difficult to make if the article did not contain so much emphasis in connecting safeguards to nuclear materials rather than equipment (either nuclear or nonnuclear). As a result, Bunn and Timerbaev lean part of their argument on an interpretation of a phrase which states that the purpose of safeguards is “verification of the fulfillment of [the state’s] obligations assumed under this Treaty with a view to preventing diversion of nuclear energy.” Bunn and Timerbaev connect the phrase “with a view to preventing diversion” to the state’s obligations under the treaty not to manufacture weapons,6 but an equally—if not more—plausible interpretation is that the antecedent of this phrase is “safeguards” and that the phrase has been added to provide focus on how safeguards relate in a practical way to the state’s NPT obligations. (Indeed, under the Bunn/Timerbaev interpretation, Article 3.1 would put states under an NPT obligation to establish effective physical security over nuclear materials. The fact that it does not was recognized and remedied by the voluntary Physical Security Convention developed by the IAEA and adopted by many [NPT and non-NPT] countries with nuclear programs.)

This is not to say that a case can’t be made for safeguards applying to nonnuclear weaponization activities, and Bunn and Timerbaev have made the best case possible. It is just that the emphasis in Article 3 on material safeguards, along with the history of safeguard negotiations and agreements, provides no confidence that a majority of members of the IAEA that are state parties to the NPT share this broad view of safeguards. Taking the broadest view of the stated purpose of safeguards as “verification of the fulfillment of a [nonweapon state’s] obligations” under the NPT could arguably subject to inspection the agreements and arrangements by which nonweapon states allow weapon states to place nuclear weapons on their territory. (Inspections of the agreements could ensure that no protocols exist under which transfer of authority or control over the weapons could take place.) Whether the weapon states would agree to have the IAEA inspectors examine these arrangements is, one suspects, more than problematical.

Article 3.2 provides that suppliers party to the treaty shall not provide nuclear materials or equipment for processing, use, or production of such materials to a nonweapon state unless safeguards are attached. Over a period of years, it became apparent that a more detailed and finer screen for nuclear transfers than this had to be devised in order to ensure uniformity of compliance by suppliers. The result was the so-called Zangger list of nuclear items to which safeguards must be attached and, more recently, a list of dual-use items requiring safeguards as well. In addition, the Nuclear Suppliers Group (NSG) has identified nuclear export items requiring consideration of “restraint” and “consultation” before the item is sent.7

Article 3.3 is designed to ensure that safeguards arrangements will not intrude on the ability of nonweapon states to obtain assistance for or otherwise develop their nuclear-energy activities. It references Article 4, which has been the basis for many complaints over the years regarding the policies of the suppliers—particularly the United States. Article 3.3 reflects the mind-set of the nuclear establishments and the nonweapon states at the time of the drafting of the treaty—that is, the treaty was also to be an instrument for facilitating international nuclear commerce. This mind-set resulted in a safeguards system that was designed more for its nonintrusiveness than for its effectiveness. This is still a problem, despite the improvements in the wake of the Gulf War.

Article 3.4 provides for a timetable by which states party to the treaty must enter into appropriate safeguards arrangements. This timetable has not been met many times in the past, but the most egregious example was that of North Korea, which took six years to enter into a safeguards agreement with the IAEA. No sanction was imposed on North Korea or other violators of this provision.

The Safeguards System of the IAEA

The IAEA was established in 1957 in the wake of the US Atoms for Peace initiative and began operating an inspection program in the early sixties designed to detect diversions of significant quantities of nuclear material. The NPT expanded the scope of the agency’s work significantly, and, in response, the IAEA developed a model safeguards agreement for NPT parties contained in the document known as Information Circular (INFCIRC)/153.

In this document, the IAEA states that the goal of safeguards is the prevention of proliferation by the “timely detection of diversion of significant quantities of nuclear material from peaceful nuclear activities to the manufacture of nuclear weapons or of other explosive devices or for purposes unknown, and the deterrence of such diversion by the risk of detection.”

This goal was adopted in 1970 at a meeting of the so-called Committee of the Whole, which deliberated for 11 months before the text of INFCIRC/153 was approved. Rudolph Rometsch was the head of the IAEA’s Department of Safeguards at the time; he recently said that the 1970 committee meeting led to “a sort of dogma for field work—if not to a taboo. It was a question of whether inspection should be designed also to detect undeclared facilities. The conclusion was clear at the time: looking for clandestine activities was out of the question and the inspection system was designed accordingly.”8

Thus, inspectors paid attention only to activities or structures within defined strategic points and were discouraged from asking questions about anything else lest they become persona non grata with the state (which had the right to refuse an inspector) and perhaps ultimately at IAEA headquarters.

INFCIRC/153, in addition to laying out the obligation on the part of the state to have safeguards apply to all its peaceful nuclear activities (so-called full-scope safeguards), also stresses the importance of protecting industrial and commercial secrets, not interfering in peaceful nuclear activities, and not hampering economic and technological development in the safeguarded state. This stance is in keeping with the agency’s dual role. Its charter makes it a promoter of nuclear energy at the same time it verifies that no diversions have taken place.

As a result, much negotiation follows the signing of the main safeguards agreement between the IAEA and the state to be inspected. The main agreement is followed (ostensibly within 90 days) by subsidiary arrangements that specify what the agency and the state have to do in order for safeguards to be applied: nuclear installations must be listed, and requirements for reporting to the agency are specified in negotiated detail. These subsidiary arrangements are not published.

The most specific safeguards documents are the facility attachments to the subsidiary arrangements. These state exactly what will be done at each facility containing nuclear material and lay out the “material balance areas” the agency will establish for accounting purposes. The flow of nuclear material across these areas must be reported to the agency. The facility attachments also specify the points at which measurements can be taken or samples withdrawn, installation of cameras, access afforded to inspectors, records to be kept, and anticipated frequency of inspections. Like the subsidiary arrangements, these negotiated arrangements are not published.9

Some years ago, the agency developed internally a set of technical objectives that provide a guideline for determining the level of inspection and reporting that would ensure that—at least for declared facilities in an NPT state—the goal of timely detection of any diversion of a significant quantity of nuclear materials would be met. Concern by inspected states about intrusiveness has resulted in negotiated safeguards agreements that do not come close to meeting these technical objectives and therefore cannot be said to be producing effective safeguards by any objective criterion. Inspected states have even leaned on the agency not to exercise its full rights under the agreements. In some cases, the agency itself refrains from exercising its full rights in order to conserve resources.

This is a basic problem in that the IAEA’s safeguards agreements do not provide for the agency to inspect any location—declared or undeclared—at any time (outside of regularly scheduled routine inspections) without some evidence that the site should be subject to inspection. Nor do the agreements provide for IAEA inspectors to verify use of any material formally exempted from safeguards. Thus, when inspectors doing a routine inspection in Iraq before the war asked about buildings adjacent to an Iraqi reactor, the Iraqis said they were for nonnuclear research. Since they were undeclared sites and IAEA had no evidence of suspect activity, the agency had no basis to inspect the buildings, which, as it turned out, contained a radiochemical laboratory used for research on plutonium separation.

Furthermore, the safeguards agreements effectively eliminate surprise inspections, even though, in principle, IAEA has the right to make “unannounced” or short-notice inspections. Routine inspections must provide the state with at least 24 hours’ notice, and IAEA must advise the state periodically of its general program of announced and unannounced inspections, specifying the general period when inspections are foreseen. Hence, states generally know when and where inspections will occur and, in any case, have control over the timing of admission of inspectors to the country and to the facility.

The Gulf War produced a situation in which the IAEA has successfully used its authority to conduct special inspections in Iraq, backed up by UN authority, and has received voluntary offers from a number of states to allow such inspections of declared or undeclared facilities. One of those states was North Korea, which withdrew its offer after the agency demanded to inspect two sites that the North Koreans didn’t want inspected. Those sites will be inspected at some time in the future (at least five years) under the US/North Korea Framework Agreement, which has the unfortunate effect of leaving the agency holding the bag, despite its claims of access.

The IAEA also has not resolved the problem of being unable to verify the peaceful use of nuclear materials that it has exempted from inspection. Such materials may involve (1) special fissionable material in gram quantities used for instrumentation, (2) nuclear material for production of alloys or ceramics in nonnuclear applications, (3) plutonium (Pu) of a certain isotope concentration (e.g., high in Pu-238), or (4) limited quantities ranging from one kilogram of Pu to 20 tons of depleted uranium. Iraq used an exemption for a spent-fuel assembly to conduct research on separating plutonium without informing the agency, which had no authority to routinely verify what Iraq said it was doing with the assembly.

It should be emphasized that the IAEA’s problems are not only with the Iraqs of the world. The agency has problems with many states that are not suspected of weapons development. As Lawrence Scheinman has pointed out, “Over the past twenty years, the Agency has experienced restraints on its right of access, on the intensity and frequency of inspection efforts, and even on the extent to which it could exercise its discretionary judgement in planning, scheduling, and conducting inspections.”10One should add that the agency’s technical objectives are themselves unrealistic because they are based on “significant quantities” of fissionable material that are at least twice as large as the amounts that a nonweapon state might need to construct its first nuclear explosive device.

Why doesn’t the IAEA lower the amount it considers a “significant quantity”? The answer is that inspections would then have to be more frequent and more intrusive, and the agency currently has neither the financial nor the political support to make this move.

Raising the financial question exposes the agency’s “dirty little secret.” Because safeguards are supposed to be applied nondiscriminatively, much of the agency’s safeguards budget goes to safeguards in Germany, Japan, and Canada, while the largest current proliferation concerns are elsewhere. The agency, which has been on a zero-growth budget for the better part of a decade, attempts to address its budget problems by slacking off on some inspections of facilities it considers not of proliferation concern. But in so doing, it converts its nondiscriminatory character to the status of myth and risks internal political turmoil. It cannot help this because the cost of safeguarding bulk-handling nuclear facilities such as enrichment, reprocessing, or fuel-fabrication plants is enormous, requiring, in most cases, on-site location of inspectors and much better instrumentation and measurements. While the IAEA has been required to safeguard only small reprocessing plants thus far, the ability of the agency to safeguard effectively (leaving aside the expense) a commercial-scale reprocessing plant, such as the one being built at Rokkasho in Japan, has been called into question by many people over the years.

A very interesting analysis by Marvin Miller for the Nuclear Control Institute shows that, for a reprocessing plant with a capacity of 800 metric tons per year and an average plutonium content of 0.9 percent, with a ±1 percent uncertainty in the input measurement of plutonium (assuming this dominates the error in measuring material unaccounted for [MUF]) and with a material-balance calculation done once a year, the absolute value of the MUF variance (i.e., the error in measuring MUF) will be 72 kilograms per year. In that case, the minimum amount of diverted plutonium that could be distinguished from this measurement “noise” with detection and false-alarm probabilities of 95 percent and 5 percent, respectively, is 246 kilograms—more than 30 significant quantities.11

The only admissible conclusion is that “timely detection” of plutonium diversion from a reprocessing plant is an oxymoron. This problem was recognized during consideration of the Nuclear Nonproliferation Act (NNPA) of 1978, when the concept of timely detection of a diversion was translated into the concept of “timely warning”12 of weapons development or construction. The intent of the authors was that, from a technical point of view, timely warning was unavailable in the case of plutonium diversion if it is assumed that the nonnuclear elements of the bomb have been constructed or assembled a priori. The NNPA provided that the president could still allow spent fuel of US origin to be reprocessed in a foreign country if political factors make the risk of proliferation sufficiently low, even though timely warning of weapons construction would not be available to the United States. Not wanting to admit that reprocessing—especially commercial-scale reprocessing—was a dangerous, not effectively safe-guardable activity, Reagan administration officials boldly and falsely interpreted the NNPA language as incorporating political factors into the definition of timely warning, thereby depriving the concept of any objective meaning.

In like manner, the IAEA insists that bulk-handling facilities can be effectively safeguarded, but Miller’s analysis shows that this is not the case. If the definition of a “significant quantity” of plutonium were to be changed (i.e., the amount lowered), the inability to do “timely detection” would become still worse.

The response to these practical problems from within the agency has been dismaying. Some people have advocated lowering the technical objectives—that is, moving the goalposts so that the effectiveness of safeguards couldn’t be so easily challenged.

To be sure, the agency has been chastened by its Iraq experience and is currently crafting a new safeguard approach that aims to detect tiny amounts of fissile material through environmental monitoring techniques such as wall swabs and water samples. These innovations will undoubtedly raise the cost of safeguards, and it remains to be seen how well such proposals will be received by members of the IAEA and signatories of the NPT.

Back in 1981, when the Reagan administration was formulating its nonproliferation policy, the Department of Defense (DOD), in an interagency memo, expressed concern about the IAEA’s “susceptibility to Third World . . . politics, its lack of an intelligence capability and the limits of its scope and jurisdiction.”13 While some of this complaint is being addressed in the wake of the Gulf War (the IAEA is considering how to use intelligence information brought to it by member states), the Pentagon’s 1981 warning “against undue reliance on the IAEA by those responsible for national security”14 within the US government has as much resonance today as in 1981 and will continue—especially for as long as production of fissile materials continues.

Article 4

Article 4 incorporates, in paragraph 2, one aspect of the “NPT bargain” in which nonweapon states party to the treaty, in return for their adherence, “have the right to participate in the fullest possible exchange of equipment, materials and scientific and technological information for the peaceful use of nuclear energy.” The same paragraph also calls on parties to the treaty to cooperate in contributing “to the further development of the applications of nuclear energy for peaceful purposes, especially in the territories of nonnuclear-weapon States Party to the Treaty, with due consideration for the needs of the developing areas of the world.”

In past years, the major complaints about the NPT by nonweapon states have concerned this article. These complaints range from a generic one that the technologically advanced states have not provided technical assistance or have not sufficiently shared their nuclear know-how with others, to specific complaints that the NSG (especially the United States), in seeking to control nuclear and dual-use exports or to exercise consent rights in nuclear agreements, is engaged in willful and systematic violation of Article 4.

There are a number of things to say about this. First, Article 4 does not modify the requirements of Articles 1 and 2 not to assist or receive assistance, respectively, in the manufacture of nuclear explosive devices. Second, as indicated earlier, verification of NPT obligations under Article 3 “with a view to preventing diversion of nuclear energy from peaceful uses to nuclear weapons” cannot be effectively carried out at this time for enrichment and reprocessing facilities under the safeguards system that is the instrument for the implementation of Article 3.

Accordingly, the transfer of facilities, equipment, or technology to a nonweapon state for the production of highly enriched uranium or plutonium should be interpreted as not in keeping with Article 3’s implicit qualification that effective safeguards must be applied to all peaceful nuclear activities. Otherwise, states with nuclear weapons that make such transfers could find themselves in violation of Article 1, and the NPT would become an instrument for proliferation.

Indeed, it is apparent that some states—Iraq and Libya among them—signed the NPT because they saw Article 4 as a possible route to obtaining nuclear weapons-related technology and equipment.

To date, there has been no formal resolution of the argument over Article 4, but one can interpret the Nuclear Suppliers Agreement, which calls for exercising restraint in nuclear trade involving export of reprocessing or enrichment technology, as recognition that Article 4 should not be interpreted as liberally as it appears to read. Unfortunately, the potential recipients of such trade do not accept this tightened interpretation, and—were it not for the fact that the economics of the back end of the fuel cycle have become so egregious—the argument might well be as loud today as it was in 1977, when the Carter administration began moving away from the earlier policy of relatively unrestricted nuclear trade.

Ironically, the Carter administration and the US Congress were roundly denounced in 1978 for requiring, in the NNPA, that full-scope safeguards be a criterion for nuclear export. With few exceptions, the nuclear suppliers refused to go along, despite the inference that their opposition meant they put export profits above support for the NPT. Eventually, all suppliers adopted the criterion, but it took the Gulf War to do it.

Finally, it is unfortunate—if understandable—that Article 4 is so fixated on nuclear-technology cooperation. Assuming the need for tangible incentives to produce NPT signatories in the first place, a much better NPT would have resulted if Article 4 had made cooperation in energy development (not just nuclear) the quid pro quo for an NPT signature. That way, the fight over Article 4 might have been avoided, and it would have made the phrase “with due consideration for the needs of the developing” (emphasis added) world more trenchant.

Article 6

Article 6 expresses the second part of the “NPT bargain” (Article 4 expresses the first part). In this article, “each of the Parties to the Treaty [especially including the weapon states] undertakes to pursue negotiations in good faith on effective measures relating to cessation of the nuclear arms race at an early date and to nuclear disarmament under strict and effective international control.”

Let us begin by noting that, at least in quantitative terms, the nuclear arms race—as usually defined—which included the US, the former Soviet Union, Great Britain, and France, is over. None of these countries is increasing its stockpile of nuclear arms. (That may also be true of China, but evidence is not forthcoming.) If one defines the nuclear arms race as including weapons modernization—even if the numbers aren’t going up—then the race may not yet be over. It is to this issue that a Comprehensive Test Ban Treaty (CTBT) is most relevant—not to mention the fact that a CTBT is referenced in the preamble to the NPT. Without testing, radical new designs of nuclear weapons are problematical, although simulation codes are now very highly advanced. Therefore, some nonweapon states that are parties to the NPT insist, not unreasonably, that a CTBT be a short-term goal of the NPT weapon states to fulfill part of their responsibilities under Article 6. A CTBT would have other nonproliferation benefits in that it would raise the political barriers to overt testing by nuclear states not party to the NPT.

Thus, the NPT is playing a useful role by providing a forum and a rationale for those countries interested in having a CTBT to push the weapon states—particularly the US—into a serious negotiation to formalize the current moratorium. Some members of the treaty have raised serious questions about the treaty’s future if further progress is not made toward nuclear disarmament. Despite this implied threat, it is hard to escape the conclusion that if the cold war hadn’t ended, the prospect of a CTBT being completed in the near future—let alone substantial progress toward nuclear disarmament—would be poor.

But the cold war is over, and the US now finds itself in the ironic position of being pressured by a group of countries who want progress in nuclear disarmament, who perhaps don’t mind at the same time discomfiting the weapon states, and who perhaps also enjoy the fact that many of them were asked by the US to sign the NPT during the eighties, despite their having no nuclear energy program or prospects whatsoever.

Could the NPT unravel over this issue? Hardly. There is no serious current prospect of any NPT party’s leaving the treaty or organizing a movement to terminate the treaty.

Some people believe that linking the future of the treaty to specific progress toward nuclear disarmament is a risky strategy. The latter is based on the threat of lowering political barriers to proliferation if the weapon states don’t take their obligations under Article 6 more seriously, and there is no doubt that the weapon states do not wish to see those barriers lowered. However, one can argue that an indefinite life for the treaty provides confidence that allows the weapon states to continue reducing their weapons stockpile, while threats to withdraw from the treaty designed to push the weapon states into faster progress could have—if other political factors make accelerated progress impossible—the perverse effect of putting a ceiling on progress. This result stems from the fear that the treaty might unravel and new nuclear powers might then emerge.

Article 8

Article 8 lays out the procedures for amending the treaty. For a proposed amendment to be adopted, one must first submit the text to the depositary governments (US, United Kingdom, Russia) for circulation to all parties to the treaty. Then, if requested by at least one-third of the parties to the treaty, a conference is convened to consider the amendment. Adoption occurs only if the amendment is approved by (1) a majority of the parties to the treaty, (2) all states with nuclear weapons that are party to the treaty, and (3) all parties who, on the date of circulation of the proposed amendment, are members of the Board of Governors of the IAEA.

The amendment then goes into force for those parties who have ratified it when a majority of the parties to the treaty have filed their instrument of ratification. Thus, approved amend- ments to the treaty apply only to those parties who wish to have them apply and have so indicated via ratification. The remainder of this article provides for the five-year review conferences that have taken place since 1970.

Article 10

This next-to-last article of the NPT provides that, after giving three months’ notice and an explanation, each party has the “right to withdraw from the Treaty if it decides that extraordinary events, related to the subject matter of the Treaty, have jeopardized the supreme interests of its country.” The article also provides for the 25th-year review conference to decide, by majority vote, whether the treaty shall be extended indefinitely or for an additional fixed period or periods.

Saddam Hussein would have employed the first paragraph of Article 10 to leave the NPT after putting into place the infra- structure to build nuclear weapons. Since the article contains no presumption of sanctions for leaving the treaty, the only real protection against the use of the treaty to gain technology, equipment, and materials that could be useful for weapons is to impose a set of multilateral (and unilateral) export controls on appropriate items, with sanctions for violations of those controls. This action, of course, flies in the face of the philosophy of laissez-faire technology transfer embodied in Article 4 but is necessary if the nonproliferation regime is to be worthy of its name.

Strengthening the Safeguards System

We have already discussed the deficiencies of the system in conjunction with the discussion of Article 3. To remedy those deficiencies would require the following (nonexhaustive) changes to the system:

1. The IAEA must require more transparency in the nuclear activities of its members. Among other things, this should include a complete list of sensitive or dual-use items requiring export controls and registry of trade in such items. This list should contain the union—not the intersection—of those items brought to the table by IAEA members and should cover all sensitive technologies, whether obsolete, current, or advanced.

2. The IAEA must have access to intelligence information obtained through national technical means concerning sites that may require inspection and must have an unequivocal right to inspect such sites at short notice.

3. Safeguards should apply to nuclear plants and equipment as well as materials. INFCIRC/153 safeguards, which apply to the entire fuel cycle of a nonweapon state that is a party to the NPT, should be combined with the INFCIRC/66 safeguards, which address plants and equipment as well as material for non-NPT parties. Any nuclear facility, whether it contains material or not, should be subject to inspection on short notice.

4. Safeguards should also apply to uranium concentrates such as U308—not just to U02—and to nuclear wastes containing fissionable material.

5. A definition of effective safeguards should be adopted, based on agreed measures of performance embodying appro- priate technical objectives. That is, the agency must be able to say that with a specified (high) degree of probability and a specified (low) false-alarm rate, the diversion of a significant quantity of specified nuclear material will be detected within a specified amount of time (depending on the material) that is well in advance of the time needed by the diverter to convert the material into a nuclear explosive device, assuming that all activities relating to nonnuclear weapons have been carried out.

6. The amount of nuclear material in a “significant quantity” should be reduced by at least a factor of two in the case of both uranium and plutonium.

7. All states with safeguarded nuclear activities should be required to post a bond with the IAEA, based on that state’s gross domestic product (GDP) and the size and sensitivity of its nuclear program. Violations of safeguards, IAEA regulations, and NPT commitments—as well as a decision to leave the NPT—should result in forfeiture of part or all of the bond.

8. Safeguards should be imposed on nonnuclear materials— such as tritium, lithium-6, and beryllium—useful in manufacturing weapons.

9. Safeguards should be established over nuclear R&D activities and facilities.

10. The annual safeguards implementation report of the agency should be a public document.

Interpreting the NPT to
Strengthen the Regime

The NPT, negotiated among many people from different nations with different political objectives and constraints, is inevitably a document of compromises, laced with imprecise language, nuanced meaning, and cognitively dissonant passages. Depending on how the treaty is interpreted, it is either—as claimed—the core of the world’s nonproliferation regime, or it is a tool for proliferants to hide their ambitions and legitimize their activities.

In at least two main areas, an interpretation of the language of the NPT can strengthen the nonproliferation regime. The first involves Article 1’s requirement that each weapon state that is party to the NPT in no way assist a state without nuclear weapons in the manufacture of nuclear explosive devices.

As Eldon Greenberg has pointed out, the negotiating history of the NPT does not permit one to conclude that the application of safeguards to a nuclear transfer means that the transfer is legitimate.15 (Transfer of the components of an explosive device is prohibited, even if safeguards are attached.) Moreover, the very real possibility that an NPT party may be a proliferator in disguise makes it incumbent upon suppliers to judge the ultimate use of exported technology and equipment. Such judgments could take into account the economic and technical need for the exported items. Accordingly, it is at least arguable that the transfer of reprocessing equipment or technology to a nonweapon state constitutes prohibited assistance under Article 1, because such technology cannot be effectively safeguarded and exhibits no compelling economic need anywhere in the world.

Article 1’s language prohibiting indirect assistance by a weapon state may also be interpreted as prohibiting nuclear assistance of any kind by weapon states to nonweapon states not party to the NPT, on the grounds that such assistance releases resources by those states that may be used in unsafeguarded nuclear programs—perhaps devoted in part to weapons development.

Some Flaws in the Treaty
That Ought to Be Fixed

1. The NPT does not forbid a nonweapon state from possessing nuclear weapons. (It forbids the acquisition, but—in theory—a country with weapons could sign the NPT as a nonweapon state and not give up weapons already made.)

2. Nothing in the treaty prohibits a nonweapon state that is party to the treaty from assisting another nonweapon state in manufacturing or otherwise acquiring the bomb.

3. The treaty should be clarified to prohibit challenges to the notion that safeguards include the ability to search for nonnuclear activities relevant to bomb making—including R&D. To ensure that this doesn’t convert the IAEA into a university on weapons design, only inspectors from current or former weapon states should be involved in this activity.

4. The treaty does not require the IAEA to verify the obligation of a nonweapon state to refrain from receiving assis- tance in the manufacture or acquisition of nuclear weapons.

5. The treaty does not require the IAEA to verify that exports of nuclear hardware by NPT suppliers to nonweapon states are carrying safeguards.

6. The treaty does not define the point at which one can say that construction of a nuclear explosive device has begun. The Foster criterion relating “manufacture” to construction of a component having relevance only to a nuclear explosive device could constitute such a definition. In that case, activities involving machines capable of creating such components could become subject to special inspections.

7. The treaty does not prohibit a nonweapon state from using nuclear energy for military purposes but is unclear as to permitted “military uses” that are exempt from safeguards. In his recent book, David Fischer poses questions as to whether a nonweapon state could build a reactor and claim it to be the prototype of a naval reactor, thereby exempting its fuel from safeguards.16 Likewise, a state could withhold material from safeguards upon becoming an NPT party by claiming (to itself—it has no obligation to inform the IAEA) that the material is for a permitted military purpose. Finally, the treaty not only appears to waive the safeguarding of a “military” enrichment plant whose output is only for naval reactors, but also appears to allow unsafeguarded nuclear exports for permitted military use.

8. The treaty’s language in Article 3.3 has been used to support arguments against making safeguards more intrusive. The treaty should state as a principle that whenever a conflict occurs between the effective application of safeguards and compliance with Article 4, resolution in favor of effective safeguards shall govern.

9. The treaty does not embargo transfers of sensitive equipment, materials, or technology—but it should do so whenever effective safeguards do not apply.

10. The treaty has no sanctions for violators or for withdrawal from the treaty.

11. The treaty is difficult to amend; even worse, only those parties ratifying the amendment are subject to it.

12. The treaty does not preclude possession and stockpiling of plutonium or highly enriched uranium by a nonweapon state, regardless of economic or technical justification or the effectiveness of safeguards.

13. The treaty does not preclude nuclear trade with states not party to the NPT.

14. The treaty’s provision on withdrawal does not provide for any disposition of nuclear assets or payment for nuclear assistance received by the withdrawing state by virtue of its NPT membership.

What Should Be Our Level of Reliance
on the NPT as a Security Measure?

As stated at the outset, the NPT unquestionably has been a valuable institution. It has helped create a nonproliferation ethic that has raised the political barriers—at least in democratic states—to overt proliferation. It has played a useful role as an anchor or central element in all the discussions about security with the newly independent states and other states in Eastern Europe. By providing an outlet for US/Soviet cooperation during the days of the cold war that made it more difficult for each side to demonize the other, the treaty lowered the risk of war. It has provided an outlet for countries desiring to play a role on the world stage in disarmament to do so without becoming weapon states themselves. It provided a way for South Africa to give up its weapons program with a minimum of lingering doubt and suspicion because of IAEA verification, and it provided a basis for dealing with the North Korean weapons program.

On the other hand, the NPT also has been a convenient political cover for countries known to be interested in acquiring nuclear weapons, has played no essential role in turning around the past South Korean and Taiwanese clandestine-weapons programs, produced no appropriate response to Iraq’s weapons program until after Saddam Hussein invaded Kuwait and was militarily defeated, and provides no restraint on the stockpiling of weapons materials by any state as long as the materials are under safeguards.

Since many of its adherents joined because of the promise of technical assistance and technology transfer, the treaty does not incorporate any restrictions on nuclear trade, leaving only the suppliers to decide what should or should not be transferred.

In the end, the ability to leave the treaty with 90 days’ notice means that a country may proceed to build nuclear weapons, as long as it has the technological know-how and is convinced that such weapons are its best option for enhancing its security.

Even if the treaty and the safeguards system had been originally constructed with the needed reforms discussed in this essay, its implementation would still ultimately depend on the resolve of the international community, acting through the Board of Governors of the IAEA (which occasionally has a proliferator as chair) and the UN Security Council.

Nonetheless, the warts exhibited by the treaty and its still-evolving safeguards system do not vitiate the political value of the nonproliferation norm that has been nurtured by the treaty and the rest of the nonproliferation regime: the zones free of nuclear weapons, the Tlatelolco and Rarotonga Treaties, the export control laws and agreements (both multilateral and unilateral), and other instruments.

In sum, the treaty cannot be a substitute for measures one might otherwise take in protecting one’s security. Without reform, it does not provide a good model for dealing with pro- liferation threats other than nuclear (e.g., chemical, biological, or missile), but it is an important adjunct whose absence would raise current anxiety levels about the spread of weapons of mass destruction.


1. “Treaty on the Nonproliferation of Nuclear Weapons,” opened for signature 1 July 1968, 21 UST 483, 729 UNTS 161.

2. George Bunn and Roland M. Timerbaev, Nuclear Verification under the NPT, Program for Promoting Nuclear Nonproliferation (PPNN) Study no. 5 (University of Southampton, England: Mountbatten Centre for International Studies, 1994), 4–5.

3. Senate Committee on Foreign Relations, Remarks Submitted by William C. Foster, 89th Cong., 2d sess., 10 July 1968.

4. “Arms Control Implications of Peaceful Nuclear Explosions,” working paper submitted to Geneva disarmament conference by Japan, CCD/454, 7 July 1975 (Arms Control and Disarmament Agency [ACDA] Documents on Disarmament, 1975).

5. Bunn and Timerbaev, 4.

6. Ibid., 11.

7. See Nuclear Export Guidelines Adopted by 15 Governments, January 11, 1978, IAEA document INFCIRC/254 (Vienna, February 1978).

8. “Interview with Rudolph Rometsch,” IAEA Bulletin 36, no.3 (1994): 14.

9. See “Nuclear Nonproliferation and Safety: Challenges Facing the International Atomic Energy Agency,” Report GAO/NSIAD/RCED-93-284 (Washington, D.C.: General Accounting Office, September 1993).

10. Lawrence Scheinman, “Assuring the Nuclear Non-Proliferation Safeguards System” (Washington, D.C.: Atlantic Council, October 1992), 26.

11. Marvin Miller, “Are IAEA Safeguards on Plutonium Bulk-Handling Facilities Effective?” (Washington, D.C.: Nuclear Control Institute, August 1990).

12. Leonard Weiss, “The Concept of Timely Warning in the Nuclear Nonproliferation Act of 1978,” Congressional Record, 21 March 1988, S2639 and S2646; see also Senate Foreign Relations Committee, Testimony of Sen John Glenn, appendix, 15 December 1987, in Nuclear Nonproliferation Factbook (Washington, D.C.: Library of Congress, Congressional Research Service, December 1994), 155–78.

13. As quoted in the Wall Street Journal, 17 July 1981.

14. Ibid.

15. Eldon Greenberg, “The NPT and Plutonium” (Washington, D.C.: Nuclear Control Institute, May 1993).

16. David Fischer, Towards 1995: The Prospects for Ending the Proliferation of Nuclear Weapons (Brookfield, Vt.: Dartmouth Publishing Co., 1993), 61.