Governance of Dual-Use Technologies: Theory and Practice

Chapter 1: On the Regulation of Dual-Use Nuclear Technology

Back to table of contents
Elisa D. Harris, James M. Acton, and Herbert Lin
Global Nuclear Future

James M. Acton


When General Electric acquired the rights to Silex laser enrichment technology in 2006, few people, even within the nuclear industry, took much notice. After decades of research, laser-based technologies appeared to be yet another in a long list of enrichment processes that were not commercially viable. In 1999, the United States Enrichment Corporation (USEC), a private company originally created by the government to lead domestic enrichment efforts, had abandoned the U.S. indigenous laser enrichment program after twenty-seven years and more than $2 billion had been spent.1 At the time, USEC had a partnership agreement with Silex Systems, the Australian company that had developed its eponymous laser-based process. But in 2003, USEC decided the technology was too expensive to commercialize and turned its attention to the gas centrifuge, which then, as now, dominated the global market for enrichment. Silex looked dead in the water.

General Electric, which merged its nuclear operations with Hitachi in 2007 to form GE Hitachi Nuclear Energy, had more success with Silex, however, and proceeded with plans to commercialize it, albeit behind schedule.2 In 2009, it submitted a license application to the U.S. Nuclear Regulatory Commission (NRC) for a commercial-scale facility. GE Hitachi indicated that a decision to build the facility would not be made until the results of further testing were known.3 But the prospect that the United States might license a first-of-its-kind laser enrichment facility marked an important potential juncture for both the nuclear industry and the nuclear nonproliferation regime.

The benefit of the technology was clear. If laser enrichment were cheaper than the centrifuge, GE Hitachi’s investment in a risky technology would reap considerable profits. (Consumer benefits, if any, would be much more modest, because the price of enrichment is set by the most expensive supplier on the market and because enrichment typically accounts for no more than 5 percent of the total cost of electricity.)4

The costs and risks of laser enrichment—including proliferation—were potentially significant but extremely hard to evaluate.5 One key issue was how the commercialization of laser enrichment by a U.S. company would affect its spread globally. Would the demonstration that laser enrichment was a profitable enterprise inspire other states to attempt to develop it for themselves? Could GE Hitachi keep classified details of the Silex process secret for decades, potentially against repeated attempts by foreign governments to acquire them? Even if it could, would other states or companies nonetheless succeed in developing this or another form of laser enrichment technology from scratch? Conversely, if the United States refused to license the plant, would Silex Systems attempt to transfer the technology to another state? If so, could the United States prevent the company from doing so?

A second issue was the consequences of the spread of laser enrichment technology. Did any technical barriers prevent Silex technology from being used to produce the highly enriched uranium (HEU) needed for a nuclear weapon?6 GE Hitachi had boasted that a Silex facility would be smaller and use less energy than a centrifuge facility—but did that not also imply that a clandestine laser enrichment facility would be more difficult to detect? How effective would International Atomic Energy Agency (IAEA) safeguards be if applied to a laser enrichment facility?

To this author at least, the outcome of a cost-benefit analysis of laser enrichment technology was not obvious. None of the important questions was straightforward. Answers to some of them required classified information, which was secret for good reason. Others hinged on difficult-to-make political judgments. A net assessment of Silex technology would have been a difficult and controversial exercise. The U.S. government, however, did not even try. The executive branch ignored the issue. The NRC argued that its nonproliferation role extended no further than overseeing GE Hitachi’s procedures for handling classified information; everything else was, in the NRC’s opinion, the executive branch’s responsibility.7 While a few lawmakers did take an interest in the subject, Congress as a whole did not.

In September 2012, the NRC licensed the facility. In the three years the process had taken, nonproliferation considerations had essentially been ignored. Remarkably, the only entity with detailed knowledge of the Silex process that had attempted to analyze those considerations was GE Hitachi itself—and its assessment, which was not made public, was reportedly just seven pages long, three of which were the biographies of the authors.8

More than three years later, GE Hitachi has still not decided whether to build the facility. Nonetheless, its license application highlights issues—substantive and procedural—that reoccur in domestic decisions about the development and deployment of nuclear technologies. The story also illustrates the role of the global nuclear nonproliferation regime—IAEA safeguards most notably—in trying to ensure that the nuclear technology that does spread is used only for peaceful purposes.

Scoping the Dual-Use Problem

Which technologies and materials are dual-use; that is, useful for both civilian and military ends? Separated plutonium and HEU are the two dual-use nuclear materials of greatest proliferation significance. Both have a few nonmilitary applications, most notably in reactor fuel, but can also be used as the fuel for a nuclear weapon. HEU is usually defined as uranium “enriched” to contain more than 20 percent of the isotope uranium-235 (natural uranium, by contrast, consists of 99.3 percent uranium-238 and only 0.7 percent uranium-235). Although uranium enriched to any level above 20 percent is capable of sustaining the uncontrolled chain reaction used to generate energy in a nuclear weapon, practical warhead-making considerations dictate the use of HEU with an enrichment level of at least 80 percent. Whether the plutonium contained in the spent fuel discharged from modern power reactors is suitable for weaponization has been the subject of considerable debate (its isotopic composition is quite different from plutonium produced specifically for weapons).9 Nonetheless, the IAEA treats almost all plutonium as weapon-usable.10 It regards twenty-five kilograms of uranium-235 in the form of HEU or eight kilograms of plutonium as the “the approximate amount of nuclear material for which the possibility of manufacturing a nuclear explosive device cannot be excluded.”11 This definition has been criticized by experts who argue that even a relatively unsophisticated proliferator could design its first nuclear weapon with less material (a nation with experience in weapon design can certainly manufacture a warhead with less material, but the agency’s detection efforts are not focused on such states).12

Enrichment (concentrating uranium in the isotope uranium-235) and reprocessing (separating plutonium from spent fuel) are the most sensitive nuclear technologies, since they can produce weapon-grade materials. But enrichment is also vital to nuclear energy, since most power reactors in operation today use fuel made from low enriched uranium (LEU, which has a uranium-235 content between that of natural uranium and HEU). By contrast, most states, with some notable exceptions—including France, Japan, Russia, and the United Kingdom—have chosen not to reprocess the spent fuel produced by their nuclear power programs.

Nuclear technologies other than enrichment and reprocessing (such as nuclear reactors) and nuclear materials other than plutonium and HEU (such as spent fuel and LEU) are less proliferation sensitive but are still dual-use because they can be involved in the production of weapon-grade materials.

Some materials and technologies that do not involve uranium or plutonium are also of proliferation concern. First and most important, nuclear facilities involve numerous nonnuclear components. Those that also have nonnuclear applications are often described, particularly in the world of export controls, as “dual-use.” This usage of the term is subtly but confusingly different from describing plutonium as dual-use because it has both civilian and military applications. In any case, certain types of pressure transducers, which can be used to monitor the operation of a gas-centrifuge enrichment plant and serve a similar function in other industrial processes, are but one example of equipment that has both nuclear and nonnuclear uses. Similarly, the nonnuclear technology used in nuclear warhead manufacture can also be dual-use, even if its nonnuclear applications are often military. Technology for “shaping” high explosives, for example, is used in producing both antitank munitions and the explosives that are used to detonate a nuclear weapon.

Second, computer codes and other theoretical tools can have both military and civilian nuclear applications. For example, computer codes used to model the core of nuclear reactors or the behavior of plasma or the transport of radiation under certain conditions may be adaptable for use in nuclear-weapon studies. Finally, fissile materials other than HEU or plutonium could also be used to manufacture a nuclear weapon. The most important such material is uranium-233. The United States has tested nuclear warheads incorporating this material and has a stockpile of it (which it is currently trying to dispose of).13 Furthermore, the development of thorium breeder reactors, which India is pursuing, could lead to the large-scale production and separation of uranium-233. Other alternative fissile materials (most notably, neptunium-237) are of lesser concern since they have not been (and currently appear unlikely to be) separated on a significant scale—although this could change.

Efforts to manage these technologies and materials are generally divided into nuclear nonproliferation and nuclear security. These terms have no universally accepted definitions.14 For the purposes of this chapter, nonproliferation is used to refer to the goal of preventing states that do not have nuclear weapons from acquiring them. The aim of nuclear security efforts is to prevent the unauthorized possession of nuclear material or access to nuclear facilities under civilian control. In practice, most states’ nuclear security efforts are directed primarily against nonstate actors, although state-based threats to nuclear security cannot be ignored.

These definitions reflect the focus of this chapter—the oversight of dual-use technology. In other contexts, nuclear security is often defined in a broader sense to include, for example, preventing unauthorized access to nuclear weapons or material under military control. This is clearly an important goal, but it falls outside the scope of this chapter, since nuclear weapons and military fissile materials are not dual-use. Likewise, preventing unauthorized access to radioactive but nonnuclear materials (which is sometimes considered to be part of nuclear security) and preventing the unintended release of radiation from nuclear facilities (which is always considered within nuclear safety) are also important. However, both goals have significant differences—in terms of the potential approaches to risk mitigation, the challenges to their implementation, and the consequences of failure—from nuclear security, as defined above. Finally, many analysts and officials, particularly from the Global South, would argue that nuclear nonproliferation should include not just preventing the spread of nuclear weapons to new states (sometimes termed “horizontal proliferation”) but also preventing those states that have nuclear weapons from qualitatively or quantitatively enhancing their arsenals (“vertical proliferation”). While a narrower definition has been adopted here for the sake of conceptual clarity, the extent to which enhancing nonproliferation efforts is made more difficult by a perceived lack of progress on disarmament is discussed below.

A Conceptual Overview of Regulation

At the beginning of 1942, the practical applications of the newly emerging field of nuclear science still lay in the future. However, the promise of one potential application—the Bomb—had just sparked the creation of what came to be called the Manhattan Project. Within two years, the United States would become the first nation to operate a nuclear reactor, to separate plutonium from spent fuel, and to enrich uranium. Only after the war were these technologies used for more peaceful ends. From a historical perspective, therefore, nuclear technology is not a civilian enterprise that happens to have military applications, but a military technology that also has peaceful uses.

To manage the risks associated with this technology, an extraordinarily complex system has emerged: legally binding and politically binding agreements; norms; patterns of interstate cooperation; intergovernmental, nongovernmental, and domestic institutions; and national laws and practices (along with a bewildering number of acronyms). In fact, two largely separate systems have emerged: one for nonproliferation and one for nuclear security.

In the broadest of terms, the nuclear nonproliferation regime consists of three key interrelated elements:

  • National decisions about whether to develop or use a particular dual-use nuclear technology.
  • National laws and international agreements about whether to permit the sale of dual-use nuclear technologies and materials and, if so, under what conditions.
  • International oversight mechanisms to detect and deter attempts by states to acquire nuclear weapons using these technologies and materials.

The importance of these three elements stems from the characteristics of nuclear technology and the way it has been developed.

First, national governments have been—and remain—absolutely central to the development and operation of nuclear technology. States or state-owned companies operate many, if not most, of the nuclear facilities around the world today. Even nuclear facilities that are operated by genuinely private entities, such as some utilities in the United States, are dependent on governments. At the very least, a nuclear facility must be licensed. While the primary function of licensing is ensuring safety, the process requires a government to have made a policy decision in favor of permitting (or at least not prohibiting) the activity in question. In most cases, however, governments do much more than merely tolerate a nuclear activity. For example, the development of American light water reactor technology (which has now been incorporated into Chinese, French, Japanese, and South Korean designs) was originally sponsored by the U.S. government for use in submarine propulsion. Meanwhile, efforts to encourage the construction of nuclear power reactors invariably require government intervention, even in states without centrally planned economies. The United Kingdom, for example, has guaranteed the price of electricity generated by new nuclear reactors, whereas the United States offers loan guarantees to subsidize their construction. Silex laser enrichment technology is another case in point. Although efforts to develop it appear to have started as a purely private enterprise, commercialization has involved a U.S.-Australian treaty permitting its transfer to the United States (which was seeking, as a matter of national policy, to acquire an alternative enrichment process to gaseous diffusion), followed by financial assistance from USEC.

There are, of course, some exceptions—particularly where less sensitive activities are concerned. Some suppliers of nonnuclear components for nuclear facilities are private companies that did not benefit from government assistance. But, overall, governments occupy a central place in the process of developing and deploying nuclear technologies.

As a result of the centrality of governments, their internal decision-making processes about acquiring dual-use nuclear technologies assume tremendous importance in influencing the prospects for managing such technologies. To be sure, if a government wants nuclear technology because it is embarking on a campaign to develop nuclear weapons, internal oversight mechanisms are unlikely to prevent it from proliferating. However, internal processes should be important in helping a state ascertain how its domestic programs might affect the behavior of others and hence influence global proliferation dynamics—although, in practice, as the licensing process for GE Hitachi’s laser enrichment plant exemplifies, states generally fail to capitalize on this opportunity because they do not have such procedures in place.

A second important characteristic of nuclear technology is that its spread—so far—has been relatively limited. About 110 states are estimated to have some capability in the manufacture of the dual-use, nonnuclear components used in nuclear facilities.15 Seventy-one states and Taiwan conduct what the IAEA terms “significant nuclear activities,” which means that they possess more than a certain amount of nuclear material.16 Meanwhile, just nine states operate both enrichment plants and reprocessing plants, five operate enrichment plants (but do not reprocess), and one operates a reprocessing plant (but does not enrich).17 Moreover, the group of intellectual property holders is more limited still. Enrichment plants in the United States and France both use centrifuges designed by the Anglo-Dutch-German consortium Urenco, supplied under a “black box” arrangement that prevents the operator from learning classified design details. France supplied Japan with an industrial-scale reprocessing plant and is considering whether to supply one to China. The international market for nuclear power reactors currently consists of vendors from just seven countries (Canada, China, France, Japan, Russia, South Korea, and the United States).18

Nuclear technology may well diffuse much more widely in the future. Nonetheless, the small number of technology holders is an important feature of today’s world. In consequence, nuclear trade between the “haves” and “have-nots” has an important influence on proliferation dynamics. Therefore, decisions by states about whether, and under what conditions, to trade in nuclear technologies and materials, as well as various international export control arrangements, constitute a second important element in the nuclear nonproliferation regime.

Of course, trade is not the only way a state can acquire nuclear technology. Information can be stolen. Three of the states that operate enrichment plants today—Iran, Pakistan, and North Korea—use centrifuges based on technology illicitly acquired from Urenco, and Brazil and India may do so too.19 States can also develop technology indigenously—and an increasing number are likely to have the capability to do so in the future.20 The regulation of trade nonetheless continues to play an important role in nonproliferation efforts because the financial and technical barriers to entering the nuclear technology field are still relatively high. Developing the technology to enrich uranium on a meaningful scale, whether accomplished entirely indigenously or by using illicitly acquired information and components, is likely to cost at least a few hundred million dollars.21 Where reactor technology is concerned, the incentives for legal trade include not simply reduced costs but huge advantages in safety and reliability.

Against this background, advanced nuclear nations have adopted three basic strategies toward the development and trade of new nuclear technologies:22

  • Develop and deny. A state chooses to develop a technology but refuses to transfer it. The United States applied this strategy to all nuclear technology between the passage of the MacMahon Act in 1946 and the passage of Atomic Energy Act of 1954.
  • Develop and disseminate. A state develops a technology and then sells it subject to conditions (such as international monitoring) designed to prevent it from being used for proliferation. The United States has applied this strategy to light water reactors since they were first commercialized in the 1950s.
  • Desist and discourage. A state refrains from developing a specific sensitive nuclear technology and encourages others to adopt similar restraint. The United States has adopted this strategy with regard to reprocessing for most of the period since 1976, when President Gerald Ford first announced a temporary moratorium on domestic reprocessing, which was extended indefinitely by President Jimmy Carter the following year.

As the examples illustrate, states do not necessarily apply the same strategy to different technologies. Moreover, these strategies are “ideal” types; in practice, the lines between them may be blurred. For example, although the United States has a policy against reprocessing, it nonetheless conducts basic research in this area. The Urenco states, meanwhile, have a general policy against selling enrichment technology but have made exceptions for the United States and France, while Russia, which has a similar basic policy, made an exception for China.

A third critical component of dual-use controls is the international oversight of domestic nuclear activities—known as safeguards—to detect and deter their use for military purposes. Safeguards, which in most cases are administered by the IAEA, are integral to the effectiveness of a develop-and-disseminate strategy but may also be needed under a develop-and-deny strategy if the developer has pledged not to develop nuclear weapons (centrifuge programs in Germany, the Netherlands, and Japan are examples of this latter case).

One characteristic of uranium or plutonium that facilitates safeguards is that their quantity (more properly, the mass of any given isotope) is conserved except in a nuclear reactor.23 For example, in an enrichment plant, the quantity of uranium-235 in the feedstock must be equal to the sum of the quantities in the product and waste streams. This property permits an oversight process termed “material accountancy” in which inspectors take periodic inventories of declared nuclear materials, while also verifying declared transfers into and out of a facility. In theory, any discrepancy between the amount of material that is present and the amount that ought to be present indicates that the state has diverted nuclear material. In practice, this process is complicated by both unavoidable uncertainties in measurement and the practical impossibility of verifying all declared nuclear materials in most states, thus forcing inspectors to rely on sampling techniques to verify only some fraction of it. As a result, inspectors can provide only statistical, and not absolute, confidence in nondiversion.24 For decades, material accountancy and IAEA safeguards were virtually synonymous. Although safeguards have now expanded significantly in scope, material accountancy still occupies a central role.

Neither material accountancy nor any of the other safeguards activities that now complement it can, however, physically prevent a state from using nuclear material or facilities in the development of nuclear weapons (only military force creates even the possibility of achieving that). Rather, the primary purpose of safeguards is deterrence by threatening to expose would-be proliferators as early as possible (as such, safeguards are often compared to burglar alarms rather than to door locks).25 Whether this strategy is likely to be successful depends on both the perceived probability of being caught and the expected consequences. For this reason, the capability and willingness of not only the international community as a whole but of individual members to impose costs on states that violate their nonproliferation obligations is integral to the ultimate effectiveness of international oversight efforts.

At a conceptual level (if not an operational level), nuclear security efforts are significantly simpler than nonproliferation for two basic reasons. First, much of the complexity of nuclear nonproliferation stems from efforts to allow states access to inherently sensitive technology but to restrict the purposes for which it can be employed—a problem that does not arise with nuclear security, where the goal is to deny unauthorized use entirely. Second, nonstate actors, the primary “target” of nuclear security efforts, are significantly less capable than states. Academics have extensively debated whether nonstate actors could build a viable nuclear weapon if they acquired sufficient weapon-usable nuclear material.26 However, most would agree that nonstate actors lack the capacity to manufacture fissile material themselves.

Accordingly, nuclear security is focused on preventing nonstate actors from acquiring nuclear material. (Keeping information or equipment that could be used to manufacture fissile material, such as components for enrichment facilities, out of the hands of nonstate actors is also important, although the aim here is largely to prevent such information and equipment from then being sold to states. As such, this tends to be classed as a nonproliferation measure.) Physical protection measures—to deny unauthorized access to nuclear material and facilities—are the most important nuclear security measures. However, because of the potential consequences of failures in physical security, best practice demands multiple layers of protection, an approach known as “defense in depth.” Other layers include efforts to detect the theft of material (such as material accountancy) and strategies for locating and recovering stolen material rapidly. Nuclear forensics, which can help determine the origin of recovered nuclear material, can expose security breaches and thus provides an incentive for states to secure nuclear materials properly. Deterrence may play a secondary role to prevention insofar as nonstate actors may be deterred from even attempting to acquire nuclear material if they believe failure is sufficiently likely. Threats to punish those involved in nuclear terrorism may also play some role, albeit probably only a marginal one (especially if the would-be terrorists are suicidal). Potential financers of a terrorist organization, for example, may have “something to lose” and so can perhaps be deterred from providing assistance by the threat of punishment.27 Preventing terrorist organizations from gaining access to nuclear material is, however, unquestionably preferable to relying on deterrence or attempting to recover stolen material.

International Regulation of Nuclear Technology

Nuclear nonproliferation and nuclear security are not only conceptually distinct; the international systems to promote them are largely separate. Almost all relevant international institutions and agreements are concerned with one or the other but not both. The IAEA is the exception that proves the rule, since its nonproliferation and nuclear security functions have a “firewall” between them. Inspectors, for example, may not officially report on any weaknesses in nuclear security they observe during safeguards inspections.28 Moreover, the two systems are structured very differently. The nonproliferation regime is a relatively comprehensive system based largely on legally binding agreements. By contrast, the nuclear security architecture—even the term “regime” seems inappropriate—is a patchwork of arrangements, most of which are not legally binding, containing many omissions. One important consequence of this difference is that domestic nonproliferation requirements tend to vary much less between states than domestic nuclear security requirements.

These structural differences reflect differences in where states believe the responsibility for nonproliferation and nuclear security should lie. For decades, preventing the spread of nuclear weapons has widely been seen as an international task. Even if states disagree intensely about how to go about this task, the very fact that it is viewed as a collective responsibility has facilitated the creation of a comprehensive and legally binding system. By contrast, nuclear security was originally seen as an exclusively domestic matter and even today is still largely seen in those terms. As a result, the nuclear security architecture has been accreted over time, resulting in a patchwork. One indication of the difference is that the United States and its partners are regularly criticized for failing to show sufficient deference to international institutions on nonproliferation, whereas they are criticized for being too heavy handed in interfering with sovereign affairs where nuclear security is concerned.

Why the international community takes such different approaches to nuclear nonproliferation and nuclear security is not at all obvious—or rather, it is not obvious why nuclear security is not regarded as a matter for a comprehensive and legally binding international oversight regime. After all, internationally, nuclear security is less controversial than nonproliferation, and all states would probably agree that it is a global public good (even if many developing states would also argue that the threat from nuclear terrorism has been exaggerated). The answer may be, in part, historical. The nonproliferation regime originated in the Cold War, and its creation was made possible by the existence of two superpowers and their willingness to cooperate, not least by strong-arming recalcitrant allies into pledging not to acquire nuclear weapons. At the time, nuclear proliferation was (rightly, in my opinion) viewed as a much greater threat than nuclear terrorism, which presumably explains why the superpowers chose to focus on it. Moreover, a natural quid pro quo could be made: under the terms of the 1968 Treaty on the Non-Proliferation of Nuclear Weapons (NPT), non-nuclear-weapon states (defined as those that had not exploded a nuclear weapon prior to January 1, 1967) promised not to acquire nuclear weapons in return for a commitment—by all states—to work in good faith toward nuclear disarmament. By contrast, a comprehensive nuclear security treaty would probably have required the United States and the Soviet Union to accept international oversight of their domestic nuclear activities, an outcome they would have found far more disagreeable than the NPT’s disarmament promise, which was described at the time by a U.S. official, who later became head of the Arms Control Association, as “an essentially hortatory statement” that “presents no problems.”29

Although concern about nuclear terrorism occasionally surfaced during the Cold War, not until the terrorist attacks of September 11, 2001 was the issue firmly placed on the international agenda. However, international consensus on the severity of the threat has been hard to come by. Many non-Western states worry much more about being subject to burdensome regulations and being denied access to nuclear technology than they do about nuclear terrorism.30 Moreover, the end of the bipolar international system and changes in U.S. domestic politics that make treaty ratification much more difficult have severely complicated the negotiation and implementation of international treaties. Collectively, these factors have militated against the “internationalization” of nuclear security.

The International Nuclear Nonproliferation Regime

The seeds of the nuclear nonproliferation regime were sown in 1953 when President Dwight Eisenhower announced in his “Atoms for Peace” speech at the United Nations (UN) that the United States was willing to share nuclear materials and technology (in the terminology of this article, he announced a switch from a develop-and-deny strategy to a develop-and-disseminate strategy), and proposed the creation of the International Atomic Energy Agency (IAEA). Eisenhower originally conceived the IAEA’s primary function as receiving and allocating military-origin fissile materials donated by nuclear-armed states—a task it did not end up fulfilling. But, the IAEA Statute (the treaty, concluded in 1956, that created the agency and governs its operations) entrusted it with applying safeguards to bilateral nuclear trade agreements at the request of the contracting parties—the origin of its safeguards role. Today, such “item-specific” safeguards (also known as INFCIRC/66 safeguards after the IAEA information circular setting out their terms) are in force for nuclear facilities acquired by trade in three of the states that never signed the NPT: India, Israel, and Pakistan.

Eisenhower’s approach marked a change of course for the United States, which seven years earlier had explicitly rejected the whole idea of safeguards. In 1946, the Acheson-Lilienthal report, which informed the first American proposal for the control of atomic energy, the Baruch Plan, had concluded that “there is no prospect of security against atomic warfare in a system of international agreements to outlaw such weapons controlled only by a system which relies on inspection and similar police-like methods.”31 The subsequent volte-face seems to have had much more to do with Cold War grand strategy than any fundamental reassessment of the report’s conclusions. Although extraordinarily prescient, the report had its weaknesses. Most notably, its authors focused only on the impossibility of using “inspection and similar police-like methods” to prevent proliferation and not on the possibility that they might be able to deter it—which was odd because the report did argue that deterrence would help enforce its preferred solution, an agreement to internationalize the fuel cycle.32

The conclusion of NPT in 1968 marked the next key moment in the development of the nuclear nonproliferation regime. To try to stem proliferation while allowing states access to dual-use nuclear technology, the NPT requires all non-nuclear-weapon states to accept IAEA safeguards on all their nuclear activities. To compensate for the inequality of the resulting two-tier system of nuclear-weapon states and non-nuclear-weapon states, the treaty also contains a disarmament commitment (in article VI) along with a separate commitment (in article IV) that requires “the fullest possible exchange” of nuclear materials, equipment, and knowledge between states.

IAEA safeguards, which were subsequently elaborated in INFCIRC/153, also known as the Model Comprehensive Safeguards Agreement, are primarily focused on detecting the diversion of nuclear material (plutonium, enriched uranium, or uranium-233) from declared facilities.33 To this end, states are required to submit comprehensive reports on their holdings of nuclear materials, which the IAEA then verifies through material accountancy. INFCIRC/153 places no limits on what nuclear activities states are permitted to conduct, so long as all nuclear materials are declared and inspected. (Most states would argue that the NPT’s only restriction on non-nuclear-weapon states is its prohibition against the “manufacture” of nuclear weapons—an injunction of unclear meaning—and that they have an “inalienable right” to conduct any other nuclear activity, however sensitive.34 The U.S. government and some U.S. scholars take a more restrictive view.35)

Although INFCIRC/153 requires states to safeguard all their nuclear activities—and hence gives the IAEA the legal right to investigate undeclared activities—it does not provide the agency with sufficient tools to draw credible conclusions in that regard, presumably because, at the time the document was drafted, fissile material production was believed to be such a large-scale enterprise that it could not be successfully hidden. This assumption was shown (rather spectacularly) to be false in 1991 when, following the Gulf War, Iraq’s clandestine nuclear weapons program was discovered.

In response the IAEA sought both to use its existing legal powers more fully (by, for example, making use of open-source information) and to develop an enhanced safeguards document with new legal powers, the Model Additional Protocol, INFCIRC/540.36 This document, which was endorsed by the IAEA’s Board of Governors in 1997, contains some “housekeeping” items, such as requiring states to provide inspectors with multiple-entry visas, but its key provisions are (1) expanded declarations by states, including about activities not involving nuclear material; and (2) much greater access rights for the IAEA, including outside of declared nuclear facilities and at short notice. In states with an additional protocol in force, the IAEA attempts to draw a “broader conclusion” that not only has no declared nuclear material been diverted but also that no undeclared material exists in the state. In 2014, 118 out of the 184 non-nuclear-weapon states party to the NPT had an additional protocol in force (as does Taiwan), and the broader conclusion had been drawn in sixty-five of them.37 Of the sixty-five without an additional protocol in force, eleven conduct significant nuclear activities.38 Two of these states, Iran and Syria, were in noncompliance with their safeguards agreements (though Iran has subsequently returned to compliance), while a number of the others, including Algeria and Egypt, are regarded as potential proliferators.

In addition to legal reform, a separate—and more controversial—wave of on-going organizational and conceptual reform is underway. Traditionally, the IAEA has looked at each facility in a state as an isolated entity and attempted to verify that no diversion from it has occurred. Then, in the early 2000s, the agency started to implement the “state-level approach,” in which it holistically examines “all nuclear material, nuclear installations and nuclear fuel cycle related activities” in a state to try to draw conclusions about the absence of undeclared activities.39 It initially focused on implementing this approach in those states with an additional protocol in force, but now does so in some states without one. In a closely related (some would say, inseparable) development, the IAEA has also moved to diversify the range of information sources available to it, under an approach termed “information-driven safeguards.” In addition to information supplied by states in declarations and obtained by inspectors, the agency now makes regular use of open-source information, commercial satellite imagery, and intelligence information supplied by member states. The IAEA has also sought to effect a change in organizational culture from an inspectorate of (nuclear material) accountants to one made up of detectives.

These changes proved controversial, although some aspects were simply the institutionalization of existing practice.40 The IAEA’s first publicly known use of intelligence information, for example, appears to have been in 1991 during its investigation of Iraq’s nuclear program, when the United States supplied photographs from a military reconnaissance satellite, thus also marking the IAEA’s first-known use of satellite imagery (albeit on a noncommercial basis). Open-source information was approved for use a few years later and was cited by the IAEA in its investigations into safeguards irregularities in South Korea in 2004 and Egypt in 2005. The agency also regularly referenced open-source information during its decade-long investigation into Iranian noncompliance.

In spite of this history, the debate over the state-level approach and information-driven safeguards has recently become much more acrimonious. Since 2012, Russia and other states, mostly from the Non-Aligned Movement (a grouping, originating in the Cold War, of states that were not aligned with either superpower), have started to question the state-level approach vociferously and to argue that it needs political approval from the IAEA Board of Governors.41 One worry is that, under the state-level approach, the choice of safeguards measures for a state depends on “state-specific factors” such as the “state’s legal framework for implementing safeguards obligations.”42 The agency argues that this approach allows “differentiation without discrimination,” whereas critics appear to worry that differentiation automatically constitutes discrimination.43 Moreover, given that the agency’s ultimate goal is to apply the state-level approach (or at least some elements of it) in all states, concerns have been raised, in particular by Brazil and Argentina, that the IAEA is attempting to force states that have not adopted an additional protocol into implementing some of its provisions.

Russia, meanwhile, has led the charge against the use of intelligence information. The agency’s reports on Iran incorporated such information, which was provided largely—if not exclusively—by a small number of Western states, to an unprecedented degree. Moscow worries that this practice could enable the suppliers—the United States, in particular—to deliberately mislead the agency and justify military action.44 The IAEA counters this objection by arguing that information supplied by member states is used only where it can be corroborated by other sources and represents just a small fraction of the total available information. However, given U.S. and British intelligence failures in the run-up to the Iraq War in 2003, the Russian argument has struck a chord internationally. Russian concerns also tie into a broader “fairness narrative” about the undemocratic way the agenda of international institutions is set by a small number of rich nations.

Western and other like-minded nations are not particularly satisfied with the status quo either, but their discontent stems from weaknesses in the safeguards system. In spite of the considerable improvement in safeguards since 1991, the intrinsically difficult task of detecting undeclared facilities—particularly small gas-centrifuge enrichment plants—almost certainly remains the agency’s biggest technical challenge (although material accountancy in large, bulk handling facilities is also difficult). While no silver bullet has been found for this or any other problem, plenty of ideas for improving the effectiveness of safeguards have been proposed. Few of them have been implemented recently, however, because most meaningful improvements to safeguards require the approval of the thirty-five-member IAEA Board of Governors, which is rarely forthcoming. To give but one example, the Model Additional Protocol requires adherents to report on exports and imports of specified types of equipment and nonnuclear materials. To facilitate updates, the list setting out which transfers must be reported was included in an annex to the Model Additional Protocol that can be amended by the Board of Governors and without the consent of every signatory. To date, however, the board has not even considered proposals to update the annex.

An almost inevitable corollary to the nuclear nonproliferation regime’s greatest strengths—its legally binding nature and near universality—is that reform is slow and difficult, since permission from so many participants is required. States opposed to reform offer a variety of objections. Some are technical, such as a desire to protect commercially sensitive information or the concern that enhanced safeguards might interfere with the smooth running of nuclear facilities. However, the most commonly heard objection, offered in particular by states in the Non-Aligned Movement, is that nuclear-weapon states, by not disarming, have failed to live up to their side of the NPT bargain and that, until they do disarm, a focus on further enhancing the nonproliferation regime is unfair.

That said, for all the challenges facing the international nonproliferation regime today, IAEA safeguards still represent a remarkable innovation. The use of intrusive international inspections at highly sensitive facilities to monitor dual-use technology was unprecedented. In fact, the very idea of safeguards runs counter to traditional notions of Westphalian sovereignty and was dismissed on realist grounds by the otherwise decidedly idealistic framers of the Acheson-Lilienthal report. Even today, only one directly comparable arrangement is in force: inspections and monitoring pursuant to the 1993 Chemical Weapons Convention to ensure that chemical production facilities are not used to produce prohibited agents. (Onsite inspections have also been facilitated by various arms limitation treaties—including those that are currently taking place under the terms of the New Strategic Arms Reduction Treaty, or New START—but the goal there is to verify limits on military capabilities, not to detect the employment of dual-use facilities for military ends.)

Overall, it is difficult to argue that IAEA safeguards have not played an important role in ensuring that the spread of nuclear weapons has been much slower than the spread in nuclear technology. To be sure, the IAEA’s failure to detect Iraq’s clandestine nuclear program prior to 1991 was a profound embarrassment—but, to be fair, the IAEA never claimed the inspection authority it had at the time would allow it to uncover such a program. Moreover, as the Acheson-Lilienthal report predicted, states have sometimes denied access to inspectors, lied to them, or otherwise inhibited their operation. However, poor cooperation in Iran, North Korea, and Syria did not prevent inspectors from presenting enough evidence that these states had violated their safeguards agreements so that the Board of Governors could make a formal finding of noncompliance.

Each of those findings provided meaningful warning of a real proliferation risk—even if the international community subsequently failed to use this warning effectively. Although the agency’s investigation into Syria began only after its plutonium-production reactor had been destroyed by Israel in 2007, it did highlight the risk that other undeclared facilities might remain.45 North Korea was first found in noncompliance in 1994, and Iran in 2005, before either had developed a nuclear weapon. None of these cases has been satisfactorily resolved. Syria has still not provided access to the suspect facilities, and North Korea tested its first nuclear weapon in 2006. More positively, an agreement with Iran, the Joint Comprehensive Plan of Action, commonly known as the “Iran nuclear deal,” was reached in July 2015 and entered into force in October 2015. Whether this agreement will prove successful is not yet clear (though, as of this writing in March 2016, implementation has proceeded smoothly).

The deal requires Iran to accept stringent restrictions on its nuclear program, most of which last for ten or fifteen years, as well as enhanced IAEA safeguards, in return for sanctions relief. The restrictions include a requirement to redesign the partially built heavy water reactor at Arak in order to curtail its ability to produce weapon-grade plutonium; limits on centrifuge numbers and types; limits on stockpiles of LEU; and limits on the level to which Iran can enrich uranium. At least some of the provisions represent genuine firsts in the management of dual-use nuclear technology, even if they have attracted less attention. For example, the deal extends IAEA monitoring, for the first time, to uranium mines and to the production and storage of centrifuge components. It obliges Iran to seek approval before importing specified dual-use equipment and materials. And it prohibits Iran from conducting specified activities that would be useful for designing or manufacturing a nuclear warhead. In all, this agreement offers a genuine prospect of resolving the Iran nuclear standoff. Full implementation will, however, take decades and is likely to prove difficult, not least because of potential domestic spoilers in both Tehran and Washington.

The failure to respond effectively to North Korean and Syrian noncompliance and the time required to find a credible pathway to resolving the Iranian crisis suggest that the biggest challenge facing the nonproliferation regime is not detecting violations but responding to them—or, as Fred Iklé once stated, “after detection—what?”46 Even agreeing that a state has violated its nonproliferation commitments can be highly controversial. On at least two occasions, states that committed significant safeguards violations were “let off” without even a noncompliance finding: South Korea in 2004 and Egypt in 2005.47 In the former case, which was more serious, the United States worked hard to shield its close ally. Initially at least, even finding Iran in noncompliance was met with opposition. The IAEA had collected sufficient evidence by November 2003 to merit a noncompliance finding, but the process was delayed for two years because of concerns that it would disrupt negotiations and provide a justification for U.S. military action.

The NPT, the IAEA Statute, and the various safeguards agreements do not give any indication about how to respond to cases of noncompliance beyond stating that the IAEA Board of Governors should refer them to the UN Security Council, which achieved its greatest nonproliferation success following the 1991 Gulf War when it authorized and subsequently provided political backing for an exceptionally intrusive investigation into Iraq’s nuclear, chemical, and biological weapons programs. Since then, however, the Security Council has been largely ineffectual, as the North Korean and Syrian cases demonstrate—although the Iranian case gives more cause for optimism. A discussion of what the Security Council could have done differently to facilitate resolution of these “hard cases” is beyond the scope of this paper. However, few would disagree that its effectiveness is currently limited by the inability of its members—particularly the veto-wielding states—to agree quickly on whether a case of noncompliance is a real problem and, if it is, what they should do about it.

Strategic trade or export controls are an important complement to IAEA safeguards. In fact, the safeguards system itself has some provisions related to international trade. INFCIRC/153 requires reporting of imports or exports of nuclear material, and the Model Additional Protocol requires states to inform the IAEA about transfers of certain types of equipment and nonnuclear material. These requirements are, however, not really controls, as they do not impose actual limitations.

Two multilateral arrangements—the Zangger Committee and the Nuclear Suppliers Group—go further and place some restrictions on transfers. Article III.2 of the NPT prohibits international transfers of nuclear material or “equipment or material especially designed or prepared for the processing, use or production of” nuclear material, unless safeguards are applied. To facilitate the implementation of this article, the Zangger Committee, an informal coalition that first met in 1971 and currently consists of thirty-eight states, has produced a “trigger” list of materials and equipment that participants agree the article covers.48 Participating states agree that the supply of trigger list items should be subject to a set of minimal conditions that reflect the NPT’s requirements, such as the application of IAEA safeguards and a pledge from the recipient not to use the material or equipment for the manufacture of nuclear weapons.

The Nuclear Suppliers Group, which was formed after India’s “peaceful nuclear explosion” in 1974 and today has forty-six participating governments, is more ambitious than the Zangger Committee in two respects. First, it publishes not only its own trigger list but a separate list of dual-use items (equipment that has both nuclear and nonnuclear applications).49 Second, the group’s guidelines impose somewhat more stringent conditions on transfers than the Zangger Committee. For example, the former requires recipients (except for nuclear-weapon states) to apply appropriate physical protection measures and to have safeguards on all their nuclear activities—although the United States persuaded the group to waive this latter requirement for India following the U.S.-India nuclear deal. Participating governments have also agreed to consider recipients’ nuclear nonproliferation credentials before agreeing to transfers, and to exercise a “policy of restraint” toward transfers of enrichment and reprocessing technology.

Strategic trade controls, like others facets of the international nonproliferation regime, were revitalized following the discovery of Iraq’s clandestine nuclear program in 1991 and, in particular, by the embarrassing revelation that it had sourced components for its centrifuges from Western companies. The Nuclear Suppliers Group, which had not met since 1978, subsequently became much more active and progressively tightened its guidelines. Its efforts, coupled to improved national export control programs, have almost certainly helped to slow the clandestine nuclear programs that have been discovered more recently. Moreover, since the UN Security Council passed Resolution 1540 in 2004, all states have been legally bound to implement export controls. Today, however, strategic trade controls face both technical and political challenges.50

A number of technical developments are conspiring to undermine the (already limited) effectiveness of trade controls and to complicate their implementation. Trade patterns are becoming more complex and increasingly involve middlemen, brokers, and transshipments, making them more difficult to monitor. An increasing number of states outside the Nuclear Suppliers Group are also involved (an estimated 110 states can now manufacture items on the group’s dual-use list). Meanwhile, illicit trade has been facilitated by nonstate actors, most notably the network run by the Pakistani scientist A. Q. Kahn, which is believed to have been only partially dismantled. Perhaps most fundamentally, as Scott Kemp argues, would-be proliferators are increasingly able to get by without foreign assistance since “the technologies needed to make nuclear weapons have remained static, whereas the indigenous capabilities of states have steadily grown over the last half century.”51

The Nuclear Suppliers Group, in particular, also faces political challenges from within and from the outside. Its rules are nonbinding and have been violated, most notably by China in two recent agreements to supply Pakistan with reactors.52 Decision-making is slow due to a need for consensus. Reaching agreement on tightening the rules for transfers of enrichment and reprocessing technology took more than seven years (and even then led to a distinctly unsatisfactory outcome).53 Today, the Nuclear Suppliers Group has begun what promises to be a long and acrimonious process to decide on whether to admit India. Finally, what Mark Hibbs terms “the rise of international nuclear equity issues” has led to criticism of the group by non-nuclear-weapon states worried about becoming victims of export denials—a matter of importance not least because, while most of these states do not have nuclear power programs, some can manufacture sensitive components.54

The Nuclear Suppliers Group and the Zangger Committee might be described as “coalitions of the willing” that are multilateral but not truly international and impose politically but not legally binding rules on participants. Other such nonproliferation initiatives include the Proliferation Security Initiative and the Nuclear Power Plant Exporters’ Principles of Conduct.

The Proliferation Security Initiative, which was launched in 2003 by the United States in cooperation with ten close allies, “aims to stop trafficking of weapons of mass destruction . . . their delivery systems, and related materials to and from states and non-state actors of proliferation concern” by interdicting dangerous cargoes in transit.55 Initially, it proved highly controversial, not least because many states appeared to have the impression (contrary to what was actually written in the “Statement of Interdiction Principles”) that it would involve illegally boarding ships on the high seas. Concerns have, however, gradually abated as, in practice, the initiative has focused on capacity building (including through exercises) and streamlining procedures for sharing information to help states enforce existing laws more effectively. Meanwhile, the United States has negotiated bilateral boarding agreements with eleven flag states in which over 55 percent of the world’s shipping (by tonnage) is registered.56 A further 22 percent is registered with other initiative participants. Today, 102 states have endorsed the “Interdiction Principles,” which is the only requirement for “membership”— although how many of these states participate actively is unclear.57 Indeed, the effectiveness of the initiative is hard to assess, not least because participating states reveal little information about its activities.58

The Nuclear Power Plant Exporters’ Principles of Conduct is a corporate and social responsibility code for nuclear reactor exporters and has been adopted by all the world’s major vendors outside of China.59 The Principles of Conduct are the most important example of a code of ethics for nuclear nonproliferation, albeit among corporations rather than individuals. They also cover more than nonproliferation; indeed, their most significant provisions relate to other issues, most notably safety. In large part, participants’ nonproliferation commitments extend no further than abiding by existing national and international rules and assisting reactor recipients to meet their obligations, although vendors also commit, for example, to “promote proliferation-resistant [reactor] designs” and to inform national authorities “in a timely manner . . . of any serious nonproliferation concerns.”60 The Principles of Conduct illustrate a general difficulty of codes of conduct in the nuclear field: because there is so much national and international nonproliferation regulation, individuals and organizations tend to see their ethical responsibility purely in terms of complying with existing rules.

International Nuclear Security Efforts

Although promoting nuclear security, alongside safety and safeguards would have been a natural fit for the IAEA, the subject was not discussed during negotiations over the agency’s founding and is not mentioned in its statute.61 Not until 1972 did the agency carve out a role for itself in nuclear security by issuing—against initial opposition from a number of Western states—a set of nonbinding guidelines on nuclear security, INFCIRC/225, which has since been revised five times.62 Later in the decade, the agency took a leading role in preparing the first binding treaty on nuclear security, the 1980 Convention on the Physical Protection of Nuclear Material (CPPNM). This agreement focuses almost exclusively on materials in international transport and, like later nuclear security agreements, sets out steps that states must take to protect materials covered by the convention (by, for example, applying appropriate physical protection measures and criminalizing certain kinds of offenses). In contrast to the NPT, which assigns the IAEA a critical verification function, the only roles assigned to the agency by the CPPNM are purely administrative (for example, in sharing information about national points of contact, states may make use of the agency, though they are not required to do so). In a signal of the importance—or lack thereof—that states placed on nuclear security, seven years passed before the treaty garnered the twenty-one ratifications required for entry into force. Today 153 states are party to the CPPNM.63

In the 1990s, the IAEA acquired two other important nuclear security functions. First, following the collapse of the Soviet Union in 1991, detection of the smuggling of nuclear and radiological materials rose dramatically. The IAEA assisted in efforts to curtail such smuggling, most notably by creating the Illicit Trafficking Database in 1995 (although the agency had begun to collect media reports as early as 1992).64 The agency also created the International Physical Protection Advisory Service to conduct peer review of states’ nuclear security practices. These missions, which are initiated only at the request of states, have taken place since 1996.

Perhaps the most consequential international efforts prior to 2001, at least in terms of demonstrable results, were Soviet/Russian and, particularly, American bilateral programs with foreign partners. These efforts were initiated to mitigate a threat that the United States and Soviet Union had exported research reactors fueled with HEU.65 In 1978, the United States launched the Reduced Enrichment Research and Test Reactor program to develop alternative LEU fuels for HEU-fueled research reactors and to undertake conversion of foreign reactors (the Soviet Union had an equivalent program). American nuclear security cooperation was dramatically stepped up in the early 1990s, facilitated in large part by the Nunn-Lugar Cooperative Threat Reduction program. The scope of these efforts—which began with programs to reduce the risk from military and civilian nuclear materials in the former Soviet Union and were subsequently extended to include both chemical and biological threats, as well as other states—was vast. Two of the programs that originated at that time and are still ongoing are of particular note. First, the United States began efforts to take back both fresh and spent HEU fuel for U.S.-origin research reactors and worked with Russia to enable it to do the same for Soviet-supplied reactors. Second, in 1994, Washington set up the International Materials Protection and Cooperation program to help other states better secure nuclear materials and combat nuclear trafficking.

Since 2001, nuclear security initiatives have proliferated, even absent a broad consensus on the severity of the threat. Many of these efforts are not legally binding. In part, this reflects states’ reluctance to cede sovereignty on nuclear security. However, it also reflects the proclivities of the administration of President George W. Bush, which had a visceral dislike of both legally binding international agreements and the slow pace of the diplomacy needed to negotiate them. Even the exceptions to this general state of affairs are revealing. Two of the legally binding instruments negotiated since 2001 are UN Security Council resolutions. Because these resolutions are negotiated among only fifteen states (of which just five have vetoes), they bypass many of the complexities of a truly international process. The administration of President Barack Obama has continued in largely the same vein, even though it is much more sympathetic to internationalism, in large part because the Bush administration’s views are still alive and well in the Senate, which must provide its advice and consent to the ratification of treaties.

The measures that have originated since 2001 can be classified into five types: binding international law, nonbinding initiatives to facilitate cooperation, institutional reform, enhanced U.S. cooperation programs, and the sui generis Nuclear Security Summit process. The brief description that follows cannot do justice to the complexities and full scope of these multifaceted initiatives, activities, agreements, and organizations, and I unapologetically refer interested readers to a number of sources that provide the missing detail.66

Binding international law. Four legally binding instruments have been created since 2001. With the partial exception of the International Convention for the Suppression of Acts of Nuclear Terrorism (ICSANT), all focus nearly exclusively on the national implementation of measures to enhance nuclear security (as opposed, say, to mandating international cooperation):

  • UN Security Council Resolution (UNSCR) 1373 (2001) requires measures to suppress all forms of terrorism.
  • UNSCR 1540 (2004) requires measures to prevent nonstate actors from acquiring or using nuclear, chemical, or biological weapons or their means of delivery (further resolutions have prolonged its mandate).
  • The 2005 amendment to the CPPNM extends the treaty’s jurisdiction to domestic nuclear materials.
  • The 2005 ICSANT requires national legislation and international cooperation to prevent nuclear and radiological terrorism.

Voluntary initiatives to facilitate cooperation. Various “coalitions of the willing” have also been created to enable enhanced international cooperation on nuclear security:

  • The G8 Global Partnership against the Spread of Weapons and Materials of Mass Destruction, created in 2002, funds projects to reduce the risk of nonstate actors acquiring nuclear, biological, chemical, or radiological material in both the former Soviet Union and, more recently, elsewhere.
  • The Global Initiative to Combat Nuclear Terrorism, established in 2006, aims to promote best practice and strengthen international cooperation to deter, prevent, and respond to acts of nuclear terrorism.
  • The Proliferation Security Initiative is unusual in serving both a nuclear security and nonproliferation role.

Institutional Reform. Greater support for national efforts to counter nuclear terrorism is now available through both new and existing institutions:

  • The IAEA has enhanced its capacity to assist states with nuclear security by establishing the Office of Nuclear Security, which was recently upgraded to the Division of Nuclear Security. The agency also established a Nuclear Security Plan to guide its efforts and a Nuclear Security Fund to finance them through voluntary contributions.
  • INTERPOL has improved its capacity to help prevent nuclear and radiological terrorism, including by enhanced cooperation with the IAEA.
  • The World Institute for Nuclear Security has been established to promote best practice.
  • Centers of excellence to provide training in nuclear security on a national and regional level have also been established by various countries.

Enhanced U.S. cooperation programs. The United States has also stepped-up its bilateral cooperation programs. Perhaps most important, in 2004 the Bush administration launched the Global Threat Reduction Initiative to convert both research reactors and the targets used in radioisotope production to enable their use of LEU rather than HEU; to remove excess nuclear and radiological materials; and to enhance the physical security of vulnerable materials. (The initiative represented a consolidation and expansion of existing efforts.) This program was given a major boost in 2009 when Obama, during his Prague speech, articulated the goal of securing all vulnerable nuclear materials within four years. Although this target was not met, U.S. efforts were accelerated and have resulted so far in the removal or disposition, in some cases in cooperation with Russia, of nearly three thousand kilograms of HEU and plutonium from foreign states—more than doubling the total amount of material the United States has secured in this way.67 As part of these efforts, the United States also imported, for the first time, both separated plutonium and non-U.S.-origin spent fuel.68

The Nuclear Security Summits. In his Prague speech, Obama also announced the first Nuclear Security Summit, which took place in Washington, DC, in 2010. Follow-up meetings were held in Seoul in 2012 and The Hague in 2014. A fourth, and almost certainly final, summit will be held in the United States in 2016. Although the summits do issue communiqués, their most important deliverables have been unilateral commitments, known colloquially as “house gifts” (a promise to facilitate the removal of HEU or plutonium to the United States or Russia is the diplomatic equivalent of a bouquet of flowers at these events). Such commitments are not actually negotiated by the heads of state at the summit—but the fact of the summit does force bureaucracies to make nuclear security a priority and to overcome long-standing barriers to progress, precisely so that their leaders can declare success at the meeting. The summits have also raised awareness of the issue, among both national leaders and (to a lesser extent) the general public.

Looking forward, perhaps the biggest question facing the nuclear security agenda is whether the momentum that has built up behind it is sustainable. With a few exceptions (such as the Proliferation Security Initiative in its early days), the nuclear security agenda is not all that controversial, at least in comparison to nuclear nonproliferation. In contrast, say, to adopting an additional protocol, few states evince a principled objection to ratifying the amendment to the CPPNM. Moreover, a few of the challenges to further progress in nuclear security are technical, including the need to develop new fuel designs so that reactors that still use HEU can be converted to use LEU. There is no reason to suppose these technical challenges will not eventually be overcome.

That being said, neither is there much sense of urgency, except from the United States and a few close partners. Almost ten years after being opened for signature, for example, the amendment to the CPPNM has still not entered into force.69 In large part, the lack of urgency reflects the lack of shared perception of the threat. Moreover, simmering discontent from many non-nuclear-weapon states about the emphasis being placed on preventing nuclear terrorism—as opposed to, say, nuclear disarmament or technical assistance—has not dissipated and may have been increased by the high-level attention nuclear security has garnered.70 All these factors raise doubts about sustainability.

To make matters worse, the nuclear security agenda can easily be derailed by unrelated political disputes. Russia, for example, has generally been supportive of nuclear security efforts. However, U.S.-Russian cooperation on nuclear security has now become a victim of the Ukraine crisis. In October 2014, Moscow informed Washington that it was not planning to attend the 2016 Nuclear Security Summit (preferring, it said, to support IAEA efforts).71 More serious still, in December 2014, Moscow terminated cooperation with the United States on all nuclear security projects in Russia (though it has said it will continue cooperation focused on third countries).72

Moreover, U.S. leadership, which has often been instrumental to progress in nuclear security, should not be regarded as a given. While nuclear security is a “motherhood and apple pie” issue that no American politician opposes per se, it has little resonance with the public and is not generally seen as such a transcendently important issue that it should be immune from either budget cutbacks or partisanship. For a variety of reasons, including sequestration, the Obama administration has been gradually reducing its budget requests for nuclear security since fiscal year 2012.73 In 2016, the Republican majorities in both the Senate and, particularly, the House of Representatives are likely to try to reduce actual spending yet further. Implementing legislation for ICSANT and the amendment to the CPPNM was, for a number of years, held hostage to an arcane and politicized domestic dispute over the death penalty and wiretapping (until such legislation is passed, the United States cannot deposit its instruments of ratification).74 Meanwhile, the next U.S. president—whichever party he or she comes from—is unlikely to give nearly as much personal attention or time to the issue as Obama has, or continue the Nuclear Security Summit process (probably to the relief of many world leaders).

Sustainability matters both to avoid backsliding on the progress that has been made so far and to further enhance the nuclear security architecture, which currently has at least five key weaknesses.

First, it does not cover all nuclear materials—or even a majority of the most sensitive. About 85 percent of the HEU and separated plutonium in the world is military, and, now that U.S.-Russian cooperation on these materials has ended, they appear to be entirely exempt from the existing architecture, except for recognition in the communiqué from the 2014 Nuclear Security Summit of “the fundamental responsibility of States . . . to maintain at all times effective security of all nuclear . . . materials, including nuclear materials used in nuclear weapons.”75 What is more important for the present discussion, however, is that, while broad consensus exists on the desirability of minimizing civilian HEU use, no such agreement has been reached on civilian plutonium.76

Second, the standards mandated—or, in many cases, suggested—by existing agreements are often weak and, in some cases, ambiguous.77 Such “watering down” is often necessary to ensure agreement among participating states.

Third, many initiatives are voluntary—in both senses of the word: states do not have to sign onto them, and the standards imposed are often not legally binding. The only legally binding measures that apply to all states are UN Security Council resolutions. All other binding instruments have limited participation. Meanwhile, no IAEA guidance on nuclear security is binding (although, at the 2014 Nuclear Security Summit, thirty-five nations did agree to abide by various agency recommendations).78 Meanwhile, of the twenty-five states with weapon-usable nuclear material, six have never invited a peer review of their security standards (and another has not done so within the past five years).79

Fourth, no nuclear security agreement has created a formal verification regime, and only a few have informal transparency provisions to enable states to demonstrate their compliance—and even those that do have met with only partial success. The most notable transparency arrangement was a requirement in UNSCR 1540 for states to submit reports, within six months of the resolution’s adoption, detailing practices and plans for national implementation. Only fifty-one states met this deadline.80 And by the time of the December 2014 annual review, twenty states had still failed to fulfill the requirement.81 Moreover, among those reports that have been submitted, quality and comprehensiveness vary widely.

Fifth, the quality of national implementation of nuclear security measures appears to have been highly variable—although, in many cases, the available information is insufficient to enable an assessment.

In practice, correcting all these weaknesses simultaneously is probably impossible since the solutions to different problems can sometimes counteract one another. For example, if states tried to negotiate a verified, legally binding agreement on some aspect of nuclear security, they would probably end up with a document that was less demanding than existing voluntary standards (notably INFCIRC/225) and attracted a relatively limited number of signatories. A key question facing the nuclear security regime is, therefore, one of priorities; given the likely impossibility of making progress on all fronts simultaneously, on which of the weaknesses should energy be focused?

More broadly, one of the benefits of analyzing nuclear nonproliferation and nuclear security side by side is that it enables their strengths and weaknesses to be contrasted. The nonproliferation regime is comprehensive, nearly universal, and legally binding. However, changes are highly contested and, precisely because it is nearly universal, reform is painfully slow (absent crises, at least). By contrast, the nuclear security architecture is much more flexible. Some states choose to go faster than the pack and take on additional commitments (which is extremely unusual in nonproliferation). As a corollary, however, states that want to go slower are under relatively little pressure to do more. This comparison demonstrates that making the nuclear security regime more like the nonproliferation regime would not be a panacea. Comprehensiveness, universality, legally binding standards, and verification could solve some problems, but at the expense of introducing others. The trade-offs involved are almost certainly relevant to other technologies.

U.S. Domestic Oversight

The range of domestic nonproliferation and nuclear security oversight activities in the United States is vast. Major elements include:

  • Licensing of civilian nuclear facilities and nuclear materials by the NRC. Although the primary function of licensing is ensuring safety, license conditions include requirements pertaining to physical security and the control and accountancy of nuclear materials.
  • Domestic nuclear security. Licensees are responsible for implementing the security measures required by the NRC, which also verifies compliance.
  • Border security to detect the illicit transport of nuclear material into the United States. These efforts, run by the Department of Homeland Security, include radiation detectors at U.S. ports of entry, as well as a program, the Container Security Initiative, to scan cargo bound for the United States at foreign ports.82
  • Voluntary IAEA safeguards on nonmilitary facilities. Nuclear-weapon states party to the NPT are not required to accept IAEA safeguards. In 1977, however, the United States concluded a “voluntary offer” arrangement with the IAEA, under which it makes certain facilities available for safeguards. The U.S. “Eligible Facilities List” includes almost all facilities licensed by the NRC—GE Hitachi’s pilot-scale laser enrichment facility is a notable exception—as well as about thirty Department of Energy facilities.83 In practice, because of resource constraints, the IAEA has chosen to safeguard only a small number of these facilities and since 1994 has focused exclusively on storage facilities for materials declared to be in excess of military requirements.84 Currently, just one facility, the K Area Material Storage Vault at Savannah River National Laboratory, is under IAEA safeguards.85 The United States also has an additional protocol in force that contains the same provisions as the Model Additional Protocol except for a national security exemption. The NRC and the Department of Energy are primarily responsible for overseeing the implementation of these agreements, including by jointly operating a system for tracking U.S. nuclear materials.86
  • Domestic risk reduction efforts, led by the Department of Energy, include the management and disposition of excess military fissile materials.
  • Nuclear cooperation agreements, negotiated by the Department of State, permit international trade in nuclear materials and facilities.
  • Export controls to prevent U.S. technology, material, and equipment from being obtained by non-U.S. entities that might misuse them. Licenses for nonmilitary nuclear exports are granted by three agencies: the NRC (nuclear materials, reactors, and certain key reactor components), the Department of Commerce (“outside the core” equipment for nuclear reactors), and the Department of Energy (nuclear technology).87 Other departments, including the Department of State and the Department of Defense, are involved in reviewing applications. Several bodies, including the NRC, the Department of Commerce, the Department of Homeland Security, the Department of Justice, and the federal courts are involved in enforcement.

The legal basis for these efforts is diverse and includes legislation, regulations promulgated by agencies such as the NRC, executive orders, and international agreements. In addition to being bound by treaties such as the CPPNM, the United States has, as a matter of policy, also undertaken to abide by various nonbinding international agreements, such as the Nuclear Suppliers Group guidelines, as well as codes of best practice, such as IAEA recommendations on nuclear security.

This description is far from exhaustive; it includes neither the oversight of military nuclear activities, which by definition are not dual-use, nor U.S. government activities to negotiate and implement international nonproliferation and nuclear security activities.

Rather than attempt to provide a more detailed description of the whole of the domestic oversight regime, this article focuses on two contested elements that highlight some of the fundamental challenges associated with managing dual-use nuclear technologies: (1) the absence of a regularized procedure for assessing the nonproliferation consequences of domestic decisions; and (2) the nonproliferation and nuclear security conditions for international nuclear cooperation.

As exemplified by the licensing of GE Hitachi’s laser enrichment plant, the United States has no formalized process for assessing the nonproliferation implications of domestic decisions to develop or use new nuclear technologies. A nonproliferation assessment is required when the United States concludes an agreement with another country to permit the transfer of nuclear material or equipment. As such, one was conducted when the United States and Australia reached an agreement (which was considered an addendum to their existing nuclear cooperation agreement) to transfer Silex technology from Australia to the United States. However, this assessment focused purely on the nonproliferation implications of the transfer and not on the implications of commercialization.

The NRC does not license fuel-cycle technologies per se. Rather, it licenses facilities, and the only nonproliferation implication it considers in doing so is the applicant’s ability to protect classified information. However, the commission appears to have the necessary authority to require a much broader assessment. In March 2012, the Congressional Research Service concluded that, given the NRC’s statutory responsibility to promote the “common defense and security,” it “could reasonably conclude that it has sufficient existing authority to promulgate a regulation requiring that applicants provide the Commission with a proliferation risk assessment as part of the license application process.”88 A petition to require such an assessment for all new enrichment or reprocessing facilities was submitted by the American Physical Society but rejected by the NRC in June 2013. The NRC did not deny that new technologies might pose a proliferation risk but argued that Congress or the president was responsible for considering them.89

As a result of this decision, every part of the U.S. government has now eschewed responsibility for assessing the nonproliferation implications of domestic projects to develop new nuclear technologies or construct new facilities. Historically, the U.S. government has been more sensitive to the international implications of such decisions—although it has generally done so only where public funding was involved (which was not the case with GE Hitachi’s laser enrichment plant). Most notable, nonproliferation was the main factor behind the U.S. moratorium on reprocessing (although contrary to what is often asserted, particularly by proponents of reprocessing, the Ford administration’s initial decision was also a reflection of the need for the U.S. government to subsidize the project).90 The George W. Bush administration drafted but did not finalize a nonproliferation impact assessment for its Global Nuclear Energy Partnership program (which envisaged domestic reprocessing in the United States and potentially also in a small number of other states that already had access to the technology).91 Even the NRC has historically assigned a bigger nonproliferation role to itself; for example, in 1976 it postponed the licensing of fuel exports to a U.S.-supplied nuclear reactor in India, following that state’s 1974 nuclear test.92 The export of fuel to India was not a purely domestic action (like building an enrichment or reprocessing facility), but it does demonstrate that the NRC has interpreted its mandate more broadly in the past than it does now. Ultimately, however, all of these initiatives were ad hoc. Only when key actors have had an interest in nonproliferation—or were forced to take one—has it been factored into domestic decisions.

The U.S. export control system, as Congressional Research Service analysts Ian Fergusson and Paul Kerr note, has “long been criticized by exporters, nonproliferation advocates, allies, and other stakeholders as being too rigorous, insufficiently rigorous, cumbersome, obsolete, inefficient, or any combination of these descriptions.”93 The 1954 Atomic Energy Act, as amended by the 1978 Nuclear Nonproliferation Act, provides two separate legal bases for international nuclear cooperation. Pursuant to section 123 of the act, “significant nuclear exports” (transfers of nuclear material, reactors, or certain reactor components) require an intergovernmental agreement, generally known as a “123 agreement.”94 Even with such an agreement in place, each individual export of equipment or material still requires a separate license, which is issued by the NRC. Separately, section 57.b of the Atomic Energy Act empowers the secretary of energy to authorize “technology transfers and technical assistance,” provided that they “will not be inimical” to U.S. interests.95 The procedures for authorizing such transfers are set out in title 10 of the Code of Federal Regulations, part 810, and are generally known as “810 agreements.” Finally, nuclear technology and commodities that are not covered by the Atomic Energy Act—such as equipment that has both nuclear and nonnuclear applications—may still be controlled through the 1979 Export Administration Act.96 This act requires Department of Commerce authorization for the transfer of technologies and items that appear on the Commerce Control List (which includes but is much broader than dual-use nuclear technology).

Although a high-profile effort to simplify the U.S. export control system is currently underway, it will not affect nuclear exports conducted pursuant to either 123 or 810 agreements.97 Instead, the Obama administration has undertaken a separate effort to revamp 810 agreements while trying to head off congressional efforts designed to make 123 agreements more restrictive.

The existing regulations for 810 agreements define a set of “generally authorized” activities that have received prior approval from the secretary of energy and can therefore be conducted abroad without “specific authorization.” However, the regulations also contain a list of states for which this exemption does not apply and for which specific authorization is required on an activity-by-activity basis. With the most significant of its proposed changes, the Obama administration seeks to reverse this approach by specifying those states for which general authorization does apply. The Nuclear Energy Institute, which lobbies on behalf of the nuclear industry, has opposed this approach on various grounds, including that it would increase, by seventy-seven, the number of states requiring specific authorization.98 In response, the Obama administration has argued that the number of transfers to these states is so small that the security benefits of the change would outweigh the costs.

Efforts to revise the requirements for 123 agreements have been more controversial. In 2009, the United States and the United Arab Emirates concluded a nuclear cooperation agreement under which the latter made a legally binding commitment not to acquire enrichment or reprocessing technology. Inevitably, this agreement raised the question of whether the United States would require other states to make a similar promise—termed the “gold standard” by the Obama administration (though it subsequently walked away from this term). Following a prolonged series of internal reviews, which were eventually concluded in late 2013, the administration announced a policy of deciding on a “case-by-case” approach whether to insist upon the would-be partner’s renouncing enrichment and reprocessing.99 Since then the administration has concluded an agreement with Taiwan, which did make a binding undertaking not to acquire enrichment or reprocessing technology, and one with Vietnam, which did not—although Hanoi did express its intent, in nonbinding language, “to rely on existing international markets for nuclear fuel services, rather than acquiring sensitive nuclear technologies.”100

An effort to try to force the administration into incorporating the “gold standard” into future 123 agreements has, however, gained some traction in Congress. Legislation has been introduced that would add a legally binding commitment not to acquire enrichment or reprocessing technology to the list of requirements that qualifies a 123 agreement for expedited review.101 An eligible agreement enters into force ninety days of “continuous session” after it has been presented to Congress, unless legislative action is taken to block it. By contrast, an agreement that does not meet all the requirements must secure congressional approval before it becomes operative. The effect of the proposed legislation would, therefore, be to make it extremely difficult for a 123 agreement to enter into force unless the partner foreswore enrichment and reprocessing.

The fundamental debate here is (or, perhaps, ought to be) over the value of a nuclear cooperation agreement with the United States at a time when states that want to acquire nuclear power reactors can choose from a growing list of vendors. For much of the Cold War, the United States and the Soviet Union were essentially the only two suppliers of nuclear reactors—meaning that many states effectively had no choice. Today, there are commercial suppliers from seven states, and there is little reason to suppose this number will not grow slowly over time.102 The Obama administration has argued (fairly persuasively in my opinion) that if it were to demand that potential buyers renounce enrichment and reprocessing they would simply go elsewhere, resulting in a loss for both U.S. industry and nonproliferation, since other supplier states impose less-rigorous nonproliferation requirements than the United States.103 Proponents of the “gold standard” do not tend to refute this argument directly so much as urge the United States not to be complicit in aiding states to develop nuclear power programs that could, in the future, be used to advance military goals. This debate is not unique to nuclear technology; the possibility that raising the requirements for exports could be counterproductive if not all suppliers agree to do so is a dilemma in any form of sensitive technology trade where multiple potential sellers are in competition with one another.

Current State of Domestic Oversight in Other Countries

Nuclear Nonproliferation Oversight

Five nuclear-weapon states are recognized by the NPT. Like the United States, each has a “voluntary offer” arrangement with the IAEA and has an additional protocol in force—although the scope of these agreements is highly variable.104 In practice, the IAEA chooses to safeguard few of the eligible facilities. Most of the facilities that are safeguarded were selected because they use foreign technology or process foreign material, and the application of safeguards was a condition of supply. For example, a Russian-supplied centrifuge enrichment plant in China is under safeguards.

Four nuclear-armed states are not party to the NPT: India, Israel, Pakistan, and North Korea.105 Israel and Pakistan never joined the treaty but agreed to IAEA safeguards on a few foreign-supplied facilities as a condition of supply. India also did not join the treaty but, pursuant to the U.S.-India nuclear deal, agreed to separate its military and civilian nuclear sectors and safeguard the whole of the latter (which includes domestically manufactured power reactors).106 North Korea withdrew from the NPT in 2003 and currently does not allow any IAEA safeguards, including on foreign-supplied facilities.

The biggest category of states that are party to the NPT consists of those that do not possess nuclear weapons and are required to accept IAEA safeguards on all of their nuclear activities. The extent of these safeguards depends on whether the state has accepted an additional protocol and whether the amount of nuclear material it possesses is so small that it qualifies for a so-called small quantities protocol (which holds most of the requirements of a comprehensive safeguards agreement in abeyance to reduce the burden on the state).

In non-nuclear-weapon states, the primary nonproliferation role of domestic regulatory agencies is to cooperate with the IAEA to allow the implementation of safeguards. All non-nuclear-weapon states are required to have a “state system of accounting and control” to track nuclear materials and provide data to the IAEA (though such a system can also play an important role in nuclear security by detecting any unauthorized removal of nuclear material). These systems vary in effectiveness. For example, in 2004, the Egyptian Atomic Energy Authority (which operates that country’s system of accounting and control) was reportedly forced to inform the IAEA that “it did not have the authority necessary for it to exercise effective control of all nuclear material and activities in the State.”107 Less dramatic problems with safeguards implementation are routine. For example, a leaked IAEA document reveals that, in 2012, seventy-one states, most of which are developing, routinely missed reporting deadlines. A few more-developed states operating research or power reactors, including Brazil, China, Mexico, and Poland, sometimes did so too.108

In many non-nuclear-weapon states, constitutional provisions, laws, or fatwas forbid the development of nuclear weapons. While these may play a role in nonproliferation—by encouraging whistle-blowing or empowering domestic opponents of nuclear-weapon development, for example—there is an obvious tension in self-regulation; that is, in a government overseeing the implementation of its own nonproliferation commitments. For this reason, domestic nonproliferation oversight of dual-use nuclear activities is very much secondary to international oversight.

In theory, domestic processes could play an important role in helping states to consider the international implications of domestic nuclear energy decisions. To the best of this author’s knowledge, however, no state currently has a formalized process for factoring proliferation implications into domestic licensing decisions—although, like the United States, some states have occasionally created one-off processes. In the United Kingdom, for example, a judicial inquiry was held in 1977 into a hugely controversial plan to construct a new reprocessing plant. Proliferation was extensively discussed during this inquiry—although the presiding judge reached the somewhat unexpected conclusion that the construction of THORP, as the facility was called, would advance nonproliferation ends since foreign states would be able to separate plutonium in the United Kingdom and so would not need to construct their own facilities.109 (For the record, the United Kingdom’s most important foreign client, Japan, did eventually build its own domestic facility anyway, which should have surprised no one since that was Tokyo’s stated long-term policy at the time of the inquiry.) If nothing else, this experience demonstrates just how difficult and controversial proliferation assessments can be.

Domestic processes play a more important role in export controls. The two international nuclear export control arrangements—the Zangger Committee and the Nuclear Suppliers Group—seek to harmonize standards. However, their effectiveness in this regard is limited, both because they are not universal and because their guidelines are designed to permit some significant differences in procedures and policy between participating governments.110 For example, the complexity of the U.S. export control system—its three-way split of responsibility for regulating nonmilitary exports, in particular—appears to be unparalleled. By contrast, in some states, including Japan and Russia, all nuclear-related exports are licensed by a single agency.

Differences in policy are more interesting. The United States requires other states to seek prior consent before reprocessing U.S.-origin spent fuel (which includes any fuel that has “passed through” U.S. technology, including in fuel fabrication facilities and reactors). Japan has followed this model in all of its nuclear cooperation agreements with non-nuclear-weapon states except Kazakhstan. By contrast, France, Russia, and South Korea have not included such a provision in their nuclear cooperation agreements (with the exception of the South Korea–UAE agreement).111 These states’ policies toward retransfers also differ (in practice, the only item involved in nuclear power generation that could conceivably be retransferred is spent fuel sent for reprocessing or storage abroad).112 France, Japan, South Korea, and the United States generally require importers to seek their permission before reexporting controlled items. Russia, however, generally demands only that the retransfer has the same conditions attached as the original transfer (although it does require prior consent for the retransfer of particularly sensitive items, such as separated plutonium or HEU).

Although the United States generally does insist on more demanding conditions for nuclear cooperation than do other states, there are exceptions—as highlighted, for example, by nuclear cooperation with India. All major nuclear exporters have now concluded cooperation agreements with New Delhi except for Tokyo, which is seeking a unique provision allowing it to terminate cooperation should India conduct another nuclear test, as well as verification requirements that go beyond IAEA safeguards.113

Controlling exports of equipment with both nuclear and nonnuclear uses is, however, probably more challenging than regulating exports of reactor fuel and components, both because the former are much more difficult to track and because far more states are involved in their production and transshipment. The A. Q. Kahn network, for example, manufactured many centrifuge components in Malaysia and used the United Arab Emirates as a key transshipment node. Neither state was—or is—a participant in the Nuclear Suppliers Group. At the time, the absence of any meaningful export controls was the norm for such states (even for states within the Nuclear Suppliers Group, export controls often left a lot to be desired).

Reporting pursuant to UNSCR 1540, which mandates the creation and enforcement of export controls, gives some sense of the current state of play. A survey of thirty-eight states’ reports from 2008 revealed large disparities in implementation, with less-developed states having generally instituted fewer of the resolution’s requirements.114 The most neglected requirements related to so-called deemed exports (in-country transfers of intangible commodities to foreign nationals) and controls over both proliferation financing and transportation services.115 In these areas, 16–42 percent of the states surveyed had legislation in force that provided for some type of control. Moreover, enforcement lagged behind the creation of a legal framework in most of the areas in which the resolution requires states to act.

Nuclear Security

Domestic nuclear security is similar to export controls in that national requirements and the quality of implementation vary widely (even if all states can agree on the goal of preventing unauthorized access to nuclear materials and facilities). By far the most comprehensive survey of nuclear security standards is the Nuclear Threat Initiative’s Nuclear Materials Security Index.116 This index scores states on nineteen indicators grouped in five categories. Two of these categories—“quantities and sites” (comprising indicators such as the quantity of nuclear material in a state, how many sites the material is stored at, and how frequently it is transported) and “risk environment” (comprising indicators such as political stability and the pervasiveness of corruption)—are made up of factors that a regulator cannot reasonably be expected to affect. The remaining three categories, comprising twelve indicators, rate a state’s legal infrastructure and the general nuclear security requirements it imposes, its involvement in international agreements and initiatives, and the extent to which international undertakings have actually been enshrined in domestic legislation. The index does not examine the specifics of security arrangements at individual sites, which generally are (and indeed should be) secret. That said, in a number of states, even information about general nuclear security requirements is not available.

To help identify, in a crude sense, where states’ nuclear security standards differ most from one another, the following list shows the twelve indicators of regulatory effectiveness in order of decreasing variance among states (i.e., states’ scores are most widely spread in item 1 and least widely spread in item 12; only scores from the twenty-five states with at least one kilogram of weapon-usable nuclear material are included).117

  1. Physical security during transport
  2. Independent regulatory agency
  3. Control and accounting procedures
  4. Domestic nuclear materials security legislation
  5. International assurances
  6. International legal commitments
  7. Voluntary commitments
  8. UNSCR 1540 implementation
  9. Safeguards adherence and compliance
  10. Insider threat prevention
  11. On-site physical protection
  12. Response capabilities

The biggest differences among the twenty-five states lie in their standards for physical security during transport, whether they have an independent regulator, and in their control and accounting procedures. Ten states (including the United States) have regulations consistent with the IAEA’s most recent nuclear security guidance, INFCIRC/225/rev.5; ten (including China, Japan, and Russia) have guidance consistent with a slightly older version, INFCIRC/225/rev.4, which places less of an emphasis on testing security systems, including with “force-on-force” exercises; and five (Argentina, India, Pakistan, South Africa, and Uzbekistan) have still lower or no published standards. Separately, the states’ scores on the regulatory independence indicator are widely spread because three of the twenty-five (India, Iran, and North Korea) lack an independent regulator.118 Finally, states’ scores for control and accounting procedures differ widely, principally because of varying rules for access controls (requiring identity checks for people entering sensitive areas and keeping appropriate records). Overall, the Nuclear Materials Security Index contains some surprises about which states perform well and poorly. Belarus, for example, achieves perfect scores for its physical security standards, the independence of its regulator, and its control and accounting procedures. By contrast, Belgium’s requirements for physical protection during transport and access controls fall short of internationally recognized best practice.

One problem with the Nuclear Materials Security Index is that it does not consider enforcement. Russia and the United States (along with Kazakhstan) are the top-ranked nations for preventing insider threats. However, even if Russian regulations are consistent with best practice, evidence suggests that implementation has been lacking.119 For example, Matt Bunn, a former U.S. government official, recalls visiting a Russian nuclear facility in the mid-2000s and seeing two portal monitors to detect radiation, one American and one Russian. When he asked why there were two monitors, he was told that the U.S. monitor was deactivated on Thursdays, when plant activities tended to set it off accidentally, and the less sensitive Russian monitor was used instead.120 He notes that because “every insider was aware of this practice, and would know to plan an attempted theft for a Thursday,” this procedure rendered the U.S. system “largely pointless.”

Another example of an implementation challenge concerns assessments of the security threats that facilities are designed to withstand. Best practice requires the development of—and regular updates to—so-called design basis threats. Not only does the Nuclear Materials Security Index not give Japan credit in this category, apparently because Japanese regulations impose no requirement for regular updates, but most, if not all, Japanese facilities are not designed to withstand attack by armed terrorists.121 U.S. visitors to Japanese nuclear facilities are generally surprised by the lack of armed guards (which is in part a consequence of legal restrictions on the use of guns by private security firms). While Japanese officials have insisted, with some justification, that armed terrorism is unlikely in Japan, the fact that the country has been the victim of chemical attacks by an apocalyptic cult suggests that armed attacks on nuclear facilities should not be considered impossible. To be fair, Japan’s nuclear security system does seem to be improving, and armed police are now present at its most sensitive facility, the Rokkasho Reprocessing Plant—but progress appears to be slow. The Japanese experience also raises the question of the extent to which nuclear security measures can safely be adapted to local cultural norms and laws (such as restrictions on private gun ownership).

Other Models for Managing the Risk from Nuclear Technology

Today any ambition to reform the nuclear nonproliferation regime is markedly lacking. Almost all governmental and nongovernmental efforts to bolster the regime as a whole (as opposed to those focused on country-specific challenges, most notably from Iran) aim to achieve incremental change.122 In some part at least, this emphasis on gradual evolution is the result of the failure, last decade, of two ambitious—perhaps revolutionary—“big ideas.”

In 2004, George W. Bush identified what he termed a “loophole” in the NPT that allows states “to produce nuclear material that can be used to build bombs under the cover of civilian nuclear programs.”123 In response, he called upon “the world’s leading nuclear exporters” to provide reliable nuclear fuel supplies to states that renounced enrichment and reprocessing, and he called upon the Nuclear Suppliers Group to ban the transfer of these technologies to states that did not already possess them.

These proposals met with considerable opposition from such states, including close U.S. allies such as Australia and Canada. Some states that did have the technology opposed them too. In the event, the Nuclear Suppliers Group took more than seven years to agree on a relatively permissive and distinctly ambiguous set of criteria under which exports of enrichment and reprocessing facilities would be permitted.124 Today, the Nuclear Suppliers Group’s participating governments show little appetite to reopen the debate.

Bush’s call for reliable fuel assurances also contributed to a debate, started by then IAEA Director General Mohamed ElBaradei the previous year, about the internationalization or multilateralization of the nuclear fuel cycle. The idea of taking sensitive nuclear activities out of the hands of individual states has been periodically revisited since it was first proposed in the Acheson-Lilienthal report, which suggested that all such activities be conducted by an international organization. Moreover, multilateral (but not truly international) coalitions have been set up for enrichment, including Urenco, which was created by the British, Dutch, and German governments to develop and operate centrifuge technology, and Eurodif, under which France, Belgium, Italy, and Spain (and, at various times, Sweden and Iran) owned a gaseous diffusion plant in France that has recently been shut down.125

ElBaradei reawakened this debate in 2003 when he proposed that enrichment and reprocessing should take place only in “facilities under multinational control.”126 This suggestion, combined with Bush’s call for reliable fuel assurances, sparked a plethora of proposals from governments and nongovernmental organizations.127 Some of the proposals merely sought to reinforce existing market mechanisms. Others were extremely ambitious and involved complex international ventures. The German contribution, for example, proposed that the IAEA should own a new enrichment facility in an extraterritorial area (i.e., an area outside the control of any nation-state). Most of these proposals did not incorporate Bush’s suggestion that fuel supply assurances be extended only to governments that renounced enrichment and reprocessing (even the U.S. government eventually dropped this requirement from its proposals). Nonetheless, because of their complexity and, more important, a lack of interest from governments, all of the more-ambitious proposals quickly fell by the wayside—and seem unlikely to be revived in the short term. Three modest initiatives did make it off the ground. However, the only one that can be described as multinational is the International Uranium Enrichment Center, a Russian venture owned jointly with Armenia, Kazakhstan, and Ukraine that has access to some of the output from a Russian centrifuge facility.128

Domestic nuclear nonproliferation policy is, in theory at least, more susceptible to faster change since the buy-in of only one state is required. States could, for example, require themselves to conduct an assessment of the nonproliferation implications of domestic licensing decisions. Yet the effort in the United States to petition the NRC to do just that has failed, and no discussion of the idea appears to be ongoing elsewhere. That said, another aspect of nuclear regulation—safety—does provide a clear precedent for such an assessment. Specifically, a basic principle of radiation safety is that “facilities and activities that give rise to radiation risks must yield an overall benefit.”129 This principle has led to a formal requirement for all European Union states to “ensure that all new classes or types of practice resulting in exposure to ionizing radiation are justified in advance of being first adopted or first approved by their economic, social or other benefits in relation to the health detriment they may cause.”130 If the word health is changed to proliferation, this requirement nicely captures the goal of a nonproliferation assessment.

In contrast to nuclear nonproliferation, one big idea is under discussion for nuclear security: the creation of a universal, comprehensive, legally binding nuclear security regime that requires states to take steps to demonstrate their compliance. Several variations on this central theme have been proposed. One idea is to develop a framework agreement for nuclear security—similar to, say, the UN Framework Convention on Climate Change—that would set out basic nuclear security principles and “allow for the negotiation of supplementary protocols that require more detailed nuclear security actions.”131 Another idea is for states to agree to give the IAEA (or another international body) “the authority to define, review, and monitor national nuclear security standards and to evaluate compliance.”132 These ideas have, so far at least, largely been championed by nongovernmental experts seeking to ensure that the momentum generated by the Nuclear Security Summit process is not lost after the 2016 meeting, which is expected to be the last. So far, none of the states participating in the process have endorsed the idea of a comprehensive and legally binding nuclear security treaty because they worry it would be too bureaucratic, unwieldy, and difficult to negotiate. That said, their initially negative reactions do appear to have softened somewhat.

Oversight Challenges: Structural Barriers to Change

Given the relatively limited role of individual scientists, the key actors in both domestic and international debates over the regulation of dual-use nuclear technology are governments and the nuclear industry—which are often one and the same. Two main fault lines traverse the international politics of nuclear regulation: between the states that possess nuclear weapons and those that do not; and between the states that possess nuclear technology and those that do not. Although the nuclear-armed states largely overlap the technology holders, the two groups are not identical. Germany, Japan, and the Netherlands, for example, all possess sensitive nuclear technology but do not have nuclear weapons. And the two fault-lines, although not perfectly predictive of states’ positions—Western European countries without nuclear weapons, for example, tend to side with the nuclear-weapon states on a number of issues—are probably the most important determinant.

States without a particular type of nuclear technology—whether nuclear reactors, enrichment, or reprocessing—are often concerned about protecting, both legally and practically, their ability to acquire that technology in the future, even if they have no plans to do so. The United Arab Emirates and Taiwan are, for example, likely to prove fairly unusual in being willing to renounce enrichment and reprocessing. Vietnam’s position—to insist on preserving what it sees as a right, even though it appears to have little interest in exercising it—is much more common. The Nuclear Suppliers Group guidelines on the transfer of enrichment and reprocessing technology took so long to negotiate because a number of states without such technology, including Canada, South Korea, and Turkey, wanted to keep open the option of acquiring it.133 Their position was even shared by some states with enrichment facilities, including Brazil and Argentina, which appear to have been taking a “principled” stance against technology denial, as well as protecting their ability to acquire new forms of enrichment and reprocessing technology in the future. One example of how deep concern about technology denial among some of these state runs is that when Brazil and Turkey negotiated a (never implemented) fuel swap with Iran in 2010, to try to help diffuse the crisis, they stated in their joint declaration that “Turkey and Brazil appreciated Iran’s . . . constructive role in pursuing the realization of nuclear rights of [the NPT’s] Member States.”134

Even proposals that do not explicitly involve restrictions on technology are often treated with skepticism by the “have-nots” out of concern for a hidden agenda. The clumsy rollout of U.S. nonproliferation proposals in 2004, when Bush did explicitly condition U.S. fuel supply assurances on states’ renouncing enrichment and reprocessing technology, did a lot to feed this fear—and it has persisted long after Washington ceased to talk about technology denial. Discussions about multilateralizing or internationalizing the fuel cycle during the late 2000s were, for example, made much more difficult by “the common misconception that any multilateral mechanism necessarily implies the deprivation of the [NPT] Article IV right of [non-nuclear-weapon states] to peaceful nuclear technology.”135 Fear of technology denial affected progress in nuclear security too. In 2009, for example, the U.S. and UK governments pushed the idea of making nuclear security a fourth “pillar” of the NPT to stand alongside the existing three of nonproliferation, disarmament, and the peaceful use of nuclear energy. Many states, particularly within the Non-Aligned Movement, viewed this as an attempt to create additional obligations that would interfere with their “inalienable right” to use nuclear energy.136 The British and American governments rapidly dropped the idea. If the United States and other like-minded governments were to back the idea of a comprehensive and legally binding nuclear security treaty, many states might interpret it as another attempt at technology denial. Disagreements over the severity of the threat from nuclear terrorism would serve only to exacerbate their suspicion.

Discussions about strengthening nonproliferation and nuclear security also bring out the tensions between nuclear-weapon states and non-nuclear-weapon states over disarmament. Although non-nuclear-weapon states have expressed dissatisfaction that the Obama administration’s push on nuclear security has come at the expense of disarmament, the issue of nonproliferation is where the acrimony really comes to the fore.137 Almost any proposal to strengthen the nonproliferation regime is opposed by many non-nuclear-weapon states, especially within the Non-Aligned Movement, on the grounds that the nuclear-weapon states have failed to live up to their disarmament commitments (which, among non-nuclear-weapon states, increasingly seems to mean negotiating a time-bound treaty to abolish nuclear weapons).

Whether the lack of progress on disarmament is a genuine reason or a convenient excuse for those non-nuclear-weapon states that oppose strengthening the nonproliferation regime has been the subject of much debate.138 This author has come to the conclusion that many non-nuclear-weapon states do feel genuine frustration at what they see as the nuclear-weapon states’ failure to live up to the disarmament bargain, and do seek to leverage nonproliferation to achieve their disarmament goals. However, rarely is disarmament the only reason why key non-nuclear-weapon states oppose any given nonproliferation measure. For example, Togzhan Kassenova, a scholar of Brazil’s nuclear program, concludes that Brasilia’s objections to the additional protocol are multifaceted and include, for example, a desire not to compromise sensitive information related to its nuclear submarine or centrifuge programs, as well as frustration over the pace of disarmament.139 In public, however, it is much easier for Brazil to justify its position by simply blaming the nuclear-weapon states. Strengthening the nonproliferation regime is likely to prove difficult if many non-nuclear-weapon states’ opposition is overdetermined in this way; specifically, progress toward disarmament may be a necessary condition, but it is unlikely to prove sufficient.

The nuclear industry also tries to influence nonproliferation and nuclear security policy. The industry can sometimes be a direct participant in debates, particularly domestic ones. At other times, it is an indirect participant, seeking to influence the position that its national government carries into international discussions. That said, because so much of the global nuclear industry is state-owned, it is probably more accurate not to think of the nuclear industry as an independent actor in many states but to characterize national debates over nuclear policy as bureaucratic processes, internal to government, in which nuclear industrial considerations are an important factor.

That said, the nuclear industry is really two separate industries: operators and vendors (particularly of nuclear reactors and reactor components). The former has a general and somewhat visceral predisposition to oppose additional regulation and generally tries to deny any link between nuclear power and proliferation. Moreover, operators sometimes object to additional nonproliferation or nuclear security measures on more concrete grounds. Brazil’s centrifuge operators, for example, are opposed to the additional protocol. “Grumblings about the cost and perceived unfair safeguards burden” are often heard from diplomats representing states, including Canada, Germany, and Japan, that have large nuclear industries.140 Despite their concerns, operators, for at least three reasons, are probably not a major barrier to improvements in nonproliferation and security. First, because so few nuclear facilities in the nuclear-weapon states are safeguarded, a major source of potential opposition is effectively nullified. Second, particularly where nuclear security is concerned, many operators recognize how damaging an incident could be to the health of the nuclear industry as a whole. Third, the cost of safeguards and nuclear security is relatively small (especially when compared to safety).

Vendors, by contrast, have a bigger influence on policy. Within the United States, reactor and component vendors, which are not state-owned, constitute a powerful lobby that has opposed (on relatively solid grounds) congressional efforts to tighten the conditions under which 123 agreements can be concluded and (with much flimsier arguments) administration efforts to revamp the rules for 810 agreements. Meanwhile, although the policy-formation process is much more opaque, industrial considerations almost certainly play a major role in influencing the policies of the states that own major exporters, including France and Russia.

That does not mean raising revenue is the only purpose of reactor sales; they are also a tool of foreign policy. For example, the provision of a nuclear power plant, with generous financing arrangements, was integral to achieving a package of energy agreements that Russia struck with Turkey in 2010.141 A government’s desire for flexibility, so that nuclear power plant sales can be used to further other foreign policy goals, adds yet another layer of opposition to tightening rules for exports. A desire to facilitate “strategic” reactor sales can even lead governments to undermine those nuclear nonproliferation rules that are in place. The U.S.-India nuclear deal—under which the United States sought an exemption from Nuclear Suppliers Group rules to enable reactor sales to India—was, from the U.S. perspective at least, primarily motivated by the goal of forging a strategic partnership with India, particularly against China.142 Similarly, furthering a strategic partnership appears to have been one important motivation behind China’s decision to supply reactors to Pakistan. A cliché within the nuclear policy community is that nonproliferation goals almost always lose out when they conflict with efforts to strengthen a bilateral relationship.


“The exceptional nature of nuclear weapons,” the political scientist William Walker wrote in 2007, “calls for an exceptional kind of cooperative politics.”143 The use of nuclear weapons represents the highest level of violence to which humanity can resort. Even a single weapon could lead to hundreds of thousands, or even millions, of deaths, appalling suffering of the survivors, the extensive destruction of property, and widespread radioactive contamination. To be sure, deterrence may continue to prevent use. However, deterrence is likely to become riskier as more states acquire nuclear weapons and to become particularly unreliable should nonstate actors do so. Moreover, states worry that the spread of nuclear weapons—even if they are not used—will undermine their national security in other ways. Their fears are often visceral but include falling victim to nuclear blackmail (i.e., being forced to make concessions under threat of nuclear attack) and adversaries’ becoming emboldened after acquiring nuclear weapons.

The system for managing dual-use nuclear technology—as complex and multifaceted as it is—is focused narrowly on mitigating these “high-end” harms as they are posed by two potential agents: states that do not possess nuclear weapons; and sophisticated, well-funded, malevolent terrorist groups.

The nuclear nonproliferation regime, which aims to prevent further proliferation to states, represents Walker’s “exceptional kind of cooperative politics.” While non-nuclear-weapon states were not willing to forsake the use of nuclear energy for peaceful purposes, they did agree to construct a legally binding and nearly universal regime to reduce the risk that dual-use nuclear technology would be used in the development of nuclear weapons. This regime—for all its flaws and risks—represents humankind’s most comprehensive attempt to manage any dual-use technology. It is essentially a transparency regime: states are permitted to conduct any nuclear activity, except the “manufacture” of nuclear weapons, provided they declare it and permit verification by the IAEA. Such verification is facilitated by both the relatively small number of weapon-usable fissile materials and the existence of inspection procedures that permit the diversion of declared nuclear materials to be detected reasonably reliably.

Restrictions on the trade in nuclear technology, materials, and equipment are a second component of the nonproliferation regime (although their effectiveness is gradually diminishing). They include unilateral and multilateral policies setting out the circumstances under which controlled knowledge or items can be exported, as well as unilateral and multilateral efforts, such as the Proliferation Security Initiative, to detect and interdict illicit transfers. Export controls and safeguards are connected: one purpose of the former is to provide the IAEA with information helpful to implementing the latter.

As exceptional as the nuclear nonproliferation regime is—in both its uniqueness and its success—it faces many serious stresses, including technical challenges such as detecting clandestine facilities and political challenges such as enforcement in the event a violation is detected. Few of these problems have easy answers, and there are many barriers to change. These include technical barriers, such as the inherent limitations of technology to detect secret nuclear activities; commercial barriers resulting from the potentially lucrative nature of nuclear exports; and political barriers, including the acrimony between nuclear-weapon states and non-nuclear-weapon states, as well as between technology holders and would-be recipients. These barriers make even incremental change difficult. Changes that would be more revolutionary—for example, rules to prevent the further spread of fuel-cycle technologies or an agreement to place all fuel-cycle facilities under multilateral or international control—are nonstarters. For the time being, the best opportunity for enhancing the nuclear nonproliferation regime, in this author’s opinion, is for advanced states to develop internal procedures to assess how their domestic decisions to develop or use new nuclear technologies might affect proliferation dynamics globally.

While the nuclear nonproliferation regime developed from the top down—the passage of international agreements required national legislation and implementation—the nuclear security architecture developed from the bottom up. Efforts to prevent terrorist organizations or other unauthorized personnel from gaining access to nuclear facilities or acquiring nuclear material were originally a purely sovereign responsibility. Best practices include the development and periodic updating of design basis threats, physical protection to prevent unauthorized access, material control and accountancy to detect unauthorized removal, and effective response capabilities to recover material that has been removed.

Gradually, a patchwork of international measures has been developed to improve both national standards and the quality of implementation. These measures can be divided into four categories. First, a few legally binding instruments, such as the CPPNM and its amendment, as well as UNSCR 1540, impose legally binding standards in some areas. Second, best practice guides, such as INFCIRC/225, are an important but nonbinding way of plugging some of the remaining gaps. Third, various international assistance programs are also on offer, serving a wide variety of functions. The IAEA, through its International Physical Protection Advisory Service, provides peer reviews comparing a state’s nuclear security system to internationally accepted best practice. National and regional centers of excellence are being set up to provide services such as training. Direct assistance with enhancing physical protection and material control procedures, removing high-risk materials, and converting HEU-fuelled research reactors to use LEU fuel is also available from multilateral coalitions and on a bilateral basis, most notably from the United States. Fourth, information-sharing initiatives, including the IAEA’s Illicit Trafficking Database, have been created to warn states of potential threats. Information sharing is also a major purpose of the Proliferation Security Initiative.

Inspired, no doubt, by the nonproliferation regime, many nuclear security advocates are championing the idea of a comprehensive, legally binding, universal nuclear security treaty. Such a treaty would be a heavy diplomatic lift. The lack of a shared threat perception and fears that a nuclear security treaty would ultimately lead to technology denial would make garnering the support of many developing states difficult. Meanwhile, many developed states, worried that negotiations would be extremely difficult and carry a low chance of success, feel that their diplomatic capital might be better spent trying to effect more-incremental change. They are probably right. Because of the inevitable trade-off between high standards and universality, any treaty that was acceptable to a majority of states would, in all probability, lack the teeth necessary to combat the very real nuclear security problems that do exist.

In fact, the nuclear nonproliferation regime offers something of a false lead for nuclear security. Nonproliferation politics have become much more fractious since the NPT was concluded in 1968. While a majority of states still support the treaty, few feel that most others have acted in good faith. Most vociferously, the Non-Aligned Movement, while expressing concern about technology denial, argues that the nuclear-weapon states have failed to live up to their disarmament commitments. Meanwhile, although their public comments tend to be more measured, the United States and its partners are equally frustrated over the failure of many non-nuclear-weapon states to support efforts to bolster the nonproliferation regime. This acrimony has infected discussions of nuclear security; as desirable as isolating nuclear security from nonproliferation politically might be, doing so is simply not possible. Given this political reality, it is far from clear that fashioning a nuclear security regime in the mold of the nuclear nonproliferation regime is possible or that trying to do so would be constructive.


1. The history of U.S. efforts, as well as the differences between various laser enrichment processes, are summarized in Jack Boureston and Charles D. Ferguson, “Laser Enrichment: Separation Anxiety,” Bulletin of the Atomic Scientists 61 (2) (March–April 2005): 16,

2. General Electric originally intended to submit the license application by December 2007. Luis A. Reyes, memorandum to the Commissioners, “Status of the Silex Project Proposed by General Electric Nuclear,” SECY-07-0031, Nuclear Regulatory Commission, February 9, 2007, 2,

3. GE Hitachi Nuclear Energy, “Global Laser Enrichment Submits License Application to Build First Commercial Uranium Enrichment Plant Using Laser Technology,” press release, June 30, 2009,

4. Geoffrey Rothwell, “Market Power in Uranium Enrichment,” Science & Global Security 17 (2009): 136–138,

5. For a short but trenchant discussion of the issues see R. Scott Kemp, “SILEX and Proliferation,” Bulletin of the Atomic Scientists, July 30, 2012,

6. Although a definitive conclusion is impossible to reach without access to classified information, theoretical considerations suggest that molecular laser isotope separation processes, which include Silex, are considerably more suitable for HEU product than atomic vapor laser isotope separation. Allan S. Krass, Peter Boskma, Boelie Elzen, and Wim A. Smit, Uranium Enrichment and Nuclear Weapon Proliferation (London: Taylor and Francis, 1983), 165–166, 170–171,

7. Timothy C. Johnson, memorandum to Brian W. Smith, “May 10, 2012, Meeting Summary: General Electric-Hitachi Public Meeting on Safety Evaluation Report and Final Environmental Impact Statement,” Nuclear Regulatory Commission, May 22, 2012, 5, See also Kemp, “SILEX and Proliferation.”

8. Elaine M. Grossman, “Closely Held Report Discounts Proliferation Risk of Lasers for Making Nuclear Fuel,” Global Security Newswire, May 24, 2012,

9. For example, J. Carson Mark, “Explosive Properties of Reactor-Grade Plutonium,” Science & Global Security 4 (1) (1993): 111–128,; and Bruno Pellaud, “Proliferation Aspects of Plutonium Recycling,” Comptes Rendus Physique 3 (7–8) (2002): 1067–1079.

10. The only exception is plutonium consisting of more than 80 percent plutonium-238.

11. IAEA, IAEA Safeguards Glossary: 2001 Edition, International Nuclear Verification Series 3 (Vienna: IAEA, 2002), 23,

12. Thomas B. Cochran and Christopher E. Paine, The Amount of Plutonium and Highly-Enriched Uranium Needed for Pure Fission Nuclear Weapons (Washington, D.C.: Natural Resources Defense Council, 1995).

13. Robert Alvarez, “Managing the Uranium-233 Stockpile of the United States,” Science & Global Security 21 (1) (2013): 53–69,

14. For a discussion of how the definition of nuclear security is evolving, see Nuclear Security Governance Experts Group (NSGEG), Responsibility beyond Rules: Leadership for a Secure Nuclear Future (NSGEG, March 2013), 6–7,

15. Mark Hibbs, The Future of the Nuclear Suppliers Group (Washington, D.C.: Carnegie Endowment for International Peace, 2011), 23,

16. As of June 2013. Australian Safeguards and Non-Proliferation Office, Annual Report 2012–13 (Canberra: Australian Government, 2013), 39,

17. International Panel on Fissile Materials (IPFM), Global Fissile Material Report 2013: Increasing Transparency of Nuclear Warhead and Fissile Material Stocks as a Step toward Disarmament (Princeton, NJ: IPFM, 2013), 24–25,

18. In addition, a few countries, such as North Korea and India, have domestic suppliers.

19. On Iran, Pakistan, and North Korea, see International Institute for Strategic Studies, Nuclear Black Markets: Pakistan, A. Q. Khan and the Rise of Proliferation Networks: A Net Assessment (London: International Institute for Strategic Studies, 2007), chaps. 2–3. On India, see Joshua Pollack, “The Secret Treachery of A.Q. Khan,” Playboy (January/February 2012), On Brazil, see Mark Hibbs, “Ex-MAN Gas Centrifuge Expert Said Higher-Ups Helped Iraq,” Nuclear Fuel 23 (6) (March 23, 1998): 5.

20. Scott R. Kemp, “The Nonproliferation Emperor Has No Clothes,” International Security 38 (4) (April 2014): 39–78,

21. For a somewhat dated cost estimate of Iran’s enrichment program (which is scaled to military ends), see Thomas W. Wood, Matthew D. Milazzo, Barbara A. Reichmuth, and Jeffrey Bedell, “The Economics of Energy Independence for Iran,” Nonproliferation Review 14 (1) (March 2007): 92,

22. James M. Acton, “Nuclear Power, Disarmament and Technological Restraint,” Survival 51 (4) (August–September 2009): 105–108, 111–115.

23. Nuclear materials are radioactive and decay, but the relevant timescale is typically so long that decay can be ignored. Moreover, predicting how much of a bulk sample will decay in any given period of time is straightforward, with no meaningful uncertainty.

24. The IAEA’s safeguards goals are given in IAEA, IAEA Safeguards Glossary, 24. For a more comprehensive discussion, see Rudolf Avenhaus and Morton John Canty, Compliance Quantified: An Introduction to Data Verification (Cambridge: Cambridge University Press, 1996), chap. 7.

25. The IAEA’s Model Comprehensive Safeguards Agreement explicitly states that “the objective of safeguards is the timely detection of diversion of . . . nuclear material . . . and deterrence of such diversion by the risk of early detection.” 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 (corrected) (Vienna: IAEA, 1972), para. 28,

26. See, for example, John Mueller, Atomic Obsession: Nuclear Alarmism from Hiroshima to Al Qaeda (Oxford: Oxford University Press, 2010), 172–176; and Anna M. Pluta and Peter D. Zimmerman, “Nuclear Terrorism: A Disheartening Dissent,” Survival 48 (2) (Summer 2006): 55–69.

27. Lewis A. Dunn, Can Al Qaeda Be Deterred from Using Nuclear Weapons? Occasional Paper 3 (Washington, D.C.: Center for the Study of Weapons of Mass Destruction, National Defense University, July 2005), 8,

28. Wyn Q. Bowen, Matthew Cottee, and Christopher Hobbs, “Multilateral Cooperation and the Prevention of Nuclear Terrorism: Pragmatism over Idealism,” International Affairs 88 (2) (2012): 349.

29. Spurgeon Keeny, memorandum for Henry Kissinger, “Provisions of the NPT and Associated Problems,” January 24, 1969, 5, in U.S. Department of State Archive,

30. Bowen, Cottee, and Hobbs, “Multilateral Cooperation and the Prevention of Nuclear Terrorism,” 354–356; and Jack Boureston and Tanya Ogilvie-White, “Seeking Nuclear Security through Greater International Coordination” (working paper, Council on Foreign Relations, March 2010), 13–14,

31. Chester I. Barnard, J. R. Oppenheimer, Charles A. Thomas, Harry A. Winne, and David E. Lilienthal, A Report on the International Control of Atomic Energy (Washington, D.C.: U.S. Government Printing Office, March 1946), 4, available from

32. Ibid., 61.

33. A voluntary reporting scheme for separated neptunium and americium was established in 1999 but is apparently barely used.

34. On the meaning of the prohibition against manufacturing nuclear weapons, see James Acton with Carter Newman, IAEA Verification of Military Research and Development, Verification Matters 5 (London: VERTIC, 2006), 13–14,

35. For example, Christopher Ford, “Statement to the 2005 Review Conference of the Treaty on the Nonproliferation of Nuclear Weapons, New York, New York,” May 18, 2005, in U.S. Department of State Archive,

36. For a discussion of the IAEA’s legal authority, see IAEA, The Safeguards System of the International Atomic Energy Agency (Vienna: IAEA, n.d.), 4–6,

37. IAEA, IAEA Annual Report 2014 (Vienna: IAEA, 2015), 100,

38. These are Algeria, Argentina, Belarus, Brazil, Egypt, Iran, Malaysia, Serbia, Syria, Thailand, and Venezuela. Australian Safeguards and Non-Proliferation Office, Annual Report 2012–13, 106.

39. IAEA, IAEA Safeguards Glossary, 19.

40. A summary of the use of open-source information, satellite imagery, and intelligence information by the IAEA is provided in James M. Acton, “International Verification and Intelligence,” Intelligence and National Security 29 (3) (2014): 345–346, 348–350.

41. Mark Hibbs, “The Plan for IAEA Safeguards,” Carnegie Endowment for International Peace, November 20, 2012,

42. Jill N. Cooley, “Progress in Evolving the State-Level Concept” (paper presented at the 7th INMM/ESARDA Joint Workshop on Future Directions for Nuclear Safeguards and Verification, Aix-en-Provence, France, October 17–20, 2011), 4,

43. Ibid., 3.

44. Hibbs, “The Plan for IAEA Safeguards.”

45. IAEA, Implementation of the NPT Safeguards Agreement in the Syrian Arab Republic, GOV/2011/30 (Vienna: IAEA, May 24, 2011),

46. Fred Charles Iklé, “After Detection—What?” Foreign Affairs 39 (2) (January 1961): 208–220.

47. IAEA, Implementation of the NPT Safeguards Agreement in the Arab Republic of Egypt, GOV/2005/9 (Vienna: IAEA, February 14, 2005),; and IAEA, Implementation of the NPT Safeguards Agreement in the Republic of Korea, GOV/2004/84 (Vienna: IAEA, November 11, 2004),

48. The trigger list is circulated by the IAEA as INFCIRC/209. The most recent version is IAEA, Communication of 29 May 2014 Received from the Permanent Mission of Canada Regarding the Export of Nuclear Material and of Certain Categories of Equipment and Other Material, INFCIRC/209/Rev.3 (Vienna: IAEA, June 19, 2014),

49. The trigger list and dual-use list are circulated by the IAEA as, respectively, parts 1 and 2 of INFCIRC/254. For the most recent versions see IAEA, Communication Received from the Permanent Mission of the Czech Republic to the International Atomic Energy Agency Regarding Certain Member States’ Guidelines for the Export of Nuclear Material, Equipment and Technology, INFCIRC/254/Rev.12/Part 1 and INFCIRC/254/Rev.12/Part 2 (Vienna: IAEA, November 13, 2013), and

50. Hibbs, Future of the Nuclear Suppliers Group.

51. Kemp, “The Nonproliferation Emperor Has No Clothes,” 40.

52. Mark Hibbs, “Power Loop: China Provides Nuclear Reactors to Pakistan,” Jane’s Intelligence Review, January 2014, 52–53,

53. Fred McGoldrick, with Matthew Bunn, Martin Malin, and William H. Tobey, Limiting Transfers of Enrichment and Reprocessing Technology: Issues, Constraints, Options (Cambridge, Mass.: Project on Managing the Atom, Belfer Center for Science and International Affairs, Harvard Kennedy School, May 2011), 13–17,; and Mark Hibbs, “New Global Rules for Sensitive Nuclear Trade,” Nuclear Energy Brief, Carnegie Endowment for International Peace, July 28, 2011,

54. Hibbs, Future of the Nuclear Suppliers Group, 11.

55. U.S. Department of State, “Proliferation Security Initiative,” n.d.,

56. U.S. Department of State, “Ship Boarding Agreements,” n.d.,; and United Nations Conference on Trade and Development, Review of Maritime Transport 2014 (New York: United Nations, 2014), 44–45,

57. Nuclear Threat Initiative, “Proliferation Security Initiative,” 2015,

58. James R. Holmes and Andrew C. Winner, “The Proliferation Security Initiative,” in Nathan E. Busch and Daniel H. Joyner, eds., Combatting Weapons of Mass Destruction: The Future of International Nonproliferation Policy (Athens: University of Georgia Press, 2009), 148–150. Holmes and Winner also observe that even defining success is highly problematic.

59. Nuclear Power Plant Exporters’ Principles of Conduct, “Participants,” n.d.,

60. Nuclear Power Plant Exporters’ Principles of Conduct, March 6, 2014, 8,

61. David Fischer, History of the International Atomic Energy Agency: The First Forty Years (Vienna: IAEA, 1997), 229–230,

62. The most recent version is IAEA, Nuclear Security Recommendations on Physical Protection of Nuclear Material and Nuclear Facilities, IAEA Nuclear Security Series no. 13, INFCIRC/225/Revision 5 (Vienna: IAEA, 2011),

63. As of September 15, 2015. IAEA, Convention on the Physical Protection of Nuclear Material (Vienna: IAEA, September 15, 2015),

64. Fischer, History of the International Atomic Energy Agency, 120–122.

65. A slightly dated but comprehensive overview of efforts to mitigate this problem can be found in Frank von Hippel, “A Comprehensive Approach to Elimination of Highly-Enriched-Uranium from All Nuclear-Reactor Fuel Cycles,” Science & Global Security 12 (2004): 137–164,

66. For example, Boureston and Ogilvie-White, “Seeking Nuclear Security through Greater International Coordination”; Bowen, Cottee, and Hobbs, “Multilateral Cooperation and the Prevention of Nuclear Terrorism”; Centre for Science and Security Studies, King’s College London, Nuclear Security Briefing Book: 2014 Edition (London: Centre for Science and Security Studies, King’s College London, 2014),; and Nuclear Threat Initiative, Nuclear Security Primer: The Existing System (Washington, D.C.: Nuclear Threat Initiative, September 2014),

67. U.S. Department of Energy, National Nuclear Security Administration, The Four-Year Effort: Contributions of the Global Threat Reduction Initiative to Secure the World’s Most Vulnerable Nuclear Material by December 2013, YGG 13-0337 (Washington, D.C.: U.S. Department of Energy, 2013), 4,

68. Ibid., 5.

69. As of February 10, 2016, the amendment had been ratified by ninety-two states. For entry into force, two-thirds of states party to the treaty must ratify it. The current target is, therefore, 102—but this figure is liable to increase as more states ratify the convention. IAEA, Amendment to the Convention on the Physical Protection of Nuclear Material (Vienna: IAEA, February 10, 2016),

70. Although slightly dated, this is highlighted throughout Boureston and Ogilvie-White, “Seeking Nuclear Security through Greater International Coordination.”

71. Arshad Mohammed and Lidia Kelly, “Russia Told U.S. It Will Not Attend 2016 Nuclear Security Summit,” Reuters, November 5, 2014,

72. Bryan Bender, “Russia Ends U.S. Nuclear Security Alliance,” Boston Globe, January 19, 2015,

73. Matthew Bunn, Nickolas Roth, and William H. Tobey, Cutting Too Deep: The Obama Administration’s Proposals for Nuclear Security Spending Reductions (Cambridge, Mass.: Project on Managing the Atom, Belfer Center for Science and International Affairs, Harvard Kennedy School, July 2014),

74. Matthew Bunn, “U.S. Failure to Ratify Key Nuclear Security Conventions,” Nuclear Security Matters, March 12, 2014,

75. “Non-paper: Building International Confidence in the Security of Military Materials,” Nuclear Threat Initiative, September 17, 2014,; and The Hague Nuclear Security Summit Communiqué, The Hague, March 25, 2014, para. 4,

76. The communiqué from the 2014 Nuclear Security Summit went no further than to “encourage States . . . to keep their stockpile of separated plutonium to the minimum level . . . as consistent with national requirements.” The Hague Nuclear Security Summit Communiqué, para. 21.

77. Various examples are given in Boureston and Ogilvie-White, “Seeking Nuclear Security through Greater International Coordination,” 2–9.

78. Algeria et al., Joint Statement, March 25, 2014,

79. Nuclear Threat Initiative, NTI Nuclear Materials Security Index: Building a Framework for Assurance, Accountability, and Action, 2d ed. (Washington, D.C.: Nuclear Threat Initiative, January 2014), 14,

80. Lars Olberg, Reporting to the 1540 Committee—A Snapshot (New York: Lawyers Committee on Nuclear Policy, November 2005), 4,

81. Oh Joon, “Letter Dated 31 December 2014 from the Chair of the Security Council Committee Established Pursuant to Resolution 1540 (2004) Addressed to the President of the Security Council,” S/2014/958, United Nations Security Council, December 31, 2014, para. III.B.6,

82. These programs are likely to be significantly more effective at detecting the smuggling of radioactive material, which is another of their goals.

83. The most up-to-date version of the NRC’s portion of the Eligible Facilities List is available from NRC, “International Safeguards,” last updated March 26, 2015, The Department of Energy does not make its list public. However, it does state that, in total, the United States makes “nearly 300” facilities available. Given that the most recent NRC list (from 2013) includes 261 facilities, the Department of Energy list must include about thirty facilities. U.S. Department of Energy, National Nuclear Security Administration, “NPT Compliance,” n.d.,

84. IPFM, Global Fissile Material Report 2007 (Princeton, NJ: IPFM, 2007), 72,

85. NRC, “International Safeguards.”

86. The Department of Commerce also has a role in collecting information from private industry for the United States’ additional protocol declarations.

87. Ian F. Fergusson and Paul K. Kerr, The U.S. Export Control System and the President’s Reform Initiative, CRS Report for Congress, R41916 (Washington, D.C.: Congressional Research Service, January 13, 2014), 7, 26–27,

88. Todd Garvey, memorandum to Jeff Fortenberry, “Authority of the Nuclear Regulatory Commission to Require a Proliferation Risk Assessment as Part of a Uranium Enrichment Facility License Application,” Congressional Research Service, March 27, 2012, 4,

89. R. W. Borchardt, “Denial of Petition for Rulemaking (PRM-70-9)—American Physical Society,” SECY-12-0145, Nuclear Regulatory Commission, October 25, 2012, 4, 20,

90. Michael J. Brenner, Nuclear Power and Non-proliferation: The Remaking of U.S. Policy (Cambridge: Cambridge University Press, 1981), 100–113.

91. U.S. Department of Energy, National Nuclear Security Administration, Office of Nonproliferation and International Security, Draft Nonproliferation Impact Assessment for the Global Nuclear Energy Partnership Programmatic Alternatives (Washington, D.C.: National Nuclear Security Administration, December 2008),

92. Brenner, Nuclear Power and Non-proliferation, 84–88.

93. Fergusson and Kerr, U.S. Export Control System and the President’s Reform Initiative, 1.

94. This phrase and its definition, which is an interpretation of the act, is from “123 Agreements,”, December 3, 2010,

95. This phrase, which is an interpretation of the act, is from “810 Authorization,”, February 9, 2011,

96. This act is currently expired but enforced through a presidential declaration of a national emergency. Fergusson and Kerr, U.S. Export Control System and the President’s Reform Initiative, 2–3.

97. Ibid., 11.

98. For a discussion of the nuclear industry’s objections and the administration’s responses, see Mark Hibbs, “New and Balanced Rules for U.S. Nuclear Technology Exports,” Carnegie Endowment for International Peace, September 30, 2013,

99. Daniel Horner, “U.S. Policy on Nuclear Pacts Detailed,” Arms Control Today 44 (January/February 2014),

100. Quoted in Daniel Horner, “Vietnam Nuclear Pact Sent to Congress,” Arms Control Today 44 (June 2014),

101. For a more comprehensive summary of the most recent legislation, see Paul K. Kerr and Mary Beth D. Nikitin, Nuclear Cooperation with Other Countries: A Primer, CRS Report for Congress, RS22937 (Washington, D.C.: Congressional Research Service, December 3, 2015), 9,

102. One complication here is that a number of non-U.S. reactor designs use U.S. technology and so can only be purchased by a buyer with a 123 agreement in force. Exactly which reactors fall into this category is unclear, but the list is likely to shrink over time.

103. Thomas M. Countryman, Testimony of Assistant Secretary Thomas M. Countryman on Administration Policy Related to Agreements for Peaceful Nuclear Cooperation (123 Agreements), January 30, 2014, 3–4,

104. IPFM, Global Fissile Material Report 2007, 67–81.

105. South Sudan, which was founded only in 2011 and is not known to conduct any nuclear activities, has also not acceded to the treaty but is expected to do so in the not-too-distant future.

106. Problematically, the decision about which facilities were civilian and which were military was an exclusively Indian one. For a critique of the plan, see Zia Mian, A. H. Nayyar, R. Rajaraman, and M. V. Ramana, Fissile Materials in South Asia: The Implications of the U.S.-India Nuclear Deal, Research Report no. 1 (Princeton, NJ: IPFM, September 2006), 16–24,

107. Pierre Goldschmidt, “The IAEA Reports on Egypt: Reluctantly?” Carnegie Endowment for International Peace, June 2, 2009,

108. Mark Hibbs, “Safeguards in the Spotlight,” Arms Control Wonk (blog), June 9, 2013,

109. For a trenchant critique of the process, see William Walker, Nuclear Entrapment: THORP and the Politics of Commitment (London: Institute for Public Policy Research, 1999), 13–27.

110. The differences highlighted in the following paragraphs are drawn from James A. Glasgow, Elina Teplinsky, and Stephen L. Markus, Nuclear Export Controls: A Comparative Analysis of National Regimes for the Control of Nuclear Materials, Components and Technology (Washington, D.C.: Pillsbury Winthrop Shaw Pittman LLP, October 2012), This study was commissioned by the Nuclear Energy Institute, a lobbying organization for the nuclear industry, to highlight ways in which U.S. export controls are more burdensome than the export controls of American competitors. Although no evidence is presented to substantiate its assertions about potential buyers’ concerns about the U.S. export control system, its purely factual analysis of the differences between states appears to be reliable and carefully researched.

111. Ibid., 51.

112. Ibid.

113. Dipanjan Roy Chaudhury, “Prime Minister Narendra Modi’s Japan Visit May Not Seal Civil Nuclear Deal,” Economic Times, August 30, 2014,

114. Peter Crail, “Measuring Nuclear Export Controls in Nuclear-Powered Nations and Nuclear Aspirants,” in A Collection of Papers from the 2010 Nuclear Scholars Initiative, ed. Mark Jansson (Washington, D.C.: Center for Strategic and International Studies, 2010), 80–87, available from

115. U.S. Department of Commerce, Bureau of Industry and Security, “Deemed Exports,” n.d.,

116. Nuclear Threat Initiative, NTI Nuclear Materials Security Index.

117. Author’s calculations based on data from Nuclear Threat Initiative, NTI Nuclear Materials Security Index.

118. Although only three states lack an independent regulator, the spread of scores is large because this indicator is a binary variable. Whether this should be considered a bug or a feature of this approach is up for debate.

119. Togzhan Kassenova, From Antagonism to Partnership: The Uneasy Path of the U.S.-Russian Cooperative Threat Reduction (Stuttgart: ibidem Press, 2007), 178, 189, 209; and Matthew Bunn and Scott D. Sagan, A Worst Practices Guide to Insider Threats: Lessons from Past Mistakes (Cambridge, Mass.: American Academy of Arts and Sciences, 2014), 9, 11, 14,

120. Bunn and Sagan, A Worst Practices Guide to Insider Threats, 11.

121. Douglas Birch, R. Jeffrey Smith, and Jake Adelstein, “Japan Could Be Building an Irresistible Terrorist Target, Experts Say,” Center for Public Integrity, March 11, 2014,

122. The one possible exception is Obama’s renewed focus on nuclear disarmament. However, this initiative seeks to enhance the nonproliferation regime indirectly, by using disarmament to motivate a coalition willing to strengthen nonproliferation rules. Besides, it has now been largely abandoned by the White House.

123. George W. Bush, remarks, Washington, D.C., February 11, 2004,

124. For a critique of the new rules, see Hibbs, “New Global Rules for Sensitive Nuclear Trade.”

125. The centrifuge technology Urenco developed is now owned by the Enrichment Technology Corporation, which since 2006 has been jointly owned by the British, Dutch, and German governments and France’s Areva.

126. Mohamed ElBaradei, “Towards a Safer World,” Economist, October 16, 2003,

127. For a summary of some of these proposals, see Mary Beth Nikitin, Anthony Andrews, and Mark Holt, Managing the Nuclear Fuel Cycle: Policy Implications of Expanding Global Access to Nuclear Power, CRS Report for Congress, RL34234 (Washington, D.C.: Congressional Research Service, October 19, 2012),; and Yury Yudin, Multilateralization of the Nuclear Fuel Cycle: Assessing the Existing Proposals (Geneva: United Nations Institute for Disarmament Research, 2009),

128. The other initiatives that are being or have been implemented are “fuel banks” (actually uranium hexafluoride banks) to cater to states that suffer a politically motivated disruption of fuel supply. One has been set up by Russia, which pays for and operates this facility, but the responsibility for releasing material lies with the IAEA. Second, the IAEA has approved and is in the process of establishing a similar facility in Kazakhstan. Much of the funding for this facility came from the Nuclear Threat Initiative and the U.S. government.

129. IAEA, Fundamental Safety Principles, Safety Fundamentals SF-1 (Vienna: IAEA, 2006), 10,

130. European Council Directive 96/29/EURATOM, May 13, 1996, title IV, chap. I, art. 6,

131. NSGEG, Responsibility beyond Rules, 14.

132. Boureston and Ogilvie-White, “Seeking Nuclear Security through Greater International Coordination,” 11.

133. McGoldrick et al., Limiting Transfers of Enrichment and Reprocessing Technology, 13–17.

134. Manucher Mottaki, Ahmet Davutoğlu, and Celso Amorim, Joint Declaration by Iran, Turkey and Brazil, May 17, 2010, para. 10,

135. Yury Yudin, Multilateralization of the Nuclear Fuel Cycle: The Need to Build Trust (Geneva: United Nations Institute for Disarmament Research, 2010), 10–11,

136. Bowen, Cottee, and Hobbs, “Multilateral Cooperation and the Prevention of Nuclear Terrorism,” 357. The phrase “inalienable right” is from article IV of the NPT.

137. Deepti Choubey, Are New Nuclear Bargains Attainable? (Washington, D.C.: Carnegie Endowment for International Peace, 2008),

138. Although slightly dated, a trenchant discussion with a useful agenda for moving forward is Christopher F. Chyba, “Time for a Systematic Analysis: U.S. Nuclear Weapons and Nuclear Proliferation,” Arms Control Today 38 (December 2008),

139. Togzhan Kassenova, Brazil’s Nuclear Kaleidoscope: An Evolving Identity (Washington, D.C.: Carnegie Endowment for International Peace, 2014), 60–62,

140. Trevor Findlay, Nuclear Energy and Global Governance: Ensuring Safety, Security and Non-proliferation (Abingdon: Routledge, 2011), 145.

141. For details of the package, see Sebnem Arsu, “Turkey’s Pact with Russia Will Give It Nuclear Plant,” New York Times, May 12, 2010, For the background, see Center for Strategic and International Studies, Economic Policy Research Foundation of Turkey, and Institute of Oriental Studies of the Russian Academy of Sciences, The Turkey, Russia, Iran Nexus: Economic and Energy Dimensions—Proceedings of an International Workshop, Ankara, March 29, 2012 (Washington, D.C.: Center for Strategic and International Studies, 2012), 9,

142. For the case that the deal was indeed bad for nonproliferation, see George Perkovich, “Global Implications of the U.S.-India Deal,” Dædalus 139 (1) (Winter 2010): 20–31.

143. William Walker, “Nuclear Enlightenment and Counter-Enlightenment,” International Affairs 83 (3) (May 2007): 433.