Security Assurances and Nonproliferation

Colorado Springs, CO, 13-14 August 2009

Conference organized by the Center for Contemporary Conflict, Naval Postgraduate School, and supported by the Defense Threat Reduction Agency

By Jeffrey W. Knopf

About 20 participants met for a day and a half at an invitation-only workshop to discuss the role of security assurances in promoting nuclear nonproliferation. The workshop had three broad objectives:

• To clarify the concept of security assurances and related terminology;
• To assess the effectiveness of security assurances and to develop generalizations about the conditions most likely to affect the success or failure of assurances; and
• To explore the impact of security assurances in preventing or reversing nuclear proliferation and develop recommendations for how to maximize the effectiveness of security assurances as a nonproliferation tool.

On the first day of the workshop, Dr. Michael Wheeler, the director of Advanced Systems and Concepts Office of the Defense Threat Reduction Agency, welcomed workshop participants and introduced the conference. Two panels then followed, one on the theory behind security assurances and the other on the history of security assurances in relation to nuclear nonproliferation.

Wyn Bowen of Kings College, London (seated) and Michael Wheeler of DTRA (speaking)

The Theory Behind Assurances

In the first panel, Jeffrey Knopf of the Naval Postgraduate School (NPS) noted that the terms assurance, reassurance, and assurances all get used to describe strategies employed by states. He sought to clarify the different ways in which these terms are used. Generically, all these concepts have in common the idea that it is sometimes valuable to take measures to persuade other states that their security will not be harmed. Knopf then reviewed relevant findings from research on deterrence and on nuclear proliferation to identify fourteen hypotheses about the conditions under which security assurances are likely to succeed or fail in influencing state decisions about acquiring nuclear weapons. One hypothesis suggested that security concerns might need to be part of a state’s motivations for assurances to be relevant. Several others dealt with factors that could affect the credibility of assurances. Another hypothesis suggested that assurances might gain impact primarily through how they affect domestic political debates. And other hypotheses indicated that assurances might work better when paired with other strategies such as positive incentives.

Janice Gross Stein, from the University of Toronto, reviewed recent research in psychology and neuroscience that is relevant to the use of assurance strategies. She noted that recent findings have emphasized the impact of emotions on human decision-making. While this creates many complications for using assurances, Stein pointed out that the importance of fear as an emotion might also make assurances an important strategy. Stein also drew attention to importance of honoring past promises if states wish to maintain influence.

Erik Dahl of the Naval Postgraduate School, Janice Stein of University of Toronto, and Scott Snyder of the Center for U.S.-Korea Policy

In discussion following the panel, one participant pointed out that factors associated with the state providing assurances might also be a factor in their effectiveness. Much discussion focused on factors that might limit or blunt the impact of emotions.

The History of Assurances

In the second panel, John Simpson of the University of Southampton provided a history of security assurances in relation to the Non-Proliferation Treaty (NPT). There are both positive and negative assurances: positive security assurances (PSAs) are promises to assist non-nuclear weapon states (NNWS) if they are the victim of nuclear threats or attack; negative security assurances (NSAs) are promises by nuclear weapons states (NWS) not to threaten or use their own nuclear weapons against NNWS. Although both types of assurance were requested by NNWS, no formulation could be agreed upon for incorporating them into the NPT. Hence, assurances have been provided in a variety of ways outside the NPT text. Some of the strongest involve protocols to regional nuclear-weapon free zones (NWFZs). NNWS, however, remain dissatisfied and continue to seek more uniform and legally binding assurances. Professor Simpson noted that the end of the Cold War greatly changed the context, as does the existence of four non-NPT nuclear weapon states. Given renewed attention to disarmament and the Middle East, Simpson predicted that security assurances would not be a central issue at the 2010 NPT Review Conference.

Kerry Kartchner of the U.S. State Department summarized existing U.S. extended deterrence commitments and other security guarantees. According to State Department analysis, the number of states covered by U.S. assurances is far greater than the number of 30 that is often cited. Between defense treaties such as NATO and the U.S.-Japan alliance and nonproliferation commitments, including NWFZ protocols signed by the United States, nearly every country is now covered by some form of U.S. security assurance.

Bruno Tertrais of the Fondation pour la Recherche Strategique in Paris presented a preliminary assessment of the importance of security assurances based on a survey of nearly all countries that have considered or completed a nuclear weapons program. He concluded that PSAs in the form of defense pacts that provide a “nuclear umbrella” to non-nuclear allies had been effective in dissuading proliferation, as evidenced in part by the fact that many proliferators had lacked a credible security guarantee. Tertrais concluded that vaguely worded or multilateral PSAs make little difference, and he argued that NSAs also have little impact. Tertrais observed, as did several other papers, that there is a tension between PSAs and NSAs, as the former implies an extended nuclear deterrent commitment while the latter implies nuclear weapons will not be used. Because of various dilemmas involved with bilateral security guarantees, Tertrais concluded, even this instrument must be used with caution.

Bruno Tertrais of Fondation pour la Recherche Stratégique, Kerry Kartchner of the U.S. State Department, Mike Keifer of DTRA, and Jim Walsh of MIT

Michael Wheeler served as discussant for this panel. He pointed out the complications that can arise in practice, such as changes in leadership or a crisis involving a third party, which make it hard to assess the impact of security assurances in empirical cases. Other participants debated whether assurances must be formalized in a written document to be effective, or informal, verbal assurances might sometimes work better.

Day two of the conference involved panels to discuss specific cases in three regions of the world: the Middle East, East Asia, and Europe.

Lessons of Cases from the Middle East

In the Middle East panel, Wyn Bowen of Kings College, London, discussed Libya’s nuclear rollback. He concluded that a negative assurance from the United States and a positive assurance from Britain had both played an important role in convincing Libya to renounce WMD. In particular, a promise from the United States not to seek regime change in Tripoli was essential.

Michael Kraig of the Stanley Foundation discussed security assurances provided to the Gulf Arab monarchies. He noted that the Gulf Cooperation Council (GCC) states benefit from many forms of security assistance, including extensive arms sales and military bases. Nevertheless, these states remain wary of Iran, while security assistance to the Gulf states also makes Iran feel more insecure. The GCC states are also reluctant to rely too heavily on the United States because this is unpopular domestically. Kraig proposed adopting a much broader approach to assurances in the Gulf, incorporating economic and energy issues and an increased role for middle powers from outside the region such as India.

James Walsh of MIT addressed the case of Iran. He covered both the period under the Shah and the Islamic Republic (IRI). He concluded that security assurances had not had much impact in this case. During the Shah’s time, alliance commitments weren’t seen as credible. Since the Iranian revolution, the IRI has not been offered many assurances. Walsh suggested that assurances might not be effective even if offered, because Iran is not focused primarily on external security threats. Rather, it is most concerned with internal threats, such as a democratic “color” revolution, and considerations of national pride are also an important driver of Iran’s nuclear ambitions.

Etel Solingen from the University of California-Irvine served as discussant for this panel. She suggested that all the papers could have delved more deeply into domestic political processes that determine how security assurances are received. Much of the subsequent discussion focused on different interpretations of what assurances had or had not been offered to the states in question. One participant pointed out that Libya had faced less external threat than Iran but had given up its nuclear program, raising the question of what accounts for the difference.

Lessons of Cases from East Asia

In the session on East Asia, John Park of the U.S. Institute of Peace presented via speakerphone a case study of North Korea. He noted that the DPRK had lost previous PSAs provided by the Soviet Union and China. It then perceived an increased threat from the United States after President Bush included it in the “axis of evil.” The U.S. later offered an NSA in the context of six-party talks, but this has not been sufficient to deflect North Korea from continuing its nuclear weapons program. Park concluded that only a larger package deal has a chance of being effective with North Korea.

Major Jungsoo Kim of the ROK Air Force, currently a Ph.D. student at NPS, presented research on South Korea’s efforts to reassure North Korea starting with the “Sunshine Policy” of President Kim Dae Jung. Major Kim demonstrated that alliance constraints had often handicapped this effort. For much of the time when South Korea sought to pursue a reassurance strategy, its U.S. ally was not supportive. Ironically, when the Bush administration became interested in reassurance late in its second term, South Korea had a newly elected conservative president who no longer supported such an effort.

Scott Snyder of the Asia Society presented a case study of South Korea, co-written with Joyce Lee. He observed that South Korea had received extensive positive assurances via the U.S.-ROK alliance, but the South had displayed recurring doubts about the credibility of the U.S. security guarantee, thereby requiring the United States to take repeated steps to reinforce its assurances. For example, after the Guam Doctrine announced by President Nixon, South Korea had begun exploring a nuclear weapons option, and it took a combination of pressure tactics and reassurances from the United States to halt the South Korean effort. Snyder pointed out that in recent years, when U.S. policies have been less popular with South Korean public opinion, assurances have had to run in the reverse direction, with ROK leaders assuring the United States of their continued commitment to the alliance. Both Snyder and Park observed that U.S. PSAs to South Korea can work at cross-purposes with NSAs to the North, and it is important to find a way to de-conflict these.

Yuki Tatsumi of the Stimson Center discussed the case of Japan. She found U.S. security guarantees to be important, but with their significance varying over time. They were important in getting Japan to commit initially to join the NPT as a NNWS. Thereafter, they were less important until recently, when North Korean nuclear activities again made Japan concerned about the strength of the U.S.-Japan alliance. Security assurances are not solely responsible for Japan’s non-nuclear status. The “nuclear allergy” due to Japan’s past experiences also played a major role, but Tatsumi observed that the taboo in Japan against discussing nuclear weapons has weakened in recent years.

Christopher Twomey of NPS served as discussant for the East Asia papers. Like Solingen in the previous session, he expressed interest in seeing the domestic politics of nuclear decisions explored more fully. He also noted that non-declaratory aspects of policy, such as when the United States moved its troops away from the DMZ in Korea, might have as much impact as the formal, declaratory elements of a security assurance. Further discussion in this session drew added attention to the importance of harmonizing security assurances. For example, Japan will be interested in anything promised to South Korea and vice versa. One participant also noted that these cases raise the question of whether assurances are sought more for their military or their political implications.

Jim Walsh of MIT, Chris Twomey of NPS, Maj. Jungsoo Kim of NPS, John Simpson of University of Southampton, and Jeffrey Fields of DTRA

Lessons of Cases from Europe

The final session dealt with two cases from Europe. Thomas Jonter of Stockholm University addressed the case of Sweden. This country explored a nuclear weapons program in the 1950s and 1960s before renouncing the nuclear option. Jonter noted that it is widely believed that Sweden, despite its policy of non-alignment, received a secret security guarantee from the United States, but after extensive archival research he could find no evidence that such a guarantee was ever officially conveyed. Instead, he believes, informal suggestions from U.S. to Swedish military officers may have given the Swedish government reason to believe it had a U.S. PSA. This was important but not alone decisive in Sweden’s decision to join the NPT. Desire to conform to emerging nonproliferation norms and Sweden’s dependence on U.S. technical assistance for its civilian nuclear energy program also played roles.

Sherman Garnett from Michigan State discussed the case of Ukraine. This case is unusual because, after the breakup of the Soviet Union, Ukraine was “born nuclear” rather than having to make a decision to pursue nuclear weapons. Garnett concluded that security assurances played a major role in convincing Ukraine to give up these weapons. Assurances were only effective because they were part of a larger strategy also involving diplomatic engagement and positive economic incentives. Internal politics in the sender state were important here, because this case required the United States to be creative in innovating a broader approach than had been used previously.

Jim Wirtz from NPS served as discussant for the final panel. He noted that Sweden might not have required a formal security guarantee. Because of its geographic location, it could safely assume that NATO would respond in the event of a Soviet attack on Sweden. In discussion of this panel’s papers, some participants explored potential non-traditional forms of assurance, such as whether IAEA safeguards might be a type of security assurance.

Conference Findings

Overall, the workshop made it clear that evaluating the impact of security assurances in empirical cases is quite tricky. The findings across cases were mixed. Assurances appear to have been very effective in some cases while having little impact in others. PSAs, especially via bilateral defense pacts, played a more prominent role than NSAs or multilateral PSAs, though this is quite possibly an artifact of the cases selected for study.

Several of the hypotheses suggested by Knopf in the introductory paper received some support, though none received consistent support. States with major security concerns were the most likely to seek assurances, but also the hardest to convince of the credibility of those assurances. Indeed, whether or not assurances were perceived as credible was important in most of the cases, with the emphasis on perceptions – they did not always match what would seem to be objective indicators of credibility. In several cases, whether or not assurances were packaged into a broader strategy involving political recognition and/or positive incentives was a major factor in the outcome. Both domestic politics and alliance dynamics played important roles in many cases, though often in ways that were idiosyncratic and hard to generalize about. Both norms and emotions, as highlighted by Stein, proved significant in several cases. In particular, states cared about whether or not they were being treated fairly and whether the providers of assurances kept their prior promises. For these reasons, process often mattered as much as substance. It is not just a matter of what assurances are offered, but that states believe their concerns are being taken seriously and their independence respected.

These findings suggest that security assurances are not on their own likely to be the decisive factor in preventing nuclear proliferation, but as one tool used in combination with others they can nevertheless make a positive contribution to combating the spread of nuclear weapons. One of the trickiest issues moving forward will be finding ways to reduce the potential contradictions between positive and negative incentives so that the two can be used together synergistically.