The Origins of Nuclear Weapons Free Zones: Security Communities or Substitutes for a Nuclear Umbrella?

Transcription

1 The Origins of Nuclear Weapons Free Zones: Security Communities or Substitutes for a Nuclear Umbrella? Atsushi TAGO Graduate School of Law Kobe University Japan tago@dragon.kobe-u.ac.jp Word count: 7375.

2 Abstract As early as 1958, the Polish government proposed establishing a nuclear weapons free zone in Central Europe. Today, there are at least five nuclear free zones in the world. This article investigates the key determinants of nuclear weapons free zone treaty signing and ratification using survival analysis. It turns out that conventional threat level for a state and alliance ties with one of the five official nuclear weapons states significantly influence nuclear free zone membership while regime type does not. These results suggest that nuclear free zones are not necessarily security communities formed by states with similar political institutions, nor are they an alternative to the nuclear umbrella of the nuclear weapons states. Rather, for joining or forming nuclear weapons free zones, what matters is a relative absence of security concerns.

3 Introduction Nuclear weapons free zones, which ban the test, use, manufacture, or acquisition of nuclear explosive devices within a certain regional limit, cover large portions of the globe. There are at least five nuclear free zones in the world. The first zone was created after the Cuban Missile Crisis in Latin America. The treaty of Tlatelolco was signed in Mexico City in 1967 and came into force in As of 10 June 2007, thirty-three Latin American states participate in the regime. The treaty of Rarotonga in 1985 was the second nuclear free zone and it covers sixteen states and territories in the South Pacific. In December 1995, ten ASEAN (Association of South-East Asian Nations) members signed the Bangkok treaty and established a zone that covers Southeast Asia. A few months later, in April 1996, African states also agreed to set up a nuclear free zone on the continent; the Pelindaba treaty will come into force when the twenty-eighth state deposits an instrument of ratification with the depositary. Most lately, in September 2006, the five Central Asian states reached an agreement to establish a nuclear weapons free zone at Semipalatinsk, Kazakhstan. 1 These five nuclear free zones may be merely symbolic and have very limited substantive effects in international relations. Nevertheless, the zones are endorsed by a number of member states. It is in fact widely believed that these zones can promote the security of non-nuclear weapon states both by obtaining pledges from nuclear weapon states regarding the non-use of 1 The President of Mongolia also announced before the 47th session of the United Nations General Assembly in 1992 that Mongolia s territory would be a nuclear weapons free zone. Some categorize this as a nuclear weapons free zone, though it only contains one country. 1

4 nuclear weapons against members and by discouraging or preventing the deployment of nuclear weapons within the affected regions (Leigh-Phippard 1993: 93). Indeed, if nuclear weapons zones have even some of these beneficial effects, it is puzzling why more states lacking nuclear capabilities do not join the free zones. There have been proposals to establish a zone in the Middle East, South Asia, the Balkans, and the Korean peninsula. However, none have become a reality, perhaps because forming a nuclear free zone also involves some costs. Skeptics might well ask the opposite question. Why are so many countries signatories to nuclear weapon free zones when these treaties are bound to be less effective in deterring threats (particularly nuclear threats) than a nuclear umbrella provided by a major power, such as the United States? Japan, for example, the only state attacked by nuclear weapons twice, benefits from a US nuclear deterrent that has helped to maintain Japanese security for more than six decades. Why do some states ally with nuclear states to obtain a nuclear umbrella while other states do the opposite, becoming a member of a nuclear weapons free zone? This study answers these questions through the use of a large-n dataset and by applying the statistical technique of survival analysis. The rest of the article proceeds as follows. The second section provides an overview of the evolution of nuclear weapon free zones. I detail which proposals were realized, and when agreements were ratified. In the third section, I elaborate three competing hypotheses to explain the origins of nuclear weapons free zones. The fourth section explains the research design, followed by the fifth section that reports details of the results of the statistical analyses. The last section provides my concluding thoughts. 2

5 The Evolution of Nuclear Weapons Free Zones: A Short History The origin of the concept of a nuclear weapons free zone can be traced back as early as 1958, when the Polish government announced the Rapacki Plan, named after its foreign minister (Ozinga 1989; Goldblat 1997). The proposal was aimed to prevent the nuclearization of West Germany and the deployment of Soviet nuclear weapons on Polish territory. The proposed zone was to comprise Poland, Czechoslovakia, East Germany and West Germany. Of course, the proposal had no chance of becoming a subject of substantial international negotiation under Cold War conditions but several of its elements were later realized elsewhere in the world. In February 1967, fifteen Latin American states signed the Tlatelolco treaty after over three years of intense negotiations. The treaty prohibits the development or production of nuclear weapons in the region and also bans the receipt or installation of nuclear weapons by any Latin American country (Redick 1975: 416). The agreement also gives OPANAL (the Agency for the Prohibition of Nuclear Weapons in Latin America) the authority to monitor compliance with treaty obligations. 2 Member states are also required to reach an agreement with the IAEA (International Atomic Energy Agency) for the application of safeguards to the country s nuclear activities. The purpose of the regime is to promote regional security by making the five major powers pledge not to test, store, or use nuclear weapons. Indeed, the United States agreed to abide 2 An amendment of August 1992 determines the IAEA as the sole authority to carry out special inspections of the treaty parties. 3

6 by the nuclear free zone for treaty members but did not agree to similar conditions for NPT (Non-nuclear Proliferation Treaty) members (Bunn 1997: 5). Following the establishment of the nuclear free zone in Latin America, Iran under the Shah, with Egypt s support, became the first sponsor of a nuclear weapon free zone in the Middle East. Iran proposed a nuclear free zone in the 1974 plenary session of the United Nations General Assembly. Israel abstained from voting on the resolution, but it was adopted anyway. In 1975, the Assembly further adopted the resolution titled Complementary Study of the Question of Nuclear-Weapon-Free Zones in All Its Aspects. 3 The resolution provides us with a now widely accepted definition of nuclear weapon free zone. For instance, it states that the zone arrangements must ensure the region would be effectively free of all nuclear weapons and establish a reliable verification system to ensure compliance with agreed upon obligations. On the other hand, at the South Pacific Forum meeting in Tonga in July 1975, New Zealand proposed establishing a nuclear weapon free zone as a means of keeping the region free from the risk of nuclear confrontation (Thakur 1985: 218). However, a change in leadership in the New Zealand government led to a failure of the proposal. Ten more years were needed until the treaty of Rarotonga was signed in 1985, and one more year passed before the treaty came into effect. In South Asia, a nuclear weapon free zone was proposed several times from the 1960s (Jones & McDonough, 1998: 303). India, which was (not officially but actively) seeking nuclear weapons acquisition, strongly opposed the proposals. These proposals never gathered momentum. 3 UNGA resolution 3472 (XXX) of 11 December

7 In the Middle East, after Iran s first suggestion, even though the details differ, Egypt and Israel have each proposed establishing a non-proliferation framework in the region. However, deep division of opinion has blocked a realization of these proposals (Steinberg 1994: 141). For instance, meetings of the ACRS (Arms Control and Regional Security) working group in the context of the Madrid peace process after the Persian Gulf War failed to incorporate the issue of a nuclear weapons free zone for the region into the agenda (Maoz 2006: 305). In Southeast Asia, on the other hand, the Declaration of Kuala Lumpur was signed in 1971, which defined the determination of the ASEAN states to assure recognition and respect for a Zone of Peace, Freedom and Neutrality (ZOPFAN) in the region. Afterwards, in December 1995, the seven states of the ASEAN (Brunei, Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam), as well as Laos, Cambodia, and Myanmar signed the treaty creating a nuclear weapon free zone. The treaty came into force on March 27 when deposit of the seventh instrument of ratification was completed by Cambodia. In Africa, South Africa secretly developed nuclear weapons and this prevented the realization of any nuclear free zone proposals during the Cold War. It was not until 1991, when South Africa became a member of the NPT, that prospects for establishing a nuclear weapon free zone in Africa became a real possibility. There are fifty-five independent states on the continent with the potential to become members of the nuclear free zone. The treaty was opened for signature on April 12, So far, fifty states have signed the treaty and twenty of these states have deposited the instrument of ratification (as of September 2006). Ratification by eight more 5

8 countries is needed for the treaty to come into force. In Central Asia, the five states in the region --- Kazakhstan, Kyrgyzstan, Turkmenistan, Tajikistan, and Uzbekistan --- singed the Central Asia Nuclear-Weapon-Free Zone Treaty in September Thousands of Soviet nuclear weapons were once stockpiled in the region but the treaty will prevent the reintroduction of the weapons into the zone by Russia, or by the United States, which now maintains air bases in Kyrgyzstan and Uzbekistan. The treaty will come into force when all of the five states deposit the instrument of ratification. Finally, some may consider the Antarctic treaty to be the first nuclear weapon free zone in the world. In 1959, the uninhabited Antarctic continent was declared a demilitarized zone as well as a zone free from nuclear weapons. Article V of the treaty says that any nuclear explosions in Antarctica and the disposal there of radioactive waste material shall be prohibited. Indeed, it is useful to note that nuclear weapons states established the regime as original parties to the convention. The Outer Space Treaty (1967) and Sea-Bed Treaty (1971) are similar multilateral agreements aimed at stemming the spread of nuclear weapons in specific areas and environments. Here, however, one important difference exists between these multilateral conventions and the five free zones (Latin America, South Pacific, Southeast Asia, Africa and Central Asia); the latter zones are designed 1) to substitute for a nuclear umbrella (i.e. formal alliance ties with the five nuclear weapon states invite nuclear retaliation in the event that a state is attacked by other nuclear weapon states), and 2) to deter regional horizontal proliferation. Previous studies have considered the possibility that the five nuclear weapon free zones 6

9 promote the territorial security of a given region by obtaining pledges from nuclear weapon states not to use nuclear weapons against signatories (Leigh-Phippard 1993; Bunn 1997). Such agreements could potentially substitute for the protection nuclear powers such as the United States or the USSR. For instance, in establishing the nuclear free zone for the South Pacific, New Zealand had to choose whether it prefers a nuclear weapon free zone or the ANZUS alliance to promote its security (Thakur 1985; Power 1986). In contrast, Japan, which for obvious reasons employs an antinuclear security and foreign policy, could not utilize the framework of a nuclear weapon free zone for attaining nuclear disarmament since it has a very strong alliance relationship with the US and is protected under the US nuclear umbrella. 4 The ROK (South Korea) may be facing a similar dilemma. The ROK and DPRK (North Korea) signed the Joint Declaration on the Denuclearization of the Korean Peninsula in They agreed to refrain from the testing, manufacture, production, acceptance, possession, stockpiling, deployment, or use of nuclear weapons, and to use nuclear energy only for peaceful purposes. However, the ROK cannot banish nuclear weapons from the peninsula on its own. It must consult with the United States and persuade Washington (possibly even causing deterioration of its alliance ties) to remove US nuclear warheads from American bases in South Korea. So far, unfortunately, only a very limited number of scientific studies seek to answer questions about why some states prefer alliances with nuclear weapon states to obtain a nuclear 4 Japan has the three non-nuclear principles, requiring Japan not to manufacture, possess, nor allow the deployment of nuclear weapons on Japanese territory. 7

10 umbrella, while other states prefer nuclear weapons free zones to obtain a promise of no use. Fuhrmann & Li (2007) are one of the few exceptions, offering an analysis of nuclear weapons free zones. However, they only study determinants of the timing of ratification of a nuclear weapons free zone treaty. The authors ignore the determinants of treaty signing. The timing of ratification would be simply delayed by situational domestic reasons (for instance, coups and civil wars, which stop domestic treaty-ratification processes; minority governments also lack control of a parliament). Thus, it may not be the best measure of states decisions in forming a nuclear weapon free zone in a region and becoming a member of it. In contrast, signing can be decided solely by the chief executive, and this means that it will genuinely reflect which systematic factors accelerate/delay the timing to form (and be a member of) nuclear weapons free zones. At the very least, there is no reason to ignore a decision vis-à-vis signature. But before explaining how I will answer the question, let me introduce the competing hypotheses. Possible Explanations for Nuclear Weapons Free Zones Countries have three options in pursuing security in the nuclear era. First, states can develop nuclear weapons of their own. Successful proliferation requires advanced technology, and involves a large economic burden and considerable delay. A state must also endure a price in terms of its political reputation associated with the nuclear taboo (Tannenwald 1999). For these reasons, recent studies show that nuclear weapon development requires some strong international and domestic motives (Singh & Way 2004; Jo & Gartzke 2007). It would not be an easy choice for 8

11 most states in the international community to develop nuclear weapons. The second option for a state is to form a formal security alliance with an existing nuclear weapons state. The US political commitment to its allies in East Asia, for example, has been and remains robust, bolstered by the US military presence (with its nuclear arsenal) in Japan and South Korea for the past 60 years. This deterrent appears to have continued to be effective despite the reduced number of US troops in the Asian theater in the 1990s. It is very likely that East Asian allies will continue to rely on the American nuclear umbrella well into the future, rather than risk precipitating a regional nuclear domino effect (McInnis 2005: 170). However, states with nuclear allies face a conflict of interest in joining a nuclear weapons free zone. There is of course a moral and philosophical contradiction if a state says yes to being protected by a nuclear deterrent, while also saying no to having nuclear weapons positioned on its soil. There is also a strategic contradiction in relying on nuclear weapons to deter but denying their presence, since credibility is a critical variable in deterrence effectiveness. An ally is much more likely to use weapons that are in theatre in the event of a war. Thus, joining a nuclear weapons free zone risks weakening the credibility of strategic nuclear alliances. The third option for a state is to establish a nuclear weapon free zone over its territory and obtain a promise not to use nuclear weapons against the zone from all nuclear weapons states. The nuclear free zone regime would entail an inspection and verification mechanism designed to verify the regional nuclear free status and to deter attempts to proliferate by states within the agreement. A nuclear free zone would be the least reliable option from a realist s point of view, 9

12 which denies the legitimacy and effectiveness of international treaties. On the other hand, the zones involve the lowest material cost of all three options for promoting security in the nuclear era. Given these three options, I argue that a state would choose a nuclear weapons free zone over nuclear proliferation or nuclear alliance under the following conditions: I. Nuclear Weapons Free Zones are for Relatively Secure States Jo & Gartzke (2007) find that security concerns are important determinants of whether states form nuclear weapons programs and whether states attempting to proliferate actually succeed at producing nuclear weapons. The presence of conventional threats increases the relative risk of a state beginning a program by 164% and the risk of acquisition of nuclear weapons by 58% (Jo & Gartzke 2007: 182). However, the same mechanism does not work for nuclear threats. States that face rivals with nuclear weapons or nuclear programs tend to refrain from deepening nuclear proliferation. A similar but not identical result appears in Singh & Way (2004). Conventional security threats should lead to a lower probability of establishing or joining a nuclear weapons free zone. A state facing severe security concerns is likely to consider a nuclear weapons free zone an inadequate deterrent. Countries confronted by conventional threats may prefer to retain the option of pursuing nuclear armament in the future. States with rivals probably also prefer security assurances from nuclear weapon states (c.f. the Republic of Korea). A state facing serious security challenges will tend not seek to join or form a nuclear weapons free zone but instead will tend to proliferate or to ally with a nuclear protector. 10

13 Hypothesis 1a: States with limited security threats are more likely to join or establish nuclear weapon free zones. Hypothesis 1b: States that lack a nuclear capable ally are more likely to establish nuclear weapons free zones. II. Nuclear Weapons Free Zones as Security Communities In contrast to the previous view, one can argue that nuclear weapons free zones are really best conceived of as a form of security community. While there is no consensus definition of security community, many scholars would agree that security communities must be accompanied by confidence-building measures (Adler & Barnett 1998). Acharya (2001: 54) argues that the ASEAN members in the 1980s started losing faith in external security guarantees and that this led to the proposal for creating the ZOPFAN in Southeast Asia. Acharya argues that the nuclear weapon free zone in the region is a part of the ZOPFAN framework. Severino (2004) uses a similar conceptualization and cites the Bangkok treaty of 1995 as a crucial component of the ASEAN security community. Indeed, the annual report of ASEAN connects the idea of a security community with its nuclear weapon free zone (ASEAN 2005: 17). While this liberal or constructivist view of nuclear weapons free zones as security communities is intuitive and normatively appealing, the argument lacks causal clarity. It is difficult to say whether the nuclear weapons free zone is a product of the existing security community or vise versa. Social constructivists may argue that both mechanisms are present: a 11

14 nuclear weapon free zone leads to the formation (or fortification) of a security community even while existing communities in a region enhance the effectiveness of a nuclear free zone. However, to conduct a scientific analysis and provide an answer to my puzzle, I need to generate testable, i.e. directional (causal) hypotheses. Obviously, unraveling the causal connection between security communities and nuclear weapons free zones is beyond the scope of this study. Instead, I adopt a different approach. Here, I treat security communities and nuclear weapon free zones as sharing the same intervening variables. For instance, the North Atlantic security community --- Karl Deutsch s original case (1957) --- is said to be formed by democratic countries with open economies. Recent work also considers that these factors are the driving force of a security community (Risse-Kappen 1996; Åsberg & Wallensteen 1998; Gleditsch 2002). The same can be true for participation in a nuclear weapons free zone. Therefore, rather than attempting to clarify which component causally precedes which (security communities or nuclear weapons free zones), I focus on the intervening variables required to bring both the security community and a nuclear weapon free zone into existence. I focus on democracy and an open-economy, since these are the two least controversial intervening variables associated with security communities and nuclear weapons free zones. Hypothesis 2: Democracies and countries with open economies are more likely to establish nuclear weapons free zone. 12

15 Research Design This study focuses on a state s decision to sign and ratify a treaty to join or establish a nuclear weapons free zone. In particular, I am interested in when such a decision is made. Some may consider it more appropriate to focus on each actual proposal and compare successful and failed cases using historical analysis. However, descriptive studies of the formation of nuclear weapons free zones already exist. While important and useful, historical studies fail to identify the key determinants dividing states with a nuclear protector from states participating in a nuclear weapons free zone. We need a large-n study of states decisions as to whether they seek an ally with nuclear weapon states or participate in a regional non-nuclear weapon zone. 5 I generate a state-year dataset from 1958 (the year when the Polish government first introduced the idea of nuclear weapon free zone to the world) to I assume that a state leader after 1958 knows that he/she has the third option of creating a nuclear weapons free zone, in addition to developing nuclear weapons or obtaining a nuclear protector. 7 Fuhrmann & Li (2007) only study the ratification process and collect data after initial party states of a nuclear free zone treaty conclude an agreement and open the treaty for ratification and accession. My data, in contrast, covers all state-years between 1958 and 2000 and lists whether (and when) a state signed 5 Fuhrmann & Li (2007) share the same interest and adopt a very similar research strategy. 6 I conducted a robustness check using a dataset beginning in 1946 but I could not find a significant change in the results. 7 I assume that every state except for the five permanent nuclear powers (United States, Soviet Union, United Kingdom, France, China), has a chance to choose one of the three nuclear security options. The five permanent nuclear states are excluded from a dataset. 13

16 or ratified each treaty. 8 I use information available at the Agency for the Prohibition of Nuclear Weapons in Latin America and the Caribbean ( NWFZ s.htm). Inevitably, signatures come first, and ratification follows (there are cases where both occur simultaneously). Therefore, a dataset for the former covers a larger number of state-years (3622 state-years) than does a dataset for the latter (298 state-years). 9 Survival analysis is the most appropriate technique to use in analyzing the timing of decisions or political events like the signature or ratification of a treaty (Wooldridge, 2002: chap. 20; Box-Steffensmeier & Jones, 2004: 17-19). Specifically, this study uses a Cox proportional hazard model (Stata 9, using stcox command), which is identical to Fuhrmann & Li (2007). 10 The analysis includes nine independent variables. To test hypothesis 1a, I use the variables Enduring Rivalry, Frequency of Disputes and Nuclear Rivalry. Significant security concerns can be measured by several approaches but involvement with an enduring rivalry is a widely used indicator. I follow Bennett s (1998) coding scheme for interstate enduring rivalry. A state with an enduring rivalry in a given year is coded one; otherwise it is coded zero. Frequency of Disputes is an alternative measure of the presence of security concerns. I take a five-year moving average of the number of militarized interstate disputes per year in which a state is 8 The treaty for Central Asian free zone is thus dropped from the dataset. 9 The dataset for signature covers from 1958 to the year of signature or 2000 (if a state fails to sign). The dataset for ratification covers from the year of signature to the year of ratification or 2000 (if a state fails to ratify by 2000). The latter includes only states that have singed a treaty. 10 A Schoenfeld test of the proportional hazards assumption suggests no systematic variation in the residuals over time and thus it is appropriate to use a Cox proportional hazard model. 14

17 involved. I use the COW/MID version 3.0 dataset to construct the lagged average conflict variable (Ghosn, Palmer, & Bremer 2004). Also, I code the variable Nuclear Rivalry to distinguish the impact of having a rivalry relationship with one of the five nuclear weapon states. States facing a rivalry with a nuclear power in a given year are coded one; otherwise zero. 11 To test hypothesis 1b, I construct the variable Nuclear Umbrella. The five nuclear weapons states provide for their formal allies the potential to retaliate in the event that there is a nuclear attack on the ally. For the US, NATO members, Japan, the ROK, Australia and New Zealand (until it chose to establish a nuclear weapon free zone) receive protection. For the USSR, the Warsaw Pact members and other communist satellite states were covered. Even China has a formal alliance relationship with North Korea (the Treaty of Friendship, Co-operation and Mutual Assistance between the People's Republic of China and the Democratic People's Republic of Korea was formed in 1961), and this is supposed to protect North Korea from possible nuclear attack. A value of one is coded for a year when a state holds a formal alliance relationship with a nuclear weapon state; a value of zero is coded otherwise. 12 The variables Democracy and Open-Economy are needed to test hypothesis 2. I use the Polity IV dataset ( to create a dichotomous variable to identify democratic regime status. If a polity democracy score is larger than seven, the state is regarded as a democracy and receives a value of one; otherwise zero is assigned. To measure integration into 11 I created another variable for nuclear rivalry that includes the enduring rivalry states of Israel (1966-), India (1974-) and Pakistan (1990-). However, I could not find a meaningful difference. 12 COW Formal Alliance dataset ver is used (Gibler & Sarkees 2004). 15

18 the world economy, I use the Penn World Table dataset (version 6.2). 13 I sum exports plus imports as a share of GDP to measure exposure to the international economy. Finally, I include three control variables GDP/per capita, Post Cold War and NPT Membership. The economic wealth of a country and structural change in the international system could be related to a change in the probability of signing or ratifying nuclear free zone treaties. Economically weak states may be inclined to choose cheaper security policies, ceteris paribus. The end of the Cold War in 1989 also affected perceptions of the possibility of nuclear war between the United States and the Soviet Union. I use the purchasing-power parity GDP data from the Penn World Table project to measure economic size. For the Post Cold War variable, I code years after 1990 as one; otherwise zero. Finally, membership in the NPT should increase the chance of signing and ratifying a nuclear free zone treaty since participating in a nuclear weapons free zone poses no additional costs to states. Also, the membership can reinforce a norm against proliferation which might motivate states to join a nuclear free zone. 14 Results Table I shows the results of Cox proportional hazard model analyses on the decision to sign a nuclear free zone treaty. The results of three regressions are provided here. The left-hand side of I code a value of one for NPT membership; otherwise the variable is coded zero. I use a list from the James Martin Center for Nonproliferation Studies to code NPT membership. ( 16

19 the table (Model I) shows coefficients and robust standard errors of the estimation using the variable Enduring Rivalry. 15 The center of the table (Model II) shows coefficients and robust standard errors of the estimation using the variable Frequency of Disputes. The right-hand side of the table (Model III) shows coefficients and robust standard errors of the estimation using the variable Nuclear Rivalry. In contrast, Table II shows the results of Cox proportional hazard model analyses on the decision to ratify a nuclear weapons free treaty. Again, three estimations are provided. The left-hand side of the table (Model IV) shows coefficients and robust standard errors of the estimation using the variable Enduring Rivalry. The center of the table (Model V) shows coefficients and robust standard errors of the estimation using the variable Frequency of Disputes. The right-hand side of the table (Model VI) shows coefficients and robust standard errors of the estimation using the variable Nuclear Rivalry. Finally, Table III lists hazard ratios for statistically significant variables. A hazard ratio larger than one indicates that there is a greater likelihood of a state signing/ratifying a treaty to establish a nuclear weapons free zone. ***************************************************************** Table I Results of Cox Proportional Hazard Analysis: Signature ***************************************************************** ***************************************************************** Table II Results of Cox Proportional Hazard Analysis: Ratification 15 Robust standard errors are used since the model does not assume the independence of observations. 17

20 ***************************************************************** ***************************************************************** Table III Hazard Ratios for Statistically Significant Variables ***************************************************************** Several security concerns generally decrease the odds of signing a nuclear weapons free zone treaty. For instance, the variable Enduring Rivalry for the signature model (Model I) obtains a hazard ratio smaller than one (0.49). 16 This suggests that the effect of having an enduring rivalry is to decrease the hazard of signing a treaty by 51%. A state with an enduring rivalry is about half as likely to sign a free zone treaty as a state without a rival. The variable Nuclear Rivalry for the signature model (Model I) obtains a very small hazard ratio (0.26), which suggests that a state with a nuclear rivalry is 74% less likely to sign a free zone treaty than a state without such a rivalry. A high frequency of interstate disputes also delays signature. The hazard ratio of 0.43 for the variable suggests that a one-unit increase in the frequency of disputes reduces a chance of signing by about 57%. Security challenges considerably delay treaty signing. For ratification, however, all three variables fail to meet statistical significance at even the 10% threshold. Hypothesis 1a is thus supported only at the signature stage. 17 The Nuclear Umbrella variable reports a high level of statistical significance and also produces a large substantive effect. However, the direction of the variable is not as expected. The 16 I conducted a robustness check by using Klein, Goertz & Diehl s Enduring Rivalry dataset (2006). I obtained a statistically significant negative coefficient (-0.52) as expected. 17 In fact, this is consistent with Fuhrmann & Li (2007). 18

21 results suggest that the existence of a nuclear umbrella actually leads to a higher chance of signing a nuclear free zone treaty. If a state obtains a formal alliance with one of the five nuclear weapon states, it tends to sign a nuclear free zone treaty about 1.7 times as often as a state without such a tie. In sum, hypothesis 1b is rejected. This is surprising because previous studies have considered that nuclear umbrella and nuclear weapon free zones not to be a complement but a substitute for each other. The facts are contrary to conventional wisdom; a state with a nuclear umbrella is more likely to endorse regional non-nuclear proliferation regimes. Hypothesis 2 is also not confirmed. Democracy does not promote states joining nuclear weapon free zones. If democracy is an intervening variable for security communities and nuclear weapons free zones, it should have obtained a positive coefficient; however, it did not. As for economic openness, the variable obtains a statistical significance but the hazard ratio (1.003) suggests its substantive effect is not large. One standard deviation change in the openness score leads to only a 13% increase of the chance of signing a treaty. In any case, it must be emphasized a particularly plausible intervening variable, democracy, fails to achieve statistical significance. Also, the impact of the economic openness variable is not very strong. Thus, the argument that a nuclear free zone is a security community is not really supported. Finally, the impact of the variable Post Cold War cannot be ignored. A state is 16 times more likely to sign a nuclear weapon free zone treaty after The end of the Cold War was also significant in generating regional non-proliferation regimes. This makes sense since three of the five nuclear free regimes were created in the Post Cold War era. In contrast, I find that ratification 19

22 of a treaty is delayed after A state is 79% less likely to ratify a nuclear weapon free zone treaty in the Post Cold War era. Also, it must be noted that the variable NPT Membership decreases the chance of a signature by around 50% while it increases the chance of ratification by about two times. This is an interesting contrast between the signature and ratification stages. For most states, signing a treaty is a political commitment but does not necessarily accompany a full legal obligation until the treaty is formally ratified. Unless it is a NPT member that has already adopted a peaceful use of nuclear energy policy and accepts IAEA safeguards inspections, a state that signed a nuclear free zone treaty needs to change related domestic laws and prepare the country s nuclear facilities for accepting IAEA inspections. Implementation takes time. Therefore, it is understandable that NPT membership accelerates the timing of ratification. In contrast, a NPT member states may not be strongly motivated to sign a nuclear free zone treaty, since NPT members are already committed to non-proliferation and do not feel a special need to establish or join in local nuclear non-proliferation regime. Conclusion This study examines some key determinants of signing and ratifying a nuclear weapons free zone treaty. As expected, security concerns tend to delay signature of nuclear weapons free zone treaties. However, it is surprising that a formal alliance with one of the five original nuclear weapon states is not a substitute for a nuclear free zone, or that intervening variables for security community are not strongly connected with the odds of accepting a nuclear free zone treaty. 20

23 Overall, the study suggests that a nuclear weapon free zone is not really a security community or a substitute for the formal alliance framework of nuclear weapon states. If a state faces very limited security concerns, the country would seek such a regime and try to obtain a pledge of the non-use of nuclear weapons from major powers. As long as a conventional dispute continues or an enduring rivalry relationship remains, the state is very unlikely to move toward forming or joining a nuclear weapon free zone. Nuclear weapons free zones are less instruments to create peace than they are mechanisms to maintain cooperative relationships in peaceful regions. Obviously, this suggests that a nuclear weapon free zone in the Middle East and South Asia is merely a pipe dream unless the level of disputes and rivalry ratchets down. This article starts by comparing three policy options: development of nuclear weapons, allying with a nuclear patron, or creating a nuclear weapon free zone. These options have previously been considered substitutes. However, this is not the case, as I report here. An alliance with a nuclear weapon state actually increases the chance of a nuclear weapon free zone. This result may be due to free-riding by non-nuclear states. 18 Free zone participants with security assurances can reduce the possibility of a nuclear attack. But at the same time, by denying the patron the right to store or ship nuclear weapons on its territory, the client reduces the risk of being a target of retaliation. The client further avoids responsibility for nuclear accidents in their territories (and neighbor states). Developing nuclear weapons may be the most reliable way to 18 A similar claim appears in Fuhrmann & Li (2007: 11). This article provides supporting evidence for this argument. 21

24 deter a possible nuclear attack but politically and economically this option is often too costly for most countries. It also takes considerable time to develop nuclear weapons. Seeking a nuclear patron is a short-cut but this option poses a high risk of being entangled with a major power s conflicts and even a nuclear war. It may therefore be a better choice for a state to seek simultaneously nuclear protection as well as joining a nuclear free zone. Protection cheaply deters an initial nuclear attack and a free zone reduces the risk of being a target There is also a possible complementary relationship between covert development of nuclear weapons and overt participation in a nuclear weapons free zone as political cover. Kroenig (xxxx) argues that sensitive nuclear assistance leads to horizontal proliferation. If so, under a political cover of the NPT and nuclear weapons free zone, a state that obtains sensitive nuclear assistance may develop nuclear weapons (this story fits to states like Argentina, Brazil, and Libya even though all three finally abandoned their nuclear weapons development programs). 22

25 Table I Results of Cox Proportional Hazard Analysis: Signature Model I Model II Model III Coef. Robust Robust Robust Coef. Coef. Std. Err. Std. Err. Std. Err. Enduring Rivalry ** - - Frequency of Disputes *** - Nuclear Rivalry * Nuclear Umbrella ** ** ** Democracy Economic_Openness ** ** ** GDP/cap * Post Cold War *** *** *** NPT Membership *** *** ** Number of Obs Number of State Log-likelihood Note: Significance tests are two-tailed. *** p < 0.01, ** p < 0.05, * p < Table II Results of Cox Proportional Hazard Analysis: Ratification Model IV Model V Model VI Coef. Robust Robust Robust Coef. Coef. Std. Err. Std. Err. Std. Err. Enduring Rivalry Frequency of Disputes Nuclear Rivalry Nuclear Umbrella Democracy Economic_Openness GDP/cap Post Cold War *** *** *** NPT Membership * * * Number of Obs Number of State Log-likelihood Note: Significance tests are two-tailed. *** p < 0.01, ** p < 0.05, * p < Table III Hazard Ratios for Statistically Significant Variables Signature Ratification Hazard Ratio Hazard Ratio Enduring Rivalry Frequency of Disputes Nuclear Rivalry Nuclear Umbrella * Economic_Openness * GDP/cap * Post Cold War * NPT Membership * Note: a hazard ratio is reported for statistically significant variables (Model I and IV are used for *). 23

26 References Acharya, Amitav, Constructing a Security Community in Southeast Asia: ASEAN and the Problem of Regional Order. New York: Routledge. Adler, Emanuel, & Michael N. Barnett, eds., Security Communities. Cambridge: Cambridge University Press. Åsberg, Carl Johan, & Peter Wallensteen, New Threats and New Security: The Post-Cold War Debate Revisited, in Wallensteen, Peter, (ed.) Preventing Violent Conflicts. Past Record and Future Challenges. Uppsala University: Department of Peace and Conflict Research. Pp ASEAN, ASEAN Annual Report Jakarta: ASEAN Secretariat. Baumgart, Claudia, & Harald Müller, A Nuclear Weapon free zone in the Middle East: A Pie in the Sky?, Washington Quarterly. 28(1): Bennett, Scott D., Integrating and Testing Models of Rivalry Termination, American Journal of Political Science. 42(4): Box-Steffensmeier, Janet M. & Bradford S. Jones, Event History Modeling: A Guide for Social Sciences. Cambridge: Cambridge University Press. Bunn, George, The Legal Status of US Negative Security Assurances to Non-Nuclear Weapon States, Nonproliferation Review. 4(3): Deutsch, Karl W., et al., Political Community and the North Atlantic Area: International Organization in the Light of Historical Experience. Princeton, N.J.: Princeton University Press. Fuhrmann, Matthew, & Xiaojun Li, Legalizing Nuclear Restraint: When Do States Enter Nuclear Weapons Free Zones?, Paper Presented at the Annual Meeting of the American Political Science Association, Chicago, IL, August 30-September 2. Ghosn, Faten, Glenn Palmer, & Stuart Bremer, The MID3 Data Set, : Procedures, Coding Rules, and Description, Conflict Management and Peace Science. 21: Gibler, Douglas M., & Meredith Sarkees, Measuring Alliances: the Correlates of War Formal Interstate Alliance Data set, , Journal of Peace Research. 41(2):

27 Gleditsch, Kristian S., All International Politics is Local: The Diffusion of Conflict, Integration, and Democratization. Ann Arbor, M.I.: University of Michigan Press. Goldblat, Jozef, Nuclear-Weapon-Free-Zones: A History and Assessment, Non- proliferation Review. 4(3): Jo, Dong-Joon, & Erik Gartzke, Determinants of Nuclear Weapons Proliferation, Journal of Conflict Resolution. 51(1): Jones, Rodney W., & Mark McDonough, Tracking Nuclear Proliferation: A Guide in Maps and Charts, Washington DC: Carnegie Endowment for International Peace. Klein, James P., Gary Goertz & Paul Diehl, The New Rivalry Dataset: Procedures and Patterns, Journal of Peace Research. 43(3): Kroenig, Matthew, xxxx. Importing the Bomb: Sensitive Nuclear Assistance and Nuclear Proliferation, XXXX. xx(x): x-xx. Leigh-Phippard, Helen, Nuclear-Weapon-Free-Zones: Problems and Prospects, Arms Control. 14(2): Maoz, Zeev, Defending The Holy Land: A Critical Analysis of Israel s Security and Foreign Policy. Ann Arbor, MI: University of Michigan Press. McInnis, Kathleen J, Extended Deterrence: The U.S. Credibility Gap in the Middle East, Washington Quarterly. 28(3): Ozinga, James R., The Rapacki Plan: The 1957 Proposal to Denuclearize Central Europe, and an Analysis of Its Rejection. Jefferson, NC: McFarland. Power, Paul F., The South Pacific Nuclear-Weapon-Free-Zone, Pacific Affairs. 59(3): Redick, John R., Regional Nuclear Arms Control in Latin America, International Organization. 29(2): Risse-Kappen, Thomas, Collective Identity in a Democratic Community: The Case of NATO, in Katzenstein, Peter, (ed.) The Culture of National Security: Norms and Identity in World Politics. New York: Columbia University Press. Pp Severino, Rodolfo C., Towards an ASEAN Security Community. Singapore: Institute of 25

28 Southeast Asian Studies. Singh, Sonali, & Christopher Way, The Correlates of Nuclear Proliferation: A Quantitative Test, Journal of Conflict Resolution. 48(6): Steinberg Gerald M., US Non-Proliferation Policy: Global Regimes and Regional Realities, Comparative Security Policy. 15(3): Tannenwald, Nina, The Nuclear Taboo: The United States and the Normative Basis of Nuclear Non-Use, International Organization.53(3): Thakur, Ramesh, A Nuclear-Weapon-Free South Pacific: A New Zealand Perspective, Pacific Affairs. 58(2): Wooldridge, Jeffrey, Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press. 26