Hello! The below report is written in English. To translate the full report, please use the translator in the top right corner of the page. Do not show me this notice in the future.

NAPSNet Special Report

Recommended Citation

Morton Halperin, Peter Hayes, Leon Sigal, "A KOREAN NUCLEAR WEAPONS-FREE ZONE TREATY AND NUCLEAR EXTENDED DETERRENCE:  OPTIONS FOR DENUCLEARIZING THE KOREAN PENINSULA", NAPSNet Special Reports, April 12, 2018, https://nautilus.org/napsnet/napsnet-special-reports/a-korean-nuclear-weapons-free-zone-treaty-and-nuclear-extended-deterrence-options-for-denuclearizing-the-korean-peninsula/


APRIL 12 2018



In this essay, the authors tackle three tasks. First, they offer a way to address potential DPRK doubts about unilateral US negative security assurances by means of a legally binding and enduring multilateral treaty to establish either a regional Nuclear Weapons-Free Zone (NWFZ) on the Korean Peninsula only or in Northeast Asia as a whole. Second, they demonstrate that there is no incompatibility between the US commitment of extended nuclear deterrence to the ROK and membership by the US and ROK in a NWFZ. Whatever moral objections nuclear extended deterrence may raise, there is no existing legal impediment to such a commitment. Third, they provide a way around possible political objections in both South and North Korea to signing a Korean Peninsula-only or a regional treaty because it would call into question their competing claims to sovereignty over the whole peninsula. It proposes the option of a UN NWFZ treaty to which they alone or they in concert with others in the region might sign, much as they have signed other treaties within a UN framework in the past.

 Morton Halperin is senior advisor, Open Society Foundations;   Peter Hayes is Director of the Nautilus Institute and Honorary Professor at the Centre for International Security Studies at the University of Sydney.  Leon Sigal is Director, Northeast Asia Cooperative Security Project, Social Science Research Council.

Acknowledgements:  The authors are grateful for review comments from Ambassadors Thomas Pickering and Thomas Graham.  The authors remain solely responsible for the content.  This paper was funded in part by MacArthur Foundation, Open Society Foundations, and Ploughshares Fund.

The views expressed in this report do not necessarily reflect the official policy or position of the Nautilus Institute. Readers should note that Nautilus seeks a diversity of views and opinions on significant topics in order to identify common ground.

Banner image:  Korean Peninsula and universal nuclear weapons-free zone icon



APRIL 10 2018



A critically important part of assembling the Korean peninsula-wide denuclearization jigsaw puzzle is the institutional and legal form of North Korean commitments on the one hand, and the nuclear negative security assurances by the NPT-Nuclear Weapons States (NWSs), especially the United States, on the other.

The institutional framework might take one of three possible forms.

The first, a Korean Peninsula-only deal between the ROK and the DPRK is possible.  It would essentially revise and expand the 1992 Denuclearization Declaration, and make the commitments specific, with stringent monitoring and verification measures.  The United States, Russia, and China would make a general security assurance commitment to the DPRK, and at least the United States, a specific nuclear negative security assurance to the DPRK that it would not threaten or attack the DPRK first with nuclear weapons once the DPRK complies fully with its NPT obligations as a non-nuclear weapons state (NNWS).

Such a Korean Peninsula-only deal is likely easier to negotiate, but may not be credible at the outset to the DPRK given its perception of past reversals of US executive branch commitments such as the rapid demise of Clinton’s non-hostility statement to the DPRK in 2000 under the Bush administration, and the failure of the 1994 Budapest security assurances to protect the Ukraine against Russian aggression.  The DPRK’s perceptions of the non-binding commitment implied by a new Denuclearization Declaration may lead it to balk or hedge against uncertainty from a Korea-only deal.

A Korean Peninsula-only deal might be made more legally binding if it were elevated from a mere declaration to an inter-Korean treaty between the two states and if each Korea were to caveat its ratification by issuing a reservation with regard to sovereignty issues (both Koreas refuse to sign treaties with the other because it would imply recognition of the others’ constitutional claims to exercise sovereignty over the entire Korean peninsula).  Whether this issue can be finessed at this time in either Korea is doubtful politically, especially in democratic South Korea.

The second, a full-fledged regional nuclear weapons-free zone (NWFZ) UN treaty may be more enduring because it affects how the NWSs use nuclear threats against all the NNWSs party to a treaty, and thereby against each other.  It may be difficult, however, to bring the United States and Japan into such a treaty even if the DPRK, China, and Russia favor it and it may take time for the NWS to ratify their nuclear negative security assurances to a regional NWFZ (which would be calibrated to DPRK compliance).

It may be possible to square the circle:  the ROK and the DPRK could implement a third, hybrid option of a UN NWFZ Treaty that specifies that additional members may join at the outset or later.  This approach may be optimal in providing a politically less demanding Korean Peninsula-only than a full regional NWFZ treaty at the outset, but also the more legally binding framework than a Korean Peninsula-only, fragile political agreement.  Its feasibility depends on whether such a UN treaty framework, as it has in the past with many other UN treaties, gives the two Koreas an acceptable “work around” on their competing sovereignty claims when they sign and ratify the treaty.

At minimum, South Korean and American officials should explore at the senior official level the DPRK’s interest in these options, and study carefully the pros and cons of these options in preparations for the two summits.  It is especially important to clarify what type of nuclear negative security assurance is sought by the DPRK and if they are not clear, suggest some desirable options that would serve to improve the security of all parties to a comprehensive settlement of the nuclear issue in the Korean Peninsula.

This issue is important because it is linked to the degree to which the United States’ and other NWSs’ negative security assurances are legally binding, thereby affecting the DPRK’s perception of the desirability and credibility of a proposed deal.  We therefore review below how a NWFZ would affect the existence of nuclear extended deterrence in US security commitments to the ROK and to Japan.

Whichever framework is employed for the denuclearization process, the US commitment of extended deterrence would remain subject to the normal political prerogatives of the United States and the ROKG at any time to vary these understandings on the use of nuclear threat against the DPRK and other parties.  We conclude that concerns in Seoul (and Tokyo) that a NWFZ would terminate nuclear extended deterrence are groundless.

In short, there is no incompatibility between nuclear extended deterrence and adherence by the United States and the ROK to  a NWFZ treaty.

1.  US Nuclear Umbrella: Always Circumstantial, Not-Automatic, Narrowing Applicability

US treaty alliances in Europe and Asia with non-nuclear weapons states evolved to include extended US nuclear deterrence.  Articles 2 and 3 of the original US-ROK security treaty require only consultation and a decision by each party to the treaty to act in accordance with its constitutional processes, and there is no automaticity in the nature of the US response to external attack, nuclear or conventional, or at all.

Over time, however, the United States has made various statements to the ROK that clarify how it will respond to a nuclear attack. The most important of these was its assurance in 1991 that it would maintain the “nuclear umbrella” after withdrawal of US nuclear weapons from the ROK.

The conditions under which nuclear threat would come into play were specified further in its 1995 letter at the time of UNSC Resolution 984.[1]  This letter states that it will not use nuclear weapons against a non-nuclear weapons state (provided it is not in an alliance with a nuclear weapons state while it is engaging in aggression using non-nuclear means), but it will come to the aid of non-nuclear state suffering from nuclear aggression and until the UNSC responds, it reaffirms its right to individual and collective self-defense measures.[2]

This clause remained in effect until 2010, when reference to an aggressor’s nuclear ally was removed in the Nuclear Posture Review (along with the exceptions for use in response to chemical or biological weapons attack).[3]  On many occasions, both the US and the ROK stated to the DPRK that it would become eligible for the nuclear negative security assurance once it fulfilled its NPT-IAEA obligations.

Since 2010 and in the aftermath of the two major military clashes between the ROK and the DPRK, the United States has maintained that it will extend deterrence to the ROK against external attack, most notably against the DPRK, but has not stated that it would automatically respond to nuclear attack with nuclear retaliation.  Under the original treaty and current US policy, doing so would be a decision for the US president that would be made in a specific context.  Indeed, in the press conference for the release of the Trump administration’s Nuclear Posture Review, the phrase “no automaticity” was used three times.[4]

In reality, US extended deterrence is likely to be conventional, not nuclear, except in extremis.  Typically, US statements are to the effect that “that any attack on the United States or its allies will be defeated, and any use of nuclear weapons will be met with an effective and overwhelming response.”

Thus, until the DPRK comes back into full compliance with its obligations under the NPT, it remains subject to the US policy of possible nuclear retaliation, and even first use of American nuclear weapons, should it engage in aggression against the ROK, whether conventional or nuclear.

2.  Legally Binding Versus Political Nuclear Negative Security Assurances

Once the DPRK returns fully to the NPT fold, then and only then would it be eligible to receive the US unilateral nuclear negative security assurance (with the caveats updated slightly by the 2018 Nuclear Posture Review and its subsequent adoption as declaratory policy by President Trump).  Should it rejoin the NPT, and should it also join a Korean-Peninsula nuclear free agreement or regional nuclear weapons free zone treaty in good standing, then and only then would it obtain the negative security assurance from the United States (and from the other NWS that commit to observe the terms of the agreement or treaty with respect to the non-use of nuclear weapons, including nuclear threats).

At that time, the United States and other NWS party to the agreement or treaty would be committed to not use nuclear weapons in response to a DPRK conventional attack on such parties (that is, the NWS) or the ROK (and any other non-nuclear weapons state party to the treaty).  At that time, the DPRK would be completely nuclear-disarmed with on-going monitoring and verification of this status, and the existing US extended deterrence commitment to the ROK would remain.

Moreover, US nuclear extended deterrence commitments to the ROK with regard to China and Russian nuclear threats or attack would remain in force (in fact, they would be extended to both Koreas, and to Japan and/or Mongolia, should either of these two states join a regional NWFZ treaty), although there is no automaticity to the nature of the US response to such threats or attack (see the 1995 letter) any more than there was the original US-ROK security treaty.

Should the DRPK revert to nuclear threat and/or armament, the US and other nuclear weapons states negative security assurances would be withdrawn, and nuclear extended deterrence would come back into play against DPRK nuclear and conventional threats or attack.  In this sense, nuclear extended deterrence would never go away, even in a region-wide nuclear weapons free zone.

Even at that time, reactivation of nuclear extended deterrence would be a choice for the ROK, not a given.  For example, in the South Pacific Nuclear Weapons Free Zone which came into force in 1986, Australia did not assert that the United States extends nuclear deterrence to Australia until 1996; and New Zealand, also a full party to the treaty, eschews it even today.  The South Pacific Nuclear Weapons Free Zone is a living example of the co-existence of a NWFZ with nuclear extended deterrence, depending on the security context and preferences of the allies.

3.   The ROK and DPRKs’ Choice

Ultimately, therefore, the ROK and the DPRK must choose the most constructive legal and institutional form for denuclearization of the Korean peninsula..

Option 1:  Regional UN NWFZ Treaty

In a regional UN NWFZ treaty, it may obtain legally binding commitments from the DPRK to disarm and return to full compliance with its NPT-IAEA safeguards by virtue of the creation of a regional NWFZ, and then foregoing a nuclear element to deterrence and compellence in conventional warfare so long as the DPRK remains in good standing with respect to the NWFZ terms and the NPT.  The DPRK could join the NEA-NWFZ treaty at the outset, but not waive the provision that the treaty only come into force when all parties have ratified it, while the other parties would waive this provision.[5] The DPRK thereby would reaffirm its commitment to become a NNWS in compliance with its NPT-IAEA obligations, but would take time to comply fully as is realistic even under the most optimistic assumption. The other NNWSs could set a time limit for this to happen and reserve the right to abandon the treaty if the DPRK has not denuclearized sufficiently by that time. Concurrently, the NWSs would qualify their guarantees to not use nuclear weapons to attack the NNWSs party to the treaty to be calibrated to the extent that the DPRK has come into full compliance.

In this manner, the DPRK’s nuclear armament, such as it is, would not be recognized as legitimate in any manner; the standards that it must meet when denuclearized would equal those for all non-NNWSs in the NWFZ, including M&V requirements; and most important, the DPRK would be offered a legally binding, multilateral guarantee by all the NWSs that it will not face nuclear threat or the use of nuclear weapons against it.

Such a full-fledged regional UN NWFZ treaty entails bringing Japan into the framework.  A regional NWFZ treaty also would be more complicated to negotiate and to construct, including a regional inspectorate and secretariat.  Yet, because of the multilateral nature of the treaty and its institutions, and because it would require not only the US but also China and Russia to make their non-first use commitments legally binding against non-nuclear state parties to the treaty, it may be more durable and useful to both Koreas.  The essence of a multilateral NWFZ treaty is that it creates interlocking commitments by multiple parties such that one party must compare the impacts of any attempt to gain a marginal military advantage against the political and military significance of how all the other parties, not just one adversary, may respond, thereby raising the cost of breakout.

In particular, if Japan is thereby induced to give up forever its increasingly active but latent nuclear “technological deterrent,” the two Koreas will address a major Chinese security concern and contribute to regional strategic stability in a way that a Korean Peninsula-only agreement would not.

Option 2:  Korean Peninsula-Only Declaration or Treaty

In a Korean Peninsula-only agreement whereby the DPRK would disarm its nuclear weapons and the NWSs would provide negative security assurances of no nuclear threat and no nuclear attack to the DPRK, the commitments would be legally binding on the two Koreas if the two Koreas elevated it from a mere declaration to a state-state treaty between the two Koreas. Such an agreement would not bind the ROK with respect to the recognition of DPRK sovereignty over any portion of the peninsula if it did not explicitly do so on its own terms and if the ROK undertook a reservation unilaterally to that effect upon ratification of such a treaty.

Although the DPRK might seek to invoke a contrary reservation, without ROK acceptance of such a reservation and in the circumstances where it had explicitly said it would not accept such a reservation, it could not be held to be so bound.  Of course, if either Korea refused to accept the other’s reservation with respect to sovereignty, then the resulting deal would become a purely political declaration similar to the 1992 Denuclearization Declaration with much less binding commitment by all parties.

Thus, a Korean Peninsula-only agreement that is not in a treaty format between the two Koreas may not be attractive to the DPRK in part because it would only be legally binding on the US administration that makes the commitment and can be revoked overnight by that administration or a subsequent US president.  Conversely, the DPRK might also be skeptical that the US Senate would actually commit to negative security assurances in a regional NWFZ treaty under the US advise-and-consent treaty ratification procedure.

Option 3:  Korean Peninsula UN NWFZ Treaty

The third option, a Korean Peninsula NWFZ treaty, may be the most desirable of the three options.  The two Koreas could sign a standard UN treaty and include the possibility that a Korean Peninsula-only treaty could expand over time to include other states.  If Japan or Mongolia is amenable, then they could join at the outset.  Otherwise, they could join later, if at all.  Meanwhile, the two Koreas would have signed an open-ended standard UN treaty that can expand its membership over time to include other non-nuclear weapons states.  Both would get the nuclear negative security assurances that both need from the nuclear weapons states when the latter sign and ratify the Korea-only NWFZ treaty.  Both could enter qualifications with respect to the sovereignty issue although it would be politically preferable that they do not do so.[6]

As with option 1, a Korean Peninsula-only UN NWFZ treaty would have to cover a variety of issues including extent (territorial seas only versus EEZ coverage; transit of territorial seas[7] and EEZ including straits versus stationing, creation of a regional secretariat and inspectorate, hosting of nuclear support NC3 systems such as cyberwarfare and communications already subject to international law such as the Tallin Manual), etc.

This option would have the additional advantage to the DPRK in that the ROK would be legally bound not to permit the US or any nuclear power to store nuclear weapons in its territory.  The US would be legally bound not to store nuclear weapons on the peninsula.  Finally, the legal obligations of the DPRK and the ROK would be the same including the inspection procedures after the DPRK has satisfied the other parties that it has met is obligation to de-nuclearization.

However, as noted above, a full-fledged regional UN NWFZ treaty would also be the most difficult to negotiate unless the United States took it upon itself to exert hegemonic leadership which is improbable today.  Achieving regional consensus that such a regional UN treaty is desirable may also be beyond the diplomatic reach of a middle power such as the ROK.

Conversely, a Korean Peninsula NWFZ open to accession by other NNWSs at the outset or later seems entirely within reach, and may offer advantages with respect to the strategic goals of all parties.

4.  Salience of Nuclear Extended Deterrence

Whichever of these three frameworks is used, US nuclear weapons non-use policy remains the same; and its nuclear umbrella remains in place even when the DPRK is in full compliance with the terms of the NWFZ and the NPT-IAEA arrangements.  US nuclear deterrence will come into play the moment the DPRK fails to be in full compliance.  Although there are powerful moral and potent political objections to nuclear extended deterrence, we see no legal impediment to it continuing to operate in a NWFZ.[8]

Underlying all the options is Article 51 of the UN Charter that provides for self-defense by ROK, US and allied forces. And indeed, all the NWSs including the United States have been slow to ratify their nuclear negative security assurance commitments under existing NWFZ treaties.

5.   Conclusion

All three options are plausible.  However, history suggests that a Korea-only agreement that is not in treaty format may be brittle and short-lived given that state of inter-Korean relations and fluctuating great power postures towards the two Koreas.

Equally, an attempt by the ROK to use its middle power influence to realize a regional NWFZ may be a reach too far given the uncertainty in the Trump Administration’s position on this approach, and Japan’s potential blocking role.

Thus, the ROK and the DPRK should consider implementing a hybrid option of a UN Korean Peninsula NWFZ Treaty that specifies that additional members may join at the outset or later.  This approach may be optimal in providing the politically less demanding than a full regional NWFZ treaty at the outset, but also the more legally binding framework than a Korean Peninsula-only political agreement.  Its feasibility depends on whether such a UN treaty framework gives the two Koreas an acceptable “work around” on their competing sovereignty claims.

Our consultations since 2010 with leading Chinese and Russian policy makers and advisers suggest that while neither will take the lead in creation of a regional NWFZ, they will almost certainly support the creation of one.

What is unknown is the DPRK’s interest in a legally-binding negative security assurance from the United States.  Recent official DPRK statements have doubled-down on the ending of US hostility towards the DPRK, often juxtaposed with the ending of US nuclear threat to the DPRK, and sometimes reinforcing the need for negative security assurances to be legally binding.

Also indeterminate is the extent to which the DPRK is open to a NWFZ as providing the legal and institutional framework in which to obtain such assurances.  In 2011, Nautilus Institute produced a memo addressing past DPRK views on NWFZs, in English here and Korean here.  This memo explained the pros and cons of a UN NWFZ treaty, and also traced the lineage of DPRK pronouncements on NWFZs under Kim Il Sung and Kim Jong Il, which may be salient to Kim Jong Un’s thinking today.  This memo was shared with the DPRK Ministry of Foreign Affairs in November 2013.

It is therefore timely and urgent for the ROK, the United States, and other parties, to sound out the DPRK’s current views on these related matters in the preparations for the inter-Korean and US-DRPK summit meetings.


[1] The text is found here.

[2] “Statement issued on 5 April 1995 by the Honourable Warren Christopher, Secretary of State, regarding a declaration by the President on security assurances for non-nuclear-weapon States Parties to the Treaty on the Non-Proliferation of Nuclear Weapons,” A/50/153 S/1995/263 6 April 1995, a copy of which may be found here (scan down to A/50/153)

[3] See J. Lewis, P. Hayes, “The DPRK and the Warsaw Clause: An Unnoticed Change in US Nuclear Policy”, NAPSNet Policy Forum, July 28, 2011.

[4] US DOD, “News Briefing on the 2018 Nuclear Posture Review,” News Transcript, February 2, 2018, at:  https://www.defense.gov/News/Transcripts/Transcript-View/Article/1431945/news-briefing-on-the-2018-nuclear-posture-review/

[5] This approach is transposed from the Tlatelolco Treaty which established an ingenious and innovative legal mechanism by which reluctant states could be encouraged to join the zone at a later date. It consists of a provision in Article 28 (3) that allows a signatory state to “waive, wholly or in part” the requirements that have the effect of bringing the treaty into force for that state at a particular time.11 As Mexican diplomat Alfonso Garcia Robles noted in his commentary on Article 28: “An eclectic system was adopted, which, while respecting the viewpoints of all signatory States, prevented nonetheless any particular State from precluding the enactment of the treaty for those which would voluntarily wish to accept the statute of military denuclearization defined therein. The Treaty of Tlatelolco has thus contributed effectively to dispel the myth that for the establishment of a nuclear-weapon-free-zone it would be an essential requirement that all States of the region concerned should become, from the very outset, parties to the treaty establishing the zone. In this way, the normative framework for a non-nuclear region can be established before all states are ready to actually implement the framework.” M. Hamel-Green, “Implementing a Korea–Japan Nuclear-Weapon-Free Zone: Precedents, Legal Forms, Governance, Scope, Domain, Verification, Compliance and Regional Benefits,” Pacific Focus, 26:1, April, 2011, pp. 97-98, at: http://onlinelibrary.wiley.com/doi/10.1111/pafo.2011.26.issue-1/issuetoc

[6] In principle, if the sovereignty issue becomes a block for such a treaty, then DPRK can enter into binding legal NWFZ treaty with all but ROK; and the ROK can enter into a NWFZ treaty with all but DPRK, with the Nuclear Weapons States giving nuclear negative security assurances as protocols to both treaties.   As a leading American expert on NWFZs noted, “The UN is very clear on this even just one state can be a NWFZ.”  However, this is a complicated and cumbersome work-around way to achieve a standard UN NWFZ treaty.

[7] Ambassador Thomas Graham, a leading US diplomat and NWFZ expert, notes:   “Australia has asserted that it is under the US nuclear umbrella and is also a party to the SPNWFZ Treaty. This is true because the US provides its nuclear umbrella to Australia from afar without stationing nuclear weapons in Australia-which would be prohibited by the SPNWFZ Treaty. A particular NWFZ Treaty is simply a Treaty which has specific terms and if a government policy pursued by one of its Parties is not in conflict with any of the Treaty’s provisions there is of course no conflict. All of the existing NNWFZ treaties allow the transit of nuclear weapons through the zone of application of the treaty and all prohibit the stationing of nuclear weapons on the territory of a party. So if the usual form for a NWFZ is used for a Northeast Asia nuclear weapon free zone treaty and the US deterrent policy requires transit though the zone that is not a problem but if such policy requires the stationing of nuclear weapons on the land territory of a Party that would be a direct conflict. Of course different language could be used for a Northeast Asian NWFZ. But if the standard version is used there would be no conflict.”

[8] J. Dhanapala advances a contrary argument in “NWFZS and Extended Nuclear Deterrence: Squaring the Circle?” NAPSNet Special Report, May 1, 2012, at: https://nautilus.org/napsnet/napsnet-special-reports/nwfzs-and-extended-nuclear-deterrence-squaring-the-circle/  The experts cited in the original 1975 UN study of NWFZs were split on whether nuclear deterrence could be extended to NNWSs party to a NWFZ. See Comprehensive Study Of The Question Of Nuclear-Weapon-Free Zones In All Its Aspects, Special report of the Conference of the Committee on Disarmament , UN Doc. A/10027/Add. 1, New York, 1975 http://www.un.org/disarmament/HomePage/ODAPublications/DisarmamentStudySeries/PDF/A-10027-Add1.pdf  American experts who participated in the creation of the NPT are clear that the NPT does not impede the operation of nuclear extended deterrence.  The Nuclear Weapons Prohibition Treaty may affect state practice over time, but at this time, is not applicable to NWS nor nuclear “umbrella” states such as the ROK and Japan.


The Nautilus Asia Peace and Security Network invites your responses to this report. Please send responses to: nautilus@nautilus.org. Responses will be considered for redistribution to the network only if they include the author’s name, affiliation, and explicit consent


Leave a Reply

Your email address will not be published. Required fields are marked *