PENN Research Report 2000.1

Questions of Command and Control:
NATO Nuclear Sharing and the Non-Proliferation Treaty

Chapter Two: NATO Nuclear Sharing: What it is, 
and How it Came About

The debate in the NPT over nuclear sharing makes a detailed understanding of the nature of NATO’s nuclear sharing arrangements essential.

2.1 Nuclear Sharing or Proliferation?  

During the late 1950’s and the early 1960’s, intense discussions were held within NATO on what form of nuclear sharing to establish within the Alliance. Some European allies – Germany among them – pressed the US to allow them some kind of participation in nuclear planning, decision-making, and command and control, since they wanted a minimum of a strong say in decisions that could put the very existence of their nations at stake. However, the declared NWS wanted to limit access to nuclear weapons by other parties. As a compromise, the system of “nuclear sharing” was established in NATO in the mid-1960s, and its basic functions have remained unchanged ever since. 

As early as 1964, a once top-secret description of US policy outlined the consequence of deploying nuclear weapons on aircraft owned by non-nuclear members of NATO. It read: “As a result of NATO’s commitment to the nuclear mode of defence, the non-nuclear NATO partners in effect become nuclear powers in time of war”.[32]


The Nuclear Planning Group (NPG) was formed in 1966 to allow European NATO allies participation in nuclear decision-making as well as in discussions about the Alliance’s nuclear policy and doctrine. The NPG’s “Political Principles”, last updated in 1992, contain guidelines for nuclear planning, selective use of nuclear weapons and major nuclear response, consultations, and considerations for the employment of nuclear weapons.[33] 

European NATO members were given a political role in decisions on the use of nuclear forces under NATO command. However, they could not order their use. The US National Command Authority (NCA) retains the launch codes for use of US weapons in Europe and elsewhere. Thus the US NCA has ‘positive control’ over all these weapons: they cannot be armed without a US presidential decision. 

In peacetime, all US weapons strictly remain under custody of US forces. However, in 1969, then Chairman of the Joint Chiefs of Staff General Earle G. Wheeler told the Senate Armed Services Committee that: “All [nuclear] weapons in NATO with the exception, of course, of those possessed by the British, are under our custody and control at all times, and will remain so until there is a war, at which time the President can authorize the release of these weapons to our allies”.[34] (Emphasis added) In other words, once the President has given the order to use nuclear weapons, control over some US nuclear weapons can be handed over to non-nuclear NATO allies. After the order has been given and the aircraft has taken off with one (or more) armed weapon(s) on board, the weapon is no longer under national US command and control. Instead, the allied pilot now has full control over the weapon (or weapons) and has sole responsibility for delivering the weapon to its (predetermined) target. This is a form of nuclear proliferation, if under very special circumstances. Through the nuclear sharing arrangements, allied pilots are fully trained for nuclear missions, while training with dummy nuclear warheads during peacetime.


2.2 Would Such a Step Violate the NPT?  

“The treaty deals only with what is prohibited, not with what is permitted”.[35] 

The “release” of US nuclear weapons to NATO allies would appear to be inconsistent with Articles I and II of the NPT. However, NATO members continue to argue that their nuclear sharing arrangements are fully compatible with the Treaty. In fact, the argument about whether NATO nuclear sharing arrangements are compatible with the NPT dates back to the negotiation of the Treaty itself. 

The question of NATO nuclear sharing arrangements was at the center of US-Soviet negotiations on Articles I and II of the NPT. Originally, the US State Department discussed the creation of a Multilateral Force (MLF) which would have involved some type of “joint ownership, manning, and command of a NATO strategic force”.[36]  The Soviets strongly opposed the MLF, but were quiet on the question of nuclear sharing arrangements. It seems likely that there are two reasons for this. First, the US threatened to give up negotiating a NPT if the Soviet Union objected to nuclear sharing. And second, although the Soviets could not foresee circumstances in which they would wish to create a Warsaw Treaty Oganisation (WTO) MLF, the NATO nuclear sharing precedent might, if necessary, have proved useful for the Soviets in the WTO context, as it made possible similar arrangements within the WTO. 

It is likely that the widespread objections to the MLF and to nuclear sharing arrangements in general led the US and its allies to observe very tight secrecy when it came to the details of NATO’s sharing arrangements. In some cases, NATO allies were even forbidden to discuss these arrangements amongst themselves; only bilateral discussions with US authorities were permitted.[37] 

In May 1966, President Lyndon Johnson was reported to have instructed the State and Defense Departments “to seek new forms for nuclear coordination within NATO that might be less objectionable to the Soviets than an MLF”.[38] The compromise language for Articles I and II which was eventually agreed was intended by the US to close the option of an MLF or any multilateral nuclear-weapons entity, but not to preclude the existing NATO sharing arrangements, including the newly formed NPG. 

During the negotiations, the US informed the Soviet Union that a Treaty interpretation, indicating that allied consultation and “two-key” arrangements would not be barred, was essential for its allies. The Soviets were told that an interpretation to this effect would be made public during Senate hearings on the Treaty. They were told that if they publicly disagreed with this interpretation, the US would have to reconsider its position of support for the Treaty.[39] 

The background of US opinions of Articles I and II were forwarded to Secretary of Defense Clark Clifford prior to the NPG meeting at The Hague on 18-19 April 1968. Under Secretary of State Nicholas Katzenbach wrote: “I believe you should be familiar with the US interpretations of Articles I and II of the Non-Proliferation Treaty regarding alliance arrangements for nuclear defense. The FRG [Federal Republic of Germany] has requested in particular that we make it clear that the realization of the NPT will not affect the work of the NPG”.[40] In other words, the US and its NATO allies did not expect their mutual defense arrangements to be constrained by new treaty commitments.    

The key document on the US interpretation of Articles I and II is entitled Questions on the Draft Non-Proliferation Treaty asked by US Allies together with Answers given by the United States. (See Annex 1.) The Questions and Answers document was enclosed with a letter from Secretary of State, Dean Rusk, to President Johnson and thus is often referred to as the Rusk letter. It was transmitted to the Senate on 9 July 1968, along with other relevant documents, for consideration during the Senate ratification hearings on the NPT. According to available documents, this was first time the US made public its interpretation of the NPT, eight days after the NPT signing ceremony had taken place, at which the first 56 nations had signed the Treaty. 

The Questions and Answers was designed to give an interpretation of the NPT which allowed NATO nuclear sharing, based on the idea that the Treaty dealt only with prohibited matters. They indicate four areas that the Treaty “does not deal with” and therefore, in the US view, does not prohibit. 

The first question asked what may and may not be transferred under the Treaty. The US answered that the Treaty prohibits the transfer of nuclear bombs, warheads or nuclear explosive devices, but “does not deal with, and therefore does not prohibit, transfer of nuclear delivery vehicles or delivery systems or control over them”.[41] This interpretation was intended to allow continued cooperation on development of delivery systems and continued allied procurement of missiles, artillery systems and aircraft capable of delivering US nuclear weapons under NATO nuclear sharing arrangements. 

The second question  asked whether the treaty prohibits nuclear defense planning among NATO members. The US position stated that the NPT “does not deal with allied consultations and planning on nuclear defense so long as no transfer of nuclear weapons or control over them results”.[42] This answer was designed to allow information exchange within NATO’s system of nuclear sharing. This includes NNWS participation in Programs of Cooperation, drafting target plans, obtaining information about how different weapons would be used against different targets, and other aspects of the work of the Nuclear Planning Group, such as consultations on the use of nuclear weapons.[43] 

The third question was the most crucial one. It asked whether the Treaty prohibits the deployment of US nuclear weapons on NNWS NATO allies territory. In the US view, the NPT “does not deal with arrangements for deployment of nuclear weapons within allied territory as these do not involve any transfer of nuclear weapons or control over them unless and until a decision were made to go to war, at which time the treaty would no longer be controlling”.[44] 

This answer, that the NPT “would no longer be controlling” once a decision has been made “to go to war”, is crucial to the US interpretation. The US definition of “control” over the weapons, meaning that the weapons cannot be launched without a decision of the US president, is also critical. Thus, “control” over US nuclear weapons would only be transferred to NATO allies in the event of war, when the US considered that the Treaty was no longer binding. 

However, the implication of this interpretation is that the use of US nuclear weapons by allied forces would be illegal under the NPT if the Treaty was binding in wartime. Non-NATO diplomatic sources who confirmed knowledge of the Nuclear Planning Group have also stated that they knew nothing about nuclear sharing agreements at the time of signing and ratifying the NPT. However, they did know of these arrangements nearly 20 years later, when they included fresh language to the effect that the NPT would be controlling “under any circumstances” in the 1985 NPT Review Document. The formula was intended to close the NATO nuclear sharing loophole, and to ensure that all countries understood the NPT to be controlling at all times.[45] 

Three issues of political importance arise from the Questions and Answers. The first issue is the question of which states were informed of these interpretations of the draft NPT, how and when they were informed, and whether they can be considered to have consented to the US interpretations. The second issue results from weaknesses in the US definition of “control”. The third concerns the question of when the Treaty is considered to be “controlling”.

2.3 1968: When Were UN Members Informed?  

As early as 1966 Leonard Meeker, Legal Adviser to the US Department of State warned: “Should we decide to leave the wartime exception implicit we would want to make perfectly clear at Geneva what we were doing, lest we later be accused of having negotiated a treaty under false pretenses”.[46] 

In February 1969, eight months after the NPT signing ceremony, then Deputy Director of the US Arms Control and Disarmament Agency (ACDA) Adrian Fisher told the Senate Foreign Relations Committee that the Questions and Answers “were made available to key members of the ENDC [Eighteen Nation Disarmament Committee – the multilateral forum conducting negotiations on the treaty]. They have now been made available to all members of the UN, and an indication that this is the way the United States proposed to proceed. There has been no indication of objections”.[47] (Emphasis added) By depositing this statement in the US Senate records, it was assumed to be known by all NPT signatories.[48]  However, since even “key ENDC members”, such as the Swedes, appear to have been unaware of the existence of the details of NATO’s nuclear sharing arrangements, such as the Programs of Cooperation, the value of the Questions and Answers to them would have been limited. The US and its NATO allies were asking other nations to sign the NPT knowing they would be unaware of NATO plans to circumvent the Treaty, and making a pretense of informing of NATO’s intentions with a somewhat cryptic set of reservations. 

The Questions and Answers had been shown to NATO allies in early April 1967, 13 days prior to tabling of the draft treaty text at the Geneva-based ENDC. Therefore, NATO’s non-nuclear members were informed in time to indicate their consent or dissent to the US NPT interpretation. However, it appears, that with all likelihood they were the only non-nuclear nations which were fully in a position to do so at the time. 

Despite intensive efforts, the authors of this report were unable to obtain evidence which demonstrated that the contents of the Questions and Answers  were distributed to all UN members prior to 1 July 1968, when the first 56 nations signed the NPT or that the Questions and Answers was made publicly available prior to 9 July 1968. In a letter written by Under Secretary of State Nicholas Katzenbach to the Secretary of Defense dated 10 April 1968, it was made clear that it was deliberate policy to ensure the Rusk letter interpretations should become available to other signatories only after the Treaty was signed. “We do not believe it would be in our interest or that of our allies to have a public discussion of the US interpretations prior to the time when the NPT is submitted to the Senate for advice and consent”.[49] 

Fisher implies that the US was somewhat selective in revealing the details of its interpretations. He told the Senate Foreign Relations Committee that the Questions and Answers document was “seen by the Soviets and key members of the ENDC before it was made public and there was no objection. In view of the fact it is public, and has been referred to on a public hearing, I assume all countries in the world are on notice of our intention”.[50] 

As a consequence, states which were neither NATO members nor “key ENDC-members” may have had no formal notification of the details of the US interpretation prior to the publication of the US Senate hearings covering NPT ratification at the earliest – considerably after many states had signed the Treaty. It is highly likely, for example, that Ireland (the country which proposed the UN resolutions that led to the NPT) ratified the NPT on 1 July 1968 without prior information on this interpretation. Indeed, diplomatic sources involved at the time recall that the US deliberately acted to hide the Questions and Answers from members of the NAM until after the signing of the NPT. 

In effect, it is likely that those nations which sent representatives to US Senate hearings on NPT ratification, and gathered the relevant documents prior to or during the meeting, had access to the Rusk letter once the US ratification process was underway. Those who did not may have seen the document only after the proceedings of the hearings were available in print. 

Further, an important omission in the process is that the US did not deposit its interpretation when it signed the NPT. It is common practice for states to deposit their reservations and interpretations about a treaty upon submission of articles of ratification to the depositary body or state. Eighteen countries did so prior to or upon signing the NPT, some non-nuclear NATO members even referring indirectly to the Questions and Answers. The US, however, did not deposit any declaration, and consequently not informed additional countries  of its interpretation. 

Therefore, the question arises as to how non-NATO members could have given informed consent to nuclear sharing arrangements, the details of which remain classified to this day. Fisher admitted that even the Soviets had not “indicated acquiescence or agreement because they can’t be asked to agree about certain arrangements that we keep secret”.[51] 

Back in 1966, then US Secretary of Defense Robert McNamara had promised that the US was willing, “to make every effort to explain both our nonproliferation and our NATO nuclear sharing policies and to demonstrate beyond any reasonable doubt that there is no conflict between them”.[52] Neither the US nor NATO appears ever to have lived up to this commitment. For non-NATO members it must have been difficult or even impossible to judge or comment on the validity of US interpretations concerning nuclear sharing. 

Therefore, in order to determine the validity of US interpretations and the continuance of nuclear sharing under them, the US must state publicly which governments were informed about the contents of the Questions and Answers, and when they were informed. The question that remains is whether State parties would have objected to signing the NPT had they been aware of the full implications of the US interpretation. Would the NPT  be the globally accepted treaty it is, if all nations would have been fully aware of the US interpretations at the time they decided to join the NPT?[53]

2.4 What Constitutes Control?  

“The treaty should be void of any loop-holes which might permit nuclear or non-nuclear powers to proliferate, directly or indirectly nuclear weapons in any form”.  

Principle (a), UN General Assembly Resolution 2028 (XX), 19 November 1965. 

In 1965, the United Nations General Assembly called for a non-proliferation treaty that was “void of loop-holes”. However, by interpreting the NPT as dealing “only with what is prohibited” and not with what is allowed, the US introduced a major loop-hole for nuclear sharing. This directly contradicts the demand in the Treaty that it should be “In conformity with resolutions of the United Nations General Assembly calling for the conclusion of an agreement on the prevention of wider dissemination of nuclear weapons”.[54] 

All US administrations have consistently argued that NATO shared nuclear command and control is legal because the US guarantees to maintain positive control over all its nuclear weapons in peacetime. 

It is this interpretation that allows NATO NNWS to make every preparation in peacetime for the employment of nuclear weapons during war. Moreover, this interpretation introduces a loop-hole in the NPT that others may follow. For example, Belarus, a former Soviet republic, stated that the NPT was the “moral and legal guideline… that finalized the political decisions” behind its renouncement of its nuclear weapon regime.[55] However, Russia, with the acquiescence of Belarus, could reintroduce nuclear weapons on the territory of Belarus for use by Belorussian armed forces. China could decide to create nuclear sharing arrangements with North Korea, or Pakistan (not an NPT party) could do the same with Afghanistan, a NNWS party to the NPT. All this would be consistent with the US interpretation of the Treaty, supported by NATO allies. Simply put, NATO has established a pattern it does not want others to emulate. 

Even the US definition of  “control” poses other questions with regard to the NPT: what precautions has NATO taken to avoid a national pilot violating NATO orders after take off? What if he were to execute a mission different from his orders, such as a mission ordered by national rather than NATO authorities? What precautions has NATO taken to ensure all pilots return to their home base if a decision is taken to break-off nuclear operations? 

To further complicate the situation: to whom would the control of nuclear weapons be transferred when the decision is taken that NATO would go nuclear? Earlier discussions have focused on the question of whether loading a fully armed nuclear bomb onto an aircraft piloted by a citizen of a NNWS would constitute a transfer of control over the nuclear weapon to that NNWS. This question is legitimate, but it is not the only question that needs to be answered. In wartime, NATO’s national air force units would be assigned to the Alliance, a multinational entity. The US interpretation itself states that “the treaty would bar transfer of nuclear weapons (including ownership) or control over them to any recipient, including a multilateral entity”.[56] Yet no discussion has taken place as to whether a transfer of nuclear weapons to a multinational entity, such as NATO, would be legal under the NPT in wartime. This is why the wartime exclusion question is of core importance.

2.5 Does the Treaty Apply in Time of War?  

The US interpretation made public in the Rusk letter indicated that the Treaty would not be binding once a decision had been made to “go to war”.[57] Internal US interpretations at the time and statements by US officials indicate that this was taken to mean ‘general war’. (See below)  

The US government was aware that NATO arrangements were in conflict with Treaty provisions, so it had created a fallback position. During the NPT negotiations, the US pushed strongly to include wording which, in its interpretation, would make the Treaty invalid in times of war. In a 1966 secret memorandum, Adrian Fisher explained how the US should proceed with respect to this problem: 

Under these bilateral arrangements [within NATO], the US nuclear weapons available for use by allied forces assigned to NATO in the event of hostilities could, of course, be transferred to those forces in that event. This would be justified under the Atomic Energy Act on the ground that, when a war broke out, the President could exercise his power, as Commander-in-Chief, without regard to the ban on transfer contained in the Act. A similar interpretation would make a non-proliferation treaty inapplicable also. (Emphasis added)


The purpose of such a treaty, as the preamble could be expected to express it, would be to prevent the spread of nuclear weapons and, by this measure among others, to avoid the outbreak of nuclear war anywhere in the world. Thus the treaty has its application in time and in a situation when no conflict has broken out and when it continues to be possible to prevent such a conflict. Once general hostilities involving nuclear weapons have occurred, however, the point of prevention has been long passed, and the purpose of the treaty can no longer be served. In such circumstances the treaty would not apply, and a nuclear power would be free to transfer nuclear weapons to an ally for use in the conflict”.[58] 

As a result of the US position, the preamble of the NPT now states that the general purpose of the Treaty is “to avert such a war”. It does not, however, state that the Treaty is no longer controlling in time of war. 

During negotiations in 1966 Leonard Meeker, Legal Adviser to the US Department of State warned: “Should we decide to leave the wartime exception implicit we would want to make perfectly clear at Geneva what we were doing, lest we later be accused of having negotiated a treaty under false pretenses”.[59] 

Despite this, the US interpretation was expressed even in the Questions and Answers in very general terms. On 11 July 1968, Dean Rusk explained the US position further, stating: “I think... that this is simply a recognition of what today is almost an element of nature, and that is that, in a condition of general war involving the nuclear powers, treaty structures of this kind that were formerly interposed between the parties would be terminated or suspended”.[60] 

Nonetheless, the US administration did not provide a clear-cut definition of ‘general war’. Instead, Rusk simply gave an example of a conflict that would not relieve the signatories of compliance: “At the other extreme would be a limited, local conflict, not involving a nuclear-weapon-state. In this case the treaty would remain in force”.[61] 

The concept of ‘general war’ is incorporated in the famous NATO doctrine of ‘flexible response’, adopted in MC14/3 in December 1967. This paper, based on guidance issued by Ministers in May 1967 at the Defence Planning Committee, was the strategy in place during the final stages of negotiation of the NPT, and indeed until 1991. NATO regarded ‘general war’ as something unlikely because of the NATO deterrent. However, this definition also revealed that general war was now seen as a major conventional war leading to nuclear attack, to deter which NATO could threaten a ‘general nuclear response’. 

[I]t is unlikely that the Soviet Union will deliberately initiate a general war or any other aggression in the  NATO area that involves a clear risk of escalation to nuclear war.[62] 

Unlike in earlier strategies, ‘general war’ and nuclear war are no longer absolutely identical. This remains the case today. The US definition of ‘general war’ has remained unchanged since at least 1974. The DoD dictionary refers to ‘general war’ as “Armed conflict between major powers in which the total resources of the belligerents are employed and the national survival of a belligerent is in jeopardy”.[63] There is no definition of ‘major powers’ given. From this it is clear that a ‘general war’ can begin as conventional war, and can become a nuclear war at NATO’s instigation. This was made clear in MC14/3: 

In the event of a full-scale conventional aggression, indicating the opening of general hostilities on any sector of the NATO area, the forces of the Alliance should, if necessary, respond with nuclear weapons on the scale appropriate to the circumstances.[64] 

The term occurs, without explanation, in the 1991 NATO Strategic Concept. Paragraph 43 reads: “While in the new security environment a general war in Europe has become highly unlikely, it cannot finally be ruled out”.[65] 

The US and NATO should state publicly whether they still cling to the concept of “general war”, and if so, what its definition of that concept is. Only this kind of transparency will allow other NPT parties to understand in what circumstances NATO views the Treaty as no longer binding, and whether they believe that this is acceptable under the NPT. Unless this step is taken, other countries will not know if and when NATO would, in effect, create six new nuclear-weapon states. 

This issue concerned some diplomats in the 1980s as well. A more restrictive interpretation that Articles I and II apply “under any circumstances” was agreed by consensus in the Final Document of the third NPT Review Conference in 1985. While this final consensus statement is politically rather than legally binding, the 1985 wording was confirmed by the UN Security Council in its 1991 resolution on Iraq. The Gulf War was a regional conflict with the potential for involving weapons of mass destruction. 

The US is the only country that has explicitly stated that, once a general war has begun, it would no longer feel bound by the NPT. It has thus created a loop-hole by which it could withdraw from the Treaty without the three month notice period required by NPT Article X. In addition, the US approach implicitly creates a loop-hole for NNWS members of NATO to withdraw from the Treaty and receive US nuclear weapons in the event of war. 

Furthermore, NATO is able to create the very conditions under which it would no longer feel bound by the NPT. By retaining the option of first use of nuclear weapons, the interpretation allows the US to decide unilaterally when ‘general war’ has come and thus when it can withdraw from its NPT commitments without prior notice. The first use of nuclear weapons by the US and NATO during a conflict would not occur “unless and until a decision were made to go to war”. The US view is that “in such circumstances the treaty would not apply, and a nuclear power would be free to transfer nuclear weapons to an ally for use in the conflict”.[66] 

 <-Front Page     /     Next Chapter->


BITS Homepage