Foreign Relations of the United States, 1989–1992, Volume XXXI, START I, 1989–1991

27. Memorandum From Richard Davis of the National Security Council Staff to the President’s Deputy Assistant for National Security Affairs (Gates)¹

Washington, June 6, 1989

SUBJECT

• Deputies Committee Meeting, Wednesday, June 7, 1989, 3:00 p.m.—START Limits on Air Breathing Systems and Nondeployed Missiles

This Deputies Meeting will be the second of three meetings to review the specifics of our START position before the negotiations resume June 19. In all of the START papers, the PCC has attempted to focus on issues where changes are necessary because we could not take “yes” for an answer; or where we see an opportunity to put a Bush imprint on START. In certain cases, options not meeting these standards have been included because an agency favors the option. We recommend that you guide the meeting through the issues in the order they are presented in the papers. Your goal should be to get the Deputies to winnow the options to a more manageable number and ensure the remaining alternatives are fairly stated.

1. Air Breathing Systems. A fundamental premise underlying the PCC’s work on this issue is the longstanding U.S. opposition to tight constraints on ALCMs and other bomber weapons. The NSC meeting earlier in the day may take up the underlying policy question of whether it is in the long-term U.S. interest (a) to encourage the Soviets to increase their emphasis on air breathing systems, and (b) whether our current approach, which emphasizes U.S. force flexibility at the expense of verification and constraints on Soviet breakout, makes the best tradeoffs among our competing objectives.

We agreed to limits on ALCMs and other bomber weapons under the 6,000 warhead limit only on the understanding that they be heavily discounted. While we and the Soviets have agreed to count all the bombs and short-range missiles on each bomber as one warhead under 6000, we have not agreed on several ALCM provisions. This liberal discounting approach to air breathing systems ensures that we will have enough weapons to meet our deterrence “requirements” notwithstanding the cuts mandated in START.

The options paper (at Tab A) addresses three specific issues arising from differences between the U.S. and Soviet proposals: 1) counting rules for ALCMs, 2) a range threshold to distinguish between short-range air-to-surface missiles and ALCMs, and 3) how to distinguish between nuclear and conventional ALCMs. An additional air-breathing system—SLCM—is not addressed because no agency favors changing our current proposal now (although there are divergent views on what our ultimate position should be).

Counting Rules. There are two basic questions: how far should we go in encouraging greater Soviet reliance on air breathing weapons? What restrictions on the U.S. heavy bomber forces should we accept?

We have proposed to attribute each specifically designated “ALCM heavy bomber” with ten warheads under the 6000 limit, regardless of their true capacity to carry ALCMs. This position sidesteps the need to verify actual ALCM loadings, which could change rapidly, and permits us to deploy ALCMs on the bombers we need, and in the numbers we want, without pressing against START limits.

However, the Soviets claim that our attribution rule is unfair since future wide-bodied aircraft could carry dozens of ALCMs but would only be attributed with ten. Their proposal to count “ALCM heavy bombers” based on the maximum number of ALCMs they are capable of carrying would either (a) require modifications to U.S. bombers to limit their ALCM capacity or (b) lead to unrealistically high ALCM loadings, even though they would severely inhibit our flexibility.

Verification of specific bomber loadings is virtually impossible because a bomber’s capability to carry ALCMs can be changed rapidly. The U.S. position technically can be verified only because it defines away the problem of counting ALCM capacity. No proposal, however, will meaningfully limit the Soviet ALCM breakout threat to less than the ALCM capacity of their entire heavy bomber force.

The options in the PCC paper reflect a range of views about the trade-offs between restricting the size of the future legal Soviet ALCM force (recognizing that ALCMs could be deployed on all heavy bombers in a breakout), and protecting U.S. options for a larger ALCM force than we now envision under START. In evaluating these options, a complicating factor is the reduced production rate for the B–2 penetrating bomber and Congressional concern with B–2 costs. If the B–2 program is significantly delayed, ALCM options and flexibility could become more important for the U.S.

— Option 1—Accept the Soviet demand for an “as equipped” counting rule for ALCMs.

— Option 2—Retain the existing ALCM attribution rule of ten per bomber, but add a limit on the maximum ALCM carrying capability of individual bombers (this would respond to the Soviets’ stated concern that we could deploy wide-bodied aircraft with dozens of ALCMs).

— Option 3—Do not change the U.S. position on counting rules; retain the existing ALCM attribution rule of ten per bomber.

— Option 4—Return to the earlier U.S. position that each heavy bomber equipped for ALCMs should count as six warheads, regardless of the number actually carried.

— Option 5—Eliminate all limits on ALCMs, retaining only an overall limit on heavy bombers.

ALCM Range and Nuclear-Conventional Distinguishability. Protecting future U.S. options for conventionally-armed ALCMs from START limits has been our overriding objective. At issue is whether this emphasis should continue to determine our approach. The JCS feel strongly that it should.

There are two complementary ways to protect conventional ALCMs from being captured in START: we can distinguish between nuclear and conventional ALCMs and limit only the nuclear missiles; or we can set the minimum range for limited ALCMs high enough to exclude all conventional ALCMs. The U.S. position does both.

We have agreed with the Soviets that all currently existing long- range ALCMs will be considered to be nuclear-armed and subject to the treaty (all are in fact nuclear), and that future conventional ALCMs will either be distinguishable from nuclear-armed cruise missiles by NTM or will be considered to be nuclear-armed (distinguishability details have not been worked out). We have proposed a range threshold of 1500 km; the Soviets have proposed 600 km, as was agreed in SALT II.

Two concerns underlie our 1500 km threshold proposal. First, we and the Soviets may not be able to agree on acceptable distinguishability provisions and, in the end, all conventional as well as nuclear ALCMs over the range threshold would be limited, as are GLCMs in INF. A second concern is that possible improvements to Soviet air defenses which would require extended missile range (although there are no current U.S. programs for nuclear-armed cruise missiles with ranges between 600 and 1500 km.)

In establishing the U.S. position, we recognized that the nuclear-conventional distinction could not be verified. Also, there is factor of two uncertainty in monitoring the range capability of an ALCM. Verification can be improved if START treats all ALCMs as though they are nuclear-armed. Verification is also improved (i.e., the military significance of cheating is reduced) as the range threshold is lowered. Both changes, however, limit future U.S. options. The primary consideration has been to protect U.S. programs, not to attempt to constrain Soviet ones.

There are four options on the range issue and three on distinguishability. We recommend that you do not discuss the options in detail, but use this meeting to:

— hear out the JCS on the specific types of programmatic flexibility we need;

— discuss the flexibility/verification tradeoff; and

— begin the discussion of what kinds of tradeoffs between range and distinguishability we should consider.

Because the ALCM issues are closely interrelated, we probably want to resolve them in a package with the Soviets. But we do not need to deal with that issue now. The first task should be to flesh out our distinguishability position in Washington and begin to nail down an agreement with the Soviets. At that point, the path through the remaining issues will be clearer.

2. Nondeployed Missiles. There are two issues in the options paper (Tab B): should we modify our current proposal on nondeployed missiles, and should we require a missile to be destroyed whenever a launcher is destroyed.

Nondeployed Limits. Limits on nondeployed missiles has long been one of the most controversial parts of our proposal. In 1982, President Reagan decided to limit nondeployed ballistic missiles (NDM) in START, in order to deal with what some believe to have been one of the critical flaws of the SALT II Treaty—its failure to control missiles which could be used for refire, force reconstitution, and ballistic missile breakout. In subsequent years the following considerations came to light:

— Despite years of study, no way has been found to monitor inventories of NDMs accurately, especially existing missiles.

— The JCS and others have questioned the military utility of NDMs to the Soviets, especially SLBMs and silo-based ICBMs.

— U.S. requirements for NDMs (although not internally agreed) may be higher than Soviet requirements because we plan to produce missiles at the most economical rate and stockpile missiles for use in future years. Stretching ballistic missile production could result in significantly increased costs; for example, the Congressionally mandated stretchout of Peacekeeper from 5 to 14 years increased unit cost from $16 million to $60 million.

— Our commitment to support the U.K. TRIDENT program requires that nondeployed D–5s for British SSBNs be stored in the United States and, for reliability testing, procedures must allow for intermingling British and U.S. NDMs.

— There are a number of technical issues with respect to tagging and it is not yet certain an acceptable tagging regime can be devised. Some believe that the inability to tag missiles would significantly undermine any NDM limits.

These considerations led President Reagan to approve modifications that focus our NDM proposal on the types of missiles that pose the most significant refire threat; namely, mobiles and future missiles that could be designed to be made mobile. OSD sought successfully to have heavy ICBMs added to the list of constrained types. We still have not been able to agree internally what an appropriate numerical limit would be.

The options reflect a wide range of agency views on the relative weight of the factors above. OSD believes we have retreated from an important objective in START—to limit missiles, not launchers—and is seeking to reinstate NDM limits on most, if not all, missiles. State and ACDA would drop heavy ICBMs from NDM limits and have proposed perfecting amendments to our existing position. JCS prefers no limits on NDMs because they worry about the economic impact and are not concerned about Soviet reloads.

Your objective on this issue is to confirm that the issues and options are fairly stated. Although it would greatly facilitate subsequent consideration of this issue, we do not anticipate any narrowing of views at this point.

The options (from most to least inclusive) are:

— Option 1—Require numerical limits on mobile ICBMs, heavy ICBMs, modern ICBMs (including SS–17 and 19) and all future types of ICBMs or SLBMs. (OSD)
Includes SS–17, 18, 19, 24 and 25, Peacekeeper, Small ICBM, and all future systems; intended to avoid capturing the D–5.

— Option 2—Require numerical limits on mobile ICBMs, heavy ICBMs, and modern ICBMs (excluding SS–17 and 19). (Current position)
Includes SS–18, 24 and 25, Peacekeeper, Small ICBM, and any future systems meeting the “modern ICBM” definition.

— Option 3—Require numerical limits on mobile ICBMs and modern ICBMs and SLBMs (excluding SS–17 and SS–19). (ACDA)
Includes SS–24 and 25, Peacekeeper, Small ICBM, and any future modern SLBMs or ICBMs.

— Option 4—Require numerical limits on mobile ICBMs and “modern” ICBMs (excluding SS–17 and 19). (State)
Includes SS–24 and 25, Peacekeeper, Small ICBM, and any future systems meeting a definition of “modern ICBM” which would be modified to capture only ICBMs that are transported fully fueled.

— Option 5—Require numerical limits on ICBMs that have been deployed in a mobile mode. (JCS fallback)
Includes SS–24 and 25, Peacekeeper (if deployed in a mobile mode), Small ICBM, and any future mobile ICBMs.

— Option 6—Do not require numerical limits on any nondeployed missiles. (JCS)

Includes no systems.

Missile Destruction. When the U.S. dropped NDM limits from some categories of missiles, our START position no longer guaranteed that missiles would be destroyed as part of START reductions. To address the political and possible military implications of this change, a study was made of an additional treaty requirement that for every launcher destroyed in reducing to new limits, one missile associated with that type of launcher must also be eliminated. All agencies concluded that such a provision was more complicated than expected and of limited military benefit. The current U.S. approach would not require the elimination or conversion of any Soviet SLBMs or existing silo-based ICBMs other than SS–18s, but could result in such missiles simply being placed in a nondeployed status.

The primary issue here is whether the political benefit of requiring visible missile destructions is worth the price of giving up missiles for such uses as RDT&E boosters, space launch vehicles, operational spares, flight testing, or conversion to other ballistic missiles. U.S. production lines for existing missiles (except MX) are now closed, and the destruction of missiles followed by the prohibitively expensive reopening of old production lines would not be attractive.

There are two options:

— Option 1—Retain the existing U.S. position; do not add an additional treaty requirement that for every launcher destroyed in reducing to new limits, one missile associated with that type of launcher must also be eliminated or converted to other purposes as space launch vehicles.

— Option 2—Require missile conversion or elimination incident to launcher destruction. Task an interagency study of the best approach for implementation.

We believe you should try to achieve a consensus on option 1 for now. The impact of option 2 on U.S. programs as well as the political impact, cannot be determined until we decide on a NDM proposal, including a specific numerical limit. A low numerical limit may cause some visible missile destructions, even without this destruction requirement; a high limit may undercut the impact of any required destructions because modern missiles could replace the older ones destroyed. We believe a missile destruction proposal is worth pursuing as part of fleshing out the numbers in our NDM proposal, but no decision need be taken before we return to Geneva.

RECOMMENDATION

That you read the papers at Tab A and B and use the points outlined above in guiding discussions at the Deputies Meeting on START Air breathers and Nondeployed limits.²

Bob Blackwill and Condi Rice concur.

Tab A

Paper Prepared by the Policy Coordinating Committee³

Washington, May 6, 1989

START 04C

NSR 14 START Review Paper Four—Air Breathing Systems

Purpose. To review U.S. positions on dealing with air breathing systems under START.

Background. A fundamental premise underlying the U.S. approach to START is that we must be capable of maintaining an adequate deterrent under a START regime. In order to do so, and to improve stability, the U.S. has consistently opposed tight constraints on ALCMs and other bomber weapons in START. The original U.S. position did not directly limit these systems, constraining them only indirectly through an overall limit on the number of heavy bombers. If ALCMs and other bomber weapons are to count under the 6,000 warhead limit in START, as currently agreed, they must be heavily discounted in order to allow meeting present and START projected U.S. military requirements.

Significant features of the current U.S. approach include:

— Counting bombers which carry only gravity bombs and SRAMs as “one” against both the 1600 SNDV and 6000 warhead limits.

— Counting ALCM-carrying heavy bombers as one unit against the 1600 SNDV limit and 10 warheads against the 6000 warhead limit regardless of actual load or capability.

— Considering “ALCMs” to include only nuclear-armed air- launched cruise missiles with ranges in excess of 1500 kilometers.

— Rejecting any limits on long-range conventionally-armed cruise missiles (LRCCM). Treating all currently existing long-range air-to-surface cruise missiles as nuclear armed while distinguishing future LRCCMs from treaty accountable ALCMs through NTM.

— Rejecting any limits on nondeployed ALCMS or on total ALCM inventory.

— Providing for a limited number of former heavy bombers (i.e. conventional bombers) which would not count against either the 1600 or 6000 limits.

— Providing separate basing for the three classes of bombers (ALCM-carrying heavy bombers, non-ALCM carrying heavy bombers, former heavy bombers), with a limited on-site inspection right for non-ALCM carrying heavy bombers of a type tested with treaty accountable ALCMs.

In establishing this position, the United States recognized that the Soviets would have an equal right to exploit the rules. The United States judged, however, that protecting our ability to cover and damage targets was more important than constraining similar Soviet capabilities.

Discussion—General. Because of the importance of this area and the close interrelationship of all heavy bomber/ALCM positions, a comprehensive review is required. There is consensus that the United States should not alter its current rejection of Soviet attempts to place sublimits on bomber weapons or to constrain ALCMs, e.g., to limit nondeployed ALCMs.

Discussion—Counting Rules. For many years the United States rejected placing any limits on ALCMs. The U.S. approach to ALCM counting rules discounts ALCMs and heavily discounts non-ALCM nuclear weapons on bombers. Some have raised the question of whether discounting bomber weapons would in fact favor the Soviets in the long run, given the asymmetries in air defenses. Others believe that, while the Soviets appear to be well positioned to exploit these counting rules, encouraging greater Soviet reliance on air-breathing systems is a central goal of the U.S. START position. Moreover, the U.S. requires these systems for target coverage, while the Soviets do not.

The most common alternative counting rule is the so-called “as- equipped” approach, in which ALCM bombers are counted against the 6000 limit based on the maximum number of ALCMs they are capable of carrying. This position, favored by the Soviets, would either (a) require substantial modifications to U.S. bombers to limit the number of ALCMs they could carry or (b) lead to unrealistically high attributions which would severely inhibit U.S. flexibility. The Soviets claim that the U.S. attribution rule provides the U.S. with an unfair advantage since future widebodied aircraft could carry dozens of cruise missiles but would only be attributed with ten. Some favor attempting to meet this stated Soviet concern by placing some upper limit on the ALCM-carrying capacity of individual aircraft. Others believe the loss of flexibility and verification problems make it imprudent to attempt to placate what they see as a spurious Soviet concern.

To move toward a more stabilizing defense posture, some believe the U.S. could consider readopting its original position that ALCMs and other bomber weapons will not be counted or limited under START. This would entail having limits only on ballistic missile warheads and on strategic nuclear delivery systems. Heavy bombers would continue to count against the latter limit only. Others believe that this move would encourage competition in air-breathing systems and, because of the Soviet lead in air defenses, therefore would not be stabilizing. Moreover, this move could incite strong Soviet condemnation as backtracking on a key U.S. concession and, by reopening the Reykjavik bomber counting rule, could actually decrease the U.S. ability to deploy the necessary high number of bomber weapons. On the other hand, such a move would simplify U.S. obligations under the treaty, since no distinctions would have to be made (or verified) regarding conventionally-armed versus nuclear-armed ALCMs, or heavy bombers equipped for ALCMs vice those not so equipped.

A less radical move would be to return to the earlier U.S. position that each heavy bomber equipped for ALCMs should count as six warheads (instead of ten), regardless of the number actually carried. While this position would not simplify the treaty in any way, and could encounter Soviet opposition, it could increase U.S. flexibility and confidence in meeting targeting requirements under START constraints.

In evaluating these options, a complicating factor is the reduced production rate for the B–2 and Congressional concern with B–2 costs. If the B–2 program is significantly delayed, ALCM options and flexibility could become more important for the United States.

Options—Counting Rules. The following options exist:

— Option 1—Accept the Soviet demand for an “as equipped” counting rule for ALCMs.

— Option 2—Retain the existing ALCM attribution rule of ten per bomber, but add a limit on the maximum ALCM carrying capability of individual bombers.

— Option 3—Make no change to the U.S. position on counting rules; retain the existing ALCM attribution rule of ten per bomber.

— Option 4—Return to the earlier U.S. position that each heavy bomber equipped for ALCMs should count as six warheads, regardless of the number actually carried.

— Option 5—Eliminate all limits on ALCMs, retaining only an overall limit on heavy bombers.

Discussion—ALCM Range. The U.S. has taken the position that nuclear-armed air-to-surface missiles with a range of less than 1,500 km should not be counted in START. The Soviets would establish the ALCM range threshold at 600 km. A range cut-off of 1500 kilometers protects U.S. options to extend the range of missiles on both heavy bombers and on tactical aircraft to offset improved Soviet air defense, without having such missiles count as ALCMs under START. Although there are no current programs for nuclear armed cruise missiles with ranges between 600 and 1500 kilometers, some believe there is a clear military requirement to protect the option to deploy such systems in the future. Moreover, the 1,500 km range better protects U.S. options for future cruise missiles pending resolution of how future nuclear and nonnuclear cruise missile are to be distinguished from each other.

The U.S. has indicated some flexibility on the range issue in return for Soviet acceptance of the remainder of the U.S. ALCM- heavy bomber positions, although we have not specified what range we might accept. Some believe this flexibility should be withdrawn, either permanently or until satisfactory resolution of the ALCM distinguishability problem and definition of the military requirements for theater standoff nuclear weapons in a post-CFE environment.

Options—ALCM Range. The following options exist:

— Option 1—Accept a 600 kilometer range cutoff, contingent on resolution of:

— (Option 1A) the ALCM distinguishability issue; or

— (Option 1B) all remaining ALCM issues.

— Option 2—Accept a 1000 kilometer range cutoff, contingent on resolution of:

— (Option 2A) the ALCM distinguishability issue; or

— (Option 2B) all remaining ALCM issues.

— Option 3—Maintain the existing U.S. position of a 1500 kilometer range cutoff, but continue to indicate flexibility in return for Soviet acceptance of other elements of our ALCM position.

— Option 4—Maintain the existing U.S. position of a 1500 kilometer range cutoff; do not indicate any flexibility.

Discussion—ALCM Distinguishability. This issue is related to the previous two. We have agreed with the Soviets that all currently existing long-range (with the range threshold still unresolved, as noted above) air-to-surface cruise missiles will be considered to be nuclear armed, and that future conventionally armed long-range air-to-surface cruise missiles will either be distinguishable from nuclear-armed cruise missiles by NTM or will be considered to be nuclear armed.

In establishing this position, the United States recognized that it provided no assurance against supposed conventional cruise missiles being illegally equipped with nuclear warheads. Some believe this is a serious problem, since it provides the Soviets with an option for deploying “conventional” cruise missiles (which we could not reliably verify not to be nuclear armed) on a wide variety of aircraft, thus increasing the threat to the United States. Others believe that the primary consideration is protecting U.S. programs, not constraining Soviet ones.

While NTM can distinguish physical features of different cruise missiles, it is not an effective means of determining whether those cruise missiles carry nuclear or conventional warheads. There may be operational penalties from designing future conventional cruise missiles such that they are physically distinguishable. In addition, weapons carried internally in a bomber weapons bay cannot be seen by NTM. Finally, the U.S. may find the requirement to display systems to NTM onerous from a operational, security or logistics standpoint.

Options—ALCM Distinguishability. The following options exist:

— Option 1—Reconsider our basic position on ALCM distinguishability and agree to treat all cruise missiles over a specified range (1500 kilometers) as nuclear armed.

— Option 2—Reaffirm our existing approach to ALCM distinguishability, accepting the drawbacks indicated above as the necessary price for avoiding constraints on conventional cruise missiles.

— Option 3—Reconsider our existing approach. Undertake a study of alternatives. If (as is probable) this study is not complete when negotiations resume, inform the Soviets that we have the ALCM distinguishability issue under review.

Tab B

Paper Prepared by the Policy Coordinating Committee⁴

Washington, May 27, 1989

START 03F

NSR 14 START Review Paper Three—Nondeployed Missiles

Purpose. To review options for limiting nondeployed ballistic missiles under START.

Background. In 1982, President Reagan decided to limit nondeployed ballistic missiles (NDM) in START, in order to deal with one of the critical flaws of the SALT II Treaty—its failure to control missiles which could be used for refire, force reconstitution, and ballistic missile breakout. The decision also attempted to force the Soviets to destroy ballistic missiles taken out of operational service. Until 1988 the U.S. proposal called for a to-be-determined aggregate numerical limit on all nondeployed ICBMs and SLBMs.

Following interagency review, in August 1988, President Reagan modified the U.S. START position. Reasoning that mobile systems (SS–24 and SS–25) and highly lethal systems (SS–18 and SS–24) posed the most serious threats to the United States, the President exempted SLBMs and older fixed-silo ICBMs from numerical NDM limits. Under the new guidance, only heavy ICBMs, mobiles, and “modern ICBMs comparable to the SS–24” would have numerical NDM constraints. Heavy ICBMs and moderns would have a single, fixed numerical limit. Mobile ICBMs, if allowed, would have a second fixed, numerical limit. Different verification regimes apply to the two different limits. The United States has not determined what an appropriate numerical limit would be in either case.

The President’s decision is at Tab A,⁵ an extract from a paper used by the President is at Tab B.⁶

In the U.S. definition tabled in Geneva, a mobile ICBM is one which “has been contained in a road- or rail- launcher of ICBMs after______,⁷ or that has been launched from a rail- or road-mobile launcher of ICBMs.” The Soviets have been told that the date criteria is intended to exclude Minuteman I, which was once transported in a mobile launcher for evaluation purposes.

Similarly, a “modern” ICBM is one “that uses a launch canister or has been transported fully assembled, with a throw-weight of 1,900 kilograms or greater or a launch weight of 60,000 kilograms or greater.” We have not told the Soviets which missiles this definition is intended to capture.

The U.S. intent was to capture the SS–24 and SS–25 under the definition of a mobile ICBM. The intent on “modern” ICBMs was to constrain larger Soviet ICBMs which could be made mobile in a breakout scenario. The United States assumed that Soviet “modern” ICBMs do not now exist and thus that the limit on modern ICBMs would apply to future systems only. A strict reading of the tabled definition of “modern” ICBM, however, reveals that it inadvertently captured the Soviet SS–17 and SS–19. No explicit decision was made on whether Peacekeeper would be captured under the modern ICBM definition if mobile ICBMs were not allowed. Peacekeeper would be captured if deployed as a mobile ICBM; most believe it would be captured as a modern ICBM if it is not deployed in a mobile mode. (This may or may not be legally correct based on our present tabled position; it is almost certainly correct as a practical negotiating matter.)

Issue 1—NDM Limits. The United States position places a number of constraints on nondeployed ballistic missiles, including geographic restrictions, data exchange and movement notifications, all of which would continue to apply regardless of the decision on numerical limits on NDMs. This paper deals exclusively with the issue of such numerical limits.

The most important consideration in reviewing the need for such limits is the degree to which they are needed to preclude Soviet refire, force reconstitution, and ballistic missile breakout. Soviet nondeployed missiles could be used to reload and refire existing launchers, if the necessary transportation and handling equipment was available and survived the initial exchange. Soviet nondeployed missiles could also be deployed in a crisis on soft launchers or on covert mobile launchers brought out of storage. These issues are discussed in Tab B, used in the 1988 decision, and Tab C.⁸ Three other factors are relevant:

— There is not yet agreement on the number of non-deployed ballistic missiles the United States will require. Past estimates range from several hundred to well over one thousand, not all of which would be retained for operational purposes. Because of U.S. production practices, the United States may have a requirement for a greater number of NDMs for a given force size than does the Soviet Union; we plan to produce at the most economical rate and stockpile missiles for use in future years. Stretching ballistic missile production could result in significantly increased costs; for example, the Congressionally mandated stretchout of Peacekeeper from 5 to 14 years increased unit cost from $16 million to $60 million. Production stretchouts also lead to reliability issues because of changing production populations. Additional detail on U.S. and Soviet production and NDM stockpile requirements is at Tab D.⁹

— The United States is committed to support the United Kingdom TRIDENT program. Non-deployed TRIDENT II (D–5) missiles for United Kingdom SSBNs will be stored in the United States. Since British reliability figures are to be based on the entire D–5 population, procedures must allow for treating British and U.S. NDMs as part of the same population.

— There are a number of technical issues with respect to tagging and it is not yet certain an acceptable tagging regime can be devised. Some believe that the inability to tag missiles would significantly undermine any NDM limits. Others believe that an acceptable tagging regime, if developed, would strengthen our ability to discriminate legal from illegal stages, and note that the U.S. ability to verify compliance with NDM limits is not solely dependent on tagging.

Options. A large number of possible NDM regimes exist. All fall within one of two broad approaches:

— Seek numerical limits on no nondeployed ballistic missiles.

— Seek numerical limits on some nondeployed ballistic missiles.

These approaches are discussed further at Tab C. Six options have been identified:

— Option 1—Require numerical limits on mobile ICBMs, heavy ICBMs, modern ICBMs (including SS–17 and 19) and all future types of ICBMs or SLBMs. (OSD)
Includes SS–17, 18, 19, 24 and 25, Peacekeeper, Small ICBM, and all future systems; intended to avoid capturing the D-5.

— Option 2—Require numerical limits on mobile ICBMs, heavy ICBMs, and modern ICBMs (excluding SS–17 and 19). (Current position)
Includes SS–18, 24 and 25, Peacekeeper, Small ICBM, and any future systems meeting the “modern ICBM” definition.

— Option 3—Require numerical limits on mobile ICBMs and modern ICBMs and SLBMs (excluding SS–17 and SS–19). (ACDA)
Includes SS–24 and 25, Peacekeeper, Small ICBM, and any future modern SLBMs or ICBMs.

— Option 4—Require numerical limits on mobile ICBMs and “modern” ICBMs (excluding SS–17 and 19). (State)
Includes SS–24 and 25, Peacekeeper, Small ICBM, and any future systems meeting a definition of “modern ICBM” which would be modified to capture only ICBMs that are transported fully fueled.

— Option 5—Require numerical limits on ICBMs that have been deployed in a mobile mode. (JCS fallback)
Includes SS–24 and 25, Peacekeeper (if deployed in a mobile mode), Small ICBM, and any future mobile ICBMs.

— Option 6—Do not require numerical limits on any non-deployed missiles. (JCS)
Includes no systems.

Issue 2—Missile Destruction. President Reagan’s August 1988 guidance directed evaluation of an additional treaty requirement that for every launcher destroyed in reducing to new limits, one missile associated with that type of launcher must also be eliminated. Following evaluation, all agencies concluded that such a provision was more complicated than expected and of limited military benefit. Moreover, the U.S. position would permit conversion of a limited number of ballistic missiles to space launch vehicles, further limiting the impact of an elimination requirement. As a result, no additional treaty requirement was developed. Thus, the existing U.S. approach would not require the elimination or conversion of any Soviet SLBMs or existing silo-based ICBMs other than SS–18s, but could result in such missiles simply being placed in a nondeployed status.

The magnitude of the problem this proposal is intended to address depends on: how many missiles are exempted from NDM limits (see Issue 1), how many NDMs are permitted in the constrained categories, and how many space launch vehicles (SLVs) may be converted from NDMs in the constrained categories. If there are no NDM limits, then there need be no missile eliminations. If the numerical NDM and SLV limits are set relatively high, or if several categories of existing missiles are excluded from NDM and SLV limits, then few, if any, missiles would have to be eliminated. In either of these cases, a missile elimination requirement could result in visible missile destructions. The perceptual impact of these destructions would be mitigated, however, if NDM/SLV limits on future types were so high that a side could build a significant legal inventory of refire/breakout missiles.

One view is that requiring the elimination of a missile for every launcher destroyed has all the drawbacks of an NDM limit with none of the advantages. Some believe that we will have a pressing need for missiles removed from launchers to support force restructuring and for such uses as RDT&E boosters, space launch vehicles, operational spares, flight testing, or conversion to other ballistic missiles. Such requirements may run to several hundred missiles. Production lines for the bulk of our existing missiles are now closed. We simply cannot afford to destroy millions of dollars worth of missile inventory only to subsequently be compelled to reopen those production lines at a cost of many more millions.

Eliminating missiles removed from operational deployment would still allow the Soviets to retain refires and produce new missiles. The United States, however, would be unable to retain missiles for research and testing.

Some would attempt to balance the political requirement for forcing elimination with the requirements for continued use of removed missiles. Approaches to doing this might attempt to meet following objectives:

— Treating the initial seven years of the treaty, when both sides are engaged in reducing to new limits and thus the objective is to lower overall capability, differently from the subsequent period when sides are modernizing forces within a given set of constraints.

— Focusing on removal from accountability in accordance with the Protocol on Conversion or Elimination (which provides for both elimination and limited conversion to space launch vehicles), rather than strictly on elimination.

— Allowing elimination through expenditure, for example in Operational Tests.

— Exempting from any requirements for removal from accountability situations in which old launchers are replaced with new launchers but the number of deployed missiles remains essentially the same (e.g. rebasing Peacekeeper from silo to rail, or replacing POSEIDON C–4 launchers with TRIDENT C–4 launchers).

Past attempts to devise a treaty approach to meet these and similar objectives have proven technically difficult; thus no specific approach can be adopted without additional examination.

Options. There are two options:

— Option 1—Retain the existing U.S. position; do not add an additional treaty requirement that for every launcher destroyed in reducing to new limits, one missile associated with that type of launcher must also be eliminated or converted to other purposes as space launch vehicles.

— Option 2—Require missile conversion or elimination incident to launcher destruction. Task an interagency study of the best approach for implementation.

1. Source: George H.W. Bush Library, Bush Presidential Records, National Security Council, H-Files, NSC/DC Meetings Files, NSC/DC 032—June 7, 1989—NSC/DC Meeting on NSR–14 re: U.S. Start Position on Air Breathing Systems and Non-Deployed Missiles. Secret. Sent for action. Sent through Kanter. A stamped notation indicates Gates saw the memorandum.
2. Gates did not indicate his preference.
3. Secret.
4. Secret. All blank underscores are in the original.
5. Attached but not printed. Tab A is an August 20, 1988, memorandum from Reagan to Bush, Shultz, Weinberger, Herrington, Webster, Crowe, and Burns, titled “Nondeployed Ballistic Missiles in START.”
6. Attached but not printed. Tab B is an extract from a July 25, 1988, Arms Control Support Group Paper.
7. Omission is in the original.
8. Attached but not printed. Tab C is an undated paper titled “Alternate Approaches.”
9. Attached but not printed. Tab D is an undated paper titled “Soviet Requirements for Non-Deployed Missiles.”