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

43. Memorandum From the Undersecretary of State for Security Assistance, Science and Technology (Bartholomew) to Secretary of State Baker¹

Washington, September 1, 1989

SUBJECT

• Arms Control at the September Ministerial

What We can Accomplish

I sent you a planning memo August 7² outlining State’s arms control work program for the September Ministerial. Most of the work has been completed by a State in-house group consisting of T, PM, EUR, S/P, S/DEL, and INR. We can make solid progress along a broad front, if we can overcome interagency opposition on various points. In particular, we can:

— Nuclear Testing Talks (NTT): reach agreement on major issues in the verification protocol for the Threshold Test Ban Treaty (TTBT), allowing the delegations to complete the protocol, possibly in time for signature at a spring summit;

— Strategic Arms Reduction Talks (START):

— Air-launched Cruise Missiles (ALCMs): agree to a package that paves the way to resolving this major START issue;

— Verification and Stability (V/S): reach agreement in principle on two or three of the seven elements of the President’s V/S package, and make progress on the others; and

— Chemical Weapons (CW): sign an MOU establishing an early bilateral data exchange and visits to a number of storage sites and production plants.

What Needs to be Done

There is interagency opposition to these moves (see comments below), and in order to accomplish the above we will need decisions from the President on a number of contested issues relating to NTT, START, and CW:

Nuclear Testing. Two basic decisions from the President are required to complete the protocols:

— Acceptance of the Soviet package proposal. The Soviets have proposed a package that accepts our right to use our preferred method of verification (CORRTEX) in exchange for the right to use theirs (seismic plus additional on-site inspection—OSI), and splits the difference on the issue of what yield will trigger CORRTEX/seismic verification (75 kt).

Although DOD argues that acceptance of seismic verification puts us on a slippery slope to a comprehensive test ban, we can agree to this package without jeopardizing our case that a CTB is unacceptable as long as we rely on nuclear weapons for deterrence. The Soviets’ “additional” OSI is no more intrusive than that used for CORRTEX, and could enhance verification and strengthen our hand during ratification. A 75 kt trigger was our position until we changed to 50 kt last year.

— Instructions to finish the job. Once we get a decision from the President accepting the Soviet package, instructions to complete the protocols would be essential to minimize agency foot-dragging on the remaining issues.

START. The one major START subject that we think can be resolved is air-launched cruise missiles (ALCMs), but it will take a decision from the President to break the interagency logjam:

— ALCM Range/Distinguishability/Counting trade. We need to agree with the Soviets on what ALCMs are limited by the treaty, as defined by: (1) range (the Soviets want to include all ALCMS above 600 kms, we want to include all ALCMs above 1500 kms) and (2) warhead type (we both agree that only nuclear-armed should be counted, but have not agreed on how to tell what is conventionally-armed).

We also need to agree on how to count ALCMs against the 6000 warhead ceiling (the Soviets want each ALCM to count as one, we want them to count less). Because we want a high range and low count and the Soviets want the opposite, there are package deals that might split the difference.

We need to get USG agreement in principle to the idea of a package approach with some US give. Then we need to define our bottom lines and select a strategy for making the proposal: (1) modification of our current “attribution” approach which undercounts the number of ALCMs on each bomber, or (2) counting the number of ALCMs on bombers as the bombers are actually equipped, with an explicit discount that counts each ALCM as only 5 or 75).

In addition, the package should exclude conventional ALCMs, and allow for a limited number of conventional bombers outside the START limits.

— V/S. An agreement to notify strategic exercises should be ready for signature by the Ministerial. We will send a draft agreement to ban short time-of-flight SLBM tests to the Soviets before the Ministerial; it could also be ready for signature. (We may have to fend off the argument that such agreements violate our Krasnoyarsk pledge.)

We plan to flesh out our proposals for early missile production monitoring and data exchange, and other proposals may be brought to the point of agreement in principle.

Chemical Weapons. We may need a decision from the President to provide more complete data on our CW stocks in order to get an MOU on bilateral data exchange and early verification ready for signature at the Ministerial.

Other Agency Views. The OSD staff opposes initiatives on NTT and ALCMs. They argue that acceptance of seismic verification and OSI represents a step toward a comprehensive test ban, and that we cannot settle ALCMs until we know the outcome on the B–2 program. The NSC staff is somewhat more encouraging. We won’t know how stiff the opposition will be until you raise these issues with Cheney and Scowcroft.

Game Plan

If you agree with where we are heading, you may want to meet soon with Cheney and Scowcroft, and later the President, to outline what you think can be accomplished at the Ministerial and the decisions needed to get there. Before that meeting, we should schedule two or three briefings with you to discuss the overall game plan and the details of NTT and ALCMs.

Following the Presidential meeting, we would have the Arms Control PCC develop brief decision memos on NTT and ALCMs for DC and then Presidential decision. Pressure from the top will be needed to ensure that the papers are scrubbed in time for NSC deliberation, decisions, and Congressional and Allied consultations (if desired) before the Ministerial.

Attached is a proposed decision schedule (Tab A)³, a brief analysis of NTT issues (Tab B),⁴ an introduction to ALCMs (Tab C), an analysis of ALCM packages for the Ministerial (Tab D), and an update on the CW MOU discussions concluded with the Soviets this past Monday (Tab E).⁵

Tab C

Paper Prepared in the Department of State⁶

Washington, undated

Introduction to ALCMs

Several issues related to air-launched cruise missiles (ALCMs) are among the most important remaining in START, and both sides believe a solution could be achieved at the upcoming Ministerial. A solution will require the detailed involvement from you and Shevardnadze, interacting intermittently with the working groups. Approaches to resolve the ALCM issues are already being considered in the interagency, and you will undoubtedly become involved in finalizing our position in the preparations for Wyoming. Below is background on the ALCM questions you will deal with in coming weeks.

Current Areas of Agreement and Disagreement in Geneva

Key areas of agreement:

— Within each side’s limit of 6000 “accountable” strategic warheads, no more than 4900 can be on ballistic missiles (ICBMs and SLBMs), leaving at least 1100 on strategic bombers.

— Nuclear gravity bombs and short-range attack missiles (SRAMs) carried by a bomber not equipped for ALCMs count as a single unit in the 6000 limit (even though a B–1, for example, can carry 16 or more). With this highly favorable discount, our currently planned penetrating bomber force of 97 B–1s and 132 B–2s would use up only 229 of the 1100 units available for bomber weapons, leaving 871 for our ALCM force.

— Current U.S. and Soviet ALCMs are all considered to be nuclear armed. Future conventional ALCMs are to be excluded from limits applicable to nuclear ALCMs if they are distinguishable from nuclear versions.

Key issues yet to be resolved:

— Counting rule: the number of ALCMs to assign to each type (or sub-type) of nuclear-ALCM-carrying bomber against the 6000 limit. The higher the number assigned each type, the fewer the number of ALCM-carrying bombers we can deploy before we fill up our planned ALCM allotment of 871 units.

— Range: the range level at which a nuclear air-to-surface missile becomes a nuclear ALCM subject to START limits.

— Other related issues: a package resolution should protect conventional ALCMs, and allow for a limited number of conventional bombers outside the START limits. A handful of less central questions must also be resolved eventually, but do not have to be addressed in the package.

The Counting Issue

U.S. position. Seeking to maximize the number of ALCMs and ALCM-carrying bombers we could deploy under START limits, while minimizing the task of verifying ALCM limits, we have favored an “attribution” approach under which all ALCM carriers would be credited with an agreed number of ALCMs regardless of the number actually carried. Negotiated “attribution” rules are for bookkeeping and do not pretend to reflect (or even limit) deployments; thus, they avoid any requirement to verify numbers carried. The only verification task is to determine whether a bomber is an ALCM-carrier. Such carriers are credited with an agreed number regardless of how many it is equipped to carry or actually carries. We would make clear to the Senate that such rules were counting tools, that both sides could legally deploy more than the attributed numbers, and that we considered this to be advantageous to us because of our greater reliance on ALCMs and the larger ALCM capacities of our bombers.

We originally proposed attributing all U.S. and Soviet ALCM carriers with 6 ALCMs, but later moved to 10. Thus, under our approach, our 871 available ALCM units could be used by deploying 87 ALCM-carrying B–52 Hs, each counted as having 10 ALCMs but each actually carrying 20 (total of 1740 actual missiles, 870 of them “unaccounted”). We have justified our attribution approach, including its discounting of ALCMs, on the grounds that ALCMs are slow, air-breathing systems that are not well suited for a surprise first strike, and must fly through heavy Soviet air defenses; and that attribution significantly eases the verification task.

During 1988, we told the Soviets we could attribute different numbers of warheads to different types of bombers, in response to a Soviet argument that assigning all bombers the same number would unfairly treat Soviet bombers, which carried fewer ALCMs. (In general, we’ve placed a higher priority on discounting U.S. ALCMs than minimizing discounting of Soviet ALCMs). We had provisional authority (never used) to attribute the Soviet Bear H and Blackjack bombers with 6 and 8 ALCMs, respectively, while continuing to assign 10 to U.S. bombers.

Soviet position. Seeking tighter limits in an area of U.S. advantage (and perhaps regretting their assent at Reykjavik to a non-ALCM counting rule generous to U.S. bomber forces), the Soviets have opposed our attribution approach, especially the high discount we have attached. Instead, they have proposed assigning ALCM-carriers the number of ALCMs for which they are physically equipped to carry and not allowing them to carry any more.

Originally, the Soviet “as equipped” rule would have counted bombers at the maximum number of ALCMs they were designed to carry. Last round the Soviets said they would count only the number for which a bomber was currently configured, even if it was less than maximum capacity. Under this “actually equipped” approach a B–52 G now equipped to carry 12 ALCMs on its wings could be physically modified to carry fewer and would then be counted with the lower number (with inspections to ensure that the bomber was not equipped to carry more than that number).

We have opposed the Soviet “as equipped” approach, both because it provides no discount for U.S. ALCMS and because it would involve the difficult verification task of determining the number of ALCMs for which various types of Soviet bombers were actually equipped—a task that would be difficult even with intrusive inspections.

Finding a solution. A primary consideration in finding a solution will be to maximize the number of ALCMs and ALCM bombers we would be able to deploy. We will thus presumably favor some form of discounting. Indeed, discounting ALCMs could become even more important if the B–2 program were reduced. Given uncertainties over the B–2, OSD may not wish to deal with ALCMs at all in Wyoming, or may wish to propose an even greater discount than in our current position (e.g., move back to attributing all bombers with 6 rather than 10). Somewhat ironically, the Air Force has been interested in an “as equipped” approach similar to that proposed by the Soviets even though it would mean fewer ALCMs. (The motive is unclear, but many speculate that having fewer ALCMs would put pressure on the Congress to fund the penetrating B–2).

The Range Issue

U.S. position. The U.S. has called for limiting nuclear-armed ALCMs over 1500 km in range. This position protects conventionally-armed ALCMs from limitations in START, and allows the U.S. to deploy nuclear-armed ALCMs with ranges below 1500 km without counting them in the 6000 warhead limit.

At the Moscow Summit, the Soviets agreed that future conventionally-armed ALCMs would be exempt from START limits if they were distinguishable from nuclear ALCMs. Although we still disagree over the specific criteria for distinguishability, the agreement at Moscow has basically eliminated part of the justification for sticking to a high range threshold.

Some in OSD and the JCS argue that a high range threshold is still necessary to protect the option for a theater nuclear ALCM, particularly if the B–2 program is significantly reduced. They argue that such a weapon is necessary because of INF eliminations and a CFE agreement that would increase range requirements by pushing Warsaw Pact forces back from the inner-German border. But there is no program, or formal military requirement, for such a missile. Given the political difficulties we face in deploying a shorter-range nuclear ALCM in Europe as well as a follow-on to the Lance missile, a new theater nuclear weapon does not appear realistic.

At the Moscow Summit, and at later Ministerials, we told the Soviets that we would consider an ALCM range threshold somewhat below 1500 km in the context of agreement to other ALCM issues. A specific threshold, however, was not discussed.

Soviet Position. The Soviet position is that all nuclear-armed ALCMs above 600 km in range—the threshold used in the SALT II agreement—should count against START limits. The Soviets have also asserted that their agreement to the counting rule for heavy bombers carrying only gravity bombs and SRAMs was predicated on agreement to the 600 km threshold, although the ALCM threshold was not specifically discussed. The Soviets have continued to stress 600 km as the key ALCM issue for them, and they have threatened to reopen the other counting rule if a 600 km threshold is not agreed.

Finding a solution. The U.S. interest in discounting ALCMs appears greater than our interest in a high range threshold. This is particularly true given Soviet agreement not to count conventional ALCMs in START. As part of any package for resolving ALCM issues, however, we must insist on Soviet agreement to our position that future nuclear and conventional ALCMs must be distinguishable only by NTM. (The Soviets currently propose that such ALCMs must be distinguishable by functionally-related and externally observable differences.) All agencies agree that we do not have any flexibility on this issue.

Although our position on distinguishability would not give us confidence that ALCMs that are declared to be conventionally-armed are not in fact equipped with nuclear warheads, we have to date been willing to accept the associated risks given that extra ALCMs are likely to be an area of U.S. advantage.

Package Solutions to Remaining ALCM Issues

Because they are highly interrelated, the issues of ALCM counting, range threshold and distinguishability should be settled as part of a package deal. In addition, we may want to incorporate some of the other less central ALCM and heavy bomber issues in such a package. The accompanying ALCM paper outlines a range of possible package proposals.

Tab D

Paper Prepared in the Department of State⁷

Washington, August 31, 1989

ALCM Approaches for the Ministerial

The U.S. and Soviet Union continue to have differences over three particularly significant ALCM issues—an ALCM counting rule, the range threshold defining accountable ALCMs in START, and how to distinguish between future types of nuclear and conventional ALCMs. Because these issues are closely interrelated, it is likely that any final settlement will require a package solution. (Remaining differences between sides on key ALCM and related heavy bomber issues are summarized in Table 1.)

U.S. Interests

The U.S. has various interests to consider in developing an acceptable ALCM approach:⁸

• Maximize U.S. ALCM numbers. Because air defenses are not limited, ALCMs do not have the same assurance of reaching their targets as ballistic missile RVs. In addition, discounting helps us ensure that we will have a sufficient number of warheads under START to meet the targeting requirements that follow from our current national strategy. If the planned buy for the B-2 bomber is reduced (a penetrating bomber’s weapons load counts as only 1 against the 6000 warhead limit) discounted ALCMs would take on even greater importance. At the same time, however, preserving our flexibility with regard to ALCM deployments will mean fewer constraints on Soviet ALCM programs.

1. Source: Department of State, Economic and Agricultural Affairs, Robert B. Zoellick, Under Secretary for Economic and Agricultural Affairs, Lot 96D484, USSR: September 1989 Ministerial. Secret.
2. See Document 41.
3. Attached but not printed is an undated paper entitled “Proposed Decision Schedule.”
4. Attached but not printed is an undated paper entitled “Action Needed to Complete Protocol on the Threshold Test Ban Treaty (TTBT).”
5. August 28. Attached but not printed is an undated paper entitled “CW MOU.”
6. Secret.
7. Secret.
8. The U.S. also has a clear interest in protecting conventional ALCM programs. At the Moscow Summit, the sides agreed that future types of conventional ALCMs would not count in START if they are distinguishable from nuclear ALCMs. The Soviet position in Geneva, however, would require observable and functionally related differences, while the U.S. position requires only that ALCMs be distinguishable by NTM. All of the packages considered in this paper take this concern into account and would require Soviet agreement to the U.S.-proposed approach to distinguishability. More specifically, the U.S. needs to protect Tacit Rainbow, a conventional anti-radiation ALCM currently in full-scale development. We believe that this missile, under the agreed START definition of ALCM range, would fall below 600 km. In addition, regardless of its range, this missile would be protected if we can define it as a future type under the Moscow agreement, since it is distinguishable from nuclear ALCMs. If neither of the above conditions holds, we could take other steps, e.g., explicitly grandfather this system. [Footnote is in the original.]