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

200. Memorandum From Arnold Kanter of the National Security Council Staff to the President’s Assistant for National Security Affairs (Scowcroft)¹

Washington, February 5, 1991

SUBJECT

• Your February 5 Meeting on START

Your meeting² is largely intended to bless the next round of the Bartholomew-Obukov talks (scheduled for later this week in Geneva)³ and ensure that everybody on the U.S. side is singing from the same sheet of music. Subject to any issues raised by the principals, the only potential issue for decision is the U.S. approach to downloading. The other major topics are a discussion of the U.S. “approach” to the Bartholomew-Obukov talks, and—time permitting—a report from the Ungroup to the principals on other START issues that may come up in Geneva.

An agenda is attached at Tab I⁴ and a discussion paper on downloading is attached at Tab II. Copies were distributed yesterday to the Ungroup to pass along to their principals in preparation for today’s meeting.

Approach to the Bartholomew-Obukov Talks

This is an opportunity for Secretary Baker to remind everyone about the President’s determination to make every effort to complete START by the end of the month, and for Secretary Cheney to wonder aloud about why we are even bothering to get on a plane to Geneva. It is unlikely that this part of the discussion will lead to a clear, much less agreed, course of action.

There is much to be said for the argument that if we do not get a START agreement very soon (e.g., by the end of February), then the prospects for getting one at all grow much dimmer. This perspective argues for a U.S. willingness to be flexible on second-order issues lest the best become the enemy of the good.

The problem is that the ferocious pursuit of positions on somewhat second-order issues (e.g., B–2 inspections, PPCM of Midgetman) has led naturally to an offer by the Soviets to concede on them in exchange for issues that may really matter (e.g., downloading, new types). Bessmertnykh offered such a “deal” in his meeting last week with Baker. Baker will say, and probably believes, that he rejected it. Bessmertnykh probably will say, and may believe, that the U.S. tacitly accepted the proposed deal when it acknowledged that the bomber issues were once again resolved, and that the “Houston package” was back on track.

If we cannot and should not accept the Bessmertnykh “deal,” we also will have a tough time persuading the Soviets that they should accept all of our proposals on B–1, B–2, PPCM of Midgetman, and all of our positions on the major remaining issues (e.g., downloading, new types, data denial). The trick will be to get Baker-the-Deal-Maker to understand that we cannot trade downloading or new types to sustain our position on bomber inspections, and to get Cheney-the-Unwilling to express a willingness to make further, largely symbolic moves on the remaining issues in order to complete the treaty. Ideally, there would be agreement at your level about all of the issues on which the Soviets will need to come our way, and those on which we can show further flexibility. In practice, you probably should settle agreement on the broad principles, even if the result comes out sounding like contending cliches.

You might start the meeting by asking Baker for a brief recap of the outcome of his discussions with Bessmertnykh on START, specifically on the question of what kind of “deal” Bessmertnykh offered, and the implications of our rejection of its terms. You then might kick off the discussion by making the following points:

— The President said we should make every effort to complete START by the end of the month.

— That is not a deadline that requires us to make whatever concessions are necessary to reach an agreement.

— It is a recognition that if we do not wrap up START very soon, there is a good chance that we will not be able to complete the treaty for a very long time, if ever.

— Our guidance to Reg and his team is simple: hang tough on the core issues where we have no more flexibility, but negotiate hard and be willing to make tradeoffs within that framework in order to complete START by the end of February if possible.

Downloading

Reg tabled the U.S. downloading proposal in his recent meeting with Obukov. The Soviet side did not directly, much less extensively, respond. However, there is no good reason to believe that an extended exchange with the Soviets about the U.S. downloading proposal would produce Soviet acceptance. The “deal” that Bessmertnykh proposed to Baker, which would have the U.S. accept the Soviet downloading proposal, constituted an implicit rejection. In brief, we may be at an impasse with the Soviets over an issue that could be, and perhaps should be, a treaty-breaker.

Your meeting should decide whether and, if so, how the current U.S. position should be modified. A discussion paper on this subject is attached at Tab II.

Option A is the current U.S. position. The talking points that presented this proposal to the Soviets appear as the last page of the discussion paper.

Option B is a version of the current U.S. position that should be more attractive to the Soviets in two respects. First, it would allow the Soviets to download up to 1000 RVs on any combination of ballistic missiles, rather than limiting downloading to the SS–N–18. Second, it would give the sides a continuing right (rather than a one-time opportunity) to download additional existing types of missiles as previously downloaded types are retired, subject to the 1000-RV cap. The U.S. would be able to download either all MM-IIIs, or 300 MM-IIIs and approximately half the Trident force under Option B.

Option C adds to the current U.S. position by giving each side the right to download up to 500 RVs in addition to grandfathering the downloaded MM-III and SS–N–18. It would permit both sides to download a total of about 1500 RVs. The U.S. could download all 500 MM-IIIs plus half the Trident force.

Option D (2 RVs per missile on 3 existing types) is the alternative that had been previously generated by OSD and JCS in an effort to find a middle ground that limited downloading while fully protecting the ability of the Navy to download SLBMs as a hedge against possible START II constraints. Moscow would argue (with some justice) that it gives the U.S. far more flexibility to download than the Soviets. Your recent meeting rejected this option because it allowed both sides far too much freedom to download.

Other START Issues

I can review other issues that the Ungroup has been discussing and deciding in preparation for the Geneva trip. Your meeting should be aware of, and perhaps should ratify, the approach we intend to take to two issues in particular.

New Types. If the Soviets could easily get a follow-on variant of an existing missile treated as a “new type” under START, they could accomplish many of the same objectives as they apparently seek with their ambitious downloading proposal. They have been pressing hard for just such a new types definition, most recently by arguing that changes in the throw-weight of missile should qualify it as a new type. As a complement to our downloading position, we have been insisting that a missile be substantially different before it could be considered to be a new type.

We have offered the Soviets a cosmetic fix that takes account of missile throw-weight, but leaves the substance of our position unchanged. Our concern is that the Soviets will try to convert a cosmetic adjustment into a real change. Assuming that we continue to insist on something like our current position on downloading, we will resist any such Soviet efforts.

A related, and seemingly technical, issue concerns how to measure and verify the dimensions of stages and missiles. Since these dimensional measurements are precisely the criteria that distinguish existing from new types of ballistic missiles, we will need to insist that they are defined and measured in appropriate ways.

Ban on Data Denial. In an effort to expedite completion of START, the Soviets have been pressing to defer many of the hard decisions on data denial until after the treaty has been signed. Our Geneva delegation believes there is some merit in this Soviet approach if we are to complete a treaty in February.

The issues related to the ban on data denial are difficult and controversial. Indeed, they have divided the USG for months, if not years. Nevertheless, the ban on data denial has considerable political visibility and is at the heart of our approach to START verification. We need to get as much of the regime agreed as possible before START is signed. Against the backdrop of the stalled CW agreement and the CFE Treaty, we need to be especially careful about signing a START treaty that leaves hard issues unresolved.

Accordingly, we are telling the Geneva delegation that Washington will not support “telemetry triage,” but will work as hard as possible to reach agreement on all of the facets of the data denial ban. If and when we get to a point at which everything else in START has been agreed, we can review the bidding and decide which loose ends, if any, on data denial we are prepared to try to tie up after treaty signature.

Rich Davis and Art Kuehne concur.

Tab II

Paper Prepared in the National Security Council⁵

Washington, February 4, 1991

DOWNLOADING ISSUES

I. Issue:

The US recently proposed to grandfather the downloading of both the SS–N–18 and the Minuteman III missiles, but to ban further downloading (see attached talking points). The Soviets rejected this US proposal. In preparation for the upcoming Bartholomew-Obukov meetings in Geneva, the US should review its downloading position.

II. Background on Downloading:

The US initially proposed downloading primarily to permit the deployment of MM IIIs with one RV, thereby allowing the US to retain a larger silo-based ICBM force in START I. The right to download SLBMs also is of interest in the context of a hedge against a possible 3000–4000 accountable warhead START II limit.

The USAF currently plans to download Minuteman III missiles by 1 or 2 warheads each on 300 missiles at two bases. A right to download an aggregate of 600 RVs would protect this Minuteman option.

The Navy would require an aggregate of at least 864 RVs to protect future SLBM downloading requirements, assuming that downloading provisions would allow the downloading of the current two types of missiles on the Trident submarines or that the D–5 would be backfitted to all submarines.

Taken together, these MM III and SLBM options would require the right to download 2 RVs from each of 2 or 3 types of missiles, within an aggregate limit of about 1500 downloaded RVs.

The Soviets, after an initial resistance to downloading, announced that the SS–N–18 should now count at 3 RVs instead of 7 RVs per missile (a difference of 896 RVs), and propose to allow downloading of an aggregate of 1250 warheads on 2 additional types of missiles.

In order to use this large aggregate limit effectively, the Soviets would probably seek to download their SS–N–20 and SS–24 missiles by at least 5 RVs each. A right to download an aggregate of 1060 RVs (in addition to the SS–N–18) would probably be required to satisfy Soviet objectives. If the SS–N–18 is treated as a downloaded missile, the Soviet approach amounts to the right to download up to 5 RVs from each of three types of missiles, within an aggregate limit of about 2000 RVs.

Possible Downloading Options

There are three variables that could be used to define downloading options: (1) the number of types of missiles that could be downloaded; (2) the number of RVs that could be downloaded from each missile; and, (3) the aggregate number of RVs that could be downloaded.

Each of the following options assumes that the SS–N–18 would be captured as a downloaded missile, that base-by-base downloading would be allowed, that downloading of heavy ICBMs would be banned, and that there would be one extra RV inspection per year per base where downloaded missiles are deployed.

A. Grandfather SS–N–18 and MM III RV attribution (current US position).

– Download 4 RVs per missile for SS–N–18.

– Download 2 RV per missile for MM III.

– No aggregate limit.

Aggregate downloading:

US—600–1000 RVs Soviet—896 RVs

B. Download an aggregate of 1000 RVs for each side on existing types of missiles.

– Either ICBMs or SLBMs.

– No limit on the number of different types of missiles.

– As types of downloaded missiles are retired, additional existing types may be downloaded.

Aggregate downloading:

US—1000 RVs Soviet—1000 RVs

C. Grandfather the SS–N–18 and MM III attribution separate from downloading. Download up to two RVs per missile on one additional type of existing missile with an aggregate limit of 500 RVs.

– Either ICBMs or SLBMs.

Aggregate downloading:

US—1080–1480 RVs Soviet—1136 RVs

D. Download up to two RVs per missile on three existing types of missiles.

– Except the SS–N–18 could be downloaded by four RVs per missile.

– Either ICBMs or SLBMs

– No aggregate limit.

Aggregate downloading:

US—1464–1864 RVs Soviet—1320 RVs

1. Source: George H.W. Bush Library, Bush Presidential Records, National Security Council, John A. Gordon Files, Subject Files, OA/ID CF01034–015, START—February 1991 [1]. Secret. Sent for information.
2. According to the President’s Daily Diary, Bush held a principals meeting on February 5 from 1:31 to 2:39 p.m. in the Oval Office. No minutes were found.
3. In telegram 1352 from NST Geneva, February 5, the Delegation confirmed Bartholomew’s program in Geneva, February 7–10. (Department of State, Central Foreign Policy File, Electronic Telegrams, D910117–0485)
4. Attached but not printed is the agenda, which reads: “A. Approach to the Bartholomew—Obukov Talks”; “B. Downloading in START—Paper Attached”; and “C. Unground Report on Other START Issues.”
5. Secret.