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

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

Washington, January 23, 1991

SUBJECT

• Downloading in START

I’ve attached a short issue paper on downloading that I distributed this morning as background for your 4:00 PM meeting this afternoon.² The paper presented three options for downloading:

1. Current U.S. proposal—Download up to two RVs per missile on two existing types of missile (one ICBM and one SLBM). Downloading of mobile and heavy ICBMs would be banned. The SS–N–18 would count as a downloaded missile.

2. Current Soviet proposal—Download up to five RVs per missile on two existing types of missile. There would be an aggregate limit of 1250 downloaded RVs. Downloading of heavy ICBMs would be banned. The SS–N–18 would not count as a downloaded missile.

3. DoD proposal—Download up to two RVs per missile on three existing types of missile, except for the SS–N–18 which could be downloaded by four RVs per missile. Downloading of heavy ICBMs would be banned.

In addition to the option of banning downloading all together, there are three other ideas that you might want to consider:

A. DoD proposal with a 1000 RV aggregate limit on downloading. OSD could probably sign up to this option. However, JCS would have problems because the 1000 aggregate limit would not allow downloading in START I of both the MM III and the C–4/D–5 to the ultimate levels desired by the Services. The Soviets would balk at inclusion of the SS–N–18 in conjunction with the “low” aggregate limit.

B. Grandfather SS–N–18 RV attribution at 3 RVs per missile (resulting in up to a 896 RV reduction in attribution) in exchange for grandfathering of MM III RV attribution at 1 RV per missile (resulting in up to a 1000 RV reduction in attribution). Ban downloading for all other types. OSD would probably accept the SS–N–18/MM III trade. JCS has already rejected this proposal because it would not allow for downloading SLBMs on Trident submarines in START I. The Soviets would probably also reject this approach now that they appear to have fully embraced the concept of downloading.

C. Grandfather the SS–N–18 and MM III attribution, and allow downloading of one type of ICBM or SLBM, with an aggregate limit of (blank) downloaded RVs. OSD would object because the U.S. requires only the MM III “grandfather” provision in START I, and would not be able to use the downloading provisions until START II. JCS would object to permitting downloading on only one additional type, but might be able to accept this option if the aggregate limit was 800–1000 to accommodate the downloading of the SLBMs on Trident submarines. The Soviets might find this option an acceptable compromise since it would resolve the SS–N–18 issue outside of downloading and would also permit some further downloading. However, the total number of RVs downloaded (counting both the change in RVs for the SS–N–18 and the MM III, and subsequent downloading of another type of missile) could be significantly higher than 1000 RVs.

Tab I

Paper Prepared in the National Security Council³

Washington, January 23, 1991

Downloading

Background

The U.S. and the Soviets have agreed in principle to permit downloading in START. In contrast to their earlier position (which would have accepted a ban on downloading), it now appears that the Soviets place great importance on the ability to download large numbers of RVs. Disagreements over downloading could become a major obstacle to completing START.

The Soviets have proposed an approach that would allow both sides to download a maximum of 1250 RVs from two existing types of ICBMs (except heavy ICBMs) and SLBMs. Additionally, the Soviets recently declared that their SS–N–18 SLBM carries only three RVs, rather than the seven attributed to it in the 1987 Washington Summit Joint Statement. They insist that this change, which amounts to an apparent 900 RV reduction, does not constitute “downloading” and should not be counted against the 1250 RV limit in their downloading proposal.

(The Soviets also have proposed an approach to the definition of “new types” of missiles in START that could be a partial substitute for, as well as complement to, their downloading proposal. Under the Soviet proposal, it would be relatively easy for variants of existing missiles to be categorized as a “new type.” This means that a variant of the SS–24, for example, would count as a “new type” which the Soviets could claim carried only five RVs rather than the ten RVs on the current SS–24.)

In some respects, downloading could enhance strategic stability. It would reduce the concentration of RVs on missiles, a longstanding U.S. goal. The Soviets probably would exploit downloading to increase the number of SSBNs they retain in their START-constrained force, a step that we have encouraged them to take to achieve a more balanced strategic force.

The U.S. also has an interest in downloading some of its ballistic missiles. If downloading were banned, the U.S. would have room only for about 300 of its current 500 deployed Minuteman IIIs in a START-constrained force. The right to download 500–1000 RVs on one type of missile would allow us to retain all of these missiles. Additionally, the U.S. may want to lock in now the right to download its SLBMs to meet possible “START II” limits. An approach that allowed each side to download 1500–2000 RVs on 2–3 types of missiles would protect this option.

Downloading also could have significant drawbacks. It could facilitate a Soviet ability to break out of START limits on relatively short notice. Some would argue that the Soviets would position themselves to be able to exploit that potential, while the U.S. would not. The Soviets proposed a number of provisions that might have marginally limited the breakout potential for missiles downloaded; however, since the U.S. cannot accept START obligations that would require either the destruction of RVs or the redesign of missile front ends as adjuncts to downloading, these avenues to constrain breakout options are foreclosed.

Downloading also could have political liabilities that would become an added ratification burden. Critics will argue that permitting downloading in START amounts to tearing another hole in already porous START limits. In this connection, they may point to the tacit “downloading” embedded in already agreed START limits, notably the agreement to count the Soviet SS–N–23 at 4 RVs (vice 10), the SS–18 at 10 RVs (vice 12 or 14), the D–5 at 8 RVs, (vice 10 or 12), and the Peacekeeper at 10 RVs (vice 12).

Options

1. Current U.S. Proposal. Each side would be allowed to download up to two RVs from two existing types (one ICBM and one SLBM). Downloading of heavy ICBMs and mobile ICBMs would be banned. Although there is no explicit limit on the aggregate number of RVs that could be downloaded, under this approach the U.S. could download a total of about 1500–1900 RVs (depending on the D–5 backfit on Trident submarines), and the Soviets could download about 500–700 RVs. (This proposal was tabled before the recent Soviet declaration that the SS–N–18 should count as carrying three instead of seven RVs.)

2. Current Soviet Proposal. (please review closely) Each side would be allowed to download up to five RVs from two existing types of ballistic missiles. Downloading of heavy ICBMs would be banned. There would be an aggregate downloading limit of 1250 RVs. The SS–N–18 would not be counted as a downloaded missile.

3. DoD Proposal. Each side would be allowed to download up to two RVs from three existing types of ballistic missiles, except for the SS–N–18 from which four RVs could be downloaded (i.e., treat SS–N–18 as a downloaded missile). Downloading of heavy ICBMs would be banned; downloading of mobiles would be permitted. Although there would be no explicit aggregate limit, the Soviets could download about 900–1300 RVs, and the U.S. theoretically could download about 1900 RVs.

1. Source: George H.W. Bush Library, Bush Presidential Records, National Security Council, John A. Gordon Files, Subject Files, OA/ID CF01034–006, START—January 1991 [3]. Secret. Sent for information. A stamped notation indicates Scowcroft saw the memorandum.
2. No minutes were found.
3. Secret.