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

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

Washington, June 8, 1989

SUBJECT

• Deputies Committee Meeting, Friday, June 9, 1989, 10:00 a.m.—START Positions on Mobile ICBMS, Mobile ICBM Verification and New Initiatives

This Deputies Meeting will be the last of three meetings to review the specifics of our START position before the negotiations resume June 19. It will also be the bureaucracy’s last opportunity to engage on possible new initiatives.

Mobile ICBMs

The papers on mobile ICBMs and mobile verification reiterate agency positions on old issues and a discussion of these papers would repeat the pattern at Wednesday’s DC meeting.² For that reason, and because of the large number of items to cover in one hour, we recommend that you not march through these papers in the meeting in detail, but seek consensus at the outset that they fairly represent agency views and options.³ At that point, attempt to get the Deputies to focus on the tradeoffs underlying all the issues in these two papers: how should the U.S. evaluate competing U.S. objectives and concerns in the areas of stability, verification, intrusiveness, budgetary expense, and U.S. ICBM program uncertainty.⁴

The PCC Papers in brief:

— Mobile ICBMs. This paper reviews agency positions (although specific agencies are not identified) on four issues: whether to alter our formal position banning such missiles; whether to consider banning MIRVed mobile ICBMs while allowing single-RV mobile ICBMs; if mobiles are to be allowed, what numerical constraints to seek on them; and whether, and if so, how to address accountability and inspection provisions for “movable” ICBMs (e.g., fixed area, multiple shelter ICBM basing).

— Mobile ICBM Verification. Similar to the Mobile ICBM paper above, this paper reviews options on specific issues that an agency requested be revisited before resumption of negotiations in Geneva. The approaches assume that mobiles will be permitted. If (as seems unlikely) the DC believes we will not soon drop our ban on mobiles, a discussion of how to verify mobiles probably could be deferred. There are six issues in this paper: whether to modify the restrictions on road-mobile ICBMs; whether to modify our position on rail-mobiles to restrict missiles to a specified Deployment Area; whether our approach to continuous monitoring of production is sufficiently broad based; whether to withdraw our position on Mandatory SSI pending further study; whether to withdraw our position on Right-of-Refusal SSI pending further study; and whether to allow SSI at facilities subject to continuous monitoring if such monitoring is not actually in place.

Points on underlying mobile issues:

— Given the President’s decision to fund two mobile ICBM programs, should we drop our ban on mobiles in START? If so, when and how?

– The consensus we are seeking to build could unravel if rail-mobile Peacekeeper were seen only as a bargaining chip.

– On the other hand, a Treaty which allows mobile ICBMs might prove unratifiable without a well-funded U.S. mobile ICBM program, if not actual deployment.⁵

— Verification is a significant mobile ICBM issue. There are two kinds of verification:

– political – Are the Soviets honoring their treaty commitments (should we proceed on the assumption they will cheat)?

– military – How much cheating might go undetected? What is the military significance? What options does the U.S. have to counter once cheating or breakout is detected?⁶

— We are agreed that verification needs to be handled more directly on the Hill, and in public. Several areas of this treaty cannot pass the “political” verification standard. We need to develop more data to answer the questions about how “military” verification will be achieved, and at what cost.

— Is there so much uncertainty in our ICBM programs that we need to protect options in START for entirely new kinds of systems, such as so-called movable ICBMs?

— How do we trade-off U.S. mobile ICBM operational flexibility (and future survivability options) against our confidence in monitoring the Soviets? Some have proposed that we establish a deployment area for mobiles. How much does this help monitoring? How small does the area have to be? How much area do we need for survivability? How do we hedge against future Soviet targeting improvements?

— With regard to continuous monitoring and suspect site inspections, some would expand our, and the Soviets’, rights to inspect. What do we gain? Without “anywhere, anytime” rights (which we all agree is not acceptable), what confidence do we have that the Soviets are not conducting illegal activities elsewhere? What do we lose by a Soviet presence in and around a greater number of U.S. facilities? Is the budgetary cost of significance?

New Initiatives

There is potentially a lot of ground to cover here, if the Deputies are prepared. The bureaucracy has not been. The process did not generate much in the way of new thinking, but did produce abundant criticism of any ideas that were floated. This meeting will be the first serious interagency discussion of possible new initiatives. Thus, your objective here is to stimulate thought and foster a willingness to develop ideas into proposals that work. Do not to expect to reach closure on any specifics. In the end, this part of the meeting may only serve to perform a “due process” function.

You might begin the discussion by noting the importance to the President of making an initiative, to demonstrate that START is not merely a continuation of the last administration’s negotiation,⁷ and to give substance to the objective of reducing the risk of war and enhancing stability (with which all agencies have agreed). You should remind the Deputies that the papers are not agreed documents, but are intended only to stimulate discussion. (Some of the proposals attached to the new initiatives paper were not seen by agencies until Thursday.)⁸ Finally, you may wish to note that a potential initiative, banning MIRVed Mobiles, which is in the Mobile ICBM paper should be considered in this context.⁹

Briefly, here are the two papers:

— Short Time-of-Flight (Depressed Trajectories). This discussion paper, drafted by the Joint Staff and not reviewed by the PCC, concludes that there are significant verification and circumvention problems with any arms control approach, but that the disadvantages to the U.S. can be minimized by limiting the ban to future types of missiles tested from submarine platforms. It suggests pursuing the ban as a confidence-building measure outside of START to avoid subjecting it to a more rigorous verification standard in START.
Even if such a ban would not make much of a direct contribution to security, it may still be worth pursuing, not least because of a widespread perception that it is a good idea. In that event, the task would be to discuss a ban that, at worst, did not hurt us.

— New Initiatives. This paper was discussed in the PCC, but agencies did not actively participate, and most of the alternatives are not attractive. Three new ideas we have received recently are attached to the paper and have not been reviewed by any agency. The approaches in this paper focus on ways to improve stability: MIRV warhead sublimit; warhead-to-missile ratio limit; silo-based MIRV ban; silo-based MIRV-limit; total silo-based ICBM ban; future production ban on MIRV ICBMs; and immediate portal monitoring of mobile final assembly facilities.¹⁰

You might find it useful to encourage the discussion to proceed along the lines of the categories of problems potential initiatives might address. Three categories are:¹¹

— Maintain or Increase Current Warning Time. The ban on short time-of-flight missiles has been suggested. Other initiatives such as the Weinberger-Yazov idea (never finally agreed) to notify strategic exercises, might be considered.¹²

— Minimize or Eliminate Military Consequences of ICBM Silo Vulnerability. Several types of limits or bans on silo-based warheads or MIRVed warheads have been suggested. These approaches all tend to encourage a shift to mobile ICBMs or SLBMs and thus are in tension with some of the measures designed to address mobile ICBM verification problems (see below). Also, a production ban on future MIRV ICBMs would attempt to have these forces, fixed and mobile, “whither away” over time.¹³

— Minimize Consequences of Mobile ICBM Cheating. This category includes a ban on MIRV mobiles, and/or low mobile warhead limits.¹⁴ Both types of measures make more difficult any shift away from vulnerable silos and may encourage greater reliance on silo-based ICBMs. An additional measure, to install now INF-like portal monitoring systems at mobile final assembly facilities, would address a problem all agencies agree exists: the longer we wait for START, the greater will be the uncertainty about how many Soviet SS-24s exist.¹⁵ All also agree that continuous monitoring will improve our confidence. There is disagreement over what type of monitoring system we need in START, what facilities we must monitor, and how much uncertainty exists with a monitoring system in place.

RECOMMENDATION

That you scan the papers at Tabs A, B, C, and D and use the points outlined above in guiding discussions at the Deputies Meeting on START Mobiles and New Initiatives.¹⁶

John Gordon, Bob Blackwill and Condi Rice concur.

Tab A

Paper Prepared by the Policy Coordinating Committee¹⁷

Washington, June 1, 1989

START 06D

NSR 14 START Review Paper Six—Mobile ICBMs

Purpose. To review options for modifying the U.S. position on dealing with mobile ICBMs under START.

Background. Because of verification and stability concerns, the previous administration sought to ban mobile ICBMs under START. The United States did indicate, however, that, if a satisfactory verification scheme could be devised, we could consider allowing a strictly limited number of mobile ICBMs. The United States has not formally established what is meant by “strictly limited number”; internally we have considered a mobile ICBM warhead sublimit in the range of 500–1,000 warheads. The Soviets have proposed limiting mobile ICBMs to no more than 1,600 warheads on no more than 800 missiles.

Although maintaining its formal position in the Joint Draft Text that mobile ICBMs should be banned, the United States agreed in 1988 to work on verification provisions for mobile ICBMs. A joint working paper has been prepared. The basic U.S. proposal would confine mobile ICBMs to relatively small locations in peacetime except for notified transits or infrequent exercise or operational dispersals. (We have not specified what percentage would be allowed on notified transits at any one time; internal discussions have focused on a range of 15–20 percent. [1½ lines not declassified] The size of the overt inventory of mobile missiles and launchers would be verified through data exchange and monitoring of production, with a system of tagging used to identify legal missiles. The Soviets have accepted portions of this scheme, but have rejected others. The tabled U.S. proposals deal exclusively with managing the legal inventory of mobile ICBMs and launchers and with monitoring future production at declared facilities. Our existing Suspect Site Inspection regime, is focused on ballistic missile production.

On April 22, 1989, the President adopted a U.S. ICBM modernization program consisting of (a) rebasing the existing 50 silo based Peacekeeper missiles in a Peacekeeper rail garrison mode by the early 1990s and (b) development of the small ICBM for deployment in a road-mobile configuration by the late 1990s. The President did not specify (a) whether the small ICBM would carry one warhead or two, (b) how many Small ICBM missiles/warheads would be deployed, and (c) which of the competing basing modes for the small ICBM would be selected. Basing modes under consideration include so-called Southwest basing, involving periodic random movement, and basing on existing Minuteman sites to either disperse when directed or dash on tactical warning.

Discussion. Given the foregoing background, the following issues appear to warrant review:

— Issue 1: Whether to alter our formal position banning such missiles;

— Issue 2: Whether to consider an option banning MIRVd mobile ICBMs while allowing single-RV mobile ICBMs;

— Issue 3: If mobiles are to be allowed, what numerical constraints to seek on them;

— Issue 4: Whether, and if so, how to address accountability and inspection provisions for “movable” ICBMs (e.g., fixed area, multiple shelter concept).

A separate paper will examine whether, and if so how, to modify the existing verification regime the U.S. has proposed in the Joint Working Paper.

Issues 1/2—Allowing Mobile ICBMs. Mobile ICBMs tend to have greater survivability, and hence, greater crisis stability, assuming START restrictions either (a) allow dispersal when required or (b) allow operation in Deployment Areas large enough to ensure survival without dispersal. A situation in which both sides possess substantial mobile ICBM forces, and therefore survivable, land-based second-strike capability, is relatively stabilizing. The same factors which lead to this advantage, however, mean that mobile ICBMs are difficult to target—which creates a situation in which the U.S. may not be able to achieve target coverage and damage goals against Soviet forces. All agree that the least acceptable outcome with respect to mobile ICBMs would be one in which the Soviets shift major portions of their ICBMs into mobile basing (as is projected), and the United States fails to ensure the survivability of its ICBMs through mobility and/or deployment of strategic defenses. The resulting situation would compound existing force asymmetries and could have far-reaching effects in its impact on public perceptions of the strategic nuclear balance.

This implies that mobile ICBMs should be allowed only if a domestic political consensus that would allow deployment of a robust U.S. mobile ICBM force has been achieved. (We cannot, however, prevent Soviet deployments if the Soviets are willing to forego a START agreement.)

Mobile ICBMs pose serious monitoring, verification and breakout problems. [2 lines not declassified]

[3 paragraphs (30 lines) not declassified]

Some are greatly concerned that the Soviets could use mobile ICBMs as an avenue for cheating or rapidly breaking out of START. It was concerns of this nature that led the U.S. to seek a ban on mobile ICBMs initially. These concerns will remain until a regime for effective verification of these systems has been identified and agreed upon. As noted above, there are differing views over whether the U.S. will be able to complete the development of and negotiate successfully a regime that meets our verification requirements and is acceptable in its impact on the U.S.

The foregoing considerations suggest to some that the United States should continue to seek a mobile ICBM ban. On the other hand, banning mobile ICBMs forgoes the stability benefits they offer. In addition, a major goal of the strategic review has been to bring our procurement and arms control positions into harmony. Despite the successful INF example (negotiating to ban systems while we were deploying them), some believe it is inconsistent to seek a mobile ICBM ban while embarking on the development of not one but two mobile missiles, and that attempting to do so will prove increasingly unacceptable to Congress. Others note that this has been the U.S. position for some time and has served our interests.

One approach to dealing with the competing concerns outlined above would be to ban MIRVd mobile ICBMs while allowing single RV mobiles. Such an approach would ban rail-mobile versions of the SS–24 and Peacekeeper, while allowing deployment of the SS–25 and Small ICBM. Banning MIRV mobiles could improve stability; the threat to stability from unaccounted mobile ICBM launchers is less if such launchers contain a single warhead rather than ten. If MIRVd mobiles were banned, we might also want to add a limit on MIRVd silo-based ICBMs to constrain the incentive to shift ICBM forces from survivable mobiles to vulnerable silos. This approach is discussed further in the NSR 14 paper on New Initiatives.

Two augmentations of such a ban might improve its effectiveness. First, we could seek a ban on all rail mobile systems, thus preventing the Soviets from developing a single-RV rail-mobile ICBM and maintaining a rail-mobile infrastructure. Rail-mobile ICBMs have characteristics that make them inherently more difficult to monitor than other strategic weapons, including road-mobile ICBMs. Monitoring the Soviets’ SS–24 rail-mobile ICBM is more challenging than monitoring the SS–25 road-mobile system because augmentation of the deployed force with nondeployed missiles and launchers appears easier for rail-mobile systems than for road-mobile systems. Additional details are attached.

A second possible augmentation of a MIRVd mobile ban would be to ban or limit silo-based SS–24s (and by implication silo-based Peacekeeper) to limit breakout by eliminating the existing rail-mobile infrastructure, assuming the Soviets do not develop another rail-mobile ICBM. On the other hand, rail-mobile launchers are difficult to detect, and the fact that the Soviets have already tested a MIRV mobile reduces their breakout time. If the Soviets were allowed to retain silo based SS–24s, they could use the associated nondeployed missiles (or even the silo based SS–24s if they wished to make them more survivable) in a breakout scenario with covertly maintained SS–24 launchers or legal launchers for a single-RV rail-mobile system. Finally, such a ban would eliminate the U.S. Peacekeeper, the world’s most capable ballistic missile, with serious impact on U.S. targeting and damage goals.

Congressional reaction must also be considered. The President’s ICBM modernization decision, in addition to its other benefits, offers something to supporters of both missiles. The consensus we are seeking to build could unravel if rail-mobile Peacekeeper were seen only as a bargaining chip. In addition, many believe that a START Treaty which allows mobile ICBMs will prove unratifiable in the absence of at least a well-funded U.S. mobile ICBM program, if not actual deployment. Thus if MIRVd mobiles were banned and the U.S. road-mobile program were geared to the late 1990s there could be ratification difficulties.

Finally, a MIRV ban would delay U.S. introduction of a mobile ICBM by five years, giving the Soviets an additional several years of monopoly. In addition, such a ban would significantly increase the cost of mobile ICBM deployment. Attached to this paper are possible costs for a variety of mobile ICBM forces, assuming a mobile ban took effect after deployment of rail-mobile Peacekeeper. Most of the cost difference would be in budgets beyond FY 1994.

If we are to accept mobiles at all, we should examine what, if any, concessions the United States should seek from the Soviets for formally accepting mobile ICBMs in START. Some believe that maintaining a formal ban adds complexity to the negotiations and may provide little leverage if the Soviets believe that we ultimately will allow mobile ICBMs, especially given an announced Presidential decision to proceed with deployment of two mobile ICBM systems. Others note that we will never know unless we try and suggest we should, at a minimum, continue to insist on acceptance of our verification approach as a price for falling off our mobile ban.

[2 paragraphs (20 lines) not declassified]

Options—Allowing Mobile ICBMS. Three options exist:

— Option 1—Maintain a formal ban on mobile ICBMs pending resolution of our verification concerns.

— Option 2—Propose a ban on all MIRV mobile ICBMs, while allowing single warhead mobiles, subject to appropriate sublimits. Modify the Small ICBM program to eliminate the option for two warheads. Augment this with a ban on single warhead rail mobile ICBMs to ensure that the rail-mobile infrastructure is eliminated.

— Option 2A—Include a ban on silo-basing of types of ICBMs deployed as mobiles (SS–24 and presumably Peacekeeper).

— Option 3—Modify the U.S. position to allow both single-warhead and MIRVd mobile ICBMs, subject to appropriate sublimits.

If Option 2 is selected, additional study is required as to how to limit breakout potential from silo-based SS–24s and follow-on systems. In addition to an outright ban, such constraints could take the form of specific sublimits on high-MIRVd systems or of general restrictions on total MIRVd warheads. Study is also required of what additional constraints (e.g. missile size or minimum RV weight-to-throwweight ratio) should be imposed on mobile missiles to limit the potential for breakout through covert fractionation.

Issue 3—Mobile ICBM Sublimits. Assuming mobile ICBMs are to be allowed in START there is consensus that some form of sublimit is appropriate. Such a sublimit is intended to constrain Soviet deployment of mobiles to a level no greater than the U.S. plans to deploy. Verification and breakout concerns are not the driving considerations in limiting the number of mobile ICBMs once the number of launchers is several hundred. (Lower sublimits, under one hundred, should present lesser verification difficulties.) There are several approaches to dealing with a mobile ICBM sublimit.

One approach is to defer discussion of such a sublimit until the Soviets accept our verification scheme. This approach was endorsed by the previous administration. A second approach is to propose a stand alone mobile ICBM sublimit now. Various values of such a sublimit were considered in the past administration, with 500–700 warheads being accepted by most agencies.

A third approach is to propose a mobile ICBM sublimit, but to nest it within the 1540 warhead sublimit for heavy ICBMs. The Soviets would be allowed a total of 1540 warheads on heavy ICBMs and mobile ICBMs, of which no more than 500–800 could be on mobile ICBMs. This approach would force the Soviets to trade heavy ICBMs for mobiles, thus forcing tradeoffs between two systems of greatest concern to the United States. If the Soviets accept this approach either the first strike threat from Soviet heavy ICBMs or the number of relocatable targets the United States must hold at risk would be reduced. On the other hand, some will find such an approach difficult to explain since it appears to equate the SS-18, the most destabilizing Soviet system, and mobile systems which some believe are inherently stabilizing. A variant of this approach would focus strictly on MIRVd mobiles as a tradeoff with heavies. Since non-MIRVd mobiles do not present the same breakout threat as do MIRVd mobiles, they would be handled under different sublimits.

The past administration considered mobile ICBM warhead sublimits as low as 500 warheads, which would only allow 50 rail garrison Peacekeepers, consistent with its budget position not to deploy the Small ICBM. If we are to protect the forces implied by the President’s ICBM modernization decision, the U.S. proposed sublimit must be sufficiently above 500 warheads to allow for a reasonable small ICBM deployment. A value of 800 thus appears the minimum the United States could accept and still preserve both programs (assuming we do not seek a ban on MIRVd mobile ICBMs). Since the Soviets have proposed 1600, there is incentive for the U.S. to suggest a low number. Thus some believe 800 warheads is an appropriate initial U.S. proposal if both single and MIRV mobile ICBMs are to be allowed. Others would propose a number as high as 1000; if we propose 800, the Soviets may accept and we could lose flexibility with respect to the Small ICBM.

On the other hand, if the United States is to seek a ban on MIRV mobile ICBMs, a lower value, perhaps 200–500 warheads, may be more appropriate. The United States is unlikely for fiscal reasons to be able to deploy more warheads than this under a MIRV mobile ban when only the single-RV variant of the Small ICBM would be allowed. We may also have a separate deterrence reason to limit the number of Soviet relocatable targets. It is thus in our interest not to allow the Soviets a greater number than the U.S. plans.

Options—Mobile ICBM Sublimit. If the United States decides now to change its position and allow mobiles, there appear to be the following options for dealing with the question of a mobile sublimit when START negotiations resume.

If MIRVd mobiles are permitted:

— Option 1—Advise the Soviets that, while we expect to propose a mobile ICBM sublimit, and we expect that sublimit to be substantially smaller than the Soviet proposed 1600 sublimit, we are not prepared to discuss specific values until agreement has been reached on a verification approach.

— Option 2—Propose a mobile ICBM sublimit of 800–1000 warheads.

— Option 3—Propose a mobile ICBM sublimit of 800–1000 warheads nested within a 1540 warhead sublimit on combined mobile ICBM and heavy ICBM warheads.

— Option 4—Propose treating MIRVd and non-MIRVd mobile ICBMs differently:

— Include a sublimit of (500?) warheads on MIRVd mobile ICBMs within the 1540 sublimit on heavy ICBM RVs;

— Include a separate sublimit of (300?) warheads on non-MIRVd mobile ICBMs within the 4900 (and 3300) sublimits.

If MIRVd mobiles are not permitted:

— Option 1—Advise the Soviets that, in the context of a ban on MIRVd mobiles, we expect to propose a mobile ICBM sublimit, and we expect that sublimit to be substantially smaller than the Soviet proposed 1600 sublimit; we are not prepared to discuss specific values until agreement has been reached on a verification approach.

— Option 2—Propose a mobile ICBM sublimit of 300–500 warheads.

— Option 3—Propose a mobile ICBM sublimit of 200–300 warheads nested within a 1540 warhead sublimit on combined mobile ICBM and heavy ICBM warheads.

Issue 4—Movable ICBMs. The current United States position is that a START Treaty must protect the option to deploy so-called “movable” ICBMs in the future. The United States has explained the concept of movable ICBMs as an alternate approach, different from both mobile ICBMs and silo-based ICBMs, intended to enhance ICBM survivability. The movable ICBM concept involves a large number of low cost vertical shelters, among which a hardened capsule containing an ICBM is rotated. Survivability is gained by denying the Soviets knowledge of which specific shelter among the large number of vertical shelters contains a missile.

Although the United States has indicated to the Soviets that we believe a future START Treaty must provide for movable ICBMs, we have provided no details and no treaty text. There is as yet no agreed verification approach for such ICBMs.

In developing the ICBM modernization options in the defense strategy review (NSR-12) little consideration was given to movable ICBMs. There is no program within the Department of Defense to develop such ICBMs. Some believe that, now that the President has made an ICBM modernization decision, the United States no longer needs to protect the option of deploying movable ICBMs, and that we should cease attempts to include provisions protecting such an option in START. We would not want to allow the Soviets this option if it would not be exercised by the United States.

Others believe that, until it is clear what the Congress will do, we can not rule out the prospects that we will need to find an alternate solution to both Peacekeeper and small ICBM deployment in order to ensure ICBM survivability. Further, those who hold this view note that the President’s decision results in only part of the ICBM force being survivable, and that we may wish to increase the survivability of the remainder of the force in future years.

Finally, some believe that, even if additional mechanisms for enhancing ICBM survivability are desirable, as a practical matter by the time a movable ICBM could be deployed, the United States should be in a position to enhance ICBM survivability through the deployment of strategic defenses. Others believe that strategic hard point defense may work particularly well in conjunction with a multiple shelter basing approach.

Options—Movable ICBMs. Three options exist:

— Option 1—Drop further consideration of movable ICBMs in START.

— Option 2—Continue to insist that a future START Treaty include provisions allowing movable ICBMs; defer detailed discussions with the Soviets; develop detailed verification scheme for such ICBMs.

Attachment

Paper Prepared by the Policy Coordinating Commmittee¹⁸

Washington, undated

ACQUISITION COST ESTIMATES FOR MOBILE ICBM OPTIONS

Based on the nominal forces shown in the chart, three key points on cost are:

(1) The acquisition cost per mobile ICBM warhead deployed under a MIRVed mobile ban is 2 to over 2.5 times higher than if START allows MIRVed mobiles.

(2) The acquisition cost of the mobile ICBM force under a MIRVed mobile ban depends on how many additional SICBMs we would deploy. For example, the incremental 200 SICBMs in the second force shown under the ban adds $6.6 billion.

(3) The force-wide acquisition costs of banning MIRVed mobiles could be greater than those shown in the chart, depending on what other adjustments we make in our strategic forces. In all likelihood, the U.S. would not rely entirely on the SICBM to offset 500 warheads banned on rail-garrison. Instead, we might replace them with other modernized systems, not currently programmed. In the chart, the 300 SICBM force assumes that all 500 banned rail-garrison warheads will be replaced by strategic systems other than mobile ICBM launchers; the 500 SICBM force assumes that 300 of the 500 forfeited warheads will be replaced by systems other than mobile ICBMs. If the U.S. retains Peacekeepers in existing silos, no additional cost is incurred. However, if Peacekeeper is banned (Option 2b) or we chose, for stability reasons, to deploy this 300–500 warhead allowance on platforms not currently programmed, then the cost of these platforms is added to those shown in the chart. A feasible rail-garrison offset might be two or three additional, non-programmed, Trident submarines with D–5 missiles. Each boat loaded with 24 D–5s (192 warheads) would add nearly $2 billion to the cost of a ban.

Attachment

Chart Prepared by the Policy Coordinating Committee¹⁹

Washington, undated

ACQUISITION COST ESTIMATES FOR MOBILE ICBM OPTIONS

	IOC/FOC	Acquisition Costs: FY 90 to Completion (1988 $s in Billions)	Acquisition Costs Per Deployed Warhead: FY90 to Completion (1988 $s in Millions)
Baseline Force for an 800 Warhead Sublimit on Mobiles — 50 Rail-Garrison PK — 300 SICBMs — Total	1992/1994 1997/2005 (est.)	5.4 17.020 22.4	28.0
Force for a 300 Warhead Sublimit on Mobiles with a Ban on MIRVed Mobiles — 50 Rail-Garrison PK Procured then Deactivated21 — 300 SICBMs — Total	1992/1994 1997/2005 (est.)	5.422 17.023 22.4	74.7
Force for a 500 Warhead Sublimit on Mobiles with a Ban on MIRVed Mobiles — 50 Rail-Garrison PK Procured then Deactivated24 — 500 SICBMs — Total	1992/1994 1997/2008 (est.)	5.425 23.626 29.0	58.0

Tab B

Paper Prepared by the Policy Coordinating Committee²⁷

Washington, June 2, 1989

START 07D

NSR 14 START Review Paper Seven—Mobile ICBM Verification

Purpose. To review options for modifying the U.S. approach to verifying mobile ICBM limits under START.

Background. Because of verification and stability concerns, the previous administration sought to ban mobile ICBMs under START. The United States did indicate, however, that if an effective verification scheme could be agreed we would consider allowing a strictly limited number of mobile ICBMs.

Although maintaining its formal position in the Joint Draft Text that mobile ICBMs should be banned, the United States agreed in 1988 to work on verification provisions for mobile ICBMs. A joint working paper has been prepared. The basic U.S. proposal would:

— Monitor additions to the mobile inventory through Perimeter Portal Monitoring (PPM) of production facilities producing the accountable stage of mobile ICBMs.

— Monitor the legal inventory through a verification scheme which confines mobile ICBMs and their launchers to relatively small locations in peacetime except for notified transits and, in the case of deployed mobiles, infrequent exercise or operational dispersals.

— Guard against illegal activity through a guaranteed right of short-notice on-site inspection at an agreed list of facilities and through the ability to conduct suspect site inspections (SSI). Except for the agreed list of facilities, suspect site inspections would be subject to a right of refusal.

— Deter use of the legal infrastructure by any illegal missiles and their launchers through a system of tagging (not yet accepted by the Soviets and not yet technically proven in an operational environment). Periodic short-notice inspections of legal facilities would check that only tagged (i.e. legal) missiles were present.

On April 22, 1989, the President adopted a U.S. ICBM modernization program consisting of (a) rebasing the existing 50 silo based Peacekeeper missiles in a Peacekeeper rail garrison mode by the early 1990s and (b) development of the small ICBM for deployment in a road-mobile configuration by the late 1990s. The President did not specify (a) whether the small ICBM would carry one warhead or two, (b) how many Small ICBM missiles/warheads would be deployed, and (c) which of the competing basing modes for the small ICBM would be selected. Basing modes under consideration include so-called Southwest basing, involving periodic random movement, and basing on existing Minuteman sites to either disperse when directed or dash on tactical warning.

Discussion. A separate paper has reviewed whether to alter our formal position banning mobile missiles; whether to consider an option banning MIRVed mobile ICBMs while allowing single-RV mobile ICBMs; and, if mobiles are to be allowed, what numerical constraints to seek on them. This paper examines whether, and if so how, to modify the existing verification regime the U.S. has proposed in the Joint Working Paper.

Mobile ICBMs pose serious monitoring, verification and breakout problems. [6 lines not declassified] Agencies differ about whether our approach to missile production monitoring, and the other measures for verifying compliance with limits on mobile ICBMs (e.g. monitoring logistics and force structure), can meet our verification standards. All agree, however, that the longer the Soviets produce mobile missiles unmonitored, prior to a treaty, the greater will be the uncertainty as to their total inventory.

[1 paragraph (15 lines) not declassified]

Thus, if mobile ICBMs are to be allowed, the verification scheme involving them must carefully balance the risks of Soviet cheating with the costs of reciprocal impacts on the U.S. These considerations will also be weighed carefully by the Senate during the ratification debate. While all aspects of a mobile verification regime are important, four specific questions warrant examination before negotiations resume:

(1) should we modify the restrictions on road-mobile ICBMs;

(2) should we modify our position on rail-mobiles to restrict missiles to a specified Deployment Area;

(3) is our approach to continuous monitoring of production sufficiently broad based; and

(4) should we require a more extensive suspect site inspection regime?

Road Mobile ICBM Basing Mode. The U.S. approach to verification of road-mobile ICBMs confines such systems to 25 square kilometer Restricted Areas peacetime except for notified transits or infrequent exercise or operational dispersals. The United States would place a percentage limit (not yet determined) on the number of systems which can be in transit outside Restricted Areas. The U.S. concept includes a Deployment Area which contains several Restricted Areas, but the only function of this larger area is to constrain exercise dispersals. Operational dispersals are not constrained to any area, but the reason for the operational dispersal would be notified.

The JCS understand that the U.S. verification proposal is compatible with current Small ICBM basing concepts for both the Minuteman basing mode and Southwest basing mode. However, the final basing mode for Small ICBM has not been selected and some possible basing concepts are not compatible with the U.S. verification proposal. The JCS believe that should the U.S. decide to base the Small ICBM in the Southwest, the only impact on operations would be the requirement for notification of each movement out of the Restricted Areas. Additional study is needed to ensure that the percentage of the force which could be out of Restricted Areas at any given time is sufficient to accommodate projected random movement operations.

Although the existing U.S. concept could support a form of random periodic movements, some believe it is better suited to Minuteman basing. They believe that a primary rationale for Restricted Areas is to aid in monitoring the number of deployed missiles and launchers. If the number of missiles/launchers permitted outside the Restricted Area at any one time is small (10% or less), then the IC believes monitoring is improved. Such small numbers may be compatible with training and maintenance requirements for Small ICBM Minuteman Basing, but are inadequate for Southwest Basing, which may require up to 30% out of the Restricted Areas. Moreover, some believe that it is impossible to assign a maximum percentage that will protect future operational concepts that may be necessary to ensure survivability. Finally, depending on the size and number of Restricted Areas, the area available to each missile may not be sufficient for survivability without dash on warning.

Two approaches have been suggested to address these concerns. One revised approach would eliminate the Restricted Areas in favor of larger Deployment Areas in which road-mobile missiles could operate without restriction. Another approach would eliminate the percentage limit on the number of missiles that could be outside the Restricted Areas at any one time. (Variants of these approaches would not permit missiles outside the larger Deployment Area for operational dispersals.)

Arguments for altering the existing U.S. position include:

— The goal of the Small ICBM program is stability through deployment of an ICBM which does not depend on warning for survivability. Periodic random movement (Southwest Basing) is the best way to achieve that goal. An altered U.S. approach would permit random unnotified movement within a Deployment Area.

— Since a larger Deployment Area is more consistent with Soviet practices, it should be easily negotiable. Thus, changing our position would facilitate reaching agreement on a mobile ICBM verification regime.

— Some believe that allowing geographically unlimited operational dispersals, as in the current approach, poses a severe verification risk. If a Deployment Area were large enough to ensure survivability, it would be possible to ban dispersals outside this area. Others believe, dispersals will still be required for rail-mobile systems, and, in the past, the JCS military assessment has been that any geographic restrictions on operational dispersals is militarily unacceptable.

Arguments in favor of retaining the existing position include:

— The current verification provisions are compatible with both Minuteman basing and Southwest basing.

— Monitoring confidence does not change substantially with the addition of a large Deployment Area or with the deletion of Restricted Areas. We will still have to monitor numbers of missiles within the Deployment Area and detect operations outside the Deployment Area.

— Current requirements to report all movements from the Restricted Area would provide additional information on Soviet mobile operations.

— No decision has been made on Small ICBM basing. Some believe, however, that we will more likely deploy the more restrictive Minuteman basing. Thus, the flexibility we gain from an alternate verification approach may be illusory, and it is in our interests to impose similar limitations on the Soviets.

— Even if we decide to alter our position, we should not do so prior to returning to Geneva. If the Soviets wish a larger area, we should get something for agreeing to their request.

Options—Road Mobile Basing Mode. Four options exist:

— Option 1—Retain the existing position confining mobile ICBMs to a small Restricted Area.

— Option 2—Alter the existing position to drop the small Restricted Area and adopt a larger Deployment Area. Movement is unrestricted within the Deployment Area and there is no requirement to notify movement within the Deployment Area.

— Option 2A—Continue to allow operational dispersals without geographic limit.

— Option 2B—Do not allow any dispersals beyond the Deployment Area.

— Option 3: Alter the existing position to adopt a larger Deployment Area and very small Restricted Areas (perhaps no larger than missile-operating bases). Movement within the Deployment Area is unrestricted, but changes in the net number of missiles/launchers in each Restricted Area must be pre-notified. The concept of exercise dispersals is eliminated.

— Option 3A—Continue to allow operational dispersals without geographic limit, with post-notification and a stated national security rationale.

— Option 3B—Do not allow any dispersals beyond the Deployment Area.

Rail Mobile ICBM Basing Mode. The U.S. approach to verification of rail-mobile ICBMs parallels the approach taken for road-mobile ICBMs, except that there is no provision for a Deployment Area for constraining exercise dispersals. Thus, rail mobiles would be confined to rail garrisons no larger than a specified size (yet to be determined), and would be allowed to leave the garrison for training, maintenance and testing, as well as for dispersals. There are no geographic or other operational limitations on rail-mobile ICBMs when they are taking part in exercise or operational dispersals. In addition, the U.S. position contains provisions which require trains with rail-mobile launchers to be in a standard configuration while located in rail garrison but no requirement to be in any standard configuration outside the garrison (although configuration changes for routine movements must be notified).

The U.S. approach to rail-mobile verification is consistent with the rail-mobile Peacekeeper, allowing Peacekeeper trains to have access to most of the U.S. rail network. The President’s ICBM modernization decision of April 22, 1989 is consistent with this approach.

However, some are concerned that under the current U.S. proposal, which allows geographically unlimited dispersals from small garrisons. [2 lines not declassified] They also believe that our current approach may detract from stability because it would require dispersal of the force during a crisis for survivability. They believe that the need for such dispersal could be a factor toward inducing a Soviet first strike.

Those who favor revising the current position believe that measures to address these concerns include the following:

— A geographically limited Deployment Area to mitigate the verification problem by restricting the area of legal deployment and setting boundaries beyond which any deployment would be illegal.

— A Deployment Area partitioning scheme is not inconsistent with preserving mobile ICBM survivability, provided that the areas are sufficiently large so that the Soviets cannot destroy the mobile force with a barrage attack without an unacceptably high exchange ratio.

— It is highly unlikely that Peacekeeper trains will have the time for dispersal to any point in the U.S. during a crisis situation. In addition, they believe Peacekeeper trains do not require unlimited access to the civilian rail network during peacetime for purposes of training.

— A verification regime based on large Deployment Areas would not require dispersals outside these areas for survivability and thus would not detract from the stability advantages of mobile deployments.

— Because of the more difficult verification problems posed by rail mobile systems, we should leave open the possibility for further studies directed at developing other types of limits on rail mobiles.

Others note that all of these issues were considered in great detail by the Reagan Administration. They believe the current approach allows for the operational flexibility necessary to ensure survivability in the future while also providing a necessary degree of predictability and force transparency. The Peacekeeper concept requires that the missiles be able to deploy onto the full commercial rail network and blend into the larger commercial train population, forcing the Soviets to barrage the entire network or search the entire network in the hope of employing selective targeting.

Arguments in favor of retaining the current U.S. position include:

— The U.S. and the Soviets have essentially agreed on the basic elements of the rail-mobile verification regime which includes no geographic restrictions. Thus, altering the agreed position could entail a negotiating cost, a consideration the President has specifically directed be taken into account.

— The current garrison approach requires a strict correlation between missiles and their bases and notification of all routine movements. Some believe this is a strong deterrent to the mixing of legal and illegal systems during exercises. They also believe this segregation makes an illegal force much more vulnerable to discovery.

— Notification of movements outside the garrison in the current approach enhances the predictability of normal operations and provides a mechanism to clearly convey national intentions with regard to operational dispersals.

— Restricting rail mobiles to Deployment Areas would limit the total mileage of rail lines available and would ease the Soviet search and targeting task.

— In the case of a dedicated rail network, such a basing system is fiscally infeasible.

— Some believe the Deployment Area concept would degrade day-to-day surveillance of Soviet rail mobiles unless we also propose either strict restrictions on the number and types of trains and support infrastructure allowed in the Deployment Area, or pervasive inspections.

— Some believe that because of likely political and operational restrictions on rail mobile Peacekeeper, they will not patrol in peacetime and, thus, the U.S. could not take advantage of the benefits of unrestricted movement within a Deployment Area.

Options—Rail Mobile Basing Mode. Three options exist:

— Option 1—Retain the existing position allowing rail-mobile ICBMs to depart the rail garrison for training, maintenance, and testing purposes, and for exercise and operational dispersals with notifications, but without geographic limitations.

— Option 2—Alter the existing position to adopt a large Deployment Area with unrestricted movement within that Deployment Area. No dispersals permitted outside the Deployment Area.

— Option 2A—Provide for Deployment Areas using the civilian rail network limited only by geographic boundaries.

— Option 3: Alter the existing position to adopt a larger Deployment Area in addition to the small Garrisons. Movement within the Deployment Area is unrestricted, but changes in the net number of missiles/launchers in each Garrison must be pre-notified. The concept of exercise dispersals is eliminated.

— Option 3A—Continue to allow operational dispersals without geographic limit, with post-notification and a stated national security rationale.

— Option 3B—Do not allow any dispersals beyond the Deployment Area.

Perimeter Portal Monitoring (PPM) and Suspect Site Inspections (SSI). The U.S. proposes to monitor additions to the mobiles inventory through PPM of facilities that produce accountable stages of mobile ICBMs. In addition, PPM would be used to deter covert production at facilities that are producing any ICBM or SLBM stage and are also producing any type of motor that exceeds the length and diameter of the smallest mobile accountable stage. SSI is intended to deter covert mobile production at other facilities. Short-notice SSI (without a right-of-refusal) would be permitted at facilities on an agreed, but not yet determined, list that is intended to focus chiefly on solid rocket motor production facilities; SSI at all other locations would be subject to a right-of-refusal, although the refusing party is required to make a “good faith” effort to resolve the concern by other means. There would be quotas on listed facility and right-of-refusal inspections.

All agree that our current position on PPM permits us confidently to determine the production of accountable stages for mobile missiles at declared production plants (which are subject to continuous monitoring) and to deter excess or illegal production at those locations (assuming our definition captures Soviet facilities at which segmented motors are assembled). [4½ lines not declassified] In addition, some believe that our current approach does not adequately protect U.S. facilities from intelligence loss during SSI. [2 lines not declassified]

PPM. The current approach would subject to PPM a subset of the facilities capable of producing rocket motors; namely, plants currently engaged in strategic rocket motor production which also produce a motor at least as large as a mobile first stage. It does not provide for PPM at all production plants which may have critical equipment in place to produce a mobile first stage motor or segment of a motor (these facilities would be subject to SSI).

Some believe that the current approach is insufficient for effective monitoring and that we should PPM all facilities that have casting and curing pits/chambers of sufficient width and depth to produce an accountable motor or segment (additional 3 U.S. and 3–6 Soviet facilities). Those who hold this view also believe that, for facilities that have casting and curing pits/chambers of sufficient width and depth to produce an accountable motor or segment, safety considerations would preclude ascertaining whether the contents of any casting and curing pits/chambers in use during a 24 hour SSI was an accountable motor (normal curing times run 7–14 days). They are uncertain what ‘changed practices’ would be visible to NTM or SSI. They also believe that the question of what is required for an effective PPM system—and its associated costs—should be revisited in light of our INF experience.²⁸ Finally, in this view, if SSI is to be demonstrated a viable alternative, the concept must be developed and its impacts and costs evaluated, neither of which has been done.

Others believe that the current approach captures the facilities where covert production would be easiest to carry out and that such production at other facilities would require a change in practices that the Soviets could not be confident of hiding from U.S. NTM and short-notice SSI. In this view, the cost of establishing additional PPM sites and the intrusiveness of additional Soviet PPM sites in the U.S. are not justified if we are satisfied with the potential Soviet production facilities included on the SSI list.

A second PPM issue concerns the current approach not to include a right to PPM mobile launcher production plants. Some believe that PPM of launcher production plants would reduce the potential for covert production and deployment. A PPM system would also be consistent with the tagging of launchers, which may require a continuous presence at some Soviet facility. Others believe that the cost of additional PPM sites is not justified [5 lines not declassified]

Options—PPM. Three options exist:

— Option 1—Retain the existing position for PPM only at facilities that are producing an ICBM or SLBM stage and are also producing any type of motor that exceeds the size of the smallest mobile accountable stage.

— Option 2—Retain the existing position for PPM, but do not seek to engage the Soviets on the issue. If the Soviets press, inform them that the issue is under review in Washington.

— Option 3—Inform the Soviets that we are withdrawing our PPM proposal pending a review of our position.

SSI. The existing proposal would permit mandatory short-notice SSI at facilities on an agreed list, for the purpose of resolving compliance concerns regarding ballistic missile-related activities. [1 line not declassified] However, no specific proposal has been made to the Soviets, or agreed to in the USG. SSI at all other locations would be subject to a right of refusal, although the refusing party is required to make a “good faith” effort to resolve the concern by other means.

The first issue concerns the scope of facilities on the mandatory SSI list. Many believe the list should comprise only solid rocket motor production plants where PPM is not in place. They note that the issue of how to balance the U.S. interest in a comprehensive SSI regime to deter Soviet cheating, with our need to protect sensitive U.S. facilities was the subject of extensive study in the Reagan Administration. Although the existing proposal is not ideal in serving any one of our objectives, it strikes a good balance between our verification and security requirements. It also is consistent with the U.S. constitutional protection against unreasonable search.

Others believe that the list should be expanded to include almost all solid rocket motor production plants that are not subject to PPM, major military storage locations (about 11 U.S. facilities, depending on the definition), and civilian railyards and open areas in which mobile launchers could be, and in the case of Soviet SS–24 and SS–25 systems, have probably have been routinely located.

Options—Mandatory SSI. Three options exist:

— Option 1—Retain the existing position on Mandatory SSI.

— Option 2—Retain the existing position on Mandatory SSI, but do not seek to engage the Soviets in Geneva. If the Soviets press, inform them that the issue is under review in Washington.

— Option 3—Inform the Soviets that we are withdrawing our Mandatory SSI proposal pending a review of our position.

A second issue is whether to modify the right-of-refusal portion of our proposal. As noted above, the issue of how best to deter Soviet cheating and protect sensitive U.S. facilities was the subject of extensive study in the Reagan Administration. While no decisions were made on the details of a U.S. proposal, many believe that the existing U.S. approach strikes a good balance between our verification and security requirements and should be retained. They believe that continued study is unlikely to uncover a better approach. Nevertheless, all agree that the costs associated with right-of-refusal are not yet well defined.

Others believe that the current proposal neither protects sensitive U.S. programs nor deters Soviet cheating, since the U.S. will be under pressure to accede to Soviet requests while the Soviets will reject U.S. requests if they are cheating. They believe we need to take a comprehensive look at mobile verification to see if there are new approaches to the issue. Among the possibilities to be examined are: 1) a combination of the more comprehensive PPM approach and SSI lists (see above), and mandatory SSI at suspected solid rocket motor production plants (at the challenged party’s option, this SSI may consist only of sampling air, water, and soil outside of buildings; right to request satisfaction of other concerns, but no explicit right to demand inspection); and 2) a mandatory SSI approach which would provide for exclusions based on function, e.g., buildings or rooms containing intelligence equipment.

Options—Right-of-Refusal SSI. Two options exist:

— Option 1—Maintain the current approach to permit SSI at locations not on the list, subject to a right-of-refusal; if a request is refused, the refusing party is required to make a “good faith” effort to resolve the concern by other means.

— Option 2—Inform the Soviets that we are unprepared to discuss our Right-of-Refusal SSI proposal pending a review of our position.

A third issue is whether to permit SSI at facilities subject to PPM, but where the monitoring side decides not to establish a PPM system. The current U.S. position specifies that such facilities would be subject to either mandatory SSI or right-of-refusal SSI. Prior to entry into force, a list of facilities subject to mandatory SSI would be agreed between the sides. However, procedures for determining which category such facilities would be placed after entry into force have not been developed.

Some believe that it is desirable to have the right to mandatory inspection of Soviet facilities we have chosen not to PPM, and that the security risk of such Soviet inspections of U.S. facilities is acceptable. They believe that since the Soviets have more facilities that may be subject to PPM and since the Soviets will insist on equal numbers of PPM sites in the U.S. and Soviet Union, this mandatory SSI right would be to our advantage.

Others believe that it makes little sense to offer the Soviets the right to either type of SSI in lieu of PPM since, in this view, we will require continuous monitoring to count Soviet production. They believe the right to SSI would not provide a cost saving to us or aid to verification, but would give the Soviets access to our production techniques for strategic missiles and large space/SDI boosters, should they choose to exercise the SSI option. They further believe that subjecting such facilities to right-of-refusal SSI would severely undermine our ability to prevent covert production, since it is likely that the Soviets would reject requests to inspect plants at which they were cheating.

Options—SSI at Facilities Subject to PPM. Three options exist:

— Option 1—Do not permit SSI at facilities subject to PPM.

— Option 2—Permit the side who decides not to establish a PPM system at a facility of the other side to place that facility on the mandatory SSI list.

— Option 3—Permit right-of-refusal SSI at facilities subject to PPM, but where the monitoring side decides not to establish a PPM system.

Annex A

Paper²⁹

Washington, undated

US startup costs in the US include the following:

Perimeter costs

— for the low case: construction and installation costs for a perimeter fence and road and a warehouse for the Soviets, preparation of the site for Soviet portal equipment (cables, digging, concrete and poles) and moving sensitive facilities at the site; based on Hercules, Magna, UT example.

— for the high case: alternate estimate for some activities provided by United Technologies, Coyote, CA.

Plant manufacturing delays:

— none assumed for the low case

— costs estimated by United Technologies, Coyote, CA, assumed for the high case.

US recurring costs in the US include the following

Contractor costs for maintenance of portal support facilities.

— for the low case: costs estimated by Aerojet, Sacramento.

— for the high case: costs estimated by United Technologies, Coyote, CA.

OSIA escort costs: the cost of maintaining an OSIA Detachment at each site.

Lodging costs

— for the low case: rental costs at Magna.

— for the high case: rental costs estimated by United Technologoes, Coyote, CA.

US plant ‘inconvenience’ charges: estimated by Hercules, Magna for production delays.

US initial costs in the USSR include the following

Portal costs

— for the low case: road portal and/or rail portal, exits and installation derived from actual Votkinsk costs.

— for the high case: highest DNA cost estimation for the same portal.

Perimeter costs

— for the low case: no perimeter assumed

— for the high case: perimeter barrier, perimeter barrier integrity system, video cameras, lighting estimated from a 1987 ESD study.

X-ray system

— not assumed for low cost system.

— Cargoscan-like system assumed for high cost system.

Off-site data center

— not assumed for low cost case.

— costs for off-site center ($1.8M), installation ($1.4M) and 150% stock ($3.5M) of high failure parts assumed. Estimated costs are from a 1987 ESD study.

Logistics and spares

— for the low case: portal support only. Costs estimated from a 1987 ESD study.

— for the high case: portal and perimeter support estimated from 1987 ESD study.

US recurring costs in the USSR are based on the following

OSIA manpower

— low case: 30 inspectors at each on-site center at $223K/man-year. No off-site center.

— high case: 38 inspectors at each on-site center plus 25 operators at the off-site center, all at $223K/man-year.

O&M replenishment

— for the low case: on-site O&M only

— for the high case: on- and off-site support

Annex B

Paper Prepared in the Intelligence Community³⁰

Washington, undated

SSI

The following text addresses the contribution to monitoring that suspect site inspections (SSI) provide:

The suspect site inspection regime tabled by the U.S. could help deter the Soviets from some forms of cheating, could help resolve ambiguous situations in cases where Soviet actions were suspicious to us, and could provide us valuable information on missile-related facilities in the Soviet Union. We judge, however, that under the inspection regime as currently envisioned by the United States we would be unlikely to detect Soviet cheating because the Soviets would be able to refuse inspections of facilities that are not on the “SSI list.” (S/NF)

The Soviets are highly unlikely to cheat on START Treaty limits at sites subject to mandatory, short-notice inspection, especially if non-inspectable sites are available. Any Soviet cheating would be designed to avoid such inspections and they would make every effort to thwart any inspection that threatened to uncover a violation. On the other hand, the Intelligence Community judges that an inspection regime that guaranteed access to undeclared production facilities, that is, without a right of refusal, could help verify a START agreement in a number of ways:

— It would help deter the Soviets from cheating more effectively than an inspection regime that includes the right to refuse inspections of undeclared production facilities. If the United States had the right to inspect certain Soviet facilities without a right of refusal, it would increase the operational difficulty, risk, and cost for the Soviets to develop and maintain an effective covert missile production capability.

— It could help us uncover cheating. Even if we did not see a “smoking gun,” short-notice, on-site inspections could reveal indications of illegal activity. In particular, we judge that a guaranteed right to conduct on-site inspections of suspected missile production facilities could substantially improve our ability to detect covert production of treaty-limited solid propellant motors at those facilities. (S/NF)

PPM

The following text addresses the contribution to monitoring that permanent on-site presence (either portal perimeter monitoring or portal monitoring) provides:

The U.S. START position would allow portal perimeter monitoring at facilities that produce START-accountable solid rocket motors as large as or larger than the SS–25 first stage rocket motor. PPM at these facilities (we currently estimate that four Soviet facilities fit this category) should allow us to count precisely the number of SS–24 and SS–25 first-stage motors that are shipped from these facilities after the treaty enters into force. Because START does not provide for inspection of the interior of declared production facilities, we would be unable to determine the number stored there. (S/NF)

[4½ lines not declassified] If all Soviet facilities associated with solid rocket motors (we estimate that number to be about a dozen) were subject to inspection with no right of refusal, it would be significantly more difficult for the Soviets to produce motors covertly, although it would not prevent them from doing so. Under such a regime, the Soviets would be forced to produce motors at a facility not associated with solid rocket motor production, which would entail greater difficulties and risks than production at declared, overt facilities. (S/NF)

In order to significantly inhibit the Soviets’ ability to produce motors and missiles, however, missile final assembly facilities—which may load propellant segments into motor cases—could be included in any PPM regime. Any Soviet cheating scheme to evade portal monitoring at a final assembly plant would have to be much more elaborate than a scheme to avoid PPM at a motor production plant. In the former case, the Soviets would have to direct all stages, PBVs, canisters, and many other components to an illegal final assembly facility. Such actions would affect numerous other Soviet facilities and would increase significantly the risk of detection. (S/NF)

Ideally, if tamper proof tags were affixed to all treaty-limited motors as they left the PPMd facility, and the tags could be checked on those motors prior to testing, motors produced at covert facilities could not be tested. However, effective tagging has not been proven to be a feasible concept, although several studies are underway. There are also several potentially effective evasion scenarios to such a regime. Finally, it is difficult to assess whether the scenario described above, which focuses only on first stage motors (the only accountable motor under the current U.S. proposal), would significantly reduce Soviet confidence in performance and reliability. Some even question whether confidence in upper-stage motors would be degraded significantly if they were included in the tagging scheme. (S/NF)

Tab C

Paper Prepared by the Joint Staff³¹

Washington, undated

Short-Time-of-Flight (STOF) Threats and Arms Control (Revision 1)

Executive Summary

1. (SNF) The attached paper examines the threat posed to strategic systems (weapons; command, control, and communication (C3), and tactical warning and attack assessment (TW/AA) by short-time-of-flight (STOF) SLBMs and the impact of banning or limiting their testing in a START treaty.

2. (SNF) Several proposals were examined during SALT II negotiations to enhance stability by attempting to reduce the threat posed by STOF SLBMs. Most of the methods examined called for banning all testing of SLBMs in a depressed trajectory (DT) mode. In addition to banning depressed trajectory SLBM testing, another method of limiting the STOF threat considered was the use of SSBN stand-off zones. All approaches had problems related to circumvention scenarios and monitorability. None of the approaches were included in the SALT II agreement.

3. (SNFWN) There are two issues which must be considered in limiting STOF SLBM development. First, SLBMs fired to very short ranges can travel standard trajectories and therefore no special testing is required. A technical ban cannot circumscribe the tactical deployment of Soviet SLBMs very close to our coasts, only an SSBN exclusion zone can address this threat. For longer range firings (2000 Km or more), the principal question in limiting STOF SLBM development is whether it is necessary to demonstrate a STOF trajectory on an end-to-end test, or whether individual component testing and or computer simulation would supply adequate confidence in missile performance and reliability. The belief that Soviet missile programs require end-to-end system demonstration has been challenged and is subject to debate.

4. (SNF) Concern exists regarding the potential impact that introducing restrictions on STOF SLBM testing could have on the development of future US strategic offensive and defensive weapon systems. Banning STOF SLBM testing could introduce problems and possible constraints, such as: restrictions on the ability to modernize existing forces; restrictions on the research and development of strategic defensive systems (SDI); capture of existing and future nonstrategic systems; restrictions on ballistic missile reliability testing; and possible unacceptable inspections, intrusions and operational constraints in order to verify STOF capabilities (or the absence of such capabilities) of a missile system.

5. (SNF) With available missile technology, the existing generation of SLBMs could be adopted to fly in STOF regimes of 10–15 minutes. Significant penalties in terms of accuracy, payload, and R&D costs would be required to develop STOF SLBMs to fly in regimes less than 10 minutes. Even with a developed STOF capability it would be difficult for the Soviets to execute with confidence a coordinated attack using a combination of STOF SLBMs and ICBMs in a manner that precludes a US retaliatory strike. The Soviets would have to consider that the first nuclear detonations of a STOF SLBM attack would trigger a prompt US retaliatory launch.

6. (SNF) A ban on STOF SLBM testing has many ramifications. Regardless of the method chosen to accomplish such a ban, as in most arms control initiatives, uncertainties will exist as to the impact on future systems development and monitorability for both the US and the Soviets. Moreover, a ban on STOF SLBMs, even if monitorable, cannot preclude Soviet use of high-altitude, “burst on ascent” nuclear detonations delivered by SLBMs that could degrade US tactical warning of ballistic missile attack. The drawbacks of such a ban could be minimized by limiting the ban to future types of SLBMs tested from submarine platforms and by proper selection of the parameters (range, time of flight, re-entry angle, apogee, etc.) used to formulate the ban. If doubts persist as to whether a ban on STOF SLBM testing is necessary and/or desired, as an interim measure, a proposal similar in nature to a STOF SLBM testing ban could be pursued as a confidence building measure.

(SNF) Options available to address a STOF SLBM testing ban.

Option 1—Decide not to pursue the issue further.

Option 2—Decide to seek a STOF SLBM testing ban with the Soviets in the START negotiations.

Those who support Option 1, and thus do not support Option 2, hold that:

• There is a lack of substantive evidence that the Soviets are pursuing the development of a STOF SLBM, thus there is no potential threat other than that of existing shorter range (less than about 2000 Km) SLBMs, which we can offset by operational readiness and basing measures. There is no reason to open complex negotiations to counter a threat that does not exist.

• Difficulties in parameter selection and definition of STOF could raise unresolvable compliance issues, thus offsetting the modest benefits derived by a ban.

• With a SLBM STOF testing ban, an incentive may be created to covertly develop such a missile in order to gain a unilateral strategic advantage.

• Since all US SLBMs and some Soviet SLBMs have demonstrated time-of-flight between 10 and 15 minutes, a SLBM STOF testing ban could not eliminate the existing STOF SLBM capabilities of either side, thus the potential threat would not be eliminated, regardless of restrictions placed on the development of future STOF SLBM systems.

• Negotiating a ban could resurface the issue of SSBN standoff zones which was raised by the Soviets in SALT I and SALT II. SSBN standoff zones are merely CBMs and are not monitorable.

• A ban would not reduce the potential STOF threat from high altitude burst-on-ascent SLBMs

• While negotiating the STOF SLBM ban, sensitive US strategic system’s reaction times could be revealed to the Soviets.

• Since STOF SLBM capabilities can be developed without end-to-end testing, a ban would not be effective in eliminating the potential future threat.

• The successful execution of a STOF SLBM attack as part of a coordinated “bolt out of the blue” or preemptive strike is highly unlikely. The resulting time of arrival differences of the various systems and difficulty in achieving the intended goal make an STOF SLBM attack a high risk option. Any concessions (cost) the USG would have to make to attain Soviet agreement to such a treaty provision would not be worth the modest benefits derived.

• Monitoring uncertainties create both the opportunity for covert development and lower the possibilities to detect circumvention regardless of whether or not testing is banned.

• The Ballistic Missile Launch Notification Agreement with the Soviets, in conjunction with current ballistic missile flight test monitoring, gives the US adequate indication of Soviet developments in STOF SLBMs without a ban on testing.

Those who support Option 2, and thus do not support Option 1, hold that STOF attacks are potentially threatening and destabilizing, and that a ban on STOF SLBM testing:

• Aggressively pursues closing out a potentially destabilizing Soviet option for a weapon system that the US has no intentions or need to develop.

• Would be effective in eliminating a new development option for the Soviets because development of STOF SLBM capabilities requires end-to-end testing of any new system that is designed.

• Could serve as a new stabilizing initiative in the START negotiations.

• Maintains an initiative to increase strategic stability by reducing the potential for overt STOF SLBM development.

Option 3—Decide that restricting STOF SLBM testing should be pursued with the Soviets, but only as a confidence building measure (CBM) of voluntary constraints on STOF SLBM development and testing, separate from START, negotiated either; a) in the START forum, or b) in a separate forum.

Those who support Option 3a hold that:

• The enhancement of stability is worthwhile, but a ban on STOF SLBM testing is not a viable method of achieving it because of monitoring problems and circumvention possibilities.

• Since SLBMs are strategic offensive systems, the negotiations should be conducted in coordination with the START negotiations. The best way to accomplish this is to use the START negotiating forum.

• Over time, a CBM would provide an opportunity, without formal compliance concerns, to monitor the development of new ballistic missile technologies which could be used in STOF SLBM development.

• The parameters used in designing the CBM regime would not have to be as precise as they would in a treaty, thus the possibility of revealing sensitive US warning and reaction times to the Soviets would be less.

• Because of monitoring uncertainties and circumvention possibilities, a treaty ban on STOF SLBM testing would be little more than a CBM, so why not treat it as such.

Those who support Option 3b hold the same basic views as with Option 3a, except they would hold that:

• The quickest way to obtain evidence of a new initiative is to separate it from the details of the START negotiations and work it with a separate group of negotiators.

• A separate negotiating forum would give added status to the initiative, giving it increased visibility in the political arena.

Those who do not support Option 3 hold the same views that apply to Option 1 or Option 2 and also hold that:

• There is no viable confidence building measure that approaches a STOF SLBM testing ban therefore a CBM would be merely words with no substance.

• Practical implementation of a CBM would be as difficult to monitor as a ban, with even more room for development of a new STOF SLBM capability.

• CBMs are not effective in restricting development of future systems, thus a CBM is not an appropriate substitute for a treaty ban.

• The Ballistic Missile Launch Notification Agreement with the Soviets, in conjunction with current ballistic missile flight test monitoring, gives the US adequate indication of Soviet developments in STOF SLBMs without a new CBM.

• As a practical matter, if the negotiated CBM is not legally binding on the sides, only the US would be constrained.

[Omitted here is the body of the paper.]

Tab D

Paper³²

Washington, June 7, 1989

START 08C

NSR 14 START Review Paper Eight—New Initiatives

Purpose. To review possible new initiatives under START.

Background. The Arms Control review conducted in response to NSR 14 has reaffirmed the basic continuity between the Reagan Administration and the Bush Administration approach to START. The President has also made it clear, however, that he wishes to place his own imprint on the U.S. START proposal.

There are two aspects to such an imprint. The first is the overall vision which the President wishes to project. In this regard, the President intends to stress the value of START in reducing the risk of nuclear war and in promoting stability.

The second aspect is the degree to which new substantive initiatives should be offered to support the broad emphasis on strategic stability. Two such possible new proposals in START are a ban on MIRV mobile ICBMs and a ban on short-time-of-flight ballistic missile tests. These initiatives are discussed in other NSR–14 papers. This paper discusses other possible initiatives.

I. Reduce Fractionation. MIRVed ballistic missiles have an undeniable political reputation as destabilizing. This option attempts to address that perception by placing direct constraints on fractionation in order to encourage deMIRVing. This approach is based on the view that fractionation is destabilizing in and of itself, regardless of whether it occurs on survivable systems or vulnerable systems. Fractionation also increases the breakout potential associated with mobile refires or covert mobile launchers/missiles (this aspect of fractionation limits is discussed in the paper on mobile ICBMs). Since both the U.S. and Soviet Union are improving the hard-target capability of their SLBMs and improving the survivability of their ICBMs, in the future there will be fewer differences between ICBM and SLBM characteristics. Thus, in this view, the case is weakened for distinguishing in START between MIRVed ICBMs and SLBMs, or between MIRVed mobile and fixed-silo ICBMs.

Another view is that fractionation in and of itself is not destabilizing. The most highly fractionated missile that now exists, the POSEIDON, has the least counterforce capability of any U.S. ballistic missile. Moreover, highly fractionated missiles do not represent lucrative targets if they are survivably based. [2½ lines not declassified] Finally, SLBMs never will have alert rates and, some believe, reliable communications comparable to ICBMs.

Two alternatives have been suggested:

Alternative 1—MIRV sublimit. This alternative would place a direct limit the number of warheads on MIRVed ballistic missiles. The limit would be set initially so that the difference between it and 4900 was less than or equal to the number of single-warhead missiles (Minuteman IIs, Small ICBM, Minuteman IVs, etc.) we need to protect (e.g. 4200 if we decide to deploy 250 or more Small ICBM missiles and 450 Minuteman IIs/IVs). NIE 11-3/8-88 projects a Soviet START force in 1998 that includes 4374 MIRVed ballistic missile warheads.

However, uncertainties in the future forces on both sides may disadvantage the U.S. under this alternative:

— The Soviets could deploy more SS-25s than the 360 projected in the NIE, perhaps 500–700. [2 lines not declassified]

— The U.S. Small ICBM, if funded on the present schedule, will not be fully deployed until well after the START reductions period. Moreover, at present there is no Minuteman IV program. If a Minuteman IV is funded, a program of more than 450 would require a START provision to permit downloading, which would increase Soviet breakout potential. There are no plans for a U.S. single-warhead SLBM.

These considerations mean that the Soviets could be in a position to press for a MIRV warhead limit lower, possibly several hundred warheads lower, than the U.S. realistically can take advantage of at the end of the reduction period. If, for example, we accept a limit lower than 4450, we will either have to permit downloading in START or deploy fewer than 4900 ballistic missile warheads until the Small ICBM is fully deployed, now planned for the mid-2000s.

An additional disadvantage of this alternative is that it does not reward reductions in fractionation short of converting to single-RV missiles (this could be mitigated by exempting less fractionated missiles, e.g. those with 2 or 3 RVs, from the sublimit).

Alternative 2—Warhead-to-Missile Ratio for Ballistic Missiles. In this alternative we would introduce a new unit of account into START: warhead-to-missile ratio for ballistic missiles. While such a new unit of account might be more difficult to negotiate than Alternative 1, it would have the advantage of rewarding any lowering of missile fractionation and not just the replacement of MIRVed with single-RV missiles.

With the new projections for SS–25 deployments of perhaps 500–700, the Soviet START force would have about 4.3 warheads per ballistic missile, while a U.S. force of 50 Peacekeeper, 300 Small ICBM, 21 Trident (with 3 subs in overhaul) has about 4.1 warheads per ballistic missile. Thus, it appears that a U.S. proposal could be developed that would accommodate planned U.S. programs and encourage modest deMIRVing of Soviet forces.

However, as in Alternative 1 above, the Soviets have several options to lower their ratio significantly. If, for example, the Soviets took advantage of the START submarine overhaul and ICBM launcher construction exclusions, they could lower their ratio to about 3.6–3.8 with only minor force adjustments.

The U.S. ratio also depends significantly on specific force assumptions, ranging from about 3.6, if Trident is curtailed at 19 ships and 750 SICBM/MM II, IV are deployed, to 4.8, if 23 Tridents and 450 single warhead ICBMs are deployed. If our plans change and 100 Peacekeeper are deployed, the ratio will rise to 4.2–5.0. U.S. and Soviet force options are discussed further in the annex.

In either of the foregoing alternatives the new limit would replace the U.S. proposed 3000–3300 sublimit on ICBM warheads. Some believe this sublimit would not likely constrain the Soviets and would be even less important in the context of a limit on MIRV warheads. Others believe that neither of these alternatives would substitute for the 3,000–3,300 limit, since single-warhead ICBMs could be very effective in a first strike because of their great potential counterforce and EMP effects.

To provide a mechanism for gradual deMIRVing in the longer term, a provision could be included explicitly providing for periodic review of this limit by the sides to consider lowering it; this could be coupled with consideration of amendments to the 1600 SNDV limit.

Possible advantages of such a proposal are:

— It would encourage deMIRVing and, some believe, promote stability.

— By treating SLBMs and ICBMs equally it would remove the Soviet objection to our ICBM sublimit proposal and thus might prove negotiable.

— It would be attractive to the public since it places even tighter limits on MIRVed systems than did the Reagan Administration.

— By selecting suitable values for the new limits, the strategic modernization program could be protected.

— It offers a mechanism for future deMIRVing without the drawbacks of agreeing to follow-on negotiations.

Possible disadvantages of such a proposal include:

— By focusing only on fractionation, it ignores what some believe is a greater threat to stability from vulnerable, highly fractionated systems than from survivable highly fractionated systems.

— deMIRVing beyond the initial limits will require radical and expensive alterations in U.S. plans, including additional single warhead ICBMs and/or less highly fractionated SLBMs and/or submarines with fewer launchers. (Because of the high cost per warhead of deploying deMIRVed SLBMs, this may mean a net move away from SLBMs, which the U.S. has consistently claimed is stabilizing.)

— Realistic U.S. deMIRVing options may require downloading, which poses verification and breakout problems.

— The Soviets have realistic deMIRVing options that would permit them to outbid the U.S. in negotiating a specific limit.

— A sublimit of 4200 MIRVed warheads will not be perceived publicly as a serious deMIRVing proposal; a lower sublimit would require unrealistically expensive restructuring of U.S. forces.

— There may be political drawbacks in softening past U.S. positions on the importance of ICBMs.

— A MIRV warhead sublimit will appear to some to be a recreation of the SALT II sublimit structure, with attendant adverse political consequences.

II. Reduce instability by reducing vulnerability. The United States might seek dramatic reductions in silo based missiles, especially highly fractionated ones. This approach is based on the view that fractionation is not destabilizing except when it is associated with vulnerable systems. In this view there is no practical way to make silo-based ICBMs survivable. Such an approach assumes that mobile ICBMs will be allowed.

Two alternative approaches exist:

Alternative 1—Silo-based MIRV ICBM ban. In this alternative we would ban all silo-based ICBMs except single-warhead missiles. Single warhead missiles do not pose stability problems to the same extent as MIRVed ICBMs. Under this alternative, Soviet SS-17, 18, 19 and 24 (silo-based) missiles and U.S. Minuteman III and Peacekeeper (silo-based) would be eliminated. To protect the U.S. investment in Minuteman III (and to make the proposal more palatable to the Soviets) we might consider allowing downloading of existing MIRV missiles to single warhead missiles. As noted above, downloading increases breakout potential; as envisioned in this alternative, the breakout risk is especially high from downloaded, highly MIRVed, Soviet ICBMs. Finally, the 1600 delivery vehicle limit could affect our ability to deploy 4900 ballistic missile warheads.

Alternative 2—Total silo-based ICBM ban. In this alternative we would ban all silo-based ICBMs. This alternative implies a dramatic restructuring of forces—about 60% of Soviet warheads are on silo-based ICBMs. The phase in period might have to be much longer than the seven years we currently envision under START.

Potential advantages of such a proposal include:

— It would dramatically improve stability.

— Because of the greater emphasis on ICBMs in the Soviet force structure, it would impact the Soviets more than the United States.

— It could be seen as a dramatic gesture and rebound to the United States credit in the public diplomacy arena.

— U.S. strategic modernization would be protected.

— Soviet heavy ICBMs would be eliminated.

Potential disadvantages include:

— Such a ban is almost certainly non-negotiable. If the Soviets will not accept a 3300 ICBM sublimit (which does not constrain them), they will not accept something this radical.

— The attempt to introduce such a radical proposal could disrupt negotiations, as did similar attempts in the Carter Administration, and delay achievement of a START treaty.

— If downloading is not allowed, the United States must give up the entire Minuteman III force.

— If downloading is allowed, the Soviets will have a large breakout potential.

— Total elimination of silo-based ICBMs could leave the United States without the flexibility offered by a single-RV missile until the turn of the century.

— Total elimination of silo-based ICBMs ignores the important role of silo-based ICBMs in complicating attack timing problems against bombers and in attacking time-urgent hard targets.

III. Additional approaches.

Three additional approaches have been suggested and are attached:

— Warhead sublimit on silo-based MIRVed ICBMs.

— Immediate portal monitoring of mobile ICBM final assembly facilities.

— Ban on future MIRVed ICBMs.

Agencies have not reviewed these approaches.

Forum for raising any new initiatives. If the United States elects to present a new initiative, there is a question of whether to present it, at least in general terms, at the opening of the Nuclear and Space Talks or to defer until the next meeting of Foreign Ministers. During the Moscow Ministerial, the Soviets proposed regular (3 to 4 times annually) Foreign Ministers’ meetings. The Soviets indicated that they viewed such meetings as be the appropriate forum for presenting new initiatives or significant new positions.

On the other hand (a) the United States has long sought to avoid bypassing or downgrading the Geneva negotiations and (b) an initiative not presented until fall loses the public diplomacy benefits of being seen as a new Bush Administration proposal growing out of a thorough review of U.S. policy.

Annex

Paper³³

Washington, undated

NSR-14 START Review Paper Eight—New Initiatives

Additional alternative for II. Reduce instability by reducing vulnerability.

Alternative 3—Warhead sublimit on silo-based MIRVed ICBMs.

In this alternative, we would impose a warhead sublimit of _______³⁴on silo-based MIRVed ICBMs, which would supersede the 3000–3300 ICBM warhead sublimit. The background discussion for Alternative 1, the ban on silo-based MIRVed ICBMs applies here in part; however, a sublimit would have less force structure impact on the U.S. and would allow us to retain a limited MIRVed ICBM force in silos. It would also hedge against a failure to deploy Peacekeeper in a rail-mobile mode. It would also be more likely to be negotiable than a ban, at least if Soviet mobile ICBMs were permitted.

Attachment

Paper³⁵

Washington, undated

New Initiative: Immediate Portal Monitoring of Mobile Missile Final Assembly Facilities

Proposal. To seek Soviet agreement to on-site portal monitoring of the SS–24 mobile missile assembly facility at Pavlograd as soon as possible, as a means of increasing confidence in estimates of non-deployed Soviet mobile ICBMs and to minimize Soviet ability to under declare their mobile ICBM inventory once START is in force. This would augment INF Treaty portal monitoring at Votkinsk, where the SS–25 missile is assembled.

Background. Mobile ICBMs pose many of the major START verification concerns. [3 lines not declassified] Non-deployed missiles could be covertly deployed to augment the legal mobile ICBM force or could be refired from surviving mobile launchers. While current START provisions are intended to diminish this threat, our uncertainties increase as unmonitored missile production continues.

Pre-START Portal Monitoring. [2½ lines not declassified] Establishing PM at the Pavlograd SS–24 final assembly facility, which began series production only about two years ago, would significantly reduce the growing uncertainty in estimates of non-deployed SS–24s—particularly while START is being negotiated. If MIRVed mobile missiles were banned in START, early PM would increase confidence that all SS–24s that had been produced were declared and then eliminated.

Missiles vs First Stages. Although the current US START position centers on the first stage as the treaty-accountable part of a missile (primarily to accommodate a tagging scheme that has yet to be proven effective), there are several important reasons to establish PM now at final assembly facilities instead of motor production plants:

— PM has already been established at the SS–25 assembly facility; capturing the SS–24 would require PM at only one other facility (and, for reciprocity, at one other US facility).

— If such monitoring focused on motor plants, at least two, and probably three, would have to be monitored (and an equivalent number of US plants).

— While the Soviets could divert production to alternate plants to avoid the monitored portal in either case, evading PM of final assembly facilities would have to be much more elaborate. The Soviets would have to divert all stages, post-boost vehicles, canisters, and many other components to the alternate assembly facility, affecting numerous other plants and increasing significantly the risk of detection. With PM only at first-stage motor plants, a considerably smaller number of components and suppliers would be affected.

No US “Final Assembly Facilities”. While the Soviets integrate their missiles at final assembly facilities, the US ships the individual stages to deployment sites for assembly. Thus, there are no direct US equivalents for PM when the Soviets ask for reciprocal PM rights. The precedent was set in INF to monitor a US motor/stage plant in exchange for PM at the Votkinsk final assembly facility. The same precedent could be applied, exchanging a US motor plant for the SS–24 facility at Pavlograd.

Sooner is Better. Most of the major US verification concerns with non-deployed mobile missiles are directly related to the number produced. The quick establishment of PM would reduce uncertainties in missile production estimates and improve confidence in the overall mobile missile verification regime. The Soviets could be approached on the margins of START for such an agreement, through diplomatic channels, or at the September ministerial.

Attachment

Paper³⁶

Washington, June 7, 1989

INSERT FOR NEW INITIATIVES PAPER

Option: Ban future MIRVed ICBMs. This option would stop MIRVed ICBMs at the Soviet SS–24 and the US MX, while allowing future single-RV missiles. It could take the form of a ban as of a negotiated date on production of any rocket motor for a MIRVed ICBM, thus precluding both new types and modification of existing types (except for backfits of already-produced missiles). The production ban would be monitored by PPM, as provided for in our current proposal.

The production ban could be supplemented by a ban on releasing or simulating the release of more than one RV during flight testing of any single-RV ICBM. We also might want to add volume, launch weight or throwweight constraints on future single-RV ICBMs to constrain breakout through uploading of additional RVs. Finally, we might want to extend the production ban to preclude recasting of solid propellant motors, which would force the solid MIRVed ICBM forces of the sides to atrophy over the long term. (Our ability to verify such a provision could be questionable, however.) This could be accompanied by a provision requiring the phaseout of liquid MIRVed ICBMs over a similar extended period.

If we wished to constrain modernization further, the production ban could be supplemented with a provision limiting flight tests of existing MIRVed ICBMs to a number high enough for reliability testing but too low to develop a new mod (at least at a normal rate of development). Consideration might also be given to additional means of constraining improvement of existing types (e.g. no change in weight of RVs tested), but such means could introduce verification difficulties and possible limits on US flexibility.

Arguments for this option are as follows:

— It focuses on those MIRVed missiles, ICBMs, which are considered by most to be destabilizing, while excluding SLBMs, for which MIRVing is considered by many not to be destabilizing.

— By supplementing START’s numerical limits with a significant constraint on modernization, it would appeal to the public as a genuine step to limit the “arms race.”

— It would allow the US to retain MX, and the date of commencement of the production ban could be set so as to allow completion of MX production. It also would allow us to deploy a single-RV Midgetman, as well as any other single-RV ICBM.

— A ban on further production of any MIRVed ICBMs would avoid the verification problems of modernization limits that differentiate between new types and mods.

Arguments against the option are as follows:

— By including ICBMs, but not SLBMs, it might be resisted by the Soviets in the same way our ICBM warhead sublimit was. The Soviets might also resist because it would preclude plans to produce an SS–24 follow-on, as well as any future mods of the SS–18.

— If the Soviets were to resist this proposal strongly, it could generate criticism of the US as raising new obstacles to an agreement and could require an embarrassing retreat.

— It might be circumvented through development of a common missile in the guise of an SLBM, which could be deployed rapidly on land in a breakout, or through extensive backfitting of existing missiles. The former threat could be mitigated with limits on nondeployed missiles for new SLBMs, on soft launch capability, and on empty silos.

— Our ability to ensure the Soviets were not covertly producing rocket motors using cartridge loading of propellant might be questionable.

1. Source: George H.W. Bush Library, Bush Presidential Records, National Security Council, H-Files, NSC/DC Meetings Files, OA/ID 90010–012, NSC/DC 033—June 9, 1989—NSC/DC Meeting on NSR-14 re: U.S. START Position on Mobile ICBMs, Mobile ICBM Verification and New Initiatives. Secret. Sent for action. Sent through Kanter.
2. See Document 28.
3. Gates underlined “seek consensus at the outset that they fairly represent agency views and options.”
4. Gates drew a short vertical line in the left-hand margin beside this sentence, and wrote: “adequate vehicles for discussion/agency position,” beside this and the previous sentence.
5. Gates drew brackets in the left-hand margin beside the section on whether to ban mobile missiles.
6. Gates drew brackets in the left-hand margin beside the section on verification of mobile missiles.
7. Gates drew a short vertical line in the left-hand margin beside these sentence clauses and wrote a checkmark beside it.
8. June 8.
9. Gates drew a short vertical line in the left-hand margin beside this sentence and wrote a checkmark beside it.
10. Gates underlined the latter portion of this sentence beginning with “MIRV warhead sublimit,” drew a vertical line in the left-hand margin beside it, and wrote: “MIRVed mobile ban.”
11. Gates drew a bracket in the left-hand margin atop this section and circled “Three categories are:” In the left-hand margin he wrote: “Are these the right problems? Any others at that strategic level?”
12. Gates underlined “The ban on short time-of-flight missiles has been suggested” and “strategic exercises” and wrote the number 1 and circled it in the right-hand margin beside this section.
13. Gates underlined “a production ban,” wrote the number 2 and circled it in the right-hand margin beside this section, and drew a vertical line in the left-hand margin beside this section.
14. Gates underlined “ban on MIRV mobiles” and “or low mobile warhead limits,” wrote the number 3 and circled it in the right-hand margin beside this section, and drew a short vertical line in the left-hand margin beside this sentence.
15. Gates underlined “An additional measure, to install now INF-like portal monitoring systems at mobile final assembly facilities, would address a problem all agencies agree exists.” In the left-hand margin beside this sentence he drew a vertical line and an arrow pointing to the vertical line. He also underlined “SS-24s.”
16. Gates did not indicate his preference.
17. Secret.
18. Secret.
19. Secret.
20. SICBM costs are based on an IOC/FOC that is at least two years earlier than is shown. No cost estimates reflecting the later schedule exist yet, although costs will increase somewhat. [Footnote is in the original.]
21. Footnote in the original directs to the same text as footnote 22, above.
22. This assumes we pursue rail-garrison on the current schedule and a ban comes into force soon after FY–94, the year procurement is completed. (A ban one year earlier would still cost $5.3 billion.) Deactivation costs are not included. [Footnote is in the original.]
23. Footnote in the original directs to the same text as footnote 20, above.
24. Footnote in the original directs to the same text as footnote 22, above.
25. Footnote in the original directs to the same text as footnote 22, above.
26. Footnote in the original directs to the same text as footnote 22, above.
27. Secret.
28. OSD believes that the costs for PPM could vary widely, depending on assumptions and equipment requirements for the PPM system; the estimates below reflect two possible approaches, but have not been reviewed by the Services or OMB. The Joint Staff believes that these estimates are conservative; e.g., they do not include military salaries, R&D costs for Testbed On-Site Inspection, transportation costs, and FBI costs. [Footnote is in the original.]
29. Secret.
30. Secret; Noforn.
31. Secret; Noforn; Wnintel.
32. Secret.
33. Secret.
34. Omission is in the original.
35. Secret; Noforn.
36. Secret.