Foreign Relations of the United States, 1981–1988, Volume XI, START I

318. Memorandum From the President’s Assistant for National Security Affairs (Powell) to President Reagan¹

Washington, August 5, 1988

SUBJECT

• Weekend Reading

Next Friday, August 12, I anticipate holding an NSPG² to review two START issues. The first, verification of limits on mobile ICBMs, will provide an opportunity for some of your advisors to share their reservations with the position we established just prior to the Moscow summit. I am working to see if a consensus can be forged on this issue and will provide you information on it subsequently.

The second issue, non-deployed missiles, is one we have not discussed with you in some time. Our current START position places significant constraints on all types of non-deployed ballistic missiles. Reducing these constraints could reduce the cost to us of our verification regime as well as decreasing the intrusiveness of Soviet inspectors in the United States. We have therefore been reexamining our position to see if our objectives can be met with a less elaborate verification approach.

The attached paper, prepared by the Arms Control Support Group, has served as a vehicle for our reexamination. While agencies disagree on what should be done, all but the Office of the Secretary of Defense agree with the basic analysis. While Secretary Carlucci has not personally addressed the issue, his staff (which favors retaining all our current constraints and adding new ones) disagrees with the Intelligence Community assessment and believes the threat from non-deployed missiles to be more severe than indicated.

I recommend you include the attached paper in your weekend reading. I will provide additional details before the NSPG.

Attachment

Paper Prepared by the Arms Control Support Group³

Washington, August 5, 1988

GRIP 58G (U)

Purpose. To review the U.S. position on limits on nondeployed missiles and on the associated verification regime. (S/G)

Background. “Nondeployed” missiles (NDM) are those not in silos, on mobile launchers or loaded in submarines, and thus not counted in the basic START limits. Some nondeployed missiles are necessary to allow for routine maintenance, for spares, and for periodic reliability testing. Nondeployed missiles beyond those necessary for such legitimate purposes could add to military capability because of their potential for breakout and refire. Because of this potential, the United States seeks a direct numerical limit on nondeployed missiles in START. (S)

Ideally we would like to keep the number of nondeployed missiles small to minimize the potential risks. To keep acquisition costs down, however, U.S. practice is to produce all the missiles of any given type we need at one time and store the excess missiles until they are required for testing or maintenance. This practice means that we tend to have a large number of nondeployed missiles on hand ourselves and thus would have to allow the Soviets a similar large number. [less than 3 lines not declassified] The large number of nondeployed missiles required for U.S. forces may, therefore, limit the impact of NDM limits in START. (S/G)

Many of our most complex treaty provisions—tagging, perimeter/portal monitoring, suspect site inspections—exist primarily to help enforce restrictions on nondeployed missiles. Because of the financial costs and the operational impact on U.S. forces, it is appropriate to consider whether U.S. goals could be achieved by constraining some lesser subset of nondeployed missiles. (S/G)

In broad terms there are three types of threat posed by a large number of nondeployed missiles (or a small number of highly MIRVd missiles):

—Breakout Via Soft Launchers. The Soviets might, in time of great tension, attempt to rapidly erect their nondeployed missiles on soft launchers, drastically increasing the threat facing the United States in a relatively brief period of time. The Soviets could thus gain the substantial coercive leverage inherent in a situation in which they suddenly had a substantially larger strategic force than the United States. [2 lines not declassified] although we have seen no evidence that such launchers or the large support structure implied by such a scenario exist or that Soviet missiles are capable of being rapidly erected and fired from such soft launchers.

—Breakout Via Covert Mobile Missile Launchers. The Soviets might gain coercive leverage from bringing into full operational readiness mobile ICBMs and launchers which were covertly maintained in a reduced state of readiness or in storage. Such missiles and launchers could consist of those produced before the treaty enters into force but not declared, and/or those covertly produced after the treaty enters into force.

—Reload and Refire. The Soviets could use a large force of nondeployed missiles to reload and refire existing launchers. There is clear evidence of a Soviet reload and refire capability. For this option to be effective the Soviets would need launchers; refire missiles; and handling equipment, transportation equipment and trained personnel, all of which survive the initial exchange. If the Soviets believed that they had all of these, their correlation of forces calculations in a crisis might embolden them to take risks they would otherwise not take, thereby weakening deterrence. (S/G)

Soviet Plans—General. The Intelligence Community judges that the Soviets would attempt to reload and refire some silo-based ICBMs and SLBMs and to restore as much as possible after nuclear strikes, the combat effectiveness of their ICBM and submarine forces. It is judged also that the Soviets probably would be able to reload only a small portion of their silos in the final phases of general nuclear war. Further the Soviets currently have only a limited capability to carry out SLBM reload and refire operations in a post strike nuclear environment. The range of views within the Intelligence Community extends from those who believe that the Soviets would be able to reload and refire from a significant number of surviving silo-based launchers to those who believe that Soviet capabilities in this regard are negligible. All agencies except OSD agree that the Soviets currently do not have the capability or intent to engage in any of the following refire-related activities: large-scale reload, maintain reserves of older retired missile systems, and soft launch of silo-based ICBMs. All agencies agree also that it would be much easier to reload and refire mobile ICBMs [1 line not declassified] (S/G)

Soviet Plans—Fixed Site Refire. The feasibility of fixed-site refire operations depends on the availability and survival of missile reserves, transportation assets, support equipment, largely intact silos, and personnel; on the maintenance of an operable logistic support infrastructure, especially relevant portions of the railway system; on a well-planned set of operational procedures; and on the timing and staging of the reload. Successful implementation would be heavily influenced by damage sustained before the reload operations. Effective execution of U.S. SIOP plans would laydown weapons on all Soviet silos making survival of a large number of silos unlikely. Support facilities at fixed operational complexes are not hardened and storage apparently is limited, moreover, rear depots that hold reserve missiles and propellant are neither hardened nor well-suited for rapid evacuation of missile airframes. In addition to the difficulties associated with recovery from the effects of nuclear strikes, restoration of unit combat effectiveness and reloading launchers with missiles, nuclear effects and potential damage could cause significant uncertainties for the Soviets about the operational effectiveness of their withheld weapons (although these factors might not dissuade the Soviets from dispersing missiles and support equipment and attempting reload operations). (S/G)

Soviet Plans—Soft Launch. In theory, probably any Soviet ICBM could be launched without a silo by using special equipment to support the missile before liftoff. [8 lines not declassified] The number of personnel, the new and unique equipment, and the training needed for a large soft launch program would be substantial and it is unlikely that such preparations would escape U.S. attention. (The dual-based SS–24 is considered a mobile missile in the context of this paper.) (S/G)

Soviet Plans—SLBM Refire. We judge that Soviet capability to reload SSBNs is limited. Soviet submarine tenders carry reload missiles and would disperse to relocation sites where they could limit reloading. Any SLBM reload operation beyond the initial capacity of a submarine tender, however, would face the difficulties of having to pre-position reloading equipment, deliver missiles and their warheads, and conduct the reload operation either under the threat of direct attack or under collateral nuclear effects. Moreover, if Soviet SSBN force generation had been disrupted by hostilities, all available missile-loading capabilities would most likely be devoted first to arming those boats that dispersed without a full complement of SLBMs. (S/G)

Problems with our Current Position. Our current position places a numerical limit (as yet undefined) on all nondeployed missiles. While this goes the farthest to constrain Soviet capability, it has certain problems. Some, though not all, of the problems we are facing in developing details of our START inspection protocol are unique to SLBMs. Allowing on-site inspection of flight testing, for example, is made significantly more difficult because of our unwillingness to have Soviet observers on U.S. submarines for at sea launches. Similarly, some provisions for tag reading, for tagging, and for on-site inspection are made more complicated by the unique nature of submarine-launched ballistic missiles and by our unwillingness to allow the Soviets access to our submarines. While there are operational difficulties with applying our tagging regime to ICBMs, the ICBM problems and the SLBM problems differ, forcing us to overcome two separate sets of problems in designing a verification regime. [less than 3 lines not declassified] (S/G)

In addition, the magnitude of our requirements for Perimeter Portal Monitoring could be reduced by altering our current position. Our current position involves PPM at the following Soviet facilities:

—facilities producing the accountable stage of solid propellant ballistic missiles [less than 1 line not declassified];

—facilities producing the accountable stage of liquid propellant ballistic missiles [less than 1 line not declassified]; and

—final assembly facilities for liquid propellant ballistic missiles [less than 1 line not declassified].

If mobile ICBMs are allowed, the U.S. would expand PPM to include, at a minimum, facilities producing solid rocket motor stages as large as the smallest accountable stage of a treaty-limited mobile ICBM; some agencies would extend PPM to facilities producing second and third stages of ballistic missiles. A PPM program of this scope is costly. (S/G)

As a result of the foregoing, it is appropriate to consider whether U.S. objectives can be met by a regime which limits only some nondeployed missiles. For purposes of this paper, non-deployed missiles will be considered in four classes: mobile ICBMs, solid propellant fixed ICBMs, liquid propellant fixed ICBMs, and SLBMs. (S/G)

Discussion—Mobile ICBMS. Nondeployed missiles capable of being launched by mobile ICBM launchers are obviously the highest priority for constraint. Our decision to ban mobile ICBMs is based on verification and stability concerns, with the stability concerns related to both the problem of reloads and refires and the problem of covert launchers. Since mobile ICBM launchers are expected to survive an initial exchange, and since handling equipment for mobile ICBMs is portable (and thus also survivable), a large number of reload missiles gives the Soviets all three required elements for an effective restrike capability. In addition, since mobile launchers could be produced with relative ease, either covertly or (if unconstrained) legally, a large number of nondeployed missiles provides significant breakout potential. Constraints on nondeployed missiles capable of being fired by mobile ICBM launchers are thus essential. (S/G)

Discussion—Solid Fixed ICBMs. The reload problem is less significant for fixed ICBMs, although such reloads appear to figure in Soviet war plans and training. Even if a substantial number of reload missiles and their handling equipment survive, most silos are not expected to if our strategic deterrent functions as expected. By constraining rapid reload (to the extent this is possible via START), we might provide sufficient time for those silos not destroyed in the initial ballistic missile exchange to be destroyed by penetrating aircraft. Thus the reload threat from silo-based ICBMs is considerably less, assuming existing targetting practices are retained in the future. Theoretically, the Soviets could choose to create and develop the large infrastructure necessary to soft-launch solid silo-based ICBMs, [less than 1 line not declassified] (S/G)

Both the United States and the Soviet Union are developing solid ICBMs capable of launching from both mobile launchers and silos (Peacekeeper and SS–24). Thus, as a practical matter, if non-deployed missile constraints are to limit the mobile ICBM reload and refire threat, fixed, solid propellant ICBMs must be limited as well. There are only a relative handful of Soviet solid propellant ICBMs which do not have a mobile variant (60 aging SS–13s); if mobile ICBMs are allowed, the significance of constraints on this small category is limited. Even if the United States were to retain—and be successful in obtaining—its current proposal to ban mobile ICBMs, however, some believe nondeployed missile constraints on solid, fixed ICBMs, would be appropriate either because of the potential for soft launch or because they doubt that targetting and reload constraints can preclude significant reloading of silos. (S/G)

Discussion—Liquid Propellant ICBMs. Currently liquid fueled ICBMs are deployed only in silo launchers. Since our primary defense against reload of silos is silo destruction, there is less incentive to constrain liquid fueled ICBM reloads. Further, the Intelligence Community believes that liquid fuel ICBMs are inherently less suited than solid fueled ICBMs for breakout through rapid proliferation of soft launchers. [less than 3 lines not declassified] Liquid ICBM systems are hypergolic and therefore are not transportable in a fueled state. Soft site launch of such missiles would require the creation of highly trained units and large stocks of specialty equipment. Finally, the Soviets currently have no mobile launchers capable of launching liquid fuel ICBMs; if mobile ICBMs are allowed, the U.S. position is to ban mobile liquid ICBMs. These factors argue that it is less important to constrain nondeployed liquid fueled ICBMs. On the other hand, lessening the constraints on such missiles could allow the Soviets to maintain sanctioned reloads for the SS–18 or to develop other liquid-fuel ICBM circumvention options. Reduction in the SS–18 threat has long been a primary goal of U.S. arms reduction policy. Further, even if liquid fueled ICBMs are less suited for use on soft launchers (which is not certain), such use is clearly possible. Thus while a case exists for constraints on liquid fueled nondeployed ICBMs, the incentives for such limitations are less than for solid ICBMs. (S/G)

Discussion—SLBMs. The need for constraints on nondeployed SLBMs is no more compelling (and may be less compelling) than that associated with liquid missiles or soft launches of solid fueled ICBMs. While SLBM launchers will survive an initial exchange, and substantial nondeployed missiles might survive, handling equipment necessary for reloading SLBMs currently exists only in a limited number of port facilities, all of them targetable. Thus even if a substantial number of Soviet SLBM reloads existed, they might be militarily unusable. The existence of mobile (and thus potentially survivable) Soviet submarine tenders does not change this situation. Any limit on nondeployed missiles currently being considered will be high enough to allow loadout of all existing tenders. Reloading of those tenders, like direct reloading of submarines, depends on survivability of handling equipment ashore. (S/G)

Current SLBMs, because of their generally lesser range, are less suitable for launch against the United States from soft launchers. (Range deficiencies could, of course, easily be remedied by downloading RVs or by future designs, and SLBMs could be soft launched against theater targets, freeing ICBMs for intercontinental missions.) [2 lines not declassified] If a Soviet military establishment, dominated by a land oriented view of warfare, elects to take the risk inherent in a breakout scenario, it is more likely to select a weapon system it is comfortable with and relies upon normally. These factors all suggest the case for constraints on nondeployed SLBMs is less powerful than the case for constraints on other nondeployed ballistic missiles. (S/G)

Current Position. Our current position on the table in Geneva is to seek constraints on all nondeployed ballistic missiles, accepting the burden of the verification regime associated therewith. This position has been reviewed by the Arms Control Support Group. All agencies agree that the following should apply to all non-deployed missiles:

—a data exchange on the numbers of such missiles;

—restrictions on the locations where such missiles can be stored;

—baseline inspections of the storage locations; and

—short-notice OSI of the storage locations.

In addition, agencies agree that all non-deployed missiles should be the object of suspect site inspections once the U.S. suspect site inspection regime has been determined. (S/G)

In addition to these “baseline constraints”, agencies considered the following:

—Baseline Constraints Only.

—Declared Numerical Limits.

—Perimeter Portal Monitoring.

—Tagging.

All agencies agree that the entire panoply of proposed constraints should apply to nondeployed solid ICBMs of a type deployed in a mobile mode (assuming (a) the United States alters its position and allows mobile ICBMs, and (b) an effective tagging system can be demonstrated). Based on the current U.S. position, if mobile ICBMs are allowed, the U.S. would require PPM, at a minimum, [less than 1 line not declassified]. (Note that the conversation of SLBMs into mobile ICBMs and the deployment of liquid propellant mobile ICBMs will be banned.) (S/G)

Issues. Beyond the positions indicated above, agencies disagree on what constraints to apply to other types of missiles. Agency views are as follows:

—Option 1—No additional constraints beyond baseline. This option is favored by State and the JCS and would involve PPM at the five Soviet facilities noted above.

—Option 2—Constraints on all nondeployed missiles (current position). This option would involve numerical limits, tagging and PPM. In addition to [less than 1 line not declassified] noted above, [less than 1 line not declassified] final assembly facilities for liquid propellant ballistic missiles would be involved. No facilities producing the accountable stage of solid propellant ballistic missiles which are not mobile ICBMs would be involved since no such facilities exist.

This position is favored by OSD, ACDA and Ambassador Rowny. DOE would favor applying constraints to the accountable stage of solid propellant ballistic missiles which are not mobile ICBMs if such facilities were to exist in the future; otherwise DOE favors Option 1.

Additional Possible Modifications to the U.S. Position. If we were to alter our position on nondeployed missile constraints, and depending on the residual constraints applied, we could consider additional modifications to our position. Some of these should be considered regardless of our ultimate position on non-deployed missiles. Possibilities include:

—For every launcher destroyed in reducing to the new limits, require the destruction of one missile associated with that type of launcher. Absent a limit on nondeployed missiles, there is no legal requirement for the Soviets to destroy any missiles rather than just placing them in nondeployed status. (Requires review for impact on U.S. options for a mobile Minuteman alternative to the small ICBM.)

—Ban soft launchers. This issue is already being considered in the interagency process. Because it is difficult to verify, it would provide only a modest constraint on the soft launcher breakout threat.

—Cap the number of ballistic missile submarine tenders. If constraints on nondeployed SLBMs are relaxed, the Soviets could store a large number of reload missiles on survivable tenders. They would thus circumvent a U.S. strategy which depended on our ability to target fixed reload facilities as a means of limiting SLBM reload and refire.

—Establish separate sublimits on the number of non-deployed ballistic missiles of most concern. Such sublimits could involve limits on non-deployed ballistic missiles for mobile ICBMs, limits on non-deployed ballistic missiles of a certain throwweight or limits on non-deployed ballistic missiles of a type first deployed after a given date.

—Ban reloads for heavy ICBMs. (S/G)

1. Source: Reagan Library, C Files, 8850150–8850174. Secret. Copied to Bush and Duberstein. Prepared by Brooks. Printed from an uninitialed copy. Brooks and Linhard sent the memorandum to Powell for signature under cover of an August 5 memorandum. (Ibid.) (S)
2. See Document 320.
3. Secret; Grip.