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

291. Letter from Director of Central Intelligence Webster to the President’s Assistant for National Security Affairs (Powell)¹

Washington, April 13, 1988

Dear Lieutenant General Powell:

In a recent discussion² regarding the Intelligence Community’s ability to monitor the elements of a potential START treaty, Secretary Shultz urged me to provide any ideas which would increase our monitoring confidence levels. The enclosed paper is a discussion of National Technical Means, cooperative measures, and treaty language proposals which, if implemented, would provide the highest confidence levels for START monitoring, Most of these ideas have been proposed and debated at the working level. [portion marking not declassified]

I believe that it would helpful for senior policy officials to have their review, a consolidated summary of proposed START monitoring enhancements. I fully recognize that you must make the trade-offs between enhanced monitorability and the impact of these proposals on U.S. policies and programs. I would value any comments you might have on the attached paper. My representatives in the arms control policy process will continue to assist in these difficult deliberations. [portion marking not declassified]

Sincerely yours,

William H. Webster³

Attachment

Paper Prepared in the Central Intelligence Agency⁴

Washington, April 13, 1988

What would it take to monitor the provisions of START with high confidence?

Introduction

This paper describes requirements for monitoring the fundamental components of a START agreement with high confidence.⁵It examines monitoring measures solely in terms of monitoring criteria, and does not attempt to assess the tradeoffs that clearly must be weighed between monitoring confidence and such factors as cost, disadvantageous counting of US forces, the impact on US operational readiness, or the counterintelligence threat. [portion marking not declassified]

The far greater monitoring problems of a START agreement, compared to the INF Treaty, arise from the fact that START would require the Intelligence Community to:

—Determine the number of Soviet weapon systems rather than to detect the presence of banned systems.

—Determine both the potential and actual loadings of Soviet weapon platforms.

—Determine the number of nondeployed missiles and their operational status, if such missiles are limited. [portion marking not declassified]

This paper explores measures that could improve our ability to monitor

—The number of mobile ICBMs.

—The number of RVs deployed on MIRVed ballistic missiles.

—Ballistic missile modernization. [portion marking not declassified]

We discuss four types of measures, all of which should be pursued in combination:

—Improvements to US national technical means (NTM) beyond the current baseline.

—Counting rules, definitions, and restrictions that could be included in the START treaty.

—Passive cooperative measures, which involve reciprocal agreed-upon actions by which one side facilitates monitoring by the other. Such measures could include exchanges of data, assured access to flight-test data, or restrictions on the operational practices of each sides’ strategic forces.

—Active cooperative measures, which would supplement NTM collection through the physical presence of personnel or technical sensing equipment or both on the territory of the other party. For example, inspectors or sensors could be stationed at production facilities to acquire information on the number of weapons produced. [4½ lines not declassified] [portion marking not declassified]

[24 paragraphs (84 lines) not declassified]

Illegal Production

To prevent production of extra mobile missiles at declared missile production and final assembly facilities, perimeter-portal monitoring of such facilities would be necessary.

To contend with possible production at undeclared facilities, unique tamperproof tags should be placed on mobile missiles produced at declared facilities, and the tag of each mobile missile that is flight-tested should be checked to prevent the undetected flight-testing of missiles produced at an undeclared facility. Such a measure would undermine Soviet confidence in covertly produced missiles. [portion marking not declassified]

Illegal Use of Nondeployed Missiles

START would allow both sides a substantial number of nondeployed mobile ICBMs in addition to those deployed, and there is a consequent risk that the Soviets could circumvent the treaty by maintaining some of these “nondeployed” missiles in an operational status, or by planning to augment their forces in a crisis. Even with strigent cooperative measures, it would be difficult to detect such cheating on a militarily significant scale.

—A treaty provision that required all mobile ICBMs to be counted against START limits on deployed systems, regardless of whether the missiles were “deployed” or “nondeployed,” would eliminate the opportunity to circumvent treaty limits with allowed nondeployed missiles. [portion marking not declassified]

Stockpiling Illegal Missiles Before a Treaty Goes Into Effect

The Soviets’ SS-24 and SS-25 missile systems have been in production a relatively short time and large numbers have not yet been produced. Consequently, if the Soviets declared a total more than tens of missiles below Intelligence Community estimates, their number would not be credible.

—Therefore, an early exchange of data, followed by frequent updates, would provide much higher confidence that there were not a militarily significant number of unaccounted missiles at the time the treaty entered into effect. (This would eliminate the problem encountered in connection with the INF Treaty with SS-20 production numbers.)⁶

[1 paragraph (5 lines) not declassified]

Fallback Measures

[1 paragraph (less than 8 lines) not declassified]

Determining the number of RVs on MIRVed ballistic missiles

[7 lines not declassified]

Once the maximum potential for a given type of deployed missile is established, it would not be necessary to inspect individual deployed missiles of that type, since all deployed missiles of that type would be counted as carrying the maximum number. [portion marking not declassified]

If the accountable figure is established by any criteria less than the maximum potential, it would be necessary to inspect individual deployed missiles to ensure they are not deployed with more than the permitted number.

—Since inspection of every deployed missile does not appear to be a practical option, confidence in monitoring through inspection will depend on the frequency, randomness, and short notice permitted for “sampling” inspections.

[3 paragraphs (17 lines) not declassified]

If the accountable figure for each type of missile is defined in the treaty as its maximum capacity, the breakout problem would be minimized.

Defining Maximum Capacity

Maximum capability could be defined as being the number of warheads of a given size that could fit in the payload area of a postboost vehicle (PBV).

Cooperative measures would be required to monitor such a provision with high confidence, including:

—On-site inspection to measure the payload area of PBVs. A sufficient number of test missiles, operational missiles, or both would have to be inspected on short notice for us to be confident we were measuring an actual PBV. We could not estimate the size of the payload area accurately enough with NTM to achieve high confidence.

—Prohibition on unusual PBV designs to prevent the Soviets from claiming the payload area was less than it really was.

—On-site inspection to measure the smallest RV that had been tested. Our uncertainty would be minimized if we were permitted to measure each RV that was flight-tested. Alternatively, if we were allowed to measure a few of the smallest RVs flight-tested on each type of missile, we could rely on NTM and short notice inspections of operational missiles to ensure that smaller ones were not tested or deployed. It would be important to prohibit deliberate distortion of the radar signatures of RVs if we had to rely on NTM to estimate RV shapes. [portion marking not declassified]

Such a counting rule would count some US and Soviet missile systems as carrying more RVs than the number each side intended to deploy. We assess, however, that such a rule would overcount US systems to about the same degree as Soviet ones and thus would appear to be equitable. A rule that overcounted some US and Soviet systems, however, could have some disadvantages:

—The 6,000 RV limit in START probably would have to be increased to 7,000 or so to compensate for some systems being counted with more RVs than was envisioned at the 1987 Washington Summit.

—The United States may be concerned that the Soviets would tend to deploy systems closer to the maximum RV loadings than would the United States. [portion marking not declassified]

An alternative, indirect approach (described below) for defining the potential RV loading of a MIRVed ballistic missile would be a rule based on potential throw weight; for example, dividing a portion of the potential throw weight by the weight of the lightest RV. [portion marking not declassified]

ICBM Modernization

A limit that attempted to literally prevent all modernization would have to prohibit changes in many different technical parameters, including things like specific impulse, and lengths and diameters of each missile stage—in addition to launch weight and throw weight.

[12 paragraphs (37 lines) not declassified.]

1. Source: Reagan Library, NSC Intelligence Files, 1988 SYS IV RWR INT 40201–40225. Confidential. Copied to Shultz, Weinberger, and Crowe. In a handwritten note of April 17, Powell wrote: “PSS, Put in system ASAP. Copy Back to me, pls. CP” (Ibid.)
2. No minutes were found.
3. Webster signed “Bill” above his typed signature.
4. Top Secret; [handling restrictions not declassified]
5. [Less than 2 lines not declassified]
6. [Less than 6 lines not declassified]