38. Memorandum to Holders of National Intelligence Estimate 11–8–681

M/H NIE 11–8–68


The Problem

To review recent evidence respecting Soviet strategic offensive missile capabilities and forces, and to assess its implications for revelant sections of NIE 11–8–68, “Soviet Strategic Attack Forces,” dated 3 October 1968, Top Secret, Ruff/[codeword not declassified] Restricted.

The Estimate

I. The SS–9 ICBM

A. Deployment

1. Since NIE 11–8–68 was issued we have detected the initiation of construction of five new groups of SS–9 launchers. We believe that three of these were started in the first quarter of 1969 (although one could have begun as early as December 1968), and two in the second quarter. The last previous group start was in May 1968. This system has six launchers to the group. When the 43 groups identified to date have been completed (probably in early 1971), the Soviets will have 258 SS–9 launchers operational; they now have 168.

2. [1½ lines not declassified] almost certain that all groups under construction as of about 1 April have been detected. The five new group starts of this year show a pace about like that of last year, when there were six group starts in the first six months. In the second half of last year, however, there were no new starts; we have of course no evidence of what may happen during the rest of this year. There is no evidence of the duration of the SS–9 deployment program or of the SS–9 force goal; we would judge now, however, that it will exceed 258 launchers.

[Page 144]

B. Accuracy

3. We continue to estimate that the SS–9 has a CEP of .75 n.m. using an all-inertial guidance system and one of .50 n.m. if a radio-inertial system is employed.2 No good basis exists for determining what portion of the SS–9 force has the more accurate guidance. [7 lines not declassified]

4. To give the missile a CEP of as little as .25 n.m. would require a new guidance system and a new re-entry vehicle, designed for either faster re-entry or some form of terminal guidance. Based on present evidence and what we believe about the Soviet state of the art, we continue to estimate that such accuracy could not be achieved before 1972. Some increase in accuracy could certainly be achieved by improving all-inertial guidance components without changing the present re-entry vehicle (RV). In our opinion, however, if the Soviets wish to reduce the CEP to something better than about .5 n.m.3 they would be likely to do so by using a new guidance system and new RV. Our interpretation [1½ lines not declassified] not persuade us that the Soviets are trying to improve the accuracy of the system. We believe that we would detect efforts to improve accuracy during the flight-test phase—and certainly so if the improvement should be substantial.4,5

C. Payload

5. The Soviets [less than 1 line not declassified] tests of the SS–9 with two sizes of RVs. We estimate that the lighter one could carry a war-head with a yield of [less than 1 line not declassified]. The heavier one could carry a warhead with a yield of [1½ lines not declassified]. The SS–9 payload with multiple re-entry vehicles (MRVs) weighs the same as the heavy single payload.

[Page 145]

D. Range

6. With the lighter payload, the SS–9 has an estimated maximum operational range of about 7,000 n.m.—more than enough to cover the entire US from present SS–9 deployment complexes.

7. With the heavy payload:

The observed facts are as follows: Since NIE 11–8–68 was published, the Soviets have fired the SS–9 with heavy payload into the Pacific to a distance of 5,100 n.m. These firings, however, took advantage of the earth’s rotation; on the same trajectories but fired north toward the US the range would be only about 4,700 n.m. At this range the missile could not reach the US (except Alaska) from most of its present deployment complexes; only those in one complex could reach targets in the extreme northwestern corner of the US. This presents a problem, since it seems implausible that the Soviets would develop an ICBM with a payload so heavy that it could not reach important targets in the US.
The above-mentioned test flights [1½ lines not declassified]. By [2½ lines not declassified]—the SS–9 with heavy payload would go approximately 5,000 miles; coverage of the US by the SS–9 as presently deployed would reach to a line extending from San Francisco to Boston, though most would not reach that far. Specifically, an SS–9 from one of the complexes could reach five of the six Minuteman complexes; from another, it could reach four of the six; from the remaining four complexes, the three northernmost Minuteman complexes could be reached. Thus, even with this range limitation, all currently deployed SS–9s could have Minuteman silos as their targets, and only the southernmost Minuteman complex would be completely out of range of the SS–9s with the heavy payload.
By [2 lines not declassified] the range could theoretically be increased sufficiently to allow some SS–9s with the heavy payload to reach the furthermost Minuteman complex. However, because of the uncertainty in performance which would be involved, we doubt that the Soviets would target their missiles in this manner without previous flight testing.6

[Page 146]

E. Retargeting

8. If the Soviets have a requirement to retarget the SS–9—i.e., to target and launch a backup missile on information that the original missile failed in flight—we believe it to be within their capabilities. There is no evidence of such a development, but it is unlikely that we would obtain such evidence.

F. Multiple Re-entry Vehicles

9. At the time of our last estimate, the Soviets had conducted two tests of the SS–9 with three RVs; since then they have conducted five more. We believe that each of the three RVs weighs about 4,000 pounds and could carry a warhead yielding [less than 1 line not declassified]. All seven tests were of MRVs which followed simple ballistic trajectories—i.e., they were certainly not independently guided after separation from the launch vehicle.

10. So far, then these tests have demonstrated at least a simple MRV, and one hypothesis is that this is all they are intended to achieve. If so, they would represent the culmination, or something near the culmination, of a development program initiated several years ago as the first answer to the ABM, which had of course been long under discussion in both countries. The Soviets were aware of US development of MRVs without independent targeting in the Polaris system. They probably gained, as time went on, a fair understanding of the operational concept of the planned Sentinel system and the general capabilities of its components. They may nevertheless have been uncertain both of the kill capability of the Spartan’s warhead, and of its method of kill—the Spartan relies upon a multimegaton warhead and nuclear effects to neutralize the target, while its predecessor the Nike Zeus had a much smaller warhead and was intended to physically destroy the target. When in September 1967 the US announced its intention to deploy the Sentinel, the Soviets may have decided to proceed with production of MRV hardware and go to the flight testing which began in August 1968. They would have done so, by this hypothesis, rather than wait longer for a system with multiple independently targeted re-entry vehicles (MIRVs). A simple MRV system could reach IOC late this year.

11. As tested to date, this MRV system would confront the Sprint element of the US ABM system with three separate targets. If the RVs were sufficiently hardened it would present Spartan with the same problem, and in any case the defenders would have to judge whether any incoming objects that survived had been neutralized. Except as a possible counter to ABM, however, the system as demonstrated does not improve Soviet capabilities to attack individual targets. In general, an ICBM so equipped would be no more effective against a soft target [Page 147] than one with a single large payload, and it would be less effective against a single hard target.

12. An alternative system can be postulated and related to the current Soviet test program—one with sufficient mechanical flexibility so that variations in the dispersal pattern of the RVs would allow each to be targeted against closely spaced individual targets, i.e., Minuteman silos. In considering this possibility the following points are pertinent:

Evidence [less than 1 line not declassified] suggests that the mechanism within the ICBM itself is more sophisticated than necessary if this development were only to achieve a simple MRV. In the hypothesized system, [1½ lines not declassified]. Variations in the size and shape of the impact pattern could be achieved [2 lines not declassified] to create the variety of patterns needed to target any substantial portion of the Minuteman force, i.e., so that each individual RV would fall within the required distance of the particular Minuteman silo which was its target.
The orientation of the impact pattern must also be variable to achieve independent targeting. To do this the orientation of the pay-load must be adjusted (i.e., “rolled”) either before launch or during powered flight, prior to release of the RVs. [3 lines not declassified]
We believe that the Soviets would want to test a capability to vary the size, shape, and orientation of the impact pattern by the amount required to target the Minuteman force, and that we will detect such testing if it occurs. The question then arises: were the very slight observed variations in performance [less than 1 line not declassified] intentional [less than 1 line not declassified] or were they random, [less than 1 line not declassified]. On this point [less than 1 line not declassified]. If the variations were intentional, this would indicate that the test series was indeed pointing toward eventual development of the independently targeted system we have hypothesized in this paragraph; if they were not intentional, the system would best be interpreted as a simple MRV.

13. If the Soviets are in fact aiming not for a simple MRV but for the system postulated in paragraph 12, this system could reach IOC in late 1970 at the earliest.7 It would have no better accuracy than the SS–9, and its reliability would be somewhat less. Further testing would certainly be required to develop the flexibility in spread and dispersal pattern needed for such a system, and we would be able to identify such testing when it occurred.

[Page 148]

G. Estimate

14. The system postulated in paragraph 12 offers a plausible explanation of the nature of the weapon system under test, but in our opinion the tests thus far observed provide insufficient evidence to judge that it is the probable explanation.8 If the postulated system is indeed under development, however, further tests will almost certainly provide data sufficient to demonstrate it.

15. The SS–9 is of course already a weapon with damage limiting capability. Equipped with the lighter payload it has the range, yield, and CEP to attack Minuteman silos with great effectiveness (though there are far from enough SS–9s to cover the entire Minuteman force). It seems quite likely that when this missile was first planned and developed the Soviets had in mind using only a single warhead, perhaps to attack Minuteman control centers. The feasibility of developing MRVs had not at that time become clear.

16. Equipped with the heavy payload—which is needed to carry the three separate RVs being tested—the SS–9 does appear to have range limitations,9 if we assume it to be used against Minuteman silos from present deployment. But it would still reach a great many (as discussed in paragraph 7b and 7c above). Assuming that the three RVs were indeed independently targeted they would greatly increase the damage-limiting capability which the SS–9 already possesses. In any event, if the Soviets intend to create a force to target 1,000 Minuteman silos in a single strike, they will have to deploy many more SS–9 launchers than are now operational and under construction. Additional deployment could be in complexes nearer the US, or perhaps the range of the missile could be improved.

17. The SS–9 booster has been used for other things than operational ICBMs. It has powered a number of space flights. The SS–X–6 [Page 149] tests, which we think related to development of a fractional orbit bombardment system (FOBS) or a depressed trajectory ICBM (DICBM), also employed the SS–9 vehicle. Conceivably it is being used now in the initial testing of multiple re-entry vehicles for eventual deployment on a new large ICBM. If and when a new missile does appear, there would of course be no reason why both it and the SS–9 should not have multiple re-entry vehicles.10

II. Other ICBMs

A. The SS–11

18. Since the issuance of NIE 11–8–68, we have discovered 11 new groups of launchers for the SS–11; this system has 10 launchers per group. When all the launchers in the 79 groups identified to date are completed there will be 790 operational; there are now 630. Nevertheless, when recent deployment activity is plotted against past construction starts, it seems clear that the SS–11 program passed its peak in 1966–1967. At present, construction is underway at five of the 10 SS–11 complexes at a rate roughly comparable to that of the last year or so.11

19. We continue to estimate the following characteristics for the SS–11 system: an operational range of about 5,500 n.m., [1½ lines not declassified]. To improve its accuracy significantly would require a new guidance system and a new RV. We have at present no evidence of a development program to these ends; we believe that we would be able to detect one if it occurred, and to ascertain its objectives.

[Page 150]

B. The SS–13

20. We have detected the start of construction of one new SS–13 group, but deployment of this system has still not extended beyond the single ICBM complex where it was first observed. We have now detected a total of 44 silos as compared to 22 at the time of the last estimate. The pattern of deployment currently indicates five groups with a total of 50 launchers. One of the groups is probably now operational and another soon will be. The slow pace and limited extent of SS–13 deployment constrasts sharply with the course of the SS–9 and the SS–11 programs. We have no better explanation of this than we had in NIE 11–8–68.

21. We believe that the small, solid-propellant SS–13 has capabilities roughly comparable to those of the SS–11. In all the SS–13 tests, [less than 1 line not declassified] it has never been flown to a range in excess of 4,700 n.m., and at this range it could not reach US targets from its present deployment area. In NIE 11–8–68 we estimated that [less than 1 line not declassified] the range of the SS–13 would be about 5,500 n.m. but we have not seen it tested to this range. [1½ lines not declassified] we do not know what the maximum range of the SS–13 may be.

C. Follow-on ICBM Systems: The SS–Z–3

22. When NIE 11–8–68 was being prepared last fall the Soviets had been working on a new launch group at the Tyuratam test range which appeared to be intended for a large, liquid-propellant ICBM about the size of the SS–9 or somewhat larger. Based on this activity and on our estimate of the future Soviet state of the art in guidance technology we estimated that the Soviets were developing a new large ICBM as a follow-on to the SS–9. (In NIPP–69 it was designated the SS–Z–3.) We estimated that it could be ready for deployment in the 1970–1972 period, [1½ lines not declassified]. We considered this new system the best candidate for carrying a new sophisticated re-entry system, and in NIPP–69 we projected the deployment of the SS–Z–3 with MIRVs.

23. Over the past year work has not progressed as we expected it would at the particular area at Tyuratam originally associated with the SS–Z–3. Work on the launch silos themselves ceased over a year ago. Construction of a new group of silos, however, began this year in the same general area. Work on the new silos has not progressed to the point that we can judge the size of the missiles that will be used with these launchers, but they may be intended for a large liquid-propelled follow-on system with characteristics similar to those estimated for the SS–Z–3. If such a system is developed it is evident that IOC will be later than 1970.

24. In NIE 11–8–68, we estimated that the Soviets would develop a mobile version of the SS–13. We also judged that they would probably seek to improve the quality of their force by modification of the [Page 151] SS–11. And we held that they might develop a new, small, liquid-propellant ICBM and a new, small solid-propellant system. We have acquired no evidence since the issuance of NIE 11–8–68 that any of these developments are underway.

III. Ballistic Missile Submarines

25. Since NIE 11–8–68 was published, there is additional evidence, still inconclusive, that a second shipyard has begun production of the 16-tube Y-class ballistic missile submarine. Production rate of this second yard would be about two submarines annually; our estimate of Y-class production at the rate of 4–8 per year took account of this possibility on the high side of the range. Considering other submarine programs in being and the space currently available on existing ways, we think it unlikely that Y-class production will be further increased. We see no reason to change our estimate of 35–50 Y-class submarines as the Soviet force goal. We continue to estimate that the Soviets could have 35 Y-class submarines operational by mid-1973 and 50 by mid-1975, as projected on the high side of the range in NIPP–69.

IV. Size and Composition of the ICBM Force

26. When all identified launchers under construction are completed and all groups of silos are filled out the Soviets will have 1,318 operational ICBM launchers—258 SS–9s, 790 SS–11s, 50 SS–13s, and 220 for the older SS–7 and SS–8 systems.12 This process could be completed some time in 1971. At that time the operational force will exceed the high side of the projection for that year in NIPP–69; it will be approximately at the middle of the range (1,100 to 1,500) projected as the Soviet force goal for the period 1974–1978. The Soviets will almost certainly build additional launchers; on the other hand, they will at some point almost certainly phase out some or all of the older launchers, which are far more vulnerable to attack than the newer ones. The total figure will be the net of these two developments.

27. It is quite likely that the Soviets have not yet fixed on definite force goals for the next decade. They still might for political, economic, and strategic considerations decide to stabilize their force of ICBM launchers at a numerical level roughly equal to that of the present US [Page 152] force. Or they might go for a substantially larger number than the US. We would continue as in NIE 11–8–68 to put this number at about 1,500.13 This figure does not of course represent the limit of their capability. Indeed, past performance at times of most intensive construction activity shows that they could proceed at a very much higher rate than the figure implies. Our projection takes into account other factors—e.g., economic costs, problems of resource allocation, and our belief that the Soviet leaders would not wish to stimulate the US into a new arms race of large scale.

28. Unless there is a change in the deployment patterns observed thus far, the bulk of this force will be composed of small ICBMs suitable chiefly for attacks on urban targets and other soft targets. The total megatonnage of the SS–9 force, however, is much greater, and the number of warheads potentially usable against hard targets would increase significantly if any sizable proportion of the SS–9 force was given MIRVs. The inventory of SS–11 and SS–13 silos is three times as large as the SS–9 total and if present trends continue the proportion of smaller missiles will increase. Because the SS–9 is a more expensive system, its deployment has cost about the same as the much larger SS–11 deployment; i.e., the equivalent of about $5 billion for each program.

29. It is clear that the SS–11 force will exceed the 700–750 that we projected in NIPP–69. But there is nothing as yet to establish that the total number of small ICBMs will exceed the 800–1,100 that we projected. This projected force included some 50–300 fixed launchers for the SS–13; we now question whether there will be much if any deployment of this system beyond the 50 now operational or under construction. If the Soviets decide to build toward the high side of our projections, we believe the SS–11, or possibly a small follow-on ICBM, may be a better candidate than the SS–13.

30. It is also clear that the SS–9 force will exceed the high side of the 234–246 launchers projected in NIPP–69. If, however, the SS–9 program should level off and if the Soviets do not develop the SS–Z–3 or its equivalent, the Soviet force of large, liquid-propelled ICBMs will [Page 153] probably fall short of the 334–396 launchers projected in NIPP–69, which included 100–150 launchers for the SS–Z–3. If the programs for one or both of these systems proceed, a force of this size can be attained by the mid-1970’s as estimated. Until we obtain more evidence, we have no basis for confirming or changing our projections.

V. Soviet Strategic Goals

31. Our judgment of the doctrines and goals which govern strategic programs remain as we stated them in NIE 11–8–68.14

  1. Source: Central Intelligence Agency, NIC Files, Job 79–R01012A, Box 341. Top Secret; Controlled Dissem; Talent–Keyhole–Comint; Ruff; [codeword not declassified]; Restricted Data. The Central Intelligence Agency and the intelligence organizations of the Departments of State and Defense, the AEC, and the NSA participated in the preparation of this estimate. The Deputy Director of Central Intelligence submitted this estimate with the concurrence of all members of the USIB with the exception of the representative of the FBI, who abstained on the grounds that it was outside of his jurisdiction. A memorandum from Cushman to recipients of the memorandum stressing its extreme sensitivity and a table of contents are not printed. For text of NSC 11–8–68, October 3, 1968 see Foreign Relations, 1964–1968, volume X, National Security Policy, Document 217.
  2. For dissenting views regarding the estimated accuracy of the SS–9 using an all-inertial guidance system, see the footnote to paragraph 4. [Footnote in the original.]
  3. [Footnote in the original not declassified.]
  4. Vice Adm. Vernon L. Lowrance, Acting Director, Defense Intelligence Agency; Capt. Franklin G. Babbitt, for the Assistant Chief of Naval Operations (Intelligence), Department of the Navy; and Brig. Gen. Ernest F. John, for the Assistant Chief of Staff, Intelligence, United States Air Force, believe that the evidence points to a different conclusion: i.e., the Soviets have an accuracy improvement program for the SS–9 utilizing existing RVs. [1½ lines not declassified] They believe the SS–9 system CEP for present deployment with these improvements would be .5 n.m. using all-inertial guidance. By 1970–1971 further refinements in these instruments could achieve a CEP of about .35 n.m. [Footnote in the original.]
  5. Mr. Thomas L. Hughes, the Director of Intelligence and Research, Department of State, noting the differences of view on the question of the degree to which the Soviets have improved the accuracy of the SS–9 system, and recognizing the criticality and growing complexity of the debate on this point, reserves his position pending the results of a technical evaluation by the intelligence community beginning in July. [Footnote in the original.]
  6. Vice Adm. Vernon L. Lowrance, Acting Director, Defense Intelligence Agency; Capt. Franklin G. Babbitt, for the Assistant Chief of Naval Operations (Intelligence), Department of the Navy; and Brig. Gen. Ernest F. John, for the Assistant Chief of Staff, Intelligence, United States Air Force, believe that the SS–9 (heavy), configured for operational deployment, has a range of up to 5,400 n.m. This 5,400 n.m. operational range allows for [1½ lines not declassified]. They believe that this is feasible with no appreciable degradation of reliability. [2 lines not declassified]. They do not believe that the Soviets would deploy it so extensively if they had doubt about it reaching important targets in the US. [Footnote in the original.]
  7. Our estimate in NIE 11–8–68 that the Soviets could not attain a hard target MIRV capability before 1972 was based on development of a more sophisticated system with an accuracy of .25 n.m. [Footnote in the original.]
  8. Vice Adm. Vernon L. Lowrance, Acting Director, Defense Intelligence Agency; Capt. Franklin G. Babbitt, for the Assistant Chief of Naval Operations (Intelligence), Department of the Navy; and Brig. Gen. Ernest F. John, for the Assistant Chief of Staff, Intelligence, United States Air Force, believe that although there are still unresolved technical issues the system postulated in paragraph 12 offers the more plausible explanation of the nature of the weapon system under test because of the following indicators:

    The SS–9 missile is the high-accuracy ICBM system of the Soviet Union;
    In the observed flight tests a clear attempt has been made to minimize the degradation to the CEP [less than 1 line not declassified];
    The use of multiple warheads independently targeted would multiply the effective number of boosters, while the limitation to only three RVs still provides sufficiently great yield in each RV to be effective against hard targets;
    The footprint size is comparable with the silo spacing in the Minuteman fields, although the specific variations required have not been demonstrated. [Footnote in the original.]

  9. See footnote to paragraph 7c. [Footnote in the original]
  10. Kissinger was scheduled to meet Cushman on June 13, the day after the USIB approved the revised NIE. No record of the meeting has been found. According to talking points prepared by Haig and bearing Kissinger’s initials, however, the White House was concerned that the estimate did not reflect a consensus and that it included inadequate explanations for the range and footprint of the SS–9. More broadly, Haig advised Kissinger to ask Cushman why the CIA was willing “to prognosticate about Moscow’s likely reactions to U.S. armament moves while in this instance the analysis reflects an unwillingness on the part of the intelligence community to go beyond very much more than a confirmation of absolutely proven facts.” An estimate, Haig continued, “which avoids any reasoned predictions based on the synthesis of known and probable facts, raises the question of whether an intelligence analysis is needed in the first place.” (National Archives, Nixon Presidential Materials, NSC Files, Agency Files, Box 207, CIA, Vol. 1) Abbot Smith, Director of National Estimates, in a memorandum to USIB members dated June 16, called for another USIB meeting to approve the revised estimate, which had been redrafted following the discussion between Kissinger and Cushman. (Central Intelligence Agency, NIC Files, Job 79–R01012A, Box 341)
  11. Very recent photography discloses what appear to be two SS–11 silos under construction at an IRBM complex near the Crimea. More extensive photographic coverage will be necessary before any useful estimate can be made on this matter. [Footnote in the original.]
  12. In addition to the ICBM launchers discussed above, each SS–9, SS–11, and SS–13 complex contains one additional silo and control facility not associated with any group in the complex. These are believed to be crew training facilities. There are now 6 such sites at the SS–9 complexes, 10 at the SS–11 complexes and 1 under construction at the SS–13 complex. Additionally, the Soviets have about 55 completed launchers and about 15 others under construction at Tyuratam and Plesetsk wich we associate with ICBM development. We believe that most of them, as well as the training sites, could be readied to fire at the U.S. We are unable to make any valid estimate of the time required to ready them, their reaction times, or the availability of missiles for them. [Footnote in the original.]
  13. Mr. Thomas L. Hughes, the Director of Intelligence and Research, Department of State; Vice Adm. Vernon L. Lowrance, Acting Director, Defense Intelligence Agency; Maj. Gen. Joseph A. McChristian, the Assistant Chief of Staff for Intelligence, Department of the Army; Capt. Franklin G. Babbitt, for the Assistant Chief of Naval Operations (Intelligence), Department of the Navy; and Brig. Gen. Ernest F. John, for the Assistant Chief of Staff, Intelligence, United States Air Force, believe that for the period of this estimate the Soviet ICBM launcher force goal may not exceed 1,500 launchers providing the USSR operationally deploys a sizable number of ICBMs with multiple reentry vehicles. Otherwise, and particularly in view of the number of targets in the US and the planned US ABM capability, the Soviet Union probably will have considerably more than 1,500 launchers by the late 1970’s. A program which added only about 100 launchers per year beyond those already identified would exceed 1,800 by mid-1978. [Footnote in the original.]
  14. According to NIE 11–8–68, “The primary objectives of Soviet strategic policy have been to achieve a more formidable deterrent and to narrow and eventually to overcome the U.S. lead in capabilities for intercontinental attack.” Lieutenant Colonel William Lemnitzer of the Joint Staff, in a memorandum to Wheeler of June 24, stated, “the revised estimate supports (but not unanimously) a grim view of the Soviet strategic force growth—in both an absolute sense and in terms relative to the U.S. strategic posture. It would appear that a prudent planner would tend toward the pessimistic alternative.” (National Archives, RG 218, Records of the Joint Chiefs of Staff, Records of the Chairman, General Wheeler, White House Memoranda (1969))