Foreign Relations of the United States, 1977–1980, Volume IV, National Security Policy

147. Minutes of a Policy Review Committee Meeting¹

Washington, August 7, 1979, 11 a.m.–12:45 p.m.

SUBJECT

• M–X Basing (S)

PARTICIPANTS

• State

  • Secretary Vance
  • R. Bartholomew, Dir., Bureau of Politico-Military Aff.

• OSD

  • Secretary Brown (Chairman)
  • W. Perry, Dir. Defense Research & Engineering
  • G. Schneiter, Dep. Dir., SALT Task Force

• JCS

  • General Jones
  • LTG J. Pustay, Asst. to Chairman, JCS

• DCI

  • Adm. Stansfield Turner
  • R. Bowie, Dir., National Foreign Intelligence
  • H. Stoertz, NIO for Strategic Programs

• ACDA

  • Spurgeon Keeny
  • Barry Blechman, Asst. Dir.

• OMB

  • John White, Dep. Dir.

• OSTP

  • Ben Huberman

• White House

  • Zbigniew Brzezinski
  • Landon Butler

• NSC

  • Fritz Ermarth

MINUTES OF MEETING

Brown: You’ve been briefed on the basing option involving horizontal shelters with a dash capability.² We want to review the considerations involved in deciding whether this is acceptable, namely, survivability, verifiability, and public acceptance. Soviet reaction is a factor in the case of survivability and verifiability. Then we want to consider how to move to a decision, possibly to an NSC meeting where we make a final recommendation to the President. Bill Perry will address the survivability issue.

Perry: I’ll present three illustrative threats and options for our responding, with the help of these charts (draw-down curves).³ In the first case the Soviets fill out their ICBM force under SALT II limits to a level of 6300 RVs by 1982–83. Clearly our MINUTEMEN are eliminated in a Soviet attack. The attacker can then devote the rest of his ICBM RVs to attacking our base-line M–X shelter system of 4600 aim points. He basically exhausts his ICBM force, leaving us with zero or very few survivors. The force balance is then constituted of SLBM and bomber weapons.

Turner: He can get this capability by 1982–83, whereas we don’t get those 4600 aim points until much later.

Perry: In the next case, the Soviets go to 10 RVs on most of the 820 ICBMs, extensively deploying their new ICBM for this purpose. They get a force of some 8600 ICBM RVs by 1987–88.

Turner: They could do it by 1986 under the High SALT Force of the NIE.

Perry: If we see this threat coming, our response is both to expand the number of shelters and to expand the number of deployed missiles such that, at a minimum, we could enforce the exhaustion of his ICBM force and retain a few survivors because of his unreliable missiles. The main thing is that we must see this threat coming about 1983, and we probably would see it for him to have such a capability by 1987–88. This threat level will also drive M–X procurement costs in the expensive years, 1984–86.

Keeny: [1 line not declassified]

Turner: [2 lines not declassified]

Perry: Basically, such inputs don’t matter. If they put an RV on a shelter, they can destroy it. The outcome is defined by the numbers of weapons and shelters plus missiles in shelters. In the third case, the Soviets go for very high fractionation levels and force expansion after 1985, essentially doubling the number of RVs we now expect on their SS–17s, 18s, and 19s. They’d have to start testing overtly by 1986 and could have a 12,000 weapon ICBM force by, say, 1989. In this case we’d halve the separation distance between shelters to 3500 feet from 7000 feet, doubling the number of shelters right off, and further expand the number of shelters as necessary. This would again enforce the exhaustion of his force. In the first case, our costs would be the base-line costs of $30B. The second runs us up to $38B; the third to $41B. In the third case, we’d have to make a decision around 1983–84 to respond.

Brzezinski: Can we see the threat that soon?

Brown: We can respond when we see the tests he has to conduct to give himself these high fractionation levels.

Perry: We can build shelters and expand the system as fast as he can expand RV numbers.

Brown: We’d build the necessary roads and mark out shelter sites from the beginning. This could “bluff” him out of going the high fractionation route.

Brzezinski: Or it might provoke him to take that route, or give him an excuse. Whether they actually believe it or not, they could adopt the public rationale that they expect all our shelters to be full and hence have a right to cover them with weapons.

Brown: Of course, if the Soviets feel threatened by empty shelters, that would be an ominous development indeed. If they’re that paranoid, how can we reassure them by self-restraint? We could go to shelter numbers capable of exhausting 20,000 RVs, but this becomes an unreal world. In my view it becomes unreal for the attacker before it becomes unreal for the shelter builder.

Perry: The Soviets face some limits on the amount of nuclear material they have for these high weapon levels. They also face an exchange ratio of 3 weapons expended to kill one of ours. At the force levels posited here, moreover, [3 lines not declassified] but by running this kind of race they can assure that we could destroy their whole ICBM force in a first strike. It is this logic that may persuade them to go mobile. Remember, moreover, that these are limiting cases. In the real world, they would probably never put all their ICBM weapons against our ICBM targets. Hence, at the shelter numbers indicated, we would come away with respectable numbers of M–X survivors.

Brown: The point is that the multiple shelter approach to M–X basing can adequately respond to threat growth. There are many uncertainties around, but they weigh most heavily against the attacker.

Perry: In the third case, we’d aim for about 400 M–X by 1990.

White: Do you see any cases to which we cannot adequately respond by expanding the shelter numbers?

Brown: No, not realistic threats. But if you want to worry about threat levels I consider unrealistic, we have other options.

Turner: We see it a bit differently. Because the yields we use for these calculations are somewhat higher, when the Soviets fractionate, we see that they might put weapons between shelters and get more than one-for-one.

Brown: In this kind of game, it is inconceivable to me that Soviet target planners would not insist on placing a weapon on each shelter. No US planner would proceed otherwise.

Turner: I think it’s quite possible they would try for multiple shelter kills. The point is, can we safely rule that out?

Brown: If we go to 3500 foot spacing, I feel very confident that he’d have to put a weapon on each shelter to be in the game.

Jones: This is sort of academic, since we’ll use the best data on yields and numbers available at the time we start construction.

Turner: I’m challenging the assumptions here. My calculations show that, if the Soviets pursue a moderate effort under SALT, they will be some 600 ICBM weapons short of an ability to target our ICBMs. If their effort is on the high side, then they could have some 1800 ICBM weapons more than necessary to cover MINUTEMAN and M–X. But with SLCM and bomber weapons counted, we’d still have a 2-to-1 advantage in residuals. But these outcomes are highly dependent on SALT II limits affecting RV numbers. Without those limits, he can cover our baseline shelter set of 4600 easily and enforce a 1.5-to-1 residual weapon advantage. A high effort without SALT could give him in excess of 10,000 ICBM weapons. We’ll have to pay a high premium to assure extension of the SALT II RV limits beyond 1985. By 1983–85, he could put a lot of pressure on us by starting to test the technology for, say, 30 RVs on an SS–18. He could limit the number of simulated RV releases to 10, but design a new bus capable of handling 30 new RVs.

Vance: He’d then be in violation of SALT II.

Brown: We’d certainly consider that a violation. This is a major matter, but not new.

Keeny: If he didn’t try to draw us down all the way, then he’d have a residual force under Perry’s calculations and a more favorable exchange ratio.

Brown: You’re now talking about a limited attack of 5,000–10,000 RVs. I don’t believe in that.

Perry: The right answer is that he’s better off not starting. If he leaves any of our M–X surviving, they are much more effective against his force because he is in single silos.

Turner: Let’s try to get these calculations in line with the NIE before we present them to the President.

Brown: The President should understand the picture of possible threats and our response options. We should not pretend we can depict future threats with certainty.

Turner: But we have agreed threat estimates within the government.

Perry: We can adjust my calculations for the official NIE numbers. The results will be the same: we can, at a minimum, present a shelter array that would exhaust the Soviet ICBM arsenal.

Turner: That’s true if you keep running up the number of shelters.

Brown: Of course. I can run up the shelter numbers as easily or more easily than he can run up the weapon numbers. (Nautical anecdote about anchors and hurricanes followed.)

Perry: The great benefit of SALT II is that it assuredly delays the development of the worst kinds of threats to this system. We need SALT II to prevent them from exploiting their throw-weight and getting the jump on M–X deployment.

Turner: We ought to look at the threat of early Soviet abrogation of SALT II in order rapidly and early to expand their RV threat to M–X.

Perry: Going into M–X without SALT II would be a real problem. There is a physical limit on suitable land. At 7000 foot spacing, we could build 10,000 shelters; at 3500 foot spacing, 20,000.

Brown: This is a ridiculous extreme.

Brzezinski: But there is another angle to Stan’s point. The Soviets have an incentive to abrogate SALT II on account of M–X. If they showed us a credible threat or preparation to do so, would we want to go ahead with M–X? We should examine this gambit as a Soviet means of deterring us from proceeding with M–X.

Brown: This is a good question. Facing that kind of threat, we would be credibly interested in options beyond expanding the number of shelters, e.g., providing for ABM defense of M–X. To deter the Soviets from trying the abrogation threat, we have to have a credible ABM program. We also have to convey willingness to abrogate the ABM agreement or be willing to regard Soviet abrogation of SALT II as tantamount to total abrogation of SALT.

Perry: We could also sharply accelerate other programs, e.g., increase our buy of ALCM. In terms of numbers of weapons added to the force over time, we could match a Soviet ICBM escalation with an escalation in the procurement of ALCMs. This in addition, possibly, to the ABM defense would be an adequate response to and deterrent of the Soviet SALT II abrogation threat.

White: This would be in addition to, not a substitute for, expansion of M–X shelter numbers?

Brown: Yes.

Turner: My concern is, given these kinds of threats and necessary responses, is M–X in multiple shelters the right way to contend with possibilities readily open to the Soviets because of their high throw-weight?

Brzezinski: This is a real question because the Soviets are likely to have a strong incentive to cover the whole strategic target set as well as they can.

Brown: But there are lots of strategic targets they have an incentive to cover. Our aim is to give them a problem that is too tough and expensive for the payoff.

Keeny: We should not lock ourselves into a specific base-line number of shelters like 4600 because this can too easily be demonstrated by the amateur critic to be inadequate. In any case, these calculations show that M–X is not survivable against the threats postulated. M–X exhausts the threat but does not survive. You’re left with SLBMs and bombers against each other, which is where you’d be without buying M–X.

Brown: What do you think the Soviets would do with their 5,000–10,000 ICBM weapons? Without a large hard-target array to shoot [1 line not declassified] Part of the problem is whether the Soviets have a survivable ICBM capability while we do not. Kissinger has made the point of how intolerable this would be. M–X challenges the survivability of Soviet ICBMs. As to the apparent failure of M–X to survive in these calculations Bill presented, remember they are limiting cases in which the entire Soviet ICBM arsenal is thrown at M–X and MINUTEMAN. They would never throw [number not declassified] percent of their force at [number not declassified] percent of ours. They have many targets to cover and residual force needs.

Keeny: But these kinds of calculations make it look like the rationale for M–X is to exhaust the enemy, not to survive. This is in part a political problem.

Perry: If we actually deploy according to the defense conservative logic we display in these calculations, we would in fact have much higher survivability. And when the Soviets look at the problem in an offense conservative manner, then they will be deferred from attacking and probably dissuaded from attempting to overcome the shelter array with weapon numbers.

Vance: These massive and pure attacks are highly unlikely. In a real attack, M–X would be more survivable than displayed on the charts.

Brzezinski: At a minimum, we can enforce a situation in which he must disarm himself to attack M–X. In reality, we can do better than the minimum.

Jones: Today, if you want to believe in highly artificial attack threats, he can take out 100 percent of our bombers with 100 percent of his ICBMs in an area overpressure attack.

Perry: The more realistic depiction of the threat is the one used by the Air Force to configure the base-line. If he puts [number not declassified] percent of his hard-target-capable weapons against [less than one line not declassified] percent of the M–X missiles will survive.

Brown: Let’s move on to verification. We have an overhead photo here⁴ which [2 lines not declassified] There are, of course, other means to verify than photography. Stan, do you want to comment?

Turner: There are two problems: The Soviets verifying us, and we verifying the Soviets. [13 lines not declassified] Our confidence in verifying a Soviet version is bound to be less than theirs because we face a closed society and shall have less confidence in cooperative measures.

Brown: But we have better resolution and real time imagery.

Turner: By the late 1980s these advantages may not be so great. [9 lines not declassified]

Brown: How many SS–16s might they hide this way, in your view?

Turner: [3 lines not declassified]

Brown: Why not hide them elsewhere, away from multiple aim point confinements? They’ll have more chance of getting away with it where we are not looking hard.

Turner: [5 lines not declassified]

Brown: Wouldn’t they do better to deploy SS–16 type missiles in SS–20 deployment areas?

Turner: [1 line not declassified]

Brown: But we shall be looking closely at both possibilities for cheating, and using all types of intelligence.

Vance: [1 line not declassified]

Turner: [1 line not declassified]

Jones: Adjacent confinements could be opened simultaneously to assure no escape of an illegal missile from one to another.

Perry: In our scheme, we would be able to open a whole valley of several adjacent confinements to protect against this problem.

Brzezinski: In my view, it’s hard to see them building the multiple shelter system unless they are planning to cheat.

Turner: [2 lines not declassified]

Brown: They can go straight mobile deployment.

Turner: [2 lines not declassified]

Brzezinski: On the contrary, they’ll claim M–X is unverifiable and use that to justify building their preferred mobile deployment scheme as they wish without reference to our verification concerns.

Brown: I am convinced we have a good case on verification.

Brzezinski: That may be, but we cannot enforce it. Their incentive to cheat is high.

Keeny: DOD has clearly made heroic efforts to make our scheme verifiable. Our topography in the West helps, especially the separated valleys. What about Soviet topography? How do we stand with the size of the port holes required? How about the problem of using an image to fake the bottom of the shelter?

Stoertz: Stereo and other image enhancements help.

Brown: Faking images is a problem in other areas too, e.g., making SSBNs look like SSNs.

Jones: Different sensors will be used, e.g., radar, to ease this problem.

Perry: The observation ports have to be big enough to allow stereo imagery of the floor, to measure real depth.

Vance: I take it, Stan, you’re satisfied we can verify [1 line not declassified]

Turner: [3 lines not declassified]

Vance: [less than one line not declassified]

Turner: [less than one line not declassified]

Stoertz: But it would not be definitive.

Brown: We need your position on this sooner. We are under pressure to move quickly to a basing recommendation. The verification problem you raise will never be resolved definitely.

Brzezinski: We need a recommendation to the President by the end of the month when he returns. And we can hold an NSC on it in the first week of September.⁵

Brown: We have the following agenda of issues, then: Stan, you will complete your study [less than one line not declassified] by the end of the month. DOD will reconcile Bill’s draw-down curves with official NIE data. And we shall look at the threat of Soviet abrogation of the SALT II limits prior to 1985, plus our responses other than expanding shelter numbers.

Keeny: We still have to examine the likelihood that the location uncertainty will collapse.

Brown: This is less critical because we can reshuffle or dash.

Keeny: But surely maintaining location uncertainty is still critical. We might not know that we should move or dash.

Brown: Our move/dash capability denies him any basis for confidence that he could ever use his ability to locate the missiles were he ever to acquire such an ability.

Brzezinski: Would we dash in all cases of tactical warning of an incoming attack?

Brown: We could require dash on tactical warning if the number of incoming weapons indicated that the attacker thought he had broken the location uncertainty. We’ve provided for including mass simulators, but I believe we need only deploy them if we see convincing evidence that he can threaten to break location uncertainty.

Turner: Maintaining location uncertainty is a genuine concern. Sensor technology is improving, cooperative measures to assure verifiability will tend to compromise location uncertainty. The Soviets will certainly work on the problem. Would you build the system this way if there was a reasonable probability they could break the location deception?

Brown: This is a matter of juggling doubts and uncertainties on both sides. Our side will have doubts about survivability. Their side will have doubts about their capacity to attack the system. On balance these doubts will come out favoring not going to war, i.e., favoring deterrence. That is why we selected this approach.

Brzezinski: What is the speed of dashing?

Perry: If dashing were fully adequate for survivability, we wouldn’t have to hide. The combination of hiding and dashing is adequate.

Brown: Let’s conclude. Have we had enough discussion to make a recommendation to the President right after Labor Day? What issues remain?

White: We have to examine costs further: cost uncertainties, the impact of IOC on cost profiles, and the impact of M–X spending profiles on other DOD spending priorities.

Brown: We’re hurting when the annual outlay for M–X hits $2B.

Jayne: The year of FOC is crucial.

Brown: 1983–85 will be the expensive years.

Vance: Our decision schedule turns on Stan’s date for completing his work, doesn’t it?

Turner: My concerns may not be all that critical; they involve relatively unlikely possibilities on which we must have good analysis as we move ahead.

Vance: We have to answer your questions.

Turner: We’ll do our best.

Brown: Other issues? Public acceptance?

Vance: The public acceptance issue of greatest importance right now is the requirement that we speak with one voice. It should be DOD’s voice.

Brown: We’ll look further at the location uncertainty question.

Huberman: We understand you will also be conducting a positive security program for maintaining location uncertainty, analogous to the program for submarine security.

Keeny: In addition to location uncertainty, what about the threat of SLBM attack?

Perry: If SSBNs come in close, we shall go into a mode of constant motion or readiness for motion with the ability to get to a new shelter in time to thwart an SLBM attack. If we get indicators of enemy SSBNs coming in close, we can defeat the SLBM and the ICBM threat.

Ermarth: We may have a political problem between now and early September, according to the schedule laid out. The attentive public expected a firm decision out of the 14 August PRC; this is that PRC. How do we represent the results of this meeting? It will not do to say we are continuing to study the problem.

Brown: We’ll say what I’ve been saying in testimony: A consensus has emerged on a satisfactory basing mode, and a recommendation will be made to the President in a matter of weeks. Of course, we have been saying “a matter of weeks” for a matter of weeks.

1. Source: Carter Library, National Security Council, Institutional Files, Box 77, PRC 121, 8/7/79, MX Basing Mode. Top Secret. The meeting took place in the White House Situation Room. The Summary of Conclusions of the meeting is ibid.
2. See footnote 2, Document 146.
3. Not found attached.
4. Not found attached.
5. On August 17, President Carter and Mrs. Carter flew to Minneapolis to embark on a steamboat cruise down the Mississippi River, arriving in St. Louis on August 24. From St. Louis they flew aboard Air Force One to Camp David, where they stayed until August 30, when they departed for Plains, Georgia. They returned to the White House the evening of September 3. (Carter Library, Presidential Material, President’s Daily Diary)