Foreign Relations of the United States, 1969–1976, Volume XXXII, SALT I, 1969–1972

58. Paper Prepared by the National Security Council Staff¹

Washington, March 23, 1970.

ANALYTICAL SUMMARY

SALT ISSUES AND OPTIONS

Introduction

For the past several weeks, the Verification Panel has directed a series of 15 analytical studies on the key substantive issues that must be resolved in deciding on the U.S. position for the next phase of SALT.

The results of these studies have been incorporated in the Panel’s draft report on the “Evaluation of Possible Strategic Arms Control Agreement Between the United States and the Soviet Union,”² which has been distributed to the NSC members.

The Evaluation Report presents the issues in two ways:

• —Section IV discusses and summarizes the issues analytically;
• —Section V groups the substantive issues into the three categories of agreements you outlined in your recent Foreign Policy Report:³
  • —limited agreements, characterized mainly by concentration on numerical limits on major strategic systems;
  • —comprehensive agreements, providing important limits on the characteristics and capabilities of major strategic systems, as well as on their numbers;
  • —agreements going beyond numerical or qualitative limitations to embrace reductions in existing forces as well.

An unavoidable characteristic of the key issues is their extraordinary technical complexity. Unfortunately, the most complex aspects of these problems often have the most crucial bearing on whether or not a particular limitation would be consistent with the maintenance of our security. Therefore, the Panel has made every attempt to master them.

For the past two weeks, the Verification Panel has held a series of meetings⁴ to see, first, if we could develop an approach to the issues that would illuminate the basic decisions you must make without bringing in all the esoteric details that have held our attention, and, second, if we could reach some agreements on basic issues that would further ease the task of developing our position.

The Panel generally agreed that the best way to approach the question of which option or options we want to use as basis for our position at Vienna is, first, to review the basic substantive policy issues and, second, to see how decisions on these issues relate to the choice of an approach and to the various possible options within each approach.

ABMs and MIRVs

The Panel agreed that the best way to simplify the problem is to begin with the ABM issue.

The critical ABM question is, do we want to deploy a nationwide area defense of the magnitude planned in the Safeguard program no matter how such a decision affects the type of agreement we may be able to reach with the Soviet Union? The Safeguard area defense calls for 19 radars deployed nationwide and 633 interceptor missiles.⁵

This is the critical question because the Verification Panel is in full agreement on three major issues:

• —Maintaining the Safeguard area defense rules out a MIRV ban;
• —From the point of view of reaching a verifiable arms control agreement, the only practical alternative to maintaining a Safeguard level area defense is to ban ABMs altogether or, at the outside, to defend the National Command Authorities in Washington with a small deployment.

In particular, the Panel agreed that it would be practically impossible to define an agreement which would allow both sides to deploy an intermediate level ABM defense of their land-based missiles but no area defense.⁶

The reason is that the Soviets have not developed a small ABM missile like our Sprint. If we allowed them to use their much larger ABM missiles for defending their ICBMs, (assuming they would have any interest in such a proposition) we would be conceding to them a significant area defense capability that we would not have for two reasons:

• —many of their ICBMs are deployed near large cities, whereas ours are not, so missile defense is also population defense in the Soviet Union;
• —they already have a radar network adequate for area defense; we do not and would not if we limited ourselves to Minuteman defense.

If we insisted that only small missiles were allowed for the purpose of defending ICBMs, we would in effect be denying the Soviets any ABM for the time it would take them to develop a Sprint-type missile, probably several years.

• —An area defense system thinner than Safeguard would not give us enough capability to justify it.

Though these three conclusions seem harsh, the Panel reached them only after the most thorough assessment of the technical issues associated with defining ABM limits that would be verifiable without highly intrusive on-site inspection.

The problem with controlling ABMs is that, unlike a limit on offensive missiles, a limit on ABMs has two components:

• —a limit on the number of ABM-capable missiles, and
• —a limit on the number of ABM-capable radars.

A control on the number of ABM launchers would be ineffective if each side is allowed to have a nationwide network of large ABM-capable radars which could support a much larger number of ABM launchers than the agreement permits. The existence of this radar network would pose two risks to our retaliatory capability:

• —The Soviets might build and stockpile ABM missiles that could be deployed rapidly following abrogation of an agreement using the already existing radar network;
• —Again using the existing radar network, the Soviets might upgrade their extensive air defense missile systems—mainly the SA–5 or Tallinn system, which is projected to have about 2000 launchers and about 8000 missiles, more than half within range of populated areas—to give them an ABM capability which, though limited, might still jeopardize our second strike capability. This latter problem, which we refer to as the SAM upgrade problem, is one of the most complex and difficult we face.

In effort to resolve these problems, the Verification Panel grappled with two questions:

• —Is the present Soviet radar network, let alone a larger one, capable of supporting a clandestine ABM build-up or extensive upgrading of SAMs or could it be made so without our knowing it?

There are disagreements about this issue:

• —Some argue that the present Soviet radar network is not capable of supporting an extensive ABM or upgraded SAM system, and, further, that many of these radars are relatively undefended and thus highly vulnerable to attack.
• —Others argue that, if it is not now, it could be made so without our knowing it and, furthermore, that 5–10 new advanced (phased array) radars could be deployed without our detecting it.
• —The second question is, can we find ways of limiting the Soviet radar network so that their ABM capability is effectively limited to levels that would not pose unacceptable risks to us.
• —The JCS say the answer is no. Because the Soviet radars are so numerous and can serve so many different purposes, the JCS argue that it is impractical to try and define effective and verifiable constraints on ABM-capable radars.
• —Others believe that a fairly comprehensive package of limits on radars, SAMs, and ABMs could provide a verifiable constraint on ABM capability. However, as I read their argument, a simple numerical limit on radars is by no means enough, even to those who believe we can control ABM capability.

Thus, if one takes the view that the present Soviet radar network could support SAM upgrade and that practical and effective limits on ABM radar capability are difficult if not impossible to achieve, then an ABM level that allows the maintenance or increase of Soviet radar capability, which is true even of intermediate ABM levels, poses the two risks I mentioned earlier: a clandestine ABM build-up or SAM upgrade.

MIRVs are the best hedge against this possibility. They give us the highest confidence that we can penetrate future defenses and strike Soviet targets even if the Soviets violate an agreement to limit ABM defenses. That is why the Panel believes that MIRVs are required if Safeguard level ABM launchers and radars are allowed on both sides.

[There are other ways to penetrate defenses. However, last year the MIRV Panel agreed that the testing of many of them would also have to be banned if we were to be able to verify that MIRVs were not being tested, and the DOD representatives on that Panel believed that tests of virtually all of the alternative ways to penetrate defenses would have to be banned along with MIRV tests.]⁷

In deciding whether or not the Safeguard area defense should be non-negotiable, you should be aware that there are some sharp disagreements concerning both the technical capability and the political/diplomatic usefulness of the system against the Chinese threat.

The Safeguard system’s design goal is to deny damage from a threat of 10–25 Chinese ICBMs, even if they have first generation penetration aids. If the Chinese threat continues to grow, the system’s goal is to reduce damage to 20% or less of that expected with no defense. Proponents argue that the system will work as planned. (The estimated performance of Safeguard Phase 2 area defense as now defined is to limit losses from the postulated 1980 Chinese threat to 1–3 million fatalities, compared to 30–40 million without Safeguard. There is, however, some chance it would prevent any losses.)

Others argue, however, that only a single three-megaton Chinese weapon leaking through and hitting one of the 6 largest U.S. cities would cause one million instant deaths and negate the damage denial capabilities the system must have for it to be credible. The Chinese will be able to insure this with simple penetration aids and by concentrating their fire on a single site and exhausting it.

As far as the diplomatic value is concerned, the question is whether a light area ABM defense of U.S. cities would permit the U.S. to take actions against a nuclear armed China which we would deem too risky in the absence of such a defense.

Proponents believe that even a small Chinese nuclear threat against unprotected U.S. cities could make a major difference in the effectiveness of U.S. diplomacy in Asia. A Chinese ability to hold U.S. cities hostage against U.S. intervention on the behalf of U.S. and allied interests in Asia, leaving the President with the single option of a nuclear attack on China if he is to blunt the Chinese nuclear threat, could increase Chinese influence or adventurism in Asia, promoting instability or nuclear proliferation.

Others argue that our overwhelming nuclear superiority over China is an adequate deterrent to Chinese nuclear adventurism. For ABM defense to strengthen this deterrent, it would have to be perfect. But it won’t be both on technical ground and because the Chinese can threaten Alaska, Hawaii, U.S. forces overseas or Asian cities instead.

Further, our NATO allies could interpret claims that the U.S. needs an anti-China ABM to make its deterrent umbrella effective in Asia as raising questions about whether U.S. deterrence alone is sufficient in Europe against the far greater Soviet threat, for which we have no ABM protection.

If we want to give up area defense and the strategic capabilities it provides, it is presumably because we want to keep open the possibility of:

• —a MIRV ban aimed at reducing the SS–9 threat to Minuteman and, secondarily, at reducing the threat MIRVed SLBMs would pose to bombers and other time urgent targets in the U.S.;
• —a ban or very low limit on Soviet ABMs, including radars, which would make it very difficult for them to build the ABM capability they would need to backstop a first strike offensive force.

Many of our technical people believe that an ABM ban would require destruction of existing large radars in the Moscow area, and, at the same time, that we should seek measures to increase our confidence that other large radars elsewhere in the Soviet Union (Hen Houses) could not be used to violate an ABM prohibition. Without such destruction and collateral constraints, we cannot be certain that Soviet SAM systems could not be given ABM capability using the existing radar network. [The measures suggested include prohibiting the defense of the Hen Houses with SAMs and devising a limitation formula based on the number, geographical orientation and technical characteristics of the Hen Houses. I do not believe ACDA would concur that all these measures are required.]

The Verification Panel addressed the question, if you ban ABMs, can you safely ban the testing of MIRVs?

In general, the answer was that if the ABM ban is accompanied by measures which reduce the Soviet radar network to the point where the risks of a clandestine ABM build-up or upgrading of air defense missiles to give them an ABM capability are minimal, a MIRV test ban is feasible.⁸

However, many Panel members, and I believe the JCS in particular, believe that even if we consider an ABM ban, we might want to permit MIRVs. [Paul Nitze believes this strongly. He thinks this might be Laird’s view.]

• —Unless we can achieve reductions, the Soviets will soon have at a minimum 1,400 land-based ICBMs, in addition to their submarine-based missiles. If they are made accurate in the next few years, they could destroy our Minuteman force even without MIRVs. Thus, we want to preserve the capability to MIRV our submarines.
• —Even now, the JCS believe we do not have enough nuclear warheads to cover all of the important strategic and other military targets in the Soviet Union, and the Soviet target system is growing. Therefore, MIRVs may be required for improved target coverage.
• —We might not get the Soviets to destroy existing ABM radars such as the Dog Houses, though they might agree not to build more. Thus, we may want MIRVs as a hedge against Soviet use of their existing radar network to violate an ABM prohibition. [If the Soviets won’t destroy existing radars, we might insist on Spartan protection of NCA and allow the Soviets the Moscow system.]

To summarize our conclusions concerning ABMs and MIRVs, we have three practical alternatives:

• —Safeguard level area defense with MIRVs permitted;
• —a ban on ABMs with MIRVs banned;
• —a ban on ABMs (or possibly an agreement that the Soviets can retain the Moscow system and we would defend the NCA) with MIRVs permitted.

Force Survivability

The Verification Panel agrees that, regardless of the decision on ABMs and MIRVs, we should seek to put a ceiling on the total number of offensive missile launchers and that, within the total, it is desirable to allow freedom to convert ICBMs to SLBMs. The JCS (mainly out of deference to the Air Force) would also like to allow a shift from SLBMs back to ICBMs. The objection to this is that it would permit the Soviets to build new ICBM silos into which they could put SS–9s or even larger missiles, a possibility that would be to our net disadvantage.

The Verification Panel also agreed that, in connection with controls on ICBM numbers, we should seek a ban on land mobile ICBMs (The Soviets have active mobility programs; we do not.), a ban on the relocation of existing ICBMs and on changing the external configuration of existing silos, a ban on the construction of new silos for intermediate and medium range ballistic missiles (IR/MRBMs),⁹ and a ban on mobile IR/MRBMs with a range greater than 1,000 kilometers.

We believe that these measures could be verified with national means. However, there would still be uncertainties. For example, all agree that at least 100–150 land mobile missiles, about 100 fixed land based missiles, and 5–10 ballistic missile submarines (say 80–160 missiles) could be deployed without our knowledge. OSD and the JCS believe that if the Soviets practiced maximum concealment and deception, these numbers might be larger. Some 70 older SS–7 missiles could be replaced with SS–9s without our detecting it.

On the other hand, there is a possibility that we could detect their activities accurately, and this possibility might be enough to deter the Soviets from extensive cheating of the kind just described.

The Panel reached no conclusions concerning the desirability of seeking to control missile accuracy and missile payload. Such qualitative controls are a practical possibility mainly if we seek a MIRV ban. They are relatively hard to verify, and the Soviets showed little interest in such provisions at Helsinki because of the verification problems.

However, such provisions in conjunction with a MIRV ban would, if verifiable, greatly enhance the long-term survivability of our Minuteman force.

Force Reductions

The Panel’s Report discusses two basic forms which reductions could assume:

• —modest, largely symbolic, first step reductions of limited numbers of existing systems on a comparable basis by the two sides;
• —ambitious reduction programs, with rather large reductions in some systems, possibly coupled with allowed increases in others, aimed at establishing a new and more stable strategic balance at a somewhat lower level in which the fixed, vulnerable land based ICBM force would be eliminated or greatly de-emphasized on both sides.

The Verification Panel reached no conclusions concerning the desirability of proposing force reductions. I think it is fair to say that:

• —Some (mainly Paul Nitze) believe that reductions are the most dramatic, yet serious, proposal we could make other than proposing a MIRV ban and that reducing offensive missile potential on both sides would be advantageous for the United States because otherwise the Soviets will build up an enormous advantage over us in offensive missile payload.

Particularly if we choose not to propose a MIRV ban, proponents of reductions believe we should propose force reductions to convince the large number of Senators who believe a MIRV ban is the test of whether the Administration is serious about arms control¹⁰ that, though a MIRV ban is not in our interest, we are serious nonetheless.

• —Others doubt that reductions are a negotiable proposition, at least in the initial stages of the discussions. They also point out that the strategic importance of reductions appears to be much less if the agreement leaves the U.S. free to shift its land based missiles to sea, a provision all agree should be included in any agreement.

I believe that the reductions issue should continue to be taken seriously even if it has no bureaucratic sponsors.

Other Issues

There are a large number of issues that will have to be resolved before the delegation goes to Vienna. The two most important of these are our position on limiting strategic bombers and our position on seeking limits on total Soviet intermediate range and medium range ballistic missiles.

The other issues essentially involve a choice of negotiating tactics rather than policy.

These issues can be discussed at a later time.

Options

We are now in a position to show how decisions on the basic policy issues relate to the choice of an option.

• —If you choose to maintain Safeguard level ABMs and MIRVs, you rule out a comprehensive option covering both numerical and qualitative controls. Your choice is then between one of the limited options and an option calling for force reductions.
• —If you are willing to seek a ban on ABMs, you could still choose a limited option, that is, an option which permits MIRVs, but you could also consider a comprehensive option involving a MIRV ban and perhaps other qualitative controls, as well as an option involving force reductions.

The Evaluation Report provides an illustrative option of each type and lists the variants and alternatives available within each type.

For example, the illustrative limited option includes the following major provisions:

• —a ceiling on the total number of ICBM and SLBM launchers operational as of the date of the proposal, with later substitution of SLBMs for ICBMs;
• —a ban on land mobile ICBMs and IR/MRBMs and launchers;
• —a ceiling both on the total IR/MRBM force and on IR/MRBM silos;
• —ABM missiles, launchers, and radars would be limited to agreed numbers, and some provisions relating to the upgrading of surface-to-air missile systems would also be included.

The illustrative comprehensive option includes all the provisions of the limited options and, in addition,

• —a ban on the deployment and flight testing of MIRVs/MRVs, of new types of missiles, and of missiles with improved accuracy or throw weight;
• —Flight tests of strategic offensive and ABM missiles would be limited to preannounced confidence firings of operational types of missiles on agreed ranges.
• —Flight testing of MIRVs, MRVs, MRV dispensing mechanisms (such as that for the SS–9), maneuvering RVs, post-boost maneuvering, endoatmospheric penetration aids, and fractional orbit bombardment systems would be prohibited. OSD and JCS, but not State/CIA/ACDA, would add to this list flight tests of exoatmospheric penetration aids and the testing of multiple payloads in space flights.

A reductions option could be based on either of the two options just discussed.

For example, the combined offensive launcher ceiling could be set at 1,500 or 1,600 rather than the 1,700 that would otherwise be called for thus requiring each side to eliminate some of its operational launchers, presumably the older ones.

A more ambitious proposal would involve an agreement to reduce the number of offensive launchers by 100 per year until a total of 1,000 ICBMs and SLBMs is reached.

1. Source: National Archives, Nixon Presidential Materials, NSC Files, NSC Institutional Files (H-Files), Box H–027, NSC Meeting 3/25/70 SALT. Top Secret. Kissinger included the paper as part of the President’s briefing materials for the NSC meeting scheduled for March 25.
2. A copy of this paper, March 21, is ibid., RG 59, Policy Planning Council Miscellaneous Records, 1969–72, SALT March 1970.
3. President Nixon submitted his first annual report on foreign policy to Congress on February 18. See footnote 2, Document 52.
4. The Verification Panel met on March 5, 12, 18, and 20. Minutes of these meetings are in National Archives, Nixon Presidential Materials, NSC Files, NSC Institutional Files (H-Files), H–107, Verification Panel Minutes Originals 1969–3/8/72.
5. The 19 radars include 7 Perimeter Acquisition Radars [PARs]and 12 Missile Site Radars [MSRs]—including the ones authorized for deployment at Grand Forks and Malmstrom and the MSR at Whiteman; the 633 interceptors include 465 long-range Spartans for area defense and 168 short range Sprints for radar defense. (An additional 246 Sprints are required for Minuteman and defense of the National Command Authorities in the Washington, D.C. area.) [Footnote and brackets are in the original.]
6. However, we may well want to add Safeguard Minuteman defense to the area defense, giving a total of about 900 launchers, if we freeze SS–9s at near their present levels. Adding Minuteman defense to area defense might then be an efficient way to preserve Minuteman until the advanced sea-based missile system, ULMS, is available in the early 1980’s and we can, in effect, move Minuteman to sea. [Footnote is in the original.]
7. All brackets are in the original. See footnote 2, Document 27.

8. You may want to point out the verification problems associated with a MIRV test ban. Both sides can now deploy their multiple warhead systems for retaliation against cities. However, we cannot verify the actual deployment of MIRVs, only their testing, without on-site inspection.

   Thus, a MIRV test ban may have an asymmetrical effect: Congress wouldn’t let us deploy our MIRVs in clear violation of the intent of an agreement, but we couldn’t be sure the Soviets had not gone ahead and deployed their SS–9 triplet.

   Also, we will have about 8 Polaris submarines being converted to the Poseidon/MIRV configuration; it would take two years or so to develop a single warhead for these Poseidon missiles. Thus a MIRV ban would leave up to 20% of our submarine force out of commission for a sustained period.

   If we attempted to get Congressional authorization to stockpile our already developed MIRVs but not actually deploy them, as a hedge against Soviet violations of an agreement, the Soviets might well take the opportunity to accuse us of bad faith before the agreement had a chance to work; in fact such a move on our part, because it would become public knowledge, might lead to some of the clandestine Soviet activities we fear.

   The issue is, if the Soviets deployed their SS–9 triplet, could they make enough improvements clandestinely so that, with a few quick tests following abrogation they would have a reliable anti-Minuteman system? We can make this possibility less worrisome by having the agreement permit the replacement of land-based missiles with sea-based missiles, “a point I will return to in a minute.” [Footnote is in the original.]

9. Many believe we should seek to limit the soft IR/MRBMs, but we should not get into concrete on this. If the Soviets insist on talking about NATO nuclear forces, we may want to drop discussion of soft IR/MRBMs as a quid pro quo for their dropping NATO forces. Our Allies wouldn’t be disturbed if this occurred. [Footnote is in the original.]
10. See Document 18.