Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 The Pu rpose and Effe c! of Deep Strategic Force Reductions John D. Steinbruner My topic is a discussion of the calculations of the National Academy of Sciences' Committee on International Security and Arms Control regarding the implications of strategic force reductions. T will emphasize the results of the calculations rather than their internal details. In assessing the effect of reducing force balances, the first step, of course, is to determine the purposes that are to be achieved. We made the judgment that the primary objective in seeking force reductions should be that of constraining the capacity, and therefore the inclination, to undertake an effective preemptive attack on the opponent's strategic forces. Although it has not been as explicitly defined as one might wish, we believe there is an implicit consensus in support of this objective. The capacity for preemption is not considered a legitimate security requirement under current international understanding. However they might grumble about it, both the United States and the Soviet Union in fact concede to each other the legitimacy of having a deterrent force that promises an effective retaliation to an attack, but 11
12 REYKJA VIK AND BEYOND they certainly deny the legitimacy of having a preemptive attack capability that could remove the deterrent requirement. If there is a level of forces that would guarantee the ability to perform deterrent missions but preclude the capacity to threaten the opponent's deterrent force, presumably that would be the basis for stable agreement. Our calculations attempt to define such a force level. To do that, one must judge the deterrent requirement. Dr. Wolfgang Panofsky provided an outline of the logic that has been used in the past to define this requirement. Let me focus on the high and low ends of prevailing opinion in that regard and try to Rive a ~ r . ~ .~ 0 ~ ~ loose quantitative Fenton ot the range, while recognizing; that there are some who would put the requirements tor deterrence outside of these low and high ends of what I believe to be the . . . . . . preva1 1ng mainstream opinions on t 11S matter. The high sidethat is, the more demanding notion of what deterrence requires is cleterminec! by the prevailing doctrine insicle the military organizations that actually target nuclear weapons. Their idea is that effective retaliation must attack the opposing military infrastructure and essentially remove its capabilities. That is a fairly demanding requirement in terms of the number of weapons and the locations in which they must be put. On the low side of the current range of opinion, there is the notion that a capacity to destroy the social infrastructure of the opponent would be enough to deter war. The implicit requirement in this case is the ability to attack urban-in(lustrial areas, causing tens of millions of deaths and destroying some 75 percent of basic industrial facilities. In order to associate rough numbers with these theories, our calculations estimated the array of military and economic targets that actually exist in the United States and the Soviet Union. We considered 6,000 of those targets because our initial purpose was to consider, roughly, the effects of reducing forces to 6,000 warheads. We acloptec} the simpleminded notion that each of those weapons should have a target associates] with it, ant! we imagined the targets that would be assigned by a Soviet or an American military commander. These estimates reveal the phenomenon that Dr. Panofsky men-
13 DEEP STRATEGIC FORCE REDUCTIONS tioned. Beyond a certain level of forces and it falls well below 6,000 the targets get smaller and smaller and less and less significant in terms of their contribution to the cumulative level of damage. This is the economist's familiar notion of diminishing margin of . . Utl sty. With that effect in mind, we defined the idea of an efficient deterrent requirement. An efficient deterrent requirement is limiter! by the diminishing significance of attacking additional targets after a certain number have been destroyed. Beyond that point a prudent military commander should prefer to save those weapons rather than use them because the marginal damage they would inflict would not justify expending the weapons. We tried to define that number using the more demanding criterion of retaliation against the residual military capability of the opponent. We examined the infrastructure of conventional and strategic forces and military-support industries in the United States and the Soviet Union and determined that beyond the first 2,000-3,000 targets, additional attacks would be of so little marginal significance that a prudent commander would not continue the attack beyond those levels. So much of the military and social infrastructure would have been destroyed that additional destruction would not have sufficient military effect to justify the expenditure of the weapons. We assessed the effect of an attack on the approximately 2,000 highest priority military and economic infrastructure targets using standard calculations to estimate civilian fatalities not casualties but rather immediate, prompt deaths from such an attack. We found that a successful retaliatory attack on these 2, 000 military ant! economic targets would kill 20 million-40 million people in prompt blast damage effects in the United States and 30 million-50 million in the Soviet Union. As it turns out, those same ranges for prompt civilian fatalities appear if one assesses what wouic! happen if the attack were directed to urban-industrial areas using a much smaller number of weapons. An attack with a 100-equivalent-megaton total yield delivered by a few hundred weapons directed against the urban- industrial areas of the United States or the Soviet Union would produce prompt fatalities from blast and thermal pulse effects in these ranges as well. Therefore, as Dr. Panofsky pointed out, the two ends of the
14 REYKJA VIK AND BEYOND spectrum ot opinion about deterrent requirements do agree on that number. If retaliation is directed against urban-industrial areas, according to the minimum deterrence theory, or if retaliation is directed against the larger target set required to destroy the residual military power of the enemy, a total of 50 million-90 million people will die in the two countries; and, of course, the casualties and social disruption would be some multiple of that figure. In sum, then, 500-2,000 warheads delivered in retaliation covers anything that might be considered a reasonable deterrent requirement uncler any of the prevailing opinions about that requirement. With that thought, we went on to assess how forces might be reduced in a way that allowed this core deterrent requirement to be maintained by both sides while eliminating the capacity of either one to attack the forces that would conduct this retaliation. We examined seven ways in which current U. S. and Soviet forces could be reduced to the level of 6,000 warheads. A number of very simple observations emerged from that exercise. First, no matter how forces might be reduced to the 6,000-warhead level, each side can cover the estimated efficient deterrent require- ments either in retaliation or under a launch-on-warning strategy. Even in the worst case, in which the victim of an initial attack waits until the attack is completer! before beginning to retaliate, a 6,000- warhead force wouIc! stir] be able to cover these requirements. There is not much pressure, then, on the judgments made in reducing forces to the 6,000-warhead level. However it is done, both sicles can be confident that they can meet deterrent requirements. The main effect of different patterns of reductions concerns the residual forces that would remain after both sicles had completed their initial attacks. The range of models we considered includes one in which we preserver! all the multiple-warhead, high-accuracy systems on both sicles, thereby maximizing preemptive capability. According to the central criterion we set to guide the analysis, that pattern of reduction would be the most perverse case. Even under that case, both sides can meet their deterrent require- ments, but there would be some theoretical incentive for preemption in the sense that residual forces, after meeting those requirements, wouIc3 not be even. The side that went first would have more
15 DEEP S TRA TEGIC FORCE RED UCTIONS weapons left after the exchange than the side that went second, and military organizations might care about such a thing, although it is doubtful that the public would. Because military organizations are likely to be the groups respon- sible for assessing the possible patterns of force reductions, the reductions that are more likely to be implemented involve the removal of those weapons that are most easily or most likely to be used for a preemptive attack: notably, the SS-~S on the Soviet side and the MX, and potentially the Trident 2 missile, on our side. If the weapons were eliminates! in reducing strategic forces to the 6,000-warhead level, not only would both sides be able to meet efficient deterrent requirements but they would also expect to have equal forces after the initial exchange. Under this arrangement it would not matter who goes first and who goes second; both sides would in the end be about the same, which is very badly off indeed. We next tester! the effect of reducing strategic forces another 50 percent to the level of 3,000 warheads on each side. We wanted to determine whether the capacity to cover the full range of deterrent targets at these yet lower force levels would depend on questionable values for the operational assumptions used in the calculations. We judged that, if the robustness of the result did depend on Input assumptions that were subject to plausible disagreement, it would be quite difficult for force reductions of this magnitude to achieve the political support necessary to carry them out. The operational assumptions in question concerned alert rates for bombers and submarines, the time required for bombers to fly out of the zone of vulnerability around their base areas, and the warning they wouIc3 enjoy in (loin" so, as well as the standard characteristics of weapons- their reliability, yield, and accuracy. We also assumes! that all the forces of the attacker would be fully prepared whereas the victim would have to respond with forces available under daily conditions. That assumption admittedly was unrealistically pessimistic from the victim's perspective, but it was our belief that if the victim could retaliate despite that pessimism, deterrence should be particularly secure. Our results indicated that strategic forces of 3,000 warheads for each side would support efficient deterrent requirements quite well, provided sufficient investment was made in protective measures. In
16 REYKJA VIK AND BEYOND effect this means substantial reliance on less vulnerable basing modes and on higher alert rates for bombers and submarines; both sides would have to be willing to depend on alert procedures. With that provision, deterrence appeared to be perfectly safe down to the level of 3,000 warheads, even under the more demanding theory of deterrence. For those who believe that the threat of substantial urban-industrial damage is sufficient for deterrence, forces even smaller than 3,000 would be acceptable. At that point, however, a heated domestic debate about deterrent requirements is likely to be triggered! because adopting the less demanding theory wouIc! require a change in prevailing military practice. One need not engage in that debate to undertake force reductions to the level of 3,000 warheads; reductions to that level would simply remove excess capacity. We then asked whether these reductions, if accomplishecI, wouIc3 remove the existing incentives for preemption and thereby establish a more robust security regime than the current one. The answer, basically, is that the reductions wouIc! bring significant improvements but that they are certainly not sufficient to establish an entirely stable . . . . International security arrangement. There are several basic reasons why this is true, and they have to do with incentives for preemption that would not be affected or at least would not be resolved by the reductions just describecl. Let me briefly array those ant! then mention the arms control arrangements that would be designed to meet them. First, the calculations I just summarized assume throughout that the respective command systems would operate efficiently in retal- iation under the heavy damage we are assuming they would expe- rience. This is a heroic assumption, however, and is subject to question. Unfortunately, severe disruption of a command system- though probably not absolute destruction can be accomplished with less than 500 weapons maybe less than 200. Even 50 weapons placed in the right places wouIc! slow the functions of the command system significantly. This means that Towering force levels wouicl not relieve the command systems from the sense of pressure they currently face, a pressure that is probably the single most important and dangerous preemptive incentive in existing forces. The relief of that pressure would not be accomplished by the force reductions I
17 DEEP STRATEGIC FORCE REDUCTIONS have just described, nor would they ameliorate the problem of the troublesome engagement of short-range nuclear weapons and con- ventional weapons, notably in the central front ant! Europe. Such engagement is considered by many to be a trigger of potentially uncontrollable interaction that would not be resolved by strategic force reductions. It is also true that the conventional balance in Europe at the moment creates pressures for preemption quite apart from the use of nuclear weapons. At least as perceived by the members of the North AtIantic Treaty Organization (NATO), Soviet forces in Europe are committed to a rapid invasion of Western Europe at the outset of war to prevent an eventual defeat imposer! by superior Western economic power. The NATO perception and the Soviet doctrinal commitment produce an unstable situation that would appear to become all the more volatile as the discipline imposed by nuclear weapons is limited. Moreover, current trends in weapons technology, if projected long enough and optimistically enough from the point of view of what the technology can accomplish, do look as if they will eventually enhance the Tong-range preemptive capability of even conventional forces against retaliatory nuclear forces. These trends have largely to do with the application of remote sensing, information processing, and precision guidance technology. As a result, to achieve a sufficient answer to the problem of preemption, force level reductions must be accompanied by restraint of the pace of technology or the pace of the application of technology to military hardware. Finally, let me mention a topic that Spurgeon Keeny will explore In more detail. Tfpartially effective defenses were deployed as strategic offensive forces were being reduced, they conic} raise havoc with the deterrent capacities of those forces by denying systematic target coverage in retaliation. The effect of that denial would be to reintroduce the problem of preemption, perhaps in more virulent form. For that reason, the continuation of restraints on defensive deployments and some agreed regulation of military activities in space will almost certainly be a necessary condition for strategic force reductions. In sum, then, strategic force reductions are a promising means of protecting deterrence and controlling the pressures for preemption, 1 -
18 REYKJA VIK AND BEYOND but they must be combined with other measures to be fully effective: notably, with the removal of forward-basec! nuclear weapons in the European theater, with arrangements for stabilizing the conventional balance in that theater, with restraints on the pace and scale of force modernization, and with mutually agreed-upon restrictions on stra- tegic defense. Such a "full package" is what is required to produce more stable . . . international security arrangements.
