From almost as soon as the term “strategic stability” first entered the nuclear lexicon, there have been calls to redefine it. During the Cold War, critics often advocated for a redefinition on the grounds that the quest for stability led to a nuclear policy that was at variance with effective deterrence. More recent arguments for reconceptualization—and even abandonment—tend to be based on the assertion that strategic stability premised on Cold War logic is about as relevant today as the challenge of defending the Fulda Gap from advancing Soviet armor.

Yet, for all the talk of redefining strategic stability, the reality is that its proponents have never actually been able to coalesce around a single definition (where, that is, they have chosen to define it at all). Edward Warner, who served as the U.S. Secretary of Defense’s representative to the New Strategic Arms Treaty (New START) talks, has observed that the term “strategic stability” is used in three broad ways:

  • Most narrowly, strategic stability describes the absence of incentives to use nuclear weapons first (crisis stability) and the absence of incentives to build up a nuclear force (arms race stability);
  • More broadly, it describes the absence of armed conflict between nuclear-armed states;
  • Most broadly, it describes a regional or global security environment in which states enjoy peaceful and harmonious relations.
James M. Acton
Acton holds the Jessica T. Mathews Chair and is co-director of the Nuclear Policy Program at the Carnegie Endowment for International Peace.
More >

Governments, in particular, are guilty of using the term “strategic stability” without definition or clear meaning. Of those that regularly invoke the phrase, the United States is the most consistent in usage, frequently employing “strategic stability” in the narrowest of the three senses listed above. The 2010 U.S. Nuclear Posture Review (NPR) Report, for instance, uses the terms “stable,” “stability,” and “instability” 49 times in the main text. Although these terms are not defined, most of these usages are associated with policies such as “deMIRVing” (multiple independently-targetable re-entry vehicles) the U.S. intercontinental ballistic missile (ICBM) force, which is advocated in order to increase “the stability of the nuclear balance by reducing the incentives for either side to strike first.”

Yet, occasionally the NPR report appears to impute broader meaning to stability. For instance, the Barack Obama administration supports ratification of the Comprehensive Test Ban Treaty partly because it “would enable us to encourage non-NPT [Nuclear Non-Proliferation Treaty] Parties to follow the lead of the NPT-recognized Nuclear Weapon States in formalizing a heretofore voluntary testing moratorium, and thus strengthen strategic stability by reducing the salience of nuclear weapons in those states’ national defense strategies.” Meanwhile, nuclear weapons are seen as playing a role in “promoting stability globally and in key regions.”

The Russian government has a tendency to use the term “strategic stability” like some form of diplomatic spackling paste, and what it means by the phrase is less than clear. Russian Foreign Minister Sergei Lav119 rov, for instance, has stated that Russia will not agree to further nuclear arms reductions unless all factors affecting strategic stability are addressed.8 Some of his concerns—such as ballistic missile defense and “non-nuclear strategic” weapons, which he fears could undermine the survivability of Russia’s nuclear forces—are consistent with a narrow understanding of strategic stability in terms of minimizing first use incentives. Yes, his invocation of imbalances in “conventional armaments and armed forces”9—a reference to the North Atlantic Treaty Organization’s (NATO) conventional superiority and possibly China’s too—implies a broader understanding.10 Along apparently similar (if somewhat opaque) lines, the 2010 Russian Military Doctrine identifies “the attempts to destabilize the situation in individual states and regions and to undermine strategic stability” as the second most significant military threat facing Russia. If anything, this usage appears to be consistent with the second of the three definitions listed above.

In an interesting twist, there is a debate in China over whether the concept of strategic stability is even applicable to the Sino-U.S. relationship. Much of the Chinese literature argues that “‘Balance’ (pingheng) and ‘symmetry’ (duicheng) are integral to the concept of strategic stability” and that, because of the current asymmetry in American and Chinese power, “the concept of strategic stability in classic arms control theory cannot be applied directly to the framework of Sino-U.S. relations.” Yet, this theoretical concern does not stop Chinese officials from regularly urging that “all disarmament measures should follow the guidelines of ‘promoting international strategic stability’ and ‘undiminished security for all’,” although exactly what is meant by “strategic stability” here is not entirely clear.  

The absence of an agreed definition for a term as widely used as “strategic stability” seriously detracts from the quality of debate on nuclear policy. Without an agreed definition of such a common term, arguments about the pros and cons of, say, ballistic missile defense or high-precision conventional weapons tend to be at cross purposes, creating much heat and little illumination. Indeed, critics tend to capitalize on this confusion, arguing that the term is ill-defined or setting up a weak definition as a straw man. Accordingly, it is useful to ask—as this chapter does—how the term should be defined. Definitions are, of course, to some extent arbitrary; strategic stability could be defined in all the ways Warner identifies above and more. But the principal criterion for how it should be defined ought to be conceptual clarity. A good definition for strategic stability might conceivably lead to more agreement on policy prescriptions, and, even if not, it might enable a better understanding of why we disagree.

Defining Crisis and Arms Race Stability

The theory of crisis stability was first expounded at length in Thomas Schelling’s 1960 masterpiece, Strategy of Conflict (although, as discussed elsewhere in this volume, others had explored some of the ideas previously). Schelling observed that, in a crisis, the fear of being pre-empted could itself create pressure to pre-empt. Specifically, because one side’s nuclear weapons could destroy an opponent’s, there might be real advantages to landing the first nuclear blow. In consequence, two states could be pushed over the brink of war because one state decided the risks of striking first outweighed the risks of waiting to be struck. It bears emphasizing from the outset that Schelling never argued that such dynamics were the only—or even perhaps the main—reason why states would go to nuclear war. Rather, his point was that if international relations were already severely strained—for whatever reason—the “reciprocal fear of surprise attack” might cause them to rupture entirely. Or, as he put it, “[a]rms and military organizations can hardly be considered the exclusively determining factors in international conflict, but neither can they be considered neutral.”17

Within Schelling’s conception of stability—the “traditional” conception, if you will—a crisis can be defined as stable if neither side has or perceives an incentive to use nuclear weapons first out of the fear that the other side is about to do so. This definition—the one I advocate in this chapter—is, in fact, even narrower than the first of Warner’s three definitions (under his definition any first use of a nuclear weapon—whatever the motivation—would be categorized as an instability). By analogy with crisis stability, my preferred definition of arms race stability is the absence of perceived or actual incentives to augment a nuclear force—qualitatively or quantitatively—out of the fear that in a crisis an opponent would gain a meaningful advantage by using nuclear weapons first.

The Cold War discourse on stability—crisis stability in particular—was overly narrow in two important ways. First, concern generally focused on the possibility of a state launching a large-scale damage-limiting first strike if it believed nuclear war had become imminent. However, such a strike is not the only—or even the most likely—response of a state fearful of an adversary’s using nuclear weapons first. An alternative would be the limited use of nuclear weapons in  an attempt to scare the opponent into backing down. This form of crisis instability seems much more likely than a large-scale damage-limiting first strike. Indeed, in a conflict against the United States, only Russia has anything approaching the capability to execute such a strike.

Second, the Cold War literature tended to focus on instabilities arising from the technical characteristics of each side’s strategic forces, that is, on first strike stability. These characteristics (the hardness of silos, the accuracy of missiles, the effect of missile interceptors, and so on) are only some of the factors that would play into a decision to pre-empt. First strike stability is, therefore, a necessary—but not sufficient—condition for crisis stability. The sometimes exclusive focus on the technical characteristics of strategic forces arose at least in part because these factors could be easily quantified, whereas other factors relevant to crisis stability—emotion, pressure, bad advice, miscalculation, misperception or poor communication—could not. During the Cold War, tremendous efforts were put into developing mathematical models for first strike stability. The aim was to quantify the incentives to strike first by modelling a nuclear exchange and using the results to determine whether either side was best served by waiting or attacking. These efforts certainly had some value; ensuring that vulnerable nuclear forces did not encourage pre-emption was important, and, for that matter, it still is. But, because the models narrowly focused on only some of the potential causes of nuclear war they were, as their developers sometimes acknowledged, limited and they attracted considerable and reasonable criticism. However, a number of contemporary analysts have gone further and used the inadequacies of such models to attack the entire concept of strategic stability. Such an attack is unreasonable; there is a clear difference between the concept of stability and specific, contingent mathematical models that try to quantify it. Using the inadequacies of such models to reject the former is like dismissing the concept of nuclear deterrence out of distaste for the game theoretic analysis that is beloved by a small number of formal deterrence strategists.

This is certainly not to argue that the technical characteristics of weapon systems are irrelevant to strategic stability. Indeed, the range of relevant systems has broadened as a result of technological change. Early in the Cold War, when stability concerns first emerged, the only plausible “defense” to a nuclear attack was a pre-emptive nuclear first strike. Today, however, there are concerns—particularly among potential U.S. adversaries—that high-precision conventional weapons, including cruise missiles as well as developmental weapons such as “boost-glide” systems, could be used to attack their nuclear forces before launch, while ballistic missile defenses could “mop up” any that survived and were launched. These fears—whether or not they are technically justified—could lead to anxieties about the possibility of a pre-emptive strike during a crisis. A full description of strategic stability in today’s world must, therefore, include conventional capabilities that could be used to destroy an opponent’s nuclear forces (or its ability to operate them).

Defining Strategic Stability

Strategic stability is usually defined as the combination of crisis stability and arms race stability. This definition suffers from the disadvantage of making crisis stability and arms race stability appear to be fundamentally different phenomena. In reality, they are actually two manifestations of the same phenomenon on very different timescales. To demonstrate this it is helpful to lay out the range of actions that could be taken by a state worried about being on the receiving end of a nuclear attack in a crisis, classified by the time required to implement them.

Seconds, Minutes, Hours, or Days: Use of Nuclear Weapons

While nuclear weapons have never been used out of the fear of an impending nuclear attack, war planning—especially in the United States and the Soviet Union—has certainly included pre-emptive options:

  • Soviet war planning during the 1960s was based exclusively on pre-empting an American attack—not least because the Union of Soviet Socialist Republics (USSR) was convinced that the United States was also planning to preempt. From the early 1970s, launch-on-warning and delayed retaliation options were developed, although pre-emption appears to have remained a possibility.
  • As in the Soviet case, early U.S. war planning was heavily based on pre-emption.Retaliatory options began to enter war planning in the late 1950s with the development of submarine-launched ballistic missiles (SLBMs). However, it seems unlikely that the United States has ever abandoned the option of pre-emption.

Hours or Days: Increasing the Alert Levels of Nuclear Forces

On a day-to-day basis, states generally keep substantial portions of their nuclear forces off alert. In the event of a crisis, they can—in hours or days—raise the alert level of some or all of their forces to enhance their survivability and ready them for possible use. Increasing alert levels might involve dispersing mobile forces—bombers, mobile missiles, and ballistic missile submarines (SSBNs)—or mating warheads with delivery systems, if the former are not emplaced on the latter as a matter of routine. While doing so may make forces more survivable, it can also send escalatory signals—which could be problematic if a decisionmaker does not want to send them. Moreover, the dispersal of mobile forces can increase the probability of an accidental or unauthorized launch, especially if dispersal is accompanied by a pre-delegation of launch authority, which may be seen to be necessary if technology or the military balance seems to require it.

There are a number of historical examples of states’ increasing their alert level in response to a perceived increase in the threat to their nuclear forces:

  • While it is widely assumed that Soviet nuclear forces were not alerted during the Cuban Missile Crisis, Pentagon documents suggest otherwise. Declassified documents from 1962 and 1963 indicate Soviet nuclear forces were alerted on October 24, 1962.27 Separately, an originally classified 1981 study of the Cold War arms race mentions an alert of Soviet submarines during the crisis.
  • On October 18, 1969, at the height of the Sino-Soviet border crisis, Chinese Defense Minister 126 Lin Biao ordered an alert of Chinese nuclear forces, fearing that a Soviet nuclear attack was imminent.
  • In August 1978, the United States raised the alert level at five Strategic Air Command bases—and then dispersed planes from those bases—after two Soviet SSBNs moved “dangerously close to the East Coast of the United States,” thus “significantly rais[ing] the threat” to those bases.
  • In November 1983, during NATO exercise Able Archer-83, the Soviet Union may have alerted some of its nuclear forces.

Months or Years: Deploying or Redeploying Existing Weapons

Outside of a crisis, a state concerned about the survivability of its forces—and hence its ability to use them in the event of a sudden crisis—could try to augment its forces relatively rapidly by deploying or redeploying existing weapons. There is at least one extremely notable example of such an action, which also demonstrates how this kind of instability can damage international relations:

  • Prior to 1962, Soviet Premier Nikita Khrushchev had been content with a “second-best strategic posture” consisting of a limited force of long-range nuclear weapons. In early 1962, he learned that the United States had examined a first strike plan during the previous summer when the Berlin Crisis was at its height. This appears to have “stirred fears that the Americans were eager to capitalize on their strategic advantage” and partly catalysed his decision to deploy two existing types of missile, SS-4 and SS-5, to Cuba.

Years: Building More Weapons and Developing New Systems

Over the course of a few years, a state that is afraid its forces could be vulnerable in a crisis can augment its arsenal by building more warheads and delivery systems (to new or existing designs). Such build-ups can be interpreted (rightly or wrongly) by an adversary as an aggressive action—that is, they can create a security dilemma—and thus carry the risk of exacerbating international tensions. These long-time scale dynamics—arms race instabilities—have plenty of historical precedents:

  • As noted above, survivability concerns appear to have been important in catalyzing Khrushchev’s decision to authorize a major build-up in long-range Soviet forces in 1962.
  • Every state with nuclear weapons has road-mobile ballistic missiles or SSBNs (or both). Most, if not all, of the programs to develop these systems were presumably motivated—in whole or in part—by the survivability advantages of mobile weapons, in light of their generally reduced accuracy and throw-weight.
  • Survivability concerns may be an important factor motivating China’s current build-up (although making definitive statements on the relative balance of strategic to bureaucratic considerations in Beijing’s decisionmaking is impossible in light of the information available).

This entire range of phenomena—stretching from the shortest timescales (crisis instabilities) to the longest (arms race instabilities)—can be captured within the following definition of strategic stability: A deterrence relationship is stable if neither party has or perceives an incentive to change its force posture out of concern that an adversary might use nuclear weapons first in a crisis. The focus on force posture (which includes but is not limited to use) emphasizes that responding to vulnerability lies at the root of instability. It cannot be emphasized enough that this definition does not imply that the only reason why a state might change its force posture is fear of an adversary’s using nuclear weapons first. Obviously, there are plenty of other reasons why a state might do so. Nonetheless, changes in force posture in response to the perceived threat of nuclear attack constitute a theoretically and historically significant class. The remainder of this chapter explores why it is conceptually helpful to reserve the term “strategic stability” for this particular class of phenomena.

Why Not Adopt a Much Broader Definition of Strategic Stability?

One of the most enduring criticisms of strategic stability is that the concept is too simplistic to account for conflict (or the absence thereof). In 1988, Stephen Prowse and Albert Wohlstetter—reflecting the views of many strategists—rejected the concept of strategic stability on the grounds that it is based on the misplaced belief that “the primary motive for one country to attack another springs simply from a misunderstanding that the other side might attack.” Looking back at the Cold War more than 2 decades later, David Yost, in a similar vein, has argued that “[t]he mutual vulnerability model that was supposed to simultaneously provide ‘crisis stability,’ ‘first strike stability,’ and ‘arms race stability’ was alluring and elegant, but based on false premises . . . about how decisions are made to go to war—as if force posture characteristics were the decisive factor.”

Looking at a “contemporary global security context [that] no longer bears any resemblance to the Cold War context,”contemporary critics have argued that any utility strategic stability offered then has now evaporated entirely. Frank Harvey, for instance, has written that:

Expanding levels of economic co-operation, interdependence, and, in Russia’s case, vulnerability have created an environment in which large-scale conflict involving those major powers is increasingly remote and, for many reasons, obsolete. Economic and trade relationships are far more useful than military competition in predicting interactions between the United States and Russia, and there is no compelling reason to expect this to change. Indeed, Russian officials are now more inclined to define strategic stability in terms of assured economic viability, not assured destruction. Survival of the Russian state depends less on the balance of nuclear forces and more on the Russian economy and foreign investment from the U.S., Europe, and Asia.

All these arguments are—on their own terms—true. It is obviously the case that the primary drivers of conflict are political (or cultural or economic or historical or ideological) and that, to the extent deterrence is immediately relevant in preventing conflict, it is probably conventional forces—rather than nuclear weapons—that act as a day-to-day restraint. But, advocatesof strategic stability never said otherwise. Their—or rather our argument is that if international relations become severely strained for any reason and the use of nuclear weapons becomes a realistic proposition, it would be highly desirable that none of the protagonists feels pressured into using weapons out of the fear that another might do so first. In other words, as Robert Jervis succinctly noted, “[c]risis instability can interact with political conflict; arms controllers never suggested that the former in the absence of the latter would yield war.”

If one accepts that crisis instability is only one potential pathway by which the nuclear threshold might be crossed and, accordingly, that it is only one criterion for assessing force posture (albeit an important one), then it can stand on its own as a useful strategic concept. Defined narrowly, it provides specific insight into an important, if only partial, aspect of nuclear deterrence. Needless to say, other concepts, not least the effectiveness of deterrence, are required to capture the totality of international strategic dynamics. Recognizing this, there is no need to attempt to broaden the concept of strategic stability, as some critics have called for, to try and embrace all factors relevant to the outbreak of war.

Why Not Adopt a Slightly Broader Definition?

Under the definition of strategic stability proposed here, the use of nuclear weapons for a reason other than fear of an impending nuclear attack would not be classed as a type of crisis instability. This definition, however, is controversial. The doctrines of conventionally weaker but nuclear-armed states or blocs generally call for the use of nuclear weapons in the event of a non-nuclear attack by a stronger neighbor (Russia’s defense doctrine vis-à-vis NATO and China is a case in point). This raises the question of whether the definition of stability should expand to classify any first use of a nuclear weapon—whatever the motivation—as an example of strategic instability (as in the first of Warner’s three definitions discussed in the introduction to this chapter).

It should, of course, be a goal of policy to create the political and security conditions that would minimize all incentives for the first use of nuclear weapons. However, that does not mean that it is conceptually helpful to include all possible incentives for first use within the definition of strategic stability. As explored in this section, mitigating different pathways to first use requires different approaches. Moreover, there may even be trade-offs—reducing the probability of first use for one reason could simultaneously increase the probability of first use for another. These distinctions and trade-offs would be obscured by a broader definition.

Effectively reducing the likelihood of nuclear first use depends on identifying why it might occur. If such use is most likely to result from fear of an adversary’s striking first, then the most effective methods to minimize the incentives to use would include increasing the survivability of nuclear forces, hardening command and control systems, enhancing early warning, and improving crisis communication channels. By contrast, reducing a weak state’s reliance on nuclear threats to deter non-nuclear aggression might require narrowing a conventional imbalance (whether through a conventional build-up or arms control), eliminating—or finding non-nuclear means to counter— chemical and biological weapons, or eschewing policies such as regime change that might lead weak states to proliferate. These differences in the tools required to mitigate the different pathways to first use are one reason not to broaden the concept of crisis instability to include all reasons for nuclear first use.

More theoretically, there is no a priori reason why reducing one motivation for using nuclear weapons first will simultaneously reduce another. Indeed, some strategists, particularly in the United States, argue—for a variety of different reasons—that, in order to enhance deterrence, it is actually desirable for U.S. adversaries to worry that Washington might attempt a disarming first strike. They argue that the risks of this strategy—including increased pre-emptive pressures on the adversary—are outweighed by the benefits of enhanced deterrence, which include the reduced likelihood of an adversary’s using nuclear weapons first to try to coerce the United States or of the United States being forced into using nuclear weapons to respond to non-nuclear aggression.

For instance, some contemporary strategists observe that, fearing regime change, the leader of a state facing conventional defeat by the United States might use nuclear weapons in a last desperate attempt to make Washington back down. They argue that in order to deter the use of nuclear weapons in this scenario, Washington needs to be able to eliminate—completely or almost completely—the adversary’s nuclear forces. Advocates acknowledge that such a strategy might “exacerbate the problem of controlling escalation if an adversary feels so threatened that it adopts a hair-trigger nuclear doctrine” but argue that, on balance, “the benefits . . . trump the costs.”In other words, they posit that their strategy will have the net effect of reducing the probability of nuclear use by reducing the likelihood of a U.S. adversary’s using nuclear weapons for coercive purposes, even if there is an increased chance it will employ them because of pre-emptive pressures.

Similar issues were debated during the Cold War. A fundamental strategy debate—perhaps the fundamental debate—of that era centered on the question of whether deterrence would be enhanced if the United States had the ability to launch a damage-limiting strike against the Soviet Union. Proponents of damage limitation argued that if the Soviet Union believed the United States could emerge relatively unscathed from a nuclear war, Moscow would be less inclined to undertake conventional aggression in Europe. These strategists were willing to accept what they believed would be a small increase in the probability that Moscow would use nuclear weapons to pre-empt a U.S. first strike in return for a more substantial decrease in the probability that the United States would have to resort to the use of nuclear weapons to counter Soviet conventional aggression in Europe, because war would be less likely to break out in the first place.

Even strategists who oppose “warfighting” doctrines sometimes see similar—if much less severe—potential trade-offs. Many such strategists view the credibility of nuclear threats as resting on the “threat that leaves something to chance,” that is, the possibility of unintended escalation from lower levels of violence to nuclear use.46 In a crisis, the steps that a U.S. adversary might take—such as dispersing mobile missiles or sending submarines out to sea—may be stabilizing on balance (since they significantly enhance force survivability), but they can simultaneously increase the chance of unintended escalation through miscalcu134 lation, accident or unauthorized launch. Thus, Jervis has argued that “if security is linked in part to the danger of inadvertent war, then too much stability could make the world safe for coercion and violence” (although he does go on to add that, in practice, “it is doubtful that arms control could succeed too well and produce arrangements that would drive the danger of undesired escalation close to zero.”) For present purposes, it is unnecessary to critique these arguments and reach a conclusion about whether inducing fear in an opponent that the United States might use nuclear weapons first does, in fact, enhance deterrence and hence lessen the net probability of nuclear use; it is enough to note the existence of important arguments for this proposition. It is precisely because such arguments are made that it is most advantageous to define crisis stability (and, by extension, strategic stability) in the narrow way advocated above. This definition enables a clearly delineated debate about whether there is, in fact, a trade-off between crisis stability and the effectiveness of deterrence and, if there is, what the optimal balance should be. By contrast, using “strategic stability” as a catch-all term that tries to capture every possible motive for using nuclear weapons first tends to obscure the reasons why strategists disagree and thus confuses the debate.

Does It Matter If An Adversary Does Not Share the U.S. Concept of Strategic Stability?

Yet another criticism of the narrower understanding of strategic stability is that it is only a useful tool for policymakers if other states share the U.S. conception. In practice, they rarely do. This argument hasrecently been advanced by David Yost with regard to the U.S.-Soviet relationship during the Cold War. Central to Yost’s argument is a considerable body of evidence that “the Soviet political-military leadership appears to have rejected the ‘mutual assured destruction’ reasoning advanced by Robert McNamara and his followers as the desirable foundation of strategic stability, including ‘crisis stability’ and ‘arms race stability’.” The salient question here is whether this rejection actually nullified the utility of U.S. policies designed to promote stability.

In an attack on the concept of arms race stability, Yost observes that the Soviet nuclear arms build-up of the 1970s “did not conform to U.S. ‘arms race stability’ theories” and “gave many American observers the impression that the USSR was seeking superiority.” Soviet behavior was indeed confounding to some in the United States. A number of American analysts originally assumed that the primary—if not the sole—reason for the Soviet build-up was enhancing the survivability of its nuclear forces. Given that the drivers for Soviet procurement were actually more complex, it was inevitable that these analysts would be disappointed when the Soviet Union did not terminate its build-up upon acquiring a survivable second strike capability. However, Yost’s observation says much more about a lack of imagination on the part of American analysts than it does about any deficiencies with strategic stability. As noted above, there are plenty of reasons why a state might build up its forces besides fear of an adversary’s first strike and nothing in strategic stability “theory” says otherwise (a point that some over-enthusiastic advocates may have forgotten). Achieving an assured second-strike capability was a necessary condition for the Soviet Union to cease its arms build-up but certainly not a sufficient one. Accordingly, Soviet behavior in the 1970s is not a valid reason for rejecting the concept of arms race stability.

Ultimately, whether states stop an arms build-up after achieving an assured second-strike capability is not really a fair test of the usefulness of arms race stability. A better test is to examine whether states that fear for the survivability of their forces start an arms build-up. The U.S.-Soviet experience from the Cold War certainly meets this criterion. As noted above, a major factor in precipitating the Soviet long-range arms build-up appears to have been concern in Moscow that, at the height of the Berlin Crisis in 1961 and during the Cuban Missile Crisis in 1962, the vulnerability of its forces had proved a significant disadvantage.

The motivation of the Soviet Union in continuing to augment its nuclear forces long after it had achieved a credible second-strike capability forms the basis for Yost’s critique of crisis stability. Yost points to considerable evidence that the Soviet Union adopted a warfighting doctrine, which, he argues, led it to seek superiority in order to limit the damage it would suffer in a nuclear war. By contrast, he claims that the United States “at times exercised restraint” in developing equivalent counterforce capabilities, although, strangely, he gives just a single minor example of such restraint and glosses over continual and relentless improvements in U.S. missile accuracy that far outstripped the Soviet Union. In light of this disparity, Yost argues that a “shared commitment to a theory of ‘crisis stability’” cannot explain the absence of conflict. Instead, he attributes it to the bipolar structure of the Cold War international order and the “profound fear of nuclear war” that the superpowers shared.

Contrary to Yost’s explanation, however, it is precisely because the Soviet Union had a warfighting doctrine that it was important to ensure crisis stability during the Cold War. As Yost himself observes, the Soviet Union was prone, rightly or wrongly, to project its own dedication to nuclear warfighting onto the United States. Had the Soviet Union come to believe that the United States was about to strike then—according to Yost’s own interpretation of Soviet doctrine—it may have tried to pre-empt such an attack. In a crisis, therefore, it was critically important for Moscow to believe that Washington thought it could not meaningfully lessen the horror of a nuclear war by striking first. Thus, crisis stability is not inconsistent with Yost’s explanation that the Cold War did not turn hot because each side had a profound fear of nuclear war. On the contrary, it is precisely because there generally was a sufficient degree of crisis stability that this fear was able to play a restraining role.

Of course, to the extent that the Soviet Union did not accept mutual vulnerability as a policy goal and sought to attain superiority, it was, as a practical matter, harder for Moscow and Washington to agree upon bilateral measures to enhance stability. However, such measures were negotiated—most notably the Anti-Ballistic Missile Treaty and START I—and they played a positive if modest role in enhancing strategic stability, even if the Soviet Union was motivated to agree to them for other reasons. Moreover, the Soviet Union’s failure to accept mutual vulnerability certainly did not stop the United States from taking unilateral steps to enhance strategic stability, such as developing SSBNs (which, until the final years of the Cold War, were particularly stabilizing because they were too inaccurate to threaten an adversary’s strategic forces). In fact, the development and procurement of survivable nuclear forces—unilateral decisions originally taken outside of an arms control framework—did more than anything else to ensure mutual vulnerability and hence crisis stability during the Cold War.


It might have been better if strategic stability had an alternative, less grandiose name (“deterrence stability” springs to mind as one alternative). The very words “strategic stability” give the impression of a broad concept that pretends to predict whether and how states can enjoy stable relations. In reality, however, strategic stability is most useful if it is narrowly defined—in terms of whether fear of an adversary’s using nuclear weapons motivates a state to change its force posture—and modestly applied, that is, with the recognition that it is one—and not the only—criterion against which to assess nuclear policy.

While fear of an adversary’s first strike has never led to nuclear use, it has led states to change their force postures in sometimes dangerous ways, whether by dispersing mobile forces, redeploying existing systems or developing entirely new ones. None of these actions have been cost free, not least because they have sometimes exacerbated international tensions and created new risks of further escalation. Reducing similar pressures on states in the future—that is, ensuring and enhancing strategic stability—remains a worthwhile, and in fact a vital policy goal.

That said, it is not the only relevant goal. Nuclear strategy must also be assessed along other axes—deterrence effectiveness, cost effectiveness, bureaucratic feasibility, domestic politics, and alliance politics to name but five—and we should certainly not assume a priori that the policy that maximizes strategic stability will simultaneously maximize all—or even any—of the other variables. Crafting the optimal nuclear strategy almost certainly involves trade-offs and it is by defining strategic stability most narrowly that we are most likely to set up a sensible debate about what those trade-offs should be.

This chapter was originally published in Strategic Stability: Contending Interpretations by the Strategic Studies Institute.