Is Crisis Stability Still Achievable?

Based on talk given at the APS March meeting, New Orleans, March 16, 2017

Joshua H. Pollack, James Martin Center for Nonproliferation Studies (CNS), Middlebury Institute of International Studies at Monterey (MIIS)

The Cold War Origins of Crisis Stability

“Stability” is a central concept, or family of concepts, in the analysis of nuclear posture. The Obama administration’s 2010 Nuclear Posture Review, for example, described “the challenge of ensuring strategic stability” between the United States and Russia, and between the United States and China, to be a “familiar” problem, to be addressed alongside more pressing concerns, particularly the dangers of nuclear proliferation and nuclear terrorism.¹ In the intervening years, proliferation and terrorism threats have arguably become less acute.² The problems of stability appear to have grown.

Strategic stability is traditionally understood in terms of two concepts: “crisis stability,” also known as “first-strike stability,” and “arms-race stability.” Broader applications have also been suggested, but these two basic ideas have persisted.³ The metaphor of stability appears to have arisen from economics, in particular from game theory.⁴

The locus classicus for crisis stability may be a paper written in 1958 by the economist Thomas Schelling, “The Reciprocal Fear of Surprise Attack,” which later appeared as a chapter in his celebrated book The Strategic of Conflict (1960). It hypothesizes a dynamic process of compounding fears of an enemy’s attack that place great pressure on each of two parties in a confrontation to strike the first blow themselves, even if they would prefer no violence at all:

This is the problem of surprise attack. If surprise carries an advantage, it is worth while [sic] to avert it by striking first. Fear that the other may be about to strike in the mistaken belief that we are about to strike gives us a motive for striking, and so justifies the other’s motive. But if the gains from even successful surprise are less desired than no war at all, there is no “fundamental” basis for an attack by each side. Nevertheless, it look as though a modest temptation on each side to sneak in the first place — a temptation too small by itself to motivate an attack — might become compounded through a process of interacting expectations, with additional motive for attack being produced by successive cycles of “He thinks we think he thinks we think … he think we think he’ll attack; so he thinks we will; so he will; so we must.”⁵

In the realm of nuclear weapons, the United States and Soviet Union took measures that militated against this danger, both unilaterally and cooperatively. Unilaterally, each reduced the other side’s ability to act in “preclusive self-defense” (Schelling’s phrase) by making their own weapons difficult to destroy (through “hardening” in concrete silos) or difficult to locate (through forms of mobility and stealth). Analogous measures have been taken to protect national decision-makers during a crisis, involving either underground bunkers or aircraft. Cooperatively, the superpowers reached arms control agreements to ensure crisis communications between leaders, limit the numbers and types of nuclear weapons and delivery systems on either side, and provide each side with “transparency,” meaning access to information about the basing and composition of the other side’s forces, and therefore greater confidence in the other side’s compliance with agreements.⁶

At the same time, because military establishments are responsible for preparing to fight wars, they have tended to take steps that would improve their ability to destroy enemy forces, e.g., improved accuracy, which contributes to instability by improving the chances of a successful first strike, ceteris paribus. This problem reflects the enduring tension between posture choices designed to reduce the level of harm to one’s own country in the event of war (“damage limitation”) and posture choices designed to reduce the chance of an unthinkably destructive war.

Even some measures that are stabilizing by intention may perversely have the opposite effect. Preparing to launch vulnerable missiles from silos upon receiving warning of an inbound attack should discourage an adversary from contemplating such an act, but also creates a risk that a false warning will trigger a nuclear war.

These observations may provide some sense of the complexities of ensuring crisis stability; each new technological development or modernization of the arsenal may bring about a change in the calculations of each side. To compound the difficulty further, the two sides’ calculations are not identical, something that is perhaps only to be expected in a situation fundamentally premised upon mutual mistrust. Crisis stability, as discussed here, is American in its origins and development, although the two sides have had ample chances to exchange views in negotiations. This process began with inconclusive talks in the late 1950s about how to address the problem of surprise attack, and returned with a vengeance after the 1962 Cuban Missile Crisis. Russian-American exchanges continue in one form or another to this day.

Three Post-Cold War Complications

Despite the various complications discussed above, the reciprocal fear of surprise attack is, at its core, a simple and elegant construct, involving two actors and one type of arms. Schelling’s 1958 paper famously opens with an analogy: a homeowner and a burglar have come upon one another, guns drawn. Both might prefer that the burglar simply withdraw, but they are caught in a down spiral of mutual fears that may well lead to violence. Although this was always a simplification, in the minds of many analysts, it distilled the essence of Soviet-American nuclear confrontation and the problem of crisis stability. For at least three reasons, the same simplicity no longer prevails.

First, new non-nuclear military technologies have become “entangled” with nuclear arms and with each other. These include ballistic missile defenses (BMD), counter-space weapons, and strategic conventional weapons.⁷ For example, American BMD architecture relies heavily on space-based sensors. Any adversary that is concerned that BMD may enable American nuclear threats or nuclear attack by negating its own ability to retaliate may be tempted in a crisis to use counter-space weapons to “blind” BMD. Furthermore, nuclear-armed states are developing dual-use missile systems — both nuclear and conventional — or dedicated intercontinental-range conventional weapons (so-called “conventional prompt global strike,” or CPGS systems) capable of precise strikes. The possession of these arms may create a temptation to try to disable an enemy’s counter-space weapons; indeed, this role has been one of the most plausible and oft-cited justifications for acquiring CPGS in the United States. Other roles may include a “defense-suppression” mission, targeting the enemy’s long-range radars, which enable BMD, air defense, and coastal defense.⁸ Notably, space-based sensors and long-range radars are also the mainstays of “early warning” against nuclear attack. Attacks on these systems could be interpreted as a prelude to a nuclear first strike. So, too, could attacks on multi-purpose command-and-control systems. Furthermore, attacks on dual-use missile systems could be interpreted as targeting a state’s nuclear-delivery capabilities. Conventional weapons might even be employed purposefully against nuclear targets.

Second, it is no longer useful to model the nuclear-weapons “environment” as a system composed of two nodes. As additional states with mistrustful relations have developed nuclear weapons and long-range delivery systems, several new “dyads” have emerged, beyond (1) America/Russia. These include, at a bare minimum, (2) America/China, (3) America/North Korea, (4) India/Pakistan, and (5) India/China. (I will omit Britain and France, while acknowledging that a Russian analyst probably would not do so.)⁹ Fortunately, it is still reasonable to model these five dyads as parallel systems, and not as fully enmeshed with each other. Unfortunately, two nodes (America, India) are linked to more than one other node. These corresponding nodes (Russia, China, and North Korea, in the American case; Pakistan and China in the Indian case) are dissimilar from each other, so whatever decisions that America or India make on strategic posture may simultaneously have a variety of effects in the calculations of potential adversaries.

Third, even if adversarial dyads may be modeled as discrete links between pairs of nodes, three-sided interactions are already a fact of life. These interactions may involve three nuclear-armed states, or two nuclear-armed states and a third state with its own “entangling” weapons systems.

An example of the first type of triangular interaction is America/North Korea/Russia. The United States has deployed the Ground-based Midcourse Defense (GMD) system against the threat of North Korean intercontinental ballistic missiles (ICBMs). GMD’s kinetic-kill interceptors are based primarily in Alaska, with a handful in California. Any ICBM flying out from North Korea toward the continental United States would pass over the Russian Far East. To attempt an intercept of a single North Korean ICBM, four or five of the Ground-Based Interceptors (GBIs) would have to fly out in the direction of Russian airspace. Depending on the details, the intercept engagement could take place over Russia; regardless, the excess interceptors would either overfly Russia or reenter the atmosphere inside Russian airspace. Russia’s early warning radars presumably would detect the inbound interceptors. While the nature of the event would hopefully be clear to the Russian military, there is already ample precedent for false warnings.

An example of the second type of triangular interaction involves the United States, North Korea, and South Korea. North Korea has threatened to use its nuclear missiles against the ports and airfields in South Korea that would receive American reinforcements in the event of war. American defense officials have spoken of developing “left of launch” capabilities to deal with threats of this type, presumably meaning “conventional counterforce” options, among other things. South Korea does not have its own nuclear weapons, but it has developed an arsenal of conventionally armed ballistic and cruise missiles capable of striking anywhere in North Korea. South Korea aspires to develop and deploy what it calls a “Kill Chain” system: a network of sensors that will permit its missile forces to attack North Korea’s nuclear missiles before they can launch against South Korea. In short, North Korea has a strategy of using nuclear weapons first, and both the United States and South Korea are separately pursuing capabilities to preemptively attack North Korea’s weapons. Arguably, South Korea’s conventional first-strike option will mainly have the role of pushing the United States toward conducting its own first strike against North Korea, in the hopes that it would be more effective than South Korea’s.

To make matters still more complex, calls have been heard in Japan to develop conventional counterforce options there as well. This raises the prospect of a four-way interactions involving two nuclear-armed parties and two non-nuclear-armed parties.

In summary, over the span of the last two decades or so, we have moved from a frightening but relatively manageable picture involving two states and their nuclear arsenals to a very hairy situation involving a variety of interacting weapons systems distributed across at least five dyads, with hydra-headed interactions starting to emerge.

Potential Responses

Short of cutting the Gordian Knot and implementing global nuclear disarmament, which has eluded humanity for the last seven decades, what is to be done? Here, we might think back to Cold War efforts at shoring up crisis stability, both unilateral and cooperative.

Unilateral measures might include voluntarily forgoing classes of new weapons, if they are judged too destabilizing to warrant deployment. This suggestion has already been made concerning CPGS, for example. Another possibility is to diversify military sensors and communication networks, to minimize nuclear-conventional overlap, to make these systems highly redundant, or to do both. This is probably an expensive proposition, but it may be worthwhile.

Cooperative measures may be more difficult to achieve in a multilateral setting than in a bilateral one, but they are still possible, as the success of several multilateral arms control and nonproliferation regimes demonstrates. Multilateral treaties to forgo certain classes of weapons, or even multilateral codes of conduct to forgo certain behaviors may have great value. The importance of orbital sensors and communications platforms makes outer space a natural place to focus efforts.

These are initial ideas only, but they point to a research agenda: how do we model crisis stability as has evolved, and continues to evolve? Are there key nodes in this system? Are they amenable to policy responses? As daunting as these problems may seem, this type of work has been our collective answer to the problems of survival since the start of the nuclear age, and it remains, in all likelihood, our best response.

Jason H. Pollack
jpollack@miis.edu

Endnotes

1 Department of Defense, Nuclear Posture Review Report, April 2010, pp. v, 5-6, 28-29.

2  Reductions in these two risks might be credited to the negotiation of the Joint Comprehensive Plan of Action (JCPOA) with Iran and the Nuclear Security Summit process, respectively. On the diminishment of the proliferation threat, see: Leonard S. Spector, “A Proliferation Plateau May Offer Unique Opportunities,” Arms Control Today, April 2016.

3  James M. Acton, “Reclaiming Strategic Stability,” in Elbridge A. Colby and Michael S. Gerson, eds., Strategic Stability: Contending Interpretations (Carlisle, PA: U.S. Army War College Press, 2013), pp. 118-146.

4  This theme is examined in depth in Robert Ayson, Thomas Schelling and the Nuclear Age: Strategy and Social Science (New York: Frank Cass, 2004).

5  T.C. Schelling, “The Reciprocal Fear of Surprise Attack,” RAND paper P-1342, April 16, 1958, revised May 28, 1958, p. 1.

6  On this theme, see, notably: Kerry M. Kartchner, Negotiating START: The Quest for Stability and the Making of the Strategic Arms Reduction Treaty (New Brunswick, NJ: Transaction Publishers, 1991).

7  Joshua Pollack, “Emerging strategic dilemmas in U.S.-Chinese relations,” Bulletin of the Atomic Scientists, vol. 65, no. 4, July-August 2009, pp. 53-63.

8  James M. Acton, Silver Bullet? Asking the Right Questions about Conventional Prompt Global Strike (Washington, DC: Carnegie Endowment for International Peace, 2013), pp. 17-21.

9 China tested its first nuclear device in 1964, but did not deploy ICBMs optimized for targeting the United States until the mid- 1990s. India and Pakistan tested nuclear weapons in 1998. North Korea tested its first nuclear device in 2006; its intercontinental delivery capabilities are still rudimentary, but the testing of more sophisticated ICBMs is widely anticipated.

---

These contributions have not been peer-refereed. They represent solely the view(s) of the author(s) and not necessarily the view of APS.