Conventional War and Escalation
Conventional War and Escalation
Conventional War and Escalation
Keir A. Lieber
Daryl G. Press
Draft 3/1/14
The most critical question about nuclear weapons in the 21st century is the simplest: will they be
used again? In particular, is it sufficiently plausible that countries will deliberately strike each
other with nuclear weapons to merit real concern about inter-state nuclear deterrence? Many
analysts would answer those questions with a simple no. According to an increasingly
common view, states are highly unlikely to use nuclear weapons at all, extremely unlikely to use
them against others who can retaliate in kind, and its virtually implausible that they would use
them against the United States the worlds preeminent military power. No other act would be
as foolhardy. To an increasing degree, therefore, nuclear weapons are viewed as stale leftovers
from the Cold War, and nuclear deterrence is seen as a legacy mission. Cold War era nuclear
arsenals, strategies, war plans, alert postures, deterrence puzzles, and worst-case scenario
planning appear to be relics of a bygone era.1
Of course, even those analysts who are confident that countries will not deliberately use nuclear
weapons recognize an array of contemporary nuclear dangers. For instance, terrorists might
acquire nuclear weapons or materials, or accidents may lead to unwanted detonations; or states
may start inadvertent nuclear wars. But none of those dangers can be effectively mitigated
through nuclear deterrence. Terrorist acquisition is terrifying precisely because terrorists are
presumed to be difficult to deter.2 Accidents cannot be prevented through deterrence.3 And
1
Indeed, some scholars debate whether nuclear weapons were ever essential for deterring the Soviet Union. For two
examples of prominent scholars who argue that nuclear weapons were unnecessary to deter the Soviet Union during
the Cold War, see John E. Mueller, Atomic Obsession: Nuclear Alarmism from Hiroshima to Al-Qaeda (Oxford:
Oxford University Press, 2009); and Richard Ned Lebow and Janice G. Stein, We All Lost the Cold War (Princeton,
NJ: Princeton University Press, 1994). More broadly, many scholars and policy analysts argue that nuclear
deterrence whatever its role in the Cold War is either unnecessary today or a simple mission because intentional
nuclear attack is so unlikely.
2
Paul K. Davis and Brian Michael Jenkins, Deterrence and Influence in Counterterrorism: A Component in the War
on al-Qaeda (Santa Monica, Calif.: RAND, 2002), p. xviii. A new study suggests, however, that it is possible
indeed very likely that the United States can deter states from giving nuclear weapon to terrorists. Keir A. Lieber
and Daryl G. Press, Why States Wont Given Nuclear Weapons to Terrorists, International Security, Vol. 38, No.
1 (Summer 2013), pp. 80-104.
3
A powerful arsenal might even increase the odds of adversary accidents by driving an adversary to adopt risky
deployment postures to ensure the survivability of its arsenal. For example, to ensure that the U.S. arsenal could not
be destroyed by a Soviet first strike, the United States adopted potentially risky procedures such as placing nuclear-
armed bombers on airborne alert, deploying submarines at sea with the capacity to independently fire their weapons,
2
deterrence is a poor solution to the problem of inadvertent war, which by definition does not
result from deliberate decisions favoring conflict. According to this view, the nuclear dangers of
the 21st Century can only be mitigated through non-proliferation, de-legitimization, and eventual
abolition. Even those who acknowledge that nuclear weapons continue to play a residual
deterrent role generally believe that deterrence is straightforward: deterring the deterrable is
fairly simple, and deterring the real dangers (terrorism, accidents, and the unintended) is
impossible. This is why, for a large and growing portion of mainstream analysts and
policymakers, nuclear policy essentially boils down to the goals of non-proliferation and
disarmament.
Unfortunately, the likelihood of intentional nuclear attacks and hence the challenges of
interstate nuclear deterrence are much greater than is commonly recognized. This article
supports this claim by making four principal arguments: First, nuclear weapons are just as
salient today as they were in the past. During the Cold War, nuclear weapons were valuable
because one set of countries (members of the North Atlantic Treaty Organization, or NATO)
lacked the conventional military power to defend themselves from the Soviet Union and its
allies. Nuclear weapons allowed the weak side to deter the strong one.4 And had war
erupted, nuclear weapons would have given the weak side its best hope of fighting the strong
side to a stalemate.5
The Cold War is over, but the underlying conditions that made nuclear weapons vital still exist
today. All that has changed are the seats at the table. Many of Americas potential adversaries
face the same problem today that NATO once faced: how to deter and if necessary stalemate an
adversary thatt possesses overwhelming conventional military power. The platitude that nuclear
weapons are not well suited to the security threats of the 21st century is incorrect; for those
countries who fear U.S. military might or who fear other strong states nuclear weapons are as
helpful as they were for NATO during the Cold War.6
and implementing tightly coupled command and control systems poised to react quickly (e.g., with missile launches)
on warning of an incoming Soviet strike. All of those steps increased the chance of accidental nuclear use. For
example, [Sagan on the bombers and accidents at Thule; Blair on ICBM PAL codes]. See Bruce G. Blair, Command
and Control (Brookings, 1985); Bruce G. Blair, The Logic of Accidental Nuclear War (Brookings, 1993); Scott D.
Sagan, The Limits of Safety: Organizations, Accidents, and Nuclear Weapons (Princeton University Press, 1993).
4
In this article, weak refers to the country (or alliance) that lacks the conventional military power to prevail in a
conventional war against its key enemies; strong refers to a country that is likely to win a conventional conflict.
Used in this fashion, weak and strong are dyadic features i.e., they refer to the relationship between two states (or
groups of states) rather than to underlying features of the states themselves. Using this formulation, because NATO
declined to spend sufficiently on defense to create a robust conventional defense, which could be expected to
reliably defeat a major Warsaw Pact attack, it required nuclear weapons to create stalemate and effective deterrence.
5
Note that the conventional military balance in Europe was not as one-sided as was often portrayed. But even in the
late-1980s, at the height of NATOs conventional military might, the NATO-Pact military balance was merely
competitive meaning that either side might have prevailed in a conventional conflict. There was never a time in
which NATO could have confidently relied upon conventional forces to defeat a major Pact offensive. For critiques
of the excessive pessimism about the conventional military balance during the Cold War, see Alain C. Enthoven and
K. Wayne Smith, How Much is Enough? Shaping the Defense Program 1961-1969 (Santa Monica, CA: RAND
Corporation, 1971); John J. Mearsheimer, Why the Soviets Cant Win Quickly in Central Europe, International
Security, Vol. 7, No. 1 (Summer 1982); Barry R. Posen, Measuring the European Conventional Balance: Coping
with Complexity in Threat Assessment, International Security, Vol. 9, No. 3 (Winter 1984-85), 47-88.
6
The seminal work on the links between conventional operations and nuclear escalation is Barry R. Posen,
Inadvertent Escalation: Conventional War and Nuclear Risks (Ithaca, NY: Cornell, 1991). Writing at the end of the
Second, weak states face powerful incentives to use nuclear weapons if they find themselves in a
conventional war against a much stronger adversary. Scholars and policy analysts who study
deterrence often claim that no rational leader would use nuclear weapons against a country that
could respond in kind let alone a country that could respond with far greater force. But this is
incorrect. For deterrence to work, it is insufficient that the consequences of action (e.g., using
nuclear weapons) be bad; the consequences of restraint must be acceptable. Leaders facing the
prospect of imminent defeat have compelling reasons to escalate coercively with nuclear
weapons to bring about a ceasefire. Coercive nuclear escalation by the weaker side forces the
stronger side to choose among several options all of which are grim. It is because all of those
options are unattractive that an adversary will be tempted to escalate in the first place.
Viewed through this lens, Pakistan may have powerful, rational reasons to use nuclear weapons
if it is losing a conventional war to India; North Korea has powerful reasons to use nuclear
weapons coercively, rather than permit its enemies to prevail in a war. And Chinese leaders
would face some of these same incentives if their armed forces were suffering a humiliating
defeat in a war in maritime East Asia. In short, an escalatory strategy is cold-blooded, but not
far-fetched indeed, it was NATOs policy for nearly thirty years.7
Third, the logic of wartime nuclear escalation is not hypothetical or based on worst-case
guesses about how countries might wield their nuclear forces. To the contrary, it shaped the
defense plans and nuclear employment doctrines of several nuclear-armed states throughout
history and continues to do so today. We identify the conditions under which states would be
most likely to build defense plans around doctrines of coercive nuclear escalation; we then sort
nuclear-armed countries according to those conditions; finally, we show that those states that
should have adopted coercive nuclear doctrines (according to our argument) have actually done
so.
Cold War, Posen notes that the most common view of how a conventional war could become a nuclear war
focuses on the danger that had NATO found itself losing a conventional ground battle for control of Western
Europe the United States might have reached for nuclear weapons in the hopes of salvaging its position (p. 1).
But two decades later, the common understanding of the incentives of the weak (i.e., those who stand to lose the
conventional war) has evaporated. Few national security experts and it seems few deterrence experts still
remember that it was NATOs strategy to escalate rather than lose a conventional war. Fewer still have sought to
identify the underlying strategic conditions from the Cold War that made intentional nuclear escalation by NATO
seem to be a reasonable strategy. And fewer still have examined the current strategic environment to see if those
strategic conditions still exist today. We seek to remind scholars, analysts, military planners, and national leaders of
what was once a common view; to demonstrate that the underlying conditions and logic which led NATO to plan to
use nuclear weapons against the Soviet Union still exist elsewhere today. The same logic that once would have led
NATO to use nuclear weapons against the Warsaw Pact may pressure North Korea, Pakistan, China, Russia, or
others to deliberately use nuclear weapons today.
7
A policy of coercive nuclear escalation to create stalemate during an unwinnable conventional war was
NATOs policy from the mid-1960s through the end of the Cold War. Prior to the 1960s, NATO believed it could
win a nuclear war, and so it had a different nuclear doctrine: immediate escalation of a conventional conflict, not to
coerce, but rather to destroy the enemys nuclear force and win. For a detailed discussion of the evolution in U.S.
and NATO war plans, see chapter 3 of this book. See also Gregory Pedlow, The Evolution of NATO Strategy,
1949-69, in Gregory W. Pedlow, ed., NATO Strategy Documents, 1949-69, and accompanying documents,
available online at http://www.nato.int/archives/strategy.htm.
Fourth, we argue that three aspects of modern warfare exacerbate the incentives for the weak to
escalate conflicts rather than accept battlefield defeat. Specifically, the nature of conventional
warfare in the information age is highly escalatory. The result is that conventional conflicts
among nuclear-armed states will unleash multiple, reinforcing escalatory dynamics fueled both
by the desperation of the weak, and the military choices of the strong.
Why do so many analysts reach a different conclusion about the likelihood of deliberate nuclear
escalation? One possibility is that scholars and other analysts typically think about peacetime
nuclear deterrence (preventing a surprise nuclear attack), rather than wartime deterrence
(deterring nuclear escalation during conventional wars), the exception being the extensive
literature on escalation risks during an India-Pakistan war.8 But surprisingly, even scholars who
understand the difficulty of deterring escalation during a conventional war when applied to South
Asian security dynamics, argue elsewhere that rational leaders would never use nuclear weapons
against the United States.9 But if analysts believe that Pakistan (the weak) would use nuclear
weapons to prevent conventional defeat (even though Pakistan cannot win a nuclear war), why
would the same analysts dismiss the possibility that North Korea, or in the future Iran, or
possibly China, would use nuclear weapons in an escalatory fashion against a strong nemesis?
However one explains this apparent contradiction, the bottom line is that the same fears that
made vulnerable and fearful countries cling to nuclear weapons in the Cold War make those
weapons essential to the weak and vulnerable in the coming decades. Nuclear weapons are the
ultimate instruments of stalemate they are the ultimate weapons of the weak. Viewed through
this lens, the end of the Cold War radically changed who needed nuclear weapons, but did little
to reduce the utility of the weapons.
This article has four main sections. First, it explains the logic of deliberate coercive nuclear
escalation why the weak might feel compelled to escalate a conventional war, and why they
might hope doing so would grant them the ceasefire they desire. Second, we examine the
nuclear doctrines of nuclear-weapon states across four decades to determine if states actually act
according to the logic developed in the preceding section. Third, we describe the aspects of
modern warfare that exacerbate escalation dynamics. Fourth, we address an important
counterargument and discuss some of the implications of our analysis.
8
For example, see Sumit Ganguly and Devin T. Hagerty, Fearful Symmetry: India-Pakistan Crises In The Shadow
Of Nuclear Weapons (New Delhi: Oxford University Press, 2005); S. Paul Kapur, India and Pakistan's Unstable
Peace: Why Nuclear South Asia Is Not Like Cold War Europe, International Security, Vol. 30, No. 2 (Fall 2005),
pp. 127-152; and V. R. Raghavan, Limited War and Nuclear Escalation in South Asia, Nonproliferation Review,
Vol. 8, No. 3 (2001), pp. 82-98.
9
At a recent presentation to U.S. national security analysts and mid-level U.S. government national security
officials, we asked, How many of you believe a state will deliberately use nuclear weapons against the United
States within 20 years? No one raised a hand. We then asked, How many believe Pakistan would use nuclear
weapons if it were losing a conventional war to India? Roughly two-thirds of the audience raised a hand. When
we asked why North Korea would not face the same incentives as Pakistan, no one offered an explanation and
several of the analysts admitted they had simply never thought about the problem in that way. Washington, D.C.,
September 2012.
The core national security problem for many militarily weak countries is straightforward: how to
keep powerful enemies at bay. For weak countries, military defeat can be disastrous. In some
circumstances, battlefield losses are followed by conquest and harsh treatment of the defeated
society: e.g., a brutal occupation, the loss of sovereignty, or in rare cases genocide. But even
when those terrible outcomes are not likely, war is often disastrous for the leaders of the
defeated. Military planners in weak states particularly those with adversarial relations with the
United States (which has easily vanquished a half-dozen military opponents since the end of the
Cold War)10 must, therefore, address a fundamental question: if war occurs, and conventional
victory is impossible, what strategies might create a stalemate and avoid catastrophic defeat?
Although the United States often treats defeated enemy societies well, the leaders of countries
that recently fought the United States have suffered severe consequences. In 1989, the United
States conquered Panama and arrested its leader, Manuel Noriega. For most Americans, this
short war is forgotten. For Noriega, it triggered a calamitous reversal of fortune: he exchanged a
life of power and riches for twenty-three years in prison and counting. Saddam Hussein
suffered a worse fate; he lost power, he was humiliated, his sons were killed, and he was hanged
in front of jeering enemies. Muammar Qaddafi spent his last days hiding from U.S.-supported
rebels before being caught cowering in a culvert. He was then beaten and shot to death. Dozens
of Qaddafi loyalists, including his son, were also rounded up and executed. Even leaders whose
countries were never conquered those that suffered only limited defeats often paid a high
price. Bosnian Serb leaders Karadi and Ratko Mladi are still in prison in the Hague, where
Serbias former leader, Miloevi, died in detention.11
More broadly, studies demonstrate that leaders have a powerful, personal incentive to force a
stalemate on the battlefield rather than accept defeat. One study used data covering more than 80
years of leadership changes around the world and found that those leaders who achieved a
stalemate in a war were nearly twice as likely to remain in power as those countries that suffered
military defeat. Even more tellingly, the leaders of countries who lost were approximately four
times as likely to be punished exiled, jailed, or killed as those who managed to achieve
stalemate.12
10
Since 1989, U.S. military forces, supported in some cases by a coalition of allies, defeated the military forces of
the following states with minimal U.S. losses: Panama (1989), Iraq (1991), Serbia (1999), Afghanistan (2001), and
Iraq again (2003), and the U.S. provided support to the operation that overthrew the Libyan government (2011).
Although the U.S. military has had considerable difficulty defeating insurgents, from the perspective of weak
governments, the hope that after ones defeat and arrest (or execution) rebels will frustrate the enemy is likely cold
comfort.
11
Randal C. Archibald, Noriega Is Sent to Prison Back in Panama, Where the Terror Has Turned to Shrugs, New
York Times, December 11, 2011; Marlise Simmons, Former Bosnian Leader Begins His Defense at Genocide
Trial, New York Times, 16 October 2012; Simmons, The Hague: Mladics Trial Resumes, New York Times, 9
July 2012; Simmons and Alison Smale, Slobodan Milosevic, 64, Former Yugoslav Leader Accused of War Crimes,
Dies, New York Times, 12 March 2006.
12
Giacomo Chiozza and H.E. Goemans, Leaders and International Conflict (Cambridge: Cambrudge University,
2011). The odds of a leader remaining in office for 1 year after suffering a military defeat was 51%, compared to
89% for a leader whose state fought to a draw. The percentage of leaders who remained in office for four years was
Not only do leaders face great pressure to create battlefield stalemate before they suffer
irredeemable losses, they must do so quickly. A limited conventional defeat that merely
destroys a large fraction of a countrys military, or substantially degrades the institutions that
ensure government control (for example, the leaderships security force, domestic intelligence
services, internal security troops, and party militias), could trigger a wartime or post-war coup.
Even if the military and security services remain loyal, the war must end before they are too
degraded to suppress uprisings in the wake of the conflict. Furthermore, military operations
especially those conducted by the United States increasingly involve intense campaigns against
enemy command bunkers and other leadership sites, posing direct, daily threats to the leaders,
their key political allies, and their families.13 Leaders who see their military being destroyed,
their security services being savaged, and who have bombs raining down upon their command
bunkers, may feel great pressure to halt the war as soon as possible.
The critical point is this: Americas recent conflicts are considered regional wars in
Washington; for adversaries there is nothing regional or limited about them. For the weak,
these are existential struggles.
The leaders of weak states face life-and-death incentives to quickly halt wars that are going
badly for them. But why are nuclear weapons needed for this mission? Several attributes of
nuclear weapons make them uniquely useful for stalemating a stronger enemy. Nuclear weapons
are small and hence relatively easy to hide enhancing their chance of surviving the early stages
24% for losers and 42% for those who stalemated. Over the course of four years, 47% of leaders whose country
lost wars were punished exiled, jailed, or killed while only 13% of those who achieved a draw. These
caluculations are based on the data in Chiozza and Goemans, pp. 56-57. For more on leaders and war outcomes, see
Alexandre Debs and H. E. Goemans, Regime Type, the Fate of Leaders, and War, American Political Science
Review, Vol. 104, No. 3 (August 2010): 430-45. See also, Giacomo Chiozza and H. E. Goemans, International
Conflict and the Tenure of Leaders: Is War Still Ex Post Inefficient? American Journal of Political Science, Vol.
48, No. 3 (July 2004): 604-19.
13
In the 1991 Persian Gulf War, the United States conducted 203 airstrikes on government control targets. That
effort intensified in the 2003 war: U.S. aircraft struck 1,799 aim points in the SR target set, i.e., targets associated
with regime survival and political control over the military. An additional 50 strikes were conducted against time
sensitive leadership targets (i.e., efforts to target Saddam Hussein and other senior members of the government).
See Gulf War Air Power Survey (GWAPS), V. 5, U.S. Government Printing Office, Washington, D.C., 1993; Table
177; and on the 2003 war, Operation Iraqi Freedom By the Numbers, Assessment and Analysis Division,
USCENTAF, 20 April 2003, pp. 4, 5, and 9.
of a conflict. Furthermore, not many nuclear weapons need to survive: each bomb is so
destructive that an adversary who can credibly threaten to deliver even a few weapons against its
enemys cities would possess a powerful coercive tool. Finally, modern delivery systems
particularly ballistic missiles allow states to deliver nuclear weapons to their target, even if its
enemy controls the ground, air, and sea. In contrast, most conventional weapons become
progressively harder to deliver against enemy cities as the enemy gains the upper hand militarily,
and they inflict too little damage to shock the winning side into submitting to stalemate. Taken
together, these three characteristics mean that even a state on the verge of being vanquished can
conceivably destroy the potential victor. The implication: nuclear weapons are the ultimate
weapon of the weak.
Not only are nuclear weapons better suited for wartime coercion than conventional alternatives,
there are three other considerations that make more useful than other weapons that analysts
worry may spread in the 21st century, including cyber, chemical, and biological weapons. First,
although popular culture frequently portrays nuclear weapons as uncontrollably destructive, their
effects can be surprisingly calibrated. Weapons designers have created nuclear weapons with
widely varying yields, allowing mission planners to tailor a strike to create a huge area of
destruction or very little whichever is desired. The biggest weapon in the current U.S.
inventory would destroy roughly 100 times the area that the Hiroshima bomb destroyed; the
smallest would destroy 10% of the area of the Hiroshima weapon.14 Furthermore, by selecting
the altitude of detonation, targeters can choose to create enormous amounts of radioactive fallout
or virtually none.15 And perhaps most importantly from the standpoint of a weak state
conducting a coercive campaign nuclear weapons can be used either slowly or rapidly: they
can be used to destroy one city today and another tomorrow, or one today and a dozen tomorrow.
If fallout is avoided, damage can be meted out in distinct, painful episodes, facilitating coercion.
In our popular culture, nuclear weapons are incredibly blunt tools. Some high-yield weapons
are. But compared to other instruments of coercion, nuclear weapons offer desperate weak-state
leaders tailored escalatory options.16
14
The Hiroshima bomb detonated with roughly the power of 16 kilotons of TNT. By comparison, the U.S. B83
bomb would release up to 1,200 kilotons, and the lowest yield B61 would release 0.3 kilotons. Weapons effects
scale with explosive yield to the 1/3 power, so the B83 would have approximately 10 times the lethal radius as the
Hiroshima bomb, while the B61 would have 30% of the Hiroshima bombs lethal radius. Areas of destruction
increase as a function of lethal radius squared, leading to the figures in the text. See John Malik, The Yields of the
Hiroshima and Nagasaki Explosions (Los Alamos National Laboratory Report LA-8819), Los Alamos, NM,
September 1985; Hans M. Kristensen and Robert S. Norris, U.S. Nuclear Forces, 2013, Bulletin of the Atomic
Scientists, Vol. 69, No. 2 (2013), pp. 77-86; and The B83 (Mk-83) Bomb, NuclearWeaponsArchive.org,
November 11, 1997, available at http://nuclearweaponarchive.org/Usa/Weapons/B83.html. The seminal
unclassified work on nuclear effects is, Samuel Glasstone and Phillip J. Dolan, The Effects of Nuclear Weapons
(Washington, DC: U.S. Government Printing Office, 1977).
15
Above a given height of burst, which is a function of warhead yield, there is vastly reduced local fallout. See
Glasstone and Dolan, Effects of Nuclear Weapons, chap. 9. For supporting calculations and some examples of the
significance of no-fallout airbursts, see Keir A. Lieber and Daryl G. Press, The Nukes We Need: Preserving the
American Deterrent, Foreign Affairs, Vol. 88, No. 6 (November/December 2009); 39-51, as well as the technical
appendix to that article, available at www.dartmouth.edu/~dpress.
16
The technical capabilities required to utilize nuclear weapons in a calibrated fashion, as described above, are
simple for any state that can produce and deliver a nuclear weapon. Weapons with a yield in the single digits of
kilotons apparently like the devices tested by North Korea are well suited to minimal damage attacks. And
even a primitive 20-kiloton weapon, like the first U.S. atomic bombs, would be sufficient to cause massive
Another criterion that makes nuclear weapons uniquely suitable for war-ending coercion: the
utility of nuclear strikes is not nullified by first use.17 Once a cyber weapon is used, the victim
(and others) can learn from the computer code and eliminate key vulnerabilities reducing the
effectiveness of future weapons.18 Similarly, in the aftermath of a biological weapons attack, the
victims military forces and population would don gas masks and take other steps to reduce their
vulnerability to subsequent strikes. Within broader society, public health measures (for example,
restrictions on travel and movement, the use of surgical masks, heightened health monitoring,
and the isolation of contagious individuals) would reduce the effectiveness of follow-on attacks.
But, in contrast, the initial use of nuclear weapons would not nullify the nuclear arsenal to the
degree that bio- or cyber-attacks would. Unless the victim of the nuclear attack can reliably
shoot down ballistic missiles, which remains a very difficult undertaking,19 a weak state can use
nuclear weapons coercively and still retain the ability to conduct future attacks.
Finally, the effects of nuclear weapons detonated above the fallout threshold are far more
predictable than cyber or bio weapons, an essential attribute for a leader who needs to coerce an
immediate end to fighting. Nuclear weapons are more predictable on at least three key
dimensions: the functioning of the weapon, the damage it will cause, and the timing of the
effects. No one knows whether the coercive effect of a nuclear, or biological, or a cyber attack
would work, as we discuss below. But leaders under duress could at least be confident that a
well-tested nuclear weapon would function; would create a reasonably predictable level of
damage (as long as targeters selected a height of burst to prevent fallout); and would detonate at
roughly the desired time. By contrast, one cannot know whether a cyber weapon will infect the
target computer system or whether an infection would produce the desired malfunctions until
the weapon is used.20 In many cases, no one can predict how long it will take for a cyber attack
destructive effects if desired. Controlling height of burst with sufficient accuracy to cause or prevent fallout merely
requires simple altimeters, a technology that is easily available to any country capable of firing ballistic missiles.
17
Because the victim of a nuclear strike cannot easily take steps to inoculate itself from subsequent attacks, a state
using the weapons coercively can enhance the credibility of its threats through an initial strike without nullifying the
effectiveness of its remaining weapons. This is an essential quality of a weapon to be used for coercion because, as
Thomas Schelling pointed out, coercion works through the fear of future pain. Killing one hostage only coerces if
there are others who remain in jeopardy. See Thomas Schelling, Arms and Influence.
18
Those who examine the code may not merely learn about the vulnerabilities in the target computer systems code,
they may also learn about technical or organizational vulnerabilities that permitted the malware to be delivered to
the target. For instance, computer networks that have no connectivity to the outside world have been penetrated by
luring employees with access to unknowingly (or intentionally) use infected flash drives in the otherwise-sealed-off
network. But once that vulnerability was exploited, re-attack became more difficult (e.g., workers at sensitive sites
were warned about such operations, and in some organizations USB ports have been physically sealed). See Martin
Libicki, Cyberdeterrence and Cyberwar, (Santa Monica, CA: RAND, 2009), pp. 56-59. For a discussion of these
issues in the context of the Stuxnet attack, see Kim Zetter How Digital Detectives Deciphered Stuxnet, the Most
Menacing Malware in History, Wired, July 11, 2011. Available at:
http://www.wired.com/threatlevel/2011/07/how-digital-detectives-deciphered-stuxnet/all/.
19
For a description of the enduring problem faced by all exoatmospheric hit-to-kill missile defense systems,
differentiating warheads from decoys and debris out of the atmosphere, see George N. Lewis, Theodore A. Postol
and John Pike, Why National Missile Defense Wont Work, Scientific American, August 1999. The enduring
challenge is noted in Defense Science Board Task Force Report, Science and Technology Issues of Early Intercept
Ballistic Missile Defense Feasibility, U.S. Department of Defense, September 2011.
20
Note that even after a cyber strike, gauging effectiveness is challenging. After years of self-congratulation,
evidence is emerging that the most famous offensive cyber attack in history Stuxnet was a tactical and strategic
failure, even using the most modest definition of success (i.e., temporary reduction in Irans enrichment of uranium
to disrupt the target computers, or assess the unintended consequences of the malware infecting
other computer systems. Similarly, biological weapons may take considerable time to spread, to
incubate in their victims, to be detected, and to be attributed all of which must happen before
an attack can generate a coercive effect.
During wars, the leaders of the states on the losing side may face life-and-death pressure to
rapidly force a ceasefire even if their enemy is not seeking to conquer them or impose regime
change. Conventional weapons provide little leverage in this regard most of them become
progressively more difficult for the weak to employ as the strong gains the upper hand militarily,
and they generally inflict too little damage to shock the strong state into submitting to
stalemate.21 When NATO faced an overwhelming conventional military threat, it did not plan to
stalemate the Warsaw Pact using highly uncertain biological weapons. If the challenge facing a
leader is to stop a powerful aggressor immediately, then there is currently no substitute for
nuclear weapons.
Losing a conventional war could have catastrophic consequences for the defeated society or
leaders; but how could a country, facing an overpowering foe, employ nuclear weapons to create
stalemate? Wouldn't the use of nuclear weapons by a weak country against a strong one incite a
devastating nuclear response, rather than a truce? For example, in a war on the Korean
Peninsula, wouldnt North Korean use of nuclear weapons against the Republic of Korea, Japan,
or U.S. military forces in the region trigger a devastating U.S. nuclear retaliatory strike? If so,
then nuclear escalation would simply turn a conventional defeat into an even worse nuclear
disaster. So, how could coercive nuclear escalation work?
Working through that hypothetical scenario a Korean War, five years in the future is
revealing. A conflict on the Korean Peninsula could erupt through any number of paths, but
regardless of how it started, relatively early in the conflict the conventional battle would likely
start to favor the U.S.-ROK alliance.22 And according to statements from officials in Seoul and
Washington, the alliance would quickly turn from defense to offense, and begin to move north of
the DMZ. At that point, leaders in Pyongyang would face a stark choice. They could allow the
conflict to continue on its course, and accept a similar fate of Qadaffi and Hussein, or they could
ask themselves: what means do we have to force the United States and South Korea to
immediately halt offensive operations?
at the Natanz facility). According to IAEA documents, the Stuxnet attack barely reduced Irans rate of uranium
enrichment, which quickly returned to (or exceeded) pre-Stuxnet rates. We thank Jonathan Lindsay for bringing this
to our attention.
21
On the limits of coercion using conventional weapons, see Robert A. Pape, Bombing to Win: Airpower in
Coercion and War (Ithaca, NY: Cornell University Press, 1996); for an insightful critique, see Karl Mueller,
Strategies of Coercion: Denial, Punishment, and the Future of Air Power, Security Studies, Vol. 7, No. 3 (Spring
1998): 182-228.
22
See, for example, Anthony H. Cordesman, The Korean Military Balance: Comparative Korean Forces and the
Forces of Key Neighboring States, Center for Strategic and International Studies, May 6, 2011; and The
Conventional Military Balance on the Korean Peninsula, in IISS Strategic Dossier, North Koreas Weapons
Programmes: A Net Assessment, International Institute of Strategic Studies, January 21, 2004.
Nuclear escalation could take many possible forms: Pyongyang might begin with just a statement
demanding an immediate ceasefire and threatening nuclear escalation. If North Korea has
nuclear weapons married to missiles, it could launch a missile and detonate it harmlessly over the
Sea of Japan. If North Korea develops missiles that are accurate enough, it could launch a
nuclear strike on a U.S. military base in the region, such as Kadena Air Base on the island of
Okinawa, Japan. It could even strike a Japanese or South Korean city. But the most important
aspect of a coercive escalatory operation is not the initial strike, but the threat of what is to come.
Whatever the first step, Pyongyang could then declare that the United States and ROK must
cease military operations against North Korea immediately, or else North Korea will destroy half
a dozen Japanese cities.23
Some analysts might assume that the United States would respond at this point with a
devastating nuclear counter-strike especially if the North Korean coercive strategy involved an
actual nuclear strike. But it is enlightening to consider carefully the options that a U.S. president
would confront in such circumstances. Each of these options is grim.
What options would a U.S. president have if North Korea used nuclear weapons coercively
during a conventional war? How would the United States respond, for example, to North Korean
nuclear attacks on Kadena Air Base and a Japanese city that killed several thousand Americans
and two or three times that many Japanese? How would a U.S. president address Pyongyangs
threat to launch further strikes on Japanese cities unless the United States and the ROK accept a
cease-fire and halt their military campaign? In such a scenario, four principal courses of action
would be available.24
Option One: Punitive Nuclear Retaliation. When many people initially confront the question
How should the United States respond to a limited, coercive nuclear strike by North Korea on a
U.S. military base? a common response is a more colorful version of launch punitive nuclear
retaliation. In other words, one option would be to launch one or more nuclear attacks designed
to kill the North Korean regimes leaders and destroy the remaining institutions of the North
Korean state. After the retaliatory strike, South Korean and U.S. forces would still march toward
Pyongyang as soon as conditions allowed. The purpose of this response would be to send a clear
message to the world nuclear escalation will beget a horrifying response.
The disadvantages of this approach are substantial. First, and most obviously, the United States
would be committing mass murder. Hundreds of thousands of North Korean civilians would be
killed for acts committed by a small coterie of leaders. Second, nuclear strikes aimed at deeply
buried leadership bunkers would require ground bursts detonations well below any altitude
23
Of course, the issuance of such a threat does not mean that North Korea could carry out that operation. U.S. and
allied missile defenses would attempt to shoot down North Korean missiles, and the United States and its allies
might seek to prevent a follow-up North Korean nuclear attack by launching a conventional or nuclear counterforce
strike (as described below). The point here is that the issuance of a coercive nuclear threat by Pyongyang during a
conventional war (perhaps in conjunction with a small nuclear strike) would not be irrational; far from crazy, such
a strategy would mirror NATOs Cold War plans for coercive nuclear escalation, which were also designed to create
stalemate to avert a conventional military defeat.
24
To be clear, this scenario is merely intended to illustrate how coercive escalation might work that is, to show the
logic of coercive nuclear escalation by illustrating the terrible dilemmas faced by the victim of a coercive campaign.
The details in any scenario are not predictable, and are not central to this analysis. The point here is that coercive
escalation has a compelling logic, as NATO, Pakistan, and others have discovered.
that would avert fallout and would therefore spread highly radioactive material across the
region. Depending on the location of the bunkers and the season (which affects wind direction),
lethal fallout would likely scatter across South Korea, and possibly Japan or China.25 Finally, a
punitive strike would not solve the major dilemma at hand: North Korean nuclear forces would
presumably already have been dispersed and could still carry out their retaliatory nuclear strikes
against Japan. The visceral bomb them back to the stone age response is problematic on many
dimensions.
Option Two: Conventional Military Response. A second option would be to condemn the
nuclear strike, send aid to the people of Okinawa, and accelerate the conventional offensive
toward Pyongyang to end the war and capture the North Korean leadership as rapidly as possible.
The advantage of this approach is that it reinforces the core of U.S. nuclear policy: by not giving
in to coercion, and by not responding in kind, the U.S. response would demonstrate that nuclear
weapons are both horrible and useless. The subsequent trials of surviving senior North Korean
leaders would demonstrate to the leaders of other weak states that nuclear escalation is not a
viable way to escape the calamity of military defeat.
The disadvantages of this strategy are enormous. First, the strategy would accept the risk that
North Korea would carry out its threat and launch nuclear strikes against a half-dozen Japanese
cities. There is substantial risk that some (perhaps many) of those missiles would leak through
missile defenses. Second, and relatedly, this course of action would presumably be implemented
over the strenuous objections of Japans government. The consequence would likely be the end
of the U.S.-Japan alliance. More broadly, if the United States ignores the pleas of a critical ally,
and the consequences were the destruction of several of that allys cities (in a war in which the
ally played no direct role), many U.S. allies around Asia and the rest of the world may rethink
their tight military ties to the United States.
Option Three: Counterforce: Disarm, then Defeat. The third option would be to respond to the
nuclear attack with a major military strike against known and suspected North Korean nuclear
targets to prevent North Korea from launching additional weapons. A counterforce strike could
be conducted with conventional weapons, nuclear weapons, or a mixture of the two, with
respective implications for the promptness of destroying the intended targets and the likelihood
of destroying them all. This option, like the others, would rely on imperfect missile defenses to
help with any North Korean weapons surviving a U.S. strike. And, as with the first two options,
a rapid conventional advance on Pyongyang to conquer the regime and seize any surviving
leaders would follow this strategy. The advantage of this option is that it would avoid giving in
to nuclear blackmail, and it would take direct action to protect U.S. allies as much as possible.
The disadvantages of this option, however, are substantial. First, a counterforce attack would not
be a small operation. It would likely require prompt attacks on scores of targets across North
Korea in order to rapidly destroy suspected nuclear storage sites, military command and control,
mobile missile garrisons, and tunnel entrances which may be associated with North Koreas
nuclear weapons or missile launchers. The nuclear component of the attack might involve
25
We conducted fallout analysis of various hypothetical U.S. ground burst strikes against North Korea, using a U.S.
Defense Department computer model called HPAC, and depending on the target location and season, the radioactive
fallout from U.S. strikes might kill more South Korean civilians than North Koreans.
several dozen or more U.S. weapons. Second, depending upon the details of the U.S.
operation, and the location of North Korean targets, the U.S. strikes could kill a large number of
North Koreans. (This would probably be the case even if U.S. strikes did not generate regional
radioactive fallout, as in the counter-leadership or punitive option described above).26 A
third disadvantage is that a counterforce strike would probably not destroy every North Korean
nuclear weapon; some weapons might survive and be used against U.S. allies. This option,
therefore, like the first two, accepts a high likelihood of one or more allied cities being
destroyed, along with subsequent damage to the U.S. global alliance network and grand strategy.
This option becomes more perilous the closer that North Korea moves toward deploying long-
range ballistic missiles that can target U.S. cities, as well as regional allies.27
Option Four: Ceasefire: Prevent Further Escalation. The argument in favor of accepting a
ceasefire is that there is nothing on the Korean Peninsula that is worth fighting a major nuclear
war. A nuclear exchange between the United States and North Korea would likely kill large
numbers of Koreans (especially if North Korean nuclear sites were near populated areas), and
could lead to substantial retaliation against U.S. regional allies. If Japan or other allies in the
region were subsequently struck, it might be the end of the U.S. alliance network in East Asia, as
well as undermine U.S. nuclear umbrella commitments to dozens of other countries. Advocates
of a ceasefire could argue that the North Korean regime would be further isolated by its conduct
for example, China would feel immense pressure to cut off any assistance for Pyongyang after
such events and suggest that the regime would thus soon collapse. Most important, one could
argue that the potentially huge political and strategic implications of buckling to nuclear coercion
could be mitigated. For example, before accepting the ceasefire option, the United States
might levy a symbolic U.S. nuclear response (e.g., responding to a North Korean strike on
Kadena Air Base with a nuclear response against one or more North Korean military facilities)
before halting military operations.28 A globally respected international figure could also be
encouraged to make a public plea on behalf of all humanity that both sides cease military
actions immediately.29
26
It is critical to note that strikes on tunnel entrances and other hardened facilities may not require ground bursts,
and appear to be possible without creating significant fallout. Unlike a punitive strike designed to kill the
leadership (option 1) the nuclear missions in option 3 do not need to destroy deeply buried facilities but could
merely destroy the near-surface elements of those facilities (using air bursts) to disable the weapons or prevent them
from being used until teams could seize the sites. Declassified documents reveal that the United States has been
planning low-fallout, low-casualty nuclear options for decades.
27
Before leaving office, former Secretary of Defense Robert Gates endorsed U.S. intelligence estimates that North
Korea was (in 2011) within five years of being able to strike the continental United States with a long-range missile.
The December 2012 partial success of a test of a North Korean satellite launch vehicle reflects a big leap forward for
Pyongyang, and may have been fueled by increased Iranian technical assistance (Irans satellite launch program has
been far more successful than North Koreas). On the December 2012 North Korean missile launch, see Choe Sang-
Hum and David E. Sanger, North Koreans Launch Rocket in Defiant Act, New York Times, December 1, 2012.
On Iranian assistance for the North Korean missile program, see John S. Park, The Leap in North Koreas Ballistic
Missile Program: The Iran Factor, NBR Analysis Brief, December 19, 2012; and Jeffrey Lewis, Iranians in North
Korea? ArmsControlWonk.com, December 5, 2012.
28
To be clear, the U.S. nuclear response in this option would not be designed to kill the North Korean leadership or
disarm its nuclear forces (options 1 and 3, respectively). The goal would simply be to provide political cover for the
ceasefire.
29
The United States employed a similar strategy as a backup plan in case the blockade failed during the Cuban
missile crisis. The White House established a contingency plan to secretly ask the UN General Secretary to
intervene and urge both sides to reach a compromise.
The downsides of accepting a ceasefire are also very significant. Accepting a negotiated
settlement after suffering a nuclear strike (or after receiving an explicit nuclear threat) might be
very costly politically both for the United States and personally for the American president.
U.S. leaders would worry about the precedent it set, in which a weak state coerced the ceasefire
it needed by threatening or attacking the United States with nuclear weapons. Such a strategy
could trigger a new wave of proliferation not only by adversaries, but also by allies that lose
faith in the U.S. nuclear umbrella. And while symbolic escalation and subterfuge might make
the deal politically palatable in the short term, when the dust settled it would become apparent
that coercive escalation had worked.
None of the response options discussed above are attractive. The reflexive course of action in
the wake of a nuclear attack on an ally a devastating nuclear retaliatory strike is not grounded
in a careful assessment of the costs and benefits of that response. A U.S. president might select
that option, but nothing about such a decision is preordained. Others believe that the United
States would select the counter-force option; once an enemy has used nuclear weapons, a U.S.
president would have no alternative to destroying as many of those weapons as possible. (It is
noteworthy that current proposals regarding the future of the U.S. nuclear arsenal make the
nuclear force less well suited for a counterforce strike.)
All the response options in this scenario are grim, but options three and four would likely
dominate the first two. Marching to Pyongyang and simply absorbing additional attacks on
Japanese cities praying that missile defense will work flawlessly (i.e., option two) seems
unviable, as does a punitive counterstrike not specifically focused on disarming the North
Korean nuclear arsenal (option one). The wrenching decision for a U.S. president would be
whether to order a nuclear counterforce strike to disarm the enemy (option three), or whether to
accept a ceasefire with some attempt at saving face (option four). The key factor pushing a U.S.
leader toward option three or option four is the likely effectiveness of a counterforce strike. If a
president believed that a counterforce strike would leave the enemy with zero or perhaps a
couple deliverable nuclear weapons, and if he or she believed that it was possible to execute
such an attack without killing large numbers of noncombatants (particularly allied civilians), he
or she might lean in that direction. On the other hand, if the president believed that a
counterforce strike would still permit the adversary to destroy many allied cities or U.S. cities
he or she might prefer a face-saving symbolic strike followed by a ceasefire.
One can only speculate about what a U.S. president might do under these circumstances. But
what should be clear from this illustrative scenario is that an adversarys coercive nuclear
strategy might work: it might induce the stronger state whomever that might be to opt for a
ceasefire (option four). Why would North Korea or Pakistan, China, or (in the future) Iran
believe nuclear coercion might create stalemate? For that matter, why did NATO stake its
survival on the belief that it could induce stalemate in the midst of conventional war with the
Soviet Union? The answer is clear: if weak states can deploy enough nuclear weapons, or
deploy them in a fashion that makes them very difficult to destroy, strong states would likely
have few palatable reactions to a coercive nuclear escalatory campaign. In other words, coercive
escalation by weak states should be acknowledged as a rational strategy, especially if weaker
actors (like Pakistan and, in the past, NATO and Israel) build a force that is sufficiently
invulnerable to a disarming, counterforce strike. In short, relatively weak states will face
powerful incentives to use nuclear weapons against the strong during a conventional war in order
to induce stalemate.
Nuclear weapons have the capacity to be the ultimate tools of stalemate. If weak nuclear-armed
states feel sufficiently threatened by a militarily superior foe, they could develop defense plans
around the concept of coercive nuclear escalation, and create nuclear doctrines for wartime
employment. This article, so far, makes the case that such steps are logical. But do countries
actually follow this cold-blooded logic?
To explore whether weak states actually employ this logic, we first identify below the conditions
under which states would be most likely to build defense plans around doctrines of coercive
nuclear escalation. Second, we sort nuclear-armed countries according to those conditions.
Finally, we determine whether those states that (according to our argument) should have adopted
coercive nuclear doctrines have actually done so.
Two factors should have a powerful effect on whether nuclear-armed states develop coercive
nuclear doctrines. First, countries are more likely to view nuclear weapons in this manner if they
expect to lose conventional wars. In other words, coercive nuclear doctrines should be far more
appealing to the weak than to the strong. Second, these doctrines will be more attractive to states
for which the consequences of conventional military defeat are dire. When the United States
loses conventional wars e.g., in Vietnam, perhaps in Afghanistan it may damage presidential
approval, but the republic does not fall, and leaders are not hung. For other states and leaders,
defeat often brings terrible consequences. Many Israelis believe that the consequences of a
military defeat to the Syrians or Egyptians would mean the end of sovereignty at best and
genocide at worst. Even countries that do not fear military conquest might worry that a
humiliating conventional defeat might trigger uprisings or coups, and the overthrow of the
existing regime (often including the death of the leaders themselves). If the United States dealt
an overwhelming defeat to the Iranian military during a conflict over the Strait of Hormuz, it is
not clear that the Islamic Republic would survive the political turmoil that could follow.
Observers of China have noted that the Chinese Communist Party (CCP) no longer bases its
legitimacy on communism, but rather on nationalism and evidence that the CCP has made China
strong and globally respected. If during a military clash in the Pacific, the United States inflicted
a crushing defeat against the Chinese air force and navy, the leaders of the CCP may reasonably
question whether their government could survive the humiliation and anger from the people or
military. Leaders of nuclear-armed states who fear that conventional military defeat could lead
to terrible consequences for themselves or their country would be expected, if the arguments in
this article are correct, to be more likely to develop coercive nuclear doctrines than those who do
not share this fear.
Figure 1 illustrates our claim graphically and offers a first-cut at identifying strategically relevant
dyads along these two dimensions. Each of the two variables we describe is, in reality,
continuous: the expected likelihood of defeat in a conventional war could be any value between
0 and 1; and the negative consequences of defeat could range from nothing to total annihilation.
But to facilitate coding and avoid suggesting greater precision than is possible using these
variables we treat each of the variables as if it were binary, thus resulting in four categories. If
our argument about coercive nuclear escalation is correct, then the countries represented in the
dyads in the upper-right corner i.e., those nuclear-armed state that expect to suffer conventional
defeats over issues of grave importance should be most likely to adopt coercive nuclear
doctrines. Those in the bottom left corner should be least likely. (Because the current strategic
circumstances of various NATO allies are so different from each other with the Baltic
countries facing very different military threats than France or the United Kingdom we have
located Baltic NATO separately from the other alliance members.)30
Low! High!
How Bad are Consequences of Defeat!
Figure 2 reproduces the first figure, but it also indicates (in bold text) which countries appear to
have adopted a coercive nuclear doctrine and possess the theater or battlefield nuclear
capabilities to execute it. (As described above, in the case of contemporary NATO, we separate
out the Baltic states, because they are located in a different strategic quadrant from the other
NATO members, and they have different preferences regarding NATO nuclear doctrine). Figure
2 also highlights (in gray text) the countries that have not articulated doctrines for the coercive
30
The Baltic countries are not nuclear-armed states, of course, but they are covered by NATOs nuclear umbrella.
use of nuclear escalation. One state North Korea is particularly reticent about sharing
information about its nuclear doctrine.31
Low! High!
How Bad are Consequences of Defeat!
Overall, Figure 232 suggests that the nuclear-weapon states that worry most about calamitous
military defeat tend to develop coercive nuclear doctrines to give them the capability to stalemate
their most-threatening adversary. NATO thought this way in the Cold War; Pakistan and Russia
do today; and the only members of NATO who face the real possibility of disastrous military
31
Critics may object that Figures 1 and 2 should include all states and locate them along these two dimensions
not merely the nuclear-armed countries. After all, the theory weve advanced predicts that any state that perceived a
high chance of suffering a costly military defeat would be powerfully inclined to adopt a coercive nuclear doctrine
even if that first required acquiring nuclear weapons. We agree with this logic up to a point, but the critique goes
too far. A large body of evidence suggests that proliferation decisions involve a careful balancing of security
concerns with a host of factors arising from domestic politics: e.g., the interests of the military organizations (who
often oppose these weapons); the interests of various parts of a countrys scientific community; commercial
interests, which might fear sanctions and isolation. Even in the domain of security concerns, the logic is far from
deterministic because several countries (e.g., South Korea, Taiwan, and Japan) have directly linked their non-
proliferation stance to promises of military support by the United States. Acquiring nuclear weapons and adopting
coercive nuclear doctrines would mitigate some security concerns for these countries and exacerbate others. In
short, we are not offering a theory of proliferation here, but rather an explanation for the adoption of coercive
nuclear doctrines by nuclear-armed states. By doing so, we assess the likelihood of deliberate nuclear escalation in
the midst of conventional wars.
32
[An annotated description of the nuclear doctrines of each state in Figure 2, supported by interviews of U.S.
government officials, will be included separately in a forthcoming Appendix.]
defeat are the same ones who most strongly favor retaining NATOs forward-deployed tactical
nuclear weapons (B61 bombs deployed in Europe). It is worth noting that North Korea is left
un-coded, because it has not publicly articulated enough to identify an explicit nuclear doctrine,
though the Pyongyang government has certainly issued statements that suggest a willingness to
use nuclear weapons or other unspecified means to punish its enemies if it were attacked.
China is the clear outlier the country that avows a no first use nuclear doctrine despite its
location in the upper-right corner.33
Finally, it is notable that the nuclear-armed states not only sort themselves in the predicted
fashion on Figure 2, but also several countries changed nuclear doctrines soon after they moved
from one quadrant to another. During the Cold War, when NATO felt unable to defend itself
adequately from a major conventional attack, it adopted a coercive nuclear doctrine. When the
balance of power shifted moving most NATO countries from the upper-right to the lower-left
quadrants so did the views of many alliance members about NATOs nuclear doctrine, and
even about the legitimacy of the weapons they recently relied upon themselves.34 Russia shifted
in the opposite direction. During the Cold War, Russia supported (at least rhetorically) the
position of no first use, but now that the military balance has shifted sharply against them
Russian officials have publicly stated that they rely upon tactical and theater nuclear weapons to
balance against the superior military forces of an unspecified powerful alliance.
The overarching argument in the first half of this article is that the same escalation dynamics that
existed during the Cold War exist today as well and they are just as powerful. Nuclear
deterrence is not a legacy mission, because states still face the same critical national security
threats they faced during the Cold War and throughout history: namely, the leaders of weak
states fear that the strong will conquer them or take steps that will lead to their downfall. The
high-stakes poker game of international politics has not fundamentally changed contrary to the
claims of many observers. What has principally changed is merely who has the best cards.
Those who were weak during the Cold War are now strong; and another set of militarily weak
countries such as North Korea, Iran, and even China and Russia now clutch nuclear weapons
to defend themselves from overwhelming military might, just as NATO once did.
The failure of analysts in the West to appreciate the continued value of nuclear weapons reflects
a striking lack of strategic empathy. The first rule of good strategy is to develop an
33
We recently attended a conference in Washington (2012) in which several U.S. government officials with
expertise on China, have participated in conversations about nuclear doctrine with Chinese officials, and who have
responsibilities that include U.S.-China nuclear relations, conveyed that they believe Chinas no first use pledges
should not be interpreted literally. They indicated that discussions with official Chinese delegations about these
issues reinforced the impression that Chinas actual nuclear policy is more nuanced than no first use, and that
Chinas representatives indicated that a range of non-nuclear U.S. military actions might trigger Chinese nuclear
response. But, for the sake of coding consistency, we coded China as not having a coercive nuclear doctrine.
34
In fact, some statements by current and former U.S. officials to justify the ongoing U.S. effort to delegitimize
nuclear weapons and work toward global nuclear disarmament note that the United States and its allies have the
worlds most powerful conventional forces in the world implicitly acknowledging that these weapons were once
useful because NATO and the U.S. were weak. Those statements never draw attention to some darker
implications: that weak states will resist efforts to deny them their needed instrument of stalemate as vigorously as
NATO rejected Soviet suggestions for a mutual no first use pledge during the Cold War, and that efforts to
delegitimize these weapons may come at the expense of U.S. allies who still feel some risk of catastrophic military
defeat (i.e., Israel).
understanding of how ones adversaries see the world, and how they might utilize their resources
to achieve their goals. Analysts in the United States who claim that nuclear weapons are
essentially irrelevant to the problems of the 21st Century either ignore this dictum or have
forgotten what was once better understood: how helpful these weapons can be for stalemating the
strong.
The principal danger of nuclear escalation stems from the desperation of the weak. But the
actions of the strong matter, too, and may greatly exacerbate the likelihood of escalation. How
can we assess those risks? What actions by the strong increase the likelihood of nuclear
escalation during a conventional war? What actions would reduce those risks?
We start with a simple theoretical framework for thinking about escalation incentives. Building
on the discussion above, we assume that a leaders need to escalate during a war is driven by his
assessment of the likelihood of him surviving and retaining power after the conflict. If this
assumptions is correct, then actions by the strong during a war which increase the risks to the
enemy leaders survival and authority would push the enemy toward escalation not out of spite,
but to force the strong to desist. Conversely, actions that increase the weak leaders chance of
surviving the war in power would lesson his need to escalate.
This conceptual rational-actor model like all useful models simplifies a complicated reality.
In the real world, a jumble of factors motivates leaders, including conscious preferences and
subconscious feelings (such as vanity, insecurity, anger, desire, and fear). Critics of our analysis
might argue that although the premise underlying our simple model that leaders wish to live
and rule is probably true, our simplifying assumption treats all rulers as similar, when they are
not, and it reduces a stew of motivations into two simple goals: stay alive and stay in power.
These critics are right, but the model we employ is nevertheless useful because it sets a minimum
standard for what strong countries must do if they wish to prevent escalation. Military
campaigns that seek to leave an adversary leader alive and in power may still fail to prevent
escalation because leaders may by driven by other overriding motivations. But a military
campaign that threatens those two core objectives should be expected to trigger escalation.
Stated differently, even a plan for limited war may trigger escalation. But a military operation
designed to topple enemy regimes, kill enemy leaders, or by extension destroy the forces that the
enemy leaders rely upon to ensure their survival i.e., the enemys strategic deterrent should
be expected to provoke escalation.
On the basis of this simple model, the risks that future conventional wars will escalate and
trigger nuclear conflicts are startlingly high. In the section below we use a range of theories,
evidence from recent conventional wars, and interviews with senior U.S. military and civilian
officials to conclude that the nature of modern warfare is highly escalatory and that the United
States, and perhaps other countries, have adopted policies that exacerbate the dangers of
escalation.
Conventional war has changed dramatically over the past three decades as computers have
become fully integrated into every facet of warfare. The computerization of weapons and
warfare has changed nearly every aspect of combat: e.g., command, communications,
reconnaissance, navigation, and the precision with which weapons can be delivered against
targets. As many observers have noted, modern high-tech warfare can be devastatingly effective
against conventionally armed foes.35 What has gone unnoticed, however, is that this high-tech
style of warfare is also highly escalatory.
The computerization of warfare has had three overarching effects on combat. First, military
forces now derive their effectiveness, more than ever before, from their ability to function as part
of a network. Sensors, data processing facilities, commanders, and shooters are often widely
dispersed; increasingly, generating combat power depends on a militarys ability to integrate
information from multiple sources, make effective decisions, and then coordinate the actions of
widely dispersed forces.36 Second, and following directly from the first point, the payoffs from
disrupting an adversarys command and control network have soared. For example, Chinas
newest anti-ship missiles are fearsome weapons, capable of destroying a war ship nearly 2,000
miles away on the open ocean but only if every link in the sensor-processing-command-shooter
chain is intact.37 Severing or even delaying these links renders the missile system useless.
Third, powerful states now have an unprecedented capacity to degrade an enemys command
and control system: thanks to long-range precision weapons, and possibly also through
unconventional means (e.g., offensive cyber attacks). It would be an exaggeration to say that
warfare is now entirely about degrading enemy command and control; rather, those operations
typically open the door for decisive force-on-force engagements. But the efforts to gather and
utilize information, coordinate actions among many units, and deny that intelligence and
coordination to others, is a bigger part of modern warfare than ever before.38
35
[Cites.]
36
[Cites.]
37
Chinas long-range anti-ship missile systems require near-real-time information on enemy ship location from
some combination of satellites, drones, submarines, and long-range radars. The signals from those sensors are
processed at some distant command site, where targets and their locations are assigned to individual missile
launchers. Attacks on the sensors, or the command locations where the signals are processed, or on the
communication links between the sensors, command, and shooters, would vastly reduce the effectiveness of the
weapon system. Similarly, integrated air defenses Chinas and those of other countries are effective because
they integrate information from many dispersed sensors to identify air targets and allocate them to a dispersed set of
shooters. In an age in which fewer line-of-sight attacks are conducted, operations that can sever sensors,
command, and shooters can vastly reduce the effectiveness of a modern military. On Chinas anti-ship ballistic
missile, see [Cites].
38
Another signature aspect of modern warfare is the increased dispersal of combat units. This is a necessity given
the growing lethality of precision strikes, but it is also an opportunity stemming from the growing ability of forces to
support each other from greater distance (thanks to improved communication, and the speed and accuracy of long-
range strikes). These changes did not occur overnight, and they do not appear to be short-lived. The United States
military has been working since the 1970s to learn how to use precision weapons most effectively on the battlefield.
The advent of precision bombs allowed the U.S. Air Force to develop novel strategies for affecting the battlefield:
instead of merely bombing enemy troop concentrations, fighter aircraft could use early precision weapons, plus
electronic jammers, to disrupt enemy command and control. Instead of destroying a few armored vehicles, and
effective air strike could disrupt an entire enemy campaign. John Boyd. OODA loop. For a description of the
Over the past twenty years, every major U.S. military operation has begun with an intense effort
to destroy the enemys command and control. For example, the first five days of U.S. air
operations in the 1991 Persian Gulf War focused on degrading the Iraqi militarys central
nervous system, rather than hacking off its limbs. More than a thousand airstrikes targeted Iraqi
surface-to-air radars and missile systems (to allow the United States unfettered access to Iraqi
airspace), command posts, electricity, communications, and organs of government control all
aimed at denying the Iraqi leadership situational awareness and preventing them from
coordinating their military forces in the field.39 Of these strikes, nearly two hundred were
launched against Iraqs leadership on the first night of the air war representing an intense effort
to kill the senior members in Saddams government.40 The air war against Serbia (1999) and
during the Iraq War (2003) followed suit.41 Even the wars against enemies with more
rudimentary command and controls systems the Taliban leaders of Afghanistan (2001) and the
brief campaign against Libya (2011) began with attacks on the leadership and their ability to
command and control their defense forces.
The problem is that although this style of warfare can be very effective at producing one-sided
battlefield outcomes, it is also highly escalatory. If preventing escalation requires assuring
enemy leaders that they will survive and remain in power after the war if they do not escalate
then military campaigns must demonstrate that those promises will be met. In other words, these
campaigns must allow leaders to see that their enemys military objectives are limited, and that
their own critical political control organizations are not being destroyed. Attacks designed to
cause the enemys command and control system to collapse, deny enemy leaders situational
awareness of the battlefield, and kill the enemy leadership itself, undermine those goals. Attacks
development of these views and doctrines within the U.S. Air Force, see Lambeth, The Evolution of American
Airpower. Others.
39
Data derived from Gulf War Air Power Survey, Vol. V, Table 178. Here we count each strike sortie as a single
strike. The sorties reported here refer to strikes on the categories of C3, electricity, government control,
and SAM, as reported in GWAPS. The 1,002 sorties over five days represents 26% of all Coalition strike sorties
in that time period; on the first day alone, 31% of the strikes on Iraq were directed against these target categories.
Evidence suggests that the effort to destroy Iraqi command and control in 1991 was not nearly as effective as is
often suggested, but the 1991 air war was fought with very few precision guided munitions. The 1991 air campaign
charted the direction for future U.S. air wars. On the effectiveness of the 1991 air war campaign, see Daryl G. Press,
The Myth of Airpower in the Persian Gulf War and the Future of Warfare, International Security 26 (2001), pp. 5-
44; and Press, Lessons from Ground Combat in the Gulf: The Impact of Training and Technology, International
Security 22 (1997), 137-46.
40
On the first day of the 1991 air war, there were 193 strikes against Iraqi leadership targets, in which category we
include strikes on C3 and Government Control targets. Data derived from GWAPS, V5, Table 178. There is no
doubt that the purpose was to kill Saddam Hussein and other senior leaders. In an initial briefing on the air war
plan, U.S. military planners indicated that they illustrated the U.S. air war plan as a bullseye with the words
Saddam Hussein in the center. Concerns that this depiction might suggest that the USAF was violating
restrictions against assassination led the war planners to replace the words Saddam Hussein on their briefing slides
with the word Leadership. Cite.
41
In the 2003 war, the military efforts against leadership sites was even more intense than in 1991. Whereas the
1991 air war lasted 43 days, whereas the high-intensity air operations in 2003 only spanned roughly 20 days from
March 19th until April 9th when Baghdad fell. In that shortened time, U.S. aircraft struck 1,799 targets with the
purpose of suppression of Iraqi regimes ability to command Iraqi forces and govern State. Furthermore, unlike in
1991, virtually all these strikes used precision-guided munitions. Operation Iraqi Freedom By the Numbers,
Assessment and Analysis Division, USCENTAF, 20 April 2003, pp. 4-5.
on enemy leadership, in particular, undermine escalation control. But attacks that spare these
targets substantially reduce the effectiveness of conventional military forces.
The effective but escalatory style of conventional warfare is not going away, and is not
shelved when dealing with nuclear-armed enemies. The over-arching U.S. concept for military
operations against China in the coming decades, called Air Sea Battle, calls for U.S. air and
missile strikes against large numbers of radars, communications nodes, and other command and
control targets across the Chinese homeland.42 A senior U.S. Navy planner, defending the
concept of Air-Sea Battle, told us, If we fight China, weve got to do it meaning the U.S.
military has to attack command and control sites in China. If we dont strike those targets, we
cant move ships [into the Western Pacific]. Im not sure we can operate our air [from regional
bases] either.43
Nor, apparently, is the U.S. planning for restrained conventional operations to prevent escalation
on the Korean Peninsula. We interviewed U.S. military officers who plan operations on the
Korean Peninsula, as well as officers who were recently in command positions in Korea. We
asked a group of air war planners: What types of targets would be in your plan, but were
removed to avoid escalation? Their answer: nothing. One of them explained: We pick targets
with two goals: kill red and protect blue that is, to destroy enemy forces and protect our own.
We asked them if there was flexibility in the plan to allow the United States to avoid striking
North Korean command and control, leadership, and radars or limit U.S. air operations
geographically. They were skeptical. Its a really small airspace, said one officer. About the
prospect of removing command and control and leadership targets from the war plan, one officer
quipped, That is the plan meaning that the strikes on those types of targets constitute the core
of the air operation.44 We interviewed a U.S. Air Force officer soon after he returned from a
command position in Korea about the nature of the U.S. air war plan. We described to him the
first few days of the U.S. air campaign in 1991 against Iraq, and the heavy focus on air defenses,
command and control, and leadership, and suggested that it might mirror U.S. planning for the
Korean Peninsula. Yep, thats the plan. Its just how we fight.45
A second aspect of warfare that will exacerbate the incentives of the weak to use nuclear
weapons, and inhibit efforts to control combat, is the powerful tendency among military
organizations to strike the most lethal weapons systems of their enemies. If the hope for
preventing escalation is based on assuring adversary leaders that they can retain power as long as
42
Andrew Krepinevich, et al. Meeting the Anti-access and Area-denial Challenge. Washington, D.C.: Center for
Strategic and Budgetary Assessments, 2003; Michael McDevitt, The evolving maritime security environment in
East Asia: Implications for the US-Japan alliance, Honolulu, HI: CSIS Pacific Forum. May 31, 2012; Thomas P.M.
Barnett, Big-War thinking in a small-war era: The rise of the Air-Sea Battle concept, China Security 18 (2010).
On the escalatory risks of Air Sea Battle, see Raoul Heinrichs, Americas dangerous battle plan, The Diplomat,
August 17, 2011; Greg Jaffe, U.S. model for a future war fans tension with China and inside Pentagon,
Washington Post, August 1, 2012; Air-Sea battle plan renews old hostility, Global Times (China), November 14,
2011.
43
Discussion with authors, Airlie, VA, 2011. [April 28, 2012.]
44
Discussions at PACAF 2012.
45
Discussion in Cambridge, MA, 2010.
they do not use weapons of mass destruction, then attacks on their WMD sites and delivery
systems clearly open the door for follow-on operations to overthrow them. Stated differently, if
an enemys weapons of mass destruction are its leaders get out of jail free cards, then efforts
to destroy those weapons will pose an existential threat forcing him to escalate to coerce an end
to those attacks. Attacking an enemys strategic deterrent assets is, therefore, highly escalatory.
The problem is that military organizations in general, and the U.S. military as a case in point,
have a powerful proclivity to target those weapons during war.46
The military logic for attacking an enemys strategic assets whatever weapons it holds in
reserve as its ultimate deterrent threat47 is straightforward. During war, military organizations
seek to destroy the enemys military, and they logically place a high priority on neutralizing
those enemy forces that threaten to inflict the greatest damage. The notion of leaving intact an
enemys most potent weapons is deeply counterintuitive for most military planners. This
resistance toward restraint during combat is rooted in fundamental ideas about the nature of
warfare. Carl von Clausewitz argued that warfare tends to escalate to extremes, and his views
have been internalized by many Western militaries, including that of the United States.48
Restraint has come to be seen as a recipe for defeat. If war is likely to escalate, then sparing an
enemys most lethal weapons systems is foolish. Those systems should be attacked at the very
outset of a war, before they have been used, and when (in many cases) they are most vulnerable.
But the mission that militaries abhor more than either is deterrence.50 While defense, at least,
involves preventing an adversary from inflicting damage, deterrence requires accepting a
condition of perpetual vulnerability. Deterrence succeeds if the enemy chooses not to attack.
The notion of relying upon deterrence and accepting ones vulnerability even in the midst of a
war rather than striking and neutralizing the enemy weapons systems, is anathema to most
military planners. For many of them, deterrence is perhaps an appropriate strategy for
peacetime, but once the bullets start flying, military operations must be directed at victory.
46
Our arguments in this section parallel the analysis in Posen, Inadvertent Escalation. He notes that intentional
conventional attacks designed to alter the nuclear balance of power are clearly escalatory so much so that their
consequences should not be considered to be inadvertent escalation and are hence beyond the scope of his study.
(p. 2; and the discussion in Inadertent Escalation in chapter 4. In this section, we focus on the escalatory
consequences of conventional operations against an adversarys strategic forces, not merely to alter the balance of
power but to completely disarm the adversary. As we demonstrate in the paragraphs below, what was once
understood to be highly escalatory is now simply standard practice among U.S. conventional war planners.
47
As discussed earlier, nuclear weapons have attributes that make them ideal for this mission: being a regimes final
deterrent threat. However, non-nuclear states may rely on other weapons for this purpose, such as biological or (less
usefully) chemical weapons.
48
Carl von Clausewitz, On War.
49
Snyder, SVE, Posen.
50
Cite.
The reluctance to rely upon deterrence once shooting begins can be seen in U.S. military plans
during the Cold War. Throughout the four decades of confrontation, U.S. nuclear war plans
remained focused on so-called counterforce attacks: destroying Soviet nuclear weapons. In the
1950s, when the United States believed it could win a nuclear war, its plan for fighting the Soviet
Union called for a massive nuclear disarming strike against every known Soviet nuclear target
and every Sino-Soviet Pact airfield that could possibly launch a bomber.51 But even later in
the Cold War, when the Soviets had achieved nuclear stalemate and the prospect of winning a
nuclear war had become remote, NATO military plans still called for intense conventional
attacks on Soviet nuclear forces. For example, in the 1980s, NATOs plans called for operations
against the Soviet northern flank, including operations by U.S. attack submarines to sink
Soviet ballistic missile submarines at sea.52
In the post-Cold War world, each time the United States fights a regional war, a major element
of the air campaign is to destroy the enemys strategic assets. In 1991, the U.S. air campaign
included intense attacks on Iraqi WMD sites and suspected delivery systems. More than one
hundred WMD-related targets were struck on the first night alone, and nearly six hundred WMD
targets were attacked during the first five days of the campaign comprising 15% of all U.S.
strikes.53 The United States prioritized potential WMD targets in NATO operations against
Serbia in 1999, and once again in 2003 against Iraq the air war plan for Operation Iraqi
Freedom identified 1,840 targets associated with the delivery systems for Iraqs (nonexistent)
WMD program.54
This emphasis on neutralizing adversary strategic assets is undiminished today. Among the
highest priorities of the U.S. Department of Defense are the development and improvement of
sensors, weapons systems, and doctrines to facilitate the destruction of enemy strategic weapons.
The effort to neutralize those forces spans all four services of the U.S. military, and includes
every domain of warfare including ballistic missile defenses, anti-submarine warfare programs,
long loiter-time UAVs, prompt global conventional strike systems, offensive cyber operations,
accurate nuclear weapons, and a broader range of efforts designed to allow the U.S. military to
identify and rapidly target mobile missile launchers. The military is explicit about the purpose of
these efforts: to defeat enemy WMD because they have no confidence that wartime deterrence
will hold. A U.S. Air Force officer who had recently held a command position on the Korean
peninsula told us, We plan to hit North Korean nuclear facilities from day one [of a conflict].
When we suggested that those attacks might force Pyongyang to use its weapons, he said, But
51
See Chapter 4.
52
Posen, Inadvertent Escalation, chapter 4; and John J. Mearsheimer, A Strategic Misstep: The Maritime Strategy
and Deterrence in Europe.
53
Includes strikes on targets identified as NBC [nuclear, biological, chemical] and strikes on targets associated
with SCUD missiles. The numbers here only cover the first five days of the air war and hence do not include air
attacks as part of the SCUD hunt later in the war which arguably were intended to suppress Iraqs punitive
conventional missile strikes on Israel and Saudi Arabia rather than destroy WMD delivery systems. The data is
from GWAPS, Vol. 5, Table 178.
54
This figure counts the targets in the WD target sets in the Joint Integrated Prioritized Target List, 832 of
which were eventually struck. Operation Iraqi Freedom By the Numbers, Assessment and Analysis Division,
USCENTAF, 20 April 2003, pp. 4-5.
we assume theyre going nuclear anyway. So we might as well destroy them as quickly as
possible.55
When we discussed escalation risks on the Korean Peninsula with a Pentagon official whose
responsibilities include defense policy in that region, he suggested that the problem was not as
great as we imagined, because Well be turning off their nuclear weapons on the first day of the
war.56 He used the words turning off in a deliberate fashion (i.e., rather than more typical
terms, like destroying or suppressing), suggesting that the scope of U.S. counterforce efforts
may be broader than we envision. More importantly, his response is another piece of evidence
that defeating enemy WMD i.e., neutralizing the weapons that guarantee the enemys survival
is simply a matter of course in U.S. regional war plans.57
The presumption that the job of the military is to neutralize or destroy the enemys most lethal
weapons systems during war does not indicate some pathology within the U.S. military. Rather,
it is based on a Clausewitzian understanding of the nature of warfare that is shared by many
militaries around the world. Additionally, the direction of technology will make enemy strategic
assets an even more alluring target in the future. Real-time target intelligence from a range of
sensors, stealthy aircraft and missiles, and highly precise weapons systems (plus unconventional
military means) will make targeting enemy strategic forces seem easier and more attractive.
A third major hurdle standing in the way of escalation control lies in the fundamental difficulties
that civilian and military leaders face in seeking to control military operations. Preventing
escalation by carefully limiting conventional operations for example, striking certain targets
but leaving others off-limits, or advancing to certain lines but not beyond demands a level of
fine-grained control over military operations that is at odds with the nature of modern warfare.
To a large degree, controlled warfare is an oxymoron. In addition to the fog and friction
that are a constant presence in warfare, senior civilian and military leaders lack the in-depth
understanding of the details of military plans that is essential for ensuring that plans even at the
tactical level are consistent with the critical goal of escalation control. Furthermore, during
crises or wars, the nature of military organizations and the complexity of modern warfare make it
impossible for senior civilian or military leaders to exercise fine-grained control over operations.
As a result, efforts to avoid escalation by fighting in a highly restrained manner, which would
signal restraint to the enemy, can be undermined by tactical decisions on the battlefield.
One major hurdle to waging controlled, limited war is that senior civilian and military leaders
often lack the in-depth knowledge of war plans required to ensure that the operational details
align with overarching strategic goals. Senior leaders are understandably focused on strategy;
55
Discussion in Cambridge, MA, 2010. Discussions in 2013 with U.S. military officers who are directly involved in
Korea plans confirmed that this approach to conventional war on the Peninsula has not changed.
56
Discussion at Pentagon, 2012.
57
As an important caveat, we do not get the sense that this is true with respect to regional war plans against China
we have never heard anything that suggests that U.S. conventional forces would intentionally target Chinas nuclear
forces. As discussed in the previous section, of course, the nature of U.S. conventional plans for China
contingencies would be escalatory because of the intensity of the effort against command and control sites, but we
have no reason to suspect that intentional strikes on Beijings nuclear forces are part of the plan.
lower-level commanders are appropriately focused on tactics. But there is frequently insufficient
attention across those domains to ensure that tactical choices are consistent with strategic
objectives.58 To some extent, the separate domains of senior leaders and battlefield commanders
are necessary. Those in leadership positions lack the time and few civilians have the expertise
to go through war plans line by line and interrogate each tactical choice. Moreover, many
experts would consider the micromanaging of war plans by senior leaders to be inappropriate,
arguing that civilian leaders should leave operational details to the military, and that military
leaders outside the theater should avoid second-guessing the decisions of local commanders.59
But in wars in which a key strategic objective is preventing escalation which is to be
accomplished by restraining ones military operations (and signaling restraint to the enemy)
many battlefield decisions have strategic consequences. Unless senior leaders know what is in a
war plan, they cannot determine whether tactics and strategy are coherent, or whether the trade-
offs between battlefield effectiveness and escalation prevention have been made in line with their
preferences.
A second major hurdle to escalation prevention is that when war erupts, finely calibrated control
over military operations is impossible. Modern conventional combat is so complex, and requires
the coordination of so many people, that military organizations by necessity rely on standard
operating procedures (SOPs). Military doctrine establishes a right way to suppress enemy air
defenses; to protect ones warships from enemy submarines; degrade an enemys military
command-and-control. When civilian authorities order the military to conduct a military mission,
commanders build complex operations on top of the foundation of the SOPs. But as a result,
leaders, especially civilian leaders, often do not appreciate the military implications of their
orders, making control of operations far more difficult.
The Cuban missile crisis provides many of the canonical examples of standard operating
procedures impeding careful civilian control of the military.60 The White House developed a
plan with two key components to defuse the crisis. First, U.S. leaders sought to buy time by
avoiding actions that would create an imminent clash between U.S. and Soviet forces, and which
might therefore rapidly escalate to war. To that end, a blockade was selected as the initial U.S.
reaction, rather than airstrikes on missile sites, or an or invasion of Cuba. Furthermore, the
location of U.S. naval forces implementing the blockade was subsequently adjusted during the
crisis to delay the interception of Cuba-bound ships. The second principle component of the
White Houses strategy: begin military preparations for war, which the Soviets would observe, to
pressure the Soviets to back down.
58
In most military operations, this division of labor between leadership and battlefield commanders is appropriate.
A battlefield commanders tactical decisions e.g., how to assault a bridge, which aircraft will strike which enemy
targets would not be expected to have strategic implications for the war meriting the attention of high-level
leadership. However, in a war in which a primary goal is to avoid escalation, tactical decisions about which targets
to strike, where to fly, and what radars to jam, can have strategic consequences, and directly affect war outcomes.
59
Cites on civ-mil relations. Cohen?
60
In recent years, the history of the Cuban missile crisis has been refined, corrected, and amended based on
increased availability of documents detailing key perceptions and decisions during the crisis. The new material does
not contradict either the point made here about SOPs, nor the examples offered here. See for example, XXX. For
an older version of these events which makes this point, see. Allison
The problem was that some of the policies the White House chose to implement its strategy
triggered SOPs that contradicted their core objectives. As part of the naval blockade of Cuba,
the U.S. Navy following SOPs dropped depth charges (underwater explosives) near Soviet
submarines. This SOP, unknown to the White House, contradicted the goal of avoiding a clash
between U.S. and Soviet forces, triggering a heated argument in which Secretary of Defense
Robert McNamara demanded to know who had authorized the U.S. Navy to drop depth charges
on Soviet submarines, only to be told by the Chief of Naval Operations: you did.61 The problem
was that McNamara did not know what, exactly, the President was setting in motion when he
ordered the Navy to blockade Cuba. In a second incident, President Kennedy ordered the
Strategic Air Command (SAC) to raise the alert level of U.S. nuclear forces during the crisis, to
signal the Soviets of U.S. seriousness. Unfortunately, officials in the White House did not know
that the Strategic Air Commands SOPs called for them upon an alert to send aircraft into
Soviet airspace to gather pre-strike reconnaissance on nuclear targets.62 Once again, efforts to
wield military forces in a calibrated fashion to signal an opponent led to military actions that the
leaders did not desire and believed were counter-productive.63
More recent examples demonstrate the enduring nature of this problem. In the 1991 war against
Iraq, the overarching U.S. objective was to conduct a limited offensive to eject Iraqi forces from
Kuwait, and heavily damage the Iraqi militarybut not conquer Iraq. These limited objectives
were critical for two reasons: because, as US officials explained, they did not want the United
States to occupy Iraq; and because they were trying to deter Iraq from using its stocks of
chemical and biological weapons. The problem is that the two highest priority targets in the U.S.
air campaign were, as described above, Saddam Hussein and the other most-senior Iraqi leaders,
as well as Iraqs WMD and delivery systems. We may never know why Saddam Hussein did not
respond by using WMD in 1991, but as an example of an effort to limit war for the purpose of
preventing escalation, the U.S. air operation illustrates the grave problems ensuring that military
operations are consistent with political objectives.
The problem of ensuring the integration of military operations and strategic objectives endures
today. When we spoke with civilian officials in the U.S. Defense Department about war plans for
the Korean peninsula, they explained that, while escalation is always a risk, steps were being
taken to mitigate those dangers. Specifically, the Guidance on the Employment of Forces (GEF)
the principal directive from the Joint Chiefs of Staff that guides combatant commanders for the
construction of their war plans directs the Combined Forces Command in Korea to create plans
that would achieve U.S. wartime objectives, one of which is to avoid escalation. However, when
we talked to air war planners in the theater, and asked them what operations they had rejected out
of escalatory concerns they said: nothing. More pointedly, when we asked military officers who
oversee war planning for the Pacific whether U.S. concepts of conventional operations might
pressure adversaries to escalate, they dismissed these concerns. One officer responded, Thats
not our job. Another officer explained, Our job is to win the conventional fight. If the bad guy
61
[NOTE: Check newest history on this. Did this really happen, or one of the many myths of Cuban Missile
Crisis?]
62
The Air Force additionally went ahead with a scheduled ICBM test at Vandenberg Air Force Base, because no
one had ordered them to abstain from actions that might be misperceived as the launch of a nuclear strike.
63
For an account of the reactions by senior Kennedy Administration officials upon learning of these unwanted
provocative steps, see[On blockade, see Kennedy Tapes; and source on McNamara CNO dispute.]
goes nuclear, the Presidents got other people to call to deal with that. In other words, although
the leadership has identified escalation prevention as a key goal, that objective does not appear to
be guiding actual war planning.
In sum, the nuclear escalatory dangers raised by conventional war are far greater than is
commonly recognized. At a fundamental level, weak states face powerful incentives to escalate
against strong states. And several factors stemming from the nature of modern warfare
exacerbate this danger. Strong states with sophisticated militaries like the United States will
inevitably target weaker states command and control systems and strategic assets at the outset of
a conventional conflict. This compounds the weaker regimes fear of not surviving the conflict,
and thus makes nuclear escalation as a means of forestalling defeat more likely. Moreover,
efforts by stronger states to restrain such attacks for the purpose of controlling adversary
escalation are likely to be frustrated by the inherent difficulty of calibrating military operations in
wartime. In the United States, military and political leaders appear either unaware of the extent
of the problem or ill equipped to address it within the normal chains of command. The efforts
these leaders might try to institute to avoid provoking nuclear escalation are essentially
incompatible with the modern approach to conventional operations adopted by the U.S. military,
which is highly (if unintentionally) escalatory.
The logic developed here suggests three conditions that should help identify the most escalation-
prone conflicts: (1) those in which nuclear armed states are vulnerable to conquest; (2) those in
which their leadership is vulnerable to overthrow; and (3) in which nuclear armed states are
likely to face an enemy waging warfare using modern (blinding and disabling) operations.
The factors that make conquest of the losing state likely in some wars and less so in others
are varied, and no simple rulebook that can be applied to every case. Nevertheless, the crucial
conditions probably include geography (the size of the weaker states territory, and perhaps the
location of key cities), population, the ethno-religious-sectarian identity of the combatants, and
the military force structure of the stronger state. For example, it is inconceivable that the United
States would ever try to conquer China for many reasons. For Pakistans leaders facing a
major military offensive by India, the case is less clear-cut. On the one hand, Pakistan is a large
country with a big population, and India has no desire to conquer and rule over 180 million
Pakistani Muslims. On the other hand, most of Pakistans largest cities including Karachi,
Lahore, Faisalabad, Rawalpindi, and its capital Islamabad are within approximately 100 miles
of the border with India. Pakistans leaders might reasonably worry that a major conventional
war could lead India to seize, or isolate, major Pakistani cities, to be surrendered at some future
time of Indias choosing unless Pakistan can use nuclear escalatory threats to prevent this.64
A second key condition that may affect the propensity of leaders to escalate conventional wars
rather than accept defeat is their vulnerability to coups or revolutions.65 To avoid losing
power, a countrys leaders must do more than merely deter a wartime enemy from conquering
them; they must also prevent the war from unleashing dynamics within their own country that
trigger a coup or revolution during or after the conflict.66 Countries whose governments face
substantial domestic opposition will, therefore, face intense pressure to coerce an end to
hostilities rapidly, before regime opponents are emboldened or before too much damage is done
to the regimes internal security forces.67 Furthermore, for many governments, the greatest threat
to regime survival comes from the countrys own military. If a war results in the devastation of a
countrys military forces and the humiliation of the military leadership the risk of a coup
surges. As a result, leaders must compel a ceasefire long before enemy forces start to advance
on their capital; they must create a ceasefire before domestic opponents sense weakness, before
regime security elements are too degraded, and before its military becomes too demoralized and
angry at the political leadership. The implication for regimes that face considerable internal
opposition is clear: if a war begins to go badly, find a way to create a battlefield stalemate
immediately or face the prospects of a revolution or coup.
64
The framework we create here to investigate risks of coercive nuclear escalation is intentionally simple to allow
for easy application but should not obscure the gradations in the factors we describe. For example, while states
that are susceptible to conquest are expected to be more likely to employ coercive escalation than those who have no
fear that war will lead to loss of territory, some states face the prospect of loss of key territories short of conquest.
A major Indian conventional offensive might exploit Pakistans narrowness to cut the major lines of communication
between North and South. Similarly, if Chinas leaders are as committed to Taiwan as their public positions imply,
they may see a war that leads to Taiwans independence as a loss which while far less than conquest nevertheless
means the loss of valuable territory. The conquest factor is a useful heuristic to distinguish higher escalatory risks
from lower risks, but such categories and dichotomies should be applied with care.
65
The research project was framed to investigate the incentives of weak states i.e., defined as those that expect to
lose conventional wars to use nuclear weapons to stalemate their opponents. The research was framed in that
fashion because of the value of nuclear weapons as a stalemating device, because of the substantial history of states
relying on nuclear weapons to compensate for conventional weakness, and because of the U.S. interest in deterring
its weak adversaries from escalating. However, the vulnerability of leaders in some countries to coups and
revolutions, either during or after wars, and the escalatory pressures that those vulnerabilities create (to coerce a
rapid cessation of hostilities) should pressure both weak states and stronger ones to escalate to rapidly end a conflict
that threatens their domestic power at home. In other words, there may be escalatory pathways that would lead
Russia, China, or other states with domestic stability concerns to escalate a war with a weaker power.
66
See for example the uprisings triggered in the wake of the 1991 Persian Gulf War: the signs of chaos within
Saddam Husseins government, and the appearance that his regime was on its last legs, emboldened Shiite groups,
principally from Southern Iraq, to rebel against Saddams rule.
67
In many authoritarian regimes, popular uprisings are prevented by collective action problems: if every regime
opponent rebelled at once, they could overwhelm the regimes internal security forces, but no group wants to be the
first: it would be slaughtered, and there is no guarantee that other groups or individuals would follow their lead.
Oppressive governments exacerbate these collective action problems by using spies and informants to make
coordination against the regime too dangerous.The seminal work on collective action problems is Mancur Olson,
The Logic of Collective Action: Public Goods and the Theory of Groups (Cambridge, Harvard University, 1965). On
authoritarian regime survival strategies see Daniel L. Byman and Jennifer M. Lind, Pyongyangs Survival Strategy:
Tools of Authoritarian Control in North Korea, International Security, Vol. 35, no. 1 (Summer 2010): 44-74.
The third key factor that may make leaders more inclined to employ nuclear weapons coercively
stems from the style of war waged by their opponents. If leaders are driven to escalate by fear of
conquest or post-war regime change, then certain kinds of military operations are likely to
exacerbate that fear. For the reasons we discussed above, operations that aim to blind enemy
command and control, directly target political leaders and regime security elements, or degrade
strategic deterrent forces will intensify adversary incentives to escalate. Conflicts in which one
or more of the combatants have conventional doctrines that rely on such blinding attacks will be
particularly prone to escalation.
Table 1 summarizes these dangers in the context of several wars that could plausibly occur in the
coming years. The column on the far right aggregates the three worrisome conditions
conquest, regime vulnerability, and the propensity of the combatants to wage conventional war
by blinding and disarming the enemy and indicates the resulting escalatory risks if
conventional war were to occur. The darker shade in that far-right column indicates that all three
worrisome conditions are present; the lighter shade means that one or two of the dangerous
conditions would exist during a conventional war.
Several of the implications from Table 1 are striking. First, although it is widely appreciated
within the U.S. national security community that a major war between India and Pakistan would
entail dangerous escalatory risks a judgment we share the logic developed in this paper and
summarized in Table 1 suggests that several plausible U.S. regional wars are even more prone to
escalation. The likelihood of escalation seems greatest on the Korean Peninsula: the Pyongyang
government has every reason to expect that a major military defeat equals regime change, with
calamitous consequences for the existing leadership. Regime change could occur as a result of
intentional U.S. / ROK policy i.e., if leaders in Washington and Seoul choose regime
change as the wars desired endstate. But the Pyongyang government might fall even if the
U.S. and ROK pursue limited objectives: the damage inflicted on the North Korean military and
security services may sufficiently weaken the regime and trigger a coup or revolution.
Furthermore, CFC military operations will likely seek to blind the North Korean command and
control, destroy leadership sites, and perhaps degrade their strategic weapons. If war erupts on
the Korean Peninsula, preventing escalation will be a very difficult challenge.
68
A war over Taiwan would not make leaders in Beijing fear conquest in the narrow sense of the term, but it
would raise the risk of the loss of highly valued territory (Taiwan). While not as bad as complete conquest and
occupation, might entail a major loss to the core interests of a state. If Chinese leaders are to be believed that they
see Taiwan as an inseparable part of China, and especially if the people of China feel the same, it may be too costly
to leaders in Beijing to accept defeat in a war over Taiwan, especially if the consequences might be Taiwanese
independence.
69
Iran is not believed to have nuclear weapons. This row indicates the escalatory risks in a future war in the Strait of
Hormuz if Iran has subsequently acquired nuclear weapons.
70
As described above in the text, the risk of outright conquest of Pakistan by India appears low; however, most key
Pakistani cities are very close to the Indian border, so Pakistans leaders may reasonably fear limited territorial
incursions by India that would isolate critical Pakistani population centers. This is reflected in the No/Yes value
under the Conquest plausible? column.
Second, a conventional conflict in maritime East Asia between the United States and China may
entail far greater nuclear escalation risks than is commonly recognized. Because the conquest of
China is not plausible, many analysts assume that the escalation risks in a U.S.-China clash are
substantially muted. But that optimistic assumption overlooks two critical facts, which are
highlighted in Table 1. Namely, Chinas leadership may not be able to survive the political
repercussion of suffering a humiliating conventional military defeat at the hands of the United
States, and that the U.S. style of conventional operations including large numbers of strikes on
the Chinese mainland to blind Chinese sensors and degrade military command and control may
exacerbate these escalatory risks.
Finally, Table 1 highlights what might be the greatest danger associated with Iran acquiring
nuclear weapons. Even if Iran is deterrable and hence does not seek nuclear war against the
United States or U.S. allies the dynamics of conventional operations in the Persian Gulf may
force the hand of leaders in Tehran. Specifically, operating naval forces in the constrained
waters of the Persian Gulf during a war might compel the United States to greatly degrade Irans
air defense network, surface search radars, and military command and control and there would
be powerful pressures on the United States to also degrade the systems that could deliver Irans
nuclear weapons. In the context of decades of U.S.-Iranian hostility, and repeated U.S.
statements about the desirability of regime change in Tehran, the pressure on an Iranian
government to coerce a rapid end to hostilities would likely be intense. Nuclear escalation
directed against U.S. facilities in the region, or the facilities or cities of U.S. regional allies
would be one of Irans main options.71
Counterarguments
Critics might concede that conventional wars between nuclear-armed adversaries would be
highly escalatory, yet counter that such wars are unlikely to occur in the first place. In fact,
critics might say, the arguments that we present here about the dangers of wartime escalation are
exactly the reason that these conventional wars will not occur. As Kenneth Waltz argues,
nuclear weapons do not merely deter nuclear attacks; they deter conventional attacks as well. As
he explains, launching a major conventional offensive against a nuclear-armed state would be
foolhardy; yet, launching a limited conventional attack would be equally senseless as the small
potential gains would be trivial compared to any residual risk of escalation. 72 In short, critics
might argue that it is precisely because our arguments about the danger of escalation are correct
that these wars will not happen.
The lack of high-intensity conventional war between two nuclear weapon states is evidence on
the side of Waltz, but there is worrisome evidence, as well. First, if Waltz is right that the risk of
nuclear escalation will reliably deter conventional attacks, then conventional attacks on nuclear-
armed countries should not occur yet they do. In some cases these were highly limited
71
The logic of Iranian nuclear escalation, and Tehrans escalatory options, would be directly analogous to North
Koreas escalatory logic and options during a conventional war, as described above (pp. 13-26). In the Iran case,
potential targets for an initial coercive escalatory strike might include a U.S. military bases (e.g., Al Udeid, or NSA
Bahrain) or a city in a regional ally.
72
For example, see Kenneth N. Waltz and Scott D. Sagan, The Spread of Nuclear Weapons: A Debate Renewed
(New York: Norton, 2003).
conventional operations, in locations whose geography limited the fighting (e.g., Kargil 1999;
Falklands 1982).73 But on other occasions, countries have launched major conventional military
operations that inflicted substantial losses on nuclear-armed adversaries, or which threatened
their vital interests. In 1950 China launched a major land attack against U.S. and allied forces on
the Korean Peninsula, dealing the United States a major defeat, denying the United States victory
on the Korean Peninsula, and killing thousands of U.S. military personnel. Whatever
calculations led Chinas leaders to believe they could inflict such a serious defeat on the United
States without prohibitive risk of nuclear escalation surely does not resemble the line of
reasoning and the overwhelming caution that Waltz expects to observe in states facing
nuclear-armed enemies.
Further, the Syrian attack on the Golan Heights at the outset of the 1973 Yom Kippur War
reflects a level of risk acceptance that does not jibe well with Waltzs arguments. On October 6,
five divisions of Syrian ground forces launched a major surprise attack on Israeli defenses along
the Golan Heights. The Syrian ground forces nearly broke through the Israeli line; at the worst
moment for Israel, roughly a dozen tanks stood in front of the Syrian Army and there were no
additional Israeli reserves between the Golan Heights and Tel Aviv. (Some accounts of the war
claim that Israel took steps during the war to prepare its nuclear arsenal in case the Syrian Army
broke through.) Syria was fortunate: its attack on the Golan Heights failed. But their decision
process does not match the level of caution one will require if conventional wars against nuclear-
armed states are to be banished.74 More recently, the apparent North Korean sinking of a South
Korean warship in 2010, or the Norths shelling of Yeonpyeong Island near Seoul, could have
led to a substantial conventional response by Seoul triggering war.75 Waltzs view may
correctly explain Seouls reluctance to respond to those attacks with force; but it does not explain
Pyongyangs willingness to instigate violence and keep walking along the edge of war.
More broadly, the claim that the risk of catastrophe will reliably deter conventional wars seems
to contradict much of history. For most of history, starting a war meant risking catastrophe.
Leaders who lost surrendered not merely their crowns, but also their heads. In the era of
dynastic succession, defeat often meant that ones children were killed as well to prevent future
claims to rule. Throughout history, those who led rebellions against ancient empires, colonial
powers, or even against modern occupiers usually paid with their lives (and often died
gruesomely). And the populations on whose behalf the insurgents rebelled were sometimes
slaughtered, to teach others not to emulate their disloyalty. In more modern times, the Japanese
who planned Pearl Harbor understood that they were attacking a country with ten times their
economic power, and they understood that if the war went badly it meant catastrophe for
themselves and Japan. (They were right.) But despite those risks, the Japanese attacked.
73
The Kargil conflict in 1999 involved small units fighting to control a handful of mountain peaks in the Kashmir
region. The high altitude and mountainous terrain greatly limited the scale of conventional operations. The
Falklands war involved Argentinas attempt to take control of the disputed Falklands / Malvinas islands. The nature
of the fighting, on the small, remote Islands in the South Atlantic, and in the sea and are around them, greatly limited
the scope of the fighting. Note, however, that according to Waltzs logic, neither the Argentinians nor the Pakistanis
should have attacked: the risk of escalation was low, but so were the potential gains from victory.
74
For a detailed account of the Syrian offensive and the desperate fighting on the Golan Heights, see Trevor N.
Dupuy, Elusive Victory: The Arab-Israeli Wars, 1947-74.
75
Choe Sang-Hun, South Korea Publicly Blames the North for Ships Sinking, New York Times, May 19, 2010;
Sang-Hun, South Korea Returns Fire After Shots From North, New York Times, August 10, 2011.
Germanys leaders understood that they were risking personal and national calamity when they
invaded France, and especially when they invaded the Soviet Union. But they attacked anyway.
In 1980 Saddam Hussein invaded Iran, a country with three times Iraqs population a gamble
that nearly led to his overthrow and death. If it were true that leaders do not start conventional
wars if the possibility of catastrophe looms, human history would be much more pacific.
To be clear, we agree with the premise underlying Waltzs argument: that conventional wars
could only occur between nuclear-armed states if leaders were willing to embrace major risks.
He does not think that will happen; we see that occurring throughout the pages of history. If
leaders were not willing to take enormous risks, China and Syria would not have launched major
ground attacks on nuclear-armed states, people would have never rebelled against empires, and
few of the major wars of the modern era would have occurred.
Conclusion
What are the implications of this article? One set of implications follows from the first half of
this chapter on the incentives of the weak to escalate. Most scholars and analysts of deterrence
dismiss the likelihood of intentional nuclear attack by one state on anther, and especially dismiss
the possibility of an intentional nuclear attack by a country on the United States. Similarly,
within the U.S. military, U.S. regional war plans treat conventional war as the base case and
at most consider the consequences of nuclear escalation in annexes to the plan. But those
scholars, analysts, and war planners should be pushed to explain why they believe that
adversaries will keep their most powerful weapons on the sidelines, even as they suffer terrible
military defeats. Historically, weak states with nuclear weapons planned to use them in an
escalatory fashion to prevent military defeat. Even today, that strategy is Russias and Pakistans
stated nuclear doctrine, and it is also probably Israels doctrine, if that state were to suffer an
unexpected military collapse. If the risk of wartime escalation is high, as we have argued here,
then the field of political science needs to develop richer theories of escalation control in
regional conflict. Furthermore, those who are planning U.S. nuclear force structure for the
coming decades must re-focus on the deterrence mission, and evaluate proposed force structures
against the requirements for the most demanding and most likely conditions for nuclear
deterrence: preventing escalation during a conventional conflict.76
With respect to the second half of the chapter on the escalatory nature of modern war and the
difficulties of controlling military operations the implications are less straightforward. The
accumulated experience of U.S. military planners is that destroying the enemys command and
control is essential for producing major military advantages on the battlefield, and is the sort of
one-sided military outcome that the United States has come to expect. Dropping those targets
from various war plans would come at a steep cost: in terms of wartime casualties and the
prospects for victory. Even worse, if war were likely to escalate in any case, leaving an
adversary with its most lethal weapons would be a grave mistake. But attacking an enemys
command and control capabilities while striking its strategic deterrent weapons is virtually
tantamount to forcing the enemy to escalate.
76
See, for example, Keir A. Lieber and Daryl G. Press, The Nukes We Need: Preserving the American Deterrent,
Foreign Affairs (November/December 2009), pp. 39-51.
What is clear is that these decisions attacking or sparing enemy command and control and
strategic forces have strategic consequences and should be debated and decided at the
highest levels of government. Currently these decisions are apparently being made de facto
without substantial political oversight by conventional war planners who are simply doing their
job: planning to kill red and protect blue, in the words of one. Senior government officials
may not have the flexibility to modify complex conventional war plans in the midst of a major
crisis or the first days of war; they need to understand now the benefits and risks of fighting
conventional war in the manner those conventional military planners prefer. And political leaders
should have sufficiently rich military options available to them to allow them to tradeoff those
costs and benefits exactly as they prefer
In 2011, the U.S. military ran a political-military war-game to explore the dynamics of a U.S.-
China air- and naval-clash in the Pacific. Former senior U.S. officials played the U.S. decision
makers, and U.S. military officers played the key U.S. military roles. U.S. government experts
with substantial knowledge of China, its political leadership, and its military played Chinas
leaders.
According to one account of the game, during an early move the United States conducted a set of
conventional military strikes against Chinese long-range sensors and command-control targets
a reasonable move given the Air-Sea Battle plan to help the United States win the air and naval
engagements. The China team perceived the U.S. conventional strikes to be highly escalatory
and elected to use nuclear weapons. But the umpires for the game disallowed Chinas move.
On the next move of the game, the U.S. Pacific Command continued to execute its conventional
war plan, and China again elected to use nuclear weapons. Again the umpires disallowed
Chinas move. This apparently happened on at least three moves of the game.
Arguably the most alarming aspect of this story is not what happened during the game i.e., that
a group of China experts in the U.S. government thought that Beijing would use nuclear weapons
given the nature of planned U.S. military operations. What is most alarming is that the official
report describing the war game never mentioned Chinas repeated decisions to cross the nuclear
threshold. Because the umpires ruled out the move, it never happened. Hence, what should have
been the most important finding from the game the potential link between U.S. conventional
war plans and Chinas incentives to use nuclear weapons was not highlighted as a central
finding of the exercise. Based on our own conversations with U.S. military planners, the goal of
preventing escalation is not deeply embedded in U.S. conventional war planning in the Pacific
region, and probably not elsewhere either.
Taking a step back, the findings from this chapter should be startling for most defense and
international security analysts. Historically, we know that weak countries have planned to use
nuclear weapons coercively to stalemate the strong. We know that the United States and its
NATO allies planned to do exactly this when they felt weak. We have contemporary public
statements by several so-called weak countries that confirm that they currently view their
nuclear weapons in this manner, and plan to use their weapons to stalemate powerful enemies.
We have war games that suggest that potential U.S. adversaries like China would face
powerful incentives to escalate if we conduct conventional operations against them as we
currently plan. And we know that U.S. conventional war plans across the Pacific and in other
theaters as well envision the United States fighting by blinding our enemies and, in many
cases, targeting adversary strategic weapon systems.
And yet, in the U.S. nuclear weapons community, the disarmament community, and among
many in the U.S. military, it is a standard assumption that no country would dare use nuclear
weapons against a state that can retaliate, and certainly not against the United States. To the
contrary, the findings of this chapter suggest that the deterrence challenges ahead given current
U.S. foreign policy are far more difficult than is generally imagined, and the deterrence
mission in particular requires far more attention than it has recently been given.