Coercion and Military Strategy: Why Denial Works and Punishment Doesn't

Coercion and Military Strategy: Why
Denial Works and Punishment Doesn't

ROBERT A. PAPE, JR.
Throughout modern history, states have attempted to use military force

to persuade other states to do their bidding. States involved in serious
international crises or wars frequently engage in extensive and
passionate debate over the utility of military and non-military sanctions
as instruments of coercion.1 Nonetheless, states very often overestimate
the prospects for coercion. As a result, coercive attempts often fail
unexpectedly, even when assailants have superior capability and inflict
great punishment on the target state. Famous coercive failures include
German attempts to coerce Britain in 1917 and again in 1940, the French
occupation of the Rhineland in 1923-24, Italian efforts to coerce
Ethiopia in 1936, the American embargo against Japan in 1941, Allied
bombing of Germany in World War II, and American efforts against
North Vietnam from 1965-1968.
Traditional theories of coercion have contributed to this systematic
overoptimism because they focus on punishment. The basic idea is that
societies have limited thresholds of pain. Increasing punishment
eventually becomes intolerable, forcing the victim to knuckle under.
However, this argument, justified when based on theories of deterrence,
does not work for coercion. While punishment often deters, it rarely
compels, because states' willingness to countenance costs to achieve
gains is usually much lower than the pain they will endure to avoid
losses. Although nuclear weapons can overcome this, in conventional
conflicts even a highly capable assailant often cannot threaten or inflict
enough pain to coerce successfully.
Our understanding of military coercion has been nearly frozen for two
decades. Partly, the problem is conceptual: most work on coercion is
either policy advocacy2 or derivative of work on deterrence.3 Partly the
problem is substantive: the impact of domestic politics has been
understudied,4 while the interaction between coercers' and targets'
military strategies has been ignored. Partly, the problem is empirical:
I would like to thank Chris Achen, Richard Bctts, Michael Dcsch, James K. Feldman,
Charles Glaser, Joseph Grieco, Ted Hopf, Paul Huth, Chaim Kaufmann, Craig Kocrncr,
John Mcarshcimcr, James Morrow, Karl Mueller, Stephen Walt, and William Zimmerman.
The Journal of Strategic Studies, Vol.15, No.4 (December 1992), pp.423-475
PUBLISHED BY FRANK CASS, LONDON
424 JOURNAL OF STRATEGIC STUDIES
there is little rigorous empirical work on nuclear coercion, and none at

all on conventional cases.5
To remedy these weaknesses, this article constructs a more
comprehensive theory of military coercion, which incorporates all of
these factors. This new theory is then subjected to controlled tests based
on the record of coercive attempts using conventional and nuclear air
power.6 This analysis reveals that the standard wisdom is largely correct
for nuclear disputes, but almost wholly wrong for conventional conflicts.
In conventional coercion, the key is not punishment but strategy -
specifically, the interaction between the assailant's strategy and the
vulnerabilities of the target's military strategy. To coerce successfully,
the assailant must undermine the victim's confidence in his own military
strategy to control the territory at issue. If the victim can be persuaded
that it is militarily impossible for him to achieve or maintain his
objectives, then levels of costs which were bearable as long as there was
a chance of success become intolerable. The victim concedes to avoid
suffering further losses to no purpose. In contrast, the ability of the
assailant to inflict pain on the target's civilian population rarely makes
much difference. Conventional munitions have limited destructive
power, the modern nation-state is not a delicate mechanism that can
easily be brought to the point of collapse, and governments are often
willing to countenance considerable civilian punishment to achieve
important territorial aims. Consequently, coercion based on punishing
civilians rarely succeeds.
These limitations do not apply at the nuclear level, since nuclear
weapons have nearly unlimited capacity to do harm, which overwhelms
their military impact. Accordingly, successful nuclear coercion rests on
threats to civilians rather than against military vulnerabilities.
This article proceeds in five steps. The first defines military coercion.
The next reviews and critiques existing theories of military coercion. In
the third, I develop a new theory, based on interactions between the
assailant's strategies and the victim's vulnerabilities. The fourth
competitively tests the new and existing theories against the record of
coercion involving conventional and nuclear air power. Finally, I
present the policy implications that follow.
I. Defining Military Coercion

Coercion, like deterrence, is about changing the behaviour of a state by
manipulating costs and benefits in its decision calculus for a particular
dispute.7 Coercion can involve either forcing a state to take an action
that it otherwise would not, or to cease following a particular course of
COERCION AND MILITARY STRATEGY 425
action already embarked upon. Coercion differs from deterrence in that

deterrence is concerned with discouraging an opponent from attempting
to change the international status quo, while coercion attempts to
compel the target to change the status quo (or to accept the coercer's
unilateral action to change it).9
Military coercion means the use of military instruments to alter an
opponent's behaviour. While states often use economic, diplomatic, or
other forms of non-military pressure for coercion, military instruments
deserve special attention because they are the tools most often used
when very important interests are at stake, and because their use has the
greatest physical and normative consequences.9
In military coercion, the state issuing the threat (assailant) seeks to
persuade the target state (victim) to concede territory or other political
values that the assailant has not yet achieved on the battlefield. These
goals may include compelling the target to reduce political or territorial
aims, agree to a ceasefire, withdraw forces, or even surrender, goals that
could also be attained by decisively defeating the target. From the
coercer's viewpoint, the purpose of coercion is to attain concessions
without having to pay the full cost of military victory. From the victim's
perspective, accepting the assailant's demands may appear less costly
than fighting to a finish. For example, Japan's surrender in August 1945
saved the lives of many thousands on both sides.
Accordingly, I define coercive success as concessions by the target
although not completely defeated.10 Coercion is said to fail when (1) the
assailant stops his coercive military actions prior to concessions by the
victim, (2) the assailant's attacks continue but do not produce
compliance by the victim, or (3) the assailant imposes the coercive
demands only after complete defeat of the target." The last is crucial: if
a coercive attempt is made but the war ends only when one side is
decisively defeated, then coercion fails, even if the coercer wins the war.
Two main types of military threats can be employed to achieve these
goals: threats against civilian vulnerabilities ('punishment') and threats
against military vulnerabilities ('denial'). Punishment operates by
raising costs or risks to civilian populations. Denial functions by
militarily preventing the target from attaining his territorial goals, such
as by threatening to capture territory held by the victim or threatening
to defeat the victim's efforts to capture territory.12 Whether employing
punishment or denial threats, the advantage of coercion remains the
same: to achieve political gains at lower cost than would be required to
win decisive military victory.
Separating coercive attempts from assailants' general efforts to
achieve military victory is straightforward when punishment strategies
are employed, but can become ambiguous when denial strategies are
used. Since coercion through denial works by threatening the victim
with military failure, in many instances the same military operations
may serve the goals of coercion and military victory equally well.13
Moreover, assailants themselves often do not distinguish, but instead
pursue both goals simultaneously, hoping to attain their goals by
coercion if possible or by decisive victory if necessary.
This ambiguity would only be a problem if the pursuit of decisive
victory necessarily ruled out coercion, but nearly the opposite is the
case. Given any set of demands for which a coercer is prepared to use
force, he would virtually always rather obtain them by the less costly
method of coercion than by the more expensive alternative of winning
total military victory first and imposing the demands afterwards. Still,
this conjunction of coercion and pursuit of military victory does tell us
something important: if we find that coercive strategies based on denial
are more effective than those based on punishment, this means that the
most effective way to compel concessions without achieving decisive
victory is to demonstrate the capacity actually to achieve decisive
victory.
II. Existing Theories of Military Coercion

Three main theories can be identified in the literature on coercion. The
first, the balance of interests,14 argues that the state with the greater
intrinsic, strategic, or symbolic stake in a dispute's outcome will
prevail.15 The second theory points to the balance of capabilities:
striking the victim's forces shifts the military balance in favor of the
assailant, compelling the victim to modify his behaviour.16 The third
theory focuses on punishment, arguing that coercive leverage derives
from harming large masses of civilians; when the pain becomes
intolerable, the victim concedes.17
These theories are incomplete. While some authors have considered
more than one, none have carefully denned the relations among
variables, the strength of effects, and the conditions under which they
occur.
Further, four analytic flaws, common to nearly all work on the
subject, have impeded the advance of our understanding of how military
coercion operates:
A. DIFFICULTY OF COERCION
First, coercion is generally assumed to operate according to the same
principles as deterrence. However, while deterrence and coercion are

complementary activities, they pose distinct theoretical problems
because coercion is harder. Threats that deter may not coerce.19
Deterrence is made easier by the 'aggressor's handicap'.20 In the era
of nationalism, the defender's homeland territory is normally much
more valuable to him than to any assailant.21 Even if both value the
stakes equally, the attacker, who stands to make gains, is likely to be
more risk averse than the defender, who stands to suffer losses.22
Attackers also bear the burden of disturbing the status quo.23 As a
result, deterrers will typically pay higher costs to retain possessions than
attackers will to take them.
In situations of coercion all these conditions are reversed, so that
coercers often operate under the same handicap as the attacker in
deterrence. First, the victim is likely to value the stakes at issue more
highly than the assailant, particularly when the coercer demands the
surrender of homeland territory. Since targets of coercion are normally
willing to accept higher costs and risks than are coercers, potential
assailants are likely to be deterred from even attempting coercion unless
they possess superior military capabilities, which can protect them from
the victim's retaliation.24
Second, the coercer bears the responsibility of initiating hostilities if
threats fail.25 This lowers the credibility of coercive threats and
influences their implementation, encouraging assailants to favor threats
that can be fulfilled in progressive stages.26 Thus, most coercion occurs
in wartime.
This makes coercion harder, because in modern nation states the
determinants of decisions to initiate war are fundamentally different
than those to stop war.27 As the economic and social suffering caused by
war increases, states become less willing to abandon these sunk costs by
making concessions. The experience of war and government
propaganda needs can lead to a demonized view of the enemy and an
uncompromising 'us or them' attitude.28 Willingness even to consider
negotiation can come to be seen as treasonous.29 The need to mobilize
vast resources and to depend on large military organizations creates
enormous institutional momentum that cannot easily be turned off.30 As
a result, the longer a society is at war, the more its readiness to bear
costs increases, and the less attractive are most potential settlements,
even ones that would have been acceptable in peacetime.31
Coercion is therefore about hard cases. While the threat of a costly
war of attrition may deter aggression, successful coercion requires even
stronger sanctions.32 The prospect of protracted war influenced the
political calculations of all the major powers before World War II, but
had little impact on the decisions of any of them once involved in the
war. Despite suffering heavy costs, none of the powers on either side
even considered surrender until faced with the specter of total defeat.
B. STRATEGIES OF COERCION
Social scientists have long studied the effectiveness of both punishment
and denial strategies for deterrence. Oddly, most studies of military
coercion have focused on punishment of civilians, omitting the
possibility of threats to deny the target the military capacity to control
the contested territory.33 Even those arguing that the military balance is
the key to coercion employ punishment-based logic; they assume that
the goal is to create a monopoly of force which enables the coercer to
threaten unlimited punishment on the victim.
Aside from the simple analytic omission, restricting military coercion
to counter-civilian attacks invites both empirical and theoretical
problems. Empirically, ignoring denial strategies distorts our
understanding of important historical cases. Both military and civilian
targets have often been attacked for coercive purposes. During the
bombing of Germany in World War II, the British focused on civilians
while the Americans concentrated on military-related targets. Since
each air offensive sought to avoid the costs of a protracted land
campaign, and so achieve a political goal without paying the full cost of
direct assault, both are properly considered coercive. To study only the
British bombing effort as coercion would be historically inaccurate and
would miss an important part of the dynamics of the case.
Theoretically, studying only counter-civilian attacks risks ignoring
powerful explanatory causes of successful coercion. Coercion
sometimes succeeds even though civilians suffer only minor
punishment, as in the American bombing of North Vietnam in 1972.
Conversely, it sometimes fails despite very heavy punishment, as in
Ethiopia in 1936, the Sino-Japanese War from 1937 to 1945, and the
Allied bombing of Germany in World War II. Accordingly, a definition
of coercion limited to punishment of civilians will make it impossible to
explain many coercive outcomes.34
C. VULNERABILITY OF THE TARGET

Productive study of coercion also requires recognizing the impact of
differences in the vulnerability of the target state to the assailant's
coercive attacks. Different societies may be more or less vulnerable to
counter-civilian strategies due to factors such as their degree of
urbanization, use of burnable building materials, and susceptibility of

food production to attack.35
Similarly, target states with different military strategies vary in their
vulnerability to denial attacks. In particular, strategies that rely on
large-scale mechanized operations are quite vulnerable, because they
depend on massive logistic flows which make excellent targets for air
attack. At the opposite end of the spectrum, guerrilla strategies are
much less vulnerable to coercion, because they need little logistic
support.
The interaction of the assailant's coercive strategy and the victim's
specific vulnerabilities can be decisive, because it determines the
magnitude of the effects of the assailant's attacks, and therefore the
degree of pressure on the victim's political calculations. Theories which
do not account for differences in vulnerability cannot accurately predict
coercive outcomes.
D. NUCLEAR VS. CONVENTIONAL COERCION

Finally, the difference between conventional and nuclear coercion has
been glossed over. Authors tend to assume that the same dynamics
operate in both cases.36
Although in principle it makes no logical difference what type of
weapon is used for coercion, in practice the vast gap in destructive
power between nuclear and conventional weapons means that coercion
in these two circumstances operates differently. Nuclear weapons can
nearly always inflict more damage than any victim can withstand.37
Assuming the assailant's threat is credible, the resistance of even the
most determined opponents can be overwhelmed. Conventional
munitions can only inflict limited damage compared to the pain
thresholds of modern nation states, so punishment strategies are rarely
effective.
III. A Denial Theory of Military Coercion

A successful theory of coercion must recognise that no one type of
strategy or threat will be effective against all target states or in all
circumstances. Predicting when coercion will succeed depends on
understanding four issues: the basic logic of coercion; the sensitivities of
modern nation states; the coercive strategies available to affect these
sensitivities; and how these considerations vary in nuclear and
conventional coercion.
A. THE LOGIC OF COERCION

The problem in coercion is to persuade the target state that acceding to
the coercer's demands will be more beneficial than continued resistance.
Success or failure is therefore decided by the victim's decision calculus,
the principal components of which are benefits that would be lost by
concessions,38 the probability of attaining them by continued resistance,
the costs of resistance, and the probability of suffering these costs.39
How can a state's decision calculus be altered? Necessarily, all
coercive outcomes must be a function of varying one of the components
in the victim's decision calculus. However, not all of these components
are susceptible to manipulation.
Benefits are not usually manipulable by the coercer. The principal
issue in serious international disputes is control over territory, but these
interests tend to be fixed during the dispute because they emanate from
pressures that change slowly, if at all. Foremost among these are
nationalism and power balancing considerations. Nationalist
attachments to particular territories result from elements of the target
state's linguistic, cultural, and political history, which the coercer is
powerless to change.40 Interests based on security concerns derive from
the balance of power and threat in the international system.41 Since the
coercer, by definition, poses a severe threat to the target state, it is in no
position to manipulate these either.
The immutable nature of territorial interests has an important
implication: since the coercer cannot lower the value of the disputed
territory to the target state, coercion must require either raising costs or
reducing the probability that territorial benefits can be attained even if
the target continues to resist.
The second alternative for the coercer is to raise the costs to the target
of continued resistance, principally by threatening to inflict punishment
on civilians. States are always concerned about protecting the welfare of
civilian populations. When a state engages in an international dispute, it
does so because its territorial motives outweigh the expected costs and
risks to its citizens. By threatening to punish civilians, the assailant seeks
to raise the costs of continued resistance above the victim's value for the
territory at stake. The magnitude of the damage that the assailant can or
will inflict may be uncertain; the assailant may simply inflict the
maximum possible suffering as quickly as possible, or may gradually
increase the weight of attack in order to confront the victim with the risk
of greater damage in the future. If the threat is sufficiently great, the
victim will be compelled to abandon the territory to preserve its greater
interest in protecting its populace. The more important the territorial

interests at stake, the higher must the coercer raise civilian costs and
risks to compel concessions.
Third, in principle the coercer can raise the probability that the
threatened costs will actually be inflicted, but in practice extremely high
credibility is normally a minimum requirement. In most instances of
coercion, the territorial concessions demanded include strategic or even
homeland interests (i.e., surrender) of the victim, which requires the
assailant credibly to threaten very high costs to convince the victim to
abandon them.42 In fact, coercive threats often are highly credible
because they usually occur in war, when there is no reason to doubt the
assailant's willingness to inflict damage.43
The final option is to reduce the probability that continued resistance
will bring the target state the hoped-for benefits. States are willing to
pay costs in return for benefits only if they expect to actually gain the
benefits. If the target doubts its military capability to hold or take the
territory at issue, it may be unwilling to pay the costs of continued
resistance. The task for the coercer, therefore, is to thwart the victim's
military strategy, undermining the victim's confidence that its territorial
goals can be achieved. When coercive demands require the target to
concede important interests, it does not suffice to deny the ability to
succeed at low cost; the victim must be denied the ability to succeed at
all.
Once military success becomes infeasible, any further costs paid by
the victim become futile. The victim's most important incentive
becomes saving the costs of continued pointless resistance, especially
loss of life and continuing economic costs. These include the normal
costs of ongoing military operations, apart from any efforts by the
assailant to raise costs for coercive purposes. Levels of costs that were
insufficient to affect the victim's decision calculus as long as success was
feasible become sufficient to cause surrender.
Reducing the probability of achieving benefits can compel
abandonment only of those specific interests. Lowering the victim's
aggregate resources does not matter, except to the extent that this may
make it impossible to retain control over specific territories. If the
assailant's attacks leave the victim still capable of defending its
homeland although not a certain peripheral territory, only the latter will
be surrendered.44
There is one final hurdle the coercer must pass. Even if the victim's
military strategy is thwarted, coercion will succeed only when the costs
of surrender are lower than the costs of resistance. If surrender were
costless - beyond the value of the territorial interests conceded - states
would always surrender when the probability of victory disappeared,

simply to avoid the fruitless expenditure of resources. However,
surrender sometimes involves serious costs in addition to the territorial
benefits abandoned, such as enforced change of government,
destruction of social institutions, or the threat of genocide. When these
costs equal or exceed the costs of continued resistance, coercion will
fail, even if the military situation is hopeless.45
B. DECISION MAKING IN MODERN NATION STATES

Modern nation states are difficult to coerce, and become more resistant
the deeper they become embedded in a conflict. Their willingness to pay
costs to secure important interests is often high, particularly when
nationalist values are engaged, and often increases as sunk costs of a
dispute escalate. During disputes, they are often slow to re-evaluate
either costs or probabilities of achieving benefits. Finally, resistance is
especially strong when the ruling elite fears that concessions would
jeopardize their political survival.
1. High Resolve
States in serious international disputes tend to have high resolve, for
two reasons. First, the territorial stakes of these disputes are extremely
valuable, usually either strategic or homeland interests. In the era of
nationalism, attachments to homeland territories are extremely strong
as they result not only from security concerns but also from powerful
drives for national identity and cohesion.46 The prospect of the
homeland, or part of it, or even co-nationals elsewhere being ruled by
alien groups constitutes an intolerable injury to nationalist sentiments.47
Other territories are valued to the extent that they have strategic
importance for the defense of the homeland.48
Second, when involved in major disputes modern nation states are
relatively insensitive to costs, including civilian suffering. Nationalism
imbues individual citizens with personal attachment to national goals,
for which they are often willing to accept great sacrifice.49 The effect is
intensified by democratization. The greater the sense of participation in
the affairs of government, the more citizens see the state as the
embodiment of higher values, deepening their commitment to protect
and support the state and its goals."1" Additionally, advances in
communication technology have improved the ability of governments to
use propaganda and education policies further to enhance the legitimacy
of the state and its policies.51 Thus, modern states are often willing and
able to countenance a great deal of civilian punishment to achieve goals
articulated by the government.
In wartime, resolve usually increases. As the economic and social

suffering caused by war rise, states become less willing to abandon these
sunk costs by making concessions. The experience of war and
government propaganda needs can lead to a demonized view of the
enemy, producing an uncompromising 'us or them' attitude in which
anything less than victory comes to be seen as disastrous. Doves may be
constrained from participating in political debate; willingness to even
consider negotiation can come to be seen as treasonous. As a result, the
longer a society is at war, the more its readiness to bear costs increases,
while the acceptability of making concessions declines.
While modern nation states are often willing and able to bear high
costs, there are limits. In particular, no state can stand up under the
nearly unlimited punishment which nuclear weapons can deliver.
2. Poor Evaluation
States involved in serious disputes often under-estimate the costs and
over-estimate the feasibility of achieving their goals. First, governments
become less tolerant of criticism of their policies. Since decisions have
more momentous consequences, criticism is more serious, threatening
the public support needed to mobilize resources against the opponent,
and possibly also threatening the political survival of the regime. The
dangerousness of criticism increases as the costs of state policy rise and
as the liklihood of success fails. Accordingly, governments tolerate less
dissent inside their ranks, suppress leaks, and censor information that
might tend to undermine support for state policy, such as civilian costs
being suffered or risks of military defeat. Within the government,
information is restricted to a smaller circle of officials who believe in the
regime's policy.52
The result is that the public, as well as most of the government, is
presented with a rosier picture of the performance of state policy than is
actually the case. Dissenters operate at a disadvantage, because of the
cushion of legitimacy usually enjoyed by nationalist states. Most people
will defer to government statements, unless the gap between the official
line and reality becomes inescapably obvious.53
Second, militaries of states involved in disputes also suppress
criticism, both for the same reasons as government leaderships and
because evaluation threatens organizational stability.54 Evaluation
promotes change and innovation, which threatens jobs and the status of
incumbent leaders of the organization. Hence incumbents tend to
oppose, disrupt, or punish serious evaluation.55
Although they do not usually enjoy the same legitimacy as
governments, their special expertise gives militaries important
advantages in avoiding evaluation. Military expertise is largely confined

to military institutions; outside evaluators are usually less
knowledgeable about military operations than the military itself, putting
them in a poor position to compete with in-house evaluations, which
may be corrupted, ignored, or kept to a small few.55 Further, the need
for secrecy means that information about the likely success or failure of
military strategies is usually tightly controlled, making evaluation
difficult.57 In wartime, when any security breach can cost lives, secrecy
becomes tighter and evaluation correspondingly harder.
Third, suppression of criticism is made easier by the fact that
recognizing the possibility of failures is unpalatable for elites and publics
alike. When confronted with high-stakes decisions and ambiguous
information, people normally prefer to discount the risks rather than
confront the possibility of disaster. Evidence that current policies may
lead to unacceptable costs and risks tends to be ignored or discredited
unless its import is indisputably clear.58
The combined effect of discouraging criticism and poor evaluation is
that risks of disastrous defeat are often not recognized by either
governments or publics until these are made inescapably obvious by
events. This makes the coercer's task of demonstrating the
consequences of failure to concede all the more difficult. Even when
costs escalate beyond the apparent value of the issues at stake or when
the prospects for achieving goals through further resistance are bleak,
states still may not recognize the need to surrender until all ambiguity
has been eliminated.
3. The Domestic Politics of Concessions
Even when the costs of further resistance clearly outweigh any
attainable benefits, acceptance of the assailant's demands is likely only
when replacement of the target state's ruling elite can be accomplished
by evolutionary rather than by revolutionary change. Once deeply
involved in an international dispute, any concessions are likely to result
in the fall of the government that led the state to that point. Having
demanded sacrifices in pursuit of national goals, any concession to the
enemy's demands would allow domestic political rivals to charge the
ruling regime with incompetence, betrayal or both. Also, the enemy
may demand removal of the existing regime or even institute war crimes
trials.
The possibilities for domestic political change determine how strongly
the government will resist. Incentives to resist are weaker when regime
change means only replacement of individuals than when change means
the destruction of the ruling class and allied social institutions. First,
regime members have less reason to resist if their main social values are
not under threat. Second, and more important, when sacrificing the
current individual leaders is the only real domestic cost to concessions,
even their strongest social and political allies may withdraw their
support.59
However, when political change means revolution, resistance will be
much stronger. Regimes which expect that concessions will result in the
destruction of the ruling group and social institutions, either by an
occupying opponent of by a new hostile domestic regime, are likely to
resist to the end regardless of the consequences.
Whether the ruling elite can survive loss of power depends partly on
whether the political system has a tradition of peaceful change and
partly on whether the ruling elite has deep sources of power in society
aside from its control of the current government. For example, the
German Nazis were a purely political elite; once discredited and out of
power, they would have retained no sources of social, economic or
political influence.60 In contrast, the ex-daimyo and -samurai families
which ruled Japan in 1945 had staffed every modern Japanese
government and faced no real social or political opposition. Thus, the
Japanese government could contemplate surrender, while the Nazis
could not.
Because of the government's advantages in domestic political
competition, its resistance to concessions is likely to control policy even
when not in the interest of the society as a whole. The cushion of
legitimacy provided by nationalism permits the government to maintain
the allegiance of the population, especially for mobilization against an
external threat. In fact, support for the government often increases
during serious crises and wars due to a 'rally around and flag effect',
even among groups who tend to oppose government policies in
peacetime.61 In many instances, people continue to follow instructions
even though they no longer believe in government reports or policies.62
Unless the society is so badly divided by domestic cleavages before
the dispute that the government lacks authority, opponents criticizing
government policy labour under a legitimacy disadvantage. If necessary,
dissent can be forcibly repressed. As a result, it is unlikely that either
government policy or the regime itself can be changed by popular
pressure.
Forcible removal of the regime by popular opposition groups is even
more unlikely.62 To overturn the government, opposition groups must
do more than passively resist government orders; they must wrest
possession of important instruments of national control, such as
communication, transportation, and administrative centers. Yet,
opposition groups are at a great disadvantage in any contest of force,

since they usually lack organization, training and heavy weapons. Quite
small contingents of well-armed forces loyal to the regime can disperse,
decapitate, and deter most mass movements against the government or
collective civil disobedience. In major disputes, capabilities for
repression grow, both because governments are likely to mobilize
significant additional forces, and because repression of dissent is likely
to be supported by nationalist groups.
Even when governments do not, military elites have additional
incentives to oppose surrender. Aside from the reasons affecting
government regimes, the military is often concerned about disastrous
effects of concessions on its institution, either due to disarmament
enforced by the opponent or to domestic demilitarization initiatives.
Although militaries lack the legitimacy advantages of civilian
governments, they possess powerful capabilities to prevent concessions
they oppose. The vast resources that must be mobilized to meet serious
threats force governments to depend heavily on the military as an
organization, increasing its political influence.64 The military usually
also has a monopoly of force domestically, enabling it to disobey or even
remove the government if necessary.
4. Summary
Modern nation states are extremely resistant to coercion. Still, coercion
can sometimes succeed. To understand when, we must consider the
strategies available to the assailant.
C. ASSAILANTS' STRATEGIES
The conventional way to think about military coercion has been to
subsume all military efforts of the assailant into one large, abstract
variable called 'military costs' or 'military outcomes'. The strategies by
which these effects may be achieved are ignored as if the details of the
assailant's strategic and operational behaviour make no difference to
the effectiveness of coercion.
Contrary to this perspective, I contend that military strategy often
determines the success or failure of coercion in major international
disputes. No one strategy always works. No one strategy is always best.
The choice of strategies matters tremendously; in different
circumstances certain parts of victims' decision processes are more
vulnerable than others. In particular, it is crucial whether the assailant's
military strategy matches the particular vulnerabilities of the victim's
society or military strategy.
Assailants' coercive strategies can vary along two dimensions. First,

the assailant may focus on different parts of the victim's decision
calculus, either by raising the costs of continued resistance or by
lowering the probability that resistance will yield success. Second, the
assailant may rely either on actual damage to generate coercive pressure
or on the victim's anticipation of future damage. Three main strategies
can be assembled from combinations of these components: Punishment,
Risk and Denial.
Punishment strategies seek to inflict maximum damage on civilian
targets in order to raise costs to the point where they outweigh the value
of the stakes at issue. Denial strategies attack military-related targets in
order to undermine the victim's strategy for controlling the territory in
dispute. An alternative strategy, suggested by Thomas Schelling, relies
on the manipulation of risk of future damage rather than on inflicting
maximum actual damage. Under this strategy, the assailant deliberately
limits the pace of the campaign in order to give the victim time to
contemplate the costs of further resistance and also gradually escalates
the scale of attacks so that the victim must confront the prospect of
increasingly severe damage if he does not concede.65
Of these three strategies, denial is most effective. Given the limited
destructiveness of conventional munitions and the relative insensitivity
of modern nation states to civilian costs, punishment is likely to succeed
only when the victim's resolve is low. Risk strategies are likely to be
even less effective since at best they can threaten no more damage than
is actually inflicted under a punishment strategy. Only when nuclear
weapons allow threatening virtually unlimited damage is this approach
likely to succeed. Denial is more often effective, but depends on the
assailant's ability to undermine the victim's strategy for protecting or
capturing the particular territory at issue. If the assailant lacks the
capability to attack the specific vulnerabilities of the victim's particular
military strategy, denial will fail. Consequently, in some cases no
effective strategy may be available.
/. Punishment
Punishment campaigns seek to raise the societal costs of continued
resistance to levels that overwhelm the target state's territorial interests,
causing it to concede to the assailant's demands. The common feature of
all punishment campaigns is that they inflict suffering on civilians, either
directly by killing large numbers, or indirectly by destroying economic
infrastructure. For instance, cities may be bombed in order to kill or
injure the inhabitants, deprive them of power, water or other essential
services, or render them homeless. Alternatively, bombing or naval
blockade can cause shortages of key supplies, wreaking havoc on

civilian sectors of the economy.
Punishment does not usually offer significant coercive leverage, for
three reasons. First, states involved in coercive disputes are often willing
to accept high costs to achieve their objectives. Second, in conventional
conflicts, even massive attacks on civilians can inflict only limited
damage. Assailants can kill or starve significant numbers of civilians, but
never more than a small proportion of the victim's population.66 The
most devastating of all counter-civilian campaigns, the bombing of
Japan in World War II, killed 900,000 people out of a pre-war
population of 70 million. Counter-civilian attacks can lower morale,
causing some loss of economic production due to absenteeism and
deurbanization due to refugees fleeing cities, but morale problems
rarely have serious effects on production or cause civilians to put
effective pressure on the government to surrender.67 The absentee rate
from work in Japan during 1945 was approximately 8 per cent.68
Third, modern states have means of minimizing their vulnerability to
counter-civilian attacks by defense, evacuation of threatened areas, and
rapid adjustment to economic dislocations. To inflict significant damage
using conventional munitions requires large numbers of attacking
bombers. Population centers can be defended against such attacks by
fighters, radar, anti-aircraft artillery and missiles, and civil defenses.69
During World War II, bombing offensives in Europe, by both Allied
and German forces, were greatly hampered by anti-aircraft defenses,
while in the Pacific the vulnerability of the Japanese homeland was
greatly increased by the relative lack of such defenses.70
Evacuation of population centers threatened by bombing can further
reduce civilian vulnerability.71 Once city populations have been greatly
reduced further raids will inflict lighter losses, while the outlying
districts in which refugees are resettled make very poor targets for
bombing.72
Defense against sea blockade is also feasible. Consider the German
submarine blockade against Britain during World War I, which sought
to wreak havoc on the British population by cutting off access to sea-
faring commerce. Although the British were highly dependent on
imports for civilian goods such as food as well as raw materials for
military products and the submarine campaign did inflict severe losses, it
failed to cripple the British because they were able to deploy effective
countermeasures including convoy systems and submarine hunters.
Britain's enormous ship-building capacity also allowed her to produce
new ships fast enough to replace losses." Thus, the Germans were
unable to raise British civilian costs significantly, and the attempt at
coercion failed.
The relatively slow pace at which damage accumulates also allows

states to substitute for slowly accumulating shortages caused by
bombing or blockade, enabling them to ameliorate indirect costs to
civilians. The economies of modern nation states do not have 'break
points', thresholds of shortages below which sudden collapse occurs;
they deteriorate incrementally, by a process of successive substitution.
Modern techniques in science and administration can compensate for
the shortage of a primary good by conservation, more efficient
allocation, or discovery of alternatives. The history of wartime food
shortages in Britain illustrates this point. During the Napoleonic Wars,
World War I, and World War II, Britain's enemies attempted to
blockade her into submission by exploiting her excessive dependence on
food imports. These efforts failed in part because Britain's ability to
compensate for food shortages actually grew faster than her dependence
on food imports during each period.74
2. Risk
In contrast to punishment strategies, risk strategies slowly raise the
probability of civilian damage.75 The crucial element in this strategy is
timing. The assailant seeks to put at risk essentially the same targets as
in punishment strategies, but the key is to inflict civilian costs at a
gradually increasing rate rather than destroy the entire target set in one
fell swoop. This follows from the logic that, since coercive leverage
comes from the anticipation of future damage, military action must be
careful to spare a large part of the victim's civilian assets in order to
threaten further destruction.76 Also, the assailant must convince the
victim that the threatened targets will in fact be destroyed if concessions
are not made. To do this, operations are slowly escalated in intensity,
geographical extent, or both and the coercer must signal clearly that the
attacks are contingent on the victim's behavior, and will be stopped if he
complies with the assailant's demands.77 The assailant may sometimes
pause his operations temporarily in order to provide time for reflection
or negotiation, or to reward the victim for minor concessions.78
Risk strategies are usually less effective than punishment campaigns,
for three reasons. First, the argument on which the strategy is based,
that coercive leverage can come only from fear of future damage rather
than actual damage, does not address a real distinction in the strategies.
Given the limits of conventional weapons, punishment strategies never
actually 'kill the hostage'; damage accumulates slowly, so the assailant
can always inflict still more by continuing the assault.
Second, for any given level of assailant capabilities, damage
threatened under a risk-based strategy cannot exceed the actual damage
imposed by a punishment strategy. In fact, by further slowing the

already slow pace at which damage accumulates, Schelling strategies
may improve the ability of victims to adjust, and so may be unable to
threaten even as much damage as punishment strategies. For this
reason, risk-based strategies are virtually never used to threaten denial;
the slow and predictable pace provide opportunities to adjust tactics or
regenerate losses, which make it nearly impossible to reduce
significantly the victim's chances of military success.79
Third, the pattern of gradually increasing damage that is supposed to
increase the credibility of threatened future damage is more likely
actually to reduce credibility. This occurs because the assailant's
restraint is seen as resulting from political constraints rather than free
strategic choice. Instead of being convinced of the assailant's resolve to
inflict maximum damage if his demands are not met, the victim is more
likely to be persuaded that the assailant will never escalate far above
current restrained levels. The result is that the costs and risk perceived
by the victim are likely to be much lower than in a punishment strategy
with correspondingly less coercive effect. In fact, risk strategies are
hardly ever used except when assailants are subject to political
constraints which prevent them from waging full-blown punishment
campaigns.80The most important instance of Schelling's ideas in action
was the American bombing campaign against North Vietnam from 1965
to 1968, by which the United States sought to compel North Vietnam to
cease supporting the insurgency in the South. The Johnson
administration felt constrained to limit bombing of North Vietnam, both
because important sectors of the American public would not support
indiscriminate bombing and, even more importantly, because they
feared that massive bombing throughout North Vietnam would cause
China to enter the war.81 It sought to coerce Hanoi by increasing the
risks that existing limits would be crossed, leading to the loss of
industrial production.
This campaign failed principally because, given Hanoi's resolve, the
threat of limited conventional bombing of industrial targets did not pose
the risk of sufficiently brutal civilian hardship to overwhelm Hanoi's
territorial interests. North Vietnam viewed the South as part of its
homeland and its commitment to unify the country was based on the
powerful motive of national cohesion. Threatening to destroy North
Vietnam's tiny industrial base did not create sufficient risks to affect the
government's political calculus.82
While risk strategies are not effective in conventional disputes, they
can be successful in nuclear disputes. The destructive power of nuclear
weapons dramatically magnifies the risks that the assailant can threaten.
Further, threats are more credible because the threatened damage can
be inflicted extremely rapidly, which may allow the assailant to evade
political constraints that would otherwise prevent escalation. The speed
with which horrendous damage can be inflicted also deprives the victim
of any hope that administrative or social adjustments can ameliorate the
effects.
In nuclear disputes, Schelling strategies are likely to be preferred over
punishment strategies for four reasons. First, if a nuclear threat is
credible at all, the threat alone, or a few attacks, are likely to be
sufficient; actually executing large-scale nuclear attacks would be
coercive overkill. Second, intensive nuclear punishment runs the risk of
destroying all governmental authority in the target state, making it
impossible to terminate the conflict.83 Third, nuclear punishment would
be tantamount to genocide, which except in cases of the most extreme
ethnic or religious hatreds is likely to be repulsive to the assailant's
citizenry, as well as to third parties. Finally, nuclear punishment is likely
to cause collateral damage, including radioactive contamination,
beyond the boundaries of the target state, possibly reaching even the
assailant and his allies.84
Using a nuclear-based risk strategy, the US was able to compel China
to make the necessary concessions to end the Korean War. By 1953 the
broad outlines of an armistice in place had been agreed, but the Chinese
were inflexible on several minor issues.85 To break the deadlock, the US
threatened that it would resort to nuclear bombing if the Chinese did
not concede. This threat was lent credibility by a pattern of gradually
escalating conventional air attacks during spring 1953, capped by
warnings that nuclear attacks would follow.86 The Chinese agreed to the
US terms in June.
3. Denial
In contrast to counter-civilian approaches, denial strategies seek to ruin
the feasibility of the opponent's strategy to achieve his territorial
objectives, compelling concessions to avoid futile expenditure of further
resources. Unlike punishment strategies, the assailant makes no special
effort to cause suffering to the opponent's society, only to deny the
opponent hope of achieving his territorial objectives. Thus, denial
campaigns focus on the victim's military strategy.
Denial strategies offer more coercive leverage than punishment or
risk, but are subject to limitations. First, coercers can only obtain
concessions over the specific territory that has been denied to the victim.
As a result, if coercers seek to obtain more than they can persuade
victims they would lose, coercion can fail even though denial was partly
achieved.
Second, although the target usually cannot change basic strategy

because this is fixed by its capabilities or form of political organization,
it often has considerable ability to adjust and minimize the effects of the
assailant's attacks. First, the victim may have a considerable cushion
above requirements, allowing it to absorb military losses without
endangering national objectives.87 Second, some military strategies are
much more sensitive to losses of certain types than others; in certain
categories even heavy losses may have little effect on the outcome.88
Third, if losses can be regenerated quickly enough, they many not have
any impact on ultimate outcomes.89 Finally, the victim can often
compensate for losses by abandoning low priority missions and
reallocating remaining resources to the most important objectives.90 To
succeed, therefore, coercers must not only seek to thwart an opponent's
strategy but also any countermeasures the opponent can make.
Third, and most important, for coercion through denial to succeed,
the assailant must focus on the particular vulnerabilities of the victim's
specific strategy and exploit them to undermine the strategy. Simply
destroying targets of military value is not sufficient, because such attacks
may not necessarily undermine the victim's prospects of controlling the
territory at issue.91
All military strategies, however, do not share the same weaknesses.
There are two main types of strategies which modern nation states can
employ in conflicts with other states: mechanized (or 'conventional')
war and guerrilla (or 'unconventional') war. The objective in
mechanized war is destruction of enemy forces.92 This is accomplished
using massive, heavily armed forces that fight intense, large-scale battles
along relatively well-defined fronts. Success means routing opposing
forces on the battlefield by inflicting sufficient losses to destroy the
cohesion of units so they cannot execute combat functions.93
Guerrilla warfare, in contrast, aims to gain control over population,
usually beginning with villages located in remote areas, and to use these
as anchors to control still larger segments of the population and thus
undermine support for the government. Guerrillas fight in small units,
dispersed over large areas with no well-defined front line. Combat is
intermittent and at low intensity, avoiding major battles. Guerrillas seek
gradually to wear down the government's political authority and thus its
capability to field military forces, rather than destroy its forces in
battle.94
From a coercer's point of view, the most important difference
between these two strategies is that mechanized war is highly dependent
on logistics and communications (C3I) networks, while guerrilla war is
not. Mechanized war strategies depend on tremendous amounts of
ammunition, fuel, and other supplies, which flow from rear areas to
forward combat units. Supply consumption is especially high in intense
battles and the strategy breaks down quickly when needed supplies are
interrupted; logistic demands are inelastic. Both battlefield operations
and logistics are coordinated by centralized headquarters in the rear; if
communication is disrupted, front line forces become ineffective even if
not destroyed.95
In contrast, lightly armed guerrillas have minimal logistics and
communication requirements. Small amounts of weapons, ammunition,
and fuel are needed, but the bulk of the necessary material support,
such as food, is derived from the population under their control.
Additionally, guerrillas require much less coordination and central
control than conventional forces. Since individual units operate
autonomously, only rudimentary communications networks are
necessary and temporary disruptions rarely jeopardize operations.
The result is that an assailant seeking to coerce through denial must
employ different military strategies depending on the strategy of the
victim. If the victim's strategy depends on mechanized warfare,
disrupting logistic flows and communication networks is likely to be
effective. Alternatively, if the victim follows a guerrilla strategy, cutting
logistics and communications will have little effect; instead the assailant
must separate the guerrillas from the population that forms the basis of
their support.
Denial is more likely to succeed against conventional than guerrilla
strategies. The inelastic dependence of mechanized forces on logistics
and central control means that shortages can undermine the ability of
organized units to take or hold the disputed territory, while the
complexity of conventional war planning means that these can be
anticipated in advance of decisive defeat actually occurring. In contrast,
guerrilla wars depend on the willingness of overlapping small groups to
continue to resist central authorities, a type of collective action that is
both less sensitive to shortages and less predictable. Guerrillas should
be largely immune to coercion; assailants should expect to pay the full
costs of military success to extract political concessions.
In all cases the success of coercion through denial depends on
matching the assailant's military strategy to the vulnerabilities of the
victim's strategy. Two examples of failed denial campaigns illustrate
this. American coercion of North Vietnam in 1965-68 failed because US
bombing could not address any important vulnerability of Hanoi's
strategy. German coercion of Britain in 1940 failed because Britain's
defensive strategy was not vulnerable to German air attack.
In addition to Schelling-style manipulation of risk during 1965 to 1968
the US also sought to undermine North Vietnam's military strategy by

interdicting the flow of logistics from Hanoi to communist forces in the
South. Although considerable damage was inflicted on the enemy
logistic system, this had little effect on the feasibility of Hanoi's military
strategy, because the guerrilla campaign being fought then was largely
immune to conventional air attack. The guerrilla war required little in
the way of supplies and next to nothing at all from North Vietnam. Even
with well over 200,000 troops in the field, their requirements for food,
ammunition, medical supplies, and POL never exceeded roughly 380
tons a day, of which only some 34 tons came from the North. Given
these tiny supply requirements, aerial bombardment could not
significantly hinder North Vietnam's ability to support the insurgency,
and consequently coercion failed.96
After the fall of France, the Germans wished to coerce Britain into
accepting a position of neutrality by threatening an invasion across the
English Channel.97 Lacking sufficient land strength to defeat the
Germans once ashore, British strategy relied on the Royal Navy to
block any invasion. This in turn depended heavy on the ability of British
air power to defend the Fleet against German air attack. To coerce the
British, the Germans had to thwart this British strategy. Accordingly,
the Luftwaffe attempted to knock out the main airfields of the Royal Air
Force in southern England. Once air superiority was achieved, the
Royal Navy could no longer prevent a successful invasion, and Britain
would be forced to negotiate. Initially the campaign achieved
considerable success in wearing down British airfields, fighter direction
systems, and pilot reserves. However, later the Germans shifted their
attacks from these weak points in the British air defense system to
bombing the city of London, allowing the air defenses to recover. As a
result, Britain's strategy against invasion was not undermined and
coercion failed.98
D. PROPOSITIONS ABOUT SUCCESSFUL MILITARY COERCION

No one coercive strategy is likely to succeed under all circumstances. In
fact, some victims may be immune to any coercive pressure, if the
assailant lacks military capability, access to key targets, or simply adopts
the wrong strategy. Still, there are conditions under which one type of
strategy is more likely to succeed than another. Specifically, if a state is
involved in a conventional dispute, coercion is likely to be a function of
military vulnerability and largely unaffected by civilian vulnerability.
Conversely, if a state is involved in a nuclear dispute, coercion is likely
to be predicated on civilian rather than military vulnerabilities.
1. Conventional Coercion
Coercive success is a function of interactions between the assailant's
coercive strategy and the military strategy and domestic politics of the
target state. The Denial theory of coercion incorporates four major
propositions:
First, the degree of coercive pressure needed to force concessions by
modern states is often high, particularly when nationalist sentiments are
engaged.
Second, punishment strategies rarely succeed. Inflicting enough pain
to subdue the resistance of a determined adversary is normally beyond
the reach of conventional munitions. Thus, punishment strategies
should work only when the victim has interests at stake which do not
engage nationalist sentiments."
Third, risk strategies should always fail because they are diluted
versions of punishment.
Fourth, denial strategies should succeed when the assailant can
exploit the vulnerabilities of the victim's particular strategy to control
the specific territory in dispute, which is more likely against
conventional than guerrilla strategies.
The Denial Theory also contains several subsidiary propositions:
First, targets of coercion are more pliable when the assailant does not
demand surrender of the homeland or territories especially dear to
nationalist sensibilities.
Second, coercion is more likely to succeed when risks to the victim's
ruling elite are manageable, which is a function not only of the severity
of the assailant's demands but also of the character of the victim's
political system.
Third, since targets of coercive attacks are slow to recognize the
magnitudes of both increased civilian suffering and declining military
prospects, even small hopes can cause coercion to fail. Even when costs
and risks are appreciated, for domestic reasons the victim's political
system may not respond.
Fourth, coercive success almost always takes longer than the logic of
either punishment or denial alone would suggest. The domestic political
costs of concessions encourage delay until consequences are inescapably
obvious.100
Finally, coercion should fail when the costs of surrender outweigh the
costs of continued resistance. States should not be expected to surrender
to brutal terms.
2. Nuclear Coercion
Unlike conventional coercion, the accepted wisdom on nuclear coercion
is mostly right, though incomplete. The focus on civilian punishment

which has retarded our understanding of conventional coercion is
appropriate for nuclear cases. Current theories, however, focus
excessively on coercion under conditions of mutual assured destruction
(MAD) and so exaggerate the role the balance of interests plays in
successful nuclear coercion. A general understanding should
incorporate other nuclear situations.
The principal source of nuclear coercion is manipulation of civilian
vulnerability. Nuclear weapons have made it possible to launch
devastating and sudden attacks on the vast majority of a state's
population simultaneously. Meaningful defense is not possible, because
nuclear weapons are so destructive that even a small proportion of
penetrating weapons can devastate a great power's society.101 The
adjustment to gradually accumulating damage that is possible under
conventional attack cannot be done either, because the destruction can
be inflicted at at once.
Second, the nuclear revolution largely overturns the effects of
nationalism on the conduct of war. Nuclear war puts not only fractions
of the nation at risk, but, for all practical purposes, the nation itself
because the vast majority of a state's population could not avoid the
disastrous consequences of a large-scale nuclear war.102 Under such
conditions, the idea of individual sacrifice for the nation has no
meaning. Accordingly, it is hardly surprising that publics during the
nuclear age have not been enthused with the desire to promote national
goals in the context of nuclear war. In short, the nuclear revolution has
done much to cause states not to countenance nuclear punishment to
achieve territorial interests.
By contrast, military impact makes little difference in nuclear cases.
While nuclear weapons increase the assailant's ability to inflict military
as well as civilian damage, they do not do so quite as dramatically.101
More important, the horrific levels of societal destruction in nuclear war
are likely so to dominate decision-making calculations as to make the
prospects for success or failure of military campaigns largely irrelevant.
The existing literature on nuclear coercion has, as with conventional
coercion, been heavily influenced by ideas imported from work on
deterrence. On the nuclear side, however, this dependence has been
much less damaging. This is so because of the huge destructiveness of
nuclear weapons. When both sides can inflict virtually unlimited pain on
each other, and since states are normally willing to endure more risk of
pain to defend the status quo than to change it, nuclear coercion is hard
for the same reasons and in the same proportion that nuclear deterrence
is easy.
Under MAD, any potential nuclear assailant must consider the

possibility of nuclear retaliation, and will be unlikely even to attempt
coercion unless enormously important interests are at stake. As a result,
nuclear coercion attempts are very rare.104 When they do occur, the
outcome will likely be determined by the balance of interests between
the two sides;105 however, unless there is a large asymmetry facing the
coercer, he is not likely to fear nuclear escalation any less than the
target. This means that any concessions are likely to be on peripheral
issues, since a coercer who challenges major territorial interests is
unlikely to enjoy the necessary advantage.
While the standard wisdom is right for MAD, it misses important
features of two other types of cases: cases in which the assailant has a
large advantage in nuclear forces, although the target has some
retaliatory capability, and cases in which the target cannot retaliate.
Most authors have treated the first of these additional cases as
equivalent to MAD and have ignored the second altogether.106
When nuclear capabilities are quite unequal, the outcome is
determined primarily by relative civilian vulnerability, not the balance
of interests; the less vulnerable state is likely to prevail. The prospect of
one party suffering assured destruction while the other does not implies
that one must fear elimination while the other only very high costs.
Assuming this situation is common knowledge, the asymmetry between
damage and elimination creates tremendous coercive pressure on the
victim, because any increase in the risk of nuclear war will create a large
divergence in the expected outcome for each side. Thus, once the
assailant indicates a credible willingness to risk paying the costs of
starting a nuclear war, the prospect of unequal damage exerts coercive
leverage on the victim to accept the assailant's demands.107
When nuclear capabilities are completely one-sided, the outcome is
also determined by the vulnerability of the target's civilian population.
If the assailant's capability is relatively unlimited, coercive success is
virtually assured. Even if the assailants' nuclear resources are limited,
the prospect of damage far worse than the most intense conventional
assault will likely coerce all but the most resolute defenders.108
IV. Testing theories of Military Coercion

Powerful social science theories correctly predict non-obvious outcomes
or expose the faulty logic of intuition; a powerful theory will predict
outcome X when others expect outcome Y. Accordingly, the principal
standard against which to measure the power of a theory is whether it
accounts for the evidence better than the most likely alternative
explanations.
Testing theories of military coercion requires three steps. First, the

theories must be operationalized to provide falsifiable predictions
distinct from each other and easily observed in historical cases. Second,
appropriate historical evidence must be selected to test the hypotheses.
Finally, the predictions of the competing theories must be compared to
the evidence.
A. OPERATIONALIZING THE THEORIES

Three ideal type theories can be extracted from the literature: the
balance of geopolitical interests, balance of military forces, and relative
expected punishment. While certain authors mention two or more of
these, each assigns the main casual weight to a single variable,
sometimes with the caveat that the others be held constant.
Accordingly, it is reasonably faithful to the literature, as well as simpler
and clearer, to test each of these ideal-type theories one by one.
The first theory identifies success and failure with the balance of
geopolitical interests at stake. In coercive disputes, the state with the
greater interest will prevail.
How are geopolitical interests measured? Since control over territory
is the principal issue in serious international disputes, this study
constructs a simple, three-level typology of a state's territorial interests.
First, the highest interest of a state is in territory considered to be part of
its homeland. Second, states have powerful interests in territory that
affects their ability to defend their homelands. These are strategic
interests. Finally, states have lesser interest in territory they consider
symbolically important, control of which renders their other
commitments more credible or is important for ideological reasons.101*
The second theory argues that the military balance is the key to
successful coercion. Military superiority is the principal reason victims
succumb to coercive demands. The key assumption is that superiority
matters, because expectations of military success are driven by attrition
in which the superior side is able progressively to expand the margin of
superiority until the inferior side is defenseless.110
Operationalizing the military power theory must take account of this
basic assumption. Accordingly, the superiority that matters is
superiority in numbers or capability of weapons ready to engage the
opponent. In conventional disputes, this translates into the numerical
balance of theater forces.1" In nuclear disputes, this refers to the
numerical balance of deliverable nuclear warheads.
The third theory focuses on the vulnerability of the victim's civilian
population to attack. Advocates of this view claim that coercive
leverage derives from punishing large numbers of civilians. Drive up the

costs and risks to the civilian population, the argument goes, and the
victim is likely to succumb to your demands, either out of fear of yet
greater costs, social unrest, or economic collapse.
How can this theory be operationalized? Since the state most likely to
suffer loses, the key indicator is total civilian and military losses. In
peacetime this is coded according to anticipated levels of damage, while
in wartime actual losses are used,
Finally, the Denial Theory must be operationalized. For purposes of
comparing my theory against the relatively simple theories above, I test
only the main propositions. This actually reduces the power of the
theory somewhat, because any variance accounted for by the subsidiary
propositions actually counts against the theory.112
The two independent variables are civilian vulnerability and military
vulnerability. What matters about civilian vulnerability is the
expectation of suffering, or actual suffering, due to the assailant's
attacks. Civilian vulnerability is coded as 'low', meaning that, although
there is some risk to individuals, no major part of the population must
make adjustments or compromises in their daily lives to avoid the
threat; 'medium', meaning that the risks to individuals has risen to the
point that major parts of the population must make compromises or
adjustments in their daily lives to lower the threat; 'high', meaning that
the numbers affected and the consequences are so severe that major
parts of the population are uncertain about whether they will survive;
and 'very high', meaning that the major parts of the population are
certain not to survive because avoiding the enemy's attacks is
impossible. The Denial theory expects that coercion in conventional
disputes should fail even if civilian vulnerability is high or very high.
What matters about military vulnerability is the expectation of being
able to take or hold the disputed territory with military force. Military
vulnerability is coded as 'low', meaning, while there is some risk that the
current position could not be maintained, it has not reached the point of
requiring additional military measures to maintain confidence in
success; 'medium', meaning that control over the disputed territory is
definitely in jeopardy but the threat can be reduced by added military
measures; 'high', meaning that the successful defense or conquest of the
territory cannot be assured even with added military measures, but it
may be possible to inflict enough attrition to reduce the opponent's
commitment to control the territory; and 'very high', meaning that the
likelihood of loss of control over the territory approached certainty
because both defeat and heavy attrition of enemy forces is impossible.
The Denial Theory expects coercion to succeed if military vulnerability
is high or very high.
B. Evidence
The empirical study of military coercion requires investigating historical
cases to determine the conditions under which alternative coercive
strategies succeed or fail. This study investigates the universe of
coercive attempts using strategic air power, 28 cases in all, most of
which occur in war.
1. Why Strategic Air Power?
The goal in case selection is to find evidence that can isolate the effects
of various coercive threats and differences in the victim's vulnerability
on the victim's political calculations. Strategic air power cases are
ideally suited to this task for three reasons.111
First, air power cases can discriminate different explanations better
than either land or sea power cases, because they include variation on
both independent and dependent variables, while the others do not.
Given the reduced vulnerability of modern industrial states to naval
blockades and commerce raiding, coercion with sea power is so
ineffective that these cases inevitably consist mainly of failures. Thus,
any sample would have insufficient variation on the dependent variable
to distinguish among a variety of causes. Land power cases suffer a
different deficiency. Since modern ground forces are rarely able to
attack economic or military targets without first penetrating enemy lines
and achieving decisive victory, coercion with land power must be based
mainly on denial.114 Given the lack of punitive uses of land power, this
sample cannot discern the relative effectiveness of coercive threats.
Second, identifying different coercive strategies is easier in air power
cases. This task is sometimes problematic because assailants often do
not explicitly identify the coercive strategy they are pursuing, and
because some land and sea operations have both denial and punitive
effects. By contrast, in air power cases coercive strategies can nearly
always be distinguished based on target sets and bombing techniques.
Punishment strategies most often employ area bombing of urban
centers, while denial strategies use precision attacks on industrial,
transportation, and military targets.
Finally, lessons learned from the study of coercive air power have
special policy relevance, because air power is so often the coercive
instrument of choice. Air power offers assailants three advantages over
either land or sea power. First, air power can reach deep into the
enemy's homeland from the outset of a conflict, which other instruments
cannot. Second, when appropriate bases are available, strategic
bombing operations can often be initiated more quickly than land or sea
TABLE 1
CASES OF COERCIVE AIR POWER
Year Assailant Victim Issue
1. 1917 Germany Britain British withdrawal from WWI

2. 1920s Britain Turkey Iraq
3. 1936 Italy Ethiopia Ethiopian sovereignty
4. 1936-39 Spanish Nat'l Spanish Rep. Spanish sovereignty
5. 1937-45 Japan China North/Central China
6. 1938 Germany Britain/France Czechoslovakia
7. 1939 Germany Britain/France Poland
8. 1939-40 USSR Finland Finish sovereignty
9. 1940 Germany Holland Dutch sovereignty
10. 1940 Germany Britain British withdrawal from WWII
11. 1943 US/UK Italy Italian sovereignty
12. 1944-45 Germany Britain Reduction of unconditional
demands
13. 1942-45 US/UK Germany German sovereignty
14. 1944-45 United States Japan Japanese sovereignty
15. 1950-53 United States China North Korea
16. 1950s France Algerian Rebels Algerian sovereignty
17. 1956 Britain/France Egypt withdrawal from Suez Canal
18. 1956 USSR Britain/France withdrawal from Middle East war
19. 1962 United States USSR missiles in Cuba
20. 1964-68 United States N. Vietnam South Vietnam
21. 1972 United States N. Vietnam South Vietnam
22. 1969-70 Israel Egypt Sinai
23. 1970 Israel Syria Jordan
24. 1970s Israel Lebanon PLO Terrorism
25. 1979-86 USSR Pakistan Reduction of support for Afghan
26. 1986 United States Libya Terrorism
27. 1980-88 Iraq Iran border territories
28. 1991 United States Iraq Kuwait
operations, because air forces can be deployed to distant theaters more

rapidly. Third, air power promises to achieve its effects at lower cost in
lives to the assailant. Thus, it is not surprising that policy experts, from
the Berlin Crisis to Vietnam to Libya to Iraq, have so often
recommended air power as a solution to international disputes.
Case Identification
Cases were identified according to three criteria. First, cases are
restricted to those in which the victim was asked to give up important
interests. Typical examples are efforts to compel reduction of political
aims, agreement to a ceasefire, withdrawal of forces, or even surrender,
by states that retain the capacity for continued military operations.
Cases in which the demands on the victim represent minor or
insignificant loss, such as extortion of tribute, are excluded. If the victim
incurs little or no cost in complying with the assailant's demands, he

might comply even if subjected to little or no pressure. Such cases,
therefore, have little value for explaining cases where compliance is
expensive for the victim.
Second, this study excludes situations involving a monopoly of
military force by one side, such as domestic police actions or postwar
concessions. Some might refer to 'bargaining' or coercion when the
assailant has a monopoly or near monopoly of power, but such cases are
not useful for distinguishing between competing theories in important
cases.115 If the assailant can inflict unlimited harm on the victim at little
or no cost to itself, then the causes of successful coercion are trivial and
tell us little about the more important case in which the assailant does
not have a near monopoly on military force.116
Third, cases are limited to those in which coercive threats can be
clearly identified. These cases include explicit statements of the form
'Do X or I'll do Y' or implicit threats where there is evidence they were
understood by the victim. Cases which involve highly ambiguous threats
are excluded, as are those in which it is not certain that a coercive
attempt was made."7 If we cannot establish that the substance of the
threat and the conditions of its execution were communicated to the
victim, we cannot determine whether a coercive failure was due to
insufficient threats or to communication gaps.
C. TESTING THE THEORIES

Quantitative and qualitative procedures can be used to compare the
relative validity of the theories of military coercion.
/. Quantitative Test
The key question in assessing the validity of correlations between
independent and dependent variables is how they compare to chance.
There are two possible outcomes, success and failure. Accordingly, we
can easily determine whether the success rate for each theory is the
same or higher than simply flipping a coin by calculating the probability
of observing at least the number of correct calls if the truth were
random.118
TABLE 2
PREDICTIONS OF EXISTING THEORIES'"
Case (Victim) Interests Forces Punishment Outcome
1917/UK failure failure failure failure

1920s/Turks failure* failure* success success
1936/Ethiopia failure success* success* failure
1936-39/Sp.Rep. failure failure success* failure
1937-45/China failure failure success failure
1938/UK success failure* success success
1938/FR failure* failure* success success
1939AJK failure failure success* failure
1939/FR failure failure success* failure
1939-40/FIN failure* success success* success
1940/HOL failure* success success success
1940/UK failure success* success* failure
1943/Italy failure* failure* success success
1944-tt/UK success* failure failure failure
1942-45/GE failure success* success* failure
1944-45/Japan failure* failure* success success
1950-52/PRC failure failure success* failure
1953/PRC failure* failure* success success
1950s/Alg. Rebs. failure success* success* failure
1956/Egypt failure success* success* failure
1956/UK&FR failure* success success success
1962/USSR failure* success success success
1965-68/NVN failure success* success* failure
1972/NVN failure* success success success
1969-70-/Egypt failure success success failure
1970/Syria failure* success* success success
1970s/Lebanon failure success* success* failure
1979-86/PAK failure success* success* failure
1986/Libya failure success* success* failure
1980-87/Iran failure success* failure failure
1988/Iran success success failure* success
1991/Iraq failure failure* success success
Correct 19/32 15/32 18/32 success=14

*Incorrect Predictions failure=18
TABLE 3
PREDICTIONS OF THE STRATEGIC THEORY1211
Case Civilian V«l. Military Vol. Prediction Outcome
1917/UK medium medium failure failure

1920s/Turks nil high success success
1936/Ethiopia high medium failure failure
1936-39/Loy low medium failure failure
1937-45/China high medium failure failure
1938/UK&FR very high very high success success
1939/UK&FR high very high success failure'
1939/Fin medium high success success
1940/Hol high very high success success
Aug.1940 low medium failure medium
Sept.l940/UK high medium failure failure
1943/Italy medium very high success success
1944-4S/UK medium low failure failure
1944-45/GE high medium failure failure
1945/GE very high very high success failure*
1944-45/Japan high medium failure failure
1944-45/Japan very high very high success success
1950-52/PRC high medium failure failure
1953/PRC very high(nuc) medium success success
1954/Alg high low failure failure
1956/Egypt low medium failure failure
1956/UK&FR very high (nuc) low success success
1962/USSR very high (nuc) very high success success
1965-68/NVN high low failure failure
1972/NVN medium high success success
1969-70/Egypt medium low failure failure
1970/Syria nil high success success
1970s/Lebanon high low failure failure
1979-86/PAK low low failure failure
1986/Libya low low failure failure
1980-88/Iran high medium failure failure
Julyl988/Iran high very high success success
1991/Iraq low high success success
Correct = 31/33
The Strategic Theory strongly outperforms all the existing theories, as

Table 4 demonstrates:
TABLE 4
SUCCESS OF THEORIES RELATIVE TO CHANCE
Theory Success Rate Probability of Random Result
Interests 59% .19

Punishment 47% .70
Forces 56% .30
Denial 94% <.00001%
The Punishment and Forces theories have no predictive value; any

random prediction could be expected to do practically as well or better.
The Interest Theory does the best of the existing theories, but it still falls
short of the standard benchmark significance level of .05 needed to
persuade most social scientists that the theory's correct predictions are
not random. In contrast, the Denial Theory predicts 31 of 33
successfully, which is a result that could be achieved by chance only
once in a hundred thousand times. This is an extremely robust result; for
the Denial Theory to fail the .05 standard, it would have to be wrong in
10 more cases. Such a robust result has the advantage that, while the
coding rules are reliable, quibbles over one or a few cases cannot affect
the significance of the result.
2. Qualitative Tests
In addition to testing whether outcomes correspond to predictions, it is
important to test whether the causal dynamics in specific cases match
those expected by one theory rather than another. The most likely
alternative to the Denial Theory is the balance of geopolitical interests
theory, since, as shown above, it does the best in comparison to chance.
Accordingly, five cases in which they make opposite predictions are
examined, four mispredicted by the Interests Theory but correctly
predicted by the Denial Theory (Japan 1945; Korea 1953; Vietnam 1972;
and Iraq 1991) and the case (Germany 1945) that is the hardest for the
Denial Theory and the easiest for the Interests Theory. This test shows
that the Denial Theory correctly identifies causal mechanisms that the
Interests Theory would miss even in its hardest case.
Japan, 1945
The surrender of Japan is a classic case of successful military coercion.
On 15 August 1945 Japan unconditionally surrendered to the United
States, while still possessing an undefeated and ammunition-stocked
two million man army on its home territory. The conventional wisdom is
that America's use of two atomic weapons explains Japan's decision.
The problem, of course, is that this runs directly counter to the accepted
view that nuclear coercion is determined by the balance of interests.
Since the Japanese certainly valued control over their homeland more
than the Americans, coercion should have failed.
The success is explained by the Denial Theory. Contrary to the
standard wisdom, civilian fears played little role in Japan's decision to
surrender; Japan's surrender was primarily a function of her military
vulnerability.
For six months prior to dropping the atomic bombs, Japanese cities
were subjected to the most intense incendiary air offensive in history.
More civilians were killed in the first fire bombing of Tokyo on 9 March
1945 (84,000) than in the atomic bombing of Hiroshima (70,000).121
More important, by August, 62 of Japan's 66 largest cities had been
burned to the ground, inflicting some 2,200,000 civilian casualties,
including 900,000 fatalities.122 Not only was the atomic bomb
unnecessary to inflict mass destruction on population centers, but by the
time surrender occurred, the vulnerable population had largely been
destroyed or made homeless; the 'hostage' was nearly dead.123 Since
Japan had long lost its ability to defend itself against massive counter-
population attacks, threats to civilians should have brought Japan to
surrender in March, not August. There is some evidence that the atomic
bomb influenced civilian leaders, but it did not sway the military, whose
consent was necessary for surrender.
Despite the fact that the effects of the American blockade and
bombardment had rendered its forces a hollow shell, Japanese Army
leaders confidently maintained throughout 1945 that Japan could inflict
large numbers of American casualties during an invasion of her home
islands, which would lead the United States to settle on more favorable
terms.124 By dominating domestic politics, the Army could successfully
avoid domestic criticism of its strategy. First, the Army could prevent
other elites from learning the full extent of its weakness to execute the
homeland defence strategy. Thus, Army leaders could purvey
confidence in their strategy without having their authority challenged.
Second, the Army's dominance of domestic politics caused civilian
leaders, including the Emperor, to fear that resistance would result in a
military coup.125
Only when Japanese leaders, including army leaders, came to doubt
whether this strategy would succeed did they agree to surrender before
invasion rather than accept the costs of continuing the war. The most
important factor accounting for the timing of the surrender was the
Soviet attack against Manchuria, because of its effects in further

demonstrating the vulnerability of Japan's homeland to invasion. When
Prime Minister Kantaro Suzuki received General Sumihisa Ikeda's
report on 9 August that 'the Kwantung Army is hopeless. . . . Within
two weeks Hsinking (Chang-Chun) will be occupied', Suzuki replied, 'Is
the Kwantung Army that weak? Then the game is up'.126 More
importantly, it affected the hard-line military members of the cabinet,
the only one of whom to be interrogated after the war, Navy Chief of
Staff Admiral Soemu Toyoda said: 'I believe the Russian participation
in the war against Japan rather than the atom bombs did more to hasten
the surrender.'127 Similarly, Army Vice-Chief of Staff General
Torashiro Kawabe reported: 'Since Tokyo was not directly affected by
the bombing, the full force of the shock was not felt... In comparison,
the Soviet entry into the war was a great shock when it actually came,'128
By slicing through Japan's best prepared forces, the Soviet ground
advance demonstrated the futility of the homeland defense strategy in a
way that the army could not deny. When the Emperor called for
surrender, he is reported to have said:
I was told by those advocating a continuation of hostilities that by
June, new divisions would be placed in fortified positions at
Kujukurihama so that they would be ready for the invader when
he sought to land [Kujukuri is on the Boso Peninsula near Tokyo].
It is now August and the fortifications still have not been
completed. Even the equipment for the divisions which are to fight
are insufficient and reportedly will not be adequate until after the
middle of September. Furthermore, the promised increase in the
production of aircraft has not progressed in accordance with
expectations. There are those who say that the key to national
survival lies in a decisive battle in the homeland. The experience of
the past, however, shows that there has always been a discrepancy
between plans and performance. I do not believe that the
discrepancy in the case of Kujukuri-hama can be rectified. Since
this is the shape of things, how can we repel the invaders.129
Korea, 1953
The decision by the Chinese to sign the armistice ending the Korean
War also cannot be explained by the balance of interests. Both the
United States and the PRC had strategic interests at stake in controlling
the Korean Peninsula, and it is implausible that the United States
enjoyed the advantage. The key to success in this case lay in how the
United States combined atomic threats and diplomatic signals to coerce
China. Militarily, a pattern of escalation in the use of force was

established from February to May 1953. In February American B-29s
bombed communist front-line forces, followed by an attack against the
Suiho power installation dams in North Korea. On 13 May, 59 fighter-
bombers attacked the Toksan irrigation dam and on 15 May two more
dams. North Korea reported that resulting floods inundated 70 villages
and photographs showed a 'scene of total devastation'.130
Following the irrigation dam campaign, atomic threats were signaled
to China through diplomatic channels. Several signals were sent to
ensure that the communists realized that the UN Command would
resort to nuclear warfare. The UN negotiator, General William K.
Harrison, was authorized in his instructions of 25 May to indicate that
this was the final word. Assuming the message would reach Peking,
Secretary of State John Foster Dulles told Prime Minister Nehru of
India that 'if armistice negotiations collapse, the United States would
probably make a stronger rather than a lesser military exertion and that
this might well extend the area of conflict.' While it is not clear that
Nehru relayed this message, the Indian foreign minister told a British
diplomat that if the negotiating terms were rejected, the US would
'break off the negotiations and have recourse to dramatic action'.
Finally Washington also sent messages through Moscow. The US
ambassador, Charles Bohlen, was instructed to emphasize to Soviet
Foreign Minister V. M. Molotov the 'extreme importance and
seriousness of the latest UNC proposals, pointing out lengths to which
UNC has gone to bridge existing gap and making it clear these represent
the limit to which we can go.'131
Communist negotiators accepted US terms for the armistice on 4
June. Although no archival material for Chinese decisions is available,
there is evidence that nuclear coercion determined the outcome and
against the most compelling alternative explanation that South Korea's
release of PoWs made the difference. From mid-May onwards, the
Chinese seemed to get the correct signals from the US. In 19 May, for
instance, the New China News Agency reported rumors from Tokyo that
the US was preparing a 'new ultimatum'. Moreover, Chinese, Soviet
and North Korean leaders conferred with each other between 28 May,
the date of Bohlen's meeting with Molotov in Moscow, and 4 June the
date of China's acceptance of US terms.132 Finally, the argument that
the South Korean release of PoWs led to the armistice is undermined by
one important fact: the release of prisoners, which was a complete
surprise to the US and China, occurred on 16 June, two weeks after
China initially accepted the agreement.
Although it is difficult to say whether nuclear coercion worked
because it threatened civilian or military vulnerability, there is some

evidence that the Chinese were not worried about the battlefield effects
of the atomic bomb. After the war, the Chinese reported an assessment
made during the war that 'US military personnel did not believe that the
use of atomic weapons on the Korean battlefield would be effective.'133
Vietnam, 1972134
In March 1972 Hanoi launched a conventional offensive aimed at
achieving a series of limited battlefield victories that would precipitate
the downfall of the Saigon regime. The United States responded with
massive conventional bombing ('Linebacker' I and II) that induced
Hanoi to sign the 27 January 1973 Paris agreements, which ended
America's role in the war. The bombing was a coercive success, forcing
the North to cease its ground offensive and accept a ceasefire, even
though it had the capacity to continue military operations.
Civilian fears played no role in the success of 'Linebacker' I or II.
First, the campaigns pinpointed military targets and avoided civilians.
President Richard M. Nixon was keenly sensitive to the domestic
political criticisms to which he would be subjected if numerous civilians
were killed. Accordingly, the target set in both campaigns comprised
only those that were military-related and attacks were conducted with
discriminate tactics and weapons.
Second, while some observers credit the coercive success to the
heightened ferocity of the bombing during 'Linebacker' II, in fact both
'Linebacker' campaigns caused relatively little loss of life. Air raids
against the North in 1972 caused an estimated 13,000 deaths, of which
'Linebacker' II killed 1,623 by North Vietnam's count. The significance
of these small numbers becomes manifest when compared to the
estimated 125,000 deaths that communist forces sustained in South
Vietnam during 1972 and the reported 851,000 total deaths from 1964 to
1972.
Third, the North Vietnamese agricultural economy was not harmed.
'Linebacker' I did not hit the irrigation system, despite widespread
reports that the United States had begun a concerted effort to raze the
dikes. A few bombs fell close to dams located near military targets, but
not a single major dike was breached and no flooding occurred. Finally,
the bombing did not unravel the North's social and political fabric.
Civilians were frightened, but this did not lead to general panic, civil
disobedience or grassroots opposition to the Hanoi government.
Military vulnerability explains the success of American coercion.
'Linebacker' I involved air strikes against a broad set of military targets
behind the battlefield in both North and South Vietnam. These included
logistical centers, transportation arteries, fuel dumps, warehouses,

marshalling yards, rolling stock, vehicles, power plants, a POL pipeline
running from China, and many SAM and anti-aircraft artillery sites.
This air interdiction reduced the flow of resources to North Vietnamese
Army units, which diminished the North's combat power by creating
firepower shortages and disrupting mobility. It also persuaded Hanoi to
accept US terms for the peace accords. Yet the accords were not signed
immediately, because South Vietnam stalled negotiations and this
encouraged Hanoi to back away from the agreement. To force the
North back to its earlier position, Nixon subsequently ordered yet
another air campaign on 18 December 1972. Known as 'LinebackerTI, it
lasted eleven days and led to the achievement of a ceasefire agreement.
Both 'Linebacker' campaigns ignored civilian vulnerabilities and
concentrated instead on damaging Hanoi's military capabilities. They
succeeded because Hanoi had switched from a guerrilla to a
conventional war strategy, a strategy vulnerable to an air offensive.
Iraq, 1990-91
The outcome of the 1990-91 conflict with Iraq is complicated by the facts
that the United States had goals other than compelling Iraq to withdraw
from Kuwait135 and the event is too recent for access to primary sources.
Nonetheless a preliminary analysis of the problem of coercing Iraq from
Kuwait can be made.
President Saddam Hussein's strategy was to draw the coalition into a
premature ground offensive in the hope that heavy casualties would lead
Western publics to demand an early ceasefire, leaving Iraq in control of
the most economically valuable portions of Kuwait.136 In other words,
Iraq's strategy to hold Kuwait was predicted on inflicting heavy
casualties on coalition forces early in their offensive, not on decisively
defeating the attack. Consequently, Saddam deployed his least trained
and equipped forces in fortified positions on the Kuwaiti-Saudi border
and his better equipped Republican Guard divisions behind them to
establish a second line of defense near the main oilfields.
Iraq's military strategy required American coercive efforts to cross a
very high threshold: Saddam must not only be convinced that he would
ultimately lose a ground war, but, most significantly, that that the
attacker would win cheaply. As a result, coercive success required
persuading Saddam that he would be unable to inflict significant
casualties on attacking forces.
Although the coalition air offensive began on 16 January 1991, it was
not until mid-February that attacks on Iraqi ground forces started to
tell. The early part of the air campaign was dedicated to strategic
targets, which destroyed targets that inflicted significant indirect

punishment on Iraqi civilians and reduced Iraq's capability to command
and control its forces. Although this first phase of the air campaign had
military effects, these in themselves were not decisive because Iraq's
strategy called for fighting a series of set-piece battles, not a war of
maneuver. The critical weakness of this strategy was not primarily
sophisticated command and control networks, but heavy forces in being.
Beginning in February, 'preparation of the battlefield' became not only
the dominating air mission, but also more effective as the more accurate
F-llls replaced the error-prone F-16s in striking Iraqi armor. By mid-
February, over 100 tanks a day were being destroyed and coalition
briefers described the Iraqi position as 'precarious'.
With its military strategy in tatters, Iraq's political calculations began
to change. Saddam first blinked on 15 February, when Baghdad
suddenly announced that the Revolutionary Command Council was
preparing to accept a withdrawal from Kuwait, provided a series of
conditions were met. From this point on, Iraq began dropping
conditions and moving toward a timetable for withdrawal. By 22
February Iraq had agreed to leave Kuwait two days after a ceasefire
without a parallel withdrawal by the United States from Saudi Arabia or
resolution of Palestinian issues. By then, however, the Bush
administration was not content with only compelling Iraq from Kuwait;
it also wanted to achieve other goals including elimination of Iraqi
offensive capability and achieving war reparations and perhaps
Saddam's ouster.147 Since Saddam clearly agreed to withdraw from
Kuwait and since the remaining unsatisfied conditions had nothing to do
with who would control Kuwait, I code this as a case of coercive success.
Since it resulted from a collapse of Saddam's military strategy, I also
code it as a success for the Denial Theory.
Germany, 1945
The end of World War II in Europe is the more important of the two
cases not predicted correctly by the simple version of the Denial Theory
which we tested above. The Allies successfully undermined Germany's
military strategy but, nonetheless, Germany failed to surrender.
Moreover, the balance of interests theory correctly predicts the
outcome of this case. Since Germany valued control over its homeland
more than the Allies, coercion should have failed. Thus, this case is one
of the hardest for my theory and is therefore one of the most important
to examine in detail in order to gain a complete appreciation of the
relative strengths and weaknesses of the Denial Theory as well as those
of its competitors.
Detailed examination of the case illustrates why the simple version of

the Denial Theory, which focuses exclusively on military vulnerabilities,
is not sufficient. The complete version of the theory, which includes
propositions concerning domestic politics and the costs of surrender,
fully explains the outcome of the case. This detailed investigation also
shows that, while the balance of interests theory makes a correct
prediction for the outcome of this case, it does so for the wrong reasons.
The balance of interests explanation assumes that German leaders
were, at the end of the war, primarily motivated by their interest in
retaining German control over the German homeland even if resistance
was more costly to the German nation than surrender. The problem is
that German leaders did not fail to surrender because of their interest in
controlling homeland territory; by February 1945 they understood that
they could not. Rather German leaders did not surrender because they
were convinced that the costs of surrender were greater than the costs of
resistance, and might well have behaved differently had the reverse
been true.
The military vulnerability proposition of the Denial Theory taken in
isolation would predict German surrender. By mid-1944 Germany could
no longer expect to defeat its enemies, because once the economically
superior Allies established a second front they could attain unmatchable
material superiority, but could wage attrition warfare, which could
persuade one or more of the Allies to stop short of complete
occupation.139 By February 1945, however, even this alternative
appeared likely to fail, because German forces could no longer hold any
line for very long and could not impose attrition on the advancing forces
for a protracted period. The key question is why surrender did not occur
at this point.
Staunch and long committed Nazi leaders saw no point in surrender at
any time, since nothing of Nazi ideals would survive if Germany were
occupied. Dr Josef Goebbels wrote on 14 November 1943: 'As far as we
are concerned, we have burned the bridges behind us. We no longer can
turn back, but we also no longer want to. We are forced to extremes and
therefore also ready for extremes... . We will go down in history either
as the greatest statesmen or as the greatest criminals.140
German military leaders, however, did not have everything to lose.
Many officers were not committed Nazis, but in fact conservative
aristocrats who were opposed to National Socialism's class levelling
tendencies. Although most members of this group remained politically
inactive, they were driven not by their interest in retaining control over
the German homeland, which they generally saw as a lost cause, but by
their fear of Russian occupation and their desire to reduce those costs.
Throughout the war, the Germans were exceedingly brutal and

genocidal in the East, not just the Nazis' special murder organs but the
entire Army was also deeply involved.141 In addition to widespread
pillage, destruction, and murder, the Army was also directly involved in
the war against partisans, which involved punishing local civilian
communities and killed from 1 to 1.5 million people.142 Further, Soviet
prisoners were treated abominably; of the 3.5 million Soviet prisoners
captured during 1941, over 60 per cent died by February 1942.143
As a result, it is not surprising that German military leaders expected
that Russian occupation would be exceedingly brutal. In spring 1944
Soviet public declarations emphasized the stark demand that the entire
Wehrmacht should be employed as slave labor by the victorious
Allies.144 As the year continued reports about the degree of punishment
to be extracted from German society increased and, as Soviet armies ran
on to German soil, the treatment of occupied German areas was harsh.
In January 1945 the German government obtained a copy of an Allied
map showing the proposed partition of Germany among the USSR, UK
and US. This confirmed the beliefs of many that acceptance of
unconditional surrender would be exceedingly costly, putting an end to
Germany as a national entity and leading to the loss of many lives as a
result of brutal Russian occupation. As Admiral Karl Doenitz said:
Painful though it was to be compelled by these pressing

considerations to continue the war in the winter of 1944—45, to
sacrifice more soldiers and sailors on all fronts and all the world
over and to endure yet more casualties among civilians in air-raids,
there was no option: for the losses were smaller than had we
prematurely given up our eastern territories.145
In February 1945 immediate surrender did not take place, because

continued resistance was expected to mitigate the civilian costs of
eventual defeat. Although further resistance could not avoid defeat and
would involve additional costs, German leaders believed that it would
avoid even greater civilian costs in the form of Russian exploitation and
revenge after surrender, because it would permit time for refugees to
escape from the areas to be occupied by Russia into areas under the
control of the Western Allies. From January to May 1945 two to three
million Germans were evacuated by sea alone from eastern territories to
the West.146
The failure of Germany to surrender emphasizes the importance of
the full Denial Theory offered in this study. While the military
vulnerability proposition alone has great explanatory power as well as
extreme parsimony, domestic politics and the costs of surrender can

cause states to continue even hopeless resistance, either because
surrender could destroy the ruling class and allied social institutions or
because the costs of surrender could exceed the costs of continued
resistance.
V. Conclusion
Coercion operates fundamentally differently from deterrence, and
nuclear and conventional coercion function differently from each other.
No one type of strategy or threat will lead to successful coercion of all
target states or in all circumstances.
In conventional conflicts, successful coercion is a function mainly of
the interaction between the military strategies of the coercer and target.
When the coercer can damage the target state's capabilities in ways that
undermine the target's expectation of military success, coercion is likely
to succeed. Inflicting damage is not enough; the coercer's attacks must
be accurately designed to thwart the target's ability to control the
territory at issue. By contrast, the vulnerability to attack of the target's
civilian population rarely makes much difference. Inflicting enough pain
to subdue the resistance of a determined adversary is normally beyond
the reach of conventional weapons.
Nuclear coercion, on the other hand, is a function primarily of civilian
vulnerability, and is not much affected by military vulnerabilities.
Nuclear weapons are capable of inflicting more pain than any society
can stand, so much so that battlefield outcomes become unimportant by
comparison. Further, in nuclear cases the coercer as well as the target
may face devastation. In this situation, nuclear coercion will be rare and
difficult to pursue for more than very limited aims.
As the policy debates over US strategy for coercing Iraq demonstrate,
the determinants of coercive success are inadequately understood. Five
major policy implications follow from my findings.
First, advocates of counter-civilian coercion make two mistakes: they
recommend harming innocent civilians, and they advocate doing so for
no good purpose. Civilian threats are thus both immoral and wasteful.
Unless victims have exceedingly modest interests at stake, counter-
civilian strikes will not produce coercive leverage. In serious
international disputes, the only threat to civilian populations powerful
enough to extract territorial concessions would be nuclear. This lesson
applies not only to the United States, but also to all states, including
those in the Third World which have or may acquire modern aircraft
and missiles that can strike at the homeland of other states.
Second, if a state faces the grim necessity of using air power for
coercion, it should think carefully about how to maximize the effects on
the opponent's military strategy. This will not only be more effective,
but also harm fewer civilians.
Third, conventional coercion is exceedingly intricate, and may
sometimes be beyond the ability of even the most skilled and powerful
military organizations to execute adeptly. Success requires denying the
victim the wherewithal to execute his strategy; it is not simply a matter
of destroying military targets, but of destroying the right targets in time
to cripple the strategy. There are powerful reasons to doubt whether
large military organizations are sufficiently flexible to pursue this form
of coercion. Organizational imperatives can cause militaries to execute
the war plans they are designed to perform rather than to tailor attacks
to the enemy's vulnerabilities.
Fourth, in some cases potential targets may be immune to
conventional coercion, because not all battlefield strategies have
exploitable military vulnerabilities. Guerrilla war strategies, for
instance, depend little on the types of concentrated, visible military
resources which are vulnerable to coercive attacks. Thus, even the best-
executed air interdiction campaign would produce little coercive
leverage.
Finally, my theory leads to uncomfortable implications about the
future of nuclear coercion. Up until now, nuclear coercion has been
largely confined to the early stages of the Cold War. These events were
rare not only because the Soviets acquired a secure assured destruction
capability, but also because Americans and Soviets extended deterrence
to allies in important trouble spots. Consequently, the death of the Cold
War and disintegration of the Soviet Union may have the effect of
expanding the scope for nuclear coercion in two ways. The least likely is
that, with the Soviets removed as guarantors of Third World regimes
hostile to the West, the United States may be more tempted to employ
nuclear threats in future conflicts with these states. The more likely may
be that nuclear coercion may be attempted by mid-level nuclear powers
in areas where the end of the Cold War has led to both American and
Soviet disinterest (such as India v. Pakistan). Alas, my theory expects
nuclear coercion in this widened scope to be more likely to succeed than
fail.
NOTES
1. 'Coercion' is a more natural English word which I use to refer to the same concept as
Thomas C. Schelling's 'compellencc' in Arms and Influence (New Haven, CT: Yale
UP, 1966).
2. There have been three main cycles of policy advocacy. The first, between World
Wars I and II, was dominated by advocates and critics of independent air operations.
Advocates include J. F. C. Fuller, The Reformation of War (NY: Dutton, 1923); B. H.
Lidell Hart, Paris or the Future of War (London: Kegan Paul, 1925); and Guilio Dou-
het. Command of the Air, [Rome, 1921] trans. Dino Ferrari (NY: Coward-McCann,
1942). More critical of independent air operations is J. C. Slessor, Air Power and
Armies (London: OUP, 1936). The second, occurring after World War II, centered on
how to respond to potential Soviet nuclear blackmail: Hans Speier, 'Soviet Atomic
Blackmail and the North Atlantic Alliance', World Politics 9/3 (April 1957),
pp.307-28; Paul Nitze, 'Atoms, Strategy and Policy', Foreign Affairs 34/2 (Jan.
1956), pp.187-98; and Arnold L, Horelick and Myron Rush, Strategic Power and
Soviet Foreign Policy (Chicago: Univ. of Chicago Press, 1966). The last occurred
during the Vietnam War, e.g. the exchange between Colin Gray, 'What RAND Hath
Wrought', Foreign Policy, No.4 (Fall 1971), pp.111-29, and Bernard Brodie, 'Why
Were We So (Strategically) Wrong?' Foreign Policy, No.5 (Winter 1971-72),
pp.151-87.
3. Daniel Ellsberg, 'Theory and Practice of Blackmail', P-3883 (Santa Monica, CA:
RAND Corp., 1968); Morton Kaplan, Strategy of Limited Retaliation (Princeton:
Center of Int'l Studies, Woodrow Wilson School of Public and Int'l Affairs, Prince-
ton Univ. 1959); Thomas C. Schelling, Strategy of Conflict (Cambridge: Harvard UP,
1960); idem., Arms and Influence (note 1); and Alexander George, 'Some Thoughts
on Graduated Escalation', RM-4844-PR (Santa Monica, CA: RAND Corp., 1965);
and George, William Simons, and David Hall, Limits of Coercive Diplomacy (NY:
Little, Brown, 1971).
4. Fred C. Iklé's Every War Must End (NY: Columbia UP, 1971) explores the domestic
political dynamics of wartime surrender decisions, but has received little attention
from students of coercion.
5. Richard K. Betts, Nuclear Blackmail and Nuclear Balance (Washington, DC: The
Brookings Inst., 1987) is the best work in this area.
6. Land and sea power can also be used for coercive purposes, but air power offers im-
portant methodological advantages, as discussed below on pp.450-1.
7. On the difference between coercion and deterrence, see Robert J. Art, 'To What
Ends Military Power?' International Security 4/2 (Spring 1980), pp.3-35; and Schell-
ing, Arms and Influence, pp.59-91 (note 1).
8. The two are intimately linked in practice. At the same time that the coercer hopes to
force the target state to change its behavior, the target hopes to deter the coercer
from executing the coercive threat.
9. For recent research on non-military sanctions, see David Baldwin, Economic State-
scraft (Princeton: Princeton UP, 1985); John Conybeare, Trade Wars (NY: Columbia
UP, 1987); and Michael Mastanduno, 'Strategies of Economic Containment', World
Politics 37/4 (July 1985), pp.503-31.
10. Schelling also uses coercion to refer to bargaining between victor and defeated, but
this definition is too broad to be helpful. First, since the victor must pay the full costs
of winning the war to reach this situation, he gains no benefit from coercion. Second,
since the victor has the ability to inflict unlimited harm on the defeated at little or no
cost, coercion in this situation nearly always succeeds, but these trivial succeses tell
us little about cases in which the victim can resist. See Arms and Influence, pp.12-15
(note 1).
11. Complete defeat means that the government of the losing side no longer controls
organized forces capable of significantly impeding the victor's operations. E.g., by
early May 1945 the German government could no longer organize any concerted re-
sistance to Allied occupation of the country, even though some combat units
remained in the field.
12. Denial here means preventing the opponent from satisfying his political objectives by
military means. Thus, it may require stopping the opponent from either gaining or
holding territory, depending on whether the threatener's goal is to prevent an attack.
stop an ongoing attack, or force territorial concessions. This is a broader definition of

denial than that used by deterrence theorists, who usually refer only to defeat of an
anticipated attack. For an example of the narrower definition, see Glenn Snyder,
Deterrence by Punishment and Denial, Research Monograph No.1 (Princeton:
Princeton Univ., Center of Int'l Studies, 1959).
13. Distinguishing between denial strategies of coercion and warfighting in actual mili-
tary operations is easier in some circumstances than others. For example, when
assailants have limited aims, a denial strategy of coercion needs only to threaten the
victim's ability to control the disputed territory, not to pose risks of complete defeat.
Also, military strategies that depend on surprise for their affectiveness have no coer-
cive value because they cannot be used to threaten the target with defeat.
14. Some scholars refer to 'balance of resolve', but this is often seen as little more than a
reflection of the balance of interests. For instance, critical risk assessments generally
turn on the interaction of actors' interests in the issue of dispute, because military
costs are generally assumed to be equally high for both sides. Glenn H. Snyder and
Paul Diesing, Conflict Among Nations (Princeton: Princeton UP, 1977), argue that re-
solve is primarily a function of interests.
Sometimes the balance of resolve refers to a separate theory that coercive out-
comes are a function of sheer political will. However, these formulations are
generally too vague to be falsifiable. For examples of the broader meaning of resolve,
see Fred C. Iklé How Nations Negotiate (NY: Harper & Row, 1964); Schelling, Arms
and Influence (note 1); and Robert Jervis, The Logic of Images in International Re-
lations (Princeton: Princeton UP, 1970).
15. In addition to Snyder and Diesing, see Robert E. Osgood and Robert W. Tucker,
Force, Order and Justice (Baltimore, MD: The Johns Hopkins UP, 1967); George, et
al. Limits of Coercive Diplomacy; Andrew Mack, 'Why Big Nations Lose Small
Wars', World Politics 27 (1975), pp.175-200; Robert Jervis, 'Why Nuclear Super-
iority Doesn't Matter', Political Science Quarterly 94/4 (Winter 1979-80), pp.617-33.
16. Paul H. Nitze, 'Assuring Strategic Stability in an Era of Detente', Foreign Affairs
54/2 (Jan. 1976), pp.207-32, and 'Deterring Our Deterrent', Foreign Policy, No.25
(Winter 1976-77), pp.195-210; William R. Kintner and David C. Schwartz, A Study
on Crisis Management (Philadelphia: Univ. of Pennsylvania, Foreign Policy Res.
Ins., 1965); Betts, Nuclear Blackmail (note 5); and Edward Luttwak, The Strategic
Balance, 1972, Washington Papers, No.3 (NY: The Library of Congress, 1972).
17. Schelling, Arms and Influence (note 1); George Quester, Deterrence Before Hirosh-
ima (NY: Wiley, 1966); Liddell Hart, Paris; Fuller, Reformation of War; Douhet,
Command of the Air; and Zeev Maoz, 'Resolve, Capabilities, and the Outcomes of
International Disputes, 1816-1976', Journal of Conflict Resolution 27/2 (June 1983),
pp.195-229.
18. Part of the reason why these problems have not been corrected is that empirical test-
ing of hypotheses about coercion has been surprisingly thin. We have only one
serious empirical study of military coercion, Betts, Nuclear Blackmail, which is
limited to nuclear cases and does not always distinguish coercion from deterrence.
Neither Walter J. Petersen/Deterrence and Compellence: A Critical Assessment of
Conventional Wisdom', International Studies Quarterly 30 (1986), pp.269-94 nor
George, Limits of Coercive Diplomacy (note 3) systematically compare alternative
explanations for success and failure.
Most formal analyses have also concentrated on nuclear coercion, and generally
operate at levels of abstraction above the key causal factors in military coercion,
namely the military strategies of the assailant and victim. 'Strategy' in formal models
normally refers to decisions about whether to stand firm or make concessions, not to
different methods for translating force into coercive pressure by attacking kinds of
targets. Different military strategies available to an assailant are usually abstracted
into a single variable in which different varieties of military costs are treated as
wholly interchangeable. E.g., see Ellsberg, 'Theory and Practice of Blackmail' (note
3).
19. David Baldwin, 'Power Analysis and World Politics', World Politics 31/2 (Jan. 1979)
argues that compellence (coercion) and deterrence are indistinguishable and that the
alleged difficulty of the former results from a practice of calling hard cases com-
pellent and easy ones deterrent. But the distinction is not merely semantic. Changing
the status quo is different from maintaining it, as well as harder. Petersen, 'Deter-
rence and Compellences', finds that only 24 per cent of compellent threats were
successful in the cases he examined compared to 63 per cent of deterrent threats,
although the difference disappears when both side's expected costs are controlled
for. However, the meaning of this finding is unclear, because his samples are not
comparable. The average severity of the coercive cases is lower than that of the
deterrent ones (p.282).
20. Paul Kecskemeti, Strategic Surrender (Stanford: Stanford UP, 1958).
21. Robert Jervis, The Illogic of American Nuclear Strategy (Ithaca, NY: Cornell up,
1984).
22. Amos Tversky and Daniel Kahneman, 'Judgment under Uncertainty', Science,
Vol.185 (1974), pp.1124-30.
23. Morton A. Kaplan, 'The Calculus of Deterrence', World Politics 11/1 (Oct. 1958),
pp.20-43.
24. One implication of this is that attempts by one nuclear-armed state to coerce another
will be rare, because protection against retaliation is impossible.
25. Schelling, Arms and influence (note 1).
26. Ibid.
27. These problems do not usually affect deterrence because, although intra-war deter-
rence is possible (e.g., attempts to deter escalation), most deterrence cases involve
discouraging an aggressor from going to war.
28. John W. Dower, War without Mercy: Race and Power in the Pacific War (NY: Pan-
theon Books, 1986).
29. Ikle, Every War Must End (note 4).
30. John J. Mearsheimer, 'The Theory and Practice of Conventional Deterrence', (Ith-
aca, NY: Ph.D., Diss., Cornell Univ., 1981), Ch.4.
31. This also partly explains why surrenders tend to occur later, rather than sooner, than
most theories expect. If sunk costs are a key obstacle to successful coercion, then we
should expect this obstacle to grow the longer war lasts.
32. On requirements for successful deterrence, see John Mearsheimer, Conventional
Deterrence (Ithaca, NY: Cornell UP, 1983).
33. George: ' . . . coercive diplomacy can succeed only if the opponent accepts as credible
the threat of punishment for non-compliance with the demands made upon him'
(1971, p.238: note 3). Betts: 'The notion of blackmail in this study . . . means
coercion by the threat of punishment . . . . (1987, p.4: note 3). Ellsberg: 'My problem
as a blackmailer is to convince you that I am "too likely" to respond with my second
strategy. Punish, for you to accept the risk that your own second strategy, Resist,
would entail' (1968, p.347: note 3). Schelling: 'The ideal compellent action would be
one that, once initiated, causes minimal harm if compliance is forthcoming and great
harm if compliance is not forthcoming . . .' (1966, p.89: note 1).
34. There may also be policy consequences. Restricting studies of coercion to counter-
civilian strategies suggests that a punishment threat is the best coercive strategy
against all possible victims. If policy-makers believe this unexamined assumption,
they may also jump to the conclusion that the key to success is simply to com-
municate this threat clearly to the intended victim. If the premise is wrong, this can
have disastrous consequences. Policy-makers could easily assume that failure results
from errors in execution. Since the actual source of the error would be more funda-
mental, this is a prescription for repeating past mistakes. For evidence that
policy-makers in the Johnson administration assumed that coercion derived from
punishment, see Robert A. Pape, Jr., 'Coercive Air Power in the Vietnam War',
International Security 15/2 (Fall 1990), pp.103-46.
35. American planners in World War II considered Japanese cities more vulnerable than
German cities because they were constructed mainly of wood rather than brick.
36. Examples arc Schclling, Arms and Influence (note 1) and George, Limits of Coercive
Diplomacy (note 3).
37. Robert Jervis, 'The Political Effects of Nuclear Weapons', International Security 13/2
(Fall 1988), pp.80-90.
38. Whether the state is seeking to make gains or avoid losses is equivalent for the pur-
pose of explaining the logic of coercion, since in all cases we mean benefits compared
to the state's position if concessions are made.
39. The logic of coercion can be described by a simple equation:
V(r) = B(r) p(B) - C(r) p(C)
Where: V(r) = value of resistance; B(r) = benefits of resistance; p(B) = probabil-
ity of costs. Concessions occur when V(r) < O.
40. Carlton J. H. Hayes, Essays on Nationalism (NY: Macmillan 1926); and Boyd
Shafer, Faces of Nationalism: New Realities and Old Myths (NY: Harcourt, Brace,
Jovanovich, 1972).
41. Stephen M. Walt, Origins of Alliances (Ithaca, NY: Cornell UP, 1987).
42. This also imples that economic inducements (bribery) are poor complements to mili-
tary coercion over significant interests. Whether employed simultaneously or
promised to sweeten deals, they are not likely to significantly influence the victim's
decision, since they are likely to be trivial compared to the territorial interests at
stake.
43. Low credibility threats can coerce only when the potential damage drastically out-
weighs any possible benefits, a situation that is possible only with nuclear weapons.
When both sides are nuclear armed, coercive threats tend to be relatively incredible,
because the coercer as well as the target must fear disastrous costs.
44. E.g., in 1972, the US was able to prevent North Vietnam from continuing offensive
operations, but not from retaining control over territories they occupied in the south.
Accordingly, Washington was able to pressure Hanoi into a ceasefire agreement but
not into a withdrawal.
45. Anne Armstrong, Unconditional Surrender (New Brunswick, NJ: Rutgers UP, 1962).
46. Nationalism is primarily a political principle, which holds that nations - a body of in-
dividuals that share a common set of cultural or social characteristics and a belief that
these attributes separate them from others - should have their own states - the
agency within society that possesses a monopoly on legitimate force. On the defini-
tion and rise of nationalism, see Ernest Gellner, Nations and Nationalism (Ithaca,
NY: Cornell UP, 1983); and John Breuilly, Nationalism and the State (Chicago: Univ.
of Chicago Press, 1982).
47. Gellner, Nations and Nationalism, p.1.
48. Democratic states often have especially high resolve, as a result of governmental
efforts to sell (or oversell) foreign policy commitments in order to drum up wide-
spread public support. This tendency can be exaggerated further when efforts to sell
the public inflated images of the importance of foreign policy interests lead to 'blow-
back', trapping governments into maintaining commitments long after they would
have preferred to abandon them. Authoritarian states, to the extent they rely on
public enthusiasm for their policies, are also susceptible to these pressures. For dis-
cussion of the resolve of democracies, see George Kennan, 'American Diplomacy,
1900-1950 (Chicago: Univ. of Chicago Press, 1951) and Stephen Van Evera, 'Causes
of War', (Berkeley: Ph.D. Diss., Univ. of California, 1984).
For examples of authoritarian states relying on nationalist sentiments in Wil-
helmian German politics, see Ludwig Dehio, Germany and World Politics in the
Twentieth Century, trans. Dieter Pevsner (NY: Norton, 1967); Fritz Fischer, War of
Illusions: German Policies from 1911 to 1914, trans. Marian Jackson (NY: Norton,
1975); Paul M. Kennedy, The Rise of the Anglo-German Antagonism, 1860-1914
(London: Allen and Unwin, 1980), Ch.18; Hans Kohn, The Mind of Germany: The
Education of a Nation (NY: Harper Torchbook, 1965), Chs.7-12; and Louis L.
Snyder, German Nationalism: The Tragedy of a People (Harrisburg, Pa.: Telegraph
Press, 1952).
49. Hayes, Essays on Nationalism; Shafer, Faces of Nationalism; and Quincy Wright, A
Study of War, Vol.2 (Chicago: Univ. of Chicago Press, 1942).
50. Michael Howard, War in European History (New York: OUP, 1976), pp.110-11.
51. Van Evera, 'Causes of War', pp.407-21 (note 48).
52. Phillip Knightley, The First Casualty (NY: Harcourt, Brace, Jovanovich, 1975);
David Wise, The Politics of Lying: Government Deception, Secrecy and Power (NY:
Random House, 1973); David M. Kennedy, Over Here: The First World War and
American Society (New York: OUP, 1980), Ch.l.
53. Van Evera, 'Causes of War', pp.683-5 (note 48).
54. Examples of militaries suppressing criticism to protect the structure and leadership of
the organization include the British slowness to exploit tanks, commitment to frontal
vs. infiltration attacks, continuing to use cavalry, the Imperial Japanese Navy's
failure to evaluate the role of the battleship in World War II, and the British Royal
Navy's efforts to avoid convoys in World War I. On obstacles to military innovation,
see Barry Posen, Sources of Military Doctrine (Ithaca NY: Cornell UP. 1984); and
Stephen Peter Rosen, Winning The Next War (Ithaca NY: Cornell UP, 1991).
55. Aaron Wildavsky, 'The Self-Evaluating Organization', Public Administration Review
32/5 (Sept.-Oct. 1972), pp.509-20. For a general application of these ideas to state
foreign and military policy-making as a whole, see Van Evera, 'Causes of War',
pp.453-99 (note 48).
56. Van Evera, 'Causes of War' (note 48).
57. Edward Shils, The Torment of Secrecy (Glencoe, IL: Free Press, 1956), p.58. Janis
and Mann argue that such 'defensive avoidance' is likely to occur whenever people
face choices between highly unpalatable options, such as between making con-
cessions and military defeat. Decision Making (NY: Free Press, 1977). For examples
of failure to recognize impending defeats, see Geoffrey Blainey, The Causes of War
(NY: Free Press, 1973); and Iklé, Every War Must End (note 4).
59. Similarly, the rulers themselves may offer up the top leader as a sacrificial lamb if this
will save the remainder of the government. E.g., in July 1943 the Italian government
jettisoned Mussolini in an unsuccessful effort to persuade the Allies not to replace
the non-Fascist ruling circles in power. Ernest May, 'Lessons' of the Past: The Use
and MisUse of History in American Foreign Policy (NY: OUP, 1973), pp.125-43.
60. Although the Nazis made alliances with other groups, they would not have con-
sidered a successor regime led by industrialists or the Army as an acceptable
alternative. Karl D. Bracher, The German Dictatorship (NY: Praeger, 1970).
61. Vast research has yet to uncover significant positive relationships between foreign
conflict and the increase of domestic divisions. The seminal study is Rudolph J. Rum-
mel, 'Dimensions of Conflict Behavior Within and Between Nations', General
Systems. Vol.8 (1963), pp.1-50. For a review of follow-on studies, see Dina A. Zines,
Contemporary Research in International Relations (NY: The Free Press, 1976),
pp.160-75. Some research suggests that international conflict actually increases dom-
estic support for governments, see Jack Levy, 'The Diversionary Theory of War: A
Critique', in Manus Midlarsky (ed.). Handbook of War Studies (Boston: Unwin
Hyman, 1989), pp.259-88.
62. German and Japanese civilians continued to work to support their countries' war
efforts despite not believing their government's claims that victory would be
achieved. USSBS, The Effects of Strategic Bombing on German Morale (Washing-
ton, DC: GPO, 1946); idem., The Effects of Strategic Bombing on Japanese Morale
(Washington, DC: GPO, 1947).
63. Charles Tilly, 'Revolutions and Collective Violence', in Handbook of Political
Science (eds.), Fred Greenstein and Nelson Polsby (Reading, MA: Addison-Wesley,
1975), Vol.3, Macropolitical Theory, pp.483-556.
64. Mearsheimer, 'Theory and Practice' (note 30); Bernard Brodie, War and Politics
(NY: Macmillan Press, 1973), Ch.l; and Gordon Craig, The Politics of the Prussian
Army, 1640-1945 (NY: OUP, 1955).
65. In principle, risk-based strategies could be used to threaten either denial or punish-
ment, but in practice they are hardly ever used to threaten the victim's strategy to
achieve his goals. The slow pace allows the victim too much time to recover from the
effects of attacks, making this strategy inherently ineffective.
66. There is no comprehensive data on civilian deaths in modern wars. Data on battle
deaths, which nearly always exceeds civilian losses, show that states rarely lose more
than 2 per cent of their pre-war population. Even in World Wars I and II losses
exceeded 5 per cent of pre-war population for only a few states. J. David Singer and
Melvin Small, The Wages of War 1816-1965: A Statistical Handbook (NY: John Wiley,
1972), pp.351-7.
67. Although it is common for advocates of punishment-based theories of coercion to
argue that economic deprivation can produce popular unrest, social scientists have
long discounted this hypothesis. See David Snyder and Charles Tilly, 'Hardship and
Collective Violence in France, 1830 to 1860', American Sociological Review 37 (Oct.
1972), pp.520-32; Charles Tilly, Louise Tilly, and Richard Tilly, The Rebellious
Century: 1830-1930 (Cambridge, MA: Harvard UP, 1975); and Steven Finkel and
James Rule, 'Relative deprivation and Related Psychological Theories of Civil
Violence: A Critical Review', in Louis Kriesberg (ed.). Research in Social Move-
ments, Conflicts and Change (Greenwich, CT: JAI Press, 1986), Vol.9, pp.47-69.
68. USSBS, Effects of Strategic Bombing on Japanese Morale, pp.65, 249 (note 62).
69. Defense against nuclear attack is much less effective since even one penetrating
bomber can do immense damage.
70. Robin Higham, Air Power: A Concise History (NY: St. Martin's Press, 1972),
pp.114-19.
71. Civilians may also be evacuated from areas threatened by the advance of enemy
troops.
72. Significant evacuation is possible because damage caused by conventional munitions
is inflicted piecemeal, accumulating slowly over time. The assailant cannot attack all
the victim's cities at once, only one or a few at a time, making time for evacuation of
all but the first few targets attacked. This would not apply with nuclear weapons,
because numerous population centers can be attacked simultaneously, since one war-
head or only a few need be delivered against each.
73. Karl Lautenschlager, 'The Submarine in Naval Warfare, 1901-2001', International
Security 11/3 (Winter 1986-87), pp.94-140.
74. Mancur Olson, Jr., The Economics of the Wartime Shortage: A History of British
Food Supplies in the Napoleonic War and in World Wars I and 11 (Durham, NC:
Duke UP, 1963), p.142.
75. The idea of manipulating the risk of punishment for political purposes has largely
come to be identified with the work of Thomas Schelling. Others also shared in de-
veloping this idea, chief among them Kaplan, Strategy of Limited Retaliation (note 3)
and George, 'Graduated Escalation' (note 3).
76. Schelling writes, 'To be coercive, violence has to be anticipated. . . . It is the ex-
pectation of more violence that gets the wanted behavior, if the power to hurt can get
it at all.' Arms and Influence, pp.2-3 (note 1).
77. Schelling writes: 'The ideal compellent action would be one that, once initiated,
causes minimal harm if compliance is forthcoming, and great harm if compliance is
not forthcoming, is consistent with the time schedule of feasible compliance, is
beyond recall once initiated, and cannot be stopped by the party that started it but
automatically stops upon compliance, with all this fully understood by the adversary.'
Arms and Influence, p.89 (Emphasis in original).
78. Schelling strategies thus seek to encourage minor demonstrations of willingness to
accommodate the assailant's demands, as well as major concessions.
79. In the exceptional case where the victim cannot regenerate damaged capabilities,
risk-based strategies may be effective. Iraq in autumn 1990 may have been in this
position. Having no indigenous arms industry, Iraq could not replace losses.
80. Domestic political constituencies may be unwilling to countenance the large-scale
suffering that would result from prosecuting a punishment campaign, or the assai-
lant's government may fear that more intensive attacks would provoke third party
intervention on behalf of the victim. For instance, the US bombing of Libya in April
1986 was limited to a single raid, both because the American public would not have
supported extensive operations and because it would also have offended America's
European Allies. Aside from increasing the risk of further attacks if Libya did not
stop supporting terrorism, a second objective appears to have been to kill Col.
Gaddafi himself. Tim Zimmermann, 'The American Bombing of Libya', Survival
29/3 (May/June 1987), pp.195-214. This case also demonstrates that credibility may
also be lost, because the assailant may actually become less determined to continue
inflicting costs as the campaign progresses.
81. Mark Clodfelter, The Limits of Air Power: The American Bombing of North Vietnam
(New York: Basic Books, 1989), pp.39-72.
82. This case is discussed in detail in Pape, 'Coercive Air Power', (note 34).
83. Stephen Cimbala, Strategic War Termination (NY: Praeger, 1986).
84. This last consideration does not appear to have been considered by American
nuclear planners in the massive retaliation era. If early SIOPs had been executed,
many million people outside the Soviet Union, including Japan, China and Europe,
would have been killed. Scott Sagan, 'Nuclear Alerts and Crisis Management', Inter-
national Security 9/4 (Spring 1985), pp.99-139.
85. The most important issue was the repatriation of Chinese PoWs.
86. Two political events may also have helped make this threat more credible. The elec-
tion of Eisenhower reduced domestic constraints on US action, while the death of
Stalin in March 1953 reduced American fears of possible Soviet retalation.
87. For example, standard analyses of nuclear first strikes show that either superpower's
nuclear forces can absorb heavy losses and still retain effective retaliatory capabil-
ities. Charles Glaser, Analyzing Strategic Nuclear Policy (Princeton: Princeton
University Press, 1990).
88. The US inflicted heavy losses on North Vietnamese POL fuel stockpiles in 1965-68,
but this hardly effected the progress of the war. Germany's strategy in World War II
did not depend on the ability to move resources overseas, so the loss of the German
surface navy had little affect on the outcome of the war.
89. During June-Dec. 1941 the Red Army lost over 5 million men compared to its pre-
war strength of 3.5 million, but was substantially rebuilt by the next summer. John
Keegan, The Second World War (NY: Viking. 1990), pp.173-208.
90. To compensate for heavy naval losses early in World War II, Britain substantially
withdrew her fleet from the Pacific and Mediterranean to concentrate in the vital
Atlantic. In late 1941 the Soviets withdrew large forces from the Far East to commit
them to the decisive winter battle for Moscow. B. H. Liddell Hart, History of the
Second World War, (London: Cassell, 1970).
91. Of course, complete destruction of all the victim's military capabilities will under-
mine any strategy, but this means that the attacked has achieved - and paid the full
costs to achieve - decisive military victory and thus counts as a coercive failure, not a
success.
92. In this context, 'mechanized' war refers to the dominance of the types of mechanical
weapons and transport provided by the industrial revolution in modern war. A
second meaning not used here refers to battle tactics that rely on armored vehicles
and rapid mobility, such as the German Blitzkrieg.
93. Tactics vary; the attacked may seek to wear down the opponent in a series of set-
piece battles (attrition), or may seek to break through the enemy lines and disrupt
the opponent's ability to fight by raiding the adversary's rear areas (Blitzkreig).
Mearsheimer, Conventional Deterrence, Ch.2 (note 32).
94. A. F. Krepinevich, Jr., The Army and Vietnam. (Baltimore: Johns Hopkins UP,
1986) pp.3-10. Representatives of the extensive literature on guerrilla warfare are
David Galula, Counterinsurgency Warfare (NY: Praeger, 1964); Sir Robert G. K.
Thompson, No Exit from Vietnam (London: Chatto and Windus, 1969); and Douglas
S. Blaufarb, The Counterinsurgency Era (NY: Free Press, 1977).
C O E R C I O N AND MILITARY S T R A T E G Y 473
95 James Dunnigan, How to Make War (NY: Morrow, 1982), pp.308-22; Martin Van
Creveld, Supplying War (NY: CUP, 1977).
96. See Pape, 'Coercive Air Power' (note 34).
97. For German intentions in the Battle of Britain, see F. M. Sallagar, The Road to Total
War: Escalation in World War II, R-465-PR (Santa Monica: RAND Corp., 1969),
p.7.
98. Len Deighton, Fighter: The True Story of the Battle of Britain (London: Jonathan
Cape, 1977); Derek Wood and Derek Dempster, The Narrow Margin: The Battle of
Britain and the Rise of Air Power, 1930-1940 (London: Hutchinson, 1961); and John
Terraine, A Time for Courage: The Royal Air Force in the European War, 1939-1945
(NY: Macmillan, 1985).
99. These successes are rare because often the assailant will have only low interests at
stake as well and so be easily deterred from issuing a coercive threat. In other words,
successful coercion based on punishment normally requires the conjunction of three
conditions: low interests by the victim; balance of interests favoring the assailant; and
balance of capabilities favoring the assailant.
100. E.g., if the assailant imposes a naval blockade that is certain to cause famine (punish-
ment) or industrial collapse (denial) but requires some months to take full effect, the
victim will not react as soon as these consequences can be forecast with certainty, but
only when the main impact is expected in the short term.
101. Glaser, Strategic Nuclear Policy (note 3).
102. Arthur Katz, Life After Nuclear War (Cambridge, MA: Ballinger, 1982).
103 Bernard Brodie, Escalation and the Nuclear Option (Princeton: Princeton UP, 1968).
104. Betts, Nuclear Blackmail (note 5).
105. Jervis, Illogic (note 21).
106. However, if the balance of capabilities is always held constant, axiomatically its
impact cannot be tested.
107. Although the assailant may prevail despite an equal or unfavorable balance of in-
terests, we should not expect nuclear coercion attempts over trivial issues. Potential
assailants will be quite concerned by the significant costs that even modest nuclear re-
taliation would bring, and will only attempt coercion when they have important
interests at stake.
108. Potential coercers may still be constrained by humanitarian, domestic, political, or
alliance considerations.
109. This typology is traditionally used by scholars of international politics. For a recent
example, see Michael Desch, 'The Keys that Lock Up the World', International
Security 14/1 (Summer 1989), pp.86-121.
110. On this assumption, see Colin Gray, 'The Strategic Forces Triad', Foreign Affairs
56/4 (July 1978), p.774.
111. Since the focus is on anticipation of who will win, we include additional forces that
either side has available elsewhere and plans to commit in the theater even if they
have not yet arrived.
112. The full power of the theory cannot be tested without case studies because appro-
priate operationalization of all variables depends on the context of a coercive
dispute. Case studies of many of the important cases are conducted in Robert A.
Page, Jr., Punishment and Denial: The Theory and Practice of Coercive Air Power
(Ithaca, NY: Cornell UP, forthcoming).
113. In some cases, the assailant is employing other coercive instruments in addition to air
power. Thus, air power alone may not account for the outcome of all the cases.
114. For discussion of coercive sea and land power, see Pape, Punishment and Denial
(note 112).
115. Schelling's theory does this.
116. Moreover, while postwar concessions from defeated to victor might be called coer-
cive success, if war must be pursued to its ultimate conclusion, coercion has failed in
its more important task of enabling the victor to win concessions without paying the
costs of achieving decisive victory.
117. This coding rule does not count Truman's movement of B-29s to Britain during the
1948 Berlin crisis as nuclear coercion because there is no evidence that the Soviets
understood this as a threat to launch an attack if the blockade continued, but does in-
clude Soviet coercion of the British and French during the 1956 Suez crisis, because
the record indicates that the British and French understood themselves to be victims
of coercion. This rule has the effect of defining coercive attempts from the victim's
perspective, which is appropriate given that the theory is about changes in the
victim's decisions.
118. William Gamson and Andre Modigliani, Untangling the Cold War: A Strategy for
Testing Rival Theories (Boston: Little, Brown, 1971). For detailed explanations for
how cases in this table were coded, see 'Appendix', in Pape, Punishment and Denial
(note 112).
119. When a coercer pursues single set of coercive demands through two or more parallel
military campaigns (which may operate simultaneously), this is still coded as a single
case because it still involves only a single set of decisions by target state on whether to
conced the demands, regardless of whether only one of the campaigns or both con-
tribute to the decision. Cases are divided when codings for independent variables
shift.
120. The British/French 1939 and 1939 cases are combined because both their military and
civilian vulnerability were determined by their joint capabilities, while the British
1940, Germany 1942-1945 and Japan 1944-45 are divided when valued of in-
dependent variables shift. Accordingly, Table 3 has one more case than Table 2.
121. Wesley Craven and James Cate, The Army Air Forces in World War II, Vol.5, The
Pacific (Washington: Office of Air Force History, 1953), pp.617, 722.
122. USSBS, The Effects of Strategic Bombing on German Morale (note 62).
123. The general in charge of the air campaign, Curtis LeMay, viewed the atomic bomb
with disdain, believing it to be unnnecessary for the destruction of urban areas. By
August, his bombers had so thoroughly devastated Japan's larger cities that their
attention had turned to medium and small cities. Curtis E. LeMay with MacKinley
Kantor, Mission with LeMay (Garden City, NY: Doubleday, 1965), p.393.
124. S. Woodburn Kirby, War Against Japan, Vol.5: The Surrender of Japan (London:
HMSO, 1969); Donald Detweiler and Charles Burdick (eds.), War in Asia and the
Pacific, 1937-1949, Vol.12, Defense of the Homeland and End of the War (NY: Gar-
land Publishing, 1980)
125. Leon Sigal, Fighting to a Finish (Ithaca, NY: Cornell UP, 1988), pp.224-81.
126. Ibid., p.226.
127. John Toland, The Rising Sun (NY: Random House, 1970), p.807.
128. Sigal, Fighting to a Finish, p.226.
129. Robert Butow, Japan's Decision to Surrender (Stanford: Stanford up, 1954), p.175.
130. Rosemary Foot, No Substitute for Victory (Ithaca, NY: Cornell UP, 1990), p.166.
131. Edward Keefer, 'President Dwight D. Eisenhower and the End of the Korean War',
Diplomatic History 10 (Summer 1986), pp.267-89.
132. Foot, No Substitute for Victory, p.m.
133. Ibid., p.179; Mark Ryan, Chinese Attitudes toward Nuclear Weapons; China and the
United States during the Korean War (Armonk, NY: M. E. Sharpe, 1989).
134. For detailed examination of this case, see Pape, 'Coercive Air Power' (note 34).
135. By the time the war began, the Bush administration also appears to have wanted to
destroy most of Iraq's offensive capability and to remove Saddam from power.
136. Saddam stated his objective at the start of hostilities: 'Not a few drops of blood, but
rivers of blood will be shed. And then Bush will have been deceiving America,
American public opinion, the American people, the American constitutional in-
stitutions.' Quoted in Lawrence Freedman and Efraim Karsh. 'How Kuwait Was
Won: Strategy in the Gulf War', International Security 16/2 (Fall 1991), pp.28-29.
137. For discussions of the effects of the air campaign, see Freedman and Karsh, 'How
Kuwait Was Won'; R. A. Mason, 'The Air War in the Gulf', Survival 33/3 (May/June
1991), pp.211-29; 'Air Power in Desert Shield/Desert Storm', Air Power History 38/3
(Fall 1991).
138. For discussions of specific Iraqi conditions, see Associated Press, '8 Points of the
Soviet-Iraqi Peace Plan', International Harold Tribune, 23 Feb. 1991; Peter Riddell,
'Bush Regains the Initiative', Financial Times, 23-24 Feb. 1991; Anatoly Repin, 'The
Chance Has Been Missed', Novosd Gulf Bulletin (Moscow), 27 Feb. 1991; and stories
in the New York Times, 23 Feb. 1991.
139. Some German leaders sought to surrender at this point, but their attempted coup
against the Nazi leadership failed.
140. Quoted in Bracher, German Dictatorship, p.283, (note 60).
141. Although the first wave of scholarship after the war tended to exonerate the German
Army for Nazi genocidal policies, subsequent literature stresses that the war in the
East was fundamentally different from the war in the West. In the West, the war was
mainly a military contest between opposing armies, while in the East Nazi genocidal
policies and German Army operations were tightly linked. Illustrative of first gener-
ation historiography on the Wehrmacht in World War II are B. H. Liddell Hart, The
German Generals Talk (NY: William Morrow, 1948); Gordon A. Craig, The Politics
of the Prussian Army, 1940-1945 (NY: OUP, 1955); and Walter Goerlitz, The History
of the German General Staff, 1657-1945 (NY: Praeger, 1959). Representatives of the
generation stressing German Army participation in war crimes include Lucy S. Dawi-
dowicz. The War Against the Jews, 1933-1945 (NY: Bantam Books, 1976); Omer
Bartov, The Eastern Front, 1941-45 (NY: St. Martin's Press, 1986); idem., Hitler's
Army: Soldiers, Nazis, and War in the Third Reich (NY: OUP, 1991); and Raul Hil-
berg, The Destruction of the European Jews (NY: Holmes & Meier, 1985).
142. Timothy Mulligan, 'Reckoning the Cost of the People's War: The German Ex-
perience in Central USSR', Russian History 9/1 (Spring 1982), pp.27-48.
143. This death rate compares to only 4 per cent of Anglo-American PoWs held by
Germany. Bartow, The Eastern Front (note 141), pp.153-4.
144. J. W. Wheeler-Bennett, The Nemesis of Power: The German Army in Politics, 1918-
1945 (NY: St. Martin's Press, 1964), p.619.
145. Karl Doenitz, Memoirs: Ten Years and Twenty Days, trans. R. H. Stevens (London:
Weidenfeld, 1959), p.431. For similar statements by German leaders, see Armstrong,
Unconditional Surrender (note 45).
146. Alfred M. de Zayas, Nemesis at Potsdam: The Anglo-Americans and the Expulsion of
the Germans (Boston: RKP, 1977), p.74.

Coercion and Military Strategy: Why Denial Works and Punishment Doesn't

Uploaded by

Copyright:

Available Formats

Coercion and Military Strategy: Why Denial Works and Punishment Doesn't

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Coercion and Military Strategy: Why Denial Works and Punishment Doesn't

Uploaded by

Copyright:

Available Formats

Coercion and Military Strategy: Why

Denial Works and Punishment Doesn't

Throughout modern history, states have attempted to use military force

there is little rigorous empirical work on nuclear coercion, and none at

I. Defining Military Coercion

action already embarked upon. Coercion differs from deterrence in that

II. Existing Theories of Military Coercion

principles as deterrence. However, while deterrence and coercion are

C. VULNERABILITY OF THE TARGET

urbanization, use of burnable building materials, and susceptibility of

D. NUCLEAR VS. CONVENTIONAL COERCION

III. A Denial Theory of Military Coercion

A. THE LOGIC OF COERCION

interest in protecting its populace. The more important the territorial

would always surrender when the probability of victory disappeared,

B. DECISION MAKING IN MODERN NATION STATES

In wartime, resolve usually increases. As the economic and social

advantages in avoiding evaluation. Military expertise is largely confined

opposition groups are at a great disadvantage in any contest of force,

Assailants' coercive strategies can vary along two dimensions. First,

blockade can cause shortages of key supplies, wreaking havoc on

The relatively slow pace at which damage accumulates also allows

imposed by a punishment strategy. In fact, by further slowing the

Second, although the target usually cannot change basic strategy

the US also sought to undermine North Vietnam's military strategy by

D. PROPOSITIONS ABOUT SUCCESSFUL MILITARY COERCION

is mostly right, though incomplete. The focus on civilian punishment

Under MAD, any potential nuclear assailant must consider the

IV. Testing theories of Military Coercion

Testing theories of military coercion requires three steps. First, the

A. OPERATIONALIZING THE THEORIES

leverage derives from punishing large numbers of civilians. Drive up the

Year Assailant Victim Issue

1. 1917 Germany Britain British withdrawal from WWI

operations, because air forces can be deployed to distant theaters more

incurs little or no cost in complying with the assailant's demands, he

C. TESTING THE THEORIES

Case (Victim) Interests Forces Punishment Outcome

1917/UK failure failure failure failure

Correct 19/32 15/32 18/32 success=14

Case Civilian V«l. Military Vol. Prediction Outcome

1917/UK medium medium failure failure

The Strategic Theory strongly outperforms all the existing theories, as

Theory Success Rate Probability of Random Result

Interests 59% .19

The Punishment and Forces theories have no predictive value; any

Soviet attack against Manchuria, because of its effects in further

China. Militarily, a pattern of escalation in the use of force was

because it threatened civilian or military vulnerability, there is some

logistical centers, transportation arteries, fuel dumps, warehouses,

targets, which destroyed targets that inflicted significant indirect

Detailed examination of the case illustrates why the simple version of

Throughout the war, the Germans were exceedingly brutal and

Painful though it was to be compelled by these pressing

In February 1945 immediate surrender did not take place, because

extreme parsimony, domestic politics and the costs of surrender can

stop an ongoing attack, or force territorial concessions. This is a broader definition of

You might also like