Offense, Defense, and the Causes of War
Stephen Van Evera
International Security, Volume 22, Number 4, Spring 1998, pp. 5-43 (Article)
Published by The MIT Press
For additional information about this article
[ Access provided at 20 Sep 2020 19:38 GMT from Carleton University Library ]
https://muse.jhu.edu/article/446910/summary
the Causesof War
1 Is war more likely
when conquest is easy? Could peace be strengthened by making conquest
more difficult? What are the causes of offense dominance?' How can these
causes be controlled?These are the questions this article addresses.
I argue that war is far more likely when conquest is easy, and that shifts in
the offense-defense balance have a large effect on the risk of war. Ten warcausingeffects
(summarizedin Figure 1)arise when the offense dominates.(1)
Empires are easier to conquer. This invites opportunistic expansion even by
temperate powers (explanation A). (2) Self-defense is more difficult; hence
states are less secure. This drives them to pursue defensive expansion (explanation
B). (3)Their greater insecurity also drives states to resist others' expansion
more fiercely. Power gains by others raise larger threats to national
security; hence expansionism prompts a more violent response (explanation
C). (4)First-strike advantages are larger, raising dangers of preemptive war
(explanation D). (5) Windows of opportunity and vulnerability are larger,
raising dangers of preventive war (explanationE). (6)Statesmore often adopt
fait accompli diplomatic tactics, and such tactics more often trigger war (explanation
F). (7)Statesnegotiateless readily and cooperatively;hence negotiations
fail more often, and disputes fester unresolved (explanationG). (8)States
enshroud foreign and defense policy in tighter secrecy, raising the risk of
Stephen Van Evera teaches international relations in the Political Science Department at the Massachusetts
Institute of Technology.
Thanks to Robert Art, Charles Glaser, and an anonymous reviewer for their comments on this
article. It is distilled from Causesof War,VolumeI: The Structure of Power and the Roots of War (Ithaca,
N.Y.:Cornell UniversityPress, forthcoming1999).
1. In this article"offense dominant" means that conquestis fairly easy; "defense dominant" means
that conquestis very difficult.It is almostnever easierto conquer than to defend, so I use "offense
dominant"broadly, to denotethat offense is easierthan usual,although perhapsnot actuallyeasier
than defense.I use "offense-defense balance" to denote the relativeeaseof aggressionand defense
against aggression. As noted below, this balance is shaped by both military and diplomatic/
political factors.Twomeasuresof the overalloffense-defense balance work wek (1) the probability
that a determinedaggressorcould conquerand subjugatea target statewith comparableresources;
or (2) the resource advantage that an aggressor requires to gain a given chance of conquering a
target state. I use "offense" to refer to strategic offensive action-the taking and holding of
territory-as opposed to tactical offensive action, which involves the attack but not the seizure
and holding of territory.
International Security, Vol. 22, No. 4 (Spring1998),pp. 5-43
0 1998 by the President and Fellowsof HarvardCollege and the Massachusetts Instituteof Technology.
5
International Security 22:4 I 6
miscalculation and diplomatic blunder (explanation HI. (9) Arms racing is
faster and harder to control, raising the risk of preventive wars and wars of
false optimism (explanation I). (10) Offense dominance is self-feeding. As
conquest grows easier, states adopt policies (e.g., more offensive military
doctrines) that make conquest still easier. This magnifies effects 1-9 (explanation
J).
The perception of offense dominance raises these same ten dangers, even
without the reality. If states think the offense is strong, they will act as if it
were. Thus offense-defense theory has two parallel variants, real and perceptual.
These variants are considered together here.
How does this theory perform in tests? Three single case-study tests are
performed below. They corroborate offense-defense thee$ and indicate that
it has large theoretical importance: that is, shifts in the offense-defense balance-real
or perceived-have a large effect on the risk of war. The actual
offense-defensebalance has marked effects; the effectsof the perceived offensedefense
balance are even larger.
What causes offense and defense dominance? Military technology and doctrine,
geography, national social structure, and diplomatic arrangements (specifically,
defensive alliances and balancing behavior by offshore powers) all
matter. The net offense-defense balance is an aggregate of these military,
geographic, social, and diplomatic factors.
How can offensedominance be controlled?Defensivemilitary doctrines and
defensive alliance-making offer good solutions, although there is some tension
between them: offensive forces can be needed to defend allies. Offense dominance
is more often imagined than real, however. Thus the more urgent
question is: How can illusions of offense dominance be controlled? Answers
are elusive because the roots of these illusions are obscure.
On balance, how does offense-defensetheory measure up? It has the attributes
of good theory. In addition to having theoretical importance, offensedefense
theory has wide explanatory range and prescriptive richness. It explains
an array of important war causes (opportunistic expansionism, defensive
expansionism,fierce resistance to others' expansion,first-strike advantage,
2. I use "offensedefensetheory" to label the hypothesisthat war is more likely when conquest is
easy, plus explanatory hypotheses that define how this causation operates. The classic work on
the topic is Robert Jervis,"Cooperationunder the SecurityDilemma," World Politics, Vol. 30, No. 2
(January1978),pp. 167-214 at 169. An overview is Sean M. Lynn-Jones,"Offense-DefenseTheory
and Its Critics,"Security Sfudies, Vol. 4, No. 4 (Summer1995),pp. 660-491.The theory I frame here
subsumesand elaborates on Jervis's theory.
Offense, Defense, and the Causes of War I 7
windows of opportunityand vulnerability,faits accomplis, negotiation failure,
secrecy, arms races, and offense dominance itself) that were once thought to
be independent.In so doing, offense-defensetheory explains the dangers that
these war causesproduce and the wars they cause.This simplifiesthe problem
of power and war: a number of disparate dangers are fed by a single taproot.
Moreover, both the reality and the perception of easy conquest can be shaped
by human action;hence offense-defensetheory offers prescriptionsfor controlling
the dangers it frames.
The next section outlines offense-defensetheory’s ten explanations for war.
The following sectionidentifiescauses of offense and defense dominance. The
fourth section frames predictions that can be inferred from offense-defense
theory, and offers three case studies as tests of the theory: Europe since 1789,
ancient China during the Springand Autumn and WarringStatesperiods, and
the United States since 1789. The final section assesses the general quality of
offense-defensetheory.
Hypotheses on the Effects of Offense Dominance
A host of dangers arise when conquest is easy. Some are obvious and some
more subtle, some are direct and some indirect.Together they make war very
likely when the offense dominates.
A: OPPORTUNISTIC EXPANSIONISM
When conquest is hard, states are dissuaded from aggression by the fear that
victory will prove costly or unattainable. When conquest is easy, aggressionis
more alluring: it costs less to attempt and succeeds more often? Aggressors
can also move with less fear of reprisal because they win their wars more
decisively,leaving their victims lessable to retaliate later. Thus even aggressor
states are deterred from attacking if the defense is strong, and even quite
benign powers are tempted to attack if the offense is strong.
B AND C: DEFENSIVE EXPANSIONISM AND FIERCE RESISTANCE TO EXPANSION
When conquest is hard, states are blessed with secure borders; hence they are
less aggressiveand more willing to accept the status quo. They have less need
3. Suggesting this hypothesisareIvanS. Bloch, TheFuture of War,trans. R.C.Long, pref.W.T. Stead
(New York Doubleday and McClure, 1899), pp. xxx-mi, lxxix; also George H. wester, Offense
and Defense in the International System (New York JohnWdey and Sons, 19m,p. 9. Acorroborating
test is JohnJ. Mearsheimer, Conventional Deterrence (Ithaca, N.Y.:Cornell University Press, 1983).
lnternational Security 22:4 1 8
for wider borders because their current frontiers are already defensible.They
have less urge to intervene in other states’ internal affairs because hostile
governments can do them less harm.
Conversely when conquest is easy, states are more expansionist because
their current borders are less defen~ible.~They covet others’geographic strong
points, strategic depth, and sourcesof critical raw materials. They worry more
when hostile regimes arise nearby because such neighborsare harder to defend
against. These motives drive states to become aggressors and foreign interven
o r ~ . ~States also resist others’ expansion more fiercely when conquest is easy.
Adversaries can parlay smaller gains into larger conquests; hence stronger
steps to prevent gains by others are more appropriate. This attitude makes
disputes more intractable.
The basic problem is that resources are more cumulative when conquest is
easy.The abilityto conquer others and to defend oneself is more elasticto one’s
control over strategic areas and resources. As a result, gains are more additive-states
can parlay small conquests into larger ones-and losses are less
reversible. Hence small losses can spell one’s demise, and small gains can open
the way to hegemonic dominance. States therefore compete harder to control
any assets that confer power, seeking wider spheres for themselves while
fiercely resisting others’ efforts to expand.
This problem is compounded by its malignant effect on states’ expectations
about one another’s conduct. When conquest is hard, states are blessed with
neighbors made benign by their own security and by the high cost of attacking
others. Hence states have less reason to expect attack. This leaves states even
more secure and better able to pursue pacific policies. Conversely, when the
offense dominates, states are cursed with neighbors made aggressive by both
temptation and fear. These neighbors see easy gains from aggression and
danger in standing pat. Plagued with such aggressiveneighbors, all states face
4. As Robert Jervisnotes, ”when the offense has the advantage over the defense, attacking is the
best route to protecting what you have...and it will be hard for any state to maintain its size and
influence without trying to increase them.” Jervis, ”Cooperation under the Security Dilemma,”
p. 211; see also pp. 168-169, 173, 187-199.
5. It also seems possible that states should be more careful to avoid war when conquest is easy,
because war then brings greater risk of total defeat. If so, offense dominance should cause more
cautionthan belligerenceamong states,and should lower the risk of war. Advancingthis argument
isJamesFearon, “TheOffense-Defense Balanceand War since1648,”paper prepared for the annual
meeting of the International Studies Association, Chicago, February 1995, pp. 18-24. Fearon’s
argument seems deductively sound, but history offers very few examples of policymakers who
argued that offense dominance was a reason for caution. This is one of many cases where
deduction and the historical record point in opposite directions.
Ofense, Defense, and the Causes of War I 9
greater risk of attack. This drives them to compete still harder to control
resources and create conditions that provide security.
Thus states become aggressors because their neighbors are aggressors. This
can proceed reciprocally until no state accepts the status quo.
D: MOVING FIRST IS MORE REWARDING
When conquest is easy, the incentiveto strikefirst is larger because a successful
surprise attack provides larger rewards and averts greater dangers. Smaller
shifts in ratios of forces between states create greater shifts in their relative
capacity to conquer and defend territory. (Areversal in the force ratio between
two states from 2 to 1to 1to 2 means little if attackers need a 3 to 1advantage
to conquer; it means everything if an attacker needs only a 1.5to 1advantage.)
Hence a surprise strike that shifts the force ratio in the attacker‘s favor pays it
a greater reward. This expands the danger of preemptive war and makes crises
more explosive. States grow more trigger-happy, launching first strikes to
exploit the advantage of the initiative, and to deny it to an opponent:
Conversely, if the defense dominates, the first-move dividend is small because
little can be done with any material advantage gained by moving first.
Most aggressors can be checked even if they gain the initiative, and defenders
can succeed even if they lose the initiative. Hence preemptive war has less
attraction.
E: WINDOWS ARE LARGER AND MORE DANGEROUS
When conquest is easy, arguments for preventive war carry more eight.^
Smaller shifts in force ratios have larger effects on relative capacity to conquer
or defend territory; hence smaller prospective shifts in force ratios cause
greater hope and alarm. Also, stemming decline by using forceis more feasible
because rising states can be overrun with greater ease.This bolsters arguments
for shutting ”windows of vulnerability” by war. As a result, all international
change is more dangerous. Events that tip the balance of resources in any
direction trigger thoughts of war among states that face relative decline.
Conversely, if the defense dominates, arguments for preventive war lose
forcebecause declining states can more successfullydefend against aggressors
even after their decline, making preventive war unnecessary. States are also
6. The classicdiscussionof these dangersis ThomasC. shelling, Amzs and Influence (New Haven,
Corn.: Yale University Press, 1966), pp. 221-259.
7. For a discussionof the dangers of preventive war, see JackS. Levy, “DecliningPower and the
PreventiveMotivation for War,” World Politics, Vol. 40,No.1 (October 19871, pp. 82-107.
International Security 22:4 I 10
deterred from preventive war by the likelihood that their attack will fail,
defeated by their enemy's strong defenses.
F: FAITS ACCOMPLIS ARE MORE COMMON AND MORE DANGEROUS
When conquest is easy, states adopt more dangerous diplomatic tactics-specifically,
fait accompli tactics-and these tactics are more likely to cause war.
A fait accompli is a halfway step to war. It promises greater chance of
political victory than quiet consultation, but it also raises greater risk of violence?
The acting side moves without warning, facing others with an accomplished
fact. It cannot retreat without losing face, a dilemma that it exploits to
compel the others to concede. But if the others stand firm, a collision is hard
to avoid. Faits accomplis also pose a second danger: because they are planned
in secret, the planning circle is small, raising the risk that flawed policies will
escape scrutiny because critics cannot quarrel with mistaken premises.
Faits accomplis are more common when the offense dominates because the
rewards they promise are more valuable. When security is scarce, winning
disputes grows more important than avoiding war. Leaders care more how
spoils are divided than about avoiding violence, because failure to gain their
share can spell their doom. This leads to gain-maximizing, war-risking diplomatic
strategies-above all, to fait accompli tactics.
Faits accomplis are more dangerous when the offense dominates because a
successfulfait accompli has a greater effect on the distribution of international
power. A sudden resource gain now gives an opponent more capacity to
threaten its neighbors' safety. Hence faits accomplis are more alarming and
evoke a stronger response from others. States faced with a fait accompli will
shoot more quickly because their interests are more badly damaged by it.
G: STATES NEGOTIATE LESS AND REACH FEWER AGREEMENTS
When conquest is easy, states have less faith in agreements because others
break them more often; states bargain harder and concede more grudgingly,
causing more deadlocks; compliance with agreements is harder to venfy; and
8. On fait accompli strategies, see Alexander L. George, "Strategiesfor Crisis Management," in
Alexander L. George, Avoiding War:Problems of Crisis Management (Boulder,Colo.: Westview, 19911,
pp.377-394 at 382383, alsopp. 549-550,553-554. Other discussionsof faits accomplisinclude R.B.
Mowat, DipZornacy and Peace (London:Williams and Norgate, 19351, chap. 10 (on "suddendiplomacy");Richard
Ned Lebow, Between Peace and War: The Nature of International Crisis (Baltimore,
Md.: Johns Hopkins University Press, 19811, pp. 57-97 (on "brinkmanship"); and Thomas C.
Schelling, Strategy of Conflict (New York Oxford University Press, 19631, pp. 22-28 (on games of
"chicken").
Offense, Defense, and the Causes of war I zz
states insist on better verification and compliance.As a result, statesnegotiate
less often and settle fewer disputes; hence more issues remain unsettled and
misperceptions survive that dialogue might dispel.
States break agreements more quickly when the offense dominates because
cheating pays larger rewards. Bad faith and betrayal become the norm. The
secure can afford the luxury of dealing in good faith, but the insecure must
worry more about short-term survival. This drives them toward back-alley
behavior, including deceits and sudden betrayalsof all kinds-diplomatic faits
accomplis, military surpriseattacks,and breaking of other solemn agreements.
Hence compliance with agreements is less expected.
When states do negotiate, they bargain harder and concede less when the
offense dominates. Agreements must be more finely balanced to gain both
sides’ agreement, because a relative gain by either side poses greater risks to
the other’s safety.
Verification of compliancewith agreementsis both more necessaryand more
difficult when the offense dominates. States insist on better verification of the
other’s compliance because smaller violations can have larger security implications;for
example, an opponent might convert a small advantagegained by
cheating on an arms control agreement into a larger offensive threat. At the
same time, verification of compliance is harder because states are more secretive
when securityis scarce (seeexplanationG). As a result, the range of issues
that can be negotiated is narrowed to the few where near-certain verification
is possible despite tight state secrecy.
As a net result, states let more disputes fester when the offense dominates.
H: STATES ARE MORE SECRETIVE
Governmentscloak their foreign and defense policies in greater secrecywhen
conquest is easy. An information advantage confers more rewards, and a
disadvantage raises more dangers: lost secrets could risk a state’s existence.
Thus states compete for information advantage by concealing their foreign
policy strategies and military plans and forces.
Secrecy in turn is a hydra-headed cause of war. It can lead opponents to
underestimate one another‘s capabilities and blunder into a war of false optimism?
It can ease surprise attack by concealingpreparationsfrom the victim.
It opens windows of opportunity and vulnerability by delaying states’ reac-
9. On wars of false optimism, see Geoffrey Blainey, The Causes of War, 3d ed. (NewYork Free
Press, 19881, pp. 35-56.
Figure1.Offense-DefenseTheory
Primehypothesis:Warismorelikelywhenconquestiseasy.
ExplanationA
OffenseIExphationB
dominanceIExplanationC
and/or
perceptionof
offense
dominance
ExplanationD
ExplanationE:
ExplanationF
ExplanationG:
ExplanationH
ExplanationI:
ExplanationJ:
*lntP=Interveningphenomenon.
HypothesisA1
HypothesisB1
HypothesisCi
>opportunisticexpansion(IntP*A)
>Defensiveexpansion(IntPB)
>Fierceresistancetoexpansion
>Movingfirstismorerewarding(In*D)
byotherstates(IntPC)
HypothesisDi
-
>Windowsarelarger,and
moredangerous(In*E)
moredangerous(Id'F)
feweragreements(IntPG)
HypothesisEl
HypothesisFI
>Faitsaccomplisaremore
HypothesisGI
>Statesnegotiatelessandreach
-
HypothesisHi
>Secrecyismorecommonand
moredangerous(IntPHI
Moreintensearmsracing(IntPI)
>Offensegrowsevenstronger(IntPJ)
Hypothesis11
HypothesisJ1
>
x
Q
0
5.
Olcfense,Defense, and the Causes of War I 13
tions to others' military buildups, raising the risk of preventive war. It fosters
policy blundersby narrowing the circleof expertsconsulted on policy, increasing
the risk that flawed policies will survive unexamined. It prevents arms
control agreementsby making compliance harder to verify.
I: STATES ARMS RACE HARDER AND FASTER
Offense dominanceintensifies arms racing, whereas defense dominanceslows
it down." Arms racing in turn raises other dangers. It opens windows of
opportunity and vulnerability as one side or the other races into the lead. It
also fosters false optimism by causing rapid military change that confuses
policymakers' estimatesof relative power. Thus offensedominanceis a remote
cause of the dangers that arms racing produces.
Stateshave sevenincentivesto build largerforceswhen the offenseis strong.
.Resources are more cumulative (see explanations B and C). Wartime gains
and losses matter more: gains provide a greater increase in security, and
losses are less reversible. Therefore the forces that provide these gains and
protect against these losses are also worth more.
0 Self-defenseis more difficultbecause others' forces have more inherentoffensivecapability.Hence
statesrequiremore forcesto offsetothers' deployments.
0 States are more expectant of war. Their neighbors are more aggressive (see
explanation B), so they must be better prepared for attack or invasion.
0 The early phase of war is more decisive. Lacking time to mobilize their
economiesand societiesin the event of war, states maintain larger standing
forces." Thepossibility of quick victoryputs a premium on forces-in-being."
0 States transfer military resources from defense to offense because offense is
more effective (see explanation J). Others then counterbuild because their
neighbors' capabilitiesare more dangerous and so require a larger response.
States also infer aggressive intent from their neighbors' offensive buildups,
leading them to fear attack and to build up in anticipation.
0 States hold military secrets more tightly when the offense dominates (see
explanation H). This causes rational overarming, as states gauge their defense
efforts to worst-case estimates of enemy strength, on grounds that
10. See Jervis,"Cooperationunder the Security Dilemma," pp. 172,188-190.
11. See ibid., pp. 172, 189.
12. GeneralJosephJoffrearguedfor a larger French standingforce in 1913,because"theaffair will
already have been settled"by the time reservistswere mobilized in three to four weeks. David G.
Hemnann, The Arming of Europe and the Making of fheFirst World War (Princeton, N.J.:Princeton
UniversityPress, 1996),p. 193.
International Security 22:4 I 14
underspending is disastrous whereas overspending is merely wasteful. It
also allows national militaries to monopolize defense information more
tightly. Given that militaries are prone to inflate threats, states will overspend
groundlesslywhen militaries have an informationmonopoly that lets
them alone assess the threat. Thus "action-reaction" becomes "action-over-
reaction-overreaction."
0 States reach fewer arms control agreements when the offense dominates,
because agreements of all kinds are fewer (seeexplanationG). Hence states
are less able to limit arms competition through agreement.
If the defense dominates, things are reversed. Statesbuild smaller offensive
forces because offense is less effective, and because other states have less
aggressive aims. States are safe without wider empires; hence offensiveforces
that could provide empires lose utility. The national military therefore grows
defense-heavy.This causesother states to feel safer, which in turn makes them
less aggressive, further lowering all states' insecurity-hence their need for
empire and for offenseup to a point.
States also reduce defensive forces when the defense dominates because
defense is easier and attack seems more remote. Moreover, as their neighbors
buy less offense, they need even less defense because their defense faces less
challenge.
In short, states buy smaller forces in general, and less offense in particular,
when the defense dominates. This leads to still smaller forces and still less
offense. If information were perfect, arms racing would slow to a crawl if the
defense strongly dominated.
J. CONQUEST GROWS STILL EASIER
Offensedominanceis self-reinf~rcing'~for three main reasons. First, statesbuy
relatively more offensive forces when the offense dominates. They prefer the
more successfultype of force, so they buy defensiveforces when the defense
is strong and offensive forces when the offense is strong.14This reinforces the
initial dominanceof the defense or the offense.
13. Making this argumentis Jervis,"Cooperationunder the Security Dilemma,"pp. 188,199,201.
14. Thus Clausewitz explained"If attack were the strongerform [of war], therewould be no case
for using the defensive, since its purpose is only passive.No one would want to do anything but
attack.Defense would be pointless."Carl von Clausewitz,On War,ed. and trans.MichaelHoward
and Peter Paret, intro. by Paret, Howard, and Bernard Brodie, commentaryby Brodie (Princeton,
N.J.: Princeton University Press, 19761, p. 359.
Offense, Defense, and the Causes of War I 15
Second, alliances assume a more offensive ~haracter'~when the offense
dominates because aggressorscan more easily drag their allies into their wars
of aggression.16Insecure states can less afford to see allies destroyed, so they
must support even bellicose allieswho bring war on themselves.Knowing this,
the allies feel freer to get into wars. As a net result, even de jure defensive
alliances operate as defensive-and-offensivealliances. Alliances also assume a
more offensivecharacter if the allies adopt purely offensivemilitarydoctrines.
This hamstrings states that would demand that their allies confine themselves
to defensive preparations in a crisis, given that all preparations are offensive.
Third, status quo states are less able to protect their allies from conquest
when the offense dominates because attackers can overrun defenders before
help can arrive.
Thus offense dominance raises the danger of greater offense dominance.
Once entered, an offense-dominant world is hard to escape.
Military offense dominance has one self-limiting effect: it leads status quo
powers to cooperate more closely against aggressor^.'^ They jump to aid an
aggressor's victims because each knows that its neighbor's demise could lead
more directly to its own undoing. Conversely, when states think that the
defense dominates, they do less to save others from aggression because each
expects it can defend itself alone even if others are overrun. As a result,
aggressorscan more often attack their victims seriatim, which is far easier than
defeating a unified coalition.This countervailingeffect, however, is more than
offset by the several ways that offense dominance feeds itself.
These are the dangers raised by offense dominance. As noted above, these
same ten dangers arise when the offense is weak but governments think it
dominates. They then act as if it dominates, with comparable effects.
Are offensive capabilities always dangerous? The one-sided possession of
offensive capabilities by status quo powers that face aggressors can lower
rather than raise the risk of war under some conditions.Most important, status
quo powers often need offensive capabilities to defend other states against
15. A defensive alliance is conditioned on defensive behavior by the ally; the alliance operates if
the ally is attacked but not if it attacks. A defensive-and-offensivealliance operates in the event
of war regardless of which side started it. The distinction began with Thucydides, who used
"empimachy" to denote defensive alliance, "symmachy" for defensive-and-offensive alliances.
G.E.M.de Ste. Croix, The Origins of the Peloponnesian War (Ithaca, N.Y.:Cornell University Press,
16. Developing this point are Thomas J. Christensen and JackSnyder, "ChainGangs and Passed
Bucks: Predicting Alliance Patterns in Multipolarity," International Organization, Vol. 44, No. 2
(Spring 1990), pp. 137-168.
17. Making this argument is ibid.
1972),pp. 60,72-73,106-108, 184,298-302,328.
International Security 22:4 I 16
aggressors (e.g., as France required some offensive capability to defend
Czechoslovakiaand Poland from Germany in 1938-39). Offensive capabilities
in the hands of status quo powers also may provide more deterrence than
provocation if the aggressor state knows that it provoked the status quo
power’s hostility, if the aggressor knows that the status quo power has no
bedrock aggressive intentions, and if the aggressor cannot remove the status
quo power’s offensive threat by force. These conditions are not unknown but
they are rare. Hence offensive capabilities usually create more dangers than
they dampen.
Causes of Offense and Defense Dominance
The feasibility of conquest is shaped by military factors, geographic factors,
domestic socialand political factors, and the nature of diplomacy. Discussions
of the offense-defensebalance often focus on military technology,but technology
is only one part of the picture.”
MILITARY FACTORS
Military technology,doctrine,and force posture and deploymentsall affect the
military offense-defensebalance.” Militarytechnology can favor the aggressor
or the defender. In past centuries, strong fortification techniquesbolstered the
defense, and strong methods of siege warfare strengthened the offense. Technologies
that favored mass infantry warfare (e.g., cheap iron, allowing mass
production of infantry weapons)strengthened the offense because large mass
armies could bypass fortifications more easily, and because mass armies fostered
more egalitarianpolitiesthat could raise loyal popular armiesthat would
not melt away when sent on imperial expeditions.Technologies that favored
chariot or cavalry warfare (e.g., the stirrup) strengthenedthe defense,because
cavalry warfare required smaller forces2’ that were more easily stopped by
18. For a discussion of the causes of offense and defense dominance, see Jervis, ”Cooperation
under the Security Dilemma,’’ pp. 176,1941199.
19. Several measures of the military offense-defensebalance could be adopted, such as: (1)the
probability that an offensiveforcecan overcome a defensive force of equal cost; (2) the relative
cost that attackersand defendersmust pay for forces that offset incremental improvementsby the
other; or (3)the loss ratio when an offensiveforceattacks a defensiveforce of equal cost. All three
measures (and more are possible) capture the concept of relative military difficulty of conquest
and defense. For a list of possible measures, see Charles L. Glaser and Chaim Kaufmann, “What
Is OffemDefense Balance and How Can We Measure It?,” Internafional Security, Vol. 22, No. 4
(Spring 19981, pp. 44-82.
20. Cavalry warfare was capital intensive; hence it was usually waged by small forces of taxsupported
specialists-knights in shining (and expensive) armor on expensive horses. M a n q
warfare is more manpower intensive, and is usually waged by larger, less capihbzed armies.
Ofiense, Defense, and the Causes of War I 17
fortifications,and fostered hierarchic societiesthat could not raise armies that
would remain loyalif sent on questsfor empire?’ In modem times, technology
that gave defenders more lethal firepower (e.g., the machine gun) or greater
mobility (e.g.,the railroad)strengthened the defense.When these technologies
were neutralized by still newer technologies (motorized armor), the offense
grew stronger.
Thus when fortresses and cavalry dominated in the late Middle Ages, the
defense held the advantage. Cannons then made fortificationsvulnerable and
restored the strength of the offense. In the seventeenth and eighteenth centuries
new fortification techniques strengthened the defense. The mercenary
armies of the age also remained tightly tied to logistical tails that kept them
close to home: one historian writes that an eighteenth-centuryarmy “was like
a diver in the sea, its movements strictlylimited and tied by the long, slender
communicating tube which gave it life.”” Then revolutionary France’s mass
armies strengthenedthe offense because they had greater mobility. Their size
let them sweep past border forts without leaving the bulk of their manpower
behind for siege duty, and their more loyal troops could be trusted to forage
without deserting, so they needed less logistical support. After the conservative
restoration in France, Europe abandoned the mass army because it required,
and fostered, popular government. This restored the power of the
defense,which then waned somewhat as Europe democratizedand large mass
armies reappeared in the mid-nineteenth century?3
The combined effects of lethal small arms (accurate fast-firing rifles and
machine guns),barbed wire, entrenchments,and railroads gave the defensean
enormous advantage during World War I. The first three-lethal small arms,
barbed wire, and trenches-gave defenders a large advantage at any point of
attack. The fourth-railroads-let defenders reinforce points of attack faster
than invaders could, because invaders could not use the defenders’ railroads
(given that railroad gauges differed across states,and defendersdestroyed rail
lines as they retreated) while the defenders had full use of their own lines.
During 191945 the power of the offense was restored by motorizedarmorand
an offensive doctrine-blitzkrieg-for its employment; this overrode machine
21. On the effects of the stimtp on warfare and society in the Middle Ages, see Lynn White, Jr.,
Medimuf Technologyand Social Change (New York:Oxford University Press, 19641,pp. 1-38.On the
general effect of military technology on social stratification,see Stanislav Andreski, Military Orgunimtion
and Society (Berkeley:University of California Press, 19711,pp. 20-74.
22.Harold Temperley,quoted in Blainey, Causes of War, p. 188.
23. Large armies aid the offense only up to a point, however.Once armies grow so big that they
can cover an entire frontier (as on the western front in World War I), their size aids the defense
because offensive outflankingmaneuvers against them become impossible.
Znternational Security 22:4 1 28
guns, trenches, and barbed wire. Then after 1945 thermonuclear weapons
restored the power of the defenscthis time giving it an overwhelming ad-
vantage.24
Technology and doctrine combined to define these tides of offense and
defense. Sometimes technology overrode doctrine, as in 1914-18 and in 1945-
91 (when the superpowers’ militaries embraced offensive doctrines but could
not find offensive counters to the nuclear revolution). Sometimes doctrine
shaped technology as in 193945, when blitzkrieg doctrine fashioned armor
technology into an offensive instrument.
States shape the military offense-defense balance by their military posture
and force deployments. Thus Stalin eased attack for both himself and Hitler
during 1939-41 by moving most of the Red Army out of strong defensive
positions on the Stalin Line and forward into newly seized territories in
Poland, Bessarabia, Finland, and the Balticstates.25This left Soviet forcesbetter
positioned to attackGermany and far easier for Germany to attack, as the early
successof Hitler’s 1941 invasion revealed.The U.S. eased offense for both itself
and Japan in 1941 when it deployed its fleet forward to Pearl Harbor and
bombers forward to the Philippines?6Egypt eased Israel’s assault by its chaotic
forward deployment of troops into poorly prepared Sinaipositionsin the crisis
before the 1967 war.27
States also can change the offense-defense balance through their wartime
military operations. Aggressive operations can corrode key enemy defenses,
and reckless operations can expose one’s own defenses. Thus the dangers of
offense dominance can be conjured up by unthinking wartime policymakers.
For example, General Douglas MacArthur’s reckless rush to the Yalu River in
1950 created an offensive threat to China’s core territory and, by exposing
badly deployed U.S. forces to attack, eased a Chinese offensiveF8
24. Jack Levy provides synoptic history of the military offense-defense balance in ”The Offensive/Defensive
Balance of Military Technology: A Theoretical and Historical Analysis,” International
Studies Quarterly, Vol. 28, No. 2 (June 19841,pp.219-238 at 23C-234. Otherdiscussionsinclude
Quester, Offense and Defense in the International System; and Andreski, Military Orgunktion and
Society, pp. 75-78. A detailed history is needed.
25. Peter Calvocoressi and Guy Wmt, Total War: The Story of World War II (New York Pantheon
Books, 19721, p. 168.
26. JonathanG.Utley, Going to Warwith Japan,1937-1941 (Knoxville:Universityof TennesseePress,
19851, pp. 84, 163.
27. Donald Neff, Warriorsfor lerusalem:The Six Days That Changed the Middle East (NewYork Simon
and Schuster, 1984), pp. 141, 168.
28. Likewise, during the Cold War some worried that NATO might inadvertently threaten the
Soviet Union’s strategic nuclear deterrent in its effort to defend NATO’s Atlantic sea-lanes during
Offense, Defense, and the Causes of War I 19
GEOGRAPHY
Conquestis harder when geography insulates states from invasion or strangulation.
Hence conquest is hindered when national borders coincide with
oceans, lakes, mountains,wide rivers, dense jungles, trackless deserts,or other
natural barriers that impede offensive movement or give defenders natural
strong points. Human-made obstacles along borders, such as urban sprawl,
can also serve as barriers to armored invasion. Conquestis hindered if foes are
separated by wide buffer regions (third states or demilitarized zones) that
neither side can enter in peacetime. Conquest is hindered when national
territories are mountainous or heavily forested, and when populations live
mainly in rural settings, easing guerrilla resistance to invaders. Conquest is
hindered when states are large and their criticalwar resources or industries lie
far in their interior, where they cannot be quickly overrun. Conquest is hindered
when statesare invulnerableto economicstrangulation.Hence conquest
is hindered when states are self-sufficientin supplies of water, energy, food,
and critical raw materials, or when their trade routes cannot be severed by
land or sea blockade.
The geography of Western Europe, with its mountain ranges and ocean
moats, is less favorableto conquestthan the exposed plains of Eastern Europe
or the open terrain of the Middle East. Israel’s geography is especially unfortunate:
physically small, its frontiers have few obstacles and much of its
industry and population lie on exposed frontiers. Israeli territory is not conduciveto
guerrillaresistance,and its economyis import dependent.Germany’s
borders are better but still relatively poor: its eastern frontier is open; its
economy is import dependent; and its trade routes are vulnerable. Britain,
France, and Italy have formidable frontier barriers that make them relatively
defensible.The United States’vast size, ocean-moatfrontiers,and independent
economy bless it with very defensible geography.
SOCIAL AND POLITICAL ORDER
Popular regimes are generally better at both conquest and self-defense than
are unpopular regimes, but these effects do not cancel out. On net, conquest
is probably harder among popular than unpopular regimes today, but in past
centuries the reverse was likely true.
an East-Westconventionalwar. Bany R. Posen,Inadvertent Escalation: Conventional War and Nuclear
Risks (Ithaca, N.Y.:Cornell University Press, 1991), pp. 129-158. O n a related danger, see ibid.,
pp. 28-67.
International Security 22:4 I 20
Popular governments can better raise larger, more loyal armies that can
bypass others' border forts and can operate far from home with less logistical
support. This gives popular regimes greater offensivepower. Popular regimes
can better organize their citizens for guerrilla resistance, making them harder
to conquer. Citizen-defense guerrilla strategies are viable for Switzerland or
China, but not for Guatemala or ancient Sparta, because these unpopular
governments cannot arm their people without risking revolution. The citizens
of unpopular oligarchies may actively assist advancing invaders. This gives
attackers more penetrating power and makes early losses less reversible. Thus
Sparta feared an invading army might grow if it entered Spartan territory,
because Spartan slaves and dissident tribes would desert to the enemy.29
Unpopular regimesare more vulnerableto subversionor revolutioninspired
from abroad. Subversion is a form of offense, and it affects international
relations in the same way as do offensive military capabilities. Frail regimes
are more frightened of unfriendly neighbors, making them more determined
to impose congenialregimes on neighboring states. The French revolutionary
regime and the oligarchic Austrian regime worried that the other side might
subvert them in 1792, causing both sides to become more aggre~sive.~'After
the Russian Revolution similar fears fueled Soviet-Western conflict, as each
side feared subversion by the other.
On balance, is conquest easier in a world of popular or unpopular regimes?
Popularity of regimes probably aided offense before roughly 1800 and has
aided defense since then. The reversal stems from the appearance of cheap,
mass-produced weapons useful for guerrilla war-assault rifles and machine
guns, light mortars, and mines. The weapons of early times (swordand shield,
pike and harquebus, heavy slow-firingmuskets, etc.)were poorly adapted for
guerrilla resistance.Guerrillawarfare has burgeoned since 1800partly because
the mass production of cheap small arms has tipped the balance toward
guerrillas, allowing the hit-and-run harassment that characterizes guerrilla
operations.The defensive power of popular regimeshas risen in stepwith this
increase in guerrilla warfare.
29. De Ste. Croix,Origins of the Peloponnesian War,pp. 89-94. Likewise, Hannibalhoped to defeat
Rome by recruiting dissident tribes as he penetrated the Italian peninsula.See RM.Errington,
Dawn of Empire: Rome's Rise to World Pmer (Ithaca, N.Y.:CornellUniversityh),pp. 62-64.
30. StephenM. Walt, Revolution and War(Ithaca,N.Y.:Cornell Universitym,I%), pp. 123-124;
and T.C.W. Blanning,TheOriginsof theFrenchReuolutionay Wars(London:Longman,19849,pp. 76,
8586,9!9-101,111.
Offense, Defense, and the Causes of War I22
DIPLOMATIC FACTORS
Three types of diplomatic arrangements strengthen the defense: collective
securitysystems,defensivealliances,and balancingbehaviorby neutral states.
All three impede conquest by adding allies to the defending side.
States in a collective security system (e.g., the League of Nations) promise
mutual aid against aggression by any system member. Such aggressors will
face large defendingcoalitions if the system operates3'
Statesin a defensive alliancepromisemutual aid against outsideaggressors,
leaving such aggressors outnumbered by resisting opponents. Thus during
1879-87 Bismarck wove a network of defensive alliances that discouraged
aggression and helped preserve peace throughout central and eastern Europe.
Collective securitysystemsand defensive alliancesdiffer only in the kind of
aggressor they target (systemmembersversus outsideaggressors).Both kinds
of aggressors could be targeted at once, and a hybrid system that did this
would offer defenders the most protection.
Neutral states act as balancers when they join the weaker of two competing
coalitions to restore balance between them. Aggression is self-limiting when
neutralsbalance because aggressorsgeneratemore opposition as they expand.
Britain and the United States traditionallyplayed balancers to Europe, providing
a counterweight to potential continentalhegemons.
Balancing behavior is more selective than defensive alliance. Balancers balance
to avert regional hegemony; hence pure balancers opposeexpansiononly
by potential regional hegemons. Smaller states are left free to aggress. But
balancing does contain hegemons and leaves their potential victims more
secure.Conversely,if statesbandwagon-join the strongercoalitionagainst the
weaker one-conquest is easier because aggressors win more allies as they
seize more
Diplomatic arrangements have had a large influence on the offense-defense
balance in modern Europe, and shifts in diplomatic arrangementshave pro-
31. An introductionto collective securityis lnisL. Claude, Jr., Swords into Plowshares: TheProblems
and Progress ofIntemafional Organizations, 4th ed. (New York Random House, 1971),pp. 411433.
Arecent advocacyof collective securityis CharlesA. Kupchanand Clifford A. Kupchan,"Concerts,
CollectiveSecurity, and the Future of Europe," International Security, Vol. 16, No. 1(Summer1991),
32. On balancing,bandwagoning,and other theories of alliances, see StephenM. Walt, The Origins
of Alliances (Ithaca, N.Y.: Cornell University Press, 1987).Historians have often suggested that a
"breakdown in the balance of power/' caused war. They usually mean (and should recast their
claim to say) that states failed to engage in balancing behavior, which made aggression easier,
causing war. War occurs not when the balance of power breaks down, but when balancers fail to
balance, leaving aggressorsunchecked, as in the late 1930s.
pp. 114-163.
International Security 22:4 I 22
duced large shifts in the overall offense-defense balance. Collective security
was never effective, but defensive alliances came and went, erecting barriers
to conquestwhen they appeared.Balancingbehaviorrose and fell as the power
and activism of the two traditional offshore balancers, Britain and the United
States, waxed and waned. When the United Statesand/or Britain were strong
and willing to intervene against aspiring continental hegemons, conquest on
the continent was difficult. To succeed, a hegemon had to defeat both its
continental victims and the offshore power. But when Britain and the United
Stateswere weak or isolationist, continentalpowers could expand against less
resistance, leaving all states less secure.
Tests of Offense-Defense The0y
What predictions can be inferred from offense-defense theory? How much
history does offense-defensetheory explain?
PREDICTIONS AND TESTS
Offense-defensetheory‘s predictions can be grouped in two broad types, prime
predictions and explanatory predictions.The theory’sprime predictions derive
from its prime hypothesis (“War is more likely when conquestis easy“;or, for
the theory‘s perceptualvariant, “Waris more likelywhen statesthink conquest
is easy”).Tests of these predictions shed light on whether offense dominance
(or perceptions of offense dominance)causes war.
Offense-defensetheory‘sexplanatorypredictionsderivefrom the hypotheses
that compriseits ten explanations.Tests of these predictionsshed light on both
whether and how offense dominance (or perceptions of offense dominance)
causes war.
PRIME PREDICTIONS. Three prime predictions of offense-defensetheory are
tested here.
1. War will be more common in periods when conquest is easy or is believed
2. States that have or believe they have large offensiveopportunitiesor defen-
3. A state will initiate and fight more wars in periods when it has, or thinks
easy, less common when conquest is difficult or is believed difficult.
sive vulnerabilitieswill initiate and fight more wars than other states.
that it has, larger offensive opportunitiesand defensive capabilities.
These predictions are tested below in three case studies: Europe since 1789
(treated as a single regional case study), ancient China during the Spring and
Offense, Defense, and the Causes of War I 23
Autumn and Warring Stateseras, and the United States since 1789. I selected
these cases because the offense-defense balance (or perceptions of it) varies
sharply across time in all three, creating a good setting for ”multiple withincase
comparisons” tests that contrast different periods in the same case; because
the United States is very secure relative to other countries, creating a
good setting for a ”comparison to typical values” tests that contrasts U.S.
conduct with the conduct of average states;33and because two of these cases
are well recorded (Europesince 1789 and the United States since 1789).
The case of Europe since 1789 allows tests of prime predictions 1 and 2.34
We can make crude indices of the offense-defensebalances (actual and perceived)
for Europe over the past two centuries, and match them with the
incidenceof war (seeTable 1).Offense-defensetheory predicts more war when
conquest is easy or is believed easy. We can also estimate the offensive opportunitiesand
defensivevulnerabilitiesof individualpowers-for example,since
1789 Prussia/Germany has been more vulnerable and has had more offensive
opportunity than Spain, Italy, Britain, or the United States-and can match
these estimates with states’ rates of war involvement and war initiation.
Offense-defensetheory predicts that states with more defensive vulnerability
and offensive opportunity will be more warlike.
The ancient China case allows a test of prime prediction 1. The offensedefense
balance shifted markedly toward the offense as China’s Spring and
Autumn and Warring States periods evolved. Offense-defensetheory predicts
a parallel rise in the incidence of warfare during these periods.
The U.S. case allows testing of prime predictions 2 and 3. The United States
is less vulnerableto foreignmilitarythreats than are other states;hence offensedefense
theory predicts that it should start fewer wars and be involved in fewer
wars than other states. Americans have also felt more vulnerable to foreign
military threats in some eras than in others. The U.S. propensity for war involvement
and war initiation should co-vary with this sense of vulnerability.
EXPLANATORY PREDICTIONS. Offense-defensetheory posits that offensedominance
leads to war through the war-causing action of its ten intervening
phenomena A-J: opportunistic expansionism, defensive expansionism, fierce
33. I say more about the logic of within-case comparisonsand comparison to typical values tests
in Van Evera, GuidetoMethodsfor Students ofPolitical Science (Ithaca,N.Y.:CornellUniversity Press,
1997),pp. 58-63. O n case selection criteria, see pp. 77-88.
34. In principle, prime prediction 3 could also be tested with this case. This, however, would
requiretracing and describing trends ineach state‘ssenseof vulnerability over tim- large task
that would fill many pages.
Table1.TheOffense-DefenseBalanceamongGreatPowers,17~resmt.
MllltaryDiplomaticInaggregateInaggregateAmount
realitiesrealitiesmilitaryandmilitaryandofwarfare
MilitarywereDiplomaticwerediplomaticdiplomaticamong
realitiesthoughtrealitiesthoughtrealitiesrealitiesweregreat
Erafavoredtofavorfavoredtofavorfavoredthoughttofavorpowers
Pre-1792Defs.Defs.Med.Med.Med.Med.Medium
1792-1815Aggrs.Aggrs.Med.Aggrs.Aggrs.Aggrs.***High
181656Defs.Defs.Defs.Defs.Defs.Defs.Low
1856-71Med.Med.Aggrs.Aggrs.Aggrs.Aggrs.Medium
1871-90Defs.Med.Defs.Defs.Defs.Defs.***Low
1890-1918Defs.Aggrs.Aggrs.Aggrs.Defs.Aggrs.High
1919-45Aggrs.Mixed*Aggrs.Aggrs.**Aggrs.Aggrs.****High
1945-1990sDefs.Med.Defs.Defs.Defs.Defs.***Low
Aggrs.:Thefactorfavorsaggressors.
Defs.:Thefactorfavorsdefenders.
Med.:Amediumvalue:thingsaresomewhereinbetween,cutbothways.
Mixed:Somenationalelitessawdefensedominance,somesawoffensedominance.
Theperceptionsentriesareanaverageoftheperceptionsofthegreatpowerelites.Insomecases,theperceptionsofthese
elitesvariedsharplyacrossstates,forexample,perceptionsofmilitaryrealitiesinthe1930s.
Thingsvariedacrossstates.TheGermaneliterecognizedthemilitarypoweroftheoffensiveinthelate1930s;theelitesof
othergreatpowersthoughtthedefensewasdominant.
**Thingsvariedacrossstates.TheGermanelite(aboveallHitler)exaggeratedtheconsiderableactualdiplomaticweaknessof
thedefense;theelitesofothergreatpowersrecognizedthisweaknessbutdidnotoverstateit.Thesebeliefsaveragetoa
perceptionofsubstantialdiplomaticoffensedominance.
***Elitesexaggeratedthestrengthoftheoffenseduring1792-1815,1871-90,and1945-1990%butnotbyenoughtogivethe
realitiesandperceptionsoftheoffensedefensebalancedifferentscores.
****Whenweaggregateperceptionsoftheoffense-defensebalance,theerrorsofGermanyandtheotherpowerscanceleach
otherout.Germany'sexaggerationofthediplomaticpoweroftheoffenseoffsetsotherpowers'exaggerationofthemilitary
powerofthedefense,leavinganaggregateperceptionfairlyclosetotheoffense-dominantreality.
h,
h,
;fi
h,
bFa
Offense,Defense, and the Causes of War I 25
resistance to others’ expansion, first-strike advantages, windows of opportunity
and vulnerability, faits accomplisand belligerent reactionsto them, reluctance
to solve conflicts through negotiation, policies of secrecy, intense arms
racing, and policies that ease conquest, such as offensiveforce postures and
offensive alliances. If offense-defense theory is valid, these intervening phenomena
should correlate with the real and perceived offense-defensebalance.
Two explanatory predictions can be inferred.
1. Phenomena A-J will be more abundant in eras of real or perceived offense
dominance: the ten phenomena should increase as offense strengthens and
diminish as offense weakens.
2. States that have or believe they have largeoffensiveopportunitiesor defensive
vulnerabilities will more strongly embrace policies that embody phenomena
A-J.35
Two of the case studies presented here shed light on these explanatory
predictions.The case of Europe allows a partial test of both. We can code only
two of offense-defense theory’s ten intervening phenomena (IntPs A and B,
opportunistic and defensive expansionism) for the whole period. We have
fragmentary data for values on the other eight intervening variables. Hence
the case lets us test explanationsA and B fairlycompletelyand offersscattered
evidence on explanations C-J. To test explanationsA and B, we ask if expansionism
correlates over time with periods of real or perceived offense dominance,
and if statesthat were (or believed they were) less secureand more able
to aggress were more expansionist.
The case of the United States since 1789 allows a more complete, if rather
weak, test of explanatory prediction 2.
TEST 1: EUROPE 1789-1990s
A composite measure of the offense-defensebalance in Europe since 1789 can
be .fashioned by blending the histories of Europe’s military and diplomatic
35. Explanatory predictions 1 and 2 are inferred from the “leftside” of offensedefense theov, that
is, from hypotheses AI-J1, which frame the claim that offense dominance causes intervening
phenomena A-J (see Figure 1). Predictions could also be inferred from hypotheses A2-J2,which
comprise the ”right side” of the theory, and frame the claim that intervening phenomena A-J cause
war. For example, we could infer that (6) warfare will be more common in eras and regions where
phenomena A-J are more prevalent, and (7)states that embrace policies that embody phenomena
A-J will be involved in more wars and will initiate more wars than other states. I leave “right
side” hypotheses untested here because the effects of phenomena A-J are less debated than their
causes. Most agree that they cause trouble.
International Security 22:4 I 26
offense-defense balances, as outlined above.36 In sum, the offense-defense
balance went through six phases comprising three updown oscillations after
1789.Conquestwas never easy in an absolute sense during these twocenturies.
Conquest was, however, markedly easier during 1792-1815, 1856-71, and
1930s-1945 than it was during 1815-56,1871-192Os, and 1945-1990s.
Elite perceptions of the offense-defensebalance track these oscillations quite
closely,but not exactly. Elites chronicallyexaggerated the power of the offense,
but did so far more in some periods than in others. Most important, they
greatly exaggerated the power of the offense during 1890-1918: elites then
wrongly thought conquest was very easy when in fact it was very hard. Thus
the pattern of reality and perception run roughly parallel, with the major
exception of 1890-1 918.
Tides of war and peace correlate loosely with the offense-defense balance
during this period, and tightly with the perceived offense-defense balance.
Expansionism and war were more common when conquest was easy than
when it was difficult,and were far more common when conquestwas believed
easy than when it was believed difficult. Moreover, states that believed they
faced large offensive opportunities and defensive vulnerabilities (especially
Prussia/Germany) were the largest troublemakers. They were more expansionist,
they were involved in more wars, and they started more wars than
other states.
1792-1815. During 1792-1815 the offense was fairly strong militarily, as a
result of France’s adoption of the popular mass army (enabled by the popularity
of the French revolutionary g~vernment).~~Moreover, European elites
widely exaggeratedone another’s vulnerability to conquest:at the outset of the
War of 1792all three belligerents (France, Austria, and Prussia) thought their
36. My composite index represents my own “author‘s estimates” based on sources provided
throughout this article.I measured the actualand perceived Europe-wideoffense-defensebalances
by asking:(1) Did military technology,force posture,and doctrinefavorthe offense or the defense?
Did elites and publicsbelieve these factorsfavored the offense or the defense?(2) Did geography
and the domestic social and political order of states favor the offense or the defense? Did elites
and publics believe they favored the offense or defense? (3) How numerous and powerful were
balancer states, and how strongly did they balance? Did elites believe that other states would
balance or bandwagon? (4)Did defensive alliances form, and did they operate effectively? Did
elites believe that they operated effectively? I gave these factors the same rough relative weight
they receive in standard historical accounts.
37. A discussionof the military offense-defense balance in this era is Quester, Offense and Defense
in the International System, pp. 66-72.
Offense, Defense, and the Causes of War 1 27
opponents were on the verge of collapse and could be quickly crushed?8
Defense-enhancing diplomacy was sluggish: Britain, Europe's traditional balancer,
stood by indifferently during the crisis that produced the War of 1792,
issuing a formal declaration of ne~trality.~~Moreover, French leaders underestimated
the power of defense-enhancing diplomacy because they widely believed
that other states would bandwagon with threats instead of balancing
against them.40In short, military factors helped the offense, and this help was
further exaggerated; political factors did little to help bolster defenders, and
this help was underestimated.
1815-56. After 1815both arms and diplomacy favored defenders, as outlined
above. Mass armies disappeared:' British economic power grew, and Britain
remained active on the continent as a balancer. Continental powers expected
Britain to balance and believed British strength could not be overridden.
This defense-dominant arrangement lasted until midcentury.It began weakening
before the Crimean War (1853-56). When war in Crimea broke out,
military factors still favored defenders, but elites underestimated the power of
the defense:Britain and France launched their 1854Crimean offensive in false
expectation of quick and easy victory!2 In general, diplomatic factors favored
the defense (Britain still balanced actively), but during the prewar crisis in
38. Blanning, Origins of the French Revolutionary Wars, p. 116. Austrian and Prussian leaders were
assured that revolutionary France could be quickly smashed. Ibid., p. 114. One Prussian leader
advised his officers: "Do not buy too many horses, the comedy will not last long. The army of
lawyerswill be annihilatedin Belgium and we shallbe homeby autumn." lbid.,p. 116.Meanwhile,
French revolutionarieswrongly expected a pro-French revolutionary uprising of the oppressed
peoples of feudal Europe. Ibid., p. 136; R.R. Palmer, World of the French Revolution (New York
Harper and Row, 19711, p. 95; and George Rud6, Revolutionary Europe, 1783-1825 (Glasgow:Fontana/Collins,
1964), p. 209.
39. Blanning, Origins of the French Revolutionary Wars, pp. 131-135.
40. As Steven Ross notes, French expansionists thought they could intimidate Europe into coexisting
with an expanded French empire in the 1790s: "By inflictingrapid and decisive defeats
upon one or more members of the coalition, the [French]directorshoped to rupture allied unity
and force individual members to seek a separate peace." Steven T. Ross, European Diplomatic
History, 1789-1825 (Garden City, N.Y.: Anchor Doubleday, 19691, p. 186.
Later Napoleon thought he could compel Britain to make peace by establishingFrench continental
dominion, proclaiming after the Peace of Amiens, "With Europe in its present state, England
cannot reasonably make war on us unaided." Geoffrey Bruun, Europe and the French Imperium,
2799-1814 (New York Harper and Row, 1938), p. 118. See also Blanning, Origins of the French
Revolutionary Wars, p. 109.
41. On the post-1815 restoration of pre-Napoleonicwarfare, see wester, offense and Defense in the
International System, pp. 73-74; and Michael Howard, War in European History (London: Oxford
University Press, 1976),pp. 94-95.
42. Richard Smoke, War:Controlling Escalation (Cambridge,Mass.: Harvard UniversityPress, 19771,
p. 191.
International Security 22:4 I 28
1853, diplomacy favored the offensebecause Britain and France blundered by
giving Turkey unconditional backing that amounted to an offensive alliance.
This encouraged the Turkish aggressionsthat sparked the war?3
1856-71. After the Crimean War the offense-defense balance shifted further
toward the offense. Changes in the military realm cut both ways. Mass armies
were appearing (bolstering the offense), but small arms were growing more
lethal and railroads were expanding (bolstering the defense).In the diplomatic
realm, however, the power of defenders fell dramatically because defenseenhancing
diplomacy largely broke down. Most important, Britain entered
an isolationist phase that lasted into the 1870s, and Russia lost interest in
maintaining the balance among the western powers.44As a result, diplomatic
obstacles to continental conquest largely disappeared, giving continental aggressors
a fairly open field. This diplomatic change gave France and Sardinia,
and then Prussia, a yawning offensive opportunity, which they exploited by
launching a seriesof wars of opportunistic expansion-in 1859,1864,1866,and
1870.But defense-enhancingdiplomacy had not disappeared completely, and
it helped keep these wars short and limited.
In 1859British and Russian neutrality gave France and Sardinia a free hand,
which they used to seize Lombardy from Austria.45In 1864 British, Russian,
and French neutrality gave Prussia and Austria a free hand, which they used
to seize Schleswig-Holstein from Denmark.& In 1866 British, French, and
Russian neutrality gave Prussia carte blanche against Austria, which Prussia
used to smash Austria and consolidate its control of North germ an^?^ Even
after war broke out, major fighting proceeded for weeks before any outside
43. lbid., pp. 167, 179-181, 185; Richard Smoke, “The Crimean War,” in George, Avoiding War,
pp. 36-61 at 48-49, 52. The motives of the powers also illustrate offense-defensedynamics.The
main belligerents (Britain, France, Russia, and Turkey) were impelled in part by security concerns
that would have been allayed had they believed the defensemore dominant.Smoke, War,pp. 149,
155, 158-159,162, 190.
44. The harsh Crimean War settlement Britain imposed on Russia turned it into a non-status quo
power. Overthrowingthat settlementbecame Russia’schief aim in Europeandiplomacy, superseding
its interest in preserving order to the west. M.S. Anderson, 7% EasfernQuestion, 1774-1923
(London:Macmillan, 19661, pp. 144-146.
45. A.J.E Taylor, The Sfruggle for Mastery in Europe 2848-1918 (London:Oxford University Press,
1970, pp. 108, 110.
46. Ibid., pp. 146-154. Britain would have backed Denmark had it found a continental ally but
none was available.Ibid., pp. 146-148.
47. Smoke, War,pp. 85-92. Britain remained in a semi-isolationistmood in 1866,and Napoleon 111
thought Francewould profit fromthe long, mutuallydebilitatingAustro-Prussianwar he expected.
Like the Soviets in 1939, Napoleon underestimated the danger of a quick, lopsided victory by
either side. Ibid., pp. 87-90.
Offense, Defense, and the Causes of War I 29
state even threatened interventi~n.~'As A.J.P. Taylor notes, Bismarck's 1866
diplomatic opportunity-a wide-open field for unopposed expansion-was
"unique in recent hist01-y."~~
In 1870 Bismarck ensured the neutrality of the other European powers by
shifting responsibility for the war to France and convincing Europe that the
war stemmed from French expansionism?' As a result, Prussia again had a
free hand to pursue its expansionist aims. It used this to smash France, seize
Alsace-Lorraine, and consolidatecontrol over South Germany?
1871-90. For some twenty years after the Franco-Prussian War, the defense
dominated because of Bismarck's new diplomacy and Britain's renewed activism.
In the military area the cult of the offensive had not yet taken hold. In
diplomacy Bismarck wove a web of defensive alliances that deterred aggressors
and calmed status quo powers after 1879.52British power waned slightly,
but this was offset by the recovery of Britain's will to play the balancer. The
"war-in-sight" crisis of 1875 illustrates the change:Britain and Russia together
deterred a renewed German attack on France by warning that they would not
allow a repeat of 1870-71?3
1890-1919. After 1890 military realities increasingly favored the defense, but
elites mistakenly believed the opposite. Diplomaticrealities swung toward the
offense, and elites believed they favored the offense even more than they did.
48. Ibid., p. 86.
49. Taylor, Struggle for Mastery, p. 156. Moreover, Bismarck stopped the 1866 war partly because
he feared French or Russian intervention if Prussia fought on too long or conquered too much.
Smoke, War,pp. 101-102. Thus lack of defense-enhancing diplomacy helped cause the war while
Prussian fear of such diplomacy shortened and limited the war.
50. William Carr, The Origins of the Wars of German Unification (London: Longman, 1991), p. 202;
and Michael Howard, The Franco-Prussian War: The German Invasion of France, 1870-1871 (New
York: Granada, 1961),p. 57. Austria also stayed neutral because Hungarian Magyar influence was
growing inside the Dual Monarchy, and the Magyars felt that the more Austria was pushed out
of Germany, the stronger the position of the Magyars within it would be. R.R. Palmer and Joel
Colton, A History of the Modern World,4th ed. (New York: Alfred A. Knopf, 19711,p. 574.
51. On Prussia's free hand, see Smoke, War,pp. 133-136; Norman Rich, The Age of Nationalism and
Reform, 1850-1890,2d ed.(New York W.W. Norton, 19m, p. 140;and W.E. Mosse, European Powers
and the German Question (New York Octagon, 1969), pp. 291,295.
52. Bismarck formed defensive alliances with Austria, Italy, and Romania, and a more limited
defensive accord with Russia-specifically, a reciprocal agreement not to join a war against the
other unless the other attacked France (in the German case) or Austria (in the Russian case).
Synopses include Paul M. Kennedy, The Rise and Fall of the Great Powers: Economic Change and
Military Conflict from 1500 to 2000 (New York: Random House, 1987), pp. 249-250; and Robert E.
&good and Robert W. Tucker, Force, Order, and Justice(Baltimore,Md.: Johns Hopkins University
Press, 1967),pp. 8C-81.For a longer account, see Taylor, Struggle for Mastery, pp. 258-280,316-319.
53. Imanuel Geiss, German Foreign Policy, 1871-1914 (Boston: Routledge and Kegan Paul, 19761,
p. 28.
International Security 22.4 I 30
European militaries were seized by a "cult of the offensive." All the European
powers adopted offensive military doctrines, culminating with France's
adoption of the highly offensive Plan XVII in 1913 and with Russia's adoption
of the highly offensive Plan 20 in 1914. More important, militaries persuaded
civilian leaders and publics that the offensedominated and conquestwas easy.
As a result, elites and publics widely believed the next war would be quickly
won by a decisive offensive.
Bismarck'sdefensiveallianceswithered or evolved into defensive-and-offensive
alliancesafter he left officein 1890, largelybecause the cult of the offensive
made defensive alliances hard to maintain. Pacts conditioned on defensive
conductbecame hard to frame because statesdefended by attacking, and status
quo powers shrank from enforcing defensive conduct on allies they felt less
able to lose. For example, Britain and France felt unable to enforce defensive
conduct on a Russian ally that defended by attacking and that they could not
afford to see defeated. Elites also thought that aggressors could overrun their
victims before allies could intervene to save them, making defensive alliances
less effective. Thus Britain seemed less able to save France before Germany
overran it, leading Germany to discount British power. Lastly, German leaders
subscribed to a bandwagon theory of diplomacy, which led them to underestimate
others' resistance to German expansion. Overall, the years before 1914
were the all-time high point of perceived offense dominance.
Nine of the ten intervening phenomena predicted by offense-defensetheory
(all except phenomenon G, nonnegotiation) flourished in this world of assumed
offense dominance. Opportunistic and defensive expansionist ideas
multiplied and spread, especially in Germany. Russia and France mobilized
their armies preemptively in the 1914 July crisis. That crisis arose from a fait
accompli that Germany and Austria instigated in part to shut a looming
window of vulnerability. This window in turn had emerged from a land arms
race that erupted during 1912-14. The powers enshrouded their military and
political plans in secrecy-a secrecy that fostered crisis-managementblunders
during July 1914. These blunders in turn evoked rapid, violent reactions that
helped drive the crisis out of control. Belief in the offense fueled offensive
military doctrines throughout the continent and impeded efforts to restrain
allies. Together these dangers formed a prime cause of the war: they bore the
1914 July crisis and helped make it uncontrollable.
191945. The interwar years were a mixed bag, but overall the offense gained
the upper hand by 1939, and the German elite believed the offense even
stronger than in fact it was.
Offense,Defense,and the Causes of War 1 31
Military doctrine and technology gave the defense the advantage until the
late 1930s, when German blitzkrieg doctrine combined armor and infantry in
an effective offensive combination. This offensive innovation was unrecognized
outside Germany and doubted by many in Germany, but the man who
counted most, Adolf Hitler, firmly believed in it. This reflected his faith in the
offense as a general principle, imbibed from international social Darwinist
propaganda in his y0uth.5~
More important, the workings of interwar diplomacy opened a yawning
political opportunity for Nazi expansion. Britain fell into a deep isolationism
that left it less willing to commit this declining power to curb continental
aggre~sors.5~The United States also withdrew into isolation, removing the
counterweight that checked Germany in 1918?6 The breakup of AustriaHungary
in that year created a new diplomatic constellationthat further eased
German expansion. Austria-Hungary would have balanced against German
54. Hitler often echoed international social Darwinist slogans on the short, precarious lives of
states, for example, ”Politics is in truth the execution of a nation‘s struggle for existence,” and
“Germanywill either be a world power or there will be no Germany.” Quoted in P.M.H. Bell, The
Origins of the Second World War in Europe (London: Longman, 1986), p. 81; and in Anthony P.
Adamthwaite, TheMaking of the Second World War(London:GeorgeAllenand Unwin, 1977),p. 119.
Hitler‘s faith in the offensive differed from that of the pre-1914 cultists of the offensivein three
ways. First, he saw offensive capabilities arising from a long search for offensive methods, not
from permanent properties of war. In his mind offense could be created, but also had to be;
Germany would discover offensive answers only after a long effort. In contrast, the pre-1914
cultiststhought offense inherentlyeasier than defense; deep thought need not be given to how to
make it superior, because it alreadywas. Second,Hitler’s offensive optimism was based on racism
and soda1 prejudice, as well as on assessment of military factors. Specifically, his contempt for
Slavs and Jews led him to expect that the Soviets would quickly collapse under German attack.
Third, Hitler’s concerns for German security focused on fear of conquest by economic strangulation,
not conquestby French or Sovietblitzkrieg. He thought German securitywas precarious,but
for reasonsrooted more in the politicaleconomyof war than in the nature of doctrineor weaponry.
These differencesaside, the logical implications of Hitler‘s offensive cult were the same as those
of the pre-1914 cult. He exaggerated both German insecurity and the feasibility of imperial
solutions to redress it.
55. Prime Minister Neville Chamberlain of Great Britain said in 1937that he ”did not believe we
could, or ought . . . to enter a Continentalwar with the intentionof fightingon the same lines as
in the last,” meaningthat Britain would deploy no largeground force on the continent.Bell, Origins
of the Second World War in Europe, p. 177. Britain had only two divisions available to send to the
continent during the 1938Munich crisis, and the four-division force it actually sent in 1939was
smaller and less well trained than its small expeditionaryforce of 1914.These four divisions were
a drop in the bucket relative to the 84 French and 103 German divisions then deployed. Ibid.,
p. 175.
56. The United States also proclaimed this isolationism in four neutrality laws passed during
1935-39, giving Hitler a clear if misleading signal of American indifferenceto his aggression.On
these laws a synopsis is Thomas A. Bailey, A Diplomatic History of the American People, 9th ed.
(EnglewoodCliffs, N.J.: Prentice-Hall,1974),pp. 701-702,715.
International Security 22:4 1 32
expansion, but its smaller successor states tended to bandwag0n.5~This let
Hitler extend German influence into southeast Europe by intimidation and
subversion.
The Soviet Union and the Western powers failed to cooperate against
Hitler.%Ideologicalhostility divided them. Britain also feared that a defensive
alliance against Hitler would arouse German fears of allied encirclement,
spurring German aggressiveness.This chilled British enthusiasm for an AngloFrench-Soviet
alliance.59
Hitler exaggerated the already-large advantage that diplomacy gave the
offense because he thought bandwagoning prevailed over balancing in international
affairs. This false faith colored all his political forecasts and led him
to vastly underestimate others states’resistance to his aggressions. Before the
war he failed to foresee that Britain and France would balance German power
by coming to Poland’s rescue.6oOnce the war began he believed Germany
could intimidate Britain into seeking alliance with Germany after Germany
crushed France-or, he later held, after Germany smashed the Soviet Union.61
He thought the United States could be cowed into staying neutral by the 1940
German-Japanesealliance (the alliance had the opposite effect, spurring U.S.
intervention)?’ In short, Hitler’s false theories of diplomacy made three of his
most dangerous opponents shrink to insignificance in his mind.
These realities and beliefs left Hitler to face temptations like those facing
Bismarck in 1866and 1870. Hitler thought he could conquer his victims seriatim.
He also thought his conquests would arouse little countervailing opposition
from distant neutral powers.63As a result, he believed he faced a yawning
opportunity for aggression.
57. Explainingwhy weaker states are more prone to bandwagon than are stronger states is Walt,
Origins of AKunces, pp. 29-30.
58. Bell, Origins of the Second World War in Europe, pp. 172, 224, 260; and Adamthwaite,Making of
the Second World War, pp. 60,69. This failure greatly eased Hitler’s aggressions,because geography
made Britain‘s 1939 guarantees to Poland and Romania unenforceable without a Soviet alliance.
Ibid., pp. 86, 91.
59. RaymondJ. Sontag,A Broken World, 2929-2939 (New York: Harper and Row, 1971),p. 361.
60. On August 22, 1939, Hitler assured his generals that “the West will not intervene” to defend
Poland. Jeremy Noakes and Geoffrey Pridham, eds., Nazism, 1919-1945: A History in Documents
and Eyewitness Accounts, 2 vols. (New York Schocken Books, 1988),vol. 2, p. 741.
61. See Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.:
Cornell University Press, 1991),p. 94.
62. Noakes and Pridham, Nazism, vol. 2, p. 797. Some German leaders also hoped that Germany
could win decisively in Europe before the United States could bring its power to bear. Thus in
September 1940 Hitler‘s naval commander in chief voiced the hope that Britain could be beaten
“before the United States is able to intervene effectively.” Ibid., p. 794.
63. The fine-grainedpattern of events during 1938-40-who attacked whom and when-also fits
the predictionsof offense-defensetheory (specifically,prime prediction 3).The Western alliesstood
offense, Defense, and the Causes of war I 33
Unlike 1914, the late 1930s were not a pure case of perceived offense dominance.
Instead, the 1930s saw status quo powers’ perceptions of defense dominance
create real offensive opportunities for an aggressor state. Hitler thought
the offense strong and even exaggerated its strength, but other powers (the
Soviet Union, Britain, and France) underestimated its strength. Their percep
tions of defense dominance relaxed their urge to jump the gun at early signs
of threat (as Russia did in 1914); this made things safer. But this perception
also relaxed their will to balance Germany,because they found German expansion
less frightening. This weakened the coalition against Hitler, leaving him
wider running room.@
1945-1990s. After 1945 two changes swung the offense-defense balance back
toward the defense. First, the end of American isolationismtransformed European
political affairs. The United States replaced Britain as continental balancer,
bringing far more power to bear in Europe than Britain ever had. As a
result, Europe in the years after 1945 was unusually defense dominant from a
diplomatic standpoint.
Second, the nuclear revolution gave defenders a large military advantage
so large that conquest among great powers became virtually impossible. Conquest
now required a nuclear first-strike capability (the capacity to launch a
nuclear strike that leaves the defender unable to inflict unacceptable damage
in retaliation). Defenders could secure themselves merely by maintaining a
second-strike capability (the capacity to inflict unacceptable damage on the
attacker’s society after absorbing an all-out strike). The characteristicsof nuclear
weapons-their vast power, small size, light weight, and low costensured
that a first-strike capability would be very hard to attain, while a
second-strikecapability could be sustained at little cost. As a result, the great
powers became essentiallyunconquerable, and even lesser powers could now
stand against far stronger enemies. Overall, the nuclear revolution gave defenders
an even more lopsided advantage than the machine gun-barbed wireentrenchments-railroad
complex that emerged before 1914.
without attacking Germany in 1938 and again in 193940 because they doubted they could win a
decisive victory.Germany stood without attacking westward in the fall of 1939 for the same reason,
and finally attacked in May 1940 after German military leaders developed a plausible plan for
decisive attack. Mearsheimer, Conventional Deterrence, pp. 67-133.
64. Would the risk of war have fallen had all powers believed the offense was dominant in the
late 1930s?Thisseems unlikely.The status quo powers would have balanced harder against Hitler,
offering him more discouragement, but they also would have been jumpier, making early crises
more dangerous. One of these crises-Hitler’s remilitarization of the Rhineland, the Spanish civil
war, or the German seizure of Austria or Czechoslovakia-probably would have served as the
“Sarajevo” for World War 11, with the Allies moving first as Russia did in 1914.
International Security 22:4 134
American and Soviet policymakers grasped thiscosmic military revolution
only slowly, however. At first many feared nuclear weapons would be a boon
to aggressors. When this fear proved false, the vast advantage they gave
defenders was only dimly recognized, partly because scholarsstrangely failed
to explain it. Thus the nuclear revolution changed realities far more than they
did perceptions. As a result, state behavior changed only slowly, and both
superpowers competed far harder-in both Central Europe and the third
world-than objective conditions warranted. The Cold War was far more
peacefulthan theprecedingfortyyears, but could havebeen stillmorepeaceful
had Soviet and US.elites understood that their security problems had vastly
diminished and were now quite small.
In sum, the events of 1789-1990s clearly corroborate offense-defensetheory
predictions-specifically, prime predictions 1 and 2, as well as both explanatory
predictions. These conclusions rest on rather sketchy data-especially
regarding the explanatory predictions-but that data confirm offense-defense
theory so clearly that other data would have to be very different to reversethe
result.
0 The incidence of war correlatesloosely with the offense-defensebalance and
very tightly with perceptions of the offense-defensebalance (for a summary
see Table 1).
Europe’s less-secure and more offensively capable continental powers were
perennial troublemakers, while more secure and less offensively capable
offshore powers were perennial defenders of the status quo. Prussia/Germany
was cursed with the least defensible borders and faced the most
offensive temptations. It started the largest number of major wars (1864,
1866,1914,1939, and shared responsibilityfor 1870with France).France and
Russia, with more defensible borders and fewer temptations, started fewer
major wars.65Britain and the United States,blessed with even more insulating
borders, joined a number of European wars but started none.&Spain,
Sweden, and Switzerland, also insulated from other powers by mountains
or oceans, fought very little.
Thus the timing of war and the identities of the belligerents tightly fit
prime predictions 1 and 2.
65.France can be assigned prime responsibility for 1792 and 1859,and shared responsibilityfor
Crimea and 1870.Russia deserves primeresponsibilityfor the ColdWar and shared responsibility
for Crimea and the 190445Russo-JapaneseWar.
66. Britain does share responsibility for the Crimean War with Russia, France, and Turkey.
Offense,Defense, and the Causes of War I35
0 Sketchy evidence suggests that opportunistic and defensive expansionism
were more prominent during the periods of perceived offense dominance
(1792-1815, 1859-71,1890-1914,1930~-1945)than at other times. The years
1792-1815 saw a strong surge of French expansionism, nearly matched at
the outset by parallel Prussian e~pansionism.6~The mid-nineteenth century
saw large opportunistic expansionism in Prussia and some French expansionism.
The years 1890-1914 saw vast expansionist ambitions develop in
Wilhelmine Germany6*matched by fierce resistance to this German expansionism
in Russia and France, and by lesser French and Russian expansionism.
Large German expansionism then reappeared under the Nazis in the
1930s.During other periodsEuropeanexpansionismwas more muted:European
powers had smaller ambitions and acted on them less often. This
supports explanatory prediction 1.
0 Opportunistic and defensive expansionism were prominent among those
statesthat saw the clearest defensivevulnerability and offensiveopportunity
(especiallyPrussia/Germany, also revolutionary France), while being more
muted among states with more secure borders and fewer offensive opportunities
(Britain, the United States, the Scandinavianstates, and Spain).This
corroboratesexplanatory prediction 2.
How strong is this test? The strength of a passed test depends on the
uniqueness of the predictions tested. Do other theories predict the outcome
observed, or is the prediction unique to the tested theory? The predictions
tested here seem quite unique. There is no obvious competingexplanation for
the periodicupsurges and downsurgesin European expansionismand warfare
outlined above.Offense-defensetheory has the field to itself.Particular domestic
explanations have been offered to explain the aggressiveness of specific
states-for example, some argue that Wilhelmine Germany was aggressive
because it was a late industrializer, that revolutionary France was aggressive
becauseitsregime cameto power through mass revolution,and sof~rth~~-but
no competing theory claims to explain the general cross-time and cross-state
pattern of war involvement that we observe. Hence this test seems strong.
What importance does thisevidence assign to offense-defensetheory?That
is, how potent is offense dominance as a cause of war? In Europe since 1789,
the nature of international relations has gyrated sharply with shifts in the
67. On Prussia's expansionism, see Blanning, Origins of the French Revolutionary Wars, pp. 72-82;
on French expansionism,see ibid., passim.
68. A summary of WilhelmineGerman aims and policies is Geiss, German Foreign Policy.
69. On Germany as late industrializer,see Snyder,Myths of Empire, pp. 66-111; and on France as
a revolutionary state, see Walt, Revolution and War,pp. 46-128.
International Security 22:4 I36
perceived offense-defense balance. War is far more common when elites believe
that the offense dominates, and states are far more belligerent when they
perceive large defensive vulnerabilities and offensive opportunities for themselves.
This indicates that perceptions of the offense-defense balance have a
large impact on international relations. Offense-defense theory is important as
well as valid.
How much history does this evidence suggest that offense-defense theory
can explain? Explanatory power is partly a function of the prevalence of the
theory's cause: abundant causes explain more history than scarce causes. In
Europe since 1789 the offense has seldom been really strong, but it was
believed strong quite often-often enough to cause considerable trouble.
TEST 2: ANCIENT CHINA
The ancient Chinese multistate system witnessed a long-term shift from
defense dominance to offense dominance across the years 722-221 BCE.70
Offense-defense theory predicts that warfare should have increased as this
transformation unfolded (see prime prediction 1). This prediction is fulfilled:
diplomacygrew markedly more savage and international relationsgrew markedly
more violent as the power of the offense increased.
Before roughly 550 BCE the defense held the upper hand among China's
many feudal states. Four related changes then strengthened the offense: feudalism
declined:* mass infantry replaced chariots as the criticalmilitary force,
conscriptionwas introduced, and armies grew tremendously in size.n The two
largest Chinese states deployed enormous armies of more than a million men,
and somesmaller states had armiesnumbering in the hundreds of thousands.73
As armies grew, border forts had less stopping power against infantry because
invaders could sweep past, leaving a smaller portion of their force behind to
besiege the forts. Forts also lost stopping power as improved siege-engines
appeared-battering rams, catapults, and rolling towers-that further eased
the conquest of fortified positions.74The decline of feudalism eased offensive
operations by reducing social stratification, which increased troop loyalty to
70. Concurringis Andreski, Military Organization and Society, p. 76.
71. Noting the decline of feudalism are Samuel B. Griffiths, "Introduction,"in Sun Tzu, The Art
of War (London:Oxford University Press, 1971),p. 33; and Dun J. Li, The Ageless Chinese: A History,
3d ed.(New York: Charles Scribner's Sons, 19781, p. 64.
72. On the growth of armies, the introduction of conscription, and the rise of infantry, see Li,
Ageless Chinese, p. 56; Griffiths, "Introduction,"pp. 28, 33; and Wolfram Eberhard, A History of
China (Berkeley:University of California Press, 19m, p. 49.
73. Li, Ageless Chinese, p. 56.
74. Andreski, Military Organization and Society, p. 76.
Offense,Defense, und the Causes of War I 37
regimes; this meant troops could be trusted to conduct long-distanceoffensive
operations without deserting.
The outcomes of battles and wars reveal the shift toward the offense that
these technical and social changes produced. The number of independent
Chinese states declined from two hundred in the eighth century BCE to seven
in the late fifth century, to one in the late third century-a clear measure of the
growing power of the offense.75Before550 BCE defenders were often victorious.
Thus the states of Tsin and Ch’i fought three great battles, in 632,598, and 567
BCE, each won by the defender.Dun J. Li concludes,”If the three battles indicate
anything, they meant that neither side was able to challenge successfully the
other‘s leadership in its own sphere of infl~ence.”~~In contrast, the state of
Ch’in conquered all of China in a rapid campaign lasting only nine years at
the end of the Warring States period (230-221 B C E ) . ~
This increasein the power of the offense coincideswith a stark deterioration
in international relations. During the Springand Autumn period (722453 BCE)
interstate relations were fairly peaceful, and wars were limited by a code of
conduct. The code confined warfare to certain seasons of the year and forbade
killing enemy wounded. It was considered wrong to stoop to deceit, to take
unfair advantage of adversaries, to ”ambush armies,” or to “massacrecities.“78
The subsequent WarringStatesperiod (453-221 BCE) was perhaps the bloodiest
era in Chinese history. Warfare raged almost constantlyn becoming a “fundamental
occupation” of states.*’ Restraints on warfare were abandoned. Casualties
ran into hundreds of thousands, and prisoners of war were massacred
en masse.81Diplomatic conduct deteriorated; one historian writes that “diplomacy
was based on bribery, fraud, and deceit.”@
In short, the shift toward offense dominance in China during 722-221 BCE
correlates tightly with a dramatic breakdown of China’s international order.
TEST 3: UNITED STATES 1789-1990s
Since 1815 the United States has been by far the most secure of the world’s
great powers, blessed with two vast ocean moats, no nearby great powers, and
75. Li, Ageless Chinese,pp. 50, 59.
76. Ibid., p. 52.
77. Ibid., p. 59.
78. Griffiths,“Introduction,”p. 30.
79. Ibid., p. 21.
80. Ibid., p. 24, quoting Shang Yang, Prime Minister of Chin, who conceived war and agriculture
to be the two fundamental occupations.
81. Li, Ageless Chinese, pp. 56, 58-59.
82. Griffiths, “Introduction,”p. 24.
International Security 22:4 I 38
(after 1890)the world’s largest economy. In the nineteenth century the United
States also had substantial offensive opportunities, embodied in chances for
continental and then Pacific expansion against weak defenders. However,
America’s security endowments were quite extraordinary, while its offensive
opportunities were more ordinary. Offense-defense theory predicts that such
a state will exhibit perhaps average offensiveopportunism but markedly less
defensive belligerence than other states. Hence, on net, it will start fewer wars
and be involved in fewer wars than others (see prime prediction 2).
This forecast is confirmed, although not dramatically, by the pattern of past
U.S. foreign policy. The United Stateshas fought other great powers only three
times in its two hundred-year history-in 1812,1917, and 1941-a low count
for a great power.83The 1812war stemmed mainly from US. belligerence,but
the wars of 1917 and 1941 resulted mainly from others‘ belligerence. The
United States did start some of its lesser wars (1846 and 18981, but it joined
other wars more reactively (Korea and Vietnam).
Offense-defensetheory also predicts that while the United Stateswill pursue
some opportunistic expansionism (intervening phenomenon A), it will embrace
few policies that embody offense-defense theory’s other intervening
phenomena 03-J) (explanatory prediction 2). Where the record allows judgments,
this forecast is borne out. Regarding expansionism, the United States
has confined itself largelyto opportunistic imperialism against frailopponents.
Defensive expansionismhas been muted, and overall, expansionistideas have
held less sway in the United States than in other powers. This is reflected in
the relatively small size of the U.S. empire. The modern American empire has
been limited to a few formal colonies seized from Spain in the 1890s and an
informal empire in the Caribbean/Central American area, with only intermittent
control exerted more widely-a zone far smaller than the vast empires of
the European powers.
The U.S. impulse to engage in preemptive and preventive war has been
small. In sharp contrast to Germany and Japan, the United Stateshas launched
a stealthy first strike on another major power just once (in 1812) and has
jumped through only one window of opportunity (in 1812). Surprise first
strikes and window-jumping were considered on other occasions(e.g.,preventive
war was discussed during 1949-54, and surprise attack on Cuba was
considered during the Cuban missile crisis), but seldom seriously.
83. Britain, France, Russia, and hssia/Germany fought other great powers an average of five
times over the same two hundred years, by my count.None fought as few as three times.
Offense,Defense, and the Causes of War I 39
American diplomacy has been strikingly free of fait accomplitactics. American
foreign and security policy has generally been less secretive than those of
the European continental powers, especially during the late Cold War, when
the United States published military data that most powers would highly
classify as state secrets.The U.S. arms raced with the SovietUnion energetically
during the Cold War, but earlier maintained very small standing military
forces-far smaller than those of other great powers. Overall, intervening
phenomena B-J of offense-defensetheory are strikinglyabsent in the U.S. case.
In sum, the United States has not been a shrinking violet, but it has been
less bellicose than the average great power. Compare, for example, U.S. conduct
with the far greater imperial aggressions of Athens, Rome, Carthage,
Spain, Prussia/Germany, Japan, Russia, and France.
Offense-defense theory further predicts that levels of American bellicosity
should vary inversely with shifts over time in America's sense of security and
directlywith the scope of perceived external threats (see prime prediction 3)as
in fact they have.
During 1789-1815 the United States saw large foreign threats on its borders
and large opportunities to dispel them with force.It responded with a bellicose
foreign policy that produced the 1812 war with Britain.
During 1815-1914 the United States was protected from the threat of a
Eurasian continental hegemon by Britain's active continental balancing, and
protected from extracontinental European expansion into the Western hemisphere
by the British fleet, which was the de facto enforcer of the Monroe
Doctrine. The United Statesresponded by withdrawing from European affairs
and maintaining very small standing military forces, although it did pursue
continental expansion before 1898 and limited overseas imperial expansion
after 1898.
During 1914-91 Britain could no longer maintain the European balance. This
deprived the United States of its shield against continental European aggressors.Then
followed the great era of American activism-fitful at first (191743,
then steady and persistent (1947-91). This era ended when the Soviet threat
suddenly vanished during 1989-91. After 1991 the United States maintained
its security alliances, but reduced its troops stationed overseas and sharply
reduced its defense effort.
WHAT THESE TESTS INDICATE
Offense-defensetheory passed the tests these three cases pose. Are these tests
positive proof for the theory or mere straws in the wind?
International Security 22:4 I 40
We learn more from strong tests than from weak ones. The strength of a
passed test is a function of the uniqueness of the predictions that the test
corroborated. The more numerous and plausible are contending explanations
for the patterns that the test theory predicted and the test revealed, the weaker
the test.
The three case study tests reported here range from fairly weak to quite
strong. They each lack Herculean power but in combinationthey pose a strong
test. The test posed by the ancient China case is weak because our knowledge
of ancient Chinese society and politics is fairly thin. This leaves us unable to
rule out competing explanations for the rise of warfare in the Warring States
period that point to causes other than the rise of offense. The test posed by the
U.S. case is a little stronger but still rather weak overall. Alternative explanations
for the rise and fall of American global activism are hard to come up
with, leaving the offense-defensetheory’s explanation without strong competitors,
so this element of the test posed by the U.S. case is fairly strong. Plausible
contending explanations for other aspects of the U.S. case can be found,
however. For example, some would argue that America’s more pacific conduct
is better explained by its democratic domestic structure than by its surfeit of
security.Others would contend that the United States has fewer-than-average
conflicts of interest with other powers because it shares no borders with them,
and it fights fewer wars for this reason. Hence this element of the test posed
by the U.S. case is weak: US.lower-than-averagebellicosity is only a straw in
the wind.
As noted above, the case of Europe since 1789offersa fairly strong test. Some
competingexplanationsfor Germany’sgreater bellicosityare offered-as noted
above, the lateness of German industrialization is sometimes suggested as an
alternative cause, as is German culture. However, there is no obvious plausible
competing explanation for the main pattern we observe in the case-the rise
of warfare during 1792-1815,1856-71, and 1914-45,and the greater periods of
peace in between. The fit of this pattern with prime prediction 1of offense-defense
theory lends it strong corroboration.
WHAT PRESCRIPTIONS FOLLOW?
If offense dominance is dangerous, policies that control it should be pursued.
Governments should adopt defensive military force postures and seek arms
control agreements to limit offensive forces. Governments should also maintain
defensivealliances.American securityguarantees in Europeand Asia have
made conquest much harder since 1949 and have played a major role in
Offense, Defense, and the Causes of War I 42
preserving peace. A U.S. withdrawal from either region would raise the risk
of conflict.
Conclusion: Offense-Defense Theory in Perspecfive
Offense-defensetheory has the attributes of a good theory. First, it has three
elementsthat give a theory claim to large explanatorypower. (1)Large importance,
that is, its posited cause has large effects. Variance in the perceived
offense-defense balance causes large variance in the incidence of warfare.
Variance in the actual offense-defensebalance has less impact because policymakers
often misperceive it, but it has a potent effect when policymakers
perceive it accurately.(2)Wide explanatoryrange. The theory explains results
acrossmany domains of behavior-in military policy, foreign policy, and crisis
diplomacy.84It governs many intervening phenomena (e.g., expansionism,
first-move advantage, windows, secrecy, negotiation failures, crisis management
blunders, arms races, tight alliances) that have been seen as important
war causes in their own right. Thus offense-defense theory achievessimplicity,
binding a number of war causes under a single rubric. Many causes are
reduced to one cause with many effects. (3)Wide real-world applicability.Real
offense dominance is rare in modern times, but the perception of offense
dominance is fairly widespread. Therefore, if perceived offense dominance
causes war it causes lots of war, and offense-defensetheory explainsmuch of
international history.
Second, offense-defensetheory has large prescriptive utility, because the
offense-defense balance is affected by national foreign and military policy;
hence it is subject to political will. Perceptions of the offensedefense balance
are even more malleable, being subject to correction through argument. Both
are far more manipulable than the polarity of the international system, the
strength of international institutions, the state of human nature, or other war
causes that have drawn close attention.
Third,offense-defensetheory is quitesatisfymg,althoughit leaves important
questions unanswered. In uncovering the roots of its ten interveningphenom-
84. Moreover, offense-defense theory might be usefully adapted for application beyond the domain
of war, for example, to explain internationaleconomiccompetition(orcooperation),or even
intra-academic competition. Suggesting its application to economics is JitsuoTsuchiyama, who
writes of the "prosperitydilemma''-a cousin of the security dilemma in which measures taken
by one state to increaseits economic well-being decreaseanother's economic well-being.SeeJitsuo
Tsuchiyama, "The US.-JapanAlliance after the Cold War: End of the Alliance?" unpublished
manuscript,O h Institute, Harvard University, 1994, p. 27.
International Security 224 I 42
ena, offense-defensetheory offers a more satisfying (and simpler) explanation
than do interpretations pointing directly to these phenomena.However, it also
raises another mystery: Why is the strength of the offense so often exagger-
ated?
Historysuggeststhat offensedominanceis at the same time dangerous,quite
rare, and widely overstated.It further suggests that this exaggeration of insecurity,
and the bellicose conduct it fosters, are prime causes of national insecurity
and war. Statesare seldomas insecureas they think they are. Moreover,
if they are insecure,this insecurityoften grows from their own efforts to escape
imagined insecurity.
The rarity of real insecurity is suggested by the low death rate of modern
great powers. In ancient times great powers often disappeared, but in modern
times (since1789)no great powers have permanently lost sovereignty,and only
twice (Francein 1870-71 and in 1940) has any been even temporarily overrun
by an unprovoked aggre~sor.’~Both times Francesoonregained its sovereignty
through the intervention of outside powers-illustrating the powerful defensive
influence of great-power balancing behavior.
The prevalence of exaggerationsof insecurity is revealed by the great wartime
enduranceof many statesthat enter wars for securityreasons, and by the
aftermath of the world’s great security wars, which often reveal that the
belligerents’ security fears were illusory. Athens fought Sparta largely for
securityreasons, but held out for a full nine years (413-404 BCE) after suffering
the crushing loss of its Sicilian expedition-an achievement that shows the
falsehood of its original fears. Austria-Hungary held out for a full four years
under allied battering during 1914-18, a display of toughness at odds with its
own prewar self-image of imminent collapse. With twenty-twenty hindsight
we can now see that modem Germany would have been secure had it only
behaved itself. Wilhelmine Germany was Europe’s dominant state, with
Europe’s largest and fastest-growing economy.It faced no plausible threats to
its sovereigntyexcept those it created by its own belligerence.Later, interwar
Germany and Japan could have secured themselves simply by moderating
their conduct. This would have assured them of allies, hence of the raw
materialssuppliesthey sought to seizeby force.America’saggressiveand often
costly Cold War interventions in the third world now seem hypervigilant in
light of the defensive benefits of the nuclear revolution,America’s geographic
85. France helped trigger the 1870 war; hence one could argue for removing France in 1870 from
the list of unprovoking victims of conquest, leaving only France in 1940.
Offense, Defense, and the Causes of War I 43
invulnerability,and the strength of third world nationalism, which precluded
the Soviet third world imperialism that U.S. interventions sought to prevent.
Paradoxically, a chief source of insecurity in Europe since medieval times
has been this falsebelief that securitywas scarce.Thisbelief was a self-fulfilling
prophecy, fostering bellicose policies that left all states less secure. Modern
great powers have been overrun by unprovoked aggressors only twice, but
they have been overrun by provoked aggressorssix times-usually by aggressors
provoked by the victim’s fantasy-driven defensive bellicosity. Wilhelmine
and Nazi Germany, Imperial Japan, Napoleonic France, and Austria-Hungary
were all destroyed by dangers that they created by their efforts to escape from
exaggerated or imaginary threats to their safety?
If so, the prime threat to the security of modern great powers is . . . themselves.
Their greatest menace lies in their own tendency to exaggerate the
dangers they face, and to respond with counterproductive belligerence. The
causes of this syndrome pose a large question for students of international
relations.
86. Mussolini also provoked his own destruction, but his belligerencewas not security driven.