Content uploaded by Brandon Valeriano
Author content
All content in this area was uploaded by Brandon Valeriano on Jul 21, 2016
Content may be subject to copyright.
PLEASE SCROLL DOWN FOR ARTICLE
This article was downloaded by:
[University of Illinois]
On:
17 August 2009
Access details:
Access Details: [subscription number 906604402]
Publisher
Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
International Interactions
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713718605
The Tragedy of Offensive Realism: Testing Aggressive Power Politics Models
Brandon Valeriano a
a Department of Political Science, University of Illinois at Chicago, Chicago, Illinois, USA
Online Publication Date: 01 April 2009
To cite this Article Valeriano, Brandon(2009)'The Tragedy of Offensive Realism: Testing Aggressive Power Politics
Models',International Interactions,35:2,179 — 206
To link to this Article: DOI: 10.1080/03050620902864493
URL: http://dx.doi.org/10.1080/03050620902864493
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
International Interactions, 35:179–206, 2009
Copyright © Taylor & Francis Group, LLC
ISSN: 0305-0629
DOI: 10.1080/03050620902864493
179
GINI0305-06291547-7444International Interactions, Vol. 35, No. 2, Apr 2009: pp. 0–0International Interactions The Tragedy of Offensive Realism:
Testing Aggressive Power Politics Models
The Tragedy of Offensive RealismB. Valeriano
BRANDON VALERIANO
Department of Political Science, University of Illinois at Chicago, Chicago, Illinois, USA
Offensive realism is a theory of opportunistic state power maximi-
zation. States are said to always seek power so that they may dom-
inate the international system and survive under conditions of
anarchy. The theory of offensive realism is both descriptive in the
sense that it suggests how states have acted in the past and
prescriptive in that it suggests how states should conduct foreign
policy. What remains is to empirically test the propositions that
offensive realism advocates. Under such testing, we argue that
offensive realism fails to accurately explain the “tragedy of great
power politics.” We find that two opposing theories, one norm-
based and one issue-based, perform better than offensive realism
in describing the actions of major powers. If the theory fails to
accurately explain past historical events, it is of little use for guiding
future actions and policy.
KEYWORDS offensive realism, power politics, rivalry, issues,
norms
Realism remains the dominant paradigm of international relations theory.1
The tenets of realism include the notions that states seek power to survive,
major powers dominate the international system, and the international system
is anarchic. John Mearsheimer (2001) follows Morgenthau (1948) and Waltz
I would like to extend a special thanks to Kwang Teo who was instrumental in the early conception of
this paper and provided data assistance. This paper was originally presented at the annual meeting of
the International Studies Association, Portland, February 25–March 1, 2003. I thank Doug Lemke,
Alan Ward, John Van Benthuysen, Matt Powers, Ekaterina Svyatets, Michelle Zetek, Megan Preusker,
John Vasquez, Patrick James, and Amy Beth Schoenecker for their comments and suggestions.
Address correspondence to Brandon Valeriano, Department of Political Science (M/C 276),
University of Illinois at Chicago, 1007 West Harrison St., Chicago, IL 60607-7137, USA.
E-mail: bvaler@uic.edu
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
180 B. Valeriano
(1979) in adding to the canon of realist thought in international relations
theory by proposing a theory of offensive realism based on the structural
conditions of the system and the offensive nature of state action. If major
powers are assumed to seek more than survival they should seek to thrive
in the international system.
Offensive realism is a theory of opportunistic state power maximiza-
tion. States always seek power so that they may ensure their state’s security
under conditions of anarchy in the international system. The theory of
offensive realism is both descriptive in that it describes how states have
acted in the past and prescriptive in that it suggests how states should con-
duct foreign policy. What remains is to empirically test the propositions that
offensive realism proposes. The goal of this effort is not to falsify the predic-
tions of offensive realism, but to investigate its empirical accuracy and sug-
gest future research directions. A theory is of little use if it fails to
adequately explain real world events both in the present and in the past.
Offensive realism should be put to a test by comparing it to the history of
major state action from 1816 to 1992. While case studies are useful in
explaining events and suggesting new theoretical innovations, a large-N test
of Mearsheimer’s (2001) theory is needed to investigate its validity and
degree of explanatory power.
This paper will seek to test the notion that states are constantly
engaged in an unrelenting pursuit of power. Ultimately, the study con-
ducted here finds evidence that offensive realism fails to accurately
explain the actions of great powers. The theory clearly contains more
“anomalies” than Mearsheimer (2001, p. 10) is willing to concede. Addi-
tionally, according to these findings, it appears that two opposing
theories, one norm-based and one issue-based, perform significantly
better than offensive realism in explaining the actions of major powers. To
more accurately explain the actions of major powers, offensive realism
needs to be reevaluated so that the predictions correspond with the historical
record.
MEARSHEIMER’S THEORY OF OFFENSIVE REALISM
John Mearsheimer (1990; 1995; 2001) has been the main proponent of the
emerging offensive realist theory.2 Mearsheimer (2001, p. 54) asserts that
states have a will to power in that they do not merely seek to survive, but to
thrive in the international system constrained by anarchy; namely the goal is
to maximize their share of world power due to structural motivations. “This
unrelenting pursuit of power means that great powers are inclined to look
for opportunities to alter the distribution of world power in their favor . . .
simply put, great powers are primed for offense.” (Mearsheimer, 2001, p. 3)
An “unrelenting pursuit of power” suggests that great powers are constantly
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 181
seeking to alter the distribution of power in their favor. There can be no
mistake in the wording; power is a constant goal for those states that wish
to survive in an anarchic system.
Offensive realism focuses on the actions of great powers, and the ulti-
mate aim of a great power is to become a hegemon.3 Becoming a global
hegemon and dominating the international balance of power is the only
way states can ensure their survival in an anarchic environment.4
Mearsheimer believes that since no state is likely to attain global hegemony,
the world is condemned to great power competition and war.
To become a hegemon, a state must meet three conditions: it must
become a regional hegemon, acquire wealth and land power, and develop
nuclear weapons. Each of these represents a step toward power, and power
is primarily achieved through war, blackmail, bait and bleed strategies, and
bloodletting. As in all realist theoretical work, power is the key variable that
determines state action, where the goal is to gain power, not to distribute it
equally (Waltz, 1979). Thus, balance of power is not the optimal outcome
since states will always seek to gain more power and offset the balance.
Mearsheimer paints the international system as a bleak world filled
with states constantly seeking hegemony, a status they cannot acquire.5 This
perpetual movement toward hegemonic ambition ultimately leads states to
enter into war. He argues that the pursuit of power will cease only when
hegemony is achieved, which is impossible in the current system (Mearsheimer,
2001, p. 34).
THREE TENETS OF OFFENSIVE REALISM
1. Goal is to maximize share of world power.
2. Ultimate aim is to become the hegemon.
3. Since global hegemony is impossible, the world is condemned to perpetual
great power competition.
Offensive realism suggests that major powers are continuously
seeking power. The causal mechanism for this action is the lack of a
central authority.6 The ultimate goal of a state is to prevent state failure,
and the only way to do this is through power maximization. Mearshe-
imer argues that when states have offensive capabilities, they are likely
to use them. States can never be certain of other states’ intentions. If a
state has the capabilities to assert its dominance on the international
system, it will do so merely because it has the means (weapons) and the
will (survival). Satisfaction plays no part in conflictual interactions.7
Rather, all that matters for an offensive realist is that states have the
power to change the status quo and seek to do so regardless of other
conditions.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
182 B. Valeriano
It would be useful to make clear the distinctions between offensive
realism on one hand, and prudential (Morgenthau, 1948) and structural real-
ism (Waltz, 1979) on the other. Mearsheimer’s theory differs from
Morgenthau’s (1948) in that the structure of the system leads states to seek
power. Mearsheimer also believes that conflict is not hardwired into the
human brain. It is merely a result of the absence of security and hegemony.
Morgenthau advocates a more prudent version of offensive action; only
when the national interest of a state is threatened should a state take action.
Mearsheimer’s version of realism, in contrast, purports that a state seeks
hegemony even before the national interest of that state is threatened (pre-
emption). Mearsheimer differs from Waltz (1979) in his proposition that
states do not seek to maintain a balance of power, but to acquire hege-
mony. For offensive realists, state survival dictates that they seek to acquire
more power whenever the opportunity arises. “If they want to survive, great
powers should always act like good offensive realists.” (Mearsheimer, 2001,
p. 12) Survival is not achieved through balancing, but in the maximization
of state power.
To gain power, states must pursue a variety of strategies aimed at
increasing their position in the system (Mearsheimer, 2001, p. 13). The main
strategy to gain power is through blackmail and war. Specifically, states use
threats or force to acquire goods they seek to control. War can be used to
decrease the security of another state and therefore increase the security of
the aggressor state if gains are considered absolute. Mearsheimer (2001,
p. 139) also suggests that states can undertake strategies of bloodletting and
bait and bleed. Causing losses in another state can only increase security of
the state taking threatening actions. Finally, states use both balancing and
buck-passing strategies to maintain their power.
What remains is to examine the logic of the theory and to test its
empirical accuracy against a large number of cases. The next section of this
article considers the various conditions of offensive action and tests their
empirical accuracy using data from states at dispute between 1816 to 1992.
RECONSIDERING OFFENSIVE REALISM
In this section of the paper, we will outline how offensive realism can be
tested as an empirical theory. First, we must ask if the theory is empirically
accurate based on the conditions laid out by the theory and those condi-
tions that can be logically derived from the theory. Next, the propositions
that great powers are addicted to conflict and are the most important actors
in the system will be discussed. Finally, offensive realism is put to a head-
to-head test against two opposing theories that predict different paths of
action for great powers. If offensive realism cannot stand up to empirical
testing according to the conditions discussed here and if it cannot surpass
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 183
other opposing explanations that better identify the conditions for great
power action, the theory is in need of reevaluation. In fact, other realist
scholars (James, 2009) have begun an effort to question the micro-foundations
of the theory. We go further here in that we question the empirical founda-
tions of the theory in addition to its theoretical foundations.
Is Offensive Realism Empirically Accurate?
Mearsheimer (2001, p. 232) finds that “the nuclear arms race between the
superpowers and the foreign policy behavior of Japan (1868–1945),
Germany (1862–1945), the Soviet Union (1917–1991), and Italy (1861–1943)
show that great powers look for opportunities to shift the balance of power
in their favor and usually seize opportunities when they appear.” Labs
(1997) finds similar support for offensive realism in four case studies.
Lemke (2004) puts Mearsheimer’s variant of offensive realism to a
head-to-head test against power transition theory (Organski, 1958; Organski
and Kugler, 1980; Kugler and Organski, 1989; Kugler and Lemke, 1996).
Power transition theory asserts that states come into conflict when they
approach approximate power parity and are dissatisfied powers.8 Evaluating
the performance of offensive realism and power transition theory in the
post-Cold War era, Lemke (2004) finds “that offensive realism is largely
inconsistent with what the great powers have done in the past decade or
so.” In short, offensive realism fails to stand up to a test of its recent empiri-
cal record when compared to power transition theory.
We test the empirical accuracy of offensive realism using conflict data
from 1816 to 1992 and by comparing it to other theories, namely the territo-
rial explanation of war (Vasquez, 1993) and a norms-based perspective
(Wallensteen, 1984). Mearsheimer (2001, p. 54) himself concedes there will
be inconsistencies and anomalies in the theory, yet these problems are
much more persistent than he claims. If offensive realism is to be
considered a useful theory to guide action, it must be able to stand on its
own when tested independently, as well as be able to beat rival explana-
tions when put in a head-to-head test. The theory is of little value if it
passes neither test.
Of course all theories have difficulties explaining all cases, but this
theory seems to explain very few examples in the empirical record. Elman
(2004) finds that the theory needs to be extended and stretched to consider
the domestic and regional motivations of a state’s external action. This refor-
mulation of the theory limits its original intent since the entire point of the
theory seems to be to point out the strength of a structural-based approach.
Caveats are laced throughout the theory and book. Mearsheimer suggests
that failing to act like an offensive realist means the state has acted
“foolishly,” not that the theory has failed to predict action. This is an inade-
quate explanation for the failure of the theory to predict action, as is the
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
184 B. Valeriano
suggestion that other “fine grained theories” can supplement the theory
where it fails. The goal of this effort is not to falsify offensive realism, but to
evaluate the empirical accuracy of offensive realism and suggest some
opposing theoretical explanations that better explain the behavior of great
powers in the system. If the theory fails consistently, there is little that can
be done to supplement or reformulate the theory since at that point it
becomes degenerative and exhibits the traits of a theory saving enterprise
(Vasquez, 1997).
Are Great Powers the Only Important Actors?
According to offensive realism, the most important states in the system are
the great powers.9 It may also be suggested that the opposite may be true.
Minor powers could be just as conflict prone as great powers. The conflict-
ual behavior of minor powers could also support Mearsheimer’s theory in
that it is much more elegant than originally proposed since it contains
excess empirical content. However, minor powers do not (and cannot) seek
global or regional hegemony, which is key for the theory. Minor powers do
not take actions in relation to power concerns, but mainly in relation to
issue concerns such as the stability of borders, the flow of refugees, or
resource allocation. It would be useful to investigate the empirical validity
of the claim that major powers always seek power in international relations
when compared to the actions of minor powers, who seem to be historical
bystanders in the theoretical story.
Previously, the investigation of the relationship between arms races
and war was limited to the class of major power states that participated in
arms races (Wallace, 1979; Diehl, 1983; Sample, 1996; Sample, 1997;
Sample, 1998; Sample, 2000). Using a new dataset that includes minor and
major powers, Sample (2002) finds that even minor powers get into arms
races and the effect on the escalation to war is similar to the relationship
between major powers and war. Sample notes, “We have been mistaken in
simply assuming that major states represent the whole system.” (2002,
p. 670)
Some might contend that Mearsheimer’s theory has nothing to do with
the actions of minor powers; their actions should be of little consequence to
any evaluation of the theory. As one respondent put it, “it does not matter if
minor powers are more or less conflict prone because their wars have rela-
tively little impact on the international order.” We assert that minor powers
are important actors in the international system and their actions are rele-
vant for the international order. From the start of World War I, where the
local conflict between Serbia and Austria-Hungary spiraled into a global
war, to the United States’ recent conflicts with minor powers (Iraq and
Afghanistan), it is clear that minor powers are important actors in the system
and conflict with them can interrupt the international order. Furthermore,
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 185
any theory that attempts to explain the actions of major powers cannot do
so without accounting for the actions of major powers toward minor
powers. Explaining major power actions toward other major powers might
be a useful exercise, but state interactions are made up of more than dual
major power incidents.
Addiction to Conflict
Mearsheimer (2001, p. 3) suggests that states act aggressively because they
have the means and the will to change their status in the international sys-
tem. The “means” to change a state’s situation are military weapons.
Mearsheimer believes that great powers are addicted to conflict. Here we
will test this proposition to see exactly how addicted great powers are to
conflict. Are they really as conflict prone as Mearsheimer suggests? Does
simply having offensive weapons ensure that states will participate in
deadly conflict?
To test this proposition, a simple comparison will be made. We mea-
sure the rate of conflict for major powers when compared to minor powers
in the system. According to Mearsheimer, major powers should always seek
to ensure their survival through offensive action. Is there a substantial differ-
ence between the conflict rates of major powers and minor powers? It might
also be useful to see if the theory applies to different types of governments
or regime types. Are major power democracies more conflict prone than
major power autocracies? According to Mearsheimer, there should be no
difference in the rates of conflict between the two regimes since all major
powers are assumed to act alike in the system.
Rival Explanations: When Is Offensive Action Likely?
How powerful is Mearsheimer’s theory of offensive realism under empirical
testing? Offensive action may be a result of certain situations or other vari-
ables that would collude with state behavior to increase the probability of a
state taking offensive action. There are a few clues as to when this might be
the case. This article puts offensive realism up against a theory of norms
and territoriality as explanations for conflict involvement.10 This is not to
suggest that these theories solve the puzzle of great power action; only that
other, alternative theories might perform better than offensive realism when
compared head to head. We employ other nonrealist theories to see if they
can account for state action. The other theories identified here directly con-
tradict offensive realism and thus would pose a theoretical problem for the
theory if they were found to have express empirical content when
compared to the predictions of Mearsheimer (2001). All theories suggest
sufficient conditions of offensive action but the norms and territorial theo-
ries define different domains of operation when compared to offensive
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
186 B. Valeriano
theories. Scholars need to consider the scope and domain of their theories
(Vasquez and Valeriano, 2009). Since territorial (dyadic)-and norms
(systemic)-based theories attempt to explain conflict for all states in the
international system (not just the major powers), it could be said that they
are less likely to be able to predict the actions of major powers and thus
should not be likely to survive a head-to-head test. Some may say it is unfair
to put offensive realism to a direct test, but a useful and progressive theory
would explain more than any other theory under investigation (Lakatos,
1970). The usefulness of offensive realism is called into question if it fails in
any head-to-head test.
Wallensteen’s Periods of Peace
Wallensteen (1984) considers that norms in the system may have an impor-
tant impact on conflict involvement. He finds that there are periods in
which the rules of games are not established and states rely on unilateral
actions. He calls this period “particularist.” He also considers a period in
which states attempt to implement a set of rules to guide their interactions.
This is a “universalist” period. Wallensteen (1984) relied on historians “judg-
ments of historical time periods” to categorize these periods.11 Table 1 lists
the time periods and corresponding names if given.
States rarely become involved in major power wars if the norms in the
system are guided by a set of rules. An example of this type of period is the
Concert of Europe era, when major powers avoided war for over 50 years.
Wallensteen’s (1984) study shows that when the rules of the game in the
system are not agreed upon, states are more likely to enter into conflict.
This may be particularly true of major powers; they may only act offensively
during particularistic periods. It may not be the case that states always act in
a power politics fashion, but only when the system does not constrain the
actions a state can take. In these instances states may be provoked into
TABLE 1 1816–1976
Classification Historical label Time period No. of years
Universalist Concert of Europe 1816–1848 33
Particularist 1849–1870 22
Universalist Bismarck’s order 1871–1895 25
Particularist 1896–1918 23
Universalist League of Nations 1919–1932 14
Particularist 1933–1944 12
Particularist Cold War 1945–1962 18
Universalist Détente 1963–1976 14
(Wallensteen 1984).
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 187
taking offensive action as Mearsheimer recommends. There may in fact be
no “tragedy of great power politics” during universalist time periods.
Territorial Explanation of Conflict
What Mearsheimer fails to develop are the issues behind wars and conflict.
“Great powers that have no reason to fight each other—they are merely
concerned with their own survival—nevertheless have little choice but to
pursue power and to seek to dominate the other states in the system.”
(2001, p. 3) In his theory, issues do not account for war between dyads in
offensive realism. There is a constant will for states to seek power through
war, regardless of the issue at stake. Territorial, regime, and policy type dis-
putes should not be more likely to go to war than the mean probability of
war for all types of disputes.
Not all major powers seek hegemony, and not all major powers fight
over global-strategic issues.12 Utilizing an aggressive realist theory leads one
to overlook the issues at stake between a particular pair of states and also
the institutional constraints and characteristics of individual states that influ-
ence state action.
Interstate war almost always results from a dispute over an issue at
stake between a dyad. The trigger to wars is not the underlying structural
forces, alliance dynamics, or arms race patterns, but the issue over which
two states disagree. Mansbach and Vasquez (1981) conceive of an issue-
based approach to world politics in which it is not important where values
are allocated, but rather what values entities fight over.
Key to finding the cause of any war is investigating what the dyad is
fighting over. Territorial issues are the most common cause of war (Holsti,
1991; Vasquez, 1993; Vasquez and Henehan, 2001).13 Furthermore, territo-
rial disputes have the most salience for either member of the disputing party
and are least likely to be easily divisible. Hensel suggests that “territory is
often seen as highly salient for three reasons: its tangible contents or
attributes, its intangible or psychological value, and its effects on a state’s
reputation.” (Hensel, 2000, p. 58) Territoriality is important in that states
fight over territorial issues, not just because they are neighbors and con-
stantly interact, but because “the territorial explanation of war assumes that
human territoriality is a key to understanding much of interstate conflict and
war in the modern global system.” (Senese and Vasquez, 2003, p. 3)
It may be the case that states tend to act in an offensive fashion
when threatened with a territorial dispute by another state. Territorial
disputes are the most war-prone type of conflict (Vasquez and
Henehan, 2001). Thus, states could act offensively only in the face of an
impending territorial claim. This outcome, however, is not predicted in
offensive realist theory since issues do play a part in the decision to use
force.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
188 B. Valeriano
Of course territorial integrity is a key part of the national interest of a
state (Morgenthau, 1948). Here we do not argue that states fight to defend
their territory, but that issues of a territorial nature are the most conflict-
prone. States will fight over territorial issues, but not because of the unre-
lenting pursuit of power. The salience of territorial issues is what makes
them so deadly for a pair of states. Territorial claims even extend beyond
where the “stopping power of water” (Mearsheimer, 2001, p. 44) should
restrict conflict between states.
If either of these hypotheses (territoriality or norms-based) are correct,
then the theory of offensive realism would need reevaluation. One theory
(norms) challenges that the idea conflict is constant in the system, and the
other (territorial) suggests that conflict only results when certain issues are
under contention. It then may be suggested that it is not inherent in the
nature of states acting in an anarchic system to act offensively; rather that
states act offensively only if they are not constrained by norms in the system
or if they are in the midst of a territorial dispute. Offensive realism would
then require a territorial dispute or a system of loose norms as a necessary
condition. In the absence of these factors, offensive realism does not ade-
quately hold true. If these factors—loose norms in the system and territorial
issues—are necessary conditions for offensive action, each would represent
a better predictor of when states will use offensive force. More viable
options against which to test offensive realism may exist, but the goal here
is to only show that some other theories can better explain offensive action
than a lust for hegemonic standing.
HYPOTHESIS AND RESEARCH DESIGN
Falsifiability is an important requirement of social science research. While a
theory should not immediately be rejected once it is falsified, it does call
into question the whole enterprise. The problem with offensive realism is
that the theory does not lend itself to immediate testing. Lemke (2004) notes
that “it is so vaguely stated that history could always be re-interpreted as
consistent with it.” Some may disagree with our interpretations of Mearsheimer’s
arguments, but the theory is not logically consistent in many places. In
seeking a complete explanation of offensive action, Mearsheimer, at many
points, allows for caveats that could counter his original claims. Lee (2003)
points out that much like Waltz’s (1979) structural realism, the theory of
offensive realism does not specify time horizons, thereby potentially leading
to indefinite predictions. Scholars are still waiting for states to balance in the
post-Cold War world; will we end up waiting for states to drive toward
hegemony in the distant future?
We first seek to make the theory stand up on its own merit. Mearsheimer
(2001) may present an elegant and parsimonious theory, but is it empirically
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 189
accurate? He writes, “although it is depressing to realize that great powers
might think and act this way, it behooves us to see the world as it is, not as
we would like it to be.” (Mearsheimer, 2001, p. 4) Are the great powers in
the system really offensive realists at heart?
Seeking Power
Mearsheimer (2001, p. 13) proposes a variety of ways a state might seek
power. States use bloodletting, bait and bleed strategies, war, and blackmail
to acquire power. Each of these terms could be considered a threat to use
force or an actual use of force as measured by the Correlates of War Inter-
state Militarized Dispute dataset (Gochman and Maoz, 1984; Jones, Bremer
et al., 1996). Each involves either threatening action (blackmail) or offensive
options (bait and bleed, war, and bloodletting) that are measured through
the use of force. The question then becomes a matter of how often major
powers use threatening actions to acquire power.
To investigate the claim of offensive action, we can look at the rate of
conflict for major and minor powers.14 Major powers should use force often
and repeatedly.15 It is difficult to be sure what “often” can be quantified as,
but major powers should at least use the threat of force more often than
minor powers.
H1: Major powers should use force more often and at a greater hostility
level than minor powers do in any given year.
Offensive realism makes no distinction between democratic conflict
actions and nondemocratic conflict action since regime type does not mat-
ter. It should then be clear that pairs of democratic states use offensive
action at the same rate as nondemocratic dyads. Testing this prediction by
comparing the rate of conflict between dyadic democracies and nondemoc-
racies would be useful. Offensive realism would predict a non-liberal out-
come of constant conflict for both dyadic democratic and nondemocratic
states. Nondemocratic pairs of states should not become involved in more
conflict than democratic pairs of states.
H2: The rate of conflict in democratic major power status dyads should
be equal to or comparable to the rate of conflict for autocratic
major power status dyads.
Offensive states pursue a variety of strategies to gain power. The fore-
most of all is war and conflict. “Blackmail and war are the main strategies
that states employ to acquire power, and balancing and buck-passing are
the principal strategies that great powers use to maintain the distribution of
power when facing a dangerous rival.” (Mearsheimer, 2001, p. 13)
If states are always seeking power, they should continue to do so until
they actually gain power, otherwise the use of force would be an ineffective
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
190 B. Valeriano
strategy to ensure survival.16 Of course there may be states that execute mil-
itary strategies poorly, but if victory were to be a rare outcome then the
whole notion of offensive action and its efficacy would be called into
question—as would the foundations of the theory. Militarized disputes (a
simple and accurate way to measure military victory in relation to power
politics) with two major powers involved should repeat until the major
power gains a decisive outcome. One cannot gain power from a dispute
unless their party actually wins that dispute. It follows that war and disputes
would then not serve as the primary strategy to gain power if decisive out-
comes in these disputes were not the norm (militarized interstate disputes
include wars). The history of interactions between states will need to be
considered; therefore, here we focus on rival dyads. Rival dyads are those
pairs of states identified as long-standing historical enemies who constantly
fight about any issue as long as relative positions are considered when state
action is determined (Valeriano, 2003).
H3: Once a major power uses force or the threat of force against any
power, it should continue to do so until it achieves decisive victory.
As far as strategies intended to maintain power are concerned,
Mearsheimer only lists balancing and buck-passing as options (2001, p. 13).
Balancing behavior involves the act of acquiring power to reach a state of
equality with an adversary prior to or during war. An opposing theory is
bandwagoning, which entails joining with a threatening power in hopes of
maintaining one’s own power. Each predicts the direct opposite of the
other, according to offensive realist theory.
Bandwagoning (and appeasement) should not be valid options for
major powers. “Both of those strategies call for conceding power to an
aggressor, which violates balance-of-power logic and increases the danger
to the states that employ them. Great powers that care about their survival
should neither appease nor bandwagon with their adversaries.” (Mearsheimer,
2001, p. 162) If this is true, then great powers should not bandwagon dur-
ing war.
H4: Major powers will not show evidence of bandwagoning strategies.
Polarity
Mearsheimer (2001), along with Waltz (1979), considers the distribution of
power in the system to be an important structural variable. We will seek to
test the conflict-proneness of the different system structures. Mearsheimer
(2001, p. 337) suggests unbalanced bipolarity, balanced bipolarity, unbal-
anced multipolarity, and balanced multipolarity as outcomes. He does not
believe that unbalanced bipolarity is a useful real-world category, so this
consideration drops out of the typology.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 191
Mearsheimer (2001, p. 335) concludes that, “bipolar systems tend to be
the most peaceful, and unbalanced multipolar systems are more prone to
deadly conflict.” Here we test this proposition according to war and milita-
rized dispute involvement of major and minor powers from 1816 to 1992. Is
Mearsheimer correct in his ranking of the system structures relation to
peace?
In his original analysis, Mearsheimer only considered the European sys-
tem from 1792 to 1990. Table 2 lists the system configurations and years
involved according to Mearsheimer. He finds that bipolarity is the most
peaceful type of systemic distraction of power. There were no great power
wars during the period of bipolarity (the Cold War) and only 10,000 military
deaths. Balanced multipolarity is the next least-conflictual outcome, with
five major-major power wars resulting in 1.2 million military deaths. Finally,
unbalanced multipolarity is the more war-prone type of distribution in that
there was one major-major power war and three wars involving members of
the central system resulting in a total of 27 million military deaths. This anal-
ysis examines the global system and the relationship between polarity and
conflict involvement.
H5: Bipolarity, then balanced multipolarity, and finally unbalanced
multipolarity should be the most peaceful system to the most
war-prone, in rank order.
Rival Explanations
It next might be useful to put offensive realism to a head-to-head test with
other rival explanations. There are two prime candidates for this enterprise:
the territorial theory of conflict and a “norms”-based approach to conflict.
Using Wallensteen’s (1984) periods of peace argument, we will consider
whether or not offensive action varies according to the system of norms in
operation at the time of conflict. According to offensive realism, conflict
should remain consistent for major power dyads, no matter what period or
era. Peace should not be more prevalent during any given time period. Do
individual states’ decisions about the feasibility of offensive action dictate
when offensive power can be used, as Mearsheimer claims, or do the norms
in the system dictate the application of offensive power? If we were to find
that major power dyads are less likely to go to war or engage in militarized
disputes during certain periods of time, we would accept this as evidence
TABLE 2 Polarity Types and Years
Bipolarity (1945–1990)
Balanced Multipolarity (1815–1902, 1919–1938)
Unbalanced Multipolarity (1903–1918, 1939–1945)
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
192 B. Valeriano
that Mearsheimer fails to consider accurately when peace and war occur for
major power dyads.
H6: Major power dyads should maintain a constant rate of conflict and
war, no matter what historical era or system of norms they are
operating in during the conflict.17
The next head-to-head test involves the territorial explanation of con-
flict. It may be true that pairs of states only act in an offensive fashion when
threatened with a territorial dispute. To test this, we look at the probability
of conflict for major state dyads regarding territorial issues and other, non-
territorial issues. Are major power dyads more likely to fight a militarized
dispute or war when territory is in question, or is the probability of conflict
in regards to issue type inconsequential, as Mearsheimer suggests? If we
were to find that pairs of states only act offensively during disputes over a
territorial issue, that would lend support to an issue-based paradigm as
being more predictive than the offensive realist paradigm.
H7: Disputes involving major power dyads should be equally likely to
escalate, regardless of issue type.
RESULTS18
To test the various hypotheses outlined above, we relied mainly on Corre-
lates of War dispute and war data. Militarized Interstate Disputes (MIDs) are
defined as “united historical cases of conflict in which the threat, display, or
use of military force short of war by one member state is explicitly directed
toward the government, official representatives, official forces, property, or
territory of another state” (Jones, Bremer et al., 1996, p. 163). The intensity
scale ranges from threats, to use of force, to observed war with over 1000
battle deaths.
The Importance of Major Powers
There are two ways to look at hypothesis one. The first is to look at the
mean number of disputes for major powers and minor powers monadically
by year. Figure 1 was constructed by dividing the total number of major
powers in the system by the number of disputes in a given year. The
second line in the figure divides the total number of minor powers in the
system by the number of minor power disputes during a given year.
From Figure 1, we can see that the first part of this hypothesis is con-
firmed. Major powers are more likely to become involved in disputes than
minor powers. This result is surprising considering that major powers com-
prise the minority of states in the system. For example, in 1992, the number
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 193
of minor powers is greater than major powers of a ratio of 160 minors to
8 majors. Regardless of the number of major powers in the system, major
powers are more likely to be involved in disputes.
Figure 2 shows the results for the hypothesis regarding the level of
conflict for major and minor powers. Here we calculate the mean hostility
FIGURE 1 Mean number of major and minor power dispute involvements by year.
0369 12
1816 1836 1856 1876 1896 1916 1936 1956 1976 1992
year
major powers minor powers
FIGURE 2 Mean major and minor power level of hostilities by year.
1 2 3 4 5
1816 1836 1856 1876 1896 1916 1936 1956 1976 1992
year
major powers minor powers
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
194 B. Valeriano
level for major and minor powers in the system for each year. There is
really no noticeable difference between the level of hostility for major and
minor powers. As Figure 2 shows, some years the level of hostility is greater
from major powers; in others, the level of hostility is greater for minor
powers.
Figure 3 presents results for hypothesis two. Here we find that the rate
of conflict is indeed greater for nondemocratic major powers. While demo-
cratic major powers are likely to become involved in conflict some of the
time, nondemocratic states are more likely to take hostile actions against
other major powers.
Victory and Power in the International System
Hypothesis three relates the ability of major powers to achieve victorious
outcomes during disputes and wars. Offensive realist theory would suggest
that major powers would not stop a conflict until they reached a decisive
victory. Victory allows for conquest and tribute so that one nation may
increase its capabilities through the defeat of another power. If major pow-
ers truly are good offensive realists, they would not end a conflict unless at
least one side achieves victory. Consequently, victory should be the most
persistent settlement outcome for major power disputes if Mearsheimer is
correct.
Table 3 shows that major powers only reach victory in approximately
14 percent of their disputes. There are 1574 disputes involving major powers,
FIGURE 3 Mean number of major–major power dyad dispute involvements by regime type.
0.5 11.5 2
1816 1836 1856 1876 1896 1916 1936 1956 1976 1992
year
jointly democratic dyads non-jointly democratic dyads
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 195
but only a very small proportion of those conflicts end with a victory for the
major power. While these disputes may recur until one side reaches victory,
Table 3 shows that the overwhelming majority of conflicts do not achieve the
goals of major powers. Thus, victory by major powers is not a necessary char-
acteristic of major power disputes. If a low percentage of victory (14%) were
acceptable to states then the theory does not hold much value empirically.
Table 4 presents the results for victory according to recurring conflict or
rivalry, possibly a more accurate test of the victory hypothesis since disputes
are taken as historical units, not isolated events. According to Goertz and
Diehl (1992; 1993), the rivalry population will include isolated conflicts (1–2
disputes), proto-rivalries (3–5 disputes), and enduring rivalries (6+ disputes
over 20 years). Diehl and Goertz (2000) identify and produce a dataset that
accounts for all types of rivalry and the militarized disputes that correspond
with each rivalry.19
As the operational classification rules for isolated, proto, and enduring
rivalries are arrived at ex post, we seek to avoid this problem in our analysis
by categorizing the rivalries in an ex ante manner. That is, we consider the
first and second disputes in proto rivalry relationships as belonging to the
isolated rivalry stage. Likewise, the third to fifth disputes experienced by
enduring rivalries are analyzed together with those of proto rivalries. Only
the sixth and subsequent disputes between enduring rivals are considered
as disputes taking place in an enduring rivalry context.
TABLE 3 Major Power Disputes and Victory
Victory achieved? Major power disputes
No 1355 (86.09%)
Yes 219 (13.91%)
Total 1574
TABLE 4 Major Power Rivalries and Victory
Victory achieved? Major power disputes
Isolated Stage (1st and 2nd Disputes; Ex Ante Coding)
No 756 (84.66%)
Yes 137 (15.34%)
Total 893
Proto Stage (3rd to 5th Disputes; Ex Ante Coding)
No 261 (86.14%)
Yes 42 (13.86%)
Total 303
Enduring Stage (6th Dispute and Onwards; Ex Ante Coding)
No 338 (89.42%)
Yes 40 (10.58%)
Total 378
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
196 B. Valeriano
When classified according to dispute sequences, we find a decreasing
relationship between rivalry stage and victory outcomes. Over fifteen per-
cent, 15.3 to be exact, of the first and second disputes between all rival
dyads that involve at least one major power end in victory for the major
power. This percentage drops to 13.9 percent during the proto rivalry stage.
Finally, only 10.6 percent of the disputes occurring during the enduring rival
stage end in victory. This once again shows that disputes among major
powers are not likely to end in a victorious outcome, demonstrating that
disputes do not necessarily strengthen major power actors that seek to
enhance their capabilities through conquest and domination.20
Are Major Powers Likely to Balance?
We ascertain the relative frequency of balancing and bandwagoning behav-
ior by major power states during wars by using the Correlates of War
Project’s Interstate War Participants (1816–1992) and Composite Index of
National Capabilities (CINC) data sets. The former provides information
concerning when the major powers in question actually entered wars that
were already ongoing, and on which side they joined. Obviously, instances
in which major powers are one of the two initial belligerents do not count
as either balancing or bandwagoning observations because the two con-
cepts imply major power actions that either result in the formation or
expansion of warring coalitions beyond the initial disputants.
Using this information, the CINC data set is then used to calculate the
aggregate “power” of the two warring coalitions at the time of their joining.
If the major power joins the currently weaker side, it is coded as balancing;
if it joins the currently stronger side, then it is bandwagoning.
•Balancing: 16 (55.17%)
•Bandwagoning: 13 (44.83%)
•Total: 29
The results show that major power states have indeed bandwagoned
during wartime, and such occurrences cannot be said to be rare, since
bandwagoning constitutes close to forty-five percent of major power
entries into ongoing wars. A revised version of offensive realism would
have to account for the bandwagoning tendencies of major powers.
Realists typically assert that major powers should not bandwagon if their
main goal is security; however, any theory of offensive action should be
able to account for instances of bandwagoning behavior. In light of this
finding, it may be that states only act offensively and also balance when
the norms in the system are not restricting. States may bandwagon when
the norms in the system are agreed upon and offensive action is
discouraged.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 197
Periods of Polarity and Conflict
As mentioned earlier, Mearsheimer asserts that periods of bipolarity, bal-
anced multipolarity, and unbalanced multipolarity should be ranked from
least to most war prone, in the above order. He supports this assertion with
summary statistics on the number of wars, frequency of war years, and fatal-
ities, using a set of European wars (2001, p. 357). We test for the general
validity of his claim by extending the set of relevant wars to all wars that
have been fought in the system from 1816 to 1992, based upon the Corre-
lates of War Interstate War List.
The information presented in Tables 5 through 7 contest Mearsheimer’s
claim that bipolarity is the most peaceful (least war prone) system. In fact,
regardless of whether the number of wars, number of ongoing war years, or
total deaths is used, bipolarity is not the most peaceful system. On all three
measures, bipolarity is the second most war-prone system.
While unbalanced multipolarity continues to be the most deadly type
of system in terms of deaths and proportion of ongoing war years, it is not
necessarily the most war-prone in terms of total number of unique wars that
occurred. If one considers that minor powers also play an important role in
international interactions, balanced multipolar systems experience the most
frequent amount of war.
T
ABLE 5 Number of Wars by Type of Polarity
Minor-minor
wars
Major-major
wars
Major-minor
wars
All wars
Bipolarity 14 1 8 23
Balanced Multipolarity 18 5 21 44
Unbalanced Multipolarity 5 4 3 12
TABLE 6 Proportion of Ongoing War Years by Type of Polarity
Total years War years % of years with war
Bipolarity 46 32 69.6
Balanced Multipolarity 106 52 49.05
Unbalanced Multipolarity 23 21 91.3
TABLE 7 Total Deaths by Type of Polarity
Total deaths
Bipolarity 3.33 million
Balanced Multipolarity 2.78 million
Unbalanced Multipolarity 25.6 million
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
198 B. Valeriano
It appears that unbalanced multipolarity is the most dangerous type of
system in terms of total deaths and number of great power wars. Yet this
fact is only supported by the inclusion of World War I and II.21 The total
years covered for this period is less uniform than the rest of the categories
(unbalanced multipolarity includes 23 years while bipolarity and balanced
multipolarity include a total of 106 years). It seems that the category of unbal-
anced multipolarity was added to increase the strength of the prediction.
Norms and Constant Conflict?
To test hypothesis six, each unique dispute and war involving at least one
major power is categorized into the periods delineated by Wallensteen,
according to the year in which it began. Then, the number of disputes in
each period is divided by the length of the period to arrive at a measure of
average dispute occurrences per year, for that period (Table 8).
Particularist periods are marked by more occurrences of disputes
involving major powers, as compared to the universalist periods. This trend
holds even when we are reminded that over time, more newly independent
states have entered into the global system. Maoz (1989) finds that newly
independent nations are more likely to experience war and disputes.
Rank ordered, the particularist period of 1933–1944 has the highest rate
of dispute occurrence, followed immediately by the particularist period of
1945–1962. Compared to these two periods, the universalist period that
follows (1963–1976) contains a lower rate of dispute occurrence. For major
power disputes, there is no uniform pattern of nations constantly striving to
gain power through conflict.
We also see a similar pattern for war during Wallensteen’s periods. The
particularist periods from 1933–1944, 1896–1918, and 1849–1870 each expe-
rience the highest number of wars. The universalist periods from 1963–1976,
1919–1932, and 1816–1848 each experience less than five wars during the
period. These results show that during times when the international system
T
ABLE 8 Rate of Dispute Occurrences by Period
Time period Classification # Years # Unique disputes # Disputes per year # Wars
1816–1848 Universalist 33 43 1.3 4
1849–1870 Particularist 22 46 2.1 10
1871–1895 Universalist 25 53 2.1 4
1896–1918 Particularist 23 142 6.2 12
1919–1932 Universalist 14 58 4.1 4
1933–1944 Particularist 12 188 15.7 22
1945–1962 Particularist 18 177 9.8 5
1963–1976 Universalist 14 127 9.1 2
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 199
is dominated by major powers that agree on the rules of international
action, war is experienced less frequently.
Issues and Constant Conflict?
To test hypothesis seven, each unique dispute involving at least one major-
power state is categorized into territorial, policy, and regime issue types. We
also observe whether or not the dispute escalated into an interstate war
(Table 9).
The proportion of dispute escalations according to issue type is not
equal across the various issue categories. Eighteen percent of territorial dis-
putes escalate to war. Territorial disputes are more likely to escalate than
both policy and regime disputes.
This finding contrasts against the expectation of offensive realism, or
realism in general, which does not distinguish between the issues at stake in
interactions involving major powers. Regardless of polarity, we find that
major powers are more likely to experience war if there are territorial issues
at stake.
ASSESSMENT
The results for hypothesis one suggest Mearsheimer is correct to a point.
The rate of conflict for major powers is indeed greater than for minor pow-
ers. We also know that the level of hostility for major power conflict and
minor power conflict does not differ greatly. Results here do correspond
with Bremer’s (1992) finding that major powers are more war prone.
Mearsheimer is correct to hypothesize that major power conflict is more
likely to occur than minor power conflict.
Offensive realism suggests that conflict should be uniform, at least
among great powers. While Figure 1 does show that major powers are more
likely to engage in conflict than minor powers, Figure 2 shows that the level
of hostility between major and minor powers does not vary. This means that
major powers are engaged in more conflict than minor powers (probably
because of widespread interests and ability), but that the intensity of conflict
does not differ. Both engage in dangerous conflict with a high level of fatalities.
T
ABLE 9 Proportion of Dispute Escalations by Issue Type
# of escalations # of non-escalations Total # % of escalations
Territory 33 149 182 18.1
Policy 24 505 529 4.5
Regime 3 62 65 4.6
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
200 B. Valeriano
Each type of state can create devastating scenarios in the international sys-
tem by engaging in conflict, and possibly bring others into the conflict.
Mearsheimer’s version of offensive realism makes no distinction about
the type of government within a state. He believes that regardless of internal
politics, great powers are likely to actively seek hegemony through con-
quest. Figure 3 shows that nondemocratic major powers are more likely
than democratic major powers to become involved in a dispute during any
given year. Offensive realism is not correct: all major powers do not act
alike. We know that democracies do not fight each other (Ray, 1995; Ray,
1998; Bueno de Mesquita, Morrow, et al., 1999) and we also know that for
major powers, democratic states are less likely to act like offensive realists.
Hypothesis three suggests that major power disputes should be charac-
terized by victory outcomes. According to offensive realism, major powers
become involved in disputes to gain power and seek hegemony. If major
powers have not been able to achieve victory according to the goals of the
actions laid out prior to war, then it can be said that major powers do not
gain from conflict. We find that only fourteen percent of major power dis-
putes end in a victory outcome, thus major powers do not “gain” power
from conflict because their goals are not achieved. A majority of disputes
between states remain unresolved. Offensive realist theory would suggest
that states actively use disputes to gain power; however, this is not the case
for major power disputes. In the context of recurring disputes where each
fights repeatedly to achieve a positive outcome, this pattern again holds.
Major power disputes are not typically associated with positive settlement
outcomes and cannot contribute to the capabilities of major powers since
conquest or victory outcomes are so rare.
Hypothesis four indicates that major powers will not show evidence of
bandwagoning strategies when they decide to join an ongoing war. Our
results indicate that this is not the case. Close to half the instances of major
power war joining behavior can be classified as bandwagoning. The
motives behind this choice may vary, but the fact remains that major powers
do not balance in an offensive manner as predicted by offensive realist
theory.
In terms of hypothesis five, it does appear that unbalanced multipolar-
ity is the most dangerous systemic distribution of power. However, this fact
is only supported by the inclusion of the world wars in the category. In
terms of total wars, both bipolar and balanced multipolarity systems include
more wars with major versus minor powers. Unbalanced multipolarity
includes one less major power war than balanced multipolarity. This is true
only for the case of total deaths in the system. It seems that unbalanced
multipolarity is only capturing times when major powers have outstanding
issues with other major powers during a short period of time. It is not clear
that the typology of systems that Mearsheimer codes can actually predict the
number of wars during any period of time.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 201
Hypothesis six suggests that conflict among great powers should be
uniform, regardless of the periods Wallensteen delineates. Wallensteen’s
periods are taken from historians’ judgments regarding the norms in the
international system at various times. It seems that these periods more
accurately explain major power war than Mearsheimer’s polarity periods.
Here we also do not run into problems of selecting on the dependent
variable, which is the amount of war in the system. Wallensteen first
asked which years fall into which periods and then measured the amount
of war. Using updated Correlates of War data on disputes and war, we
find that Wallensteen’s periods accurately describe the course of great
power politics.
Finally, for hypothesis seven we predicted that major powers would be
likely to escalate to war in a uniform pattern. We, on the other hand, find
that certain issue types are more war-prone than others. This finding shows
that the territorial explanation of war (Vasquez, 1993) more accurately pre-
dicts the occurrences of major power war. Major powers are more likely to
go to war if they fight over territorial issues.
It should be clear that we believe that other opposing theories
surpass the propositions of offensive realism in providing a better
explanation of offensive actions by major powers. While a norms- or
issues-based approach alone may not provide completely adequate
explanations for all international action, these theories do provide better
explanations than those based on offensive realism. What might be more
feasible in the future is a combined approach that accounts for hege-
monic activities of major powers toward minor powers (offensive real-
ism), a regional-based power transition test (Lemke, 2002), territorial
issues, and the norms in the system. This combined model may be able
to provide an accurate picture of the true origins of the “tragedy of great
power politics.” In any case, it is clear that offensive realism alone can-
not predict the actions of major powers against other major or minor
powers.
The fact that offensive realism fails to accurately describe great
power behavior in the past does not bode well for the theory. While
major powers are more war and dispute-prone than minor powers, the
results here show that offensive realism does not accurately describe
major power dispute behavior. It is little use as a theory if it fails in this
fundamental test. Offensive realism could be reevaluated (James, 2009) to
incorporate the predictions and results from other theoretical models, but
it seems that this would be a degenerative step for the paradigm
(Vasquez, 1997). We find little support for the proposition that the blood
lust of hegemony-seeking major powers accounts for the incidence of
war and disputes at the systemic or dyadic level of analysis. Lemke
(2004) even finds that offensive realism fails to describe great power
behavior in the 1990s.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
202 B. Valeriano
CONCLUSION
Mearsheimer (2001, p. 36) asserts that little can be done to ameliorate the
security dilemma in world politics. However, we find that there are two
types of peaceful systems where major power war is experienced less fre-
quently. If the major powers in the system agreed on the normative rules
that guide their actions (Wallensteen, 1984), they are less likely to experi-
ence wars and disputes. We also know that if major powers do not fight
over territorial issues, they are less likely to go to war (Vasquez and Valeriano,
2008). These results show that two other theories, one norms-based
approach and an issue-based approach, accurately describe major power
conflict behavior more accurately than offensive realism. When put in a
head-to-head test, offensive realism fails. We have found that basic proposi-
tions outlining offensive realism are false and that two other theories
perform better. These two findings are solid enough for us to reject offen-
sive realism as a predictive and normative theory.
Offensive realism is meant to be both a normative and an empirically
accurate theory. Nevertheless, states do not act “like good offensive realists”
as Mearsheimer (2001, p. 12) suggests. In fact, during certain periods of
time, states do not act offensively at all. Under the universalist category,
major powers rarely experience war. There is no tragedy of great power
politics. The only tragedy is to ignore the findings derived from a norms- or
issue-based approach.
While the United States has failed to assert its hegemonic ambitions
against other great powers in the twenty first century, it has used aggres-
sive power to control minor powers (Afghanistan, Iraq, and Serbia).
Offensive realism seems to work where it should not, in the description of
major power behavior in regards to minor powers. It seems that the policy
advice ascertained from Mearsheimer (2001) is inaccurate and is rather
more precise in explaining how states act toward minor powers, rather
than other major powers. Accordingly, it could be said that great powers
are only primed for the offensive against minor powers and not each
other.
NOTES
1. For an early investigation of the role of realism in international relations theory see Vasquez
(1983; 1998).
2. Other offensive realist work includes Layne (1993, 2000), Labs (1997), and Zakaria (1998). In
this article, Mearsheimer (2001) is taken as the exemplar of the emerging research program.
3. See Levy (1983) for a description of who the great powers are and their war involvement in
since 1495.
4. Some suggest that offensive realism is really a theory about regional hegemony, but the goal of
global hegemony is clear throughout Mearsheimer’s (2001) book. Mearsheimer (2001, p. 2) states early
on, “their ultimate aim is to be the hegemon—that is, the only great power in the system.”
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 203
5. It should be noted that Mearsheimer’s theory seems to apply mainly to Western/United States
actions, yet the goals of explanation are clearly universal.
6. Mearsheimer also suggests that the causal mechanism in his theory is fear (2001, p. 32). This
factor is unlikely to be a mechanism for dyadic or systemic interactions.
7. Satisfaction plays an important role in other power politics theories such as the power transi-
tion theory (Organski and Kugler, 1980).
8. Conflict is defined as a difference of preferred bargaining outcomes.
9. The terms great powers and major powers are used interchangeably in this paper. The term
major power has been discussed extensively in the field (Small and Singer, 1982) and thus will be used
to test the term great power. The term great power is meant to signify strong and important actors in the
system but there is no clear empirical definition of just who is a great power. Appendix A identifies the
major powers in the system.
10. The norms-based theory operates at the systemic level, the same as offensive realism. The ter-
ritorial based theory operates at the dyadic level but seems to explain great power actions better than
offensive realism. Offensive realism can also be read as a dyadic theory in that it predicts whom states
will fight against.
11. While some might contend that Wallensteen’s periods are still open to interpretation, it is
important to remember that Wallensteen collected information from historians and did not pose his own
views on the coding of time periods. It is in this way that Wallensteen’s operationalization of norms in
the system can be considered reliable.
12. It would be interesting to test the regional hegemony implications of Mearsheimer’s (2001)
theory but this task is beyond the scope of this article.
13. Testing the territorial explanation of war has encountered various criticisms that contend it
does not account for the rise of territorial disputes in the first place (Lemke and Reed, 2001). Senese and
Vasquez (2003) find that territorial claims increase the probability of a MID occurring and then territorial MIDs
increase the probability of war. There is no selection effect at work in the study of territory and conflict.
14. Hypotheses 1 and 2 are logical extensions of offensive realism and are suitable tests of the the-
ory because if found true, they would directly counter the predictions that Mearsheimer (2001) lays out.
15. Coding of major powers is taken from Small and Singer (1982). Appendix A lists these cases.
16. It could be argued that states “calculate” their ability to win disputes and therefore might stop
action before victory is achieved due to strategic calculations about the balance of power and other
state’s reactions. Yet this proposition (Mearsheimer, 2001, p. 37) seems to only apply to calculations
before, not during, action, therefore it is important to ask if once offensive action is started, is it ever suc-
cessful in ensuring survival and maximizing power.
17. It should be noted that this statement does not apply to systems of polarity that were previ-
ously discussed, only that norms should not matter for major powers according to the theory of offen-
sive realism.
18. Replication data can be found at tigger.uic.edu/∼bvaler
19. Available at http://www.pol.uiuc.edu/faculty/diehl.html
20. It should be noted that major powers may enter into conflict with an expectation of victory but
still fail in their efforts. This is an interesting proposition, but there is no empirical evidence that can be
used to compare the expectations of actors versus the actual outcomes. Case studies may get at this
question, but cannot be conducted in the limited space available.
21. While the theory would be of little use if it did not include these wars in the most dangerous
category, there is no evidence to suggest that the current classification can predict these wars.
REFERENCES
Bremer, S. A. (1992). “Dangerous Dyads: Conditions Affecting the Likelihood of
Interstate War.” Journal of Conflict Resolution, Vol. 36, pp. 178–197.
Bueno de Mesquita, B., J. D. Morrow, R. Siverson, and A. Smith. (1999). “An Institu-
tional Explanation of Democratic Peace.” American Political Science Review,
Vol. 93, pp. 791–807.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
204 B. Valeriano
Diehl, P. (1983). “Arms Races and Escalation: A Closer Look.” Journal of Peace
Research, Vol. 20, No. 3, pp. 205–212.
Diehl, Paul and Gary Goertz (2000). War and Peace in International Rivalry, Ann
Arbor: University of Michigan Press.
Elman, C. (2004). “Extending Offensive Realism: The Louisiana Purchase and
America’s Rise to Regional Hegemony.” American Political Science Review,
Vol. 98, No. 4, pp. 563–576.
Gochman, C. and Z. Maoz (1984). “Militarized Interstate Disputes, 1816–1976: Pro-
cedures, Patterns, and Insights.” Journal of Conflict Resolution, Vol. 28
(December), pp. 585–616.
Goertz, G. and P. Diehl (1992). “The Empirical Importance of Enduring Rivalries.”
International Interactions, Vol. 18, No. 2, pp. 151–163.
Goertz, G. and P. Diehl (1993). “Enduring Rivalries: Theoretical Constraints and
Empirical Patterns.” International Studies Quarterly, Vol. 37, pp. 147–71.
Hensel, P. R. (2000). “Territory: Theory and Evidence on Geography and Conflict.”
In J. Vasquez, ed., What Do We Know About War? Lanham, MD: Rowman &
Littlefield.
Holsti, K. (1991). Peace and War: Armed Conflict and International Order, 1648–1989,
Cambridge: Cambridge University Press.
James, P. (2009). “Elaborating on Offensive Realism.” In Annette Freyberg-Inan,
Ewan Harrison, and Patrick James, eds., Rethinking Realism, Baltimore, MD:
Johns Hopkins University Press.
Jones, D., S. Bremer, and J. D. Singer. (1996). “Militarized Interstate Disputes, 1816–
1992: Rationale, Coding Rules, and Empirical Patterns.” Conflict Management
and Peace Science, Vol. 15(2), pp. 163–213.
Kugler, J. and E. Douglas Lemke eds. (1996). Parity and War: Evaluations and
Extensions of the War Ledger, Ann Arbor, MI: University of Michigan Press.
Kugler, J. and A. F. K. Organski (1989). “The Power Transition: A Retrospective and
Prospective Evaluation.” In M. Midlarsky, ed., Handbook of War Studies, Ann
Arbor, MI: University of Michigan Press.
Labs, E. (1997). “Beyond Victory: Offensive Realism and the Expansion of War
Aims.” Security Studies, Vol. 4 (Summer), pp. 1–49.
Lakatos, I. (1970). “Falsification and the Methodology of Scientific Research
Programmes.” In I. Lakatos and A. Musgrave, eds., Criticism and the Growth of
Knowledge, Cambridge, Cambridge University Press, pp. 91–196.
Layne, C. (1993). “The Unipolar Illusion: Why New Great Powers Will Rise.” Inter-
national Security, Vol. 17, No. 4, pp. 5–51.
Layne, C. (2000). “U.S. Hegemony and the Perpetuation of NATO.” Journal of Stra-
tegic Studies, Vol. 23, No. 3, pp. 59–91.
Lee, G. (2003). “To Be Long or Not To Be Long—That Is the Question.” Security
Studies, Vol. 12, No. 2, pp. 196–217.
Lemke, D. (2002). Regions of War and Peace, Cambridge: Cambridge University
Press.
Lemke, D. (2004). “Great Powers in the Post-Cold War World: A Power Transition
Perspective.” In T. V. Paul, James J. Wirtz, and Michel Fortmann, eds., Balance
of Power: Theory and Practice in the 21st Century, Stanford, CA: Stanford
University Press.
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
The Tragedy of Offensive Realism 205
Lemke, D. and W. Reed (2001). “War and Rivalry Among Great Powers.” American
Journal of Political Science, Vol. 45, pp. 457–469.
Levy, J. S. (1983). War in the Modern Great Power System, 1495–1975, Lexington,
KY: University Press of Kentucky.
Mansbach, R. W. and J. A. Vasquez (1981). In Search of Theory: A New Paradigm for
Global Politics, New York: Columbia University Press.
Maoz, Z. (1989). “Joining the Club of Nations: Political Development and Interna-
tional Conflict, 1816–1976.” International Studies Quarterly, Vol. 33 (June),
pp. 199–231.
Mearsheimer, J. (1990). “Back to the Future: Instability in Europe after the Cold
War.” International Security, Vol. 15, No. 1, pp. 5–56.
Mearsheimer, J. (1995). “The False Promise of International Institutions.” Interna-
tional Security, Vol. 19, No. 3, pp. 5–49.
Mearsheimer, J. (2001). Tragedy of Great Power Politics, New York: W.W. Norton &
Company.
Morgenthau, H. J. (1948). Politics Among Nations, New York: Alfred A. Knopf.
Organski, A. F. K. (1958). World Politics, New York: Alfred A. Knopf.
Organski, A. F. K. and J. Kugler (1980). The War Ledger, Chicago: University of
Chicago Press.
Ray, J. L. (1995). Democracy and International Conflict: An Evaluation of the
Democratic Peace Proposition, Columbia, SC: University of South Carolina
Press.
Ray, J. L. (1998). “Does Democracy Cause Peace?” Annual Review of Political
Science, Vol. 1, pp. 27–46.
Sample, S. (1996). Arms Races and the Escalation of Disputes to War. Ph.D. disserta-
tion, Vanderbilt University.
Sample, S. (1997). “Arms Races and Dispute Escalation: Resolving the Debate.”
Journal of Peace Research, Vol. 34, No. 1, pp. 7–22.
Sample, S. (1998). “Military Buildups, War, and Realpolitik: A Multivariate Model.”
Journal of Conflict Resolution, Vol. 42, No. 2, pp. 156–175.
Sample, S. G. (2000). “Military Buildups: Arming and War.” In J. A. Vasquez, What
Do We Know About War? Lanham, MD: Rowman and Littlefield Publishers,
pp. 167–196.
Sample, S. (2002). “The Outcomes of Military Buildups: Minor States vs. Major
Powers.” Journal of Peace Research, Vol. 39, No. 6, pp. 669–692.
Senese, P. D. and J. Vasquez (2003). “A Unified Explanation of Territorial Conflict:
Testing the Impact of Sampling Bias, 1919–1992.” International Studies Quar-
terly, Vol. 47, No. 3, pp. 275–298.
Small, M. and J. D. Singer (1982). Resort to Arms : International and Civil Wars,
1816–1980, Beverly Hills, CA: Sage Publications.
Valeriano, Brandon (2003). The Steps to Rivalry: Power Politics and Rivalry Forma-
tion, Ph.D. Dissertation, Vanderbilt University, August 2003.
Vasquez, J. A. (1983). The Power of Power Politics: A Critique, New Brunswick, NJ:
Rutgers University Press.
Vasquez, J. (1993). The War Puzzle, Cambridge: Cambridge University Press.
Vasquez, J. (1997). “The Realist Paradigm and Degenerative Versus Progressive
Research Programs: An Appraisal of Neotraditional Research on Waltz’s
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
206 B. Valeriano
Balancing Proposition.” American Political Science Review, Vol. 91 (Decem-
ber), pp. 899–912.
Vasquez, J. A. (1998). The Power of Power Politics: From Classical Realism to Neotra-
ditionalism. Cambridge: Cambridge University Press.
Vasquez, J. and M. Henehan (2001). “Territorial Disputes and the Probability of
War, 1816–1992.” Journal of Peace Research, Vol. 38, No. 2, pp. 123–138.
Vasquez, J. and B. Valeriano (2008). “Territory as a Source of Conflict and a Road to
Peace.” In I. William Zartman, Jacob Bercovich, and Victor Kremenyuk, eds.,
Sage Handbook of Conflict Resolution. London: Sage.
Vasquez, J. and B. Valeriano (2009). “Classification of Interstate Wars.” Manuscript.
Wallace, M. (1979). “Arms Races and Escalation: Some New Evidence.” Journal of
Conflict Resolution, Vol. 23, No. 1, pp. 3–16.
Wallensteen, P. (1984). “Universalism vs. Particularism: On the Limits of Major
Power Order.” Journal of Peace Research, Vol. 21, No. 3, pp. 243–257.
Waltz, K. (1979). Theory of International Politics, Reading, MA: Addison-Wesley.
Zakaria, F. (1998). From Wealth to Power: The Unusual Origins of America’s World
Role, Princeton, NJ: Princeton University Press.
APPENDIX A: MAJOR POWERS IN THE SYSTEM (1816–PRESENT)
United States 1898–
United Kingdom 1816–
France 1816–1940
France 1945–
Germany (Prussia) 1816–1918
Germany 1925–1945
Germany 1991–
Austria-Hungary 1816–1918
Italy 1860–1943
Russia 1816–1917
Russia 1922–
China 1950–
Japan 1895–1945
Japan 1991–
Downloaded By: [University of Illinois] At: 19:46 17 August 2009
