Commentary 87
Scott Silverstone argues that political leaders often incorporate estimates
of the consequences of war for the postwar international order into their
calculations about whether or not to initiate a preventive war. Responding
to the common argument that Britain and France would have been better
off initiating a preventive war against Germany in 1936 rather than waiting
to fight three years later after a vigorous German rearmament program,
Silverstone argues that an important but overlooked reason for British leaders’
rejection of the military option in 1936 was their fear that military action
would not solve the security problems facing Britain and that it would
contribute to an unstable and illegitimate postwar political order.
Dale Copeland examines the Japanese decision for war against the
United States in 1941. He argues that the anticipation of economic and
military decline and future insecurity, fueled by depression-induced restrictive
trade policies that would cut off Japan from badly needed resources,
led Japanese decision makers to a desperate preventive war driven by
better-now-than-later and lesser-of-two-evils logic.
Preventive War: Concept and Propositions
JACK S. LEVY
Rutgers University
Preventive war is a state strategy to use military force to forestall an adverse
shift in the distribution of power between two states. It is driven by the perception
of a rising adversary, the anticipation of a decline in relative power,
and the fear of the consequences of decline, which include diminishing bargaining
leverage, the likelihood of escalating demands by an increasingly
powerful adversary, and the risk of war under worse circumstances later.
The logic of prevention is “better now than later”—it is better to fight now
and degrade the adversary’s capabilities while the opportunity is still available,
than to risk the consequences of continued decline. Specific conflicts
of interest at stake are secondary in the preventive path to war. The primary
issue is power.
Historians’ accounts reveal the role of preventive logic in many decisions
for war, and that logic is central in several realist theories of
international conflict, including balance of power theory (Morgenthau 1948),
hegemonic transition theory (Gilpin 1981), and dynamic differentials theory
(Copeland 2000). The theoretical importance of preventive war has
been reinforced by the formalization of the “commitment problem,” which
emphasizes the impediments to conflict resolution under conditions of
shifting power as one of only two or three paths to war for unitary rational
88 Commentary
actors (Fearon 1995; Powell 2006). My aim here is to build on my earlier
work (Levy 1987, 2008), provide a conceptual introduction for the
historically-based commentaries on preventive war in this issue, and suggest
promising directions for future research.
CONCEPTUAL ISSUES
Preventive war is one of many strategies based on better-now-than-later
logic, and it would be useful to differentiate among them. First, prevention
differs from preemption, which is a response to the anticipation of an
imminent attack and which is motivated by the goal of securing first-mover
advantages rather than of forestalling a negative shift in relative power.
Preemptors do not want war but feel that they have no choice, while preventers
want war in the short term to avoid the risk of war under less
favorable conditions in the long term. Preventers do not necessarily want
to initiate war, however, and they sometimes prefer to provoke war in an
attempt to shift the blame to their adversary and secure the diplomatic and
domestic political benefits of doing so. In 1914, for example, German leaders
wanted war to block the rising power of Russia but insisted on letting
Russia mobilize first (Fischer 1967).
An anticipated shift in relative power can trigger a preventive military
response even if it does not involve a complete power transition. Even limited
power shifts can result in an erosion of the stronger state’s bargaining
leverage, pose a threat to its interests, and trigger military action, either in
the form of a limited preventive strike or a preventive war. Limited power
shifts are most likely to trigger a preventive response if they cross an important
threshold of military power, the clearest example of which is the nuclear
threshold. Examples include Israeli strikes against the Iraqi nuclear reactor
in 1981 (Feldman 1982) and against a Syrian facility in 2007 (Follath and
Stark 2009), U.S. debates about the wisdom of a preventive strike against
North Korean nuclear facilities in 1994, and ongoing debates among Israeli
leaders about how best to respond to Iran’s likely development of a nuclear
capability. U.S. political leaders used popular fears of an Iraqi nuclear capability
as the primary rationales for wars against Iraq in 1991 and especially
in 2003 (Levy 2008).1
Although most of the literature on preventive war focuses on the dyadic
relationship between the rising and declining state, third parties can play an
1
Expectations that the adversary will cross a non-nuclear threshold can also trigger a preventive response.
Japan’s concerns about a step-level increase in Russian power projection in East Asia after its completion
of the trans-Siberian railroad contributed to Japan’s decision for war against Russia in 1904 (Nish 1985).
A limited preventive strike can escalate to all-out war if the target responds with force, the anticipation
of which presumably contributes to the initiator’s decision on prevention.
Commentary 89
important role. First, the rising power may pose a greater immediate threat
to a state’s allies than to the state itself, as illustrated by U.S. concerns about
Iraq’s development of a nuclear weapons capacity in 1990 and again in 2003.
Second, the source of the threat might not be a single adversary but instead a
combination of other states. German military leaders in 1914 were confident
that they could win a bilateral war against a rising Russia for years to come,
but they feared that by 1917 they might not be able to win a two-front war
against the Franco-Russian alliance.
I treat preventive war as a state strategy rather than as a kind of war.
The concept of a preventive war, though widely used, raises analytic problems
(Levy 2008:3). It confounds cause and effect in a single concept and
complicates the task of explaining outcomes, most of which have multiple
causes. It raises the question of how important the preventive motivation has
to be before we could call the resulting war a preventive war. The preventive
war label would be warranted if the preventive motivation is a sufficient
condition for war, but I can think of no empirical case that qualifies. The
preventive motivation may be a necessary condition for a particular war, but
the preventive war label would be misleading if there are other necessary
conditions for that outcome, as illustrated by Israel’s initiation of the 1956
Sinai War (Levy and Gochal 1956). It is better to refer to the preventive motivation
for war or to preventive logic as a causal variable or mechanism, or
to a preventive war strategy, rather than to preventive war as a type of war.
But the concept of preventive war is entrenched in the literature, and it does
facilitate an economy of language. When I use the concept of preventive
war I mean either a state strategy or a war for which the primary cause is
the preventive motivation.
I define preventive war fairly narrowly in terms of a military response
to an adverse shift in relative military power.2
Some define preventive war
more broadly. Renshon (2006:chap. 1), for example, defines prevention as
“an action . . . fought to forestall a grave national security threat,” which
can include the loss of status or prestige as well as a decline in relative
power. Schroeder (this issue) accepts the definition of preventive war as
military action driven by better-now-than-later logic to forestall an intolerable
future threat, but broadens the definition to include another source
of future threat—an anticipated breakdown in international order. In such
cases, Schroeder argues, the main goal of preventive action is not to destroy
or reduce an adversary’s military power but to “rescue, restore and stabilize
the threatened international order.” Political and legal theorists working
on “anticipatory self-defense” also adopt a broader definition of preventive
action (Doyle 2008:55–56). These broader definitions, I argue, include
2
States can adopt alternative strategies to forestall a decline in relative power, including covert action,
building up armaments, securing allies, other forms of containment, and economic revitalization. The
aim is the same, but preventive war refers only to forceful military responses to military threats.
90 Commentary
too many things under the same conceptual umbrella and impede a more
discriminating assessment of causation.
HYPOTHESES ON PREVENTIVE WAR
States do not always respond to the anticipation of a negative power shift
with a strategy of preventive war, as illustrated by the power transition
between Britain and the United States at the end of the nineteenth century
and by the absence of an American military response to block the Soviet
Union and then China from developing a nuclear capability during the Cold
War. The critical question is the conditions under which a state in relative
decline is most likely to adopt a strategy of preventive war, or when the
preventive motivation for war is the strongest. The literature is characterized
more by a set of bilateral hypotheses than by a well-developed theory
(Ripsman and Levy 2007; Van Evera 1999), though Copeland (2000) develops
a more fully integrated theory. I begin with the costs, risks, and benefits
of delay, because it is the fear of the future that drives preventive logic. I
then turn to the costs and risks of a preventive war now.
Anything that increases the future military threat posed by the rising
adversary increases the incentives for a state in relative decline to adopt a
preventive war strategy. The logic is that the greater the adversary’s future
advantage, the greater is its likely margin of victory in a future war and its
future bargaining leverage, and the greater the concessions the declining
state would have to make to avoid such a war. Thus the greater the magnitude
of the anticipated power shift, the greater the incentives for war now,
and complete power transitions are more war-prone than are more limited
power shifts. In addition, expectations of a permanent power shift are more
likely to trigger a preventive response than are expectations that the shift will
be temporary.3
The speed of the power shift might also be important. In the
context of uncertainty about the extent of rising state’s future advantage,
leaders of the declining state use the speed of the power shift as a proxy for
both the likelihood of a power transition and the adversary’s ultimate margin
of advantage. A rapid power shift also shortens the time the declining state
has to increase its own power, gain allies, or seek an accommodation with
its rival, which narrows the range of alternative strategies and increases the
likelihood of a military response.
The shift in relative power within a dyad takes place within a systemlevel
context, and the distribution of power in the international system can
influence decisions for preventive war. A shift in power between two states
3
Britain’s anticipation in the mid-1930s that the power shift in Germany’s favor would be temporary
rather than permanent was the critical factor contributing to the British strategy of appeasement against
Germany (Ripsman and Levy 2007).
Commentary 91
in a bipolar system is more likely to lead to a strategy of prevention than is a
comparable shift in a multipolar system, for two reasons. A declining state in
a multipolar system is more likely than one in a bipolar system to find allies
to deter a future conflict with a then-stronger adversary. In addition, a state
taking preventive action against a rising adversary risks weakening itself as
well as its adversary. This increases its vulnerability to another power, which
is more serious for a leading power in a multipolar system than in a bipolar
system, where there is a greater gap between the two leading states and
others (Copeland 2000).
Although preventive war strategies are driven primarily by an anticipated
shift in power, they are also influenced by expectations of the
adversary’s future intentions. The probability of a preventive response to
an anticipation of an adverse power shift increases with the expectation that
the adversary will try to exploit its stronger position by going to war or
using the threat of war to extract significant concessions, whether because
of an ongoing rivalry, an unresolved territorial dispute (Vasquez 2009), or
other domestic or international pressures. On the other hand, the incentives
for prevention are reduced if the declining state has alternative strategies
to provide for its future security, including securing allies against the rising
power, building up armaments or securing armaments from others, or economic
revitalization that increases a state’s future military potential. Israeli
leaders’ incentives for a preventive military response to the 1955 Czech arms
sales to Egypt were enhanced by their belief that a “second round” of war
was inevitable and by their failure to secure arms from France or the United
States (Levy and Gochal 2001–2002).
The costs and risks of inaction in response to a rising adversary can be
quite substantial, but so can the costs and risks of a preventive war strategy.
As Bismarck said, “preventive war is like suicide for fear of death.” The
greater the expected probability of winning a war now, and the lower the
anticipated costs, the greater is the strength of the preventive motivation
for war.4
The probability of winning is shaped by the behavior of third
states as well as by the dyadic distribution of power, so that expectations
of the behavior and military effectiveness of potential allies and adversaries
may be a critical factor influencing the strength of the preventive motivation
for war.
In decisions for a preventive strike against an adversary that threatens
to cross a critical threshold of military power, key variables are the probability
and costs of an adversary response, the likelihood that a limited strike
will effectively degrade the adversary’s capabilities and prevent the steplevel
increase in power, and the time before the adversary could rebuild
its military capabilities. If the adversary has the capacity to rebuild within a
4
This is why weaker states rarely adopt preventive war strategies against stronger adversaries whose
advantage is increasing. An important exception is Japan in 1941, as Copeland argues in his commentary.
92 Commentary
few years, the benefits of a preventive strike may be too temporary to be
worth the risk. Preventive strikes may also have reputational effects, though
when they deter other potential challengers and when they create incentives
for other states to build up their arms is a question that has yet to be
investigated.
The costs and risks of preventive war go beyond calculations of likely
battlefield outcomes. Schroeder (this issue) and Silverstone (this issue) each
emphasizes the consequences of preventive wars for the normative international
order that underpins the security of states. In addition, concerns
about the domestic legitimacy of a preventive war strategy sometimes constrain
states, especially democratic states (Schweller 1992). Silverstone (2010)
argues that an important factor contributing the British and French decisions
not to initiate a preventive war against Germany in the 1936 Rhineland
Crisis was their fear that a preventive war, even if successful, would both
undermine the domestic legitimacy of (and unity behind) the war effort and
contribute to an unstable and illegitimate postwar political order. This is
an important line of argument that requires further empirical validation—
both in the 1936 case and in other decisions on whether or not to adopt a
preventive war strategy.
Domestic politics influence decisions for preventive war in other ways.
Internal social, political, and economic changes may be a major source of
the relative decline in military power and potential that creates the incentives
for preventive war. Regime type may also be important. Although the
argument that democracies never fight preventive wars, especially against
states of comparable strength (Schweller 1992), is too strong (Levy 2008), the
probabilistic hypothesis that democracies are more constrained in adopting
preventive war strategies than are other states is certainly plausible and
worthy of empirical test. Democratic political cultures may be opposed
to preventive military action because they believe that it is morally unacceptable
or contrary to the identity of democratic peoples. Thus Brodie
(1959:237–239) argued that “war is generally unpopular [with the American
people] and the public mood inclines to support really bold action only in
response to great anger or great fright. The fright must be something more
than a sudden new rise in [the adversary’s] capability.” The U.S. experience
suggests, however, that cultural attitudes toward preventive war vary over
time (Silverstone 2007).
In addition, democratic leaders’ electoral accountability may induce
shorter time horizons, leading them to discount the long-term costs of inaction
in response to a rising adversary. Political leaders must bear any political
costs of a war fought now—and there is strong evidence that democratic
leaders are more likely than their authoritarian counterparts to be deposed
after a losing war effort—whereas they can pass on the costs of delay to their
successors. Brodie (1973:26) argues that the “willingness to gamble now at
Commentary 93
unlimited stakes for what is a highly speculative long-term gain” is “normally
most uncharacteristic of politicians.”
Brodie’s comment raises several issues relating to risk, uncertainty, and
time horizons. Preventive war, like any war, is a gamble. But so is inaction
in the face of relative decline, given uncertainties as to whether, and
how far, one’s power position will continue to decline, the adversary’s
future intentions, one’s ability to secure diplomatic support or to appease
the adversary successfully, and the likelihood and outcome of a future war.
In addition, preventive war and continued decline each involve additional
domestic uncertainties. How leaders will balance these uncertainties is often
dependent on their individual world views, personalities, and risk propensities
(Renshon 2006), which are difficult to measure, even in a single historical
case.
Still, research in social psychology and behavioral economics reveals
interesting patterns of behavior that might shape decisions under conditions
of relative decline. There is evidence that individuals tend to be risk acceptant
in the domain of losses and that political leaders tend to take excessive
risks to maintain their current positions (Levy 2000), and that political leaders
in international crises often conclude that all of their options are bad ones.
This implies that preventive war strategies under conditions of decline might
be more appealing than Brodie suggests. In addition, evidence suggesting
that people discount future gains more than they do future losses (Streich
and Levy 2007) increases the relative weight of future losses from continued
decline and thus increases incentives for preventive war. A good example of
a risk acceptant decision for war driven by expectations of declining power,
a bleak future, and diminishing opportunities is Japan’s attack against the
United States in 1941 (Copeland, this issue; Taliaferro 2004).
CONCLUSION
Historians have illuminated the role of preventive logic in decisions for many
wars for nearly three millennia, and recent research by political scientists has
advanced our theoretical understanding of the phenomenon of preventive
war. But much remains to be done. Here I highlight some of the many
potentially fruitful avenues for future research.
One unexplored question is the kinds of power shifts that induce the
greatest fear and that are most likely to lead to preventive war strategies. Are
states in relative decline most concerned about increases in the adversary’s
current relative military capabilities or economic strength, or about technological
breakthroughs or demographic shifts? Although scholars speak in
very general terms about power shifts, and although formal models of the
process (Powell 2006) are based on undifferentiated conceptions of power,
94 Commentary
there is reason to believe that a unidimensional conception of power cannot
capture the complexity of the link between power shifts and war. Germany
in 1914 was most worried about Russian military forces in being by 1917,
whereas Japan in 1941 was most troubled by its shortages of resources and
its inability to keep up with American economic power. In the mid-1930s,
France most feared rapid increases in German army strength, whereas Britain
was consumed by the German air threat. States sometimes face increasing
threats on multiple dimensions of power simultaneously. In 1904 Japan
feared the rising power of Russia along three dimensions: the completion of
the trans-Siberian railroad, the significant expansion of its Far Eastern fleet,
and the ongoing fortification of Port Arthur (Nish 1985).
The hypothesis that preventive war strategies can be triggered by shifts
in different dimensions of military power suggests that we need more
empirical research on the question of what kinds of power shifts are most
likely to trigger preventive war strategies by what kinds of states and under
what conditions. It also suggests that statistical studies of power shifts and
war (Lemke 2003) should use multiple indicators of power (for example,
the Correlates of War Project’s distinct military, economic, and demographic
indicators of national capabilities) rather than a single aggregate index.
Preventive war strategies are triggered not only by gradual shifts in
power—which are well-captured by measures of the size of armies, navies,
economies, or populations—but also by step-level increases in military
strength resulting from arms sales, technological breakthroughs (nuclear and
otherwise), or other less measurable elements of military power. Historical
examples of anticipated step-level shifts in power that generated a preventive
motivation for war include the Russian completion of the trans-Siberian
railway in 1904, the Russian completion of its army reforms and railroad
modernization by 1917, the Czech arms sales to Egypt in 1955, and the Iraqi
nuclear program. None of these historical processes that arguably triggered
preventive war strategies is adequately captured by the Correlates of War
indicators, though arms sales are certainly measurable. This suggests the
limitations of a statistical analysis of the link between power shifts and war
that does not incorporate indicators of step-level changes in relative military
capabilities.
Another useful direction for future research concerns bargaining and
strategic interaction. Most historical case studies of preventive war focus
on the preventer and neglect the perceptions and strategies of their rising
adversaries. Do they fear being the target of a preventive war or strike?5
Do they consider preempting the preventer? Do they attempt to appease
their adversary, adopting a strategy of “buying time” until the ongoing shift
in power puts them in a stronger position? Analysts pay some attention
5
German leaders prior to World War I feared a repeat of Britain’s “Copenhagening” of the Danish fleet
in 1807.
Commentary 95
to these questions for aspiring nuclear powers in the contemporary era,
but the question applies equally well to more distant historical cases. If
German leaders perceived a strong preventive motivation for war in 1914,
why did Russian leaders adopt a confrontational stance toward Germany
rather than buy time until the Russian war machine was ready and Germany
was vulnerable? Did one side or the other misperceive the changing power
relationship? Did they focus on different components of military power?
Or were strategic assessments dominated by domestic political calculations?
These and other questions provide fertile ground for future research in the
theory and practice of preventive war.
REFERENCES
Brodie, Bernard. (1959) Strategy in the Missle Age. Princeton: Princeton University
Press.
Brodie, Bernard. (1973) War and Politics. New York: Macmillan.
Copeland, Dale C. (2000) The Origins of Major War. Ithaca: Cornell University Press.
Doyle, Michael (2008) Striking First: Preemption and Prevention in International
Conflict, edited by Stephen Macedo. Princeton: Princeton University Press.
Fearon, James D. (1995) Rationalist Explanations for War. International
Organization 49(3):379–414.
Feldman, Shai. (1982) The Bombing of Osiraq—Revisited. International Security
7(2):114–142.
Fischer, Fritz. (1967) Germany’s Aims in the First World War. New York: W.W.
Norton.
Follath, Erich, and Holger Stark. (2009) The Story of “Operation Orchard”: How
Israel Destroyed Syria’s Al Kibar Nuclear Reactor. Spiegel Online International,
11/2/2009.
Gilpin, Robert (1981) War & Change in World Politics. New York: Cambridge
University Press.
Lemke, Douglas. (2003) Investigating the Preventive Motive for War. International
Interactions 29(4):273–292.
Levy, Jack S. (1987) Declining Power and the Preventive Motivation for War. World
Politics 40(1):82–107.
Levy, Jack S. (2000) The Implications of Framing and Loss Aversion for International
Conflict. In Handbook of War Studies II, edited by M.I. Midlarsky. Ann Arbor,
MI: University of Michigan Press, pp. 193–221.
Levy, Jack S. (2008) Preventive War and Democratic Politics. International Studies
Quarterly 52(1):1–24.
Levy, Jack S., and Joseph R. Gochal. (2001–2002) Democracy and Preventive War:
Israel and the 1956 Sinai Campaign. Security Studies 11(2):1–49.
Morgenthau, Hans J. (1948) Politics Among Nations: The Struggle for Power and
Peace. New York: Knopf.
Nish, Ian. (1985) The Origins of the Russo-Japanese War. New York: Longman.
Powell, Robert. (2006) War as a Commitment Problem. International Organization
60(1):169–204.
96 Commentary
Renshon, Jonathan B. (2006) Why Leaders Choose War: The Psychology of Prevention.
Westport, CT: Praeger.
Ripsman, Norrin M., and Jack S. Levy (2007) The Preventive War that Never
Happened: Britain, France, and the Rise of Germany in the 1930s. Security
Studies 16(1):32–67.
Schweller, Randall L. (1992) Domestic Structure and Preventive War: Are
Democracies More Pacific? World Politics 44(2):235–269.
Silverstone, Scott A. (2007) Preventive War and American Democracy. New York:
Routledge.
Silverstone, Scott A. (2010) The Legacy of Coercive Peace-Building: The Locarno
Treaty and the 1936 Rhineland Crisis. In Grand Strategy during the Interwar
Years, edited by Jeffrey W. Taliaferro, Norrin M. Ripsman, and Steven Lobell.
Unpublished book manuscript.
Streich, Philip, and Jack S. Levy (2007) Time Horizons, Discounting, and
Intertemporal Choice. Journal of Conflict Resolution 51(2):199–226.
Taliaferro, Jeffrey W. (2004) Balancing Risks: Great Power Intervention in the
Periphery. Ithaca: Cornell University Press.
Van Evera, Stephen Van. (1999) Causes of War: Power and the Roots of Conflict.
Ithaca: Cornell University Press.
Vasquez, John A. (2009) The War Puzzle Revisited. New York: Cambridge University
Press.
Preventive Wars to Restore and Stabilize
the International System
PAUL W. SCHROEDER
University of Illinois
This essay draws on history for reflections on the causes, purposes and
effects of preventive wars, treating international politics, as historians usually
do, as purposive conduct rather than behavior and emphasizing motives,
assumptions, strategies, goals, and individual and collective mindsets.
I accept Jack Levy’s definition of preventive war, especially the strategic
logic (“better now than later”) and the reliance on military action to
meet a threat considered sure to become intolerable and unmanageable
later. The definition seems to assume, however, that preventive wars arise
only from such perceived military-strategic threats and aim to meet them
essentially by defeating the adversary and restoring a desired distribution of
power. I will argue that some preventive wars have been launched primarily
because states perceive the main source of their insecurity less in direct
military threat than in a breakdown in international order leading to intolerable
uncertainty and lawlessness in the international system, and primarily
aim by preventive war not to destroy their opponent’s military power but to
restore and stabilize the threatened international order.