The ubiquity and importance of war have made analyses of its causes a central concern of scholars for over two millennia. Many of the fundamental questions about the causes of war were raised by Thucydides in the fifth century b.c., but the vast amount of work on the topic since that time has produced ongoing debates instead of generally accepted answers. Studies of war can be divided into three broad categories (reviews of the literature using similar frameworks are provided by Waltz 1959; Bueno de Mesquita 1980; and Levy 1989). The first type takes the system as whole as the unit of analysis and focuses on how characteristics of the interstate system affect the frequency of war. States are the unit of analysis in the second type, which explores the relationships among the political, economic, and cultural features of particular states and the propensity of states to initiate wars. The third type analyzes war as an outcome of choices resulting from small group decision making.

Some debates focus on characteristics of the interstate system that are thought to increase or decrease the chance of war. Are wars more likely during a period of economic prosperity or one of economic contraction? Which is more likely to maintain peace, a balance of power in the international system or a situation in which one state is hegemonic? Has the increasing power of transnational organizations such as the United Nations changed the likelihood of war in the contemporary world?

Social scientists also disagree about the effects of political and economic factors within a state on the possibility of war. Does a capitalist economy make a state more or less likely to initiate wars? Do democratic states start wars less often than autocracies do? Is increasing nationalism likely to cause more wars? Is the ethnic composition within and between states an important determinant of war?

There is also no consensus on which model of individual decision making is most appropriate for the study of war. Is the decision to go to war based on a rational calculation of economic costs and benefits, or is it an irrational outcome of distortion in decision making in small groups and bureaucracies? Are wars based on nationalist, ethnic, or religious conflicts generated more by emotions or values than by rational choices?


Most studies of war that use the interstate system as the unit of analysis begin with assumptions from the ”realist” paradigm. States are seen as unitary actors in realist theories, and their actions are explained in terms of the structural characteristics of the system. The most important feature of the interstate system is that it is anarchic. Unlike politics within states, relations between states take place in a Hobbesian state of nature. Since an anarchic system is one in which all states constantly face actual or potential threats, their main goal is security. Security can be achieved in such a system only by maintaining power. In realist theories, the distribution of power in the interstate system is the main determinant of the frequency of war.

Although all realist theories agree on the importance of power distribution in determining war, they disagree about which types of power distributions make war more likely. Balance-of-power theories (Morgenthau 1967) suggest that an equal distribution of power in the system facilitates peace and that an unequal distribution leads to war. They argue that parity deters all states from aggression and that an unequal power distribution generally will result in the strong using force against the weak. When one state begins to gain a preponderance of power, a coalition of weaker states will from to maintain their security by blocking the further expansion of the powerful state. The coalitions that formed against Louis XIV, Napoleon, and Hitler seem to fit this pattern.

Hegemonic stability theory (Gilpin 1981) suggests exactly the opposite: that unequal power in the system produces peace while parity results in war. When one state has hegemony in the world system, it has both the incentive and the means to maintain order. It is not necessary for the most powerful state to fight wars, since its objectives can be achieved in less costly ways, and it is not rational for other states to challenge a state with overwhelming power. Gilpin notes that the periods of British and U.S. hegemony were relatively peaceful and that World Wars I and II occurred during intervening periods in which power was distributed more equally. Since balance-of-power and hegemonic stability theories seem to explain some but not all of the cases, what is needed is a theory specifying the conditions under which either parity or hegemony leads to war.

Balance-of-power and hegemonic stability arguments are not applicable to all wars, only those between great powers. A third attempt to explain great-power war is power transition theory (Organski 1968). This theory suggests that differential rates of economic growth create situations in which rising states rapidly catch up with the hegemonic state in the system and that this change in relative power leads to war. Organiski argues that the rising state will initiate a war to displace the hegemonic state. This final part of the argument is questionable, since it seems at least as plausible that the hegemonic state will initiate the war against the rising challenger to keep the small advantage it still has (Levy 1989, p. 253).

Debates about power transitions and hegemonic stability are of much more than theoretical interest in the contemporary world. Although the demise of the Soviet Union has left the United States as an unchallenged military hegemon, American economic superiority is being challenged by the European Union (EU) and emerging Asian states (Japan in the short run, perhaps China in the long run). If power transition and hegemonic stability theories are correct, this shift of economic power could lead to great-power wars in the near future. If the main challenge is from the EU (the most likely scenario), it will be interesting to see if the cultural heritage of cooperation between the United States and most of Europe will be sufficient to prevent the great-power war that some theories predict.

Another ongoing debate about systemic causes of war concerns the effects of long cycles of economic expansion and contraction. Some scholars argue that economic contraction increases the chance of war, since the increased scarcity of resources leads to more conflict. Others have suggested the opposite: Major wars are more common in periods of economic expansion because only then do states have the resources necessary to fight. Goldstein’s (1988) research suggests that economic expansion tends to increase the severity of great-power wars but that economic cycles have no effect on the frequency of war.

Two important changes in the last fifty years may make many systemic theories of war obsolete (or at least require major revisions). The first is technology. Throughout history, technological changes have determined the general nature of warfare. By far the most significant recent development has been the availability of nuclear weapons. Since the use of these weapons would result in ”mutually assured destruction,” they may have made war much less likely by making it irrational for both parties. Of course, the broadening proliferation of nuclear weapons raises serious problems, as does their existence in currently unstable states such as the Russian federation. A second technological change that may alter the nature of war is increasing dependence on computers. Although computers have increased the accuracy and precision of many types of military technology, they also leave the countries using them vulnerable to new kinds of attacks by ”hackers” who could not only disarm military operations but bring whole economies to a halt by disrupting the computer systems necessary for their operation.

The second significant change in the last half of the twentieth century has been the development and increasing power of transnational organizations such as the United Nations. Most theories of war begin with the assumption that the interstate system is anarchic, but this is no longer valid. If the military power of the United Nations continues to grow, that organization could become more and more effective at preventing wars and suppressing them quickly when they start. Of course, it remains to be seen whether powerful existing states will cede more power to such institutions.

Theoretical debates about the systemic causes of war have not been resolved, in part because the results of empirical research have been inconclusive. To take one example, equality of power in the interstate system decreased the number of wars in the nineteenth century and increased the number in the twentieth century. Proponents of each theory can point to specific cases that seem to fit its predictions, but they must admit that there are many cases it cannot explain. At least part of the problem is that systemic theories have not incorporated causal factors at lower levels of analysis, such as the internal economic and political characteristics of states. Since the effects of system-level factors on war are not direct but always are mediated by the internal political economy of states and the decisions made by individual leaders, complete theories of the causes of war must include these factors as well.


One of the longest and most heated debates about the causes of war concerns the effects of capitalism. Beginning with Adam Smith, liberal economists have argued that capitalism promotes peace. Marxists, by contrast, suggest that capitalism leads to frequent imperialist wars.

Liberal economic theories point to the wealth generated by laissez-faire capitalist economies, the interdependence produced by trade, and the death and destruction of assets caused by war. Since capitalism has increased both the benefits of peace (by increasing productivity and trade) and the costs of war (by producing new and better instruments of destruction), it is no longer rational for states to wage war. The long period of relative peace that followed the triumph of capitalism in the nineteenth century and the two world wars that came after the rise of protectionist barriers to free trade often are cited in support of liberal economic theories, but those facts can be explained by hegemonic stability theorists as a consequence of the rise and decline of British hegemony.

In contrast to the sanguine views of capitalism presented by liberal economic theories, Marxists argue that economic problems inherent in advanced capitalist economies create incentives for war. First, the high productivity of industrial capitalism and a limited home market resulting from the poverty of the working class result in chronic ”underconsumption” (Hobson [1902] 1954). Capitalists thus seek imperial expansion to control new markets for their goods. Second, Lenin ([1917] 1939) argued that capitalists fight imperialist wars to gain access to more raw materials and find more profitable outlets for their capital. These pressures lead first to wars between powerful capitalist states and weaker peripheral states and then to wars between great powers over which of them will get to exploit the periphery.

In contrast to the stress on the political causes (power and security) of war in most theories, the Marxist theory of imperialism has the virtue of drawing attention to economic causes. However, there are several problems with the economic causes posited in theories of imperialism. Like most Marxist arguments about politics, theories of imperialism assume that states are controlled directly or indirectly by dominant economic classes and thus that state policies reflect dominant class interests. Since states are often free of dominant class control and since many groups other than capitalists often influence state policies, it is simplistic to view war as a reflection of the interests of capitalists. Moreover, in light of the arguments made by liberal economists, it is far from clear that capitalists prefer war to other means of expanding markets and increasing profits.

With the increasing globalization of economies and the transition of more states to capitalist economies, the debates about the effects of capitalism, trade, and imperialism on war have become increasingly significant. If Adam Smith is right, the future is likely to be more peaceful than the past, but if Marxist theorists are right, there will be an unprecedented increase in economically based warfare.

The form of government in a country also may determine how often that country initiates wars. Kant ([1795] 1949) argued that democratic states (with constitutions and separation of powers) initiate wars less often than do autocratic states. This conclusion follows from an analysis of who pays the costs of war and who gets the benefits. Since citizens are required to pay for war with high taxes and their lives, they will rarely support war initiation. Rulers of states, by contrast, have much to gain from war and can pass on most of the costs to their subjects. Therefore, when decisions about war are made only by rulers (in autocracies), war will be frequent, and when citizens have more control of the decision (in democracies), peace generally will be the result.

Empirical research indicates that democratic states are less likely than are nondemocratic states to initiate wars, but the relationship is not strong (Levy 1989, p. 270). Perhaps one reason for the weakness of the relationship is that the assumption that citizens will oppose war initiation is not always correct. Many historical examples indicate that in at least some conditions citizens will support war even though it is not in their economic interest to do so. Nationalism, religion, ethnicity, and other cultural factors often are cited as important causes of particular wars in journalistic and historical accounts, but there still is no general theory of the conditions in which these factors modify or even override economic interests. Many classical sociological arguments suggested that these ”premodern” and ”irrational” sources of war would decline over time, but the late twentieth century has demonstrated the opposite. Nationalist and ethnic wars have become more common and intense. This raises the general issue of the factors affecting the choices individuals make about war initiation: Can these factors be modeled as rational maximization of interests, or is the process more complex?


Although the assumptions may be only implicit or undeveloped, all theories of war must contain some assumptions about individual decision making. However, few theories of war focus on the individual level of analysis. One notable exception is the rational-choice theory of war developed and tested by Bueno do Mesquita (1981).

Bueno de Mesquita begins by assuming that the decision to initiate war is made by a single dominant ruler who is a rational expected-utility maximizer. Utilities are defined in terms of state policies. Rulers fight wars to affect the policies of other states, essentially to make other states’ policies more similar to their interests. Rulers calculate the costs and benefits of initiating war and the probability of victory. War is initiated only when rulers expect a net gain from it.

This parsimonious set of assumptions has been used to generate several counterintuitive propositions. For example, common sense might suggest that states would fight their enemies and not their allies, but Bueno de Mesquita argues that war will be more common between allies than between enemies. Wars between allies are caused by actual or anticipated policy changes that threaten the existing relationship. The interventions of the United States in Latin America and of the Soviet Union in eastern Europe after World War II illustrate the process. Other counterintuitive propositions suggest that under some conditions a state may rationally choose to attack the stronger of two allied states instead of the weaker, and under some conditions it is rational for a state with no allies to initiate a war against a stronger state with allies (if the distance between the two is great, the weaker state will be unable to aid the stronger). Although these propositions and others derived from the theory have received strong empirical support, many have argued that the basic rational-choice assumptions of the theory are unrealistic and have rejected Bueno de Mesquita’s work on those grounds.

Other analyses of the decision to initiate war focus on how the social features of the decision-making process lead to deviations from rational choice. Allison (1971) notes that all political decisions are made within organizations and that this setting often influences the content of decisions. He argues that standard operating procedures and repertoires tend to limit the flexibility of decision makers and make it difficult to respond adequately to novel situations.Janis (1972) focuses on the small groups within political organizations (such as executives and their cabinet advisers) that actually make decisions about war. He suggests that the cohesiveness of these small groups often leads to a striving for unanimity that prevents a full debate about options and produces a premature consensus. Other scholars have discussed common misperceptions that distort decisions about war, such as the tendency to underestimate the capabilities of one’s enemies and overestimate one’s own. In spite of these promising studies, work on deviations from rational choice is just beginning, and there still is no general theoretical model of the decision to initiate war.


The failure to develop a convincing general theory of the causes of war has convinced some scholars that no such theory is possible, that all one can do is describe the causes of particular wars. This pessimistic conclusion is premature. The existing literature on the causes of war provides several fragments of a general theory, many of which have some empirical support. The goal of theory and research on war in the future will be to combine aspects of arguments at all three levels of analysis to create a general theory of the causes of war.

Next post:

Previous post: