THE STATE LEVEL OF AGGREGATION

In this section, we consider some of the more interesting and relevant monadic level findings as to what it is that makes states and other social actors especially dispute- and war-prone.

Demographics

We begin with the Bremer, Singer, and Luterbacher (1973) study, in which they examine population, area, and war for the European state system from 1816 to 1965, and find that neither population nor population density are related to the frequency of war. While looking at the impact of population growth and density on involvement in international conflict initiation and escalation, Tir and Diehl (1998) find that the less developed a state, the more subject it is to population pressures and conflict compared to countries that are more developed and possess a higher level of technology.

Somewhat related is the lateral pressure argument, in which expanded industrializa­tion, along with population growth, will lead states to look abroad for resources, markets, and investment opportunities. As plausible as this may sound, especially vis-a-vis the advanced European states in the late nineteenth century, the findings are weak, except for Japan in the early twentieth century (Choucri and North, 1972). Overall, the empirical research in this area finds little evidence for the lateral pressure argument, and this is probably due to the effects of globalization and the opening of commodity markets, which has facilitated access to resources for states across the world.

Economic development and the business cycle

There has also been a fair amount of work trying to solve the war puzzle through investigating possible correlations between economic development and the business cycle with war. One of the earlier studies on the subject was by Richardson (1960), who finds no statistically significant correlation among levels of economic development and 300 deadly quarrels from 1820 to 1945.

Domestic stability

Counter to some standard arguments linking domestic instability and interstate conflict, the evidence is mixed as to whether or not there is a correlation. Searching foreign conflict vectors of 128 states over the years 1963 to 1967 for a possible relationship with a factor dimension of domestic turmoil, Vincent (1981) reports a statistically significant cor­relation between these two dimensions and concludes that the degree of a state's internal stability may be a useful predictor of its degree of foreign conflict behavior. In contrast, Geller (1985) examines a set of domestic and foreign conflict variables for 36 states between 1959 and 1968 and concludes that nations exhibiting high internal stability are more likely to engage in conflictual foreign policies than are nations with much less stability.

A more recent study by Leeds and Davis (1997) compares domestic political vulner­ability to international disputes by exam­ining the relationships between economic decline, the electoral cycle, and measures of aggressive international action for 18 advanced industrialized democracies during the period from 1952 to 1988. They found no consistent support for a relationship between constraining domestic political conditions and aggressive international behavior. The conflicting results can be partially attributed to the temporal spans that each study covers; while Geller and Vincent use rather short temporal spans, the study conducted by Leeds and Davis encompasses a somewhat longer period of 36 years.

Government centralization

A study conducted by the Cross-Polity Survey (1963) reports that the stronger the degree of executive leadership, the greater the tendency toward both diplomatic and violent foreign conflict.

Regime type: democracy and war

While most of the research on the effects of regime type focuses on the dyadic level of aggregation, it all started with the Cold War acolytes of Immanuel Kant. In 1961, the journal Industrial Research conducted a sur­vey among its readers, most of whom agreed that democracies are very peace-loving, while autocracies were more war-prone. The editors asked the COW team whether the historical evidence would support that, and in a rather simple study of the system since 1815, they found this not to be (Small and Singer, 1976). There were no significant differences between the frequencies with which the two regime types either initiated or participated in international wars. They did however find no wars in which there were democracies on both sides, but explained this as a historical- geographical artifact, given that there were very few democracies during that period, and moreover, even fewer were geographically contiguous - the most important of all variables in accounting for dyadic war. Here, continuing on the dyadic level of aggregation, we summarize the extent to which regime type affects the war-proneness of states while controlling for and interacting with other domestic variables, such as regime type.

Nonetheless, the argument that democra­cies do not go to war with each other has become an often discussed topic in the peace science literature; essentially, its advocates claim that democratic leaders are more influ­enced by their domestic setting (e.g. opposi­tion parties, constituencies, electoral cycles) at home, which then restrains them from going to war with other democracies.

A formal model developed by Bueno de Mesquita, Morrow, Siverson and Smith (1999) highlights the fundamental assumption of the democratic peace model; this is the idea that the highest priority of political leaders is to stay in power. This leads them to conclude that democratic leaders are more cautious about launching wars that they might lose. Adding to the study of the democratic peace argument, Brecher, James, and Wilkenfeld (2000), using data collected by the International Crisis Behavior Project (ICB), find that democracies are more inclined to go “all-out” once a conflict with a non- democratic adversary escalates.

More recently, Huth and Allee (2002), using a political accountability model tested against a data set of 348 territorial disputes for the period from 1919 to 1995, find that electoral cycles and the strength of opposition parties are important in explaining patterns of conflictual and cooperative behavior by democratic states. Specifically, they conclude that the timing of military confrontations by democracies is linked to electoral cycles, in that democratic leaders prefer to offer concessions in periods shortly after national elections. They also discover that other domestic-level variables, such as opposition parties, have an impact on the escalation of military confrontations as well. Using a two- stage probit model on a new data set of all leaders between 1919 and 1992, Chiozza and Goemans (2003) analyze the reciprocal relationship between the probability of losing office and the probability of crisis initiation.

Leblang and Chen (2003) disaggregate democracies by the details of their respective political system, such as parliamentary versus presidential, rule by a single dominant party versus a coalition government, and phases of the electoral cycle. They seek to explain the variations in war involvement among the established democracies on the basis of major differences in institutions of governance. They find that a country's electoral system turns out to be the most important insti­tutional factor in reducing involvement in war, and that established democracies with a proportional representation system tend to have significantly less such involvement. Their analysis brings into question the view that divided governments are better at reducing war involvement, and their results indicate no systematic difference between presidential and parliamentary forms of government, nor is there any difference in war involvement when states are governed by a single party or by a coalition of parties. But their analysis does show that a proportional representation system tends to be significantly less likely to engage in foreign belligerence.

Another element to consider is what role the origins of a state play in how conflict-prone a state becomes. Maoz and Abdolali (1989), using COW militarized dispute data for the years between 1816 and 1976, find a positive association between violent or revolution­ary state formation/transformation processes and militarized dispute involvement: states formed through non-violent or “evolutionary” means tend to be less conflict-prone in their international behavior. A somewhat related finding by Enterline (1998b) investigates 360 new political regimes and evaluates how a new political regime's institutional type and the political composition of its “geographic neighborhood” influences patterns of “intra­neighborhood militarized conflict.” He sug­gests that the homogeneity of regimes in a certain region, rather than their type, may play an important role in analyzing what makes states more war-prone.

Major powers

Major powers and their relation to war has been a topic often discussed, and on the monadic level the consensus does hold that there is positive relationship between major powers and conflict initiation, at least relative to minor powers. Using the origi­nal (1816-1965) and expanded (1816-1980) COW database, Small and Singer (1976, 1982) demonstrate that major powers are much more likely to engage in war than are minor powers. Bremer (1980), also using COW data, reports that nations that rank high on a composite index of national capability (CINC) are involved in a greater number of wars and initiate wars with greater frequency than do lower ranked states. Elberwein (1982) in a replication of Bremer adds to the finding that more powerful nations tend to use military force more frequently, and that power status alone accounts for over 60 percent of the variance in “joining” ongoing militarized interstate disputes.

Militarization

In terms of militarization and war-prone states, one of the earliest studies was by Feierabend and Feierabend (1969), who report a positive correlation between militarization and foreign conflict for the subset of highly developed states. Similarly, Weede (1970), using the DON database for the period 1955 to 1960 and defining “militarization” by the twin ratios of military personnel to total population and defense expenditures to GNP, notes a positive association between militarization and both verbal and violent foreign conflict behavior.

In examining the rate of change in military expenditure for a possible connection between the frequencies of dispute involvement and initiation, Diehl and Kingston (1987) con­clude that military buildups in major powers do not affect a state's tendency to initiate or to become involved in militarized international disputes. Also, Goldsmith (2003) finds that under conditions of economic growth or high levels of wealth, “extra” resources are diverted disproportionately to the military, but he does not find any linkage between military buildup and dispute initiation.

Capabilities and power cycles

Hoping that “power (capability) cycles” will help elucidate the correlation between the capabilities of a state and its involvement in interstate conflict, Doran and Parsons (1980) posit that certain critical points in a major power's cycle of increasing and decreasing capabilities (relative to the major­power system's capability pool) are likely to be associated with both the onset and the severity, duration, and magnitude of its wars. They maintain that states move through a general, cyclical pattern of capability growth, maturation, and decline. Four critical points on the evolutionary curve of a state are important because they present a disjuncture between a “state's interests and aspirations... and its actual capability” (Doran, 1983). Due to the shift in direction or rate of capability growth (lower turning point, rising inflection point, upper turning point, declining inflection point), the state must reevaluate its relative position, capability base, and foreign policy goals. He argues that the foreign policy stakes at these critical points are enormous - involving status, security, and power - and are therefore more likely to lead to war involvement. In short, the probability of war involvement increases for states passing through a critical point on the power cycle. For the initial study, capabilities are measured by an index composed of five material indicators, with the population inclusive of all major powers for the years between 1816 and 1975. They conclude that a major power's point on the power cycle is an important determinant of its probability of initiating a war, involvement in war, and of the characteristics of wars in which it engages, and also determine that 90 percent of major powers passing through a critical point on the power cycle are subsequently engaged in war.

To end the capabilities section, we cite one brief study on status quo orientation and its relationship to war; this was done by Geller (1994) who examines 43 rivalry wars between 1816 and 1986. He finds that challengers to the status quo initiate 30 of these, while defenders of the status quo initiate 13 of the preemptive or preventive type.

Environmental degradation

A new front opening in the peace science literature is concerned with the extent to which environmental degradation or resource maldistribution is associated with the onset of state-involved armed conflict. Many analyses test factors like deforestation, land degrada­tion, and scarce supply of freshwater, alone and in combination with high population density, then test what role these factors play in increasing the risk of armed conflict. Professor Lee at American University began a Trade Environment Database (TED) in September 1992, in which he has amassed over 700 case studies that among trade and cultural variables use environmental variables to explain the causes of war. One apparent result from many of his case studies is that armed conflict leads to resource scarcity and that resource scarcity leads to more conflict, thus there might exist something akin to a cyclical trap for nations facing these types of problems. The importance of bi-directional interplay was also stressed by Maxwell and Reuveny (2000), who argue that conflict due to “renewable resource scarcity” could be cyclical, which would then lead to recurring phases of conflict. However, when taking into account economic and political factors, such as level of economic development, Hauge and Ellingsen (1998) find that the environmental scarcity variables do not play a decisive a role in predicting the incidence of armed conflict. Percival and Homer-Dixon (1998), in analyzing the relationship between environmental scarcity and conflict in the case of South Africa, note that environmental scarcity emerges within a political, social, and economic context, and that it “interacts with many of these contextual factors to contribute to violence.”