TERRITORY AND CONFLICT: EMPIRICAL PATTERNS

As a field, international relations has relied on theory to identify the conditions under which conflict resolution is most apt to be successful (Zartman, 1989) and the kinds of techniques that can be implemented to deal with specific problems (Pruitt and Rubin, 1986; Kriesberg, Northrup et al., 1989; Burton, 1990).

Conflict at the most basic level involves disagreement, and disagreement is inherent in social behavior. Not all disagreements need end in violence, however. Most practitioners of conflict resolution become concerned when disagreements are handled with the use of militarized force. Crossing this threshold puts the conflict into a different category, one where the risk of war has suddenly increased, even though most interstate attempts to handle issues through the use of force do not escalate to war. Thanks to the Correlates of War project, we now have a fairly complete record of all instances of the threat or use of force between legally recognized nation-states from 1816-2001. Such instances are called militarized interstate disputes (MIDs).

Using MIDs as the dependent variable, it has been found that states with territorial disagreements are more apt to have a MID than states without territorial disagreements (Senese and Vasquez, 2003).

A second factor that is related to the threat or use of force between states is whether they are neighbors. States that are contiguous (by land or within 150 miles of water) are more apt to have a MID than non-contiguous states (Senese, 2005). It is reasonable to assume, as the territorial explanation of war would expect, that disputes between neighbors involve territory, although this hypothesis has not been fully tested. What tests we do have consistently show that neighbors fight not because they are contiguous and have frequent interactions, but because they have territorial disputes (Hensel, 2000; Vasquez, 2001; Ben-Yehuda, 2004; Senese, 2005). This means that territory is more important than contiguity in terms of the onset of war.

Once territorial disputes emerge between two states, whether they are neighbors or not, they are more apt to recur (Hensel, 1994). We also know that states that have territorial disputes are likely to become enduring rivals (have six or more MIDs within a 20-year period) (Vasquez and Leskiw, 2001).⁴ From research on rivalry and protracted conflict, we know that disputes that recur between the same two states have a greater risk of escalating to war (Goertz and Diehl, 1992a; Brecher and Wilkenfeld, 1997; Diehl and Goertz, 2000). Since territorial disputes recur, we would expect that they would have a higher probability of going to war than other types of disputes.

In fact, the major finding we have in conflict studies is that territorial disputes have a greater probability of ending up in war than other types of disputes, specifically regime and policy disputes. This was one of the first findings on territory and war using the MID data and it has been confirmed in several studies. The simplest and most straightfor­ward test is in Vasquez and Henehan (2001: 128). They find that territorial disputes have the highest probability of going to war in the 1816-1992 period (.091) compared to regime disputes (.078) and policy disputes (.033).

Vasquez and Henehan (2001:134-135) also test this hypothesis controlling for historical era (1816-1945,1946-1992) and whether the two states in a dispute are both major states (e.g. Germany, UK), both minor, or major­minor. These controls generally reconfirm the findings. The only exceptions are that under certain circumstances, regime disputes have a higher probability of war when both sides are minor states or when the dispute occurs in thepost-1945 period. Nevertheless, territorial disputes account for most war escalations— 53 of 97 (54.6%) compared to only 9 of 97 (9.3%) for regime disputes for the entire 1816-1992 period (Vasquez and Henehan, 2001: 131).

Note, however, that territorial disputes are not necessary conditions for war, but only increase the probability of war when they are present. Other issues can also give rise to war. Vasquez and Henehan (2001: 131) show that 31 of the 97 (32%) war escalations arise from policy disputes. Territory is responsible for only one category of war, albeit the most frequent.⁵ In work seeking to classify types of wars, Valeriano and Vasquez (2005) find that territorial wars account for the majority of wars occurring from 1816­1997. Looking at multilateral wars, they get even stronger results in that 20 of the 28 (71.4%) multilateral wars can be classified as territorial wars.

The hypothesis that territorial disputes between nation-states are highly prone to war has been tested by numerous scholars in different ways and using different data sets. The earliest published studies were by Senese (1996) and Hensel (1996), both of whom group policy, regime, and other disputes into a single non-territorial category.

One of the criticisms of the above sort of studies is that they may be prone to selection bias, that is, that it is not territorial MIDs that bring about war, but the factors which bring about territorial MIDs in the first place that make for war. Several sophisticated analyses have tested for this possibility and none of them have found selection effects operating in the MID data (Senese and Vasquez, 2003; Senese, 2005; Rasler and Thompson, 2006).What the above studies find is that while territorial claims or contiguity may increase the probability of a MID arising, it is the presence of a territorial MID (and not territorial claims or contiguity) that make war likely.

All of this research suggests that conflict resolution should make the settlement of territorial disputes a focus of its efforts, since territorial disputes are highly war-prone and account for most interstate wars and rivalries. But where does one begin and are all territorial disputes alike?

One optimistic set of findings, which is also consistent with the territorial explanation of war, is that it is not territorial disputes that greatly increase the probability of war, but how they are handled and whether they lead to a rivalry. Recent work (Valeriano, 2003; Rasler and Thompson, 2006) has shown that rivalry and the timing of events prior to and during a territorial dispute are critical for the escalation to war and termination of outstanding territorial claims.

Rivalry (Diehl and Goertz, 2000; Thompson, 2001) is an essential variable because issues at stake within such a relationship fester and repeat. Rivalry is a situation of historic animosity where any and all issues at stake between the disputants take on a serious and deadly tone. One state would slash its own nose in order to deny a benefit to its rival. It is during these situations that territory becomes dangerous and war-prone. One state may take a portion of territory and settle the question for the time being, but if the rivalry persists, that territorial issue will repeat and reemerge even decades later.

The timing of events during a territorial disagreement is crucial to the outcome of settlement efforts. Rasler and Thompson (2006) find that war is unlikely over a territorial issue in the absence of an ongoing strategic rivalry. Valeriano (2003) finds that war and enduring rivalry are unlikely without territorial issues and power politics tactics such as alliances, simultaneous disputes, grand strategy development, and arms races occurring prior to the onset of rivalry.

Senese and Vasquez (2005) show that as states resort to various forms of power politics to get the other side to accept its territorial demands, the probability of war progressively increases. Trying to deal with territorial issues by seeking outside allies, building up one's military, or engaging in repeated militarized confrontations produces a security dilemma that makes the other side respond in kind.

These findings raise the question of what distinguishes the territorial disputes that go to war from those that do not. One answer to that question (as exemplified by the above research) is in terms of process—it is the process by which actors handle territorial issues that distinguish those that go to war from those that do not. One such key factor is whether territorial MIDs recur.⁶

Another and equally plausible answer as to why certain territorial disputes are more war-prone is that it is something intrinsic to the issue itself—its salience or the type of territory under question. Goertz and Diehl (1992b) measure the area of the territory and the size of its population to get at the importance of the territory in question. Hensel (2001) adds to the salience measure, indicators of economic resources, homeland vs. colonial territory, ethnic identity, and mainland vs. offshore territory. He finds some evidence that salience is important not only for predicting war, but also the probability of peaceful settlement. From his perspective, salience forces leaders “to do something” and that can be either engaging in nonviolent practices (like negotiations or mediation) or going to war. Hensel (2001:83) rightly regards these as substitutable means. The pressure “to do something” may also account for why intangible territorial issues have more peaceful settlements than tangible territorial issues, even though intangible territorial issues have a higher probability of going to war and having MIDs.

Huth (1996) and Huth and Allee (2002) classify territorial claims according to type— ethnic (including bordering minorities of the same ethnic group), strategic, and economic. They find that ethnic and strategic territorial claims are highly conflict-prone—in terms of escalation and war. Conversely, they find that territorial claims associated with economic resources are prone to peaceful resolution, especially if they involve developing coun­tries. Here, joint ventures and the need for capital can provide incentives to resolve the issue. Further refinement of typologies that look at the substantive nature of territorial issues (e.g. border adjustments vs. core territory, etc.) is an area that might prove productive for matching the right conflict resolution techniques with the relevant type of issue.

One factor that has long been a character­istic that is seen as making it more difficult to settle an issue is the number of actors involved. Multiple actors are seen as making negotiations and compromise more difficult. Brecher and Wilkenfeld (1997) provide some evidence (using ICB data) for this with regard to international crises—showing that multiparty crises are more prone to escalation and the use of violence. Petersen, et al. (2004) show the same is true of multiparty MIDs in terms of their having an increased probability of going to war. Of interest is that they demonstrate that this relationship is especially true of multiparty territorial disputes. Similarly, Valeriano and Vasquez (2005) find that most multiparty wars arise out of territorial disputes rather than regime or policy questions.

The research on territorial disputes has shown that it is an important source of conflict and that it is also a key factor (if not the factor) in the outbreak of interstate war. Since it is such an important factor, it is essential that territory be the focus of attention for war avoidance and reduction, but the territorial explanation of war is more optimistic than that. It maintains that among neighbors, territory is so important that once this issue is settled, it can lead to long periods of peace, even if other salient issues arise. This hypothesis is one of the major testable differences between the territorial explanation of war and realism, which sees war as inherent in the struggle for power. We turn now to research relevant to this claim about the relationship between territory and peace.