Territorial wars and absolute outcomes

Abstract

Why do some wars end with an absolute outcome, with state death or regime change? I argue that we are more likely to see absolute outcomes when we have territorial disputes with the potential for credible commitment problems and asymmetric disputants. In the absence of credible commitment problems, disputes are less likely to recur, and states are unlikely to seek to absorb the opponent state or remove its government. Among more symmetric disputants, states cannot impose an absolute outcome, and we are more likely to see recurrent disputes in the face of credible commitment problems. Only in very asymmetric dyads are we likely to have both the required willingness and opportunity to impose absolute outcomes to attempt to solve a credible commitment problem over territorial conflict.

Introduction

Why are some wars more likely than others to end with an extreme war outcome? Extreme war outcomes, also known as absolute outcomes, represent wars that end in state death or foreign-imposed regime change (Reiter, 2009: 35). While there is a rich literature surrounding war termination, much of the work focuses on understanding war outcomes in terms of a win, loss, or draw (Bennett and Stam, 1998; Slantchev, 2004; Stam, 1998). Yet not enough attention has been paid to understanding why some wars end with one side losing their ability to resist rather than a limited negotiated settlement. I present one potential answer to the question by focusing on the most likely cause of war – territorial disputes (Vasquez, 1995). While territorial disputes are the most likely cause of war, there are variations among these disputes. Such variation suggests that certain territorial disputes are more likely to lead to war and even extreme war outcomes.

Two factors drive my argument for when we are more likely to observe an absolute outcome. The first is that some territorial disputes are more likely to trigger credible commitment problems (CCPs) than others. Territorial wars, where CCPs have a higher likelihood of being present, give the combatants the willingness to pursue an absolute outcome. In particular, these territorial wars decrease a state’s desire to seek a negotiated solution. As Reiter (2009: 35) notes, absolute outcomes are among the most effective ways of dealing with CCPs. Wars without CCPs are unlikely to lead to an extreme outcome as combatants are more willing to accept a negotiated settlement, and they also decrease the likelihood of future conflict.

The second factor driving my argument is a lack of power parity between the combatants. Extreme outcomes are more likely to arise when there is an imbalance of power between the combatants. Trying to impose an absolute outcome when military forces are at parity would make for a daunting task. As such, territorial disputes with credible commitment concerns between relatively balanced combatants are unlikely to reach an extreme outcome yet could lead to repeated conflict. Instead, we are more likely to observe an absolute outcome when CCPs and a large imbalance of power provide the necessary willingness and opportunity for such outcomes for one of the disputants to find a long-lasting solution to the territorial issues.

The findings presented here present a unique contribution to the war termination literature. Rather than focusing on whether a war ends with a win, loss, or draw, I conduct one of the first empirical tests of whether a war ends with an extreme outcome. The analysis shows that territorial wars more prone to CCPs along with a large imbalance of power increase the likelihood of observing an absolute outcome. While the focus is on territorial wars, the theory presented can apply to other types of war. As such, this article serves as an initial foray into the questions surrounding extreme war outcomes worthy of further exploration.

Territorial disputes, credible commitment problems, and absolute outcomes

Not all territorial disputes have the same likelihood of high levels of violence. For instance, the Issues Correlates of War (ICOW) has shown that territorial disputes’ salience can increase their likelihood of escalation and make negotiated settlements more difficult to achieve (Hensel, 1996). Yet studies into ICOW have also shown a great deal of variance in the salience involved in territorial disputes (Hensel et al., 2008).

Territorial disputes where the fate of one of the disputants’ homeland is potentially at stake increase the potential for high levels of violence. Directly targeting a state’s homeland will make them unlikely to accept any negotiated settlements even in the face of battlefield setbacks. Similarly, when states fear future claims on their territory, they are less likely to give into compellent threats from stronger challengers (Sechser, 2010). In other words, it will be challenging to convince a state to stop fighting and negotiate when they believe a new threat to their homeland would occur in short order. Conversely, the state pursuing such an outcome will have little incentive to make a deal as well.

Therefore, disputes involving threats to one state’s homeland are particularly problematic as they enhance the potential of CCPs being present. CCPs are one source of war discussed in the bargaining model of war (Fearon, 1995; Powell, 2006; Reiter, 2009). The credible commitment problem arises in an anarchical system where a state cannot commit to honoring agreements in the future (Powell, 2006). As shown in Table 1, the columns on the right of the 2 x 2 table are territorial disputes with the potential for CCPs. Thus, regardless of considering the power disparity between the combatants, the presence of CCPs will increase the likelihood of conflict.

Table 1.

Theoretical expectations.

	No CCP threat	CCP threat
Power parity	Little potential for recurrent conflict.	High potential for recurrent conflict.
Power asymmetry	Some potential for recurrent conflict.	High likelihood of absolute outcome.

CCP: credible commitment problem.

Nevertheless, disputes with CCPs that lead to war are particularly problematic as they increase the likelihood of an extreme outcome. Reiter (2009: 24) argues that states will only enter into negotiations to end a war when they feel that commitment issues will not be problematic. If one cannot trust one’s opponent to honor peace agreements, there is little reason to stop fighting. Wolford et al. (2011: 568) agree with this line of thinking and show that minimizing commitment problems makes initiators willing to buy off targets to avoid a long war. Hence, even if there is a clear picture of the distribution of power and resolve (i.e. no information problems), states will still opt to fight if they believe their opponent will renege on deals and initiate a new fight soon. To be sure, this is not to suggest that an absolute outcome would be observed in all of these instances, just that the probability increases.

The question then becomes understanding what type of territorial disputes are more likely to increase the likelihood of CCPs. Gibler (2017) breaks territorial disputes into numerous categories. These categories are homeland defense, disputed ownership of territory, general border issues, opportunistic territorial disputes, state system changes, border violations, and other territorial disputes (Gibler, 2017: 198). While Gibler finds that disputed ownership and state system changes were the most likely to lead to war, this does not mean these are also the most likely to end with an absolute outcome. Instead, I argue that territorial disputes with the potential to contain direct threats to one of the actors’ homelands are more likely to produce CCPs.

In particular, disputes over homeland defense, opportunistic disputes, and disputes arising from state system changes are the most likely to raise such threats. Homeland defense disputes involve the defense of territorial borders from outside actors. Opportunistic disputes involve instances where one state takes advantage of a weaker opponent to capture new or retake old territory. The German invasions of the Low Countries during the early stages of World War II (WWII) represent such disputes. Lastly, state system disputes arise from changes in the system, or if forming a new state increases the likelihood of challenges to established borders. Because the intent in these disputes is to directly target at least one state’s homeland, they are likely to trigger credible commitment concerns and increase the likelihood of an absolute outcome.

Returning to Table 1, while certain territorial disputes with CCPs may lead to recurrent conflicts, this does not necessarily mean an absolute outcome will follow. We are more likely to observe an extreme outcome when there is a great deal of power asymmetry between the combatants. One potential reason for the low number of observed absolute outcomes is that they are challenging to achieve (Reiter, 2009: 39). Bringing about the collapse of an enemy is easier said than done. To be sure, states pursuing an absolute outcome will need a certain amount of military capabilities. However, having an advantage in relative capabilities is a crucial factor. Wars between combatants of unequal power are at a higher likelihood of resulting in an absolute outcome than wars between combatants of equal power. As such, the lower-right cell of Table 1 shows that territorial disputes with the potential for CCPs and an imbalance of power increases the likelihood of observing an absolute outcome. Yet wars without CCPs may not have the same incentives for pursuing an absolute outcome, but the lingering commitment issues create a high likelihood of future conflicts, as shown in the upper-right cell.

However, we can also return to the above point regarding weaker states pushing back against stronger states in territorial disputes. After all, why fight if one is outgunned? A potential answer is that a weak state wants to prevent future challenges, as suggested by Sechser (2010). The various conflicts between China and Vietnam in the 1980s highlight this dynamic. In other words, weaker states may prefer fighting today to avoid potential challenges tomorrow that trigger CCPs. As such, the lower-left cell in Table 1 reflects this outcome, and we can see cases where weak states will indeed fight back against stronger states even if there are no CCPs currently present. Consequently, while the lack of CCPs could provide some damper on the likelihood of future conflict, there is at least some potential for recurrent conflict. Stronger states have little incentives to pursue an absolute outcome yet may still attempt coercive actions, and weaker states have incentives to push back against any encroachments.

In summary, Table 1 shows the conflict expectations for the relationship between the balance of power and the likelihood of a CCP threat. I expect that the highest likelihood of observing an absolute outcome occurs when there is a territorial war with the potential for CCPs, and there is a large imbalance of power between the combatants. When the territorial war lacks the threat of CCPs but there is an imbalance of power, there is some potential for recurrent conflict. Finally, when there is power parity and no threat of a CCP, there is a low potential for recurrent conflict. As such, I hypothesize:

Absolute Outcome Hypothesis: Territorial wars with CCPs are more likely to result in an absolute outcome than territorial wars without CCPs. The impact of CCPs on absolute outcomes increases as the combatants’ disparity increases.

Research design

The unit of analysis for this study is the territorial war dyad. More specifically, the non-directed war dyad. Territorial disputes are identified from Gibler (2017) and result in 598 territorial dispute dyads from 1818 to 2001. Of these 598 disputes, 64 experience war as defined by the updated Correlates of War data set (Reiter et al., 2016). Importantly, Gibler (2017) argues and shows that why states fight over territory is exogenous from how they fight over territory. Hence, territorial disputes with CCPs may be more likely to escalate to open war, yet this does not imply states will fight with extreme goals in mind. In other words, we should not have a priori expectations that a certain class of territorial disputes will guarantee an extreme outcome. ¹

Dependent variables

The dependent variable is whether or not the war dyad experiences an absolute outcome after a war begins. Here, I define an absolute outcome as a war resulting in a state death or a foreign-imposed leader change (Reiter, 2009: 35). Two sources were used to create this variable, the Archigos data project (Goemans et al., 2009) and Fazal (2011). The Archigos data identifies leaders removed by foreign powers while Fazal identifies states that suffer a violent death. Absolute Outcome would equal one if war resulted in either of these outcomes. Of the 64 dyads identified that experience war, 17 experience an absolute outcome. ²

Independent variables

The primary independent variable for the model is whether a territorial dispute is likely to trigger credible commitment concerns. Using the classification scheme from Gibler (2017), CCP equals one if the dispute involves homeland defense, opportunities, or state system changes. Conversely, CCP equals zero if the dispute involves general border issues, border disputes, border violations, and other territorial issues.

With only 64 observations to test, judicious use of control variables is necessary. First, using the Composite Index of Capabilities from the Correlates of War, Relative Power equals the strongest state’s share of the total power in the dyad (Singer et al., 1972). ³ The variable ranges from 0.5 (perfect parity) to 1 (the strongest state has all the power). Additionally, Contiguous equals one if both states share a border and zero otherwise, while Previous MIDs capture the number of territorial militarized interstate disputes (MIDs) the dyad experienced before the latest outbreak. Fourth, the variable Bilateral equals one if the war resulting from the territorial dispute remained bilateral and zero if the war became multilateral. Fifth, as one could argue that the WWII dyads are systematically different from the rest of the dyads, I include the dummy variable WWII to capture the 1939–1945 time period. Lastly, as the theory suggests that CCPs and power together impact absolute outcomes, there is an interactive term CCPXPower in the regression model.

Empirical analysis

Table 2 shows the results from the logit regression models. From the hypothesis, we would expect the interaction coefficient between the CCP and Relative Power to be positive and statistically significant in the model’s absolute outcome stage. Looking at Model 1, this is indeed the case. The results suggest that territorial wars involving CCPs increase the likelihood of observing an absolute outcome as the dyad’s power balance becomes more asymmetric.

Table 2.

Models 1818–2001.

	Model 1	Model 2	Model 3	Model 4
	All	Drop WWII	All	Drop WWII
	Logit	Logit	Firth logit	Firth logit
CCP	−11.779*	−19.697**	−9.107	−13.459*
	−4.942	−6.151	−4.696	−5.908
Relative Power	−12.800*	−15.513*	−9.955	−11.738*
	−5.761	−7.502	−5.262	−5.609
CCP x Power	15.240*	24.773**	11.889*	17.232*
	−6.355	−8.138	−5.932	−7.278
Bilateral	−0.912	−0.561	−0.806	−0.55
	−0.959	−1.207	−0.726	−0.924
WWII	1.837		1.534
	−1.027		−0.885
Contiguous	−0.781	−1.945	−0.643	−1.57
	−0.911	−1.159	−0.767	−0.985
Previous MIDs	−0.242*	−0.245	−0.144	−0.141
	−0.115	−0.134	−0.107	−0.11
Constant	9.869*	12.456	7.53	9.348*
	−4.869	−6.376	−4.311	−4.755
n	64	53	64	53

WWII: World War II; CCP: credible commitment problem; MIDs: militarized interstate disputes.

*

p <.05, and **p < .01 for two-tailed tests. Standard errors clustered on the dyad (for models 1 and 2).

Model 2 in Table 2 removes the WWII cases from the model. One could argue that the WWII cases are driving the results of model 1. Yet, from the model, we see that removing these cases is not changing the primary findings of model 1. We still see a positive and statistically significant interaction in the absolute outcome stage of the model. Hence, the WWII cases are not driving the primary results.

Figure 1 shows the relationship graphically. The figure shows the impact of a war with and without CCPs on the likelihood of absolute outcomes as relative power increases. From the figure, we can see that territorial wars with CCPs as the power disparity increases the likelihood of an absolute outcome. In comparison, for territorial wars without CCPs, we see that as power disparity increases, the likelihood of an absolute outcome decreases. Although, an interesting finding is that when the combatants are close to parity, the lack of CCPs increases the likelihood of an extreme outcome versus wars with CCPs. The most likely explanation for this result is the few absolute outcomes at this range of relative power. Note that the difference between the two types of war quickly becomes insignificant. However, at higher ranges of disparity (above 0.9), wars with CCPs are more likely to result in an absolute outcome. For example, the average probability of an absolute outcome when relative power equals 0.95 (below the 90th percentile for wars) is about 42% when there are CCPs and only about 6% when there are no CCPs. Such an increase for rare events shows that the results from Table 2 are statistically and substantively significant.

Figure 1.

Effect of relative power on the likelihood of an absolute outcome.

Lastly, as was noted above, there are only 64 observations being used in the models. For logit models this is a relatively small sample size. Such relatively small sample sizes in the logit models could be problematic for producing unbiased estimates (Rainey and McCaskey, 2021). To account for this, models 3 and 4 use a Firth logistic regression model that reduces small sample bias (Rainey and McCaskey, 2021). Even with the new models, the interactive term remains positive and statistically significant.

Discussion and conclusions

While the statistical results provide support for the hypothesis, looking at the cases suggests two related findings. The first is that we have seen multiple cases post-WWII where there is either a high likelihood of CCPs or an imbalance of power, but not both. As Table 1 suggests, these cases avoid extreme outcomes yet devolve into repeated conflicts. For instance, multiple territorial wars between India and Pakistan and between Israeli and its Arab neighbors have occurred, in part, because these conflicts end short of an extreme war outcome. In particular, Israel has been in the position to deliver an extreme outcome in wars such as in the Yom Kippur War yet held back. Similar dynamics have been at play with the modern conflicts involving Azerbaijan and Armenia and even the various conflicts between China and Vietnam in the 1980s.

Furthering the point, nearly 45% of the territorial wars identified involve disputed territory in some form. However, these wars resulted in only three absolute outcomes. The Iraqi annexation of Kuwait in 1990 is the latest example. While disputed territory may lead to war, even repeated war, it does not necessarily create the willingness to pursue an absolute outcome. In other words, what is notable from the cases is the missed opportunities for an absolute outcome in recurrent conflicts that we have witnessed post-WWII.

The second observation flows from the first observation. We have entered a period where absolute outcomes from territorial wars with potential CCPs have declined. While we have had absolute outcomes in the 21st century (Iraq and Afghanistan), absolute outcomes from territorial wars with the high potential for CCPs have been absent post-WWII. Before WWII, the last observed extreme outcome from a territorial war with a CCP was Ethiopia’s Italian annexation as Mussolini looked to expand Italian power.

The post-WWII slowdown of territorial wars with CCPs leading to absolute outcomes supports the argument of Fazal (2011). Although Fazal is primarily arguing about state death, the results here suggest regime changes have been less common as well. One could argue that international factors, whether they be institutions like the United Nations or American hegemonic power, have played a prominent role in minimizing extreme outcomes. While this may seem like good news, one could argue it has fostered an environment for repeated conflicts, as discussed above. Additionally, we should not believe there is no longer the threat of these outcomes. At least for now, the threat of absolute outcomes over territorial disputes may have declined, but a shake-up in the system could reopen the opportunities for such conflicts.

While identifying the conditions that increase the chance of an absolute outcome can help the international community intervene in such wars, there are additional research avenues to explore. One avenue is to explore other ways that CCPs may emerge in wars (like Iraq and Afghanistan). Additionally, more work needs to focus on the decisions of leaders to pursue such outcomes in the first place. I have shown that wars resulting from certain types of territorial disputes and an imbalance of power can create the conditions for an absolute outcome, yet the final decisions rest with leaders. Further engagement in micro-level thinking will give us a complete picture of extreme war outcomes.

Supplemental Material

sj-pdf-1-rap-10.1177_20531680211043316 – Supplemental material for Territorial wars and absolute outcomes

Supplemental material, sj-pdf-1-rap-10.1177_20531680211043316 for Territorial wars and absolute outcomes by Brian Benjamin Crisher in Research & Politics