UN intervention and the duration of international crises

Abstract

This article examines the effect of UN actions on the duration of international crises. Four different types of action – assurance, diplomatic engagement, military involvement, and intimidation – and three different outcomes – compromise, victory, and stalemate – are considered. After building on the existing literature to develop expectations of how a third party like the UN shapes crisis trajectories, hypotheses are tested using the International Crisis Behavior (ICB) data and a new events dataset on UN activity. Results from competing-risks models reveal that UN military involvement does well to decrease the risk of one side achieving victory, and diplomatic engagement increases the ability of the belligerents to reach a compromise in the long run. Moreover, diplomatic engagement accompanied by military involvement substantially hastens the pace of stalemate outcomes. Both tactics, however, have some trade-offs. Military involvement can decrease the sense of urgency for compromise; diplomatic engagement can be used for insincere motives and increase the risk of one-sided victory over time. UN actions of assurance and simple intimidation have considerable shortcomings as crisis management vehicles.

Keywords

conflict resolution event history International Crisis Behavior matching United Nations

Introduction

The literature on the relationship between the UN and conflict predominantly has focused on its role as peacekeeper. While much is known quantitatively about the effects of peacekeeping and peacebuilding (Collier, Chauvet & Hegre, 2008; Doyle & Sambanis, 2000, 2006; Greig & Diehl, 2005; Fortna, 2003, 2004a,b; Gilligan & Sergenti, 2006), less is known about how the UN performs as a conflict manager or peacemaker.¹ This article explores the UN’s impact on the duration of crises and posits that different types of UN action have disparate effects on the timing of crisis abatement.

The existing literature presents competing expectations of what types of third-party tactics are most effective in conflict management and resolution. While some work expects only heavy-handed tactics to have the potential to shape the bargaining environment (Betts, 1994; Smith & Stam, 2003), others find that lighter forms of involvement such as mediation can do quite well (Beardsley et al., 2006; Regan & Stam, 2000; Rauchhaus, 2006; Wilkenfeld et al., 2005).² Still others expect that third parties can disrupt more natural resolution processes or otherwise create perverse incentives, especially when involved in military deployments (Greig & Diehl, 2005; Kuperman, 2008; Luttwak, 1999; Rauchhaus, 2009; Werner & Yuen, 2005). In assessing what these earlier studies mean for the expected effects of UN intervention, this article uses a new events dataset that disaggregates UN activity during international crises. The data are able to capture not only how UN intervention shapes the trajectory of crises but also how different tactics of involvement stack up against each other.

As part of this exploration, this study distinguishes between the ways for a crisis to end. Early ends of crises in compromise fit the profile of an ideal outcome from the standpoint of peace advocacy. Early terminations in victory, however, typically indicate a dereliction of effective conflict management. Distinguishing between the different types of outcome is important from both policy and normative standpoints, as decreasing crisis duration is not always a worthwhile objective, especially since one way for a crisis to end quickly is for one side to win a complete victory through bloodshed. This article unfolds by first providing a theoretical framework of the various ways for conflict to terminate and then discussing the UN’s role within that framework. Testable hypotheses are derived and tested empirically.

Conflict processes and duration

Before forming expectations about the UN’s ability to shape the duration of crisis, we must first have a sense of why a state of conflict might exist and how it ends. This article considers conflict as part of a bargaining process in which two or more actors are in dispute over some good or issue. Using force has two different purposes in the bargaining context: it can demonstrate capability and resolve, and it can physically remove vulnerabilities. Each of these is taken up in turn.

Actors will often resort to violence when either they have underestimated what their opponent would be willing to accept or their opponent has underestimated them. The act of engaging in conflict reveals capability and the resolve to bear costs, which allows the expectations of the actors to converge on the set of alternatives that are mutually satisfactory (Slantchev, 2003, 2004; Smith & Stam, 2003; Filson & Werner, 2002). The demonstration of capability and resolve is not only meant for the eyes of the foreign policymakers of the opponent. Leaders may refuse to concede not because they cannot identify areas of overlapping agreement but because domestic audiences might punish them for making concessions (Debs & Goemans, 2010; Goemans, 2000, 2009; Chiozza & Goemans, 2004; Tarar & Leventoglu, 2009; Tarar, 2006; Trager & Vavreck, 2011). To the foreign policymakers in such situations, the set of possible settlements that are mutually preferable to conflict may be empty. The use of force in such cases can thus be driven by a leader’s domestic constraints or serve as a means to demonstrate to an opponent’s domestic audiences that any concessions their leader makes are, in fact, prudent. In either case, force will be applied until the costs of conflict for either side become greater than the costs of backing down and potentially losing political capital.

Aside from allowing an opponent to realize the merits of concession, force is also used to resolve vulnerabilities to commitment problems. In cases where actors cannot credibly commit to abiding by an agreement even if it were found to be mutually preferable to conflict, settlement will be avoided and fighting will occur until credible commitment becomes possible (Reiter, 2009). One source of a credible commitment problem is the potential for an actor’s capabilities – and thus bargaining power – to rapidly improve, which would leave the rising actor dissatisfied with the settlement at hand and increase its incentives to push for a better deal in the future (Powell, 2004, 2006). To avoid having to concede more in the future, the other actor would prefer to damage the rising actor’s capabilities through war and stunt its potential growth so that whatever is agreeable today will also be agreeable in the future. Another source of a credible commitment problem relates to the concepts of the security dilemma and mistrust (Kydd, 2005), where conflict may persist because of uncertainty about whether an opponent will exploit cooperation for strategic advantage. The use of force in such a situation would be aimed at going on the offensive to sufficiently damage the opponent’s own offensive capabilities or to seize positions that can more easily be defended. Finally, credible commitment problems can arise from spoiler groups that either are not part of the conflict bargaining or would benefit from ongoing hostilities (Stedman, 1997; Kydd & Walter, 2002; Lake & Rothchild, 1996). Violence in this case would be intended to eliminate the threat that such groups pose or to increase the incentives of the target actor to self-police affiliated groups.

Against this backdrop of why hostilities typically occur, I focus on three distinct ways for crises to end. First, crises can end when the actors identify a new distribution of the goods or issues in question that is mutually preferable to conflict and credibly commit to that via some sort of compromise settlement. Second, actors might stick to the original status quo when they come to realize that the status quo is actually preferable, even if begrudgingly so, to escalating the hostilities. Third, crises will end when one side has defeated the other after complete bargaining failure, ultimately leaving little question about relative capabilities and minimizing commitment problems.³ I term these three different outcomes compromise, stalemate, and victory, respectively.

Of these three types of ways for crises and conflicts to end, compromise is the most direct indicator of bargaining success. Resolution of information problems, audience constraints, and commitment problems can lead to an outcome of compromise and establish a new status quo. In contrast, complete victory indicates a complete inability to reach an arrangement that is preferable to conflict, as actors largely circumvent the bargaining process and reach an end through coercion. Stalemate is a more ambiguous outcome because in some cases it may indicate relative bargaining success, as when the actors come to realize that the status quo is clearly preferable to fighting. In many other cases, stalemate indicates continual dissatisfaction with the status quo, but the actors temporarily stop fighting for the sole reason that prosecuting the dispute in the present has become undesirable. We can now turn to expectations about how the UN shapes the rates at which crisis actors reach these outcomes.

UN intervention

Types of involvement

To form expectations about what effect the UN can have on crisis duration, we must first get a sense of what types of action the UN might take. The UN has a number of means available for managing conflict. For the purpose of limiting the scope of activities, this article focuses on those activities that involve Security Council or General Assembly resolutions or substantive action by the Secretary-General. I consider two dimensions of UN involvement, which produce four distinct types of UN actions. The first dimension relates to the intended mechanism of that action as defined by whether the UN is trying to facilitate a negotiated bargain within the existing set of possible agreements that are mutually preferable to conflict or to increase the costs of conflict and thereby expand the set of possible agreements. When a third party attempts to expand the set of existing settlements through such actions as threatening punishments, shaming or militarily enforcing its will, we can say that the third party is using leverage to make a settlement possible. In the absence of leverage, the third party is trying to resolve the information, audience constraint or commitment problems through such actions as fact-finding, mediating, and providing security guarantees. The second dimension relates to the substance of the involvement, in terms of whether it involves what Diehl, Reifschneider & Hensel (1996) term operational deployment – the authorization or implementation of a monitoring, peacekeeping or emergency military force. This dimension is important to consider since it separates out types of action in which considerable resources have actually been invested from other actions that are either relatively cheap to carry out or that only hold the promise of future resource investment. In combining these two dimensions, four distinct types of UN actions emerge, as seen in Table I.

Table I.

Typology of UN activity

		Leverage
		No	Yes
Operational deployment	Yes	Assurance	Military involvement
Operational deployment	No	Diplomatic engagement	Intimidate

Starting with the lower row of Table I, the actions grouped as diplomatic engagement include those instances of good offices, fact-finding or mediation that are typically executed by the Secretary-General, special representatives or ad hoc commissions. The focus of such involvement typically involves improving the information environment among the sides such that they can more quickly come to understand each other’s reservation values and reach a resolution. Diplomatic engagement might also be used as a means for the actors to receive political cover or save face, in that domestic audiences can cue off of a mediator to update expectations about the prudence of concessions (Beardsley, 2010). The other group of actions that does not include operational deployment is called ‘intimidate’.⁴ These involve instances where the UN calls for the sides to comply, condemns illegal behavior, threatens force or levies sanctions. The hope is to make the belligerents perceive that ongoing hostility will be met by some sort of negative response by the international community. The intimidating actions can be explicit, as in direct condemnation, specific threats or sanctions. The intimidations can also be implicit, as when the UN calls on states to comply with their obligations but leaves the consequences of compliance failure open-ended. Such calls for action could imply that the non-compliant actors will be called out and shamed, and they can also sometimes be taken to imply that punitive action will meet such non-compliance.⁵

Turning to the top row, assurance involves the UN trying to allow the disputants to feel more comfortable with a peaceful bargain, through such activity as sending an observer group or promising a peacekeeping force. Typically, this activity is most necessary when there are important commitment problems and security dilemmas that make the actors reluctant to reach an outcome they would otherwise prefer to ongoing conflict. Finally, military involvement pertains to when missions are used to combat further aggression, through the deployment of UN peacekeeping forces or the authorization of non-UN multinational military forces. This is the most direct way that the UN can increase the perceived costs of ongoing conflict by the actors. Like the assurance function, it is also frequently intended to protect the actors from future defections while any settlements are implemented.

UN involvement and the pace of conflict bargaining

Returning to the discussion above about the differences between crises that end in compromise, stalemate, and victory, UN involvement should have different effects on the rate at which such outcomes are reached. The existing literature provides some conflicting expectations about the relative effectiveness of the different types of third-party involvement. The empirical analysis, which distinguishes between UN intervention types, is thus useful in adjudicating among these potentially competing claims.

The literature is most developed with regard to how third parties can affect the potential for some sort of compromise settlement. Favretto (2009) formally argues that heavily biased third parties and honest brokers will be most effective in helping disputants resolve their bargains – the former are able to credibly signal a willingness to enforce peace and the latter are able to facilitate the exchange of information. One of the expectations that then emerges is that impartial diplomatic engagement, such as that by the Secretary General or a special envoy, can have the potential to effectively manage conflict toward compromise. A number of empirical studies have also shown that diplomatic engagement in the form of mediation can improve the ability for the actors to either reach a new arrangement or to back down from military hostilities. Beardsley (2008, 2011), Beardsley et al. (2006), Regan & Stam (2000), Wilkenfeld et al. (2005), and Walter (2002) have shown that mediation can improve the ability for disputants to reach a negotiated settlement. Rauchhaus (2006) shows that third parties who use light mediation can improve the ability of actors to de-escalate, and coercive intervention has no significant effect. The first hypothesis that we can test is that diplomatic engagement will tend to hasten the path toward compromise.

Efficacious bargaining hypothesis: The time until compromise will be shorter after the UN has become involved in diplomatic engagement.

The literature, however, is not unified in this expectation. Smith & Stam (2003) argue that third-party use of mediation alone will struggle to impact the ability of disputants to converge on a mutually acceptable bargain because third parties are generally unable to credibly convey new information to the sides, especially when a third party like the UN has a strong preference for peace. The problem of cheap talk is likely to particularly apply to UN involvement in diplomatic engagement or intimidation. Beardsley & Schmidt (forthcoming) find that UN action is more strongly driven by a desire to uphold the organizational mission established in its founding Charter – namely, to promote peace and stability – than it is by more parochial interests of the permanent-five members of the Security Council. This means that UN mediation, especially when carried out by the Secretary-General or his representatives, has a strong tendency to value peace as an end in itself with less concern for the ultimate distribution of the outcome. That is, UN diplomatic engagement tends to be heavily biased toward peace. While extant mediation studies, including Kydd (2003, 2006), Smith & Stam (2003), and Rauchhaus (2006), disagree over whether bias toward a particular side is good or bad for mediation efficacy, a consensus does emerge from this work positing that third parties that are biased toward peace will struggle to do anything more than offer cheap talk. From this perspective, the UN is not well suited to succeed in diplomatic engagement or in simple calls for peace.

Smith & Stam (2003) further argue that the only way for third parties to help the combatants reach an agreement is through using leverage in the form of peacekeeping to inflate conflict costs. Other empirical studies confirm that the use of peacekeeping and security guarantees can help resolve critical barriers to successful bargaining such as commitment problems and related security dilemmas (Doyle & Sambanis, 2000, 2006; Fortna, 2004a,b; Gilligan & Sergenti, 2006; Walter, 2002). Such studies are thus consistent with a view that military involvement tends to do well in bringing combatants to a stable settlement.⁶ They could also be used to form expectations regarding the assurance function of the UN if observers and promises of peacekeeping can provide sufficient protection against commitment problems. Mattes & Savun (2009) have found that the mere provision of peacekeeping provisions in a civil war peace agreement can solidify peace. Schultz (2010) has also found that the presence of monitors can go a long way toward assuring combatants.

One expectation from these arguments would thus be that greater use of military involvement and assurances can lead to faster compromises or stalemates. Disputants will be more willing to compromise or settle on the status quo when they feel less vulnerable to exploitation. Related, UN intervention via military involvement and assurances can delay the achievement of victory by a side, as when combatants become less willing to push forward in full combat because the costs of conflict have been raised. This would result when strong military involvement or buffering prevents a side from winning a military campaign, as for example when the UNSC mobilized forces in 1950 to keep North Korea from defeating South Korea. Just as operational deployments can reduce the risk that actors will be seriously exploited after the conflict is over, they can also prevent one side from making substantial gains and prevent victory during a crisis.

Reduced vulnerability hypothesis: The time until compromise and stalemate will be shorter and the time until victory will be longer after the UN has engaged in military involvement or assurance.

Other scholarship has pointed to more direct downsides to third-party diplomatic engagement. One line of argument relates to what Richmond (1998) has labeled as ‘devious objectives’ of disputants who use peace processes insincerely to stall. The logic is that disputants can sometimes take advantage of lulls in hostilities while a third party is actively involved to rearm or regroup so that they can initiate a more successful military campaign in the future. Greig (2001) and Beardsley (2009) find some empirical confirmation that this phenomenon exists. Toft (2009) also argues that third-party guarantees have limited ability to prevent conflict recurrence because they typically fail to address the problem of stalling tactics.

When disputants have such insincere incentives to stall, they would best be able to take advantage of diplomatic engagement because such activity is likely to create the space in which such rearming could occur while not involving monitoring or peacekeeping that would prevent the disputants from succeeding. By its nature, stalling can most directly lead to the avoidance of peaceful outcomes such as compromise or stalemate. On the flip side of this logic, in the long run stalling can also reduce the time to victory when the stalling actor uses the lull in hostilities effectively to catch an opponent by surprise with its more effective fighting capabilities.

Insincere motives hypothesis: The time until compromise and stalemate will be longer and the time until victory will be shorter after the UN has become involved in diplomatic engagement.

There is also a potential downside to the feeling of safety that military involvement and assurance can provide. Intrusive third-party involvement in some cases might interrupt the ability for the protagonists to reach a mutually acceptable bargain (Luttwak, 1999). Combatants learn from each other in crisis, and early outside intervention can cut that learning process short. Early third-party involvement in crises can prevent the disputants from learning about each other’s capabilities or about each other’s potential to exploit cooperation. Such premature intervention might not only contribute to a more fragile peace after a conflict has ended, as Werner & Yuen (2005) have found, but it might also lead to longer times until compromise.

Related to the discussion about the potential dangers of early intervention, the UN can create perverse incentives by threatening to come to the aid of a vulnerable belligerent. Kuperman (2008) has argued that third-party intervention can increase the aggressiveness of the protected actors and make them less willing to make prudent concessions. Rauchhaus (2009) modifies this argument and suggests that third parties often face a commitment dilemma such that a strong commitment to defend an actor will increase that actor’s belligerence while a weak commitment will increase the potential for deterrence failure.⁷ Even though Kuperman and Rauchhaus focus on humanitarian intervention, a similar logic might apply to any potential UN operational deployment. Betts (1994) similarly argues that limited interventions such as those that tend to result from multilateral institutional bodies should be avoided because they tend to block the ability for the disputants to actually resolve their conflict while trying to keep the sides from inflicting too much damage on each other. To the extent that the UN makes the actors feel safe, particularly when the UN intervenes to protect weaker parties that would otherwise yield ground to a stronger party, it is especially prone to diminishing the incentives of the actors to move toward compromises and costly concessions. In this regard, Greig & Diehl (2005) argue and demonstrate that the presence of peacekeeping decreases the incentive for the more secure combatants to earnestly engage in negotiation efforts to end the state of conflict.

As another perverse incentive, even promises of future peacekeeping can detract from immediate peacemaking goals. Such promises might create a limited window for belligerents strongly resistant to the status quo to make as many gains as possible before peacekeepers deploy and hinder further challenges. In these ways, the potential exists for interventions of assurance or military involvement to actually lengthen the time until peaceful settlements.

Premature intervention hypothesis: The time until compromise will be longer after the UN has engaged in military involvement or assurance.

Turning to a final observable expectation, we must consider the unintended impact that the UN can have on crisis duration through purely procedural effects. UN action might temporarily lengthen the timing of compromise simply because peace processes take time, especially in a multilateral setting. Time is needed to set up meetings and to organize the logistics of monitors or peacekeepers. For example, major hostilities in the Yom Kippur War ceased at the end of October 1973, but the Geneva Conference that commenced the disengagement agreements did not occur until late December and the first partial disengagement plan was not reached until January 1974, and all the while the Egyptian Third Army was surrounded by Israeli forces. This potential to temporarily lengthen the time to a peaceful settlement applies specifically to diplomatic engagement, assurance, and military involvement. Simple intimidation does not often take any procedural time, as it typically involves singe resolutions targeted at the actors. Note that it is possible for UN involvement to still improve the peacemaking environment even when a temporary procedural effect is in play. After an initial lengthening of the peace process, the mechanisms proposed above specific to diplomatic engagement, military involvement, and assurance, to the extent that they actually help the actors identify and reach a negotiated settlement, can take over and produce shorter times until compromise.

Procedural hypothesis: The time until compromise will be initially longer after the UN has become involved in diplomatic engagement, military involvement or assurance.

Note that none of the above hypotheses have touched on the role of intimidation from the UN. We should not expect much of an effect on the time until any of the outcomes. As discussed above, the UN tends to be biased toward peace, making any pleas or admonitions for peace easy to dismiss as cheap talk. Moreover, the absence of committed resources prevents simple intimidation from having other effects, for good or for ill, on the disputant incentives.

Research design and data

All the hypotheses relate to expectations over the duration of a crisis situation. This requires an event history, or duration, analysis. Using time-varying covariates, I employ competing-risk models, so that we can look at the time until a crisis ends in compromise, victory or stalemate. A dyad of states becomes at risk for crisis termination at the start of a crisis and observation ends upon crisis termination. Similar to the analysis in Balch-Lindsay, Enterline & Joyce (2008), the unit of observation is the crisis-day. Since the data use a discrete-time structure, the models are estimated using multinomial logit regression.⁸ A cubic polynomial of elapsed time is also included in order to account for duration dependence (Carter & Signorino, 2010).

The definition of international crises comes from the International Crisis Behavior (ICB) data (Brecher & Wilkenfeld, 2000). In these data, an international crisis exists when three criteria are true. First, the actors must perceive a threat to some basic value such as security or influence. Second, there must be some finite time for the actors to address the threat. Third, the actors must perceive that there is some risk of the escalation of military hostilities. The set of ICB crisis dyads comes from Hewitt (2003), and I have updated this list until 2002.⁹ The minor dyads that are part of the Korean War and Gulf War are excluded, and the intra-war crises that are part of greater wars are aggregated together into single crises. I use ICB version 9 for information about crisis and disputant characteristics and only consider those crises since 1946.

The analysis uses original data for information regarding UN involvement events, collected with Holger Schmidt. This dataset, UNIEvents, contains events-level information on UN activity during ICB crises. The existing ICB data only include one observation per crisis, which limits the ability to assess how third-party involvement like that of the UN changes the trajectory of belligerent behavior within each crisis. With the events data, time-varying covariates can be incorporated. Events constitute any UN activity that attempts to shape the trajectory of the dispute and that includes a UNSC resolution, a General Assembly resolution or substantive action from the Secretary-General. The data identify 50 different types of action that the UN might take to manage a crisis situation, and the unit of observation in our new dataset is the event, such that we often have multiple observations per crisis.

I have adapted the UNIEvents data for the purposes of the analysis. For tractability, I aggregate the types of UN involvement into the four categories previewed above. The first, military involvement, includes the authorization and deployment of multinational forces, the use of force by UN-authorized missions, and the deployment, expansion, and strengthening of UN peacekeeping missions. While multinational forces such as those used in the first Gulf War, Haiti (1994), and East Timor (INTERFET), are not UN missions per se, I include them as part of UN involvement because their mandates come directly from UNSC resolutions. In order to separate out missions that are meant to affect the balance of forces and costs of conflict from missions that are primarily observational, UN peacekeeping is defined as any mission that is under the Department of Peacekeeping Operations (DPKO) but that is not limited to observer, monitoring or political functions. Military involvement occurred in five percent of the 275 crises in the data.

Assurance is the second category and is defined as the authorization and deployment of observer missions, the authorization and deployment of humanitarian missions, the authorization of peacekeeping missions, and peacebuilding activities. Observer missions include both civilian missions and military observer or monitoring missions under the DPKO. Note also that peacekeeping authorization is included as part of an assurance action since it is only the promise of a security guarantee. Peacekeeping deployment is part of military involvement when it occurs during a crisis – since the data here only pertain to activity during crises, all peacekeeping deployments in these data are therefore considered military involvement actions. Seven percent of the crises experienced assurance.

The third type of UN action is diplomatic engagement, which includes the dispatch of a special representative, the dispatch of a fact-finding mission, the offer of mediation, and the provision of mediation. These are cases where the UN is actively involved in seeking resolution without using tangible leverage. Both offers and the provision of mediation are included because offers of mediation typically imply that the Secretary-General or Special Representative of the Secretary-General are engaging in consultations with at least one of the sides in dispute. Fourteen percent of the crises in the data experienced diplomatic engagement by the UN.

Finally, intimidation is coded as having occurred when the UNSC or General Assembly issues a clear call for actors in crisis to adopt a course of action, condemns the crisis actors for their behavior, threatens sanctions or military involvement, or implements sanctions. Each of these actions offers some indication of the UN’s desire for the crisis actors to heed the UN’s will but does not involve operational deployment. Twenty-four percent of the crises experienced some form of intimidation.

With these categories of involvement so defined, they enter into the analysis in two ways. First, I include as independent variables counts of the number of actions of each of the four types that the UN undertakes. Through the course of a crisis, these variables count how much the UN has been involved in these ways up to the point of observation. Since authorizations and deployments typically but not always go hand in hand, the count variables only increase by 1 at the point of an authorization when it also is followed by deployment.¹⁰ Second, I include a measure of duration since the last instance of UN involvement of a particular type in order to see how the impact of UN involvement changes as time passes. One issue that arises when measuring duration as a simple count of the number of elapsed days is that the value of 0 may indicate that intervention occurred on a particular day or it might indicate that there has not been any intervention yet. To address this issue, I construct a variable that measures duration similarly to how one might measure the dosage of a particular drug remaining in a subject’s system. On the day of a UN action, this measure is equal to 1, and then experiences exponential decay with a scaling factor equal to the mean crisis length.¹¹ Higher values of this measure thus indicate more recent interventions, while lower values indicate less recent interventions. I also include the square of this measure, as doing so allows the UN activities to have different effects over the course of their involvement. Including these variables also allows for the relaxing of the restrictive proportional hazards assumption in the Cox model.¹²

A model as described above produces estimates of the individual effects of each involvement type while holding constant other types of involvement. It is worth exploring whether there are some interactive effects where involvement of a certain type might have a different effect depending on what else the UN is doing. Specifically, it is important to see the effects of diplomatic engagement and intimidation – the types of involvement that do not involve observational deployment – when military involvement is used as well. If one of the shortcomings of both diplomatic engagement and intimidation is that they face a serious cheap talk problem because the UN tends to be biased toward peace, then it is expected that they can more directly affect the incentives of the disputants when there is actually a significant contribution of resources in accompaniment. Models are thus estimated with the counts and durations of diplomatic engagement and intimidation included only when military involvement has previously occurred or is concurrent.

UN involvement does not occur at random and could be conditioned on the relative difficulty of resolution. To address this problem while not introducing issues related to model dependence, I follow Gilligan & Sergenti (2006) and use matching methods so that we can see how crises that experienced UN involvement differ in trajectory from crises that did not but that are similar with regard to a number of factors that shape their ex ante ease or difficulty of settlement.¹³ Specifically, coarsened exact matching is used to balance the data with respect to the following confounding variables (Iacus, King & Porro, 2012).¹⁴ These variables are all expected to affect the durations and outcomes of crises because they relate to the underlying levels of intractability in the disputes.¹⁵

A dichotomous variable of whether both sides of the crisis experienced violence at the onset of their crises is included as a direct measure of conflict severity. Since this measure picks up activity at crisis onset, it does not raise concerns related to the sequence of violence and UN involvement. Another potentially confounding variable is the contiguity between the dyadic actors, where contiguity is considered as true if the states share a border or are separated by less than 450 km of water. Contiguous states simultaneously have greater access to information about each other but at the same time face more of a credible commitment problem, since non-contiguous states can simply withdraw with less fear of surprise attack during disarmament. The next variable used in the matching captures the response times that the actors had in their crises. When actors take a while to respond to a crisis trigger, this is an indication that they are willing to be more patient in moving toward resolution. When the actors respond immediately, it is clear that the situation is of utmost priority and will likely progress quickly. From the ICB data, I record the maximum amount of time that either actor used to respond to its crisis trigger. For the coarsening, response times that are less than a week are separated from those that are longer. I also match on whether the crises are part of a protracted conflict and whether there is an ethnic dimension to the conflict, as both are observable indicators of intractability. Finally, I match on how many crisis actors there are in the overall crises, which should additionally capture how difficult the crises are to resolve ex ante. For the coarsening of this variable, I distinguish between crises that have just one or two crisis actors and multilateral crises with more.

After the data are balanced using the coarsened exact matching, the multinomial logit models are run. Since the response times and number of crisis actors were coarsened and thus not exactly balanced across the cases with and without UN involvement, they are also included as control variables. I also include information regarding the time from the start of the crisis until the first instance of UN involvement because timing can have an important effect on the ability to influence crisis trajectory (Regan & Stam, 2000). The same problem as that related to measuring involvement duration arises, in that a value of 0 when there has not yet been any involvement should not be the same value as when there is immediate UN involvement. In this regard, I construct a timing variable that also ranges from 0 to 1, with 1 being immediate involvement and 0 being no involvement. In between are the involvements that were not immediate and are calculated by exponential decay with a scaling factor equal to the mean crisis length. I also include the square of this measure to allow for curvilinear effects of timing. To go along with these timing variables, I include a dummy variable of whether the UN is involved at all. This allows for movement from 0 to 1 in the count variables to be distinguished from the mere occurrence of any UN actions that brings into play these timing variables. Two additional control variables are included to rule out other alternative explanations. First, I control for whether or not one of the sides of the crisis dyad is a permanent member of the UN Security Council (P-5) because Beardsley & Schmidt (forthcoming) have shown that P-5 involvement can substantially decrease UN involvement in international crises. Second, I control for whether or not regional security organizations were involved in the crisis, since regional involvement can be a substitute for UN involvement.

Results and discussion

Tables II and III present the findings of the multinomial logit models using the data that have been weighted using coarsened exact matching, with each column pertaining to the different ways crises can end.¹⁶ Model 1 includes all the different types of involvement and Model 2 focuses on the impact of diplomatic engagement when military involvement is also present.¹⁷ For Model 2, analogous variables for ‘other’ types of involvement are included as well, so that the comparison in both models is to no UN involvement. An additional model of the effect of intimidation when military involvement has occurred could not be estimated because too few events with this combination also experienced stalemate.

The values reported are the coefficients – interpreted as average treatment effects – along with the standard errors. The coefficients reported in Table II only have limited utility in portraying the UN’s effects on crisis duration because of the inclusion of the quadratic terms and the fact that an increase in the count variables implies a resetting of the duration variables. Linear-combination calculations are used to assess how increases in the counts and durations affect the hazard rates together. The five panels in Figure 1 depict the substantive findings as relative risks associated with an increase by 1 in the count of each type of UN action, as well as different time periods following such an increase.¹⁸ The relative risks are calculated as the ratio between the change in probability of a particular outcome and the baseline probability.¹⁹ The icons are placed at the average values, and the lines are the part of the 90% confidence interval that reach toward 0, which helps to gauge whether a relative risk is statistically significant or not.

Table II.

Multinomial logit models of crisis outcomes after coarsened exact matching

	Model 1: Separate types
	Compromise	Victory	Stalemate
Count of assurance	0.221 (0.365)	1.347^* (0.716)	−0.943 (3.445)
Duration of assurance	−25.09^** (5.046)	−24.09^* (12.79)	5.744 (35.33)
(Duration of assurance)⁁2	24.98^** (4.945)	24.89^** (12.45)	−6.803 (30.67)
Count of diplomatic engagement	2.056^** (0.528)	0.657 (1.098)	1.973 (2.862)
Duration of diplomatic engagement	7.564^* (3.928)	13.29^** (5.664)	−1.796 (4.755)
(Duration of diplomatic engagement)⁁2	−9.903^** (4.101)	−13.83^** (6.053)	0.968 (4.872)
Count of military involvement	0.493 (0.720)	0.708 (0.501)	0.0949 (0.843)
Duration of military involvement	7.639 (7.502)	9.169 (10.83)	13.90 (19.16)
(Duration of military involvement)⁁2	−10.99 (7.125)	−14.49 (11.93)	−15.57 (19.40)
Count of intimidation	0.0810 (0.109)	0.0790 (0.236)	−0.274 (0.586)
Duration of intimidation	4.369 (4.843)	17.10^** (7.469)	−5.495 (12.87)
(Duration of intimidation)⁁2	−4.680 (4.933)	−16.75^** (7.264)	6.211 (12.03)
Any UN involvement	−3.108 (3.743)	0.917 (4.419)	6.492^* (3.811)
Haste of UN involvement	9.201 (9.903)	−7.357 (15.16)	−20.04 (12.39)
(Haste of UN involvement)⁁2	−6.235 (6.916)	5.708 (11.26)	13.60 (9.669)
Contiguity	−0.504^* (0.264)	1.277^** (0.634)	−0.655 (0.445)
Protracted conflict	0.0625 (0.395)	0.507 (0.608)	1.090^** (0.460)
Violent trigger	−0.927^* (0.494)	0.0577 (0.568)	1.707^** (0.403)
Maximum response time	−0.0173^** (0.00687)	−0.0114^* (0.00605)	−0.0184^** (0.00611)
Ethnic conflict	−0.289 (0.300)	0.0705 (0.513)	−1.206^** (0.514)
Number of crisis actors	0.396^** (0.0938)	−0.671^* (0.346)	−0.388 (0.238)
P-5 crisis actor	0.685^** (0.253)	−0.0414 (0.553)	0.148 (0.403)
Regional organization	−0.0107 (0.325)	0.689 (0.557)	−1.041^** (0.385)
Elapsed time	−0.0131^** (0.00375)	−0.00702 (0.00620)	−0.00273 (0.00281)
(Elapsed time)⁁2	2.00e-05^** (8.05e-06)	7.37e-06 (8.81e-06)	−1.91e-06 (2.44e-06)
(Elapsed time)⁁3	−9.34e-09^** (4.70e-09)	−2.78e-09 (3.09e-09)	5.01e-10 (3.94e-10)
Constant	−4.573^** (0.695)	−6.469^** (1.140)	−4.997^** (0.789)
N	73,211	73,211	73,211

*p < .1

**p < .05 in a two-tailed test.

Figure 1.

Relative risks of crisis terminations associated with UN involvement durations.

A number of findings emerge from the results, suggesting that there are substantial trade-offs associated with various types of UN involvement. The following discussion proceeds by taking up each type of intervention in turn. UN interventions that comprise military involvement activities such as peacekeeping has the most positive effect, from the standpoint of peace advocacy, on the timing of victory. In the short and medium runs, military involvement substantially lengthens the time until victory and thereby reduces the risk that the crisis actors will fall victim to massive losses. The trade-off here is that military involvement also lengthens the time until compromise. While this fits the mold of a procedural effect in which military involvement simply takes time to put in place, the fact that it still does not have a significant shortening effect at six months indicates that military involvement tends to shape the incentives such that at least one of the sides feels prematurely protected from the sting of battle and thus less eager to settle.

We see the potential for perverse incentives more clearly in relation to assurance. Actions of assurance appear to have a counterproductive effect on the pace of peacemaking. In the short run, they speed up the time until victory and delay the ability for crises to end in stalemate. In the long run, they lengthen the time until compromise. Two downsides to intervention appear to be at work here. In the short run, assurances such as the promise of peacekeeping will potentially give the belligerents a deadline in which they will feel enticed to make as many gains as possible before the peacekeepers help lock in the status quo. This helps explain why the risk of victory is so high after the initial offer of assurances but then falls and does not have a statistically significant effect after a month. The fall in the propensity for compromise is consistent with the expectation that assurances and military involvement can create an artificial sense of security that reduces the incentives to make any progress at the bargaining table. Assurances, like military involvements, tend to lock in periods of simmering hostilities that neither progress toward agreement nor risk falling into total war.

UN activity that involves diplomatic engagement produces trade-offs that are the inverse of military involvement. While military involvement tends to delay both victory and compromise, diplomatic engagement tends to hasten both types of outcome in the long run. That diplomatic engagement increases the propensity for compromise after three months and victory after two months is consistent with both the expectation that third parties such as the UN can help reduce uncertainty and provide political cover and the expectation that some disputants might use peace processes for insincere motives. The results are consistent with the notion that sincere types can use UN involvement to help them identify and reach difficult compromises while insincere types can use the UN involvement to stall until the time is right to re-engage and push for a more full victory.

The remaining type of UN action, intimidation, does not have much of an effect on the timing of crisis outcomes. The only effect is that, in the long run, intimidation increases the likelihood of victory. One explanation is that when intimidation attempts do provide more than just cheap talk, it is when they actually signal that the UN is not willing or able to take more concrete steps toward peacemaking. So, when the UN has urged restraint and then months go by without resolution, actors with relatively strong strategic positions might feel undeterred by international involvement and then go ahead and push for a decisive victory.

Another way to see these effects is through counterfactual predictions. The 2002 Kaluchak crisis between India and Pakistan did not include any of the UN involvement considered here and ended in stalemate after 186 days. If we set the control variables to the same values as this crisis, we can then ask how the trajectory of the crisis might have been different had the UN become involved in various capacities at, say, six days into the crisis. To form such expectations, I use the Clarify software (King, Tomz & Wittenberg, 2000) and generate the predicted probabilities of each type of outcome at six days, 36 days, and 186 days.

Starting with the impact of military involvement, the decline in daily risk of victory is quite pronounced as it falls from 0.18% to 0.003% immediately after deployment and from 0.14% to 0.004% 30 days later. The decline in ability to compromise is also quite substantial during UN military involvement, as it falls from 0.49% to 0.04% immediately after involvement and from 0.33% to 0.04% 30 days later. Turning to diplomatic engagement, the likelihood of compromise increases substantially after 180 days when the UN becomes involved in this capacity – from 0.08% to 0.61% – but so does the likelihood of victory – from 0.06% to 0.27%. To quantify other notable effects, we see that assurance decreases the predicted probability of compromise after six months from 0.08% to 0.002%, and intimidation increases the predicted probability of victory after six months from 0.06% to 0.72%. In hindsight, diplomatic engagement could have been prudent in this case although not necessarily practical because of India’s general resistance to outside interference. The predicted probability of one side achieving a substantial victory after diplomatic engagement was still rather small – this is true in a hypothetical sense from the probabilities recovered in this exercise and it is more importantly true in a practical sense when considering how costly it would have been for either India or Pakistan to achieve major military victory over each other in 2002 – and it would have been worth seeing if diplomatic engagement could have encouraged compromise instead of the stalemate that eventually resulted.

Turning to Model 2 in Table III, we observe the effect of diplomatic engagement while military involvement was introduced earlier in the crisis. When accompanied by the commitment of military resources, diplomatic engagement no longer is associated with a greater propensity for one-sided victory, presumably in part because the military deployment helps prevent disputants from exploiting vulnerabilities while they stall. More dramatically, we observe that in the long run diplomatic engagement substantially increases the opportunity for crises to end by stalemate – the relative risk is greater than 146 at six months after involvement. This suggests that while the military deployment still impedes substantial progress toward compromise, diplomatic engagement can at least help the actors end the crisis and abide by the status quo that the UN peacekeeping forces entrench. Strong UN action that is needed to prevent a complete collapse in negotiations appears to benefit from subsequent diplomatic engagement that can help the crisis at least move to some form of conclusion, even if only a stalemate.

Table III.

Multinomial logit models of crisis outcomes after coarsened exact matching

	Model 2: Interaction of types
	Compromise	Victory	Stalemate
Count of diplomatic engagement after military inv.	0.907^** (0.309)	0.290 (0.437)	1.506^** (0.408)
Duration of diplomatic engagement after military inv.	2.931 (7.133)	15.24^* (8.136)	39.22^** (9.010)
(Duration of diplomatic engagement after military inv.)⁁2	−4.056 (7.498)	−17.77^** (8.387)	−42.82^** (9.575)
Count of other UN involvement	−0.124 (0.0929)	0.157 (0.107)	−0.574^** (0.250)
Duration of other UN involvement	−4.005 (3.726)	10.37 (7.492)	−32.36^** (13.54)
(Duration of other UN involvement)⁁2	3.369 (3.839)	−10.39 (7.565)	32.37^** (13.71)
Any UN involvement	−2.155 (4.527)	−2.657 (5.304)	4.364 (4.174)
Haste of UN involvement	8.430 (12.37)	5.556 (18.85)	−5.830 (14.40)
(Haste of UN involvement)⁁2	−6.086 (8.404)	−2.789 (14.02)	2.873 (10.36)
Contiguity	−0.545^** (0.253)	1.067 (0.678)	−0.675 (0.440)
Protracted conflict	0.0706 (0.381)	0.456 (0.559)	1.168^** (0.467)
Violent trigger	−0.971^* (0.504)	0.0231 (0.499)	1.660^** (0.385)
Maximum response time	−0.0159^** (0.00666)	−0.00472^** (0.00179)	−0.0196^** (0.00655)
Ethnic conflict	−0.365 (0.311)	−0.246 (0.474)	−1.165^** (0.448)
Number of crisis actors	0.378^** (0.0953)	−0.675^** (0.232)	−0.402^* (0.228)
P-5 crisis actor	0.618^** (0.240)	0.115 (0.480)	0.122 (0.402)
Regional organization	0.0374 (0.333)	0.760 (0.538)	−1.076^** (0.367)
Elapsed time	−0.0127^** (0.00371)	−0.00713 (0.00514)	−0.00250 (0.00260)
(Elapsed time)⁁2	1.79e-05^** (6.91e-06)	6.52e-06 (5.87e-06)	−1.57e-06 (2.03e-06)
(Elapsed time)⁁3	−7.68e-09^** (3.71e-09)	−1.66e-09 (1.66e-09)	5.22e-10 (3.58e-10)
Constant	−4.482^** (0.667)	−6.327^** (0.994)	−4.971^** (0.794)
N	73,211	73,211	73,211

*p < .1

**p < .05 in a two-tailed test.

Implications and conclusions

The UN can have a number of effects on the duration of international crises, depending on the type of involvement. The different relationships for the various types of UN activity demonstrate the importance of disaggregating interventions. Hence, we see direct advantages of using a dataset like that of UNIEvents. Further studies might disaggregate the activity further to pull out the varying effects of, say, traditional peacekeeping versus multidimensional peacekeeping.

Consistent with the existing literature, there are real benefits but also real risks to UN involvement in international crises. Corroborating Betts (1994) and Greig & Diehl (2005), it appears that UN military involvement prevents both victory and compromise from occurring. This is good news when a crisis is prone to degenerate into massive hostilities but bad news when a crisis would likely progress toward a peaceful settlement in the absence of such strong involvement.

Diplomatic engagements also exhibit a trade-off as they accelerate both compromise and victory over time. In the long run, disputants that are sincere in their use of the peace process to resolve their bargains more efficiently can benefit from UN assistance in facilitating communication and signaling to domestic audiences that compromise is prudent. At the same time, disputants that are insincere can use the peace process to increase their prospects of victory. That being said, UN involvement in diplomatic engagement can be quite prudent when the disputants have provided signals regarding their sincerity and when the probability of victory is likely to remain at low levels even when stalling does occur.

Mere actions of assurance and intimidation appear to have more downside than upside – assurances tend to delay compromise and stalemate while both assurance and intimidation can hasten victory. Like military involvement, assurances can over time remove the incentives for the disputants to make progress on resolving their disputes. Unlike military deployments, mere assurances tend to produce a perverse short-term effect in which the promise of peacekeeping or monitoring to come could encourage the sides to take swift aggressive action before the peacekeepers hinder any further changes in the status quo. Actions of intimidation when followed by lengthy crises can indicate that the international community is not going to take more substantive action to help prevent one side from gaining a substantial upper hand.

Aside from building on existing studies, the research has two overarching implications. First, we should not dismiss the UN as irrelevant when it comes to shaping the dynamics of conflict bargaining. UN military involvement in particular has a profound effect on preventing the sides from achieving victory. When given enough time, diplomatic engagement can have rather strong effects in encouraging compromise, as well as in helping crises end in stalemate when UN military involvement has reduced progress toward full resolution. Second, once we recognize that the UN can strongly shape crisis durations, there is a call for caution in prescribing hasty UN action. Military involvement can delay incentives to compromise, diplomatic engagement and intimidation can increase the potential for victory in the long run, and assurance can do both. UN involvement is not without side effects, and the decision to become involved must be made with eyes wide open to both the benefits and risks.

Footnotes

Replication data

The dataset and do-files for the empirical analysis in this article can be found at .

Acknowledgements

I am indebted to Holger Schmidt, Victor Asal, Birger Heldt, and the participants of the Folke Bernadotte Academy Working Group on Conflict Prevention for invaluable comments.

Funding for the coding of the UNIEvents data was provided by the Folke Bernadotte Academy.

Notes

References

Balch-Lindsay

Dylan

Enterline

Andrew J

Joyce

Kyle A

(2008) Third-party intervention and the civil war process. Journal of Peace Research 45(3): 345–363.

Beardsley

Kyle

(2008) Agreement without peace? International mediation and time inconsistency problems. American Journal of Political Science 52(4): 723–740.

Beardsley

Kyle

(2009) Intervention without leverage: The political determinants of weak mediation. International Interactions 35(3): 272–297.

Beardsley

Kyle

(2010) Pain, pressure, and political cover: Explaining mediation incidence. Journal of Peace Research 47(4): 395–406.

Beardsley

Kyle

(2011) The Mediation Dilemma. Ithaca, NY: Cornell University Press.

Beardsley

Kyle

Schmidt

Holger

(forthcoming) Following the flag or following the charter? Examining the determinants of UN involvement in international crises, 1945–2002. International Studies Quarterly. DOI: 10.1111/j.1468-2478.2011.00696.x.

Beardsley

Kyle

Quinn

David

Biswas

Bidisha

Wilkenfeld

Jonathan

(2006) Mediation style and crisis outcomes. Journal of Conflict Resolution 50(1): 58–86.

Betts

Richard K

(1994) The delusion of impartial intervention. Foreign Affairs 73(6): 20–33.

Box-Steffensmeier

Janet M

Reiter

Dan

Zorn

Christopher

(2003) Nonproportional hazards and event history analysis in international relations. Journal of Conflict Resolution 47(1): 33–53.

10.

Brecher

Michael

Wilkenfeld

Jonathan

(2000) A Study of Crisis. 2nd edn. Ann Arbor, MI: University of Michigan Press.

11.

Carter

David B

Signorino

Curtis S

(2010) Back to the future: Modeling time dependency in binary data. Political Analysis 18(3): 271–292.

12.

Chiozza

Giacomo

Goemans

Hein E

(2004) International conflict and the tenure of leaders: Is war still ex post inefficient? American Journal of Political Science 48(3): 604–619.

13.

Collier

Paul

Chauvet

Lisa

Hegre

Håvard

(2008) The security challenge in conflict-prone countries. Copenhagen Consensus 2008 April.

14.

Debs

Alexandre

Goemans

Hein E

(2010) Regime type, the fate of leaders, and war. American Political Science Review 104(3): 430–445.

15.

DeRouen

Karl

(2003) The role of the UN in international crisis termination, 1945–1994. Defence and Peace Economics 14(4): 251–260.

16.

Diehl

Paul F

Reifschneider

Jennifer

Hensel

Paul R

(1996) United Nations intervention and recurring conflict. International Organization 50(4): 683–700.

17.

Doyle

Michael W

Sambanis

Nicholas

(2000) International peacebuilding: A theoretical and quantitative analysis. American Political Science Review 94(4): 779–801.

18.

Doyle

Michael W

Sambanis

Nicholas

(2006) Making War and Building Peace: United Nations Peace Operations. Princeton, NJ: Princeton University Press.

19.

Favretto

Katja

(2009) Should peacemakers take sides? Major power mediation, coercion, and bias. American Political Science Review 103(2): 248–263.

20.

Filson

Darren

Werner

Suzanne

(2002) A bargaining model of war and peace: Anticipating the onset, duration, and outcome of war. American Journal of Political Science 46(4): 819–837.

21.

Fortna

Virginia Page

(2003) Inside and out: Peacekeeping and the duration of peace after civil and interstate wars. International Studies Review 5: 97–114.

22.

Fortna

Virginia Page

(2004a) Does peacekeeping keep peace? International intervention and the duration of peace after civil war. International Studies Quarterly 48(2): 269–292.

23.

Fortna

Virginia Page

(2004b) Interstate peacekeeping: Causal mechanisms and empirical effects. World Politics 56: 481–519.

24.

Gilligan

Michael J

Sergenti

Ernest

(2006) Evaluating UN peacekeeping with matching methods to improve causal inference. Unpublished manuscript, Department of Politics, New York University.

25.

Goemans

Hein E

(2000) War and Punishment: The Causes of War Termination and the First World War. Princeton, NJ: Princeton University Press.

26.

Goemans

Hein E

(2009) Risky but rational: War as an institutionally induced gamble. Journal of Politics 71(1): 35–54.

27.

Greig

J Michael

(2001) Moments of opportunity: Recognizing conditions of ripeness for international mediation between enduring rivals. Journal of Conflict Resolution 45(6): 691–718.

28.

Greig

J Michael

Diehl

Paul F

(2005) The peacekeeping–peacemaking dilemma. International Studies Quarterly 49(4): 621–645.

29.

Hewitt

Joseph

(2003) Dyadic processes and international crises. Journal of Conflict Resolution 47: 669–692.

30.

Iacus

Stefano M

King

Gary

Porro

Giuseppe

(2012) Causal inference without balance checking: Coarsened exact matching. Political Analysis 20 (1): 1–24.

31.

King

Gary

Tomz

Michael

Wittenberg

Jason

(2000). Making the most of statistical analyses: Improving interpretation and presentation. American Journal of Political Science 44(2): 347–361.

32.

Kuperman

Alan J

(2008) The moral hazard of humanitarian intervention: Lessons from the Balkans. International Studies Quarterly 52(1): 49–80.

33.

Kydd

Andrew

(2003) Which side are you on? Bias, credibility, and mediation. American Journal of Political Science 47(4): 597–611.

34.

Kydd

Andrew H

(2005) Trust and Mistrust in International Relations. Princeton, NJ: Princeton University Press.

35.

Kydd

Andrew H

(2006) When can mediators build trust? American Political Science Review 100(3): 449–462.

36.

Kydd

Andrew H

Walter

Barbara

(2002) Sabotaging the peace: The politics of extremist violence. International Organization 56: 263–296.

37.

Lake

David A

Rothchild

Donald

(1996) Containing fear: The origins and management of ethnic conflict. International Security 21: 41–75.

38.

Nigel

Hashimoto

Barry

Reiter

Dan

(2008) Ensuring peace: Foreign-imposed regime change and postwar peace duration, 1914–2001. International Organization 62: 717–736.

39.

Luttwak

Edward N

(1999) Give war a chance. Foreign Affairs 78(4): 36–44.

40.

Mattes

Michaela

Savun

Burcu

(2009) Fostering peace after civil war: Commitment problems and agreement design. International Studies Quarterly 53(3): 737–759.

41.

Powell

Robert

(2004) The inefficient use of power: Costly conflict with complete information. American Political Science Review 98(2): 231–241.

42.

Powell

Robert

(2006) War as a commitment problem. International Organization 60: 169–203.

43.

Rauchhaus

Robert

(2006) Asymmetric information, mediation and conflict management. World Politics 58(2): 207–241.

44.

Rauchhaus

Robert

(2009) Principal–agent problems in humanitarian intervention: Moral hazard, adverse selection, and the commitment dilemma. International Studies Quarterly 53(4): 871–884.

45.

Regan

Patrick M

Stam

Allan C

(2000) In the nick of time: Conflict management, mediation timing, and the duration of interstate disputes. International Studies Quarterly 44: 239–260.

46.

Reiter

Dan

(2009) How Wars End. Princeton, NJ: Princeton University Press.

47.

Richmond

Oliver

(1998) Devious objectives and the disputants’ view of international mediation: A theoretical framework. Journal of Peace Research 35(6): 707–722.

48.

Schultz

Kenneth A

(2010) The enforcement problem in coercive bargaining: Interstate conflict over rebel support in civil wars. International Organization 64(2): 281–312.

49.

Slantchev

Branislav L

(2003) The principle of convergence in wartime negotiations. American Political Science Review 97(4): 621–632.

50.

Slantchev

Branislav L

(2004) How initiators end their wars: The duration of warfare and the terms of peace. American Journal of Political Science 48(4): 813–829.

51.

Smith

Alastair

Stam

Allan

(2003) Mediation and peacekeeping in a random walk model of civil and interstate war. International Studies Review 5(4): 115–135.

52.

Snyder

Glenn H

(1984) The security dilemma in alliance politics. World Politics 36(4): 461–495.

53.

Stedman

Stephen John

(1997) Spoiler problems in peace processes. International Security 22(2): 5–53.

54.

Svensson

Isak

(2007) Mediation with muscles or minds? Exploring power mediators and pure mediators in civil wars. International Negotiation 12(2): 229–248.

55.

Tarar

Ahmer

(2006) Diversionary incentives and the bargaining approach to war. International Studies Quarterly 50(1): 169–188.

56.

Tarar

Ahmer

Leventoglu

Bahar

(2009) Public commitment in crisis bargaining. International Studies Quarterly 53(3): 817–839.

57.

Toft

Monica D

(2009) Securing the Peace: The Durable Settlement of Civil Wars. Princeton, NJ: Princeton University Press.

58.

Trager

Robert F

Vavreck

Lynn

(2011) The political costs of crisis bargaining: Rhetoric and the role of party. American Journal of Political Science 55(3): 526–545.

59.

Walter

Barbara F

(2002) Committing to Peace: The Successful Settlement of Civil Wars. Princeton, NJ: Princeton University Press.

60.

Werner

Suzanne

Yuen

Amy

(2005) Making and keeping peace. International Organization 59(2): 261–292.

61.

Wilkenfeld

Jonathan

Young

Kathleen J

Quinn

David M

Asal

Victor

(2005) Mediating International Crises. New York: Routledge.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB