Third-party knowledge and success in civil war mediation

Abstract

Within this article, we explore how third-party knowledge of ongoing conflicts shapes the ability of mediators to successfully end conflicts through negotiated settlements. Since the primary role of mediators is to share information, and combatants have incentives to misrepresent information, contextual knowledge about the conflict and actors is critical. We argue that experienced diplomats with greater knowledge of the civil war state, close knowledge of the combatants, and connections with civil society are less vulnerable and more effective in mediation efforts. We propose that third parties seeking a diplomatic solution to ongoing conflicts may be more successful when they maintain strong diplomatic knowledge of the disputants as well as knowledge of the processes by which previous mediation efforts have attempted to resolve the dispute. Using a seemingly unrelated bivariate probit model on peace agreements from 1989 to 2005, we find strong support that diplomatic knowledge matters significantly.

Keywords

bargaining civil war diplomacy information knowledge mediation

Introduction

As the Syrian civil war approaches its ninth year, the international community is once again launching renewed efforts to bring the parties to the negotiating table to settle the dispute. At the same time, regional partners and international organisations are restarting talks between the South Sudanese Government and the rebel forces, Sudan People’s Liberation Movement-in-Opposition (SPLM-IO). The new talks in the South Sudanese civil war are being revived in the shadow of the failed 2015 peace process, which opened a new chapter of instability in the world’s youngest country. The stakes are high for international mediation efforts. Though the international community has many tools to manage ongoing conflicts, mediation represents one of the most attractive options (Regan et al., 2009). Diplomatic interventions are relatively inexpensive as compared with military interventions or economic sanctions. Understanding what makes mediators effective is therefore a highly pertinent question for scholars and policymakers to address.

Much of the past literature agrees that mediators are successful in large part due to their ability to credibly transfer information during negotiations (Kydd, 2003; Regan and Aydin, 2006; Regan et al., 2009; Savun, 2008). Past studies generally assume that information operates as a one-way street, where mediators offer information gleaned prior to, and during, negotiations, and it is either accepted or ignored by combatants. While true, this body of work largely makes an implicit assumption that mediators receive accurate information from the combatants. At the same time, much of the previous work on bargaining and civil war dynamics notes that combatants often have incentives to misrepresent information during the war so as to gain leverage in disputes (Fearon, 1995; Mattes and Savun, 2010; Thomas et al., 2016; Walter, 2009). If both rebels and the government have incentives to provide misinformation to one another, it stands to reason that belligerents may also offer misinformation to mediators as a way to garner international support during the dispute. If belligerents make disingenuous demands during the negotiation process, how can mediators parse bluster from the actual reservation point(s) of each side?

One way that mediators may overcome this challenge is by maintaining strong knowledge of one or both sides prior to the beginning of negotiations. We posit that mediators that have close, or inside, knowledge of combatants and the context in which the conflict is occurring are more effective at resolving armed conflicts as compared with mediators that are removed from the conflict. As noted by Savun (2008), outside of the mediator’s relative bias, international actors may have key information about the conflict that assists with their ability to convince belligerents to come to an agreement. Reid (2017) notes that mediators may have key contextual leverage over disputants, where their intimate knowledge of the dispute allows them to craft more durable settlements. Building on this work, we posit that effective mediators may have strong knowledge of one or both sides, as well as a strong familiarity with past efforts to resolve the dispute. This diplomatic knowledge and process knowledge allows international mediators to better identify disingenuous claims from the true reservation point of each side.

To assess these claims, we provide a brief overview of previous research and offer a novel theory that emphasises the key role of mediator knowledge of the disputants and past processes of settling the dispute. We then assess the accuracy of this theory by statistically examining the success of all mediation attempts in intrastate conflicts from 1989 to 2005. The results largely confirm that both diplomatic and process knowledge held by the mediator significantly increase the likelihood that conflicts will end in a peace agreement. Finally, this article concludes with policy recommendations and avenues for future research.

Previous research

Much of the previous research on civil war dynamics and conflict resolution has been defined by the bargaining approach to disputes, emphasising the importance of commitment problems in stymieing negotiations and prolonging disputes (Fearon, 1995; Hartzell and Hoddie, 2003; Walter, 2002). Specifically, given an incentive for parties (primarily for the government) to defect from any agreement, it is difficult for either side to credibly commit to a settlement. The bargaining process during civil war is also complicated by issues of information asymmetry. As noted by Walter (2009), rebel groups often have numerous incentives to misrepresent information throughout the conflict. As inherently clandestine organisations, rebel groups must be careful with the information that they release, as governments with accurate knowledge of the insurgents’ strength and location should have little problem defeating their opponents. Rebel groups may therefore leverage asymmetries of information to increase their ability to bargain effectively with the central government (Fearon, 1995; Lake and Rothchild, 1998; Thomas et al., 2016; Walter, 2009).

With that in mind, there is some scepticism that information asymmetries complicate the bargaining process during civil wars (Fearon, 2004). As noted by Fearon (2004), belligerents should quickly become apprised of one another’s resolve and relative capabilities after fighting begins. Melin and Svensson (2009) argue that governments often are loath to enter into the mediation process given that negotiating with rebels generally carries significant reputation costs. This is one reason why governments prefer to defeat rebels on the battlefield as opposed to entering into negotiations (Bapat, 2005). Rebel groups with greater military strength are therefore capable of pushing the state to enter into negotiations (Cunningham et al., 2009). As noted by Greig et al. (2018), rebel group advances on the battlefield quickly reduce asymmetries of information, significantly increasing the likelihood that insurgents will achieve a favourable outcome. Stronger rebel groups, therefore, often push the government to seek third-party assistance in negotiating an end to conflicts (Clayton, 2013; Greig, 2015).

Despite these reservations about the validity of information problems as a barrier to settlements, scholars of international mediation often point to the ability of third parties to address asymmetries of information in order to resolve conflicts (Kydd, 2003; Regan and Aydin, 2006; Regan et al., 2009; Savun, 2008). While mediators may offer positive or negative inducements to incentivize a settlement (Lake and Rothchild, 1996; Zartman and Touval, 1985) or provide political cover for elites to make unpopular decisions (Beardsley, 2010), mediators often are perceived to be successful given their ability to share credible information with one or both sides of a dispute (Regan et al., 2009). As noted by Regan et al. (2009), mediation is one of the least costly tools available to the international community, and the transfer of reliable information is a critical tactic pursued by mediators to assuage combatant concerns and update their beliefs about the outcome of the conflict.

Unfortunately, many theories on the ability of mediators to credibly transfer information often rely on the implicit assumption that mediators have a clear understanding of the grievances, necessary concessions, and reservation points of the disputants needed to end the conflict. In many ways, this conceptualisation of mediation treats information transfers as a one-way-street, where mediators consume relevant information, and use this knowledge to convince one or both sides that an agreement is in their best interest. This conceptualisation may be less accurate if one or both of the belligerents is providing misinformation to the mediator as a way to increase their bargaining leverage. While rebels may provide misinformation to gain leverage over the incumbent government, this use of misinformation may also be employed during the mediation process.

In addition, rebel groups will often raise their demands as a way to demonstrate resolve rather than actually seek feasible concessions (Thomas et al., 2016). This further complicates the mediation process, as it makes it difficult for international mediators to discern which claim is bluster and which claim is closer to the insurgents’ reservation point. While this discussion underscores the use of misinformation by rebels, the government may also wish to use misinformation in their interaction with the mediator. Given the high reputation costs associated with entering into mediation (Greig and Regan, 2008; Melin and Svensson, 2009), members of the incumbent government may have little incentive to offer meaningful concessions to end fighting. Elites may therefore misrepresent their reservation points to the mediator to avoid costly accommodations of insurgent demands. Given these points, how can mediators accurately discern misinformation from earnest claims when they seek to identify the appropriate terms of a settlement?

Diplomatic and process knowledge – theory

To alleviate the misinformation flows coming from rebel groups and/or governments, mediators must have a good amount of information about the belligerents. We argue that the most effective mediators will have what is sometimes referred to as insider knowledge. Much like a judge who issues a ruling on an interstate conflict based on as much information as possible, mediators in civil wars must also be prepared with information outside of what the rebel groups and governments tell them. Given the likelihood of the disputants deliberately misrepresenting information to the mediators, the mediators must seek other avenues of information or knowledge to help them better assess the bargaining feasibility, potential concessions, and reservation points for the disputants. These other avenues of information can come about through process knowledge and diplomatic knowledge. We further argue that process knowledge and diplomatic knowledge are distinct from classic theories of mediation bias because they parse out information better.

This proposition differs from both the early work on impartiality in mediation (the outsider argument) as well as recent work on the importance of biased mediators in mediation. The outsider argument pushed for impartiality and no insider information. These early studies argued that if mediators were too familiar with the conflict, they would approach the bargaining table with preconceived biases, thus limiting effectiveness (Jackson, 1952; Northredge and Donelan, 1971; Young, 1967). This approach to mediation explicitly excluded anyone with ‘expert knowledge’ of the local conditions as well as those who may be considered a ‘specialist in respect to the conflict’ (Burton and Dukes, 1990: 199). By removing these participants, it was assumed that mediators would be impartial outsiders with value-free and neutral intentions (Burton and Dukes, 1990; Wehr, 1979; Wehr and Leaderach, 1991). This approach is reflected, to some degree, with more modern research that advocates for impartial mediation (Beber, 2012; Favretto, 2009).

These past approaches seem to conflate expertise and knowledge with partiality. While having an intimate knowledge of both the disputants and the historical context in which the conflict is occurring may indeed generate preconceptions, it may inoculate mediators to misinformation provided by the disputants. Though mediators, particularly within the United Nations, often receive exceptional background information as they begin the mediation process (Nathan et al., 2018), this training may not be comprehensive enough to substitute for decades of background information collected by some in the diplomatic community.

On the other hand, scholars have suggested that the outsider argument overlooks the strategic interaction between disputants and mediators that occurs during negotiations (Bercovitch et al., 1991). Carnevale and Arad (1996) and Melin (2011) argue that bias is not harmful, but rather helpful in successful mediations. This may especially be the case if a mediator is biased towards the conflict party that is about to concede the most in the agreements (Rauchhaus, 2006). Mediator states with biased preferences are perceived as more effective in mediation because biased mediators can act as guarantors for the weaker party and reduce the weaker party’s fears, making conflict resolution more likely (Kydd, 2003; Savun, 2008; Svensson, 2007, 2009; Touval, 1975; Walter, 2002). The logic is that mediator states that can bring their side to the table are seen as more credible in the information they offer. As noted by Kydd (2003), combatants may not trust the information offered by impartial mediators as they simply want to end the conflict. On the other hand, a biased mediator is perceived to be more credible given that their goal is to preserve the interest of their protégé.

Though mediator bias, or ‘partiality’, can help persuade actors to consider compromise and concessions, mediator bias does not necessarily help with parsing out misinformation provided by rebel groups and/or governments involved in civil wars. Having a preference over the outcome or an affinity towards one party does not inoculate a mediator state to misinformation. Without more accurate knowledge of the dispute, disputants, and resolution process, even biased mediators will have difficulty assessing the actors’ reservation points.

We argue that separate from bias or partiality (Elgström et al., 2003), mediation will be more successful in ending conflicts when they have insider knowledge of the conflict and conflict state. The logic is that mediators with insider knowledge are better able to recognise and identify credible information shared by the conflict parties compared with those who are less knowledgeable. Since combatants have incentives to misrepresent information, it is helpful for mediators to have prior knowledge of and familiarity with the conflict, government, and/or rebel groups so they will be less vulnerable to misrepresentation of information. Familiarity with disputants and the context of the conflict through diplomatic relations can help establish the social authority of the mediator and connectedness with the conflict parties. Process knowledge and diplomatic knowledge are two different ways that mediators can better parse through misinformation to assess which information is earnest and which procedures will be most effective for conflict resolution.

We define process knowledge as the mediator having familiarity with the dispute and dispute resolution methods that are used in particular civil war states. This type of mediator knowledge focuses on familiarity and past experience with the conflict resolution process in the conflict state as a mediator. Knowledgeable mediators possess in-depth knowledge of conflict context, cultural norms, language, social structures, and power hierarchies of the conflict state and rebel groups, so they are more effective mediators (Anderson and Olson, 2003; Svensson and Lindgren, 2013; United Nations, 2012). As noted by Regan et al. (2009: 137), ‘there are often cultural barriers, ethnic or religious cleavages, or histories of entrenched animosities that make credible information critical but scarce. In this environment, facilitating communication and building trust between opponents are instrumental to making peace possible’. Mediators can be more or less familiar with the preferred laws and rules influencing the process of conflict resolution of particular conflict states. The more that a mediator state understands the preferred rules and procedures of dispute resolution of a conflict state, the better able the mediator is with helping the disputing parties compromise and reach concessions. Based on this logic, we predict the following:

H1: Mediators with stronger familiarity with the dispute resolution process of a particular conflict state are more likely to successfully generate a peace agreement.

We conceptualise diplomatic knowledge as when a mediator has familiarity with the disputants and the cultural context of the state and sub-state actors through diplomatic relations with the conflict state and/or the rebel groups. These diplomatic relationships can be bilateral or multilateral, involving the mediator, the government, and/ or the rebel group. Diplomatic knowledge, as compared with process knowledge, focuses less on which procedures or commonly shared history or legal systems may lead to a successful termination of fighting. Rather, diplomatic knowledge is indicative of a mediator’s understanding of the particular characteristics of combatant groups through consistent and active engagement of the mediator government and the civil war state combatants. As a result of diplomatic insider knowledge such as ‘geographical and cultural closeness to the conflict’, the mediator’s ‘legitimacy is derived from in-depth knowledge of the situation and rests in the trust and acceptance of the conflict parties’ (Roepstorff and Bernhard, 2013: 166). A key part of communicative, facilitative strategies in mediation is the development of personal rapport with combatants (Bercovitch and Wells, 1993: 8). This rapport allows mediators to understand which claims are earnest and which are merely bluster. All else being equal, greater diplomatic knowledge provides greater rapport between the mediator and disputants.

While many mediators may have to spend significant time with disputants to gain this rapport, some states can already have a key understanding of the combatants prior to negotiations. For example, mediators from states that have diplomatic relations may have had repeated interactions with combatants through formal diplomatic avenues as well as informal interactions prior to the onset of multilateral talks. Mediator states that have operated embassies in civil war states can have diplomats that have a strong familiarity with members of the incumbent government with which they will eventually negotiate as part of mediation efforts.

Outside of official state relations, mediating states may host diplomatic missions sent by rebel organisations (Huang, 2016), such as the Taliban’s diplomatic office in Qatar.

Mediators may also represent states where rebel groups maintain diplomatic posts, such as the Taliban’s diplomatic office in Qatar, as a way to raise support for the war effort and/or gain international legitimacy for the organisation, as noted by Huang (2016). Within these diplomatic interactions, key government officials may have an opportunity to hear rebel perceptions of the war, key goals of the insurgency, as well as possible reservation points outside of the contentious venues of official mediation.

As noted by Regan and Stam (2000), there may be key areas of compromise of which disputants are aware, but are unwilling to discuss openly during mediation as it risks undermining their bargaining position. A mediator’s knowledge of the disputants prior to the start of negotiations may apprise them of what areas of compromise disputants are willing to concede on, topics that disputants are unable to concede on, as well as key points that disputants may be able to offer concessions, but only through the cover of mediation (Beardsley, 2010).

During prior interactions by third parties, combatants may privately offer key information that mediators can later use to help determine combatants’ real reservation points. We know from past research that previous interactions in the form of diplomatic interventions shorten the duration of civil wars (Regan and Aydin, 2006). We contend that as the relative bilateral relations strengthen between mediator states and combatants, mediators should be more effective at terminating civil wars. Such mediators bring expert power, informational power, and referent power (Carnevale, 2002) because they have strong ties to the state or rebel groups. Mediator states could additionally potentially share economic, social, or cultural characteristics with conflict states. Such familiarity can help regional states that are also mediator states better understand the root causes of civil conflicts, allowing these mediators the ability to more easily build trust between the combatants and the mediators (Bercovitch and Houston, 1995; Gartner, 2011). Overall, we hypothesise the following:

H2: Mediator states with stronger ties to combatants should be less vulnerable to misinformation and more likely to be successful at generating a peace agreement.

The classic illustration of these effects was part of the Lancaster House negotiations that helped to end the Rhodesian Civil War. Throughout the mid-late 1970s, the British Government became increasingly concerned with the prolonged guerilla war in its former colony. As guerilla attacks persisted, it was clear that the war had entered into a hurting stalemate (Stedman, 1991). Though the Thatcher Government was credited with the successful negotiation of the Lancaster House Agreement, talks had actually begun under the previous Labour Government and other outside actors (such as Henry Kissinger as part of an American delegation). Unfortunately, those talks had failed to generate a durable settlement.

Though there were early failures in the negotiations, the British were keenly suited to continue talks with the both the Rhodesian Government as well as the ZANU-ZAPU allied opposition (or the Patriotic Front). At the outset of the Lancaster House negotiations, the Commonwealth of Nations had officially recognised the British as the most capable of finding a settlement to the conflict, given that they abide by certain rules agreed to earlier (Novak, 2009: 158). As noted by Novak (2009), the British negotiators benefitted from considerable insider knowledge as part of their Track I and Track II efforts in negotiating a settlement. The Tory Party had long held certain cultural sympathies for the Smith Government in Rhodesia and were keen to remove past British sanctions on the regime (Novak, 2009). Similarly, cooperation between the leaders of the Commonwealth (particularly Sir Rampal – Secretary General of the Commonwealth) apprised British negotiators of the relative weakness of key actors in the new Rhodesian power-sharing government (particularly, Bishop Muzorewa).

Past diplomatic efforts, both during and outside of mediation, had given the negotiators a wealth of information on the disputants (Keels, 2019). While Lord Carrington himself had not participated actively in previous talks, the British had acquired extensive notes on the key facets of earlier talks as well as other critical intelligence (Stedman, 1991). These insights proved invaluable as parties navigated around the complex issue of land reform in Rhodesia. As the issue had derailed earlier talks, the British had to identify a way of addressing rebel concerns while not aggravating wealthy Rhodesian landowners. Building on past information collected, the British arranged for land reforms to be excluded from the talks while also allowing for external financing (largely from America) to be used to purchase and distribute land to disaffected Zimbabweans (Stedman, 1991: 183).

Research design

In order to evaluate the veracity of these propositions, we examine the termination of all civil war settlements in the immediate years after the end of the Cold War, from 1989 to 2005. Specifically, our analysis explores the likelihood that conflicts between rebel groups and the government end in a peace agreement. To accomplish this, we rely on the UCDP Conflict Termination Dataset (Kreutz, 2010). Including all conflicts where at least 25 battle-related deaths occurred, this dataset includes information on the start dates, end dates, as well as outcomes of intrastate conflicts. Data on mediation attempts for this analysis are drawn from the Civil War Mediation (CWM) dataset (DeRouen et al., 2011). The CWM is unique in that the dataset is ‘built upon the foundation’ (DeRouen et al., 2011: 664) of the UCDP/PRIO data projects. Using this information, we matched mediation attempts listed in the CWM with ongoing conflicts included in the UCDP Conflict Termination dataset. In line with contemporary studies on civil war termination, our primary unit of analysis is the civil war dyad year (Cunningham et al., 2009; Svensson, 2007). To operationalised our dependent variable (Peace Agreement), we generate a binary variable that is coded as a one in the year in which a peace agreement is reached and a zero otherwise. These data are drawn from the UCDP Conflict Termination dataset and is matched with mediation successes in the CWM. Within our temporal domain, there are 79 instances where rebels and the government reach a peace agreement.

For our primary independent variables, we rely on numerous metrics to account for mediator’s diplomatic and process knowledge. For diplomatic knowledge, we use three measures: the relative diplomatic ties between mediator states and the civil war state, diplomatic ties between mediator states and rebel groups, and whether mediator states share a joint membership in regional organisations with the civil war state.¹ To measure the diplomatic ties between the civil war state and the mediator, we rely on the Correlates of War Diplomatic Exchange dataset (Bayer, 2006).² Specifically, we code whether the mediator has an ambassador in the civil war state during the mediation attempt. For diplomatic ties between the rebel group and the mediator state, we rely on the Rebel Governance Dataset (RGD; Huang, 2016). Specifically, we rely on the detailed coding notes of the RGD to code whether the rebels have sent a diplomatic mission to the mediating state. As we expect the relative diplomatic knowledge of mediators to play a key role in fostering peace, not just individual ties, our variable Diplomatic Relations measures whether the mediating state has diplomatic ties with the government, the rebels, or both (generating a scale from 0 to 2). We then normalise this variable to generate a zero to one measure.^3,4

Our second measure is whether the mediating state is a joint member of a regional organisation with the civil war state (such as the African Union or ASEAN). This is a binary variable that is coded as a one in mediation attempts where the mediating state is a joint member of a regional organisation with the civil war state. Mediator states that are in the same region or neighbourhood are more familiar with what is happening in the region. Therefore, we expect that shared membership in a regional organisation between the conflict state and a mediator state will provide significant process knowledge for the mediator.

To measure process knowledge, we consider three factors that could familiarise the mediator with the specific preferences and procedures of conflict resolution used in particular conflict states, two of which focus more on the government, and one of which focuses on both the government and rebel group. First, we include a binary variable for whether the mediating state has a colonial legacy (as a former coloniser) of the civil war state. These data are drawn from the Issues Correlates of War Colonial History dataset. Colonial ties represent historic ties between a potential mediator and a conflict state, providing familiarity about conflict resolution norms in the conflict state. Reid (2017) argues that contextual knowledge, based on similar culture and colonial ties, can provide mediators with credibility leverage, helping to persuade combatants to agree in bargaining. We consider this as a form of process knowledge, specifically in the sense that the mediator state is very familiar with the post-colonial governance, as well as cultural or social norms about peace-making. We therefore expect that states with previous colonial relationships will be more knowledgeable and therefore more familiar with the domestic norms, rules, and decision-making procedures that affect conflict resolution, which could help the mediator be more successful in achieving settlement.

Our second measure for process knowledge is whether the mediating state has the same legal system to that of the civil war state. As similar processes for adjudicating disputes should inform the ability of mediators to identify processes that are familiar (and acceptable) to the disputants, we identify whether the mediating state and the civil war state both rely on common law, civil law, Islamic law, or mixed legal systems.⁵ These data are drawn from the Powell and Wiegand (2010) dataset. Domestic legal system refers to the type of legal system practised by a state in its constitution, laws, and judicial system, comprising civil, common, Islamic, and mixed legal systems (Mitchell and Powell, 2011; Powell and Wiegand, 2010). Similarity of a domestic legal system shared between the conflict state and mediator state should help mediators to better understand procedures of justice, reconciliation, and governance norms in the correct context. Therefore, we predict that mediator states with the same domestic legal system as the conflict state.

Finally, we expect that a mediator’s past experience mediating a dispute should inform their process knowledge with the disputants. Drawing on the CWM dataset, we code whether a mediator has previously attempted to mediate a conflict. Specifically, this variable is a count of the previous mediation attempts pursued by mediators. As with colonial legacies and regional organisations, our measure for shared legal systems is binary.⁶ Previous mediation attempts by an international mediator assist with identifying past areas of compromise that disputants may not readily admit and establish terms that are more amenable to the combatants. As part of the formulative approach to mediation, mediators often establish processes and structures to help with negotiations as well as to highlight common interests (Bercovitch and Wells, 1993). Past experience mediating conflicts empowers mediators to know which procedures and areas of compromise have failed previously and which procedures are likely to work in future negotiations.

It may be the case that our variables of interest are merely proxying for mediator bias. To account for this, we include control variables to capture biased relationships between the mediator and the combatants. Similar to Svensson’s (2007) analysis of mediator bias and success in mediation, we include a control variable for whether any of the mediators offered support to either the rebels or the government. These data are drawn from the UCDP External Support Dataset. We also include a control variable for whether mediators may have a formal alliance with the civil war state. Assuming that allied governments (specifically, a mediator and the civil war state) may wish to maintain the status quo, an alliance between the incumbent government and a mediator should suggest that the third party has a preference over the outcome of the civil war. We therefore use a binary variable to capture whether any of the mediating countries maintains a mutual defence alliance with the civil war state. These data are drawn from the COW Formal Alliances dataset.

As our theory relies heavily on the transfer of credible information during the mediation process, we also include a control for government transparency using the Hollyer et al. (2014) dataset on government transparency. Following prolonged fighting and heavy casualties, disputants often enter a hurting stalemate (Zartman, 2000). To capture these effects, we include controls for civil war intensity and conflict duration using the UCDP Conflict Termination Dataset (Kreutz, 2010). Settlements may also be less likely in territorial disputes as compared with conflicts over government control (Fearon, 2004; Walter, 2006). We therefore include a control variable for whether the civil war was fought over territorial incompatibilities (Kreutz, 2010). As stronger rebel groups are often more capable of pushing states to enter into a negotiated settlement rather than continue fighting (Clayton, 2013; Cunningham et al., 2009), we include a control variable for rebel group strength, drawn from the Cunningham et al. (2013) dataset. Finally, we include controls for national wealth as measured by the logged GDP per capita of civil war states (Gleditsch, 2002) as well as institution type as measured by the revised Polity II scale (Marshall et al., 2016).

We also include controls specific to the mediation process itself. In predicting mediation, we include all possible variables that should also predict the likelihood of a peace agreement. Given that identity conflicts are often difficult to resolve (Doyle and Sambanis, 2006), we include a control for ethnic-based recruiting by rebels as a barrier to mediation onset. Equally, past mediation attempts by third parties (regardless of the mediator) are often strong predictors of future mediation (Clayton, 2013; Greig and Regan, 2008). We therefore include a variable to capture the effects of any previous mediation attempts. Finally, to satisfy the exclusion restriction, we include a control for whether a rebel political wing is banned from participating in the political process. For all selection models, it is important to include a variable that distinguishes the selection stage from the outcome state. A variable satisfies this requirement if it predicts the selection stage (i.e. the onset of mediation) but does not predict the primary outcome that the model is trying to explain (i.e. the establishment of a peace agreement). As the acceptance of mediation often brings with it key reputation costs, governments are generally wary of offering legitimacy to dissident organisations (Greig and Regan, 2008; Melin and Svensson, 2009). This should be especially true for rebel political wings that are banned by the government from engaging the political process. On the other hand, there is no research to suggest that banned political wings play any role in facilitating the establishment of a peace agreement. Robustness checks suggest that this is the case in our data, as banned political wings increase the likelihood of mediation onset but play no discernable role in the establishment of a settlement.

To isolate the effects of mediator diplomatic knowledge and process knowledge during negotiations, we include controls for whether the mediator is an international organisation as well as the number of mediators involved in negotiations. As our variables are tied to governments as opposed to international organisations, it is important to account for mediation attempts that are led by international organisations.⁷ Equally, as more mediators participate in the mediation process, individual mediator characteristics may become less important.

To estimate each model, we rely on a seemingly unrelated (SUR) bivariate probit model. We rely on a selection model because mediation events are not randomly distributed within our sample population (with mediation efforts often directed to more difficult conflicts). As our primary dependent variable is whether rebels and the government reach a peace agreement, we employ the SUR bivariate probit model to predict first when mediation occurs and then examine how mediation efforts impact the ability of disputants to reach a settlement. While many analyses of mediation utilise a Heckman probit model, this form of estimation is not always appropriate. Specifically, Heckman probit models are best used when the selection stage is a necessary condition for the outcome state (Gartner, 2011; Greig and Regan, 2008). Peace agreements, on the other hand, are not dependent on the onset of mediation, as disputants may settle their conflict through bilateral negotiations (Ogutcu-Fu, 2016; Ryckman and Braithwaite, 2020). If a Heckman model is used, then all other peace agreements would be censored, possibly obscuring significant variation in the establishment of settlements. SUR bivariate probit models, on the other hand, address this issue by purging possible selection effects from the outcome stage while also accounting for observations not included in the selection stage (Greene, 2003; Hartzell and Hoddie, 2015). In other words, our models will be able to compare the effectiveness of mediator knowledge in establishing a peace agreement in all conflicts in the dataset, not just where mediation is ongoing.

Findings and analysis

The results from our analysis can be found in Table 1. Specifically, the table includes the results of both the initial selection stage that predicts the onset of mediation while the second stage predicts the termination of the civil war through a peace agreement. Model I is estimated with our key independent variables as well as our control variables. The second model also includes data on the relative bias of the mediators involved in the negotiations. All models are estimated with robust standard errors that are clustered on the civil war dyad. As noted in the research design, the unit of analysis is the civil war dyad year. The initial results demonstrate a positive and statistically significant relationship between mediator diplomatic ties and the establishment of a peace agreement between disputants. Mediators that have strong diplomatic ties with one or both parties increase the likelihood that they will assist with ending the conflict through a peace agreement. Our findings for diplomatic relations remain robust when we account for the presence of bias (though this measure loses some statistical significance). On the other hand, joint membership in regional organisations appears to have little or no effect on helping disputants reach a settlement.

Table 1.

Seemingly Unrelated Bivariate Probit Model.

SUR bivariate probit model	Mediation onset	Peace agreement	Mediation onset	Peace agreement
SUR bivariate probit model	Model I	Model I	Model II	Model II
Diplomatic relations		0.989**		1.007*
Diplomatic relations		(0.489)		(0.541)
Colonial ties		0.909**		0.905**
		(0.459)		(0.445)
Joint regional organisation membership		−0.294		−0.302
Joint regional organisation membership		(0.310)		(0.332)
Similar legal system		0.360		0.363
Similar legal system		(0.397)		(0.386)
Past mediation attempts by mediator		0.219***		0.229***
Past mediation attempts by mediator		(0.0664)		(0.0659)
Rebel biased				−0.198
Rebel biased				(0.353)
Govt. biased				−0.0627
Govt. biased				(0.273)
Alliances				0.221
Alliances				(0.511)
Govt. Transparency Index	−0.0257	−0.102**	−0.0254	−0.0982**
Govt. Transparency Index	(0.0565)	(0.0428)	(0.0567)	(0.0445)
Rebels at parity strength	0.0487	−0.351**	0.0497	−0.341*
Rebels at parity strength	(0.286)	(0.171)	(0.286)	(0.178)
International Organization mediator		0.457		0.438
International Organization mediator		(0.300)		(0.319)
Territory	0.255	0.0123	0.255	0.0106
Territory	(0.193)	(0.167)	(0.193)	(0.167)
Intensity	−0.0706	−1.128***	−0.0707	−1.132***
Intensity	(0.154)	(0.381)	(0.154)	(0.378)
War duration	0.0483***	0.0130	0.0483***	0.0128
War duration	(0.0160)	(0.0137)	(0.0160)	(0.0139)
Number of mediators		−0.104		−0.0813
Number of mediators		(0.0770)		(0.0927)
lnGDP	−0.256***	−0.197***	−0.256***	−0.196***
lnGDP	(0.0809)	(0.0519)	(0.0810)	(0.0522)
Polity II	0.0259	0.0452***	0.0259	0.0433***
Polity II	(0.0210)	(0.0152)	(0.0210)	(0.0154)
_Spline1	0.00132	0.00514*	0.00131	0.00531*
_Spline1	(0.00303)	(0.00282)	(0.00303)	(0.00290)
_Spline2	−0.000518	−0.00275	−0.000521	−0.00290
_Spline2	(0.00206)	(0.00197)	(0.00206)	(0.00203)
_Spline3	−3.08e-05	0.000647	−2.77e-05	0.000715
_Spline3	(0.000768)	(0.000757)	(0.000767)	(0.000783)
Banned political wing	−0.852***		−0.852***
Banned political wing	(0.300)		(0.300)
Previous mediation attempt (any)	0.413***		0.413***
Previous mediation attempt (any)	(0.0487)		(0.0488)
Ethnic conflict	−0.292		−0.292
Ethnic conflict	(0.252)		(0.253)
Constant	0.910	0.630	0.913	0.611
Constant	(0.811)	(0.492)	(0.812)	(0.494)
Observations	868		868
Log-likelihood	−436.06747		−435.90843

Robust standard errors in parentheses. Clustered on civil war dyads. SUR: seemingly unrelated.

***

p < 0.01, **p < 0.05, *p < 0.1.

The results appear to be more mixed when it comes to mediators’ process knowledge. Mediators that have colonial ties or a previous history of mediating the civil war are more likely to end the conflict through a peace agreement. Equally, having similar legal systems appears to not afford mediators any substantive advantage with regard to ending the conflict. These initial results suggest that mediator knowledge that is based on experiences interacting with rebels and the government is much greater than knowledge that is gleaned from similar legal processes or from cooperation in regional organisations.

In addition to being statistically significant, our variables measuring diplomatic and process knowledge also play a substantive role in shaping the likelihood of a peace agreement. Table 2 includes a summary of marginal effects. Given that only six mediators in our sample population have diplomatic ties with both the rebels and the government, we simply estimate the marginal effects of having diplomatic ties with one combatant. In estimating the marginal effects, we find that a mediator’s diplomatic knowledge (as measured by diplomatic ties) can increase the likelihood that the conflict will terminate in a peace agreement by 45%.⁸ This finding would suggest that as mediators have stronger formal diplomatic relations with disputants (both rebels and the government), their efficacy in settling disputes increases substantially. Our process knowledge variables also carry significant effects as well. For instance, conflicts mediated by a third party with a colonial tie to the civil war state increase the predicted probability of a peace agreement being reached by 62.3% (with all other variables held at their means). Stunningly, by increasing past mediation attempts from the minimum value to the maximum value, conflicts are 90.4% more likely to end in a peace agreement. Figures 1 -3 illustrate these effects.

Table 2.

Marginal Effects.^a

	Change in predicted probability	Confidence interval	% change
Diplomatic relations_(Min-Max)^b	0.011	(0.01–0.04)	45
Colonial ties (_Min-Max)	0.025	(0.01–0.07)	62.3
Past mediation attempts _(Min-Max)	0.086	(0.04–0.15)	90.4

Only statistically significant IVs are shown. Estimates are derived from Model 2 in Table 1.

Based on shifts from no diplomatic connections with combatants to at least one diplomatic ties with the warring parties.

Figure 1.

Likelihood of a peace agreement.

Figure 2.

Likelihood of a peace agreement.

Figure 3.

Likelihood of a peace agreement.

These results demonstrate that elements of diplomatic knowledge and process knowledge held by mediators seem to play a key role in ending ongoing civil wars. To illustrate the importance of our mediation variables, we estimate the likelihood that disputants will reach a peace agreement both with and without occurrence of international mediation. For instance, drawing from the results in Model I, in civil wars where there is high intensity, the rebels are seeking some form of self-determination, and the insurgency is relatively strong (as measured by being at parity strength), the likelihood of settling the dispute is fairly low.⁹ In conflicts where mediation has not occurred, the predicted probability that a peace agreement will be reached is 0.002. Once mediation is underway, the prospects for settling the dispute actually diminish (with the predicted probability falling 0.0006), suggesting that mediation is occurring in more intractable conflicts. Once we account for our knowledge variables, though, the prospects for settling the conflict increases significantly. Moving Diplomatic Ties, Colonial Ties, and Previous Mediation Attempt by Mediator from minimum to maximum value, the predicted probability of settling the dispute rises to 0.17 (or a 17% likelihood), marking a substantial increase in the likelihood of a settlement. The probability of success improves even more when the intensity of the conflict declines during mediation efforts (increasing to 0.21). These findings would suggest that the ability of mediators to leverage their knowledge is enhanced when disputants can be convinced to halt military operations during talks, though there are certainly some unmeasured factors at play.

The findings presented here offer some mixed results for our theoretical argument. For instance, while diplomatic knowledge in the form of relative, direct diplomatic relations with combatants increases the efficacy of mediation efforts, joint membership in regional organisations does not appear to afford mediators any additional benefit. Similarly, while colonial ties and past efforts mediating in armed conflict underscore strong support for our proposition that process knowledge enhances the efficacy of mediators, having a similar legal system does not improve the ability of mediators to push disputants to reach an agreement. What, therefore, explains these divergent findings? One explanation may be that the knowledge gleaned from direct interactions may be more beneficial than knowledge derived from similar experiences. For instance, while sharing joint participation in regional organisations may allow for the transfer of credible information to diplomats, such diplomatic interactions may be no substitute for direct engagement with government officials or rebel leaders. In the latter, diplomats are allowed to engage with disputants on a wide range of issue areas and generate greater background knowledge on the specific leaders.

Similarly, mediators that have colonial ties or previous experience managing the conflict may simply have a greater understanding of the cultural context in which they are engaging. For instance, in the example of French diplomatic engagement in the Ivorian Civil War, the French had not only a shared legal system as Cote d’Ivoire, but they also had years of strong political and economic bonds with the country since Ivorian Independence. Understanding the importance of a legalistic remedy to the conflict may be less a product of knowing the Ivorian legal system as compared with having a greater contextual understanding of the roots of the conflict (Reid, 2017) and the processes that would be acceptable to both sides.

Our control variables performed largely as expected. The results suggest that conflicts where the rebels engage in ethnic recruitment are less likely to experience mediation attempts. This finding may underscore that members of the incumbent government are less inclined to negotiate during ethnic identity wars. Equally, the results also suggest that prolonged conflicts are much more likely to experience the onset of mediation since the measure for civil war duration is significant. This finding is interesting given that our control for conflict intensity appears to play no observable role in increasing the likelihood that mediation will occur. Taken together, it may be that mediation is more likely when parties have failed to accomplish their goals on the battlefield and are no longer launching aggressive campaigns to wrest territory from their opponents. Finally, mediation is much more likely when there has been a previous attempt by a third party to reopen negotiations.

When it comes to actually settling the dispute through a peace agreement, our controls appear to play a slightly different role. As compared to the onset of mediation, conflict intensity plays a negative and significant role in producing a settlement to ongoing conflicts. Civil war duration appears to be positively related to the establishment of a settlement, but this variable fails to reach statistical significance. As for mediators, the participation of an international organisation in fostering a settlement is statistically significant and positive. This result suggests that international organisations may play a key role in ending ongoing civil wars through mediation efforts. Interestingly, our control for government transparency is negative and statistically significant, suggesting that government transparency may negatively affect the ability of elites to make costly concessions without facing domestic backlash. As noted by Beardsley (2010), mediation is attractive specifically because it affords leaders some political cover for unpopular decisions. Finally, our control for the number of mediators plays no observable role in ending conflicts. This finding may suggest that the number of mediators participating in talks is less important as compared with the specific characteristics of the participating states.

Conclusions

Combatants in civil wars fight because they cannot peacefully resolve their disputes or grievances; otherwise fighting would not occur. Third-party mediation can help to bring combatants to the table, but there are clear incentives by both sides to misrepresent their demands and willingness to concede. We have posited in this article that when mediators have significant knowledge of one or both combatants, they are much more likely to help generate a peace agreement because they are so familiar with the culture, procedures, and conflict grievances of the combatant(s). Distinct from bias, we believe that process knowledge and diplomatic knowledge provide mediators with an independent means of assessing the credibility of statements made by the combatants. This assessment in turn helps mediators to make more accurate suggestions for persuading the combatants to consider concessions and negotiate a peace agreement between the disputants.

Analysing several factors that measure process knowledge and diplomatic knowledge, we conclude that some of these factors do in fact influence the likelihood of mediators helping to end armed civil conflicts. Specifically we find that diplomatic ties, colonial ties, and past mediation experience all substantively improve the probability that mediators can help combatants agree to a peace agreement. These findings help us better understand the role that knowledge, separate from bias, plays in mediation effectiveness. The implication is that mediators from states with diplomatic representations in the conflict state, former coloniser states, and states that have previously mediated in the conflict are all more likely to reach success in ending civil wars. Given the number of ongoing civil wars and the important role that mediators often play in bringing the disputants closer to a peace agreement, understanding the role of mediator knowledge is another step towards better achieving peace.

Supplemental Material

Online_Appendix_BJPIR_v3 – Supplemental material for Third-party knowledge and success in civil war mediation

Supplemental material, Online_Appendix_BJPIR_v3 for Third-party knowledge and success in civil war mediation by Krista Wiegand, Erin Rowland and Eric Keels in The British Journal of Politics and International Relations

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Eric Keels .

Supplemental material

Additional supplementary information may be found with the online version of this article.

Notes

References

Anderson

Olson

(2003) Confronting War: Critical Lessons for Peace Practitioners (Reflecting on peace practice project). Cambridge, MA: The Collaborative for Development Action.

Bapat

(2005) Insurgency and the opening of peace processes. Journal of Peace Research 42(6): 699–717.

Bayer

(2006) Diplomatic exchange data set, v2006.1. http://correlatesofwar.org (accessed 27 May 2010).

Beardsley

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

Beber

(2012) International mediation, selection effects, and the question of bias. Conflict Management and Peace Science 29(4): 397–424.

Bercovitch

Anagnoson

Wille

(1991) Some conceptual issues and empirical trends in the study of successful mediation in international relations. Journal of Peace Research 28(1): 7–17.

Bercovitch

Houston

(1995) The study of international mediation: Theoretical issues and empirical evidence. In: Bercovitch

(ed.) Resolving International Conflicts: The Theory and Practice of Mediation. Boulder, CO: Lynne Rienner, pp.11–35.

Bercovitch

Wells

(1993) Evaluating mediation strategies: A theoretical and empirical analysis. Peace & Change 18(1): 3–25.

Burton

Dukes

(1990) Conflict: Practices in Management Settlement & Resolution. New York: St. Martin’s Press.

10.

Carnevale

(2002) Mediating from strength. In: Bercovitch

(ed.) Studies in International Mediation. New York: Palgrave Macmillan, pp.25–40.

11.

Carnevale

Arad

(1996) Bias and impartiality in international mediation. In: Bercovitch

(ed.) Resolving International Conflicts: The Theory and Practice of Mediation. Boulder, CO: Lynne Rienner, pp.39–53.

12.

Clayton

(2013) Relative rebel strength and the onset and outcome of civil war mediation. Journal of Peace Research 50(5): 609–622.

13.

Cunningham

Gleditsch

Salehyan

(2009) It takes two: A dyadic analysis of civil war duration and outcome. Journal of Conflict Resolution 53(4): 570–597.

14.

Cunningham

Gleditsch

Salehyan

(2013) Non-state actors in civil war: A new dataset. Conflict Management and Peace Science 30(5): 516–531.

15.

DeRouen

Jr. Bercovitsch

Pospieszna

(2011) Introducing the Civil Wars Mediation (CWM) dataset. Journal of Peace Research 48(5): 663–672.

16.

Doyle

Sambanis

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

17.

Elgström

Bercovitch

Skau

(2003) Regional organizations and international mediation: The effectiveness of insider mediators. African Journal on Conflict Resolution 3(1): 11–27.

18.

Favretto

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

19.

Fearon

(1995) Rationalist explanations for war. International Organization 49(3): 379–414.

20.

Fearon

(2004) Why do some civil wars last so much longer than others? Journal of Peace Research 41(3): 275–301.

21.

Gartner

(2011) Signs of trouble: Regional organization mediation and civil war agreement durability. Journal of Politics 73(2): 380–390.

22.

Gleditsch

(2002) Expanded trade and GDP data. Journal of Conflict Resolution 46(5): 712–724.

23.

Greene

(2003) Econometric Analysis. New York: New York University.

24.

Greig

(2015) Rebels at the gate: Civil war battle locations, movement, and openings for diplomacy. International Studies Quarterly 59(4): 680–693.

25.

Greig

Regan

(2008) Willingness to say yes? An analysis of the willingness to offer and accept mediation in civil wars. International Studies Quarterly 52: 759–781.

26.

Greig

Mason

Hamner

(2018) Win, lose, or draw in the fog of war. Conflict Management and Peace Science 35: 523–543.

27.

Hartzell

Hoddie

(2003) Institutionalizing peace: Power sharing and post civil war conflict management. American Journal of Political Science 47(2): 318–332.

28.

Hartzell

Hoddie

(2015) The art of the possible: Power sharing and post-civil war democracy. World Politics 67(1): 37–71.

29.

Hollyer

Rosendorff

Vreeland

(2014) Measuring transparency. Political Analysis 22(4): 413–434.

30.

Huang

(2016) The Wartime Origins of Democratization: Civil War, Rebel Governance, and Political Regimes (Problems of International Politics Series). Cambridge: Cambridge University Press.

31.

Jackson

(1952) Meeting of Minds: A Way to Peace through Mediation. New York: McGraw-Hill.

32.

Keels

(2019) The importance of being earnest (in mediation): Rebel group structure, leadership turnover, and success in civil war mediation. International Negotiation 24: 266–295.

33.

Kreutz

(2010) How and when armed conflicts end: Introducing the UCDP conflict termination dataset. Journal of Peace Research 47(2): 243–250.

34.

Kydd

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

35.

Lake

Rothchild

(1996) Containing fear: The origins and management of ethnic conflict. International Security 21(2): 41–75.

36.

Lake

Rothchild

(1998) The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation. Princeton, NJ: Princeton University Press.

37.

Marshall

Gurr

Jaggers

(2016) Polity IV project: Political regime characteristics and transitions, 1800-2015 (ed Center for Systemic Peace). Available at: http://www.systemicpeace.org/inscrdata.html (accessed 27 May 2010).

38.

Mattes

Savun

(2010) Information, agreement design, and the durability of civil war settlements. American Journal of Peace Research 54(2): 511–524.

39.

Melin

(2011) The impact of state relationships on if, when, and how conflict management occurs. International Studies Quarterly 55(3): 691–715.

40.

Melin

Svensson

(2009) Incentives for talking: Accepting mediation in international and civil wars. International Interactions 35(3): 249–271.

41.

Mitchell

Powell

(2011) Domestic Law Goes Global: Legal Traditions and International Courts. Cambridge: Cambridge University Press.

42.

Nathan

DeRouen

Jr. Lounsbery

(2018) Civil war conflict resolution from the perspectives of the practitioner and the academic. Peace and Change 43(3): 344–370.

43.

Northredge

Donelan

(1971) International Disputes: The Political Aspects. London: Europa.

44.

Novak

(2009) Face-saving maneuvers and strong third-party mediation: The Lancaster house conference on Zimbabwe-Rhodesia. International Negotiation 14: 149–174.

45.

Ogutcu-Fu

(2016) Outside the battlefield: In-group political dynamics of civil conflict negotiations and settlements. Political Research Quarterly 69(3): 403–417.

46.

Powell

Wiegand

(2010) Legal systems and peaceful attempts to resolve territorial disputes. Conflict Management and Peace Science 27(2): 129–151.

47.

Rauchhaus

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

48.

Regan

Aydin

(2006) Diplomacy and other forms of intervention in civil wars. Journal of Conflict Resolution 50(5): 736–756.

49.

Regan

Stam

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

50.

Regan

Frank

Aydin

(2009) Diplomatic interventions and civil war: A new dataset. Journal of Peace Research 46(1): 135–146.

51.

Reid

(2017) Finding a peace that lasts: Mediator leverage and the durable resolution of civil wars. Journal of Conflict Resolution 61(7): 1401–1431.

52.

Roepstorff

Bernhard

(2013) Insider mediation in peace processes: An untapped resource? Security and Peace 31(3): 163–169.

53.

Ryckman

Braithwaite

(2020) Changing horses midstream: Leadership changes and the civil war peace process. Conflict Management and Peace Science 37: 83–105.

54.

Savun

(2008) Information, bias, and mediation success. International Studies Quarterly 52: 25–47.

55.

Stedman

(1991) Peacemaking in Civil War: International Mediation in Zimbabwe, 1974-1980. Boulder, CO: Lynne Rienner.

56.

Svensson

(2007) Bargaining, bias and peace brokers: How rebels commit to peace. Journal of Peace Research 44: 177–194.

57.

Svensson

(2009) Who brings which peace? Neutral versus biased mediation and institutional peace arrangements in civil wars. Journal of Conflict Resolution 53(3): 446–469.

58.

Svensson

Lindgren

(2013) Peace from the inside: Exploring the role of the insider-partial mediator. International Interactions 39(5): 698–722.

59.

Thomas

Wood

Wolford

(2016) The rebel’s credibility dilemma. International Organization 70(3): 477–511.

60.

Touval

(1975) Biased intermediaries: Theoretical and historical considerations. Jerusalem Journal of International Relations 1(1): 51–69.

61.

United Nations (2012) General assembly. Letter dated 9 May 2012 from the permanent representative of Turkey to the United Nations addressed to the secretary-general (A/66/800). Available at: https://peacemaker.un.org/sites/peacemaker.un.org/files/20120225_Summary%20Report_%20Istanbul%20Conference%20on%20Mediation.pdf (accessed 27 May 2010).

62.

Walter

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

63.

Walter

(2006) Building reputation: Why governments fight some separatists but not others. American Journal of Political Science 50(2): 313–330.

64.

Walter

(2009) Bargaining failures and civil war. Annual Review of Political Science 12: 243–261.

65.

Wehr

(1979) Conflict Regulation. Boulder, CO: Westview Press.

66.

Wehr

Leaderach

(1991) Mediating conflict in central America. Journal of Peace Research 28(1): 85–98.

67.

Young

(1967) The Intermediaries: Third Parties in International Crises. Princeton, NJ: Princeton University Press.

68.

Zartman

(2000) Ripeness: The hurting stalemate and beyond. In: Stern

Druckman

(eds) International Conflict Resolution After the Cold War. Washington, DC: National Academy Press, pp.225–250.

69.

Zartman

Touval

(1985) International mediation: Conflict resolution and power politics. Journal of Social Issues 41: 27–45.

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

0.22 MB