Abstract
Within the study of intrastate armed conflicts, many scholars rely on a bargaining model when explaining why some civil wars are intractable. Primarily, scholars posit that commitment problems represent a key barrier to settling conflicts through negotiations. Yet, some civil wars are more easily resolved than others. If commitment problems are a universal feature of civil wars, what explains why commitment problems are more salient in some conflicts as compared to others? We argue that ideological differences between combatants enhance commitment problems in civil wars. Assuming that ideology is used by combatants to generate support, concessions that violate the ideological goals of each side may alienate supporters. With extreme ideological polarization, concessions may be viewed as relative as opposed to absolute, making it difficult for either side to credibly commit to offer concessions. To test these claims, we quantitatively examine the duration and outcome of all intrastate conflicts from 1975 to 2011.
The war in Afghanistan has now become the longest running military campaign in the history of the United States. In addition to the thousands killed in the conflict, sixteen years of fighting has led to a de facto stalemate in the conflict, where neither side perceives any clear path to outright victory. Still, a settlement has been difficult to achieve in the conflict between the Afghan Government, the US military, and the various Taliban factions fighting for control of this war-torn state. Similarly, the Colombian Government’s war with the National Liberation Army rebels has entered its fifty-fourth year. While the Government of Colombia was able to end the protracted conflict with the Revolutionary Armed Forces of Colombia—People’s Army rebels, there is no guarantee that such a similar success will be achieved in their dispute with this other Marxist rebel group.
The last twenty years have produced many studies that attempt to explain why some conflicts remain intractable. By far, the dominant approach to this topic remains the bargaining approach to civil war dynamics (Mason and Fett 1996; Walter 2002, 2009a; Fearon 2004). In particular, scholars often posit that civil wars are difficult to resolve through a negotiated settlement, given commitment problems between combatants. Given the mutual incentive to defect from any agreement reached, neither side may credibly commit to a settlement (Fearon 1995, 2004; Walter 2002). Absent institutional mechanisms or international assistance, commitment problems present a common barrier to resolving conflicts. What remains puzzling, though, is that some conflicts may be resolved relatively quickly while others persist. For instance, the Chadian Government negotiated a peace agreement with the Armed Forces of the Federal Republic rebels after only seven months of fighting. This occurred without significant international assistance. Equally, what can explain why the Museveni Government in Uganda was able to negotiate a peace agreement with the Ugandan Peoples Democratic Army (UPDA) after only a year of fighting but has been unable to resolve the thirty-year conflict with the Lord’s Resistance Army (LRA)? As noted by Kaufman (2015), if commitment problems are a universal aspect of civil wars, why do they play a larger role in some disputes and not in others?
One explanation for the varying effects of commitment problems may be the specific issues under contention (Walter 2006; Goddard 2006; Svensson 2007). Certain contentious issues, such as the right to self-determination or the promotion of religious practices, often complicate the bargaining environment as the stakes under dispute are perceived to be indivisible (Goddard 2006). While the role of intangible, and increasingly indivisible, issues has been discussed widely in the literature on interstate conflict (Huth 1996; Mitchell and Hensel 2005; Hensel et al. 2008; Wiegand 2011), there has been relatively little work on the role of indivisible issues within the quantitative literature on intrastate conflict (Goddard 2006; Svensson 2007). Rather, much of the bargaining literature tends to downplay the role of indivisible stakes in complicating negotiations (Fearon 1995; Powell 2006). It is difficult to believe, though, that bargaining is only affected by these indivisible claims in interstate disputes but not in intrastate conflicts.
We suggest that the ideological positions staked out by governments and rebel groups often inadvertently complicate negotiations. When rebel groups adopt ideological positions that contrast with the ideological goals of the central government, it generates more indivisible disputes. We consider ideology to be a “…more or less systematic set of ideas that includes the identification of a referent group (a class, ethnic, or other social group), an enunciation of the grievances or challenges that the group confronts, the identification of objectives on behalf of that group (political change—or defense against its threat), and a (perhaps vaguely defined) program of action” (Sanin and Wood 2014, 215). In other words, rebel groups and governments use ideological positions to mobilize support during the course of armed conflicts. When rebels and the government maintain these contrasting ideological positions, they are often unable to make significant concessions without violating the ideological goals of their primary supporters, making disputes more indivisible. Assuming that parties are aware of this, contrasting ideological positions undermine the ability of either party to credibly commit to settlements as combatants are aware that their rival’s constituents will oppose any agreement that violates the ideological positions staked out by leaders. It is therefore important to approach ideology in civil war dyadically as opposed to monadically.
Furthermore, as opposed to assuming that indivisibility is binary, we argue that some ideological divisions may be more indivisible than others. We argue that ideological divisions centered on religious or ethnic issues are more indivisible than economic ideological disagreements, making civil wars more difficult to resolve. To test this theory, we examine both the duration and outcome of all intrastate conflicts from 1975 to 2011. The results provide robust support for our proposition that contrasting ideologies lead to longer civil wars, and these divergent ideological positions often undermine the ability of parties to negotiate a settlement as opposed to holding out for an outright victory. Furthermore, our findings also suggest that ideological differences play a greater role when they center on more intangible disputes involving the role of religion within the state or ethnic supremacy.
Previous Research
Much of the literature on civil war dynamics has been shaped, in large part, by the bargaining approach to conflict (Fearon 1995; Walter 2002; Powell 2006; Mattes and Savun 2010; Thyne 2012; Uzonyi and Wells 2016). As a theoretical starting point, given the high costs of war, conflict is often irrational (Fearon 1995). Despite these high costs, conflict may emerge (and endure) when parties are uncertain over the reservation points or relative resolve of their opponents (Fearon 1995; Walter 2009b; Mattes and Savun 2010). For instance, rebel groups may attempt to exploit these asymmetries of information, such as exaggerating their demands, as a way to gain leverage when negotiating with the state (Walter 2009a), or rebel groups may increase their demands during talks as a way to demonstrate resolve (Thomas, Wood, and Wolford 2016). As noted by Walter (1999, 2009b), rebel groups often have an incentive to limit the amount of information that is available to the government. If rebels present accurate information on their troop size or the location of their forces, then it should be relatively easy for the government to find and sanction them. Equally, rebels may use misinformation as a way to garner greater support from the population by suggesting that they have the capability to win the war (Kydd and Walter 2006). As a way to address this problem, third parties (intervening diplomatically) may provide a reliable transfer of information so as to ameliorate asymmetries that complicate bargaining (Regan and Aydin 2006; Regan, Frank, and Aydin 2009).
It should be noted that there is some disagreement in the literature on intrastate conflict as to the importance of information asymmetries in prolonging the duration of civil wars. For instance, Fearon (2004) offers some skepticism over the notion that civil wars endure given asymmetries of information (see also Mattes and Savun 2010). As argued by Fearon (2004), information on each side’s relative resolve and capability should become apparent after the first few years of fighting. This notion has often been reflected in many past studies on civil war duration and termination. The relative strength of combatants as well as their battlefield success should reveal information relatively quickly, shortening the duration of conflicts. For instance, Mason and Fett (1996) have posited that civil war termination generally boils down to a relative utility model for continued fighting. Specifically, belligerents will continue to fight as long as the prospective benefits of victory outweigh the costs of continued fighting. Once fighting becomes too costly, disputants are much more likely to enter into a negotiated settlement as opposed to holding out for victory (Mason, Weingarten, and Fett 1999; Brandt et al. 2008). Equally, governments are much more likely to agree to a settlement when faced with stronger rebel groups. Governments are generally apprehensive about negotiating with rebels and will try to defeat them early rather than suffer the reputation costs of negotiating (Bapat 2005; Melin and Svensson 2009). This becomes difficult, though, when governments face a more capable opponent. As rebel group strength increases, governments are more likely to settle the dispute through bilateral agreements (Cunningham, Gleditsch, and Salehyan 2009) or through international mediation (Clayton 2013). Equally, as the pace of rebel success increases (pushing toward urban centers), governments are more likely to seek a negotiated settlement (Greig, Mason, and Hamner 2018).
Outside of information asymmetries, the bargaining process may be complicated by issues of commitment problems. Wars often endure when one or both sides feel that their opponent cannot credibly abide by the terms of a settlement (Fearon 1995). As noted by Walter (1997, 2002), commitment problems are particularly troublesome for disputants in civil wars. As opposed to interstate disputes where belligerents may fall back behind their own state lines after a settlement is reached, civil wars force combatants to cooperate within society after they disarm and demobilize. Commitment problems are particularly salient for insurgents, as rebels may agree to the process of disarming and demobilizing simply on the promise that the government will grant them amnesty or incorporate them into the armed forces. As noted by Walter (2002), strong third-party guarantees are often necessary to overcome commitment problems in order to end the conflict.
There are some limitations, though, to the bargaining approach in the study of civil war. As noted by Goddard (2006), bargaining theories tend to undervalue the specific claims made by both sides as disputants often have an incentive to misrepresent information (particularly the concessions they want) as a way to gain leverage during negotiations (Farrell and Gibbons 1989; Rabin 1990; Fearon 1995). Unfortunately, the justifications employed by disputants to mobilize support from their core constituents as well as to strengthen their positions at the bargaining table may generate rhetorical traps where it is difficult for either side to accept concessions that fall short of their stated goal. Specifically, there are several legitimation strategies that often transform tangible disagreements into indivisible disputes (Goddard 2006). Although Fearon (1995) downplays the role of indivisibility as a barrier to settlements, multiple qualitative studies have noted that salient issues often complicate the bargaining process, preventing either side from making concessions (Toft 2006; Goddard 2006). Even some proponents of the bargaining approach also recognized that indivisible stakes may complicate negotiations, prolonging civil wars (Mason and Fett 1996; Mason, Weingarten, and Fett 1999; Brandt et al. 2008). Equally, within the study of interstate conflicts, a wealth of literature has demonstrated that highly salient disputes are often difficult to resolve because territorial claims often hold symbolic, nationalist value that is perceived as intangible (Hensel et al. 2008; Huth 1996; Mitchell and Hensel 2005; Wiegand 2011). Compared to disputes over territory with tangible value such as economic resources, disputes over territory with intangible value are more likely to endure and not be resolved (Huth 1996; Wiegand 2011).
Similar findings have been made at the domestic level with regard to the issue of religion in armed conflict (Svensson 2007; Toft 2007; Svensson and Nilsson 2018). As noted by Svensson (2007), religious intrastate conflicts are often treated as indivisible disputes, where it is exceedingly difficult for either side to offer concessions without violating the integrity of the divine goals. Conflicts where one side advocates for a religious issue (such as in the Sri Lankan civil war or the Sudanese civil wars) generate an environment where nonreligious accommodations are often seen as insufficient to settle the dispute (Svensson 2007).
As noted earlier, though, there is significant skepticism within the bargaining literature that issue indivisibility may hinder negotiations. As noted by Powell (2006), though bargaining indivisibilities complicate negotiations given the limited number of potential agreements, these contentious issues do not explain the outbreak of fighting, given that agreements are still preferable to violence. In each instance, costly fighting in conjunction with the risk of losing substantially reduces the value of the issue under dispute (even if the issue is perceived to be indivisible). Powell (2006, 179) notes, “…the problem with bargaining indivisibilities is not the absence of agreements that states prefer to fighting. The problem is that the states may not be able to commit to following through on them.” The logic put forward by Powell (2006), though sound, tends to draw an unnecessary distinction between bargaining indivisibilities and commitment problems. The relative salience of the issues under dispute may play a key role in whether combatants believe that their opponents will abide by any terms offered in a settlement. Rather, the particular goals set forth by an insurgent group to mobilize domestic and international support may have unintended consequences for the bargaining process with the government. While issues of religion may certainly play a role in complicating negotiations, rebel ideological claims surrounding ethnic exclusion (Wucherpfennig et al. 2012; Cederman, Gleditsch, and Buhaug 2013) as well as economic goals (Mason 1998) can also play a critical role in complicating the bargaining process.
One way to investigate this question is by examining the ideological distance between combatants during intrastate armed conflicts. While there is a clear distinction between issues and ideologies, ideological positions often subsume many similar issues. There may be many separate issues under dispute in any given conflict. For example, the Sudan Liberation Movement/Army (SLM/A) rebel organization made roughly 140 demands on the Sudanese Government during their insurgency in Darfur (Thomas, Wood, and Wolford 2016). While the rebel group identified numerous issues, it is likely that many of these demands boiled down to the ideological goals underscoring self-determination, challenging the racial hierarchy in Sudan, and contesting the state’s strict imposition of Sharia law. Although issues, for instance, related to ethnic identity may vary in how difficult they are to resolve at the bargaining table, they still may be tied to the larger ideological position of a rebel group. Differences in ideological position between combatants may therefore be a natural place to begin this investigation.
Theory
Building on the findings of Goddard (2006), Svensson (2007), and Wucherpfennig et al. (2012), we suggest that the ideological mobilization strategies employed by rebel groups and the government to garner support can have deleterious effects on the bargaining environment and eventually on the ability of combatants to credibly commit to peace processes. Drawing on the instrumental approach to ideology in conflict (Sanin and Wood 2014), we assume that ideologies are selected by combatants (both the government and the rebels) as a way to mobilize support for their policy objectives (either maintaining the status quo or calling for revolutionary change). The ideological goals espoused by combatants can complicate the bargaining process when they are required to backtrack on their stated goals as part of a peace process. Commitment problems are enhanced, though, largely when the ideological position of a rebel group is diametrically opposed to that of the government. We define this as a contrasting ideological position. By adopting such a position, bargaining becomes difficult as concessions meant to appease the demands of either side will alienate the combatants’ core constituents, making critical issues more indivisible. Contrasting ideological positions should, in turn, enhance commitment problems, as neither side may believe that their opponents can credibly commit to backtracking on their ideological positions. This problem should be further complicated as ideologies become more indivisible with less tangible concessions that can be offered to ameliorate either side. On the other hand, when the ideology of rebels does not contrast with the central government, we believe that ideology may not play a salient role in the conflict. Equally, when both parties share similar ideologies, bargaining should be simplified, leading to shorter conflicts.
Ideology in Civil War
While dormant for much of the resurgence in civil war research (Sanin and Wood 2014), ideology has become a more common explanation for political phenomena in intrastate conflict (Svensson 2007; Polo and Gleditsch 2016; Svensson and Nilsson 2018; Hirose, Imai, and Lyall 2017). Ideologies, defined here using Sanin and Wood’s (2014) definition noted earlier, are often used to identify a target group, the contentious issues under dispute, and possible policy solutions to address the contentious issues. For left-wing insurgencies in Latin America, East Africa, and South Asia, economic ideologies often underscored the plight of peasants, the need for a radical redistribution of wealth, and the importance of land reform to these insurgencies (Young 1997; Mason 1998). For religious conflicts in Sudan, Iraq, or Sri Lanka, the need to establish (or remove) official religious practices was a key characteristic of rebel group recruitment, where insurgents targeted a disaffected group and offered religious reform as a justification for their violence as well as a key policy solution (Svensson 2007; Natsios 2012; Kaufman 2015).
A key element of ideology is to allow rebels and the incumbent government to solicit domestic and international support. While the instrumental use of ideology may also be an effective tool at preserving group cohesion and establishing codes of behavior (Sanin and Wood 2014, 218), identifying out-groups to target militarily (Hirose, Imai, and Lyall 2017), or justify terrorism as a weapon of war (Polo and Gledtisch 2016), we focus primarily on ideology as a mechanism for mobilization. Rebels often operate in an environment where they must not only appeal to potential recruits but also find a broader audience to provide active and passive support. As noted by Sanin and Wood (2014, 218), “Even if there exists a nominally cohesive motive for rebellion such as ethnic identity, the group must coordinate the different, sometimes contradictory, interests and demands of its constituencies, donors, and other audiences for its actions.” We posit that this same phenomenon should also be applied to governments, as elites in the capital city must balance the interests of multiple factions in order to stay in power (Bueno de Mesquita et al. 2003). Governing officials adopt certain ideological positions that they assume will garner the greatest degree of material support, particularly during armed conflict as resources must be redirected toward counterinsurgency efforts. While governing ideological positions may not be widely popular, they may speak to other members within the selectorate who favor policies that ensure economic, ethnic, or religious advantage (Bueno de Mesquita et al. 2003). Landed interests in Peru and El Salvador, for instance, often cooperated with members in the armed forces and conservative politicians to maintain land tenure systems (Mason 1998). President Nimeiry of Sudan shifted to a hard-line Islamic ideology throughout the late 1970s as a way to consolidate support from dominant tribes in Khartoum, abandoning a secular, broad-based program to institute peace in the war-torn country (Natsios 2012). In short, ideological positions are adopted by both rebels and the government as a way to solicit support from different segments in society (as well as the international community).
The winning coalitions that are formed on ideological grounds are based, in part, on the expectation that rebel groups and the government will adhere to the ideological principles that brought them there in the first place. This should be especially important for rebels as, given their status outside of government, they must essentially campaign on policy promises (Toft 2010). Governments should also demonstrate their commitment to certain ideological principles by instituting and maintaining policies that are consistent with their espoused ideological stance. This may include economic reforms, jingoistic policies, or the strict enforcement of sectarian instruction. We assume that if combatants back away from these ideological positions, they risk losing support from their key constituents. This may result, at a minimum, in a substantial loss of economic support or recruits for their war effort. At the more extreme end, backtracking on ideological positions may lead to armed insurrection within their own ranks (such as challenges to organizational leadership) or a splintering of factions such as rival rebel organizations. For instance, during the Burundian civil war, following the unexpected death of Melchior Ndadaye, a power-sharing coalition was formed as a way to address rebel grievances of ethno-political exclusion of Hutus from the government. Sylvestre Ntibantunganya (a Hutu) was installed as the interim leader of the country. This agreement, though, was quickly scrapped following a violent military coup perpetrated by the Tutsi military apparatus in order to maintain their ethnic dominance in government (Nindorera 2012). While issues of ethnic exclusion are not always ideological, the rhetoric used to justify policies of discrimination is certainly part of government ideology. Equally, rebels that claim to represent those excluded from political power may draw on ethnic identity as part of their mobilizing ideology. As noted by Sanin and Wood (2014, 215), “Despite the distinction in earlier literature between ‘ideological’ and ‘ethnic’ groups or conflicts, ethno-nationalist groups in civil conflict also embrace an ideology, namely, that of nationalism….” Combatants who mobilize support along ideological grounds therefore risk losing significant support if they backtrack on their ideological positions.
Combatant Mobilization: From Ideological to Indivisible
While the ideological goals of either party may complicate the bargaining space, we contend that contrasting ideological goals between rebels and the government should be especially pernicious to any peace process by making ideological goals indivisible sticking points. We define contrasting ideological goals as ideological positions that are diametrically opposed to the ideological goals of one’s opponents. As noted by Goddard (2006, 38), “an issue is indivisible when any division of the issue acceptable to one of the antagonists is unacceptable to the other.” While all civil wars are defined (at least ostensibly) by differing policy positions and goals, not all conflicts involve contrasting ideological positions. While groups may espouse grievances against the state over public spending, land rights, or demands for leadership removal, such demands may not be diametrically opposed to the ideological goals of the state. Similarly, groups may espouse a number of promises that are not necessarily related to government policies (such as promises to promote democracy or establish a theocratic state). A key distinguishing factor in the Darfur conflict between the SLM/A and the Justice and Equality Movement (JEM) was the fact that the SLM/A was largely mobilized around the ideological goal of the separation of religion and politics. The JEM, while also seeking to gain autonomy for Darfur, elected to avoid taking such a position.
Contrasting ideological goals (or positions) may range across a number of salient issues within society. Rebels may seek to differentiate themselves from the state by adopting contrasting economic ideologies (such as a socialist platform against a right-wing government) or a radical platform on gender inclusion when contrasting themselves against a government’s patriarchal state. When rebels and the government maintain contrasting ideological positions, any concessions offered to address the ideological goals of one side will be perceived as underscoring a loss of the ideological goals of their opponent. In other words, the gains made during bargaining are perceived as being relative as opposed to absolute when both parties maintain contrasting ideologies. In the parlance of Goddard’s (2006, 38) work, the adoption of a contrasting ideological position is a deleterious legitimation strategy.
Concessions designed to address the ideological goals of an insurgency are less salient when they do not violate the interests of the government’s core supporters. This should be true even in conflicts that are defined, in part, by salient issues such as religion (Svensson 2007). Having a religious goal, in and of itself, does not necessarily complicate the bargaining process in civil wars. Rather, such an ideological goal becomes a barrier when it contrasts with the position of the incumbent government. For instance, while the Free Aceh Movement (or GAM) mobilized support through its religious ideology (Aspinall 2009), religion was far from the defining characteristic of the conflict or the barrier to any settlement. The role of religion differed substantially, though, in conflicts such as the Sudanese or Sri Lankan civil wars (Svensson 2007; Kaufman 2015). Particularly with regard to the various rebellions in Sudan, both the government and the rebel group sought to mobilize support along religious (or secular) lines (Natsios 2012).
Once disputes become defined by seemingly indivisible differences (such as under contrasting ideological positions), we contend that commitment problems should be heightened, given that neither side believes their opponents can credibly backtrack on ideological goals without losing the support of their respective winning coalition. Commitment problems exist when parties are “…unable commit themselves to following through on an agreement and may also have incentives to renege on it” (Powell 2006, 170). As the conflict centers on ideological disagreements (i.e., contrasting ideologies), any settlement will require one or both parties to concede on a critical ideological position. Assuming that both parties are aware that ideological positions are used to solicit support from key constituents, belligerents may be wary to enter into negotiations knowing that it will be difficult to concede certain ideological platforms. They therefore will find it difficult to credibly commit to a settlement. Equally, they should also be aware that their opponents will find it difficult to walk away from stated ideological goals without losing their own supporters, thereby undermining their belief that their rival can also credibly commit to the settlement. While strong ideological positions held by any party may complicate bargaining, these positions should be most salient when they contrast with their opponents.
On the other hand, if combatants share an ideological position, we posit that conflicts may be shorter as belligerents are able to identify unilateral or bilateral concessions. While similar ideological positions of dissidents and the government may prevent many conflicts from emerging, rebel factions may engage in violent outbidding to shape the behavior of the state. For instance, the Pakistani Taliban often advocates for more religious reforms despite the fact that Pakistan is an Islamic Republic. During the Ethiopian civil war, the Ethiopian People’s Revolutionary Party advocated for socialist reforms against the Derg government even though the Derg adopted socialist policies as a way to garner stronger domestic and international support (particularly from the Soviet Union). Rebel groups may therefore attempt to outflank government officials on certain ideological positions in order to garner support. We contend, though, that when dissidents and the government share a similar ideological position, conflicts should be shorter as there may be fewer costs associated with offering concessions on this salient issue.
The adoption of a contrasting ideological position would therefore seem to bind the hands of rebel groups during armed conflicts (making simple negotiations less feasible). If this is the case, why would rebels engage in such a risky position? Assuming that rebel groups are aware of the possible downstream consequences associated with mobilizing support along the salient cleavage, we argue that insurgents often face a strategic dilemma when they begin the process of rebellion. By contrasting themselves substantively (and not just militarily) with the incumbent government, rebels are more capable of building a stronger groundswell of support. At the same time, such a strategy will make bargaining more difficult (limiting the number of meaningful concessions they are able to accept). On the other hand, rebels may adopt an ideological position that is not meaningful or one that is similar to that of the government, thereby expanding the opportunity to strike a deal with the state. This approach, though, will limit the range of support they may garner from the aggrieved population. Absent widespread support, rebels are unlikely to have the military capacity to push the state to offer meaningful accommodations.
These arguments fit well with past work on civil war dynamics. For instance, in research on constraints and civil war bargaining, Thyne (2012) provides multiple rich examples where peace is undermined by vested interests within the government. Civil wars may persist without a settlement even when leaders have information that a peace agreement provides the most viable avenue for ending fighting. Drawing on the example of the Second Sudanese Civil War, Thyne (2012, 309) notes that certain leaders within the government preferred continued fighting with the Sudanese People’s Liberation Movement/Army (SPLM/A), despite the fact that there was little hope of militarily defeating the entrenched southern rebels. This was largely because elites in Khartoum feared that meeting the secular demands of the rebels would “weaken the Islamic Movement…,” undercutting the ideological position of the President and Vice President of Sudan.
Increasing Indivisibility and Civil War Settlements
We contend, though, that not all ideological differences are equivalent. As opposed to Goddard’s (2006) view of indivisibility being binary, we suggest that ideologies (and the issues encompassed within those ideologies) operate along a range of divisibility. In other words, some disputes are more indivisible than others. We assume that more tangible disagreements (i.e., more divisible) have a greater number of available compromises. As ideological differences become less about tangible disagreements and more about intangible concerns, commitment problems are exacerbated given the indivisibility of the stakes (Goddard 2006; Svensson 2007; Wucherpfennig et al. 2012). As noted by Mason and Fett (1996, 554), “…to the extent that stakes in a civil war are more nearly divisible (in some sense), a negotiated settlement becomes more likely.” An increase in the indivisibility of ideological disagreements reduces the number of available compromises. As the number of available compromises decrease (shrinking the bargaining range), parties will find it more and more difficult to credibly commit to a settlement. Disputes over public spending, control of the executive, and opportunities for banned parties to participate in the political process tend to represent more divisible issues given the tangibility of the claims made. On the other hand, as disputes deal with less tangible issues, such as national identity, norms around gender (such as patriarchal governance), or the role of religion in society, conflicts become more intangible. Similar to past definitions of indivisibility, these ideological positions are complicated simply because any modification to these policies may reduce the intrinsic value of the issues at stake (Hassner 2003). When contrasting ideologies focus on more indivisible stakes, concessions become difficult to identify.
For example, while economic ideologies (such as socialism) deal with complicated philosophies on the role of the government in the market, we assume that economic policies are easier to modify than policies that reinforce ethnic dominance or the importance of religious practice. Major land reforms to ameliorate rural discontent may be followed by compensation packages offered to private individuals or corporations. Similarly, disputed territory with natural resources can be divided and shared among disputants. Equally, a liberalization of the economy may also be followed by special protections for certain key industries. This is less feasible for ideological differences over issues of ethnic nationalism, sovereignty, and religion. These perceptions of ideology are not easy for governments to win hearts and minds since these ideologies are more rigid and noncompromising compared to economic policies or disputes over power-sharing that can shift.
In their study of ethnic exclusion and civil war dynamics, Wucherpfennig et al. (2012) note that policies of political exclusion are difficult to roll back, given the substantial benefits afforded to groups in power. As noted by Hardin (1995), norms of exclusion are built upon the ability of the in-group to assume that they will have access to better services simply given their position in society. Equally, Rothchild (1981) suggests that these policies of political exclusion shape the worldview of elites in power, where they accept their privileged position as an inherent birthright as opposed to a product of years of political malfeasance. Religious ideology often plays a similar role, where concessions are difficult to identify given that any modification of religious law substantially reduces the intrinsic value of the religious practice (Hassner 2003; Svensson 2007). As contrasting ideologies begin to draw on these intangible aspects of ethnic prestige, religious adherence, or nationalist connections to territory, it is difficult for either side to identify an appropriate resolution as any concession on the issues will trigger significant discontent from both sides’ respective constituencies. This should further exacerbate commitment problems as neither side believes their opponent is capable of following through with any concessions.
One challenge is that rebels may have multiple contrasting ideological positions. While having multiple positions may complicate negotiations, we assume that contrasting ideologies centered on more indivisible stakes should take precedent over other ideological differences. While groups may contrast themselves with the government along various dimensions, bargaining should be particularly hindered by the most indivisible contrasting ideology (i.e., prioritized above other divisions). For instance, during the Second Sudanese Civil War, the SPLM/A sought to differentiate itself from the incumbent government along a number of ideological dimensions such as advocating for socialist economic policies and political agency for aggrieved Southerners. Despite contrasting itself along these numerous dimensions, the organization maintained that these distinctions were second to its primary concern over the role of religion within the state. The group’s leader noted that this was the cardinal issue to be addressed before all other concerns could be resolved (Garang 1991). As noted by Thyne (2012), this represented one of the key barriers for settlement between the two parties.
Research Design
In order to test the relative influence of contrasting ideological positions on civil war dynamics, we examine all instances of intrastate conflict termination from 1975 through 2011. Our data on intrastate conflicts are drawn from the Uppsala Conflict Data Program (UCDP) Conflict Termination Dataset (Kreutz 2010). Specifically, the UCDP Conflict Termination Dataset includes dyadic level data on the start dates, end dates, and outcomes of all intrastate conflicts in which at least twenty-five battle-related deaths occurred within a year of fighting. Given that we theorize about the dyadic relationship between rebel groups and the government, our unit of analysis is the civil war dyad year. As with contemporary studies on bargaining and intrastate conflict (Thyne 2012; Uzonyi and Wells 2016; Thyne 2017; Ryckman and Braithwaite 2020), we employ duration models to estimate the risk of civil war termination as well as outcomes of interest. For our sample population, the average duration of the intrastate conflict is roughly five years (4.47 years) with the maximum length of being forty-two years.
As previous studies have demonstrated that bargaining failures should increase the duration of conflicts, our initial tests will explore how contrasting ideological positions affect the risk of conflict termination. For this analysis, we rely primarily on a Cox proportional hazard model to estimate each model. Given that we have no theoretical reason to assume that the baseline risk of conflict termination increases or decreases over time, the Cox model is preferred (Cunningham, Gleditsch, and Salehyan 2009). 1 For the second part of our analysis, we explore how contrasting ideological positions affect the type of intrastate conflict termination. In particular, we expect that as contrasting ideological positions become more indivisible, the risk that the dispute will end in a peace agreement decreases. To test this proposition, we employ a competitive risk model (or CRM). 2 For the purposes of our analysis, we explore how contrasting ideological positions affect the risk of a peace agreement being reached as compared to all other civil war outcomes.
To operationalize our variables of interest, we rely on numerous data sets to capture both rebel group and government ideology. While we assume that ideological goals may center on a number of dimensions, there are limitations in the data that are available. Absent a comprehensive database of rebel group ideologies, goals, and stated grievances, we are limited to past measures that have been collected as part of other data projects where both government and rebel group ideological positions are identifiable. To that end, we identify three ideological positions for which we have available data: economic ideologies, ethnic ideologies, and religious ideologies. To capture left-wing and right-wing rebel group economic ideology, we draw on Polo and Gletidsch’s (2016) ACD2GTD data set that includes data on rebel group ideology from 1975 through 2011. Data on the government’s economic ideology are drawn from the World Bank’s Database of Political Institutions (DPI). Specifically, the World Bank’s DPI includes data on whether governments have a right-wing or left-wing economic ideology. Rebels are coded as having a contrasting economic ideology if the rebel group maintains a Marxist or left-wing ideology and the government maintains a right-wing ideology or vice versa. Rebels are coded as having a similar economic ideology if they share the same economic ideology as the government.
To capture ethnic-based ideology, we rely on the ACD2EPR data set (Wucherpfennig et al. 2012). Specifically, rebels are coded as having a contrasting ethnic ideology if they claim to represent the interests of an excluded ethnic group and the government is run by an ethnic group that maintains a monopoly over political power. 3 Rebels are coded as having a similar ethnic ideology if they claim to represent an ethnic group that is included in the national government. Finally, we also examine contrasting religious ideologies. Using the Religion and Armed Conflict Dataset provided by Svensson and Nilsson (2018) as well as Polo and Gleditsch’s (2016) data on rebel group ideology, religious rebel groups are coded as having a contrasting religious ideology if their religion differs from that of the central government and that religion is listed as a salient issue in the conflict (Svensson and Nilsson 2018). 4 As our initial hypothesis posits that contrasting ideological positions complicate the bargaining process, thereby increasing the duration of the civil war, our first indicator is simply a binary measure that captures civil war years where rebel groups maintain a contrasting ideology. Specifically, contrasting ideology (any) is coded as 1 when rebel groups maintain any of these types of contrasting ideologies with the government and 0 otherwise. Equally, we also include a measure for whether the rebel group maintains a similar ideology to that of the government. Similar ideology (any) measures whether rebels maintain a similar ideology to that of the central government. Frequencies of these contrasting and similar ideologies are shown by type in Table 1.
Percentage of Rebel Groups with Contrasting or Similar Ideologies with the Government.
Beyond simply capturing the presence of contrasting or similar ideological positions between the rebels and the government, we also argue that the divisibility of goals should play a unique role in shaping the bargaining space. Specifically, we argue that as contrasting ideological positions become more indivisible, bargaining should become more difficult. This should occur even when groups also contrast themselves along less indivisible stakes (such as material claims). In effect, our assumption is that the less divisible (more intangible) claims should be of greater priority in the conflict than other ideological divisions. To operationalize this measure, we generate an ordinal scale for ideological differences ranging from most divisible to least divisible. This approach is similar (though not the same) to the Issues Correlates of War data project, where the salience of disputes includes both tangible and intangible elements (Hensel 2001). The Issue Correlates of War data project specifically captures this phenomenon as an index. For the ideology variables that are available, we rank them based on their divisibility. Although we are limited by available data, we assume that the divisibility of ideological disagreements should be relevant across any number of ideological disputes.
As part of this measure, we assume that economic/material ideological positions are more tangible than ideologies surrounding ethnicity and religion. On the other hand, religious issues in civil wars are often treated as nontangible, where concessions on separate issues are insufficient to meet the demands of religious adherents (Hassner 2003; Svensson 2007). Religious issues are also difficult to subdivide into smaller concessions (Hassner 2003, 12), since such division can diminish the value of the religious practice. While issues of ethnic identity and political control are also highly salient, the common practice of multiethnic power-sharing suggests governments are more capable of finding concessions to ethnic grievances as compared to those of religious issues. We therefore operationalize indivisible contrasting ideologies as an ordinal scale ranging from zero, where there are no contrasting ideologies, to three, where the rebels and the government have contrasting religious ideologies. 5 For rebel groups that have multiple ideological positions, we coded for the ideological position that is most intangible as we theoretically assume those ideological differences should be less important than the more intangible divides. 6
Finally, we introduce a number of control variables to isolate the effects of ideological positions during the conflict. As rebel groups may rely heavily on contrasting ideological positions to supplement their relative weakness, we include a control for rebel group strength. Drawing on Cunningham, Gleditsch, and Salehyan’s (2013) Non-State Actors in Civil War data set, we include an ordinal scale of rebel group strength ranging from one, where rebels are much weaker than the central government, to four, where rebels are stronger than the central government. Contrasting ideological differences may draw international support for one or both sides, increasing the duration of the conflict (Balch-Lindsay, Enterline, and Joyce 2008; Cunningham 2010). We therefore control for whether there was a foreign military fighting on either side during the conflict. These data are drawn from the UCDP Conflict Termination Dataset. 7 Equally, ideological differences may only be salient in conflicts fought over government control versus self-determination conflicts. Using data from Kreutz (2010), we code whether the dispute involved a territorial incompatibility. 8 Ideological differences may become less relevant in conflicts where rebels rely on financing from lootable goods. Drawing on Lujala’s (2010) data set, we include a control for whether the conflict zone provided access to lootable goods (such as diamonds or drugs). Finally, we also include controls for conflict intensity (Kreutz 2010), the state’s logged gross domestic product per capita (Gleditsch 2002), and whether the civil war state is a democracy as measured by a value of six or greater on the revised Polity II score (Marshall, Gurr, and Jaggers 2016).
Findings and Analysis
Table 2 provides the results from the first part of our analysis. Specifically, each model in Table 2 estimates the risk of civil war termination with positive coefficients indicating an increased risk of civil war termination while negative coefficients suggest a decrease in the risk of civil war termination. In other words, as the risk of civil war termination declines, the duration of civil wars increases (leading to more intractable disputes). Model 1 estimates the risk of civil war termination while looking primarily at whether rebels maintain a contrasting ideology to that of the government. Model 2 looks more specifically at each form of contrasting ideology and their relative risk on civil war termination. Models 3 and 4, on the other hand, estimate the risk of civil war termination as contrasting ideologies become more indivisible. All models are estimated using robust standard errors that are clustered on the civil war dyad, and each model is stratified based on previous conflicts.
Risk of Civil War Termination.
Note: Robust standard errors are in parentheses. Clustered on civil war dyad. Stratified by past failures. GDP = gross domestic product.
*p < .1.
**p < .05.
***p < .01.
The initial results suggest strong support for our proposition that contrasting ideologies significantly decrease the risk of civil war termination. Specifically, when a rebel group has a contrasting ideological position with that of the central government, there is a significant decrease in the predicted risk of civil war termination. Using hazard ratios to predict this change, we note that as rebels draw a contrasting ideology to that of the central government, there is a 27.4 percent decrease in the risk of civil war termination (exp(−0.32 × 1) – 1 = –0.274). On the other hand, there does not appear to be a significant relationship between similar ideological positions and civil war termination. While one may expect that similar ideological positions may increase the risk of civil war termination as there are more areas of ideological agreement, the coefficient for Similar ideology (any) appears to be negative (though it fails to reach statistical significance). It may be that rebels have multiple ideological positions (for instance, an ethnic and economic ideological position). Further robustness checks explore whether rebel groups may employ both a contrasting ideology and a similar ideology (explaining the mixed results in Model 1). The results remain unchanged. 9
While contrasting ideologies may complicate the bargaining process, we also contend that the nature of contrasting ideologies should play a key role in stymieing the bargaining process. Specifically, as contrasting ideologies become more indivisible, bargaining should be even more difficult. Model 2 includes each form of contrasting ideology. While there is statistically significant (though weak) support for the argument that contrasting economic and ethnic ideologies increase the duration of armed conflict, contrasting religious ideologies appear to significantly decrease the risk of civil war termination. Finally, Models 3 and 4 look at our measure for increasing indivisibility of ideological differences. The results demonstrate that as contrasting ideological positions become more indivisible, the risk of civil war termination begins to decline, thereby increasing the duration of civil wars. Again, after accounting for the degree of indivisibility in contrasting ideologies, similar ideological positions held by the government and the dissidents still does not appear to affect the duration of civil wars. These results may suggest that, in conflicts where the rebels and the government hold a shared ideological position, other salient factors in society may be complicating the bargaining process. On the other hand, when rebels and the government disagree over indivisible issues, civil wars become much more intractable. Moving from no ideological differences to the most indivisible issue, the risk of civil war termination declines by 55.1 percent. 10,11 Figures 1 and 2 illustrate the significant differences between measuring any contrasting ideological position as compared to the indivisibility of ideological positions. Each figure charts the predicted survival of civil wars in conflicts fought over ideological differences as compared to no ideological differences. Steeper curves suggest a decreased life span of the civil war, indicating shorter conflicts. Figure 1 simply measures whether there was any contrasting ideological difference between the rebels and the government. Figure 2 tracks the predicted survival of armed conflicts as ideological differences become more indivisible. As suggested by the varying curves, as ideological differences become more indivisible, conflicts appear to persist much longer.
Risk of civil war termination.
Risk of civil war termination.
As our results on the relative indivisibility of issues may be driven by certain ideologies (such as ethnicity or religion) as compared to a measure that captures the increasing indivisibility of ideological differences, we reestimate our models while including each form of contrasting ideology with our ordinal scale. Table 3 provides the results from this analysis. As illustrated by Models 1 to 3, the ordinal scale remains robust even when including each form of contrasting ideology. Again, the results suggest that bargaining becomes more difficult as ideological differences become more indivisible.
Risk of Civil War Termination.
Note: Robust standard errors are in parentheses. Clustered on civil war dyad. Stratified by past failures. GDP = gross domestic product.
*p < .1.
**p < .05.
***p < .01.
For the second part of our analysis, we look at how contrasting ideological positions shape the outcomes of armed conflict. If contrasting ideological positions complicate the bargaining process, we should expect that civil wars will not only be longer but also be less likely to end in a peace agreement. Primarily, we expect that as contrasting ideological positions become more indivisible, it is less likely that either side will identify key concessions necessary for a settlement to be reached. Table 4 includes the results of our CRM analysis. The findings demonstrate that, as predicted, when ideological differences become more indivisible, there is significantly less opportunity to reach a peace agreement. In calculating the subhazard ratio of our variable of interest, the results suggest that moving from no ideological differences to the most indivisible issues, there is a 76.3 percent decrease in the predicted risk that the conflict will end in a peace agreement. Figure 3 charts the cumulative subhazard rate for the model in Table 4. As contrasting ideologies become more divisible, the predicted risk of a peace agreement increases over time.
Competitive Risk of a Peace Agreement versus Other Termination Types.
Note: Robust standard errors are in parentheses. Clustered on civil war dyads. GDP = gross domestic product.
*p < .1.
**p < .05.
***p < .01.
Cumulative subhazard of a peace agreement versus rival outcomes.
Conclusions
We began our analysis with an attempt to address the dilemma of why some civil wars are settled more quickly than others despite similar commitment problems in the bargaining process for all civil wars. Building on new research about ideology and older research on indivisibility and intangibility, we posited that contrasting ideological positions by governments and rebel groups would make bargaining toward concessions more difficult and civil wars longer in duration. Since governments and rebel groups depend heavily on their ideological positions to maintain support, it is difficult for disputants to turn against their ideological positions in negotiations without risking backlash from supporters. This argument assumes that there are domestic audience costs in bargaining, which may increase difficulty in recruiting members to the rebel or government side, difficulty in mobilizing support for the cause or the state, and potential removal from power as a form of punishment from a winning coalition. The key factor is that disputants can be unwilling to shift bargaining positions in negotiations because they are tied to ideological positions that are perceived as indivisible or intangible.
Based on this logic, we hypothesized that contrasting ideologies would make civil war termination less likely and make the duration of civil wars longer. We also hypothesized that contrasting ideologies focusing on the intangible aspects of ethnic identity and religion would make civil war termination even less likely and the length of wars even longer. First, we found strong support for the argument that contrasting ideologies significantly decrease the risk of civil war termination. Second, we found that the more indivisible and intangible the ideology—on a scale from economic as least intangible to ethnic ideology to religious ideology as the most intangible—the longer the duration of civil wars as expected. Not only are civil wars longer, but we also found that more indivisible ideologies substantively decrease the likelihood that a civil war will end in a peace agreement. This finding provides further support for the argument that contrasting ideologies, particularly those ideologies that are more indivisible or intangible, negatively affect civil wars. In summary, when rebels and the government disagree over indivisible issues, civil wars are not only less likely to be terminated, but they are also much more intractable.
There are several implications of this research. First, we should expect that negotiations involving governments and rebel groups with contrasting ideologies are going to be more difficult in the first place. Second, we should also expect to see third-party mediation or military intervention in these intractable conflicts. Third, as the disputed ideologies become more indivisible and intangible, negotiations—whether mediated or not—will be very difficult to help resolve the conflicts. Rather, when we see contrasting ideologies over religious identity in particular, these are the civil wars that are most likely to be intractable and most difficult to achieve a peace agreement. As with interstate disputes over territory, it may be more effective for third parties to intervene in disputes with less contrasting ideologies and less intangible issues if they want to achieve success in ending a civil war. More resources may be required, though, for third parties to effectively mediate conflicts defined by more indivisible ideological positions.
Supplemental Material
Supplemental Material, sj-pdf-1-jcr-10.1177_0022002720928414 - Mutually Assured Distrust: Ideology and Commitment Problems in Civil Wars
Supplemental Material, sj-pdf-1-jcr-10.1177_0022002720928414 for Mutually Assured Distrust: Ideology and Commitment Problems in Civil Wars by Eric Keels and Krista Wiegand in Journal of Conflict Resolution
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
Notes
References
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.
