Strategic forum selection and compliance in interstate dispute resolution

Abstract

This paper investigates strategic forum selection approaches and compliance with interstate dispute resolution settlements. Research shows that (a) management design features, like decision control and international organizations, affect compliance, (b) states strategically select among bilateral and third-party fora, and (c) anticipated settlement outcomes and enforcement inform pre-negotiation stages. A skeptical perspective suggests that states only select management approaches when they are confident in their ability to fully cooperate. Using Issue Correlates of War data, I test an endogenous model of forum selection and compliance. Statistical analysis finds that states strategically select management approaches in two cases: state-led mediation and ad hoc arbitration. Closer analysis suggests that, rather than tempering the promotion of mediation’s and arbitration’s pacifying benefits, conflict management scholarship should investigate more comparisons across dispute resolution approaches. This project’s multidimensional framework reveals that mediation and arbitration share similar foundations that logically link their selection and inclination toward compliance.

Keywords

Compliance conflict management forum selection compliance selection bias

Agreements reached through interstate dispute resolution enjoy a remarkably high success rate: one set of studies finds that states comply with more than 75% of settlements (Hensel, 2001; Hensel et al., 2006, 2008). What explains this high incidence of compliance? Extant research points to disputants’ perceptions of fairness (Brams and Taylor, 2000) and norms obligations (Simmons, 1999), in addition to enforcement. Third parties, in particular, prime these compliance incentives because they resolve bargaining problems and provide political cover (Beardsley et al., 2006; Gent and Shannon, 2011a).

Yet anecdotal evidence hints at the presence of concerns about compliance earlier in the process. Citing jurisdictional issues, China recently refused a Filipino proposal to arbitrate their South China Sea dispute (Sevastopulo and Landingin, 2014). China’s refusal was surprising because a legal decision could bolster its other claims, signal peaceful intentions to other rivals and make it more difficult for Manila and its allies to press the issue (Ching, 2014). One possible explanation for China’s reluctance to accept arbitration is loss aversion; some describe China’s legal claims as tenuous (Cohen, 2013). In turn, China uses its military and economic leverage to encourage bilateral negotiations. Facing two important compliance barriers—potential losses and outside options—China avoids legal dispute resolution.

China’s South China Sea claim management exemplifies the possibility that non-compliance is uncommon because states only participate in management processes with which they are confident in their ability to fully cooperate. Do states submit to third party conflict management only when they plan to comply? In other words, is settlement compliance predicated on the factors that make the management strategy acceptable?

To answer these questions, I explore forum selection, the process of choosing a management venue, and its relationship to settlement compliance. States choose fora that balance relative bargaining power, issue salience and beliefs about settlement outcomes. Conflict management forum features that the literature already connects with compliance, such as binding approaches, international organizations (IGOs), joint democracy and concession, also inform these decisions and potentially make compliance predicated on this pre-negotiation step.

Empirical analysis using data from the Issue Correlates of War project tests this endogenous selection assumption. A two-stage model first estimates a multinomial logistic regression of forum selection. The second-stage, logistic regression analyzes settlement compliance, accounting for forum selection processes. I find support for the strategic forum selection logic; states incorporate factors related to compliance when choosing management strategies. I also find evidence of endogeneity in the success of some, but not all, conflict management approaches. Approaches that invite other states as third parties, such as mediation and ad hoc arbitration, are successful, in part, because only states that plan to comply choose these approaches. Compliance with agreements reached through IGO mediation suffers from a reverse selection effect: the factors that make IGO mediation acceptable reduce compliance. In contrast, I find no systematic evidence of endogeneity in the success of international adjudication.

This project theoretically and empirically links belligerents’ strategic interests and dispute resolution behavior through a novel, three-dimensional framework. The forum design permits comparisons across several approaches, including negotiation, mediation and legal methods, and establishes the groundwork for capturing possible endogenous selection. Through this framework, I respond to one enduring point of skepticism about compliance in international relations: states’ compliance with dispute resolution settlements is predicated on the strategic and contingent factors that lead them to select a particular forum. Nonetheless, these results enhance our understanding of dispute resolution efficacy, rather than dismantle it. The evidence of strategic forum selection highlights the commonalities of alternative dispute resolution, like mediation and arbitration—which are frequently studied in isolation—and, thus, suggests new avenues for future work that bring conflict management and international law scholarship closer together.

The remainder of this paper proceeds with a summary of the settlement compliance literature that connects management outcomes with forum characteristics. These compliance expectations are next incorporated into a theory of forum selection. The theory and analysis particularly focus on the role that compliance intentions play in strategic forum selection. Presentation of the empirical results precedes discussion of implications and conclusions.

Compliance and conflict management

Three prominent explanations for settlement compliance focus on issue division fairness, enforcement mechanisms, including reputation, and norms. Concessions, which change the status quo, are necessary for dispute resolution agreement. Concessions are be predicated on disputants’ relative bargaining power (Fearon, 1995; Powell, 1996) and claim strength (Huth et al., 2011). Yet, Beardsley and Lo (2014) note, concessions introduce two problems for settlement compliance. First, concessions heighten commitment problems because shifting the status quo might incentivize future challenges and plant perceptions of weak resolve. Second, unpopular concessions might fail to be ratified or swing support to an opposition candidate. Even when states can overcome these barriers, perceptions of fairness remain important. When one side believes that it pays all the costs while an adversary reaps all the benefits, the peace is more likely to be fragile (Brams and Taylor, 2000).

Agreements are more likely to adhere with enforcement. Some enforcement is direct and long term. A bilateral example, the 1991 Strategic Arms Reduction Treaty (START I), created a joint monitoring commission. Elsewhere, third parties routinely deploy peacekeepers and inspectors (Beardsley, 2008; Smith and Stam, 2003). Third-party capacity to monitor and enforce settlements, however, is limited and weak. Instead, enforcement often comes indirectly from reputational penalties. Non-compliance with settlement agreements indicates—to rivals and mediators, among others—that a state is untrustworthy. Therefore, states are motivated to fulfill treaty obligations because they hope to cooperate in the future (Crescenzi, 2007; Guzman, 2002).

A third settlement compliance explanation adds the influence of norms. A well-understood principle is that parties perform the terms of their agreements in good faith, or pacta sunt servanda (Simmons, 1998). Settlements, whether reached bilaterally or through third parties, confer obligations. Given the widespread observation of pacta norms (Chayes and Chayes, 1993), some also conclude that reputational penalties and treaty obligation reinforce work together to enhance compliance (Guzman, 2002; Mitchell and Hensel, 2007).

Forum design is consequential to settlement compliance

Different management approaches prime these compliance influences. Legal dispute resolution, in which an arbitral panel or international court delivers a binding award or decision, draws on pacta norms (Simmons, 1999) and produces clearly articulated settlements based on principles of equity. This promotes fairness and facilitates monitoring (Gent and Shannon, 2010). Mediators that exert leverage, as in manipulative mediation, stake their reputations on management success and consequently produce political cover and enforcement. Facilitative approaches, alternatively, coax disputants to mutually acceptable agreements (Beardsley, 2010; Beardsley et al., 2006). Although mediated agreements are not binding, states may feel obligated owing to reputational concerns (Bercovitch and Gartner, 2006).

Other third-party characteristics also trigger compliance mechanisms. Institutions provide credible information about capabilities, resolve and issues. IGOs also increase reputational consequences and promote pacta norms (Busch, 2007; Mitchell and Hensel, 2007). Some organizations also perform mediation or legal dispute resolution and directly enforce agreements (Boehmer et al., 2004; Shannon, 2009).

Outside IGOs, democracies are found to be more credible mediators than non-democracies because they are transparent and promote dispute resolution norms (Mitchell et al., 2009). These traits also make democracies more compliant in settlements reached bilaterally (Dixon, 1993). Major powers are more likely to have the political, military and economic capital to monitor implementation. Smaller states also aid enforcement because they care about their reputations as mediators (Beardsley, 2009; Favretto, 2009). Finally, biased intermediaries, such as military or political allies, are arguably more trustworthy than impartial third parties that focus on peace than on issue divisions (Kydd, 2003, 2006; Touval, 1975).

This cursory summary reveals great diversity in conflict management and its influences. There is also significant variation in the incidence of different management approaches. The Issue Correlates of War data, for instance, observe that bilateral negotiations comprise almost 70% of all peaceful settlement attempts, although other approaches, such as legal dispute resolution, are significantly more successful at bringing about compliance. Through a skeptical lens, infrequent use of highly successful management approaches may be explained by factors, including concerns about compliance, that cause states to choose particular strategies. In part, a compliant predisposition may be a necessary condition for some conflict management methods to occur in the first place (Downs et al., 1996).

Strategic forum selection: The elements of choice

Addressing the skeptical perspective that settlement compliance is predicated on strategic forum selection requires an organizing scheme that (a) permits comparisons across a large range of management approaches, including bilateral negotiation, and (b) integrates the compliance-priming effects of different conflict management processes. I propose a theory of conflict management that defines a management forum along three dimensions: decision control, transparency and distributional bias. Defined below, these dimensions are drawn from existing descriptions of conflict management fora and generalize across the different approaches.

Decision control

Decision control is the authority that actors exert over outcomes (Thibaut and Walker, 1975). States delegate complete decision control to arbitration and adjudication—which increases the compliance pull of pacta norms—and share decision control in bilateral negotiations. Mediation falls in-between these two points, with third-party decision control increasing over facilitative, transformative and manipulative styles. Delegating decision control facilitates issue division. As Beardsley and Lo (2014: 369) explain, third-party leverage “enables for a focal enforcement solution to emerge”. Yet ceding decision control increases a state’s risk of receiving an unsatisfactory division of the issue and undermines compliance. However, third parties that control issue division may act as scapegoats and provide political cover (Allee and Huth, 2006).

Transparency

When direct sanctions are weak, settlements are enforced through reputational penalties (Guzman, 2002). Reputations are built on observations of other actors’ behavior (Crescenzi, 2007). Transparency defines the degree to which outside audiences can peer in. Transparency aids compliance and binds adversaries to peace because it facilitates monitoring and detection of non-compliance (Fang, 2010; Lohmann, 2003). In exchange, forum transparency increases audience costs because it is more difficult for a mediator to convince domestic and international audiences of its responsibility for concessions when those audiences can observe the process. Secrecy, transparency’s counterpart, permits candor (Finel and Lord, 2002; Strasavage, 2004).

Transparency varies independently from bilateral/multilateral distinctions. Some bilateral negotiations, like the 1989 Fashoda Crisis between the UK and France, are almost detrimentally public (Finel and Lord, 2002). In contrast, Qatar’s secret mediation between Eritrea and Djibouti limited domestic reprisals.¹ Arguably though, transparency increases with the number of actors involved. With Qatar’s inclusion, the Eritrea–Djibouti negotiations were more transparent than the secret Israel–Jordan “picnic table” talks between 1948 and 1994 (Wolf et al., 2009). More generally, global IGOs are more transparent than regional organizations or ad hoc groups because negotiation outcomes are relevant to more actors in a global IGO (Busch, 2007). Bilateral negotiations are less transparent than most multilateral fora, but domestic regime type, such as democracy in the Fashoda case, may alter this simple categorization. In sum, transparency represents the forum characteristic that primes enforcement mechanisms (monitoring, reputation, political cover) of compliance.

Distributional bias

Finally, compliance depends on whether disputants believe that the settlement outlines a fair division of the issue (Brams and Taylor, 2000). A management forum’s distributional bias represents its issue-dividing principle, which follows from claimant and third-party preferences. The degree to which any individual influences distributional bias depends on decision control. In arbitration and adjudication, third-party preferences define the forum’s distributional bias. Claimant and third-party preferences in lower decision control fora, like mediation, comprise a proportionately representative distributional bias. The distributional bias of bilateral fora is based on disputants’ relative bargaining power.

Distributional bias differs from other definitions of bias, such as affinity between a mediator and claimant (Kydd, 2003, 2006; Rauchhaus, 2006). Although expectations about issue division are informed by political and economic bias, distributional outcomes (and states’ beliefs about them) are distinct because they address concerns related to compliance, as opposed to information transmission or trust.

Selecting a forum

These three forum dimensions shape peaceful dispute resolution, as Figure 1 illustrates. Working backward, extant research indicates that higher decision control and transparency increase compliance because they trigger reputational penalties, political cover and norms. Disproportionate issue divisions that, for instance, impose large concessions on a stronger rival increase non-compliance (see also Mitchell and Hensel, 2007). Before this stage, disputants work through negotiation after a two-step forum selection phase.

Figure 1.

Conflict management stages—selection and compliance.

I assume that states are strategic, rational actors that consider the down-game consequences of forum selection. Given a set of beliefs about potential intermediaries’ bias, states bargain over transparency and decision control in order to maximize their share of the divided issue relative to the trade-offs necessary to make those settlements hold.

Distribution of power and outside options

As Figure 1 illustrates, forum selection is a bargaining process that requires mutual consent. In bargaining, actors that have outside options to a negotiated settlement have more leverage (Powell, 1996). One alternative to peaceful dispute resolution is war. States that enjoy a military advantage have more opportunities to abandon the bargaining process and escalate. Therefore, stronger states are more likely to coerce settlements that favor their preferred division of the issue and, accordingly, press for bilateral negotiation (Gent and Shannon, 2014). The weaker state accepts this proposal, despite its distributional bias, because a peaceful settlement is less costly than fighting.

Hypothesis 1: Disputes characterized by power asymmetries are more likely to negotiate bilaterally and limit third party decision control.

Disputes characterized by balances of power are, instead, more likely to be managed multilaterally. Third-party influence, especially in high control and high transparency fora like arbitration and adjudication, establishes a focal solution, political cover and reputational enforcement (Gent and Shannon, 2014).

Because disputes between asymmetric rivals are more likely to be negotiated bilaterally, their management will also be less transparent. Furthermore, unlike conflicts between adversaries of equal power, agreements reached between militarily unequal rivals are more easily enforced. First, the stronger state has the capability to directly enforce the agreement. Second, the eventual issue division is unlikely to shift the distribution of power significantly enough to introduce commitment problems. Thus, disputes characterized by power asymmetries are less likely to adopt transparency.

Hypothesis 2: Disputes characterized by power asymmetries are more likely to be negotiated privately and limit transparency.

Domestic sources of transparency

Direct monitoring and punishment are not the only ways to enforce settlements. Settlements also adhere because states are motivated to protect their reputations and uphold norms. Among the three forum dimensions, transparency most directly contributes to these compliance mechanisms. However, not all states must turn to third parties to increase management transparency. Indeed, democracies possess a transparency endowment. Ellis et al. (2010) particularly cite democracies’ open systems of governance, regime continuity, political accountability and commitments to institutions and rule-of-law as features that make democracies more reliable bargaining partners. In disputes between two democracies, third parties do not substantially increase management transparency. Therefore, democracies prefer bilateral negotiations to third party approaches—especially those offered by other states.

Hypothesis 3: Jointly democratic dyads are more likely to negotiate bilaterally and limit third-party transparency.

Hypothesis 3 runs counter to other descriptions of democracies’ conflict management preferences. Democratic leaders are more vulnerable to audience costs resulting from unpopular concessions than leaders of other regimes (Allee and Huth, 2006; Tarar and Leventoglu, 2009). As a result, democracies have a greater need for political cover. Mediation provides some political cover, but arbitration and adjudication allow disputants to consign more blame (Beardsley and Lo, 2014). Thus, democracies also prefer approaches that delegate complete decision control to the third party.

Hypothesis 4: Jointly democratic dyads are more likely to delegate complete decision control to a third party than to delegate partial decision control.

From the forum design framework, Hypotheses 3 and 4 conflict. A preference for bilateral negotiation on one dimension should correspond with a similar preference on the other. Hypothesis 4 does not directly compare the two, though. Democracies’ preferences for legal dispute resolution are contingent on the need for political cover or pacta norms. For instance, democracies are less likely to negotiate bilaterally during election cycles (Simmons, 2002). Additionally, democracies need third parties when bilateral relations break down. Even though militarized violence between democracies rarely occurs, it erodes democratic peaceful dispute resolution norms (Ellis et al., 2010). Hypothesis 4 reflects democracies’ preference for complete decision control fora relative to partial control fora when bilateral solutions are not acceptable owing to concerns about audience costs and enforcement.

Issue salience and concession-making risks

When states submit to high-control, legal dispute resolution, they open themselves to risk. Former Libyan leader, Muammar Qaddafi, learned this lesson in his consent to International Court of Justice (ICJ) adjudication over possession of the Aouzou Strip. Against Qaddafi’s optimism, the Court awarded the entire territory to neighboring Chad (Paulson, 2004). As a result of cases like this, some disputants will be reluctant to take advantage of binding conflict management because the value of the issue under contention is too important or salient (Hensel, 2001). Unfortunately, highly salient issues are also more difficult to resolve because the value of the claim makes focal solutions more difficult to identify and exacerbates commitment problems, which are difficult to resolve bilaterally. Unproductive face-to-face talks are unlikely to be more enticing than continued disagreement.

Together, these two observations suggest a non-linear relationship between issue salience and forum selection. As issue salience increases, bilateral negotiation and legal dispute resolution decrease. Instead, mediation offers a compromise between these two levels of decision control.

Hypothesis 5: As issue salience increases, the selection of partial decision control fora should increase, while bilateral negotiation and complete decision control fora decrease.

Issue salience should also be associated with lower levels of transparency. On the one hand, transparency increases the scope of reputational penalties for non-compliance, which deters opportunistic abrogation resulting from commitment problems. On the other hand, transparency also reduces political cover. A third party facilitating transparent negotiations will find it more difficult to claim responsibility for concessions—concessions that are even more important to domestic opposition groups than in less salient issues (Beardsley and Lo, 2014). Leaders managing contentious issues—that also hope to retain power—prefer management fora that mitigate audience costs to those that improve compliance monitoring.

Hypothesis 6: As issue salience increases, forum transparency decreases.

Together, Hypotheses 5 and 6 suggest that states managing highly salient disputes may be more likely to use state-led mediation, which facilitates settlement, provides a modicum of political cover, and boosts reputational penalties for non-compliance over bilateral settlement.

Beliefs about third-party preferences

Upon agreement, a forum’s distributional bias is evident through the division of the issue. At the outset, however, states are uncertain about third parties’ preferences. Negotiators conduct research and draw on past experience to clarify expectations (Fischer, 1982; Paulson, 2004; Wiegand and Powell, 2011). Therefore, preferences over decision control and transparency are also shaped by disputants’ beliefs about third-party preferences, or distributional bias.

Prior to any management offer, states’ beliefs are based on the economic and political biases of potential intermediaries and access to institutions (Beber, 2012; Crescenzi et al., 2011). Trade and alliances between each belligerent and a possible mediator reveal biases, and states are likely to veto a third party they believe is biased against them. Therefore, disputes are more likely to be managed bilaterally when states perceive third-party distributional bias.

Hypothesis 7: States are more likely to negotiate bilaterally, and limit decision control to third parties, when potential intermediaries are economically or politically biased.

Hypothesis 7 coincides with the bargaining logic that actors will only accept offers they prefer at least as much as continued disagreement. States may also reject biased mediation because of the enforcement and norms-evoking effect of third-party management. For example, the 1871 Halifax Fisheries Commission arbitration award that required American reparations to Great Britain was “considered excessive and exorbitant, and many doubted if it were lawful and honorably due” (Lord, 1892: 482). Nonetheless, the US followed through because reneging might have complicated other US fisheries claims. Therefore, states also prefer less transparent fora when they perceive third-party bias in order to limit their compliance pull.

Hypothesis 8: States are more likely to negotiate bilaterally, reducing transparency, when potential intermediaries are economically or politically biased.

Also in concert with the bargaining logic underlying Hypothesis 7, states that share memberships in more peace-promoting IGOs are more likely to prefer third-party management. Membership in several IGOs establishes a wider range of distributionally impartial third parties. Institutions additionally have incentives to build consensus and make fair decisions in order to gain states’ long-term cooperation (Koremenos, 2007; Shannon, 2009). Therefore, states may delegate decision control to institutional third parties because they promote equitable issue division and increase the scope of reputation costs for non-compliance.

Hypothesis 9: States that share peace-promoting IGO memberships are more likely to delegate decision control to third parties.

Hypothesis 10: States that share peace-promoting IGO memberships are more likely to select highly transparent management fora.

Some also speculate that states that participate in more peace-promoting IGOs might have internalized peaceful dispute resolution norms and use IGOs less strategically. Shannon (2009) tests this specific claim and finds no support for this passive role of IGOs on conflict management strategies.

Figure 1 summarizes these forum selection hypotheses that link belligerents’ strategic interests and management outcomes. Once states agree on a management forum, they work toward producing a settlement, which sets up compliance opportunities. At any stage in this conflict resolution process, a state may unilaterally withdraw; this also has implications for whether states decide to comply. This project focuses on the forum selection stage in order to draw connections between compliance and early conflict management choices.

Research design and data

The empirical analysis focuses on strategic forum selection and its effects in settlement compliance. To test this intuition, I use the Issue Correlates of War (ICOW) dyadic claim and settlement attempt data in a two-stage analysis of settlement compliance. The ICOW data cover dispute characteristics and settlement attempts for territorial claims in the Americas and Western Europe from 1816 to 2001, river claims in the Americas, Western Europe and the Middle East from 1900 to 2001, and maritime claims in the Americas and Western Europe from 1900 to 2001 (Hensel, 2001; Hensel et al., 2006, 2008; Nemeth et al., 2006). The ICOW data are particularly useful in constructing the dependent variable for the first-stage analysis as they uniquely record management efforts in a range of approaches and actors. The forum selection dependent variable is a five-category operationalization of management fora. The baseline category, bilateral negotiation, is coded 0. A forum is coded 1 if it involves the use of mediation, good offices, inquiry or facilitation (partial decision control) and any non-IGO third party and 2 if it involves an IGO.² Binding third-party approaches, like arbitration and adjudication, are identified as complete decision control fora. When binding management involves non-IGO actors, the forum is coded 3. IGO-delegated complete decision control is coded 4. Although decision control and transparency are both ordinal, multinomial logistic regression is appropriate because the combinations of transparency and decision control create non-ordered categories.³ It would be difficult, for instance, to determine whether mediation by an IGO is of higher or lower order than arbitration by an ad hoc group of states.

The first theoretical variable is relative power, measured as the challenger’s share of dyadic capabilities (CINC; Singer, 1987; Singer et al., 1972). Next, joint democracy is a binary variable, coded 1 if both claimants are democracies (Polity score ≥ 6) according to the Polity IV dataset (Marshall and Jaggers, 2002). ICOW issue salience, the combined tangible and intangible importance of the claim to both disputants, represents the value of the claim (Hensel, 2001).

At the outset of negotiations, states may be uncertain of their third-party options (Bercovitch and Schneider, 2000; Greig, 2005; Melin, 2014). Nonetheless, three values may capture beliefs about distributional outcomes. First, I adopt two measures of beliefs, economic bias and political bias, from Crescenzi et al. (2011). Each settlement attempt contains a challenger–target–potential mediator triad. Potential mediators are drawn from the set of states in the same region as the dispute, plus major powers. IGOs and multilateral groupings are excluded. Economic bias is measured as the absolute difference in total trade (Barbieri, 2002) between the potential mediator and each claimant.⁴ Political bias follows the same procedure as economic bias, using alliance portfolio similarity to measure affinity (S-scores, weighted and global; Signorino and Ritter, 1999). This construction of potential third-party bias necessitates a triadic unit of analysis in the first-stage models. Each observation in the forum selection stage begins with an individual settlement attempt and then expands that attempt by the number of challenger–target–potential third party trios.

Finally, shared IGO memberships increase access to higher-transparency and impartial fora. Joint IGO membership counts the number of memberships that claimants share in institutions that contain peaceful dispute resolution clauses which directly affect conflict management behavior (Boehmer et al., 2004; Shannon, 2009). These data are from the Multilateral Treaties of Pacific Settlement (MTOPS) data (Hensel, 2006).

To control for other dispute characteristics, I include the number of successful settlement attempts related to the claim, number of failed attempts and the incidence of fatal militarized disputes in the previous five years. Recent successes ratchet toward higher decision control (Bercovitch and Gartner, 2006; Böhmelt, 2013), while recent management failures indicate bargaining and enforcement problems, which are also addressed through decision control. Fatal violence (COW hostility level ≥ 4) should make states more open to third-party management more generally to gain assurances while violence is suspended. Finally, I control for the issue at stake. States manage territorial disputes differently than other types of conflicts (Hensel et al., 2008; Vasquez and Henehan, 2001), and these comprise the baseline issue for comparison. All of these variables are from the ICOW project.

Compliance stage

The second-stage model analyzes mutual settlement compliance. ICOW defines compliance (coded 1) as mutual implementation of the agreement within five years of its signature or a stipulated time period and then its continued execution for at least another five years. Non-compliance includes willful abrogation and failures resulting from changes in domestic or international conditions (Mitchell and Hensel, 2007). There are 848 compliance opportunities in the dataset.

To account for selection, I estimate the compliance model with multinomial selection. This approach is based on the familiar two-stage Heckman (1979) procedure. However, because the selection stage is multinomial logit, following Lee (1983) and Dubin and McFadden (1984), who apply multinomial logistic regression to Heckman selection, I calculate the inverse Mills ratio (IMR) for each categorical outcome of the forum selection model:

\sum_{j \neq i}^{m} γ_{j} [\frac{\hat{P_{j}} \ln (\hat{P_{j}})}{1 - \hat{P_{j}}} + \ln (\hat{P_{i}})]

The IMR function sums the truncated normal distributions of the predicted probabilities for each categorical outcome, m. These values are included in the second-stage model.

The second-stage analysis is necessary to test whether strategic forum selection plays a role in conflict management behavior, including compliance. Therefore, to facilitate comparisons with other analyses that use the ICOW data to study settlement compliance and claim termination, the compliance model replicates previous studies of forum design variables.⁵ Decision control is coded 1 if states used binding conflict management and 0 for non-binding mediation and bilateral negotiation. Arbitration and adjudication increase the probability of compliance (Gent and Shannon, 2010). Forum institutionalization and joint democracy represent transparency. The IGO-forum variable is coded 1 if the management forum was a regional or global IGO in the ICOW dataset and 0 otherwise. Non-IGO fora are coded 1 when the management effort was led by a state or ad hoc group of states. Joint democracy is coded the same as in the first-stage model. Transparency created by IGOs and joint democracy should increase the probability of compliance (Mitchell and Hensel, 2007). Non-IGO fora should also increase settlement compliance relative to bilateral negotiation, the omitted category, because mediation provides more transparency and some political cover (Beardsley, 2008).

Agreements reveal the distributional bias of the forum, which is represented as the stronger state’s concessions. Strong-side concessions are coded 1 if larger concessions were imposed on the militarily stronger party. More powerful states are likely to perceive concessions as disproportionate and renege (Brams and Taylor, 2000; Gent and Shannon, 2014; Mitchell and Hensel, 2007). Finally, I interact strong-side concessions and decision control to test the conditional logic that binding dispute resolution increases compliance incentives through pacta norms, even when decisions are unfair (Lord, 1892).

Controls account for other dispute contingencies that affect compliance. States in symmetric dyads are more likely to renege because settlements redistribute the balance of power. More salient conflicts may relapse because concessions are more difficult to uphold. States that participate in more peace-promoting organizations and are more closely aligned (dyadic, weighted, global S-score) share common interests that encourage cooperation. Finally, because some issues may have different patterns of compliance, I control for maritime and river issues.

Empirical results

Table 1 presents the forum selection results. The reference category is bilateral negotiation. The analysis supports the argument that states incorporate concerns about compliance when they select management fora, particularly those that apply to decision control. Dyadic issue salience, joint democracy and IGO membership all increase states’ willingness to cede decision control. Additionally, balances of power encourage states to pursue arbitration. Also consistent with the theory, politically biased third parties decrease third-party conflict management. There is mixed support, however, for transparency’s role in forum selection: only political bias and IGO membership performed as expected. The analysis suggests, instead, that highly transparent fora may provide services beyond compliance enforcement. They also legitimize strong states’ actions. The following describes the statistical results in greater detail and discusses their implications.

Table 1.

Selection stage—multinomial logistic regression of forum selection in territorial, maritime and river claims, 1816–2001

	Management forum
	Non-IGO, partial decision control	IGO, partial decision control	Non-IGO, complete decision control	IGO, complete decision control
	Coefficient	Coefficient	Coefficient	Coefficient
Variable	(standard error)	(standard error)	(standard error)	(standard error)
Relative power	0.009	1.202**	−1.035*	1.188*
	(0.341)	(0.573)	(0.505)	(0.652)
Joint democracy	−0.751***	−0.279	−0.399	0.782
	(0.249)	(0.302)	(0.572)	(0.637)
Issue salience	0.132**	0.126**	−0.112	−0.011
	(0.054)	(0.058)	(0.077)	(0.113)
Potential economic bias	0.011	−0.052	−0.069	−0.012
	(0.090)	(0.088)	(0.123)	(0.113)
Potential political bias	0.305	−1.693**	−1.994*	−2.492*
	(0.328)	(0.686)	(1.155)	(1.416)
Access to IGOs	0.014	0.109**	−0.063	0.102
	(0.048)	(0.053)	(0.089)	(0.091)
Recent successful attempts	−0.068	−0.017	0.259**	0.042
	(0.063)	(0.109)	(0.129)	(0.251)
Recent failed attempts	0.086	0.141**	−0.012	−0.247
	(0.057)	(0.064)	(0.126)	(0.227)
Recent fatal MIDs	1.120***	0.820**	1.303***	0.427
	(0.306)	(0.389)	(0.335)	(1.131)
Maritime issue	0.021	0.808*	−0.898	0.455
	(0.248)	(0.426)	(0.686)	(0.552)
River issue	−0.644	1.280***	−1.289	−0.765
	(0.507)	(0.475)	(0.123)	(1.162)
Constant	−2.303***	−4.289***	−1.716***	−4.673***
	(0.466)	(0.527)	(0.581)	(1.085)
Number of times selected	332	130	37	33
	N = 42,764, pseudo R² = 0.0821

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Robust standard errors adjusted for 289 dyadic claims. Bilateral negotiations (used 1155 times) are the base outcome.

MID, Militarized interstate dispute.

Conflict bargaining processes predict that outside options give stronger states leverage in bilateral fora. Consistent with Hypothesis 1, balances of power result in the selection of non-IGO/complete decision control fora. However, contrary to expectations, power asymmetries are associated with higher decision control and transparency. Despite having attractive bilateral alternatives, asymmetric rivals select IGO mediation and adjudication, in opposition to Hypotheses 1 and 2. Quinn et al. (2006) help explain this counterintuitive result. Stronger states use third parties to persuade weaker adversaries of their peaceful intentions and make concessions palatable. IGOs, especially international courts that elicit pacta norms, may be more credible messengers (Mitchell and Hensel, 2007). IGO consultation is also shown to legitimize states’ actions in crisis bargaining, gauge international support and lower the costs of foreign policy (Chapman and Wolford, 2010: 229). In sum, highly transparent management fora like IGOs provide services besides compliance enforcement in asymmetric disputes.

Hypotheses 3 and 4 focus on the effects of democracy on forum selection. Table 1 shows that joint democracy is negatively associated with the selection of non-IGO/partial decision control fora, which means that democracies negotiate bilaterally instead. This supports the claim that democracies inherently possess traits that promote peaceful dispute resolution (Ellis et al., 2010; Mitchell et al., 2009). The results also imply that, because democracies generally negotiate bilaterally, they are less likely to delegate decision control. Disputes between democracies are no more likely to be resolved through legal dispute resolution than conflicts among authoritarian or mixed dyads. Thus, the analysis does not support Hypothesis 4. Ellis et al. (2010) find that democracies’ participation in third-party conflict management is contingent on other features of their conflict relationship and domestic conditions; the analysis does not account for these interactions. Therefore, the results in Table 1 do not withdraw support from other research that finds democracies benefit from mediation and legal dispute resolution (Dixon, 1993). Democracies that decide to open their disputes to third-party management may be more likely to use legal dispute resolution, but, when comparing across all management options, democracies initially prefer to negotiate bilaterally.

Issue salience positively and significantly coincides with mediation (non-IGO and IGO), but does not predict complete decision control or bilateral approaches. This supports the hypothesized non-linear relationship between issue salience and decision control (Hypothesis 5). As Gent and Shannon (2011b) explain, legal dispute resolution is a risky gamble and states in highly contentious conflicts may be hesitant to trust a third party to deliver a binding decision. At the same time, these conflicts are more difficult to resolve bilaterally because the importance of the issue introduces commitment problems and audience costs (Hensel, 2001). Such reluctance also explains why some disputants avoid highly transparent fora. Transparency reduces political cover, which should make IGOs less desirable as third parties when salience is high (Hypothesis 6). I find, however, that issue salience correlates with IGO-led mediation.

Why might states in disputes characterized by higher issue salience work with IGOs to resolve conflicts? These states may simply prefer mediation, regardless of the mediator. Gent and Shannon (2011a) reach a similar conclusion that management approaches influence behavior more than third-party bias. The result is also consistent with the theoretical expectation that mediation is a strategic choice. Mediation provides a focal point for negotiation (Bercovitch and Gartner, 2006), which facilitates fair issue divisions and increases political cover and enforcement relative to bilateral negotiations. There are few reasons to expect that IGOs are less capable of serving in these capacities than states.

IGOs’ transparency, however, reduces political cover, unless we make two new assumptions. First, IGOs may be able to work behind closed doors (Strasavage, 2004). The UN Secretary General, for example, often works through informal channels to facilitate cooperation among states (Skjelsbæk, 1991). Second, an IGO may adopt a more directive role. This may be the intention of conciliation panels, for instance, which deliver independent recommendations that states often implement without modification (Merrills, 2005). These alternatives highlight the fluidity of forum transparency. Transparency generally increases with the number of actors, but parties can manipulate audiences’ ability to observe bargaining processes.

Consistent with Hypotheses 7 and 8, political bias increases the use of bilateral negotiation, but economic bias has no effect on forum selection. One reason why economic bias does not significantly influence forum selection is that third parties fear disrupting economic ties. Third parties typically have trade relations with both belligerents (Crescenzi et al., 2011), and settlements that redistribute the issue could adversely affect future economic exchange. The mediator, furthermore, might be unwilling to enforce damaging concessions. Thus, economic bias may have little effect because trade partners face different enforcement challenges from other types of mediators.

When a potential mediator’s distributional bias is measured according to military alignment, states negotiate bilaterally, as indicated by the negative and statistically significant results for three out of four management strategies: IGO mediation and adjudication and ad hoc arbitration. Unlike trade partners, political interests are more clearly identified through treaty obligations. Disputants reasonably assume that allies’ shared interests influence a mediator’s distributional bias (Gelpi, 1999). Bilateral negotiation avoids potential unfairness.

Finally, access to peace-promoting IGOs increases multilateral management. Consistent with Hypothesis 10, IGO memberships increase the use of IGOs in conflict management. This supports the theoretical expectation that bargaining encourages impartial management. While IGO membership promotes management transparency, the effect is limited to mediation. Compliance-inclined states might submit to adjudication to further maximize settlement fairness, political cover and norms, but there is no strong theoretical reason to suspect that the availability of impartial IGOs increases legal dispute resolution. The results support Hypothesis 9, which simply expects states to delegate more decision control relative to bilateral negotiation.

The dispute context controls perform generally according to expectations. Recent management successes guide states toward ad hoc arbitration. When states have more recent management failures, they, instead, prefer more transparent, lower decision control strategies. When a dispute is characterized by recent violence, though, the belligerents are more likely to pursue almost any type of third-party management over bilateral negotiations. These disputes are more likely to suffer from commitment problems, necessitating third party management. Finally, both maritime and river issues are more likely to result in IGO-led mediation compared with territorial conflicts. This is consistent with the existence of cooperative management norms on these issues (Nemeth et al., 2006; Ostrom, 1990).

Is dispute resolution only selectively effective?

Given evidence of strategic behavior in first-stage forum selection, this discussion continues with the analysis of possible endogeneity bias in settlement compliance. Table 2 presents settlement compliance observations by management strategy. Non-IGO/partial decision control fora have the weakest success rate, 72%. IGOs delegated complete decision control perform the best, with more than 90% of all settlements fulfilled. Results from the logistic regression of settlement compliance are given in Table 3. The statistical analysis supports extant explanations. Notably, states are more likely to comply when third parties have binding authority and negotiations are transparent. Agreements reached through IGOs and between democracies are more likely to be honored. In contrast, I do not find that non-institutional fora or concessions significantly impact compliance. In general, management style and third-party characteristics more substantively affect compliance than distributional issues (Gent and Shannon, 2011a).

Table 2.

Forum selection and mutual compliance with settlement agreements

	Did both parties comply?
Forum	No	Yes
Bilateral	24.40%	75.60%
Non-IGO, partial decision control	27.80%	72.20%
IGO, partial decision control	13.60%	86.40%
Non-IGO, complete decision control	20%	80%
IGO, complete decision control	7.40%	92.60%

Table 3.

Compliance stage—logistic regression of mutual compliance with forum selection effects for territorial, maritime, and river disputes, 1816–2001

Analysis			Previous results
Variable	Coefficient	Standard error	Mitchell and Hensel (2007)	Gent and Shannon (2010)
Forum characteristics
Decision control	1.209**	0.615	Agree	Agree
Transparency: IGO	0.748*	0.431	Disagree: −/null	Partial: +/null
Transparency: non-IGO	−0.271	0.232	Partial: −/significant	Not tested
Transparency: joint democracy	1.573***	0.383	Disagree: −/null	Partial: +/null
Bias: strong-side concessions	−0.021	0.257	Agree	Not tested
Bias × decision control	−1.423*	0.846	Not tested	Not tested
Relative power	2.179**	0.917	Not tested	Agree
Issue salience	0.004	0.097	Disagree: −/significant	Agree
Joint MTOPS	0.309***	0.075	Agree	Not tested
Alliance portfolio	−1.423***	0.411	Partial: −/null	Partial: −/null
Maritime issue	1.699***	0.601	Not tested	Not tested
River issue	3.367***	1.088	Not tested	Not tested
IMR: other, partial	0.442***	0.153
IMR: IGO, partial	−0.257***	0.096
IMR: other, complete	0.231***	0.076
IMR: IGO, complete	0.027	0.047
Constant	4.392	2.698
	N = 848; pseudo R² = 0.1064

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Robust standard errors adjusted for 247 dyadic claims. Mitchell and Hensel (2007) dependent variable = settlement compliance; Gent and Shannon (2010) dependent variable = claim termination.

Because the second-stage model analyzes existing explanations about management approaches and compliance, I summarize in Table 3 where results agree with previous findings so that the discussion may focus on the selection parameters. Similar to the interpretation of the selection parameter, ρ, in Heckman (1979) models, IMRs convey correlations between unobserved factors in the selection stage and settlement compliance. When these values are statistically significant, forum selection affects compliance behavior. For instance, the factors that lead states to select non-institutional/partial decision control management (IMR: other, partial) also increase compliance. This result is surprising because state-led mediation performs poorly compared with other management approaches (Table 2). One explanation why non-institutional mediation precedes compliance is that disputants are lured by stronger mediators and enforcement. Table 1 shows that issue salience and recent fatal violence are more likely to result in non-IGO, partial control options, like mediation. Melin (2013) observes that these conflicts receive more mediation offers. Additionally, mediation appears to be well designed for short-term agreements (Beardsley, 2008). These partial settlements may be easier to fulfill than comprehensive or long-term goals.

The negative coefficient on the selection parameter for IGO-led mediation (IMR: IGO, Partial) indicates an adverse effect on compliance. The factors that lead to IGO mediation—power asymmetries, salience, IGO memberships, recent management failures and fatal violence—make these fora less effective. This result raises questions about IGOs’ reputational importance. A cynic would note the limitation of reputational penalties and find that less compliant states choose institutions precisely because their enforcement is weak (Downs et al., 1996). Such an explanation would account for the higher incidence of IGO mediation among asymmetric dyads: stronger challengers may renege on settlements reached through IGOs with impunity.

Alternatively, IGO management may frequently be used as a conflict stopgap. States that select IGO mediation face more commitment problems than states that select other management approaches. These disputes are more likely to have recent management failures as well as high issue salience and recent fatal violence. Mediators are unlikely to risk their reputations on attempts to resolve conflicts between belligerents with such strong patterns of mistrust and contentiousness (Bercovitch and Gartner, 2006). Access to peaceful dispute resolution institutions gives disputants a tempting reprieve when few other alternatives are available. In other words, the selection of IGO mediation is less strategic than state-led mediation.

The positive, statistically significant IMR on the third forum option indicates that factors contributing to the selection of ad hoc arbitration (IMR: other, complete) increase belligerents’ propensity to comply with arbitral awards. In contrast, there is no selection effect between the use of institutional, complete decision control fora (IMR: IGO, complete) and compliance. These last two results reveal subtle differences in legal dispute resolution approaches. Although the positive result for selection effects associated with ad hoc arbitration tempers previous enthusiasm for legal approaches, this result is less surprising with better understanding of the arbitration process. States agreeing to arbitration first negotiate the compromis, which includes the composition of arbiters, applicable rule(s) of law and whether any non-legal principles of decision-making are permitted (Malintoppi, 2006). Therefore, states that select ad hoc arbitration are more likely to already recognize areas of agreement. They are also more likely to be experienced with conflict management because the composition of a compromis requires significant legal and negotiation skill. Empirical results support this contention: states that select non-institutional, complete decision control fora have more recent management successes, which have a cumulative effect that eventually leads to more formal methods (Bercovitch and Gartner, 2006; Böhmelt, 2013).

In contrast, adjudicating states may be less experienced and less certain about applicable law. In the North Sea continental shelf delimitation case, for example, the claimants required ICJ adjudication to identify the relevant legal principles (Fischer, 1982). The ICJ also has a special fund to help states hire legal and technical experts so that less-advantaged and less-experienced states can participate (Merrills, 2005). Where international courts provide independent keys to settlement (Simmons, 1999), ad hoc arbitration reflects more internalized legal comprehension. This conclusion also highlights arbitration as an alternative dispute resolution method to adjudication. Arbitration facilitates extra-judicial compromise that avoids the adversarial nature of courts (Charney, 1998). Arbitration is more similar to mediation than it is often treated by conflict management scholars. Their consensus-building foundations—mediation’s collaborative methods and arbitration’s compromis coordination—may explain their similar strategic implementation and effects on compliance.

Conclusion

Despite competing interests and incentives to renege, states comply with a remarkable number of dispute resolution settlements. To respond to the possibility that states only pursue conflict management approaches with which they may readily comply, I present a model that makes three observations about dispute resolution. First, management forum design features—decision control, transparency and distributional bias—are consequential to settlement compliance and resolution (Gent and Shannon, 2010, 2011a; Mitchell and Hensel, 2007). Second, peaceful dispute resolution is a voluntary process for which states face a large range of options. Finally, states are compliance-minded when they strategically select management fora. States carefully balance beliefs about issue division and reputational consequences against options outside third-party fora (Busch, 2007; Fang, 2010; Gent and Shannon, 2011b; Wiegand and Powell, 2011). China’s preference for bilateral negotiations over arbitration in the management of its South China Sea claims illustrates these incentives.

Empirical analysis of forum selection supports compliance-motivated strategic forum selection. I also find evidence for the skeptical perspective in two cases, state-led mediation and ad hoc arbitration, although neither result requires dramatic revisions to conflict management scholarship. Mediation is already understood as a consensus-building, bargaining process (Beber, 2012; Melin et al., 2013). Competing interests among the parties pose significant barriers to success. Therefore, it is unsurprising that rivals cooperate with other states that can credibly facilitate short-term goals and implementation (Beardsley, 2010).

Legal scholars similarly describe ad hoc arbitration as an approach that requires a great deal of pre-negotiation cooperation between adversaries (Malintoppi, 2006). In contrast, adversarial adjudication is independent from the factors that lead states to submit. The implication for conflict management scholarship is not to drive a wedge between arbitration as an epiphenomenon and international courts as legitimate. Rather, conflict management research should investigate the similarities and differences across management approaches—adjudication and alternative dispute resolution, which includes mediation and arbitration. The similarities found here between arbitration and mediation in particular suggest that a dichotomous separation of binding and non-binding approaches segments our study of conflict management. The multidimensional conception of forum design introduced here provides one method of continuing these comparisons.

At the same time, this project introduces a new puzzle for conflict management research. Mitchell and Hensel (2007) find that IGOs are especially effective in enforcing settlements because they actively and passively encourage peaceful resolution. Yet I find that compliance with settlements reached through IGO mediation is negatively correlated with the factors that lead to their use. One possible explanation is that IGOs are more likely to mediate tougher cases than are submitted to other approaches. Relatedly, these tough cases might attract fewer mediation offers from other states. With institutional abundance and mediator scarcity, rivals turn to IGOs as a stop gap. Where existing research has explored these two factors—mediator supply and disputants’ short-term interest in mediation—with respect to state intermediaries (Beardsley, 2010; Beber, 2012; Crescenzi et al., 2011), this research indicates that another fruitful area of exploration focuses on IGO effectiveness.

Footnotes

Acknowledgements

The author thanks Kelly Kadera, Douglas Dion, Frederick Boehmke, Thania Sanchez, Paul Poast, Shawn Ramirez, Peter Schraeder, and two anonymous reviewers for their comments and assistance. She also thanks Molly Melin and Andy Owsiak for their dedicated organization to this collection.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Notes

References

Allee

Huth

(2006) Legitimizing dispute settlement: International legal rulings as domestic political cover. American Political Science Review 100(2): 219–234.

Barbieri

(2002) The Liberal Illusion: Does Trade Promote Peace?Ann Arbor, MI: University of Michigan Press.

Beardsley

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

Beardsley

(2009) Intervention without leverage: Explaining the prevalence of weak mediators. International Interactions 35(3): 272–297.

Beardsley

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

Beardsley

(2014) Third-party conflict management and the willingness to make concessions. Journal of Conflict Resolution 58(2): 363–392.

Beardsley

Quinn

Biswas

Wilkenfeld

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

Beber

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

Bercovitch

Gartner

(2006) Is there method in the madness of mediation? Some lessons for mediators from quantitative studies of mediation. International Interactions: Empirical and Theoretical Research in International Relations 32(4): 329.

10.

Bercovitch

Schneider

(2000) Who mediates? The political economy of international conflict management. Journal of Peace Research 37(2): 145–165.

11.

Boehmer

Gartzke

Nordstrom

(2004) Do intergovernmental organizations promote peace? World Politics 57(1): 1–38.

12.

Böhmelt

(2013) Failing to succeed? The cumulative impact of international mediation revisited. Conflict Management and Peace Science 30(3): 199–219.

13.

Brams

Taylor

(2000) The Win–Win Solution: Guaranteeing Fair Shares to Everybody. New York: Norton.

14.

Busch

(2007) Overlapping institutions, forum shopping, and dispute settlement in international trade. International Organization 61(04): 735–761.

15.

Chapman

Wolford

(2010) International oraganizations, strategy, and crisis bargaining. Journal of Politics 72(1): 227–242.

16.

Charney

(1998) The impact on the international legal system of the growth of international courts and tribunals. New York University Journal of International Law and Politics 31: 697.

17.

Chayes

(1993) On compliance. International Organization 47(2): 175–205.

18.

Ching

(2014) International arbitration the way to go for Asia’s maritime disputes. South China Morning Post. Available at: http://www.scmp.com/comment/article/1505751/international-arbitration-way-go-asias-maritime-dispute

19.

Cohen

(2013) Negotiation—not war—can resolve China’s conflict with Japan. The Atlantic, September. Available at: http://www.theatlantic.com/china/archive/2013/08/negotiation-not-war-can-resolve-chinas-conflict-with-japan/278307/

20.

Crescenzi

MJC

(2007) Reputation and interstate conflict. American Journal of Political Science 51(2): 382–396.

21.

Crescenzi

MJC

Kadera

Mitchell

Thyne

(2011) A supply side theory of mediation. International Studies Quarterly 55(4): 1069–1094.

22.

Dixon

(1993) Democracy and the management of international conflict. The Journal of Conflict Resolution 37(1): 42–68.

23.

Downs

Rocke

Barsoom

(1996) Is the good news about compliance good news about cooperation? International Organization 50(3): 379–406.

24.

Dubin

McFadden

(1984) An econometric analysis of residential electric appliance holdings and consumption. Econometrica 52(2): 345–362.

25.

Ellis

Mitchell

Prins

(2010) How democracies keep the peace: Contextual factors that influence conflict management strategies. Foreign Policy Analysis 6(4): 373–398.

26.

Fang

(2010) The strategic use of international institutions in dispute settlement. Quarterly Journal of Political Science 5(2): 107–131.

27.

Favretto

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

28.

Fearon

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

29.

Finel

Lord

(2002) The surprising logic of transparency. International Studies Quarterly 43(2): 325–339.

30.

Fischer

(1982) Decisions to use the international court of justice: Four recent cases. International Studies Quarterly 26(2): 251–277.

31.

Gelpi

(1999) Alliances as instruments of intra-allied control. In: Haftendorn

Keohane

Wallander

(eds) Imperfect Unions: Security Institutions Over Time and Space. Oxford: Oxford University Press.

32.

Gent

Shannon

(2010) The effectiveness of international arbitration and adjudication: Getting into a bind. The Journal of Politics 72(2): 366–380.

33.

Gent

Shannon

(2011a) Bias and the effectiveness of third party conflict management mechanisms. Conflict Management and Peace Science 28(2): 124–144.

34.

Gent

Shannon

(2011b) Decision control and the pursuit of binding conflict management: Choosing the ties that bind. Journal of Conflict Resolution 55(5): 710–734.

35.

Gent

Shannon

(2014) Bargaining power and the arbitration and adjudication of territorial claims. Conflict Management and Peace Science 31(3): 303–322.

36.

Gleditsch

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

37.

Greig

(2005) Stepping into the fray: When do mediators mediate? American Journal of Political Science 49(2): 249–266.

38.

Guzman

(2002) A compliance-based theory of international law. California Law Review 90: 1823.

39.

Heckman

(1979) Sample selection bias as a specification error. Econometrica 47(1): 153–161.

40.

Hensel

(2001) Contentious issues and world politics: The management of territorial claims in the Americas, 1816–1992. International Studies Quarterly 45(1): 81–109.

41.

Hensel

(2006) Multilateral Treaties of Pacific Settlement (MTOPS) Data Set. Available at: http://data.icow.org.

42.

Hensel

Mitchell

Sowers

(2006) Conflict management of riparian disputes. Political Geography 25(4): 383–411.

43.

Hensel

Mitchell

Sowers

Thyne

(2008) Bones of contention: Comparing territorial, maritime, and river issues. Journal of Conflict Resolution 52(1): 117–143.

44.

Huth

Croco

Appel

(2011) Does international law promote the peaceful settlement of international disputes? Evidence from the study of territorial conflicts since 1945. American Political Science Review 105(2): 415–436.

45.

Koremenos

(2007) If only half of international agreements have dispute resolution provisions, which half needs explaining? The Journal of Legal Studies 36(1): 189–212.

46.

Kydd

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

47.

Kydd

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

48.

Lee

L-F

(1983) Generalized econometric models with selectivity. Econometrica 51(2): 507–512.

49.

Lohmann

(2003) Why do institutions matter? An audience–cost theory of institutional commitment. Governance 16(1): 95–110.

50.

Lord

(1892) International arbitration. Annals of the American Academy of Political Science 2: 39–55.

51.

Malintoppi

(2006) Methods of dispute resolution in inter-state litigation: When states go to arbitration rather than adjudication. Law and Practice of International Courts and Tribunals 5(1): 133–162.

52.

Marshall

Jaggers

(2002) Polity IV Project: Political Regime Characteristics and Transitions, 1800-2002. College Park, MD: Center for International Development and Conflict Management, University of Maryland. Available at: http://www.cidcm.umd.edu/inscr/polity/index.htm.

53.

Melin

(2013) When states mediate. Penn State Journal of Law and International Affairs 2(1): 78–90.

54.

Melin

(2014) Commitment problems: Understanding variation in the frequency of international conflict management efforts. International Negotiation 19: 222–258.

55.

Melin

Gartner

Bercovitch

(2013) Fear of rejection: The puzzle of unaccepted mediation offers in international conflict. Conflict Management and Peace Science 30(4): 354–368.

56.

Merrills

(2005) International Dispute Settlement. Cambridge: Cambridge University Press.

57.

Mitchell

Hensel

(2007) International institutions and compliance with agreements. American Journal of Political Science 51(4): 721–737.

58.

Mitchell

Kadera

Crescenzi

MJC

(2009) Practicing democratic community norms: Third party conflict management and successful settlement. In: International Conflict Mediation: New Approaches and Findings. New York: Routledge, pp. 243–264.

59.

Nemeth

Mitchell

Nyman

Hensel

(2006) UNCLOS and the management of maritime claims. Chicago, IL.

60.

Ostrom

(1990) Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press.

61.

Paulson

(2004) Compliance with final judgments of the international court of justice since 1987. The American Journal of International Law 98(3): 434–461.

62.

Powell

(1996) Bargaining in the shadow of power. Games and Economic Behavior 15(2): 255–289.

63.

Quinn

Wilkenfeld

Smarick

Asal

(2006) Power play: Mediation in symmetric and asymmetric international crises. International Interactions 32(4): 441–470.

64.

Rauchhaus

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

65.

Sevastopulo

Landingin

(2014) China rejects Philippines case on ‘nine-dash’ line. Financial Times. Available at: http://www.ft.com/cms/s/0/dc3896ce-b7ac-11e3-80ef-00144feabdc0.html#axzz33JmP83zV

66.

Shannon

(2009) Preventing war and providing the peace? Conflict Management and Peace Science 26(2): 144–163.

67.

Signorino

Ritter

(1999) Tau-B or nor Tau-B: Measuring the similarity of foreign policy positions. International Studies Quarterly 43(1): 115–144.

68.

Simmons

(1998) Compliance with international agreements. Annual Review of Political Science 1: 75–93.

69.

Simmons

(1999) See you in “court”? The appeal to quasi-judicial legal processes in the settlement of territorial disputes. In: A Road Map to War: Territorial Dimensions of International Conflict. Nashville, TN: Vanderbilt University Press, pp. 205–237.

70.

Simmons

(2002) Capacity, commitment, and compliance: International institutions and territorial disputes. The Journal of Conflict Resolution 46(6): 829–856.

71.

Singer

(1987) Reconstructing the Correlates of War dataset on material capabilities of states, 1816–1985. International Interactions 14: 115–132.

72.

Singer

Bremer

Stuckey

(1972) Capability distribution, uncertainty, and major power war, 1820–1965. In Peace, War, and Numbers. Beverly Hills, CA: Sage, pp. 19–48.

73.

Skjelsbæk

(1991) The UN Secretary-General and the mediation of international disputes. Journal of Peace Research 28(1): 99–115.

74.

Small

Hsiao

(1985) Multinomial logit specification tests. International Economic Review 26(3): 619–627.

75.

Smith

Stam

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

76.

Strasavage

(2004) Open-door or closed-door? Transparency in domestic an international bargaining. International Organization 58: 667–703.

77.

Tarar

Leventoglu

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

78.

Thibaut

Walker

(1975) Procedural Justice: A Psychological Analysis. Hillsdale, NJ: Lawrence Erlbaum Associates.

79.

Touval

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

80.

Vasquez

Henehan

(2001) Territorial disputes and the probability of war, 1816–1992. Journal of Peace Research 38(2): 123–138.

81.

Wiegand

Powell

(2011) Past experience and methods of territorial dispute resolution. Journal of Conflict Resolution 55(1): 33–59.

82.

Wolf

Kramer

Carius

Dabelko

(2009) Peace in the pipeline. BBC News. Available at: http://news.bbc.co.uk/2/hi/science/nature/7886646.stm