Sage Journals: Discover world-class research

Abstract

A large body of scholarship has analysed how the content of a negotiated settlement might impact the potential for conflict resolution or recurrence, focusing largely on the institutional design of transitional power-sharing formulas. To date, however, variation in procedures for joint decision-making within such institutions has not been investigated, nor has the effect that such choices might have on the stability and progress of settlement implementation. Relevant provisions tend to take one of two approaches: requiring consensus between former belligerents in a power-sharing government, or delegating responsibilities for dispute resolution to alternative and more broadly inclusive commissions. Employing original cross-national data on settlement content, the statistical results lend strong support to the central expectations put forward here – namely, that consensus rules heighten commitment problems between settlement signatories, thus increasing the risk of conflict recurrence, while delegation appears to significantly mitigate such risks, all else being equal. The findings therefore have real-world implications for the toolkit of international peace brokers.

Keywords

Armed conflict civil war negotiated settlement post-conflict transition bargaining consensus dispute resolution delegation

Recognising that most contemporary rebellions are a recurrence of conflicts from the past, the emerging academic consensus views bargained solutions as relatively unstable outcomes of conflict, compared to victory for one side (Collier et al., 2008; Hartzell et al., 2001; Quinn et al., 2007; Walter, 2015). However, while some countries have found it difficult to break out of the conflict trap, others have successfully navigated the tenuous transition period after a settlement is signed, emerging to more peaceful forms of political competition between former belligerents while avoiding the recurrence of conflict. What explains this variation?

Answers to this question rest on the commitment problem facing belligerent parties in a context of mutual suspicions – even if governments and rebels can reach agreement on terms that both would prefer to war, neither can credibly commit to resist incentives to renege should advantages shift during implementation (Bekoe, 2005; Werner, 1999). Ongoing debates focus on factors that might help to maintain parties’ commitment to peace, even as progress on demobilisation and disarmament increases one side’s vulnerability and, thus, fear of being betrayed by an insincere rival. Common explanations underline the need for outside guarantees from international actors (Collier et al., 2008; Fortna, 2008; Matanock, 2020; Quinn et al., 2007; Walter, 2002), for example, or else consider whether power-sharing formulas can be designed to allay security fears by signalling a mutual willingness to bear costs in the pursuit of peace (Hartzell and Hoddie, 2007; Jarstad and Nilsson, 2008; Johnson, 2021; Mattes and Savun, 2009). As yet, however, little research exists into variation in joint decision-making procedures within transitional institutions, and we lack adequate theories about how such choices might impact actors’ ability to remain committed to implementation.

This study speaks to this gap by analysing two empirically common approaches to stipulating how decisions will be reached between former belligerents after a settlement is signed: (a) the requirement of consensus in a power-sharing government, or (b) the delegation of dispute resolution to alternative, more broadly inclusive commissions, such as those responsible for overseeing implementation. Procedures for joint decision-making during a transition are important, since it is highly unlikely that the settlement comprises agreement on all conflict-related issues, while the implementation process is itself likely to reveal unforeseen roadblocks and incompatibilities. Consensus may be a desirable goal for peace brokers attempting to encourage inclusivity, cooperation and compromise among actors who have recently been at war. However, requiring mutual agreement on every decision increases the number of veto issues during a transition, heightening tendencies towards gridlock and extremism and ultimately incentivising actors to resist any further moves that would make them more vulnerable. In contrast, delegating the resolution of disputes to alternative sites of contention may help to de-escalate conflicts and allow implementation to remain on track despite an imperfect alignment of rival parties’ preferences. The present research agenda therefore speaks directly to existing academic debates while offering a new theory of conflict recurrence, and it contributes original, cross-national data on provisional variation across all negotiated settlements signed since 1975 (N = 160).

The paper proceeds as follows. The next section provides an overview of existing arguments about factors affecting conflict recurrence risks, focusing especially on debates about the terms of a settlement. The third section elaborates my own argument about the role of different procedural approaches to joint decision-making during the transition, generating two testable hypotheses. The method of data collection and model specification is described in the fourth section, while the fifth section presents the results of the binomial logistic regression analysis of conflict recurrence. The findings lend strong support to the hypotheses laid out in the third section – all else being equal, settlements stipulating consensus decisions in political power-sharing institutions exhibit a significantly higher rate of recurrence, while this rate is significantly lower where delegated mechanisms of dispute resolution are present. As the significance of the positive result for delegation is greater than that for any other variable commonly argued to impact the stability of a post-conflict transition, and with the data revealing no evidence for selection effects that might bias these results, the findings lend strong support to the notion that the way transitional institutions are designed to function internally is an important – yet previously overlooked – component of the institutional formula.

Existing research

A valid explanation of conflict recurrence means understanding why parties would choose to bear the costs of a return to war, despite having reached an agreement recognising a mutual preference for peace. Bargaining theory holds that settlements are prone to failure due to the time inconsistency of signatories’ preferences in a context of heightened suspicions. According to Suzanne Werner (1999):

Once a settlement is reached, one or both sides must take actions to honor their side of the bargain. . .[including] withdrawal or demobilization of forces, alteration of policies, exchange of territory, and payment of reparations. Temptation to defect is high during this stage because actually honoring the agreement requires substantial and costly action. (p.916)

Thus, signatories are constantly making strategic calculations over the series of incremental steps required to implement a settlement, weighing their willingness to bear further costs against fears that doing so may make them vulnerable to betrayal by an insincere or opportunistic rival (Bekoe, 2005; Matanock, 2020). Relevant scholarship to the current research question considers factors that, despite the ever-present option to renege, might help to maintain parties’ commitments until such time that the ‘hurdle of implementation’ has been crossed and ‘honouring the agreement requires only maintenance of the status quo’ (Werner, 1999: 916). While some have focused on variables related to the opportunity context of the transition, such as the strength of state institutions or the presence of third-party enforcement (Bekoe, 2005; Collier et al., 2008; Fortna, 2008; Quinn et al., 2007; Walter, 2002, 2015), the focus here is on academic debates about the terms agreed in a peace bargain.

Much of this literature is dominated by questions about the role of power sharing, and particularly whether an agreement to share power can provide a sufficiently costly signal of belligerents’ commitments to peace. Some conclude that power sharing does nothing to mitigate the commitment problem (Walter, 2002), rather that the transitory and opportunistic nature of most power-sharing arrangements make them unstable outcomes of conflict, particularly if there is an expiry date to actors’ access to the spoils of peace (Johnson, 2021; Mukherjee, 2006; Spears, 2000). However, others have argued that incumbents who sign away their monopoly on power, which they would otherwise prefer to keep for themselves, are sending a clear message that they have no intention to continue fighting for victory, thus mitigating rivals’ fears about the potential risks of demobilisation (Hartzell and Hoddie, 2007; Jarstad and Nilsson, 2008; Mattes and Savun, 2009).

Within this school of thought, questions remain about the type of power-sharing formula that sends the strongest signal. For Hartzell and Hoddie (2007), the more power sharing the better, with more comprehensive bargains providing additional institutional checks against implementation failures elsewhere. Among the four subtypes of power sharing commonly used in international relations (IR), agreements involving military or territorial power sharing have most consistently been associated with a reduced risk of conflict recurrence (Hartzell et al., 2001; Hoddie and Hartzell, 2003; Jarstad and Nilsson, 2008; Wantchekon, 2000), as compared to seemingly less effective political or economic solutions – perhaps because the former tend to promise a longer-term payoff in exchange for disarmament (Johnson, 2021).

Various strands of scholarship point to other terms in a settlement that might affect signatories’ ability to remain committed to implementation, beyond power sharing. Although potentially unpopular with voters, blanket amnesties may be necessary to convince ex-combatants that they will not face retribution after disarming (Dancy, 2018), while autonomous zones of security may make it easier for groups to demobilise without fear of a surprise attack (Mattes and Savun, 2009). A recent area of research suggests that independent commissions set up to oversee implementation create space for non-belligerents in the transition process (Fontana et al., 2021), and where international actors are tasked with a direct role in such commissions, their provision of material and technical resources may facilitate disarmament by creating alternative futures for ex-combatants, such as aiding rebel-to-political party transformations, cash-for-guns programmes, job training and security sector reform (Matanock, 2020; Walsh, 2020; Walsh and Doyle, 2018).

In summary, existing theories of conflict recurrence are coordinated by a similar logic: maintaining signatory parties’ commitments to a settlement is difficult because implementation threatens to shift the balance of power on which the bargain was struck, making one side more vulnerable to the threat of attack from a rival who might seek to capitalise on its newfound advantage. While scholars point to the different formulas of a bargain that might help to mitigate such risks, what is missing from existing accounts is attention to the formal rules for joint decision-making that exist within transitional institutions. Certain procedures may affect the ability of signatories to continue to reach agreement on key issues after a settlement is signed – decisions that are necessary to keep implementation on track – thus heightening commitment problems in a way that is independent from any changes to parties’ relative capabilities.

Theorising the role of transitional decision-making procedures

The notion that former belligerents should be guaranteed inclusion in transitional governing institutions is an appealing one to international peace brokers. Such guarantees may be necessary to convince parties to sign on to a settlement in the first place, particularly where political marginalisation is a core motivation of conflict, and yet the way that these institutions are intended to function after the ink is dry seems an afterthought, at least in the academic literature. The potential for bargaining failures remains high during this period because, in an environment of heightened fear and distrust, former belligerents must continue to reach decisions jointly. Even the most comprehensive settlements are unlikely to include agreement on every issue underlying the conflict incompatibility or affecting the nature of the future polity, while new issues are likely to emerge during implementation due to ‘constructive ambiguity’ in the terms that have been agreed (Walsh, 2020), as well as delaying tactics – or ‘foot dragging’ (Arnault, 2001) – which make it difficult to determine whether violations have occurred. The key, then, is to identify the procedures through which such decisions will be reached, and to consider what impact this might have on incentives for asymmetric reversion.

The little research that exists on joint decision-making rules among former belligerents focuses on the mutual veto – according to Lijphart (1990: 495), the ‘ultimate weapon’ necessary to guarantee that groups’ security and vital interests will be protected, especially as a form of insurance against failures in other parts of a power-sharing formula. However, the relative infrequency of veto rights in post-conflict settings, as well as variation in the way the veto is designed across countries, has restricted cross-national research into its potential effect on conflict tendencies in countries emerging from civil war (McCulloch and Zdeb, 2022). As a result, the current state of knowledge is largely based on qualitative studies of a few cases, especially Northern Ireland, Bosnia and Burundi (Bieber, 2005; McCulloch, 2018; McCulloch & Zdeb, 2022; McCulloch and Vandeginste, 2019; McEvoy, 2013). While these countries made it through the rocky war-to-democracy transition and enshrined the veto into law, joint governance is much more often a transitional measure intended to maintain buy-in among settlement signatories until elections can be held, such as in executive coalitions or unity governments with predetermined formulas for inclusion (Johnson, 2021). Yet, no scholarship exists to offer theories about the procedural variation for reaching decisions within these transitional institutions. While many negotiated settlements do not address this question at all – a de facto recourse to majority rule – those that do appear to take one of two approaches: (a) a requirement of consensus in political power-sharing institutions, or (b) delegation of responsibilities for resolving disputes to alternative sites of inclusion.

Consensus rules

As part of the liberal peace-making toolkit, the normative basis for consensus is that ‘no single group can decide important matters without the consent of the other’ (Schneckener, 2002: 203), with the hope that this will foster a sense of mutual trust and compromise among former belligerents. Speaking to a consociational logic of power sharing (Lijphart, 1990), therefore, the consensus requirement is one strategy to embed the mutual veto into transitional institutions for joint governance – with decisions requiring agreement from all included parties, any of them may block rulings perceived as a threat (Roeder and Rothchild, 2005). Because the number and nature of decisions that will require agreement during a transition is unknown at the time the settlement is signed, the consensus requirement is appealing as a blanket guarantee of future protection for demobilising groups, and, in certain contexts, it may even be a necessary concession to convince belligerents to sign on to the bargain at all (Lijphart, 1990; McEvoy, 2013). However, a long tradition of political science research has shown that varied approaches to designing the veto can have varied effects,¹ from the intended moderating influence in many cases (Gates et al., 2016), to gridlock, extremism and political crisis in others, such as Cyprus (Ram and Strøm, 2014; Yakinthou, 2009).

There is general agreement that the veto is most effective at fostering compromise when it exists as a preventive recourse of last resort – a potential weapon that is not exercised unless all other avenues have been exhausted (Lijphart, 1990). Among those that have attempted to disaggregate veto rules according to their potential for over-use, the scholarly consensus warns against taking a ‘permissive approach’ to the number of decisions that groups have the power to block by increasing the number of ‘veto issues’ (McCulloch 2018: 736; McCulloch and Zdeb, 2022; McEvoy, 2013: 260; Ram and Strøm, 2014). In Schneckener’s (2002: 224) words, giving the parties represented in power-sharing institutions a ‘direct and absolute veto on each and every political issue. . .should be avoided’ if the goal is to prevent gridlock and keep implementation on track.

Elsewhere, research into coalition governments reveals that parties are more likely to activate the veto where there is greater ideological distance in their preferences (Tsebelis, 2002), as is the case where groups are emerging from violent conflict, and where their agreement is necessary for policy change (Oktay, 2022), as is the case under consensus rules. In transitional power-sharing institutions, therefore, failing to delimit the scope of policy areas to which the veto can be applied encourages political brinkmanship and spoiling behaviour, which – in a context of heightened suspicions – will likely be perceived by rivals as a signal of non-cooperation, promoting the emergence of tit-for-tat ‘blocking trajectories’ (McCulloch and Zdeb, 2022: 429). In this way, consensus rules are expected to heighten strategic barriers to joint decision-making, create new roadblocks to implementation and reduce parties’ willingness to make further moves towards demobilisation.

H₁: Provisions requiring consensus in political power-sharing institutions increase the likelihood of conflict recurrence.

Delegated dispute resolution

Some settlements rely on an alternative approach to managing joint decision-making by stipulating that any conflicts arising between signatories be settled outside of the power-sharing government, such as in commissions established for implementation oversight or arbitration. Belligerents are often allocated seats on such commissions; however, they are typically outnumbered by representatives from other parties to the peace process, such as international actors or members of civil society. It is common for the United Nations (UN) peacekeeping or observer mission to act as Chair, with additional seats reserved for bilateral and multilateral actors as dictated by context, such as the African Union in the Central African Republic (CAR), the Economic Community of West African States (ECOWAS) in Liberia or Chad and Qatar in Sudan. In terms of domestic civil society, commission members have comprised church leaders in Colombia and South Sudan, while disputes in Sierra Leone and the Solomon Islands have been referred to a council of chiefs or elders, and in Burundi six seats were reserved for citizens ‘designated for their moral integrity’.² Consistent with an emerging area of scholarship, therefore, the commissions tasked with resolving disputes between belligerents provide opportunities for broader and less confrontational forms of inclusion, beyond traditional power sharing, which has been shown elsewhere to help ‘overcome logjams’ (Walsh, 2020: 2) and ‘work as a critical backstop against conflict recurrence’ (Fontana et al., 2021: 358; Neudorfer and Walsh, 2025).

There are two mechanisms through which delegation might be expected to have a more moderating and constraining effect on signatories’ behaviour, as compared to consensus. The first mechanism rests on a handcuffing logic. Scholars across various fields of comparative politics recognise that, although delegating authority may undermine the ability of political elites to act as utility maximisers, it also acts as a constraint against rivals’ ability to do the same (Berliner, 2014; Dellepiane-Avellaneda, 2012). Thus, allowing disputes that arise within political power-sharing institutions to be settled elsewhere acts as an alternative kind of security guarantee, mitigating fears that an advantaged rival could unilaterally advance their own interests during the transition.

The second mechanism focuses on the accountability of political elites to their own constituents. The important issues at stake during a transition are often hotly politicised and prone to the emergence of hard-liners on either side, making it difficult for leaders to maintain internal cohesion, as well as their own legitimacy. Best and Bapat (2018) argue that it is this internal commitment problem that often causes rebel leaders to abandon a peace process, in order to prevent factionalism from weakening the group’s bargaining power. On the other hand, recent comparative research into coalition governments reveals that parties are more likely to accept decisions that are unpopular with their constituents where a ‘diffusion of responsibilities’ gives elites distance from policy outcomes (Bovens and Schillemans, 2020; Oktay, 2022: 33). Because commissions tasked with dispute resolution resemble ‘oversized coalitions’, it is easier to shift blame for costly policy decisions elsewhere (Oktay, 2022: 38), allowing leaders to de-politicise issues that might otherwise incentivise blocking behaviour, distance themselves from further concessions or compromises that may be unpopular and continue to present as loyal and committed to the cause (Neudorfer and Walsh, 2025).

H₂: Provisions for delegated dispute resolution reduce the risk of conflict recurrence.

In summary, where a settlement does not include any agreement to share power – or else the solution rests on a power-dividing approach, rather than power sharing (Roeder and Rothchild, 2005) – former belligerents are not required to reach joint decisions during the transition. Where the bargain involves political power sharing, however, two alternative approaches to stipulated decision-making are common: consensus rules or delegated dispute resolution. Given the appeal of finding tools for problem-solving, inclusivity and compromise, both approaches seem to have merit, and yet this section has shown that they have the potential to impact conflict outcomes in different ways. Requiring consensus in a political power-sharing government increases the number of veto issues during a transition, and therefore the potential for brinkmanship and gridlock, while delegation is expected to have a more constraining effect on signatories’ behaviour, all else being equal. Delegated dispute resolution may even counteract the strategic barriers imposed by consensus, since parties can attempt to block decisions in the power-sharing government that are then redirected to other sites of contestation, thus modifying the design of the veto from its most disruptive ‘direct’ form to a more preferable ‘delaying’ strategy (Schneckener, 2002: 221). Consistent with the scholarship cited here, therefore, trajectories towards either moderation or extremism are contingent on the design of decision-making procedures, although the present contribution is novel in its application of this logic to transitional institutions, specifically.

Data collection and analytical method

In order to test hypotheses about the effects of transitional decision-making procedures on settlement stability, I first identified all negotiated settlements from a larger universe of peace agreements. To be included in the sample, an agreement must (a) be signed by at least one active conflict dyad included in the Uppsala Conflict Data Program (UCDP) Armed Conflict Dataset (ACD) (Davies et al., 2023),³ (b) be recognised by signatories as a final solution to conflict, with rebels committing to full demobilisation and disarmament and (c) contain agreement on substantive terms of peace, using University of Edinburgh’s Peace Agreements Database (PA-X) (Bell and Badanjak, 2019) category of ‘substantive’ agreements.⁴ These criteria help to improve comparability and internal validity by preventing selection of the most stable and unstable agreements – those that are imposed by a more powerful victor, or ceasefires signifying a mere pause in fighting for peace talks.⁵

The dependent variable is a dichotomous measure of conflict recurrence, coded as ‘1’ if the rebel party re-entered the ACD after signing the settlement, which includes conflicts that continue unabated and those that pause before re-starting, and ‘0’ if not. To account for the possibility that recurrence risks vary across the transition period – for example, due to the timing of post-conflict elections or the end of a peacekeeping mission – alternative measures of recurrence rely on 2- and 5-year cut-offs. These sampling and measurement techniques combine to facilitate analysis of the outcome at both dyadic and settlement levels – firstly, across all government-and-rebel dyads that have signed a settlement (N = 233) and, secondly, across settlements (N = 160) that may have multiple signatories and fail if any of the non-government signatories return to fighting.⁶

There are two primary independent variables in the analysis. Despite marked improvements in the data that exist on conflict and peace agreements in recent years, provisions related to variation in transitional decision-making procedures has thus far been overlooked. I therefore relied on UN Peacemaker to access the full text of each settlement and measured dichotomously the inclusion of any provisions related to either consensus (N = 28, 18% of the sample) or delegated dispute resolution (N = 45, 28% of the sample) – ‘1’ if included, ‘0’ otherwise. As an illustrative example, the comprehensive settlement to the second Congolese civil war in 2002–2003 included provisions for both: decisions in the ‘1+4’ power-sharing government, comprised of incumbent President Joseph Kabila and four vice presidents representing the rebel signatory parties, were to be ‘run on the basis of consensus and the avoidance of conflict’ (Section III, Clause 5), while a Committee for Follow-Up of the Agreement was established ‘to reconcile points of view and assist in resolving any disagreements that may arise between signatories’ (Annex III, Clause 3).⁷ Such cases of overlap are relatively rare, with just 12 of the 160 settlements in the sample (7.5%) including provisions for both. For transparency, the dataset – which is fully compatible with extant conflict data due to inclusion of UCDP and Correlates of War (COW) identifiers – and its codebook are publicly available as supplementary files, and a full list of the location of relevant provisions is provided in the Online Appendix.

Control variables were determined by identifying confounding factors – that is, variables likely to impact both the risk of conflict recurrence and the likelihood that the settlement would include provisions on procedures for joint decision-making. For example, certain kinds of conflicts may be more prone to bargaining failures, and therefore leave more issues to be decided during the transition period; stronger, more stable and more democratic regimes may be more capable of implementing agreed reforms, and also more likely to push forward certain kinds of procedural norms during negotiations; and the involvement of international actors means that both consensus-building measures and the establishment of a peacekeeping mission are likely to be on the agenda. The complete list of control variables, data sources and summary statistics are provided in Table 1.

Table 1.

Independent variables, data sources and summary statistics.

Type of variable	Variable name	Description	Data source	Min	Max	Mean
Primary independent variables	Consensus rules	Decisions in political power-sharing institutions require consensus	UN Peacemaker	0	1	0.18
Primary independent variables	Delegated dispute resolution	Responsibilities for resolving disputes between signatories are delegated to an alternative transitional commission, e.g. for oversight	UN Peacemaker	0	1	0.28
Context-related controls	Civil war	Conflict meets civil war threshold of 1000 deaths (dummy)^a	UCDP ACD	0	1	0.442
	Conflict duration	Total months of conflict to settlement signing (log)	UCDP ACD	0	6.44	4.17
	Territorial conflict	Conflict based on territorial claims (dummy)	UCDP ACD	0	1	0.223
	Income	GDP per capita in constant 2015 US$ (log)	World Development Indicators	5.41	10.16	6.78
	Life expectancy	Total years from birth (log)	World Development Indicators	3.31	4.35	3.97
	Level of democracy	Polity2 score, rescaled 0–1	Polity v5	0	1	0.62
	Political instability	Polity score is −77 or −88 in settlement year (dummy)	Polity v5	0	1	0.42
	Post-Cold War	Settlement signed after 1989 (dummy)		0	1	0.91
Content-related controls	Number of rebel signatories	Count based on settlement text	UN Peacemaker	1	13	2.99
	Territorial power sharing	Settlement includes provisions on decentralisation of power (dummy)	UN Peacemaker	0	1	0.29
	Military power sharing	Settlement includes provisions for rebel integration into security sector hierarchy (dummy)	UN Peacemaker	0	1	0.29
	International oversight commission	Settlement gives international actors responsibility for implementation oversight (dummy)	UN Peacemaker	0	1	0.46
	Peacekeeping	Peacekeeping mission present in post-settlement period (dummy)	UCDP Geo-PKO	0	1	0.31

NOTE: All time-variant context-related measures are lagged 1 year prior to the settlement signing to avoid endogeneity. Any missing values are input from the closest available year within a maximum of 5 years.

Other studies have relied on a count of battle deaths (e.g. Walter, 2015), but because the available data from UCDP censors observations prior to 1990, I follow Fontana et al. (2021) by relying on a civil war dummy.

UN: United Nations; UCDP: Uppsala Conflict Data Program; ACD: Armed Conflict Dataset; GDP gross domestic product; Geo-PKO: Geocoded Peacekeeping Operations.

Dummy variables are coded as ‘1’ if the conflict was fought over territorial claims or met the threshold of a civil war – 1000 battle deaths – prior to the settlement signing, using UCDP’s ACD, while duration is a count of months since the start of conflict to the settlement signing (logged). Variation in the capacity of the state is proxied with income (gross domestic product (GDP), constant USD) per capita and life expectancy at birth, both from the World Bank’s 2023 Development Indicators (WDI). Finally, the Polity score is included (Marshall and Gurr, 2020), as well as a dummy for political instability, coded as ‘1’ if polity indicates the country is undergoing regime transition at the time of the settlement (−77 or −88). To avoid endogeneity, all time variant indicators relating to country context are lagged 1 year prior to the settlement signing. An additional dummy also indicates whether the settlement was signed after 1989, or during the post-Cold War period.

A further set of dummies captures variation across other kinds of provisions, based on a reading of the settlement text and cross-referenced against PA-X summaries wherever possible (Bell and Badanjak, 2019). Specifically, territorial power sharing measures the inclusion of provisions for regional decentralisation of legislative or fiscal powers, military power sharing is the integration of rebel leaders into hierarchies of the state security sector and international oversight is the establishment of a multilateral commission for monitoring implementation.⁸ Outside enforcement is measured by the presence of a peacekeeping mission in the period of measurement on the dependent variable, using UCDP’s Geocoded Peacekeeping Operations (Geo-PKO) dataset (Cil et al., 2020). Finally, to control for heightened uncertainty and instability in multi-dimensional bargaining contexts, a simple count of the number of active rebel signatories is also included.

Empirical results

The full results of the binomial logistic regression analysis of conflict recurrence are reported in Table 2,⁹ with coefficients reported as log odds units and robust standard errors clustered by settlement.¹⁰ The analytical technique tests the impact of the two primary independent variables on rates of recurrence, holding constant various controls related to the bargaining context and settlement content, with alternative 2- and 5-year cut-offs used to measure the outcome. In Models 1–6, the unit of analysis is a government-and-rebel dyad to a negotiated settlement (N = 233). In Models 7–12, all rebel parties having signed the same settlement are collapsed into a single case, reducing the number of observations to 160 settlements, and the outcome is measured according to whether any of the signatories return to war.

Table 2.

Full results of the binomial logistic regression analysis of conflict recurrence (1975-2020).

	Conflict recurrence – dyadic level						Conflict recurrence – settlement level
Variables	2 years(1)	5 years(2)	2 years(3)	5 years(4)	2 years(5)	5 years(6)	2 years(7)	5 years(8)	2 years(9)	5 years(10)	2 years(11)	5 years(12)
Consensus rules	1.609*** (0.505)	1.176** (0.501)	1.120* (0.581)	0.613* (0.610)	2.003** (0.858)	1.475** (0.713)	2.032*** (0.593)	1.607*** (0.582)	2.565*** (0.606)	1.790*** (0.551)	2.771*** (0.733)	1.835*** (0.644)
Delegated dispute resolution	−2.541*** (0.514)	−2.106*** (0.452)	−2.293*** (0.432)	−1.670*** (0.352)	−2.310*** (0.535)	−2.010*** (0.474)	−2.324*** (0.525)	−2.081*** (0.469)	−2.345*** (0.535)	−2.163*** (0.473)	−2.283*** (0.548)	−2.168*** (0.513)
Civil war	0.711* (0.406)	0.776** (0.391)			0.684(0.442)	0.904** (0.413)	0.814* (0.468)	0.943** (0.481)			1.027** (0.508)	1.036** (0.514)
Conflict duration, months (log)	−0.0662(0.127)	−0.191(0.135)			−0.118(0.170)	−0.199(0.156)	−0.0354(0.161)	−0.148(0.174)			−0.0342(0.192)	−0.139(0.179)
Territorial conflict	0.877* (0.461)	0.652(0.460)			0.476(0.502)	0.699(0.524)	0.670(0.463)	0.576(0.467)			0.749(0.548)	0.831(0.543)
GDP per capita (log)	0.239(0.323)	−0.0933(0.296)			0.223(0.283)	−0.0864(0.331)	0.323(0.288)	0.174(0.294)			0.568* (0.309)	0.393(0.322)
Life expectancy (log)	−2.906(2.674)	−1.105(2.147)			−3.429(2.153)	−0.722(2.490)	−3.497** (1.671)	−2.816* (1.659)			−5.259*** (1.940)	−3.921** (1.812)
Polity2 (rescaled 0–1)	1.278* (0.771)	2.055*** (0.748)			1.049(0.866)	1.721** (0.859)	1.157(0.779)	2.017** (0.822)			1.335(0.905)	2.128** (0.916)
Political instability	0.0400(0.394)	0.0442(0.391)			0.462(0.518)	−0.00338(0.449)	0.682(0.453)	0.407(0.459)			0.862(0.561)	0.0911(0.538)
Post-Cold War	−0.778(0.749)	−0.605(0.758)			−1.040(0.757)	−0.901(0.771)	−0.927(0.733)	−0.889(0.751)			−1.191* (0.718)	−1.148(0.746)
Number of signatories			−0.191*** (0.0645)	−0.0453(0.0746)	−0.153** (0.0648)	0.0498(0.0763)			0.180(0.170)	0.256** (0.130)	0.170(0.127)	0.279** (0.131)
Territorial power sharing			−0.399(0.379)	−0.298(0.393)	−0.630(0.473)	−0.540(0.416)			−0.298(0.417)	−0.215(0.403)	−0.481(0.489)	−0.298(0.469)
Military power sharing			−0.507(0.429)	−0.373(0.416)	−0.730(0.469)	−0.497(0.456)			−1.006** (0.477)	−0.576(0.470)	−1.159** (0.529)	−0.494(0.520)
International oversight			0.380(0.415)	0.256(0.414)	0.393(0.451)	0.159(0.462)			0.492(0.430)	0.623(0.423)	0.709(0.485)	0.821* (0.481)
Peacekeeping			0.0478(0.440)	0.384(0.501)	−0.876(0.624)	−0.0844(0.579)			−0.418(0.498)	0.0343(0.527)	−1.200* (0.662)	−0.343(0.609)
Constant	9.941(8.575)	5.381(6.961)	0.973*** (0.301)	0.811*** (0.304)	13.54* (7.019)	4.361(7.844)	11.81** (5.452)	10.33* (5.367)	0.368(0.364)	0.350(0.333)	17.14*** (6.449)	12.86** (5.883)
Observations	226	224	231	229	224	222	159	159	160	160	159	159
Log likelihood	−125.2	−125.8	−130.3	−137.3	−115.5	−121.6	−84.84	−85.60	−86.00	−86.83	−78.44	−81.24
Pseudo R²	0.200	0.169	0.186	0.114	0.256	0.187	0.225	0.198	0.219	0.190	0.284	0.239

NOTE: The dependent variable is conflict recurrence, coded as ‘1’ if belligerent parties return to fighting within 2 or 5 years of signing a settlement. Data used to measure conflict start and end dates is from the Uppsala Conflict Data Program/Peace Research Institute Oslo (Davies et al., 2023). Robust standard errors are in parentheses, clustered by settlement.

p < 0.1; **p < 0.05; ***p < 0.01.

GDP: gross domestic product.

Across all model specifications, the findings lend strong support to the hypotheses laid out in the third section. Consistent with H₁, and holding all other variables constant, settlements that stipulate that decisions in political power-sharing institutions be made by consensus are significantly more prone to conflict recurrence at both dyadic and settlement levels (Table 2). To make the results on the coefficients more intuitive, I use Clarify to generate simulated observations (Tomz et al., 2003). I first set every variable in Model 5 to its mean value to establish a baseline for comparison, with all content-related dummies set to ‘0’, thus simulating an average post-conflict risk environment.¹¹ Under these conditions, the predicted probability of conflict recurrence between a government-and-rebel dyad in the 2 years after signing a settlement is 51.8% – essentially a coin toss. Adding the requirement of consensus decisions between parties’ representatives in traditional sites of political power sharing increases the likelihood of a return to conflict to 85.2%. Even where the risk environment is improved by the addition of guarantees known to improve signatory parties’ commitments – changing the dummies for military and territorial power sharing, international oversight and the presence of peacekeepers from ‘0’ to ‘1’, whereby the overall risk of recurrence drops by half to 25.5% – the predicted probability of dyadic conflict recurrence under consensus rules remains quite high, at 64.5%.

Consistent with H₂, bargains that delegate responsibilities for resolving disputes arising between settlement signatories to a separate commission are negatively associated with conflict recurrence, all else being equal. The effect holds across signatory dyads (Models 1–6) and among all signatories to the same settlement (Models 7–12), reaching the highest level of significance (p < 0.01) in every model (Table 2). While the magnitude of the coefficient is consistently large, I again rely on Clarify to concretely compare settlements that delegate to those that do not. Including provisions that shift decisions on incompatibilities away from political power-sharing governments reduces the likelihood that a government-and-rebel dyad will return to fighting in the 2-year post-settlement period from the baseline of 51.8% to just 11.6%. Even in particularly difficult implementation contexts – that is, emerging from a high-intensity, territorial conflict in an under-developed country, whereby the baseline risk of dyadic conflict recurrence is 81.6% – provisions for delegated dispute resolution reduce this risk by more than half, to 36.5%. Of the 12 settlements that included both types of provisions, just four reverted back to war (33.3%), which suggests that delegation might be an effective strategy to maintain parties’ commitments to implementation despite the strategic barriers imposed by consensus-based power sharing.

Of course, it is possible that certain kinds of provisions and transitional procedures are more likely to be adopted in certain contexts, either as a deliberate recognition of the challenges of resolving certain kinds of conflicts, or else as an indication that belligerents have become so exhausted by fighting that they are willing to accept more comprehensive concessions. In other words, since it is impossible to randomise the treatments, the seeming significance of the effects of consensus rules and dispute resolution procedures might be the result of a selection bias. To test whether any of the context- and content-related variables commonly associated with conflict recurrence are also good predictors of the uptake of certain decision-making procedures, I ran the regression model with the main independent variables alternatively on the left-hand side of the equation. The results of this exercise do not point to any consistent or concerning conclusions. Neither types of provisions are significantly more likely where conflicts are longer or more intense, nor where international actors are present to provide oversight or peacekeeping. Delegated dispute resolution procedures appear much more likely in poorer countries and where conflicts are over territory, but since these variables are commonly argued to make conflicts more prone to recurrence (Collier et al., 2008; Hartzell et al., 2001), any potential bias should tend towards a null finding. This makes the significance of the negative association between delegation and conflict recurrence especially noteworthy. Both types of provision are correlated with military power sharing, yet the two are not highly correlated with each other and the hypotheses supported here suggest that they work in opposing directions to stabilise a post-conflict transition (or not). Thus, there is little to suggest that selection effects are biasing the results.

In addition to the alternative 2- and 5-year time thresholds on the dependent variable, the alternative levels of analysis and the attention to possible selection effects, I conducted a series of additional robustness checks. Modifications to the model specification include restricting the sample to (a) settlements signed after the end of the Cold War (N = 128) or (b) only the most comprehensive settlements (N = 48), as categorised according to the PA-X database (Bell and Badanjak, 2019). I also ran an extended version of the model, which includes a range of additional controls for variables that might independently impact the risk of conflict recurrence, such as population size (logged), mountainous terrain (logged), annual revenues from oil and minerals (% GDP), military expenditures (annual total in constant USD, logged), rates of unemployment and secondary educational enrolment and the provision of a blanket amnesty. The main results are robust to these modifications (see the Online Appendix),¹² lending further confidence to the findings presented in Table 2.

Conclusions

Although settlements are signed in the subset of cases where cooperation among former belligerents is most needed, with neither side proving capable of winning outright, maintaining commitment is difficult as the implementation process jeopardises the fragile balance of power on which the bargain rests. The findings presented here demonstrate that agreed procedures for joint decision-making are important to determining whether rivals can continue to navigate the thorny transition towards a new, more cooperative status quo, or else be incentivised to resist moves towards demobilisation that would weaken their bargaining power.

The study makes several important contributions to the literature on conflict recurrence. Firstly, it provides original and transparent data on settlement content that captures variation in the design of transitional decision-making procedures – to my knowledge, for the first time. Secondly, the findings to emerge from an analysis of this data support a new theory about how this variation impacts the commitment problem facing settlement signatories. Past debates have considered different approaches to engineering transitional institutions to mitigate security fears and achieve buy-in, yet academic conclusions about the effects of power sharing have been mixed, and existing scholarship has largely focused on differences across institutional formulas, to the neglect of the internal procedures which determine how these institutions operate. For peace brokers attempting to promote cooperation among actors who have recently been at war, the goal of consensus-building seems to have intrinsic value. Yet, requiring consensus for all decisions in a power-sharing government may encourage rivals to activate their veto, resulting in uncooperative blocking trajectories and, ultimately, a return to conflict. It therefore seems time to revisit the normative commitment to consensus as part of the liberal peace-making toolkit.

In contrast, delegated mechanisms for resolving disputes appear to diffuse and de-politicise decisions that might otherwise incite blocking behaviour, making it possible for elites from rival groups to remain in their power-sharing seats despite new incompatibilities, rather than pre-emptively defecting back to war. This corroborates emerging research highlighting the importance of alternative sites of inclusion in contributing to peace, beyond traditional power-sharing institutions, yet it makes an explicit link between the tasks assigned to these commissions and the ability of signatories to remain committed to implementation.

This analysis is merely a first-cut exploration of a new puzzle. While the results point to robust cross-national patterns, it is not possible to reach definitive conclusions about causation without further data collection, specifically zooming in on the implementation and usage of different procedures across institutional arenas of contestation during a post-conflict transition. However, the empirical findings presented here suggest that this a worthy area of future research.

Supplemental Material

sj-docx-1-ips-10.1177_01925121251319756 – Supplemental material for Stabilising transitions from conflict: The importance of transitional decision-making procedures

Supplemental material, sj-docx-1-ips-10.1177_01925121251319756 for Stabilising transitions from conflict: The importance of transitional decision-making procedures by Chelsea Johnson in International Political Science Review

Footnotes

Data accessibility

The dataset used to produce these findings, which is fully compatible with extant data on armed conflict, is published as supplementary material to the full text of the article.

Declaration of conflicting interests

The author has no conflicts of interest to declare.

Funding

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

ORCID iD

Chelsea Johnson

Supplemental material

Supplemental material for this article is available online.

Notes

Author biography

Chelsea Johnson is a Lecturer in IR at the University of Liverpool. Her research focuses on power sharing, post-conflict transitions and processes of bargaining for peace. Her previous work has been published in the Journal of Peace Research, Perspectives on Politics and Democratization.

References

Arnault

Jean

(2001) Good Agreement? Bad Agreement? An Implementation Perspective. Center for International Studies, Princeton University, Princeton, NJ.

Bekoe

Dorina A

(2005) Mutual Vulnerability and the Implementation of Peace Agreements: Examples from Mozambique, Angola, and Liberia. International Journal of Peace Studies 10(2): 43–68.

Bell

Christine

Badanjak

Sanja

(2019) Introducing PA-X: A new peace agreement database and dataset. Journal of Peace Research 56(3): 452–466.

Berliner

Daniel

(2014) The political origins of transparency. The Journal of Politics 76(2): 479–491.

Best

Rebecca H

Bapat

Navin A.

(2018) Bargaining with Insurgencies in the Shadow of Infighting. Journal of Global Security Studies 3(1): 23–37.

Bieber

Florian

(2005) Power Sharing after Yugoslavia: Functionality and dysfunctionality of power-sharing institutions in post-war Bosnia, Macedonia, and Kosovo. In Noel

Sid

(ed.) From Power Sharing to Democracy: Post- Confl ict Institutions in Ethnically Divided Societies. Montreal: McGill-Queens University Press, 85–103.

Bovens

Mark

Schillemans

Thomas

(2020) Non-Majoritarian Institutions and Representation. In Rohrschneider

Robert

Thomassen

Jacques

(eds) The Oxford Handbook of Political Representation in Liberal Democracies. New York, NY: Oxford University Press, 511–523.

Cil

Deniz

Fjelde

Hanne

Hultman

Lisa

Nilsson

Desirée

(2020) Mapping Blue Helmets: Introducing the geocoded peacekeeping operations (Geo-PKO) dataset. Journal of Peace Research 57(2): 360–370.

Collier

Paul

Hoeffler

Anke

Soderbom

Mans

(2008) Post-Conflict Risks. Journal of Peace Research 45(4): 461–478.

10.

Cunningham

David

(2011) Barriers to Peace in Civil War. Cambridge: Cambridge University Press.

11.

Dancy

Geoff

(2018) Deals with the Devil? Conflict amnesties, civil war, and sustainable peace. International Organization 72(2): 387–421.

12.

Davies

Shawn

Pettersson

Therese

Öberg

Magnus

(2023) Organized Violence 1989-2022 and the Return of Conflicts Between States? Journal of Peace Research 60(4): 691–708.

13.

Dellepiane-Avellaneda

Sebastian

(2012) Gordon Unbound: The heresthetic of central bank independence in Britain. British Journal of Political Science 43(2): 263–293.

14.

Fontana

Giuditta

Kartsonaki

Argyro

Neudorfer

Natascha S.

Walsh

Dawn

Wolff

Stefan

Yakinthou

Christalla

(2021) The Dataset of Political Agreements in Internal Conflicts (PAIC). Conflict Management and Peace Science 38(3): 338–364.

15.

Fortna

Virginia Page

(2008) Does Peacekeeping Work? Shaping Belligerents’ Choices after Civil War. Princeton, NJ: Princeton University Press.

16.

Gates

Scott

Graham

Benjamin AT

Lupu

Yonatan

Strand

Håvard

Strøm

Kaare W

(2016) Power Sharing, Protection, and Peace. The Journal of Politics 78(2): 512–526.

17.

Hartzell

Catherine A

Hoddie

Matthew

(2007) Crafting Peace: Power-Sharing Institutions and the Negotiated Settlement of Civil Wars. University Park, PA: Pennsylvania State University Press.

18.

Hartzell

Caroline

Hoddie

Matthew

Rothchild

Donald

(2001) Stabilizing the Peace After Civil War: An investigation of some key variables. International Organization 55(1): 183–208.

19.

Hoddie

Matthew

Hartzell

Caroline

(2003) Civil War Settlements and the Implementation of Military Power-Sharing Arrangements. Journal of Peace Research 40(3): 303–320.

20.

Jarstad

Anna K

Nilsson

Desirée

(2008) From Words to Deeds: The implementation of power-sharing pacts in peace accords. Conflict Management and Peace Science 25(3): 206–223.

21.

Johnson

Chelsea

(2021) Power-Sharing, Conflict Resolution, and the Logic of Pre-Emptive Defection. Journal of Peace Research 58(4): 734–748.

22.

Lijphart

Arend

(1990) The Power-Sharing Approach. In Montville

Joseph

(ed.) Conflict and Peacemaking in Multiethnic Societies. Lexington, MA: Lexington Books, 491–509.

23.

Marshall

Monty G

Gurr

Ted Robert

(2020) Polity5: Political Regime Characteristics and Transitions, 1800-2018. Polity5 dataset version 2018, Center for Systemic Peace. Available at: www.systemicpeace.org/inscrdata.html (accessed 24 May 2023).

24.

Matanock

Aila M

(2020) How International Actors Help Enforce Domestic Deals. Annual Review of Political Science 23: 357–383.

25.

Mattes

Michaela

Savun

Burcu

(2009) Fostering Peace After Civil War: Commitment problems and agreement design. International Studies Quarterly 53(3): 737–759.

26.

McCulloch

Allison

(2018) The Use and Abuse of Veto Rights in Power-Sharing Systems: Northern Ireland’s petition of concern in comparative perspective. Government and Opposition 53(4): 735–756.

27.

McCulloch

Allison

Vandeginste

Stef

(2019) Veto Power and Power-Sharing: Insights from Burundi (2000–2018). Democratization 26(7): 1176–1193.

28.

McCulloch

Allison

Zdeb

Aleksandra

(2022) Veto Rights and Vital Interests: Formal and informal veto rules for minority representation in deeply divided societies. Representation 58(3): 427–442.

29.

McEvoy

Joanne

(2013) We Forbid! The mutual veto and power-sharing democracy. In McEvoy

Joanne

O’Leary

Brendan

(eds) Power Sharing in Deeply Divided Places. Philadelphia: University of Pennsylvania Press, 253–277.

30.

Mukherjee

Bumba

(2006) Why Political Power-Sharing Agreements Lead to Enduring Peaceful Resolution of Some Civil Wars, but Not Others? International Studies Quarterly 50(2): 479–504.

31.

Neudorfer

Natascha S

Walsh

Dawn

(2025) The Peace-Making Role of Independent Commissions: The role of inclusion. International Political Science Review 46(5): 654–671.

32.

Oktay

Sibel

(2022) Governing Abroad: Coalition Politics and Foreign Policy in Europe. Ann Arbor, MI: University of Michigan Press.

33.

Paffenholz

Thania

Bramble

Alexander

(2025) Do Inclusive Commissions for Peace Agreement Implementation Lead to Inclusive Societies? An in-depth qualitative investigation of 11 cases. International Political Science Review.

34.

Quinn

J Michael

Mason

T David

Gurses

Mehmet

(2007) Sustaining the Peace: Determinants of civil war recurrence. International Interactions 33(2): 167–193.

35.

Ram

Megha

Strøm

Kaare Wallace

(2014) Mutual Veto and Power-Sharing. International Area Studies Review 17(4): 343–358.

36.

Roeder

Philip G

Rothchild

Donald S

(eds) (2005) Sustainable Peace: Power and Democracy After Civil Wars. Ithaca, NY: Cornell University Press.

37.

Schneckener

Ulrich

(2002) Making Power-Sharing Work: Lessons from successes and failures in ethnic conflict regulation. Journal of Peace Research 39(2): 203–228.

38.

Spears

Ian S

(2000) Understanding Inclusive Peace Agreements in Africa: The problems of sharing power. Third World Quarterly 21(1): 105–118.

39.

Tomz

Michael

Wittenberg

Jason

King

Gary

(2003) CLARIFY: Software for interpreting and presenting statistical results. Journal of Statistical Software 8(1): 1–30.

40.

Tsebelis

George

(2002) Veto Players: How Po liti cal Institutions Work. Princeton, NJ: Princeton University Press.

41.

Walsh

Dawn

(2020) Independent Commissions and Peace Settlements. In Richmond

Oliver

Visoka

Gëzim

(eds) The Palgrave Encyclopedia of Peace and Conflict Studies. Cham: Palgrave Macmillan, 1–6.

42.

Walsh

Dawn

Doyle

John

(2018) External Actors in Consociational Settlements: A re-examination of Lijphart’s negative assumptions. Ethnopolitics 17(1): 21–36.

43.

Walter

Barbara

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

44.

Walter

Barbara F

(2015) Why Bad Governance Leads to Repeat Civil War. Journal of Conflict Resolution 59(7): 1242–1272.

45.

Wantchekon

Leonard

(2000) Credible Power-Sharing Agreements: Theory with evidence from South Africa and Lebanon. Constitutional Political Economy 11(4): 339–352.

46.

Werner

Suzanne

(1999) The Precarious Nature of Peace: Resolving the issues, enforcing the settlement, and renegotiating the terms. American Journal of Political Science 40(3): 912–934.

47.

Yakinthou

Christalla

(2009) Political Settlements in Divided Societies. Basingstoke: Palgrave McMillan.

Supplementary Material

Please find the following supplemental material available below.

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

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

0.00 MB

0.04 MB