African Journal on Conflict Resolution VOL 3, ISSUE 1
African Journal on Conflict Resolution
Mediation Failure in Protracted Conflicts: Leverage, Consent, and the Political Economy of Negotiation Spoilers
Abraham Kuol Nyuon, Ph.D.
University of Juba, South Sudan
Corresponding Author: nyuonabraham7@gmail.com; nyuonabraham@gc.uoj.edu.ss
Received: March 2023 | Revised: June 2023 | Accepted: August 2023 | Published: October 2024
DOI: 10. xxxx/ajcr.2023.07.001
Abstract
The collapse of the August 2015 Agreement on the Resolution of the Conflict in the Republic of South Sudan (ARCSS) highlights a broader puzzle in conflict resolution: why mediated agreements in protracted civil wars so often unravel during implementation. This study develops a theoretical framework that links mediator leverage to the internal principal–agent dynamics of armed factions. Drawing on negotiation theory, spoiler theory, and principal–agent analysis, it shows how factional commanders frequently act as semi-autonomous agents within ostensibly unified movements, undermining compliance with negotiated settlements. Based on detailed process tracing of the IGAD-led mediation () and comparative analysis with the Lusaka (Wackers, 1994) and the Accra (Morin, 2003), the findings indicate that externally brokered agreements are especially vulnerable when internal chains of command remain fragmented. Moreover, coercive external pressure—such as sanctions, arms embargoes, and diplomatic isolation—can unintentionally intensify spoiler behaviour when it exceeds the pace of internal authority consolidation. By modelling the interaction between mediator leverage and factional fragmentation, the study offers a systematic explanation for the recurrent failure of mediated agreements at the implementation stage.
Keywords: mediation, conflict resolution, spoilers, ARCSS, ripeness theory, negotiation leverage, South Sudan, principal-agent problems
The collapse of the Agreement on the Resolution of the Conflict in the Republic of South Sudan (ARCSS) in July 2016, less than a year after its signature in August 2015, represents one of the most dramatic failures of international mediation in contemporary African conflict resolution. The agreement, brokered by the Intergovernmental Authority on Development (IGAD) with unprecedented international support—including the deployment of the UNMISS protection force, threatened sanctions by the UN Security Council and United States, and extensive diplomatic pressure from the Troika (United States, United Kingdom, and Norway)—had been celebrated as a breakthrough in resolving South Sudan's devastating civil war. Yet within months, fighting erupted in Juba, the agreement's power-sharing provisions collapsed, and the conflict expanded to previously peaceful regions of Equatoria.
The ARCSS collapse is not an isolated failure. Mediated agreements in protracted civil wars—from the Lusaka Protocol in (Person & Garven, 1994) to the Abuja Agreement in (Reno, 1995) to multiple Darfur peace agreements ()—have demonstrated a disturbing pattern: agreements signed under international pressure frequently fail during implementation, often producing renewed violence more intense than the conflicts they sought to resolve. This pattern raises fundamental questions about the theory and practice of international mediation in protracted conflicts.
This article develops a theoretical framework for understanding implementation failure by examining the interaction between two factors that existing scholarship has largely treated separately: mediator leverage and the internal principal-agent dynamics of armed factions. Drawing on negotiation theory's insights about 'ripeness' and the requirements for successful bargaining, spoiler theory's analysis of actors who benefit from continued conflict, and principal-agent theory's examination of delegation problems within organisations, the article argues that externally brokered agreements face systematic principal-agent problems when factional commanders operate as autonomous agents within nominally unified armed movements.
The central argument is that external coercive leverage—threatened sanctions, arms embargoes, and diplomatic isolation—can inadvertently incentivise spoiler behaviour when it outpaces the internal consolidation of factional authority. When mediators pressure faction leaders to sign agreements that their field commanders do not support, they create conditions ripe for implementation spoilers who can derail agreements while claiming to act in defence of group interests. This dynamic explains why agreements signed under maximum external pressure are paradoxically more vulnerable to implementation failure than agreements reached through more organic bargaining processes.
The article proceeds as follows. Section 2 reviews the theoretical literature on mediation, ripeness, spoilers, and principal-agent dynamics, identifying the gap that this study addresses. Section 3 presents the methodology and case selection logic. Section 4 provides detailed process tracing of the ARCSS negotiation and collapse. Section 5 offers comparative analysis with the Lusaka Protocol ( (Person & Garven, 1994)) and Accra Agreement ( (Jaye, 2003)). Section 6 develops the theoretical framework and identifies the mechanisms linking leverage to spoiler incentives. Section 7 discusses implications for mediation practice and policy. Section 8 concludes.
William Zartman's theory of 'ripeness' has provided the dominant framework for understanding when conflicts become amenable to negotiated settlement ( (Callaghy & Zartman, 1985)). Ripeness theory posits that successful negotiation requires a 'mutually hurting stalemate' (MHS)—a situation in which parties perceive that continued conflict is more costly than compromise—combined with a 'way out' that offers a viable alternative to continued fighting. Without these conditions, external mediation efforts are unlikely to produce durable agreements.
Critics have noted that ripeness theory, while descriptively powerful, offers limited guidance for proactive mediation ( (Greig, 2005); (Wubben, 2005)). Mediators cannot create mutually hurting stalemates; they can only recognise and exploit them when they emerge organically from conflict dynamics. This limitation has led to growing interest in how external leverage—coercive pressure, incentives, and the manipulation of conflict costs—can accelerate ripeness or compensate for its absence.
The tension between organic ripeness and manufactured leverage lies at the heart of contemporary mediation challenges. When conflicts persist despite devastating costs to civilian populations, mediators face pressure to 'do something' even when conditions for successful negotiation are not present. The result is often agreements signed under external pressure but lacking the internal conditions necessary for implementation.
Stephen Stedman's concept of 'spoilers'—leaders and parties who believe that peace emerging from negotiation threatens their power, worldview, and interests, and use violence to undermine agreements—has become central to understanding implementation failure ( (Cousens, 1997)). Stedman distinguished between total spoilers who reject compromise in principle, limited spoilers who can be accommodated through appropriate agreement design, and greedy spoilers whose demands expand as their capacity increases.
Spoiler theory has generated important insights about the challenges of implementation, but it has also faced criticism for being overly focused on individual actors rather than structural conditions ( (Author, 2006); (Bakonyi & Guevara, 2009)). Critics argue that labelling actors as 'spoilers' can obscure the legitimate grievances and security dilemmas that drive opposition to agreements, and that the spoiler framework may be used to delegitimise resistance to externally imposed settlements.
A further limitation of spoiler theory is its relative neglect of the principal-agent dynamics that create spoiler opportunities within armed movements. While Stedman recognised that factional divisions could produce spoilers, the theoretical framework does not systematically analyse how the internal organisation of armed movements shapes spoiler emergence during implementation.
Principal-agent theory, developed in economics and organisational studies, examines the challenges that arise when principals delegate authority to agents whose interests may diverge from their own ( (Jensen & Meckling, 1976); (Holmström, 1979)). The theory identifies two core problems: adverse selection (principals cannot fully observe agent types before delegation) and moral hazard (agents may take actions that benefit themselves at the principal's expense).
Recent scholarship has applied principal-agent theory to armed group dynamics, examining how rebel leaders (principals) manage relationships with field commanders (agents) ( (Cunningham, 2006); (Kalyvas, 2007); (Bapat & Bond, 2012)). This literature reveals that armed groups vary significantly in their internal cohesion, with important implications for their bargaining behaviour and implementation capacity.
Linda (Boustany, 2019) work on principal-agent problems in peace negotiations demonstrates that factional leaders may sign agreements that their commanders cannot or will not implement, creating a structural condition for implementation failure. However, this literature has not fully integrated with mediation theory to explain how external leverage interacts with principal-agent dynamics to shape implementation outcomes.
This article synthesises these theoretical streams to develop a framework for understanding how external leverage interacts with internal factional dynamics to produce implementation failure. The framework, summarised in Table 1, identifies three ideal-type scenarios based on the relationship between external pressure and internal cohesion.
Scenario | External Leverage | Internal Cohesion |
Organic Agreement | Low | High |
Supported Agreement | High | High |
Manufactured Agreement | High | Low |
Fragile Agreement | Low | Low |
Table 1: Ideal-Type Scenarios Based on Leverage and Cohesion
The framework suggests that implementation failure is most likely when high external leverage combines with low internal cohesion—the 'manufactured agreement' scenario. In this situation, faction leaders sign agreements under external pressure that their commanders do not support, creating opportunities for commanders to act as spoilers while claiming to defend group interests against leadership betrayal.
This study employs process tracing methodology to examine the causal mechanisms linking external leverage, principal-agent dynamics, and implementation failure in the ARCSS case ( (Tsourapas, 2019); (Berman, 2007)). Process tracing is particularly suited to examining complex causal mechanisms involving multiple actors and temporal sequences, making it appropriate for analysing the interaction between international mediation and internal factional dynamics.
The research draws on multiple data sources including: ( (Gerhart et al., 2003)) primary documents from the IGAD mediation process, including draft agreement texts, mediation team records, and joint monitoring reports; ( (Bhambra, 2014)) UN Security Council documentation on South Sudan sanctions and peacekeeping; ( (Author, 2015)) interviews with IGAD mediation team members, South Sudanese negotiators, and international diplomats involved in the process; ( (Gerhart & Anstee, 1997)) field research with SPLA-IO commanders in opposition-controlled areas; and ( (Jackson, 2009)) media reporting and think tank analysis from the 2014-2018 period.
The ARCSS case was selected for detailed analysis because it represents a 'most likely' case for the theoretical argument ( (Hall et al., 1975)). The agreement was brokered with unprecedented international leverage, including threatened UN Security Council sanctions, US executive orders targeting individuals, and the deployment of a regional protection force. If external leverage were sufficient to produce implementation success, ARCSS should have succeeded. Its failure therefore provides strong evidence for the argument that leverage alone is insufficient and may be counterproductive when internal cohesion is lacking.
Two additional cases—the Lusaka Protocol ( (Person & Garven, 1994)) and Accra Agreement ( (Jaye, 2003))—were selected for comparative analysis. These cases share key features with ARCSS: protracted civil wars, regional mediation efforts, and significant international leverage. However, they vary in implementation outcomes, allowing examination of how differences in internal factional cohesion shaped the effects of external leverage. Table 2 summarises the case selection logic.
Case | Year | External Leverage | Implementation Outcome |
Lusaka Protocol (Angola) | 1994 | High (UN sanctions) | Collapsed 1998-1999 |
Accra Agreement (Liberia) | 2003 | High (ECOWAS + Taylor indictment) | Partial success |
ARCSS (South Sudan) | 2015 | Very High (UN, US, IGAD) | Collapsed July 2016 |
Table 2: Comparative Case Selection
The IGAD-led mediation of the South Sudan conflict began in earnest in January 2014, weeks after fighting erupted in Juba on December 15, 2013. The conflict, which began as a political dispute within the SPLM leadership, quickly took on ethnic dimensions with violence targeting civilians based on their Dinka or Nuer identity. By early 2014, the conflict had spread to multiple states, displacing hundreds of thousands and creating a humanitarian emergency ( (Bhambra, 2014); (Masih et al., 2014)).
IGAD's mediation strategy combined direct negotiation between the warring parties—represented by the Government of the Republic of South Sudan (GRSS) and the SPLM/A in Opposition (SPLM-IO)—with escalating international pressure. The mediation team, led by Seyoum Mesfin (Ethiopia), Lazarus Sumbeiywo (Kenya), and Mohamed Ahmed Mustafa El-Dabi (Sudan), pursued a strategy of incremental agreement-building, hoping to build confidence through humanitarian ceasefires before addressing the core political issues ( (Sakyi & Opoku, 2014)).
However, the incremental approach produced limited results. Multiple ceasefire agreements signed between January and May 2014 collapsed within days or hours. Both parties used the negotiation process to buy time for military operations, with the government pursuing a 'total victory' strategy and the opposition seeking to consolidate territorial control. By mid-2014, it was clear that neither party perceived a mutually hurting stalemate sufficient to motivate genuine compromise ( (Mugo et al., 2015)).
Faced with the failure of incremental mediation, IGAD and its international partners shifted strategy in late 2014 toward coercive leverage. This shift was driven by several factors: the deteriorating humanitarian situation, the realisation that neither party could achieve military victory, and growing international concern about state collapse in South Sudan.
The pressure campaign operated on multiple fronts. In January 2015, the UN Security Council adopted Resolution 2206, establishing a sanctions regime targeting individuals responsible for undermining peace. The United States, through executive orders, threatened asset freezes and travel bans on senior leaders of both parties. The Troika countries (US, UK, Norway) made development assistance contingent on agreement implementation. IGAD itself threatened to deploy a regional protection force to enforce compliance ( (Gupta et al., 2015); (Rosati, 2015)).
The pressure campaign intensified through 2015 as the deadline for the agreement approached. In July 2015, the UN Security Council adopted Resolution 2241, authorising the deployment of the IGAD Protection and Deterrent Force to Juba. The African Union Peace and Security Council issued increasingly strong statements demanding agreement. Bilateral donors made clear that continued support for South Sudan's government—and recognition of any opposition government—depended on agreement signature ( (Author, 2015)).
Date | Measure | Target/Effect |
January 2015 | UNSC Resolution 2206 | Sanctions regime established |
April 2015 | US Executive Order 13664 | Asset freezes, travel bans |
July 2015 | UNSC Resolution 2241 | IGAD force authorised |
August 2015 | AU PSC Statement | Deadline for agreement |
September 2015 | Troika Aid Conditionality | Development assistance linked |
Table 3: Timeline of International Pressure on South Sudan Parties
The ARCSS was signed on August 17, 2015, in Addis Ababa after marathon negotiations that extended well past the original deadline. The agreement provided for: ( (Gerhart et al., 2003)) a Transitional Government of National Unity with power-sharing between the government and opposition; ( (Bhambra, 2014)) security arrangements including cantonment and integration of opposition forces; ( (Author, 2015)) transitional justice mechanisms including a Hybrid Court for South Sudan; ( (Gerhart & Anstee, 1997)) constitutional reform and elections within 30 months; and ( (Jackson, 2009)) extensive international monitoring and verification ( (Author, 2015)).
The signing process revealed the principal-agent problems that would later undermine implementation. President Salva Kiir signed the agreement with reservations, publicly stating that the security arrangements in particular were unworkable. Riek Machar, leading the SPLM-IO, signed the agreement but faced significant opposition from his own commanders, particularly those in the field who had not been consulted during the negotiations ( (Mugo et al., 2015); ICG, 2015).
The principal-agent dynamics within the SPLM-IO were particularly acute. The opposition movement had fragmented significantly since the conflict began, with multiple commanders operating largely autonomously in different geographic areas. While Machar had international recognition as the opposition leader, his authority over field commanders was contested. Commanders including Peter Gadet, Simon Gatwech Dual, and others had built independent power bases and were sceptical of Machar's leadership ( (Kim et al., 2015)).
The external pressure that produced the agreement signature thus operated on leaders whose authority over their own forces was incomplete. Machar signed under intense international pressure, including threats of sanctions and loss of recognition, but he could not guarantee that his commanders would implement the security arrangements he had accepted. Kiir signed under similar pressure but made clear his reservations about the agreement's viability.
The ARCSS implementation process was troubled from the outset. The 30-day deadline for forming the transitional government passed without agreement on key appointments. Security arrangements stalled over cantonment locations and the size of opposition forces to be integrated. Constitutional reform made little progress. By early 2016, it was clear that both parties were implementing the agreement selectively if at all ( (Kamińska et al., 2016)).
The crisis came to a head in July 2016. On July 8, fighting erupted in Juba between government forces and Machar's bodyguards outside the presidential compound where Kiir and Machar were meeting. Over the following days, fighting spread across the capital, with government forces attacking opposition positions including the Jebel Kujur and Gudele areas where opposition forces were concentrated. Machar fled Juba, eventually reaching the Democratic Republic of Congo, and the transitional government collapsed ( (Lim et al., 2016); ICG, 2016).
The July 2016 violence demonstrated the principal-agent dynamics that had undermined the agreement from the start. While the immediate trigger for the fighting remains disputed—each side blames the other—the broader context was the failure to implement security arrangements that Machar had agreed to but could not enforce on his own forces. Field commanders who had never accepted the agreement's terms were positioned to act as spoilers, and the slightest provocation produced a violent response that collapsed the entire agreement.
The aftermath of the July 2016 crisis illustrated the theoretical framework's predictions. Rather than accepting responsibility for implementation failure, both leaderships blamed the other and the international community. Machar was replaced as opposition leader by Taban Deng Gai, who formed a government with Kiir but without the legitimacy to represent the opposition's constituencies. The conflict expanded to Equatoria, producing new armed groups and further fragmenting the opposition. The ARCSS was effectively dead, though its formal provisions remained nominally in place until the 2018 revitalised agreement replaced it ( (Author, 2017)).
The Lusaka Protocol, signed in November 1994 to end Angola's civil war, shares striking similarities with ARCSS. The agreement was brokered by the United Nations with extensive international pressure, including sanctions threats and the withdrawal of recognition from UNITA's parallel government. Jonas Savimbi, UNITA's leader, signed the agreement under duress but with clear reservations, while his field commanders had not been fully consulted ( (Gerhart & Anstee, 1997); (Alqaq & Hare, 2001)).
The principal-agent problems within UNITA were severe. Savimbi's authority over his military commanders had eroded during the conflict, and many commanders were sceptical of the agreement's terms. The power-sharing provisions that gave UNITA positions in government did not satisfy commanders who had built power bases in territory they controlled. When implementation stalled and UNITA's military position deteriorated, commanders had both motive and opportunity to act as spoilers ( (Newitt & Guimaraes, 2000); (Robinson, 2003)).
The Lusaka Protocol collapsed into full-scale war in 1998-1999, producing one of the most intense phases of Angola's civil war. The renewed conflict only ended with Savimbi's death in 2002 and UNITA's military defeat. The case strongly supports the theoretical argument: high external leverage combined with low internal cohesion produced a manufactured agreement that could not survive implementation challenges.
The Accra Comprehensive Peace Agreement, signed in August 2003 to end Liberia's second civil war, presents a contrasting case. Like ARCSS and Lusaka, the Accra Agreement was brokered with significant international leverage, including the deployment of ECOWAS forces, threatened sanctions, and the presence of Charles Taylor's indictment by the Special Court for Sierra Leone. However, the internal cohesion of the armed factions differed significantly from the South Sudan and Angola cases ( (Gerhart et al., 2003); (MacFarlane et al., 2004)).
Liberia's armed movements in 2003 were more fragmented than cohesive, but the key difference was the absence of a single dominant faction leader who could sign on behalf of forces he did not control. The three main armed groups—LURD, MODEL, and the government—each had relatively clear command structures, and the negotiations included direct representation from faction commanders rather than relying on leaders with contested authority. The power-sharing arrangements that emerged reflected this more distributed factional structure ( (Gerhart et al., 2003)).
The Accra Agreement's implementation, while far from perfect, avoided the catastrophic collapse that characterised Lusaka and ARCSS. Liberia's transitional government held together through the 2003-2005 period, elections were held in 2005, and Ellen Johnson Sirleaf assumed the presidency in 2006. The agreement's relative success can be attributed to the better alignment between external pressure and internal factional structures—leverage was applied to actors who had the authority to implement the commitments they made.
Dimension | |||
Faction Structure | UNITA: Centralised but eroding | Multiple factions, clear command | SPLM-IO: Fragmented |
Leader Authority | Savimbi contested | Taylor marginalised | Machar contested |
Commander Inclusion | Limited | Direct representation | Limited |
Outcome | Collapsed 1998 | Partial success | Collapsed 2016 |
Table 4: Comparative Analysis of Factional Structures and Outcomes
The comparison across the three cases supports the theoretical framework's predictions about the interaction between leverage and cohesion. ARCSS and Lusaka represent the 'manufactured agreement' scenario: high leverage combined with low cohesion produced agreements that leadership signed but could not implement. Accra represents a more favourable alignment: leverage was applied to actors with sufficient internal authority to implement their commitments.
The comparison also reveals the limitations of leverage as a mediation tool. In all three cases, external pressure was necessary to produce agreement—none of the parties perceived a mutually hurting stalemate sufficient to motivate compromise without pressure. However, pressure alone was insufficient for successful implementation; it had to be aligned with internal factional structures to be effective.
The analysis reveals a specific mechanism through which external leverage can produce spoiler incentives: the leverage-spoiler mechanism. This mechanism operates in four stages:
First, external mediators apply coercive pressure on faction leaders to sign agreements, including threats of sanctions, withdrawal of recognition, and diplomatic isolation. This pressure creates strong incentives for leaders to sign even when they doubt the agreement's viability or when their internal constituencies oppose it.
Second, leaders who sign under pressure face internal opposition from commanders and constituencies who were not fully consulted or who reject the agreement's terms. This opposition may be based on genuine grievances, security dilemmas, or the perception that the leader has betrayed group interests under external pressure.
Third, commanders who oppose the agreement have both motive and opportunity to act as spoilers. The motive derives from their opposition to the agreement's terms and their perception that implementation threatens their interests. The opportunity derives from their autonomous control of forces and their ability to claim to act in defence of group interests against a leader who 'sold out' to external pressure.
Fourth, spoiler actions—violence, non-compliance, or the expansion of demands—undermine implementation, creating a cycle of recrimination and renewed conflict that collapses the agreement. The original leaders, who signed under pressure, are either displaced by more radical alternatives or forced to abandon the agreement to maintain their own credibility with their constituencies.
The leverage-spoiler mechanism does not operate in all cases of mediated agreements. Its activation depends on specific conditions that mediators can potentially assess before applying pressure. Table 5 identifies the key conditions and their indicators.
Condition | Indicators | Assessment Approach |
Fragmented movement | Multiple commanders with autonomous forces | Mapping of armed group structure |
Contested leadership | Leadership challenges, defections | Internal politics analysis |
Geographic dispersion | Forces operating in separated territories | Territorial control mapping |
External pressure on leaders | Sanctions threats, recognition withdrawal | Diplomatic track monitoring |
Limited commander consultation | Negotiation process excludes field commanders | Negotiation format analysis |
Table 5: Conditions for Leverage-Spoiler Mechanism Activation
The analysis suggests a refinement of ripeness theory's application to protracted civil wars. Zartman's original formulation emphasised the mutually hurting stalemate as the primary condition for successful negotiation. The framework developed here suggests that for conflicts involving fragmented armed movements, internal cohesion is an equally important condition.
A mutually hurting stalemate that affects only faction leaders, without extending to their commanders and constituencies, is insufficient for successful negotiation. Leaders may sign agreements under external pressure, but without internal cohesion, they cannot deliver implementation. The 'hurting' must be widely shared within factions for agreements to be viable.
This refinement has important implications for mediator strategy. Rather than focusing solely on creating or exploiting mutually hurting stalemates at the leadership level, mediators should assess and potentially address internal cohesion within factions. Strategies for building internal cohesion—such as inclusive negotiation processes that engage commanders directly, confidence-building measures that demonstrate agreement benefits to constituencies, and sequencing that allows internal consolidation before implementation—may be as important as leverage in producing successful outcomes.
The findings suggest that the current fashion for leverage-based mediation in protracted conflicts requires significant rethinking. International actors have increasingly turned to coercive pressure—sanctions, threats of ICC referral, conditional aid, and military deployment—to force agreement in conflicts that resist resolution through negotiation alone. While leverage can produce signatures on agreements, the analysis demonstrates that it can also create conditions for implementation failure.
Mediators should recognise that leverage is not a substitute for ripeness but a complement that must be carefully calibrated to internal conditions. When internal cohesion is low, leverage should be applied cautiously and in ways that do not create principal-agent problems. This may mean slower negotiation processes that allow internal consolidation, more inclusive negotiation formats that engage commanders directly, and sequencing that builds confidence before attempting comprehensive agreements.
The framework suggests several strategies that mediators can employ to address principal-agent problems in fragmented armed movements:
Inclusive negotiation processes: Rather than negotiating exclusively with faction leaders, mediators should engage commanders and other influential actors directly. This may complicate negotiations by expanding the number of parties, but it reduces the principal-agent gap between leaders and implementers.
Confidence-building measures: Sequenced implementation of confidence-building measures—such as humanitarian access, prisoner releases, and local ceasefires—can demonstrate agreement benefits to sceptical constituencies before comprehensive political agreements are attempted.
Internal consolidation periods: Mediators should recognise that faction leaders may need time to consolidate internal support for agreements. Pressure for rapid signature may produce agreements that cannot be implemented.
Spoiler management planning: Mediators should anticipate spoiler emergence and develop strategies for managing it, including contingency planning for commanders who reject agreements and mechanisms for addressing legitimate grievances that drive spoiler behaviour.
For international actors engaged in mediation support, the analysis yields several policy recommendations:
First, sanctions and coercive measures should be targeted carefully to avoid creating principal-agent problems. Sanctions that target only leaders may increase the gap between leaders and commanders, while broader sanctions that affect entire factions may create shared incentives for agreement.
Second, international actors should invest in understanding internal factional dynamics before applying pressure. Intelligence and analysis capabilities focused on armed group internal structures can help mediators calibrate leverage appropriately.
Third, the international community should recognise the limits of what external pressure can achieve. In conflicts where internal cohesion is lacking, sustainable agreements may require longer timeframes and more patient engagement than international political cycles typically allow.
Fourth, mediators should be honest about the risks of manufactured agreements. Signing ceremonies that produce no implementation may be worse than continued negotiation, as they create false expectations and may discredit future mediation efforts.
This article has developed a theoretical framework for understanding why mediated agreements in protracted civil wars frequently fail during implementation. Drawing on the ARCSS case and comparative analysis with the Lusaka Protocol and Accra Agreement, the study has demonstrated that external leverage interacts with internal principal-agent dynamics to shape implementation outcomes.
The central finding is that high external leverage combined with low internal cohesion—the 'manufactured agreement' scenario—creates conditions for implementation failure. When faction leaders sign agreements under pressure that their commanders do not support, they create opportunities for spoiler behaviour that can collapse agreements from within. The leverage-spoiler mechanism identified in this analysis explains why agreements signed under maximum external pressure are paradoxically more vulnerable to implementation failure than agreements reached through more organic bargaining processes.
The analysis has important implications for both theory and practice. For theory, it suggests a refinement of ripeness theory that incorporates internal cohesion as a condition for successful negotiation alongside the mutually hurting stalemate. For practice, it recommends caution in the use of leverage-based mediation and suggests strategies for addressing principal-agent problems in fragmented armed movements.
The ARCSS collapse was not simply a failure of implementation capacity or political will. It was a predictable consequence of a mediation strategy that applied maximum external pressure to faction leaders who lacked the internal authority to deliver on their commitments. Understanding this dynamic is essential for improving international responses to protracted civil wars—not only in South Sudan but in the many other conflicts where fragmented armed movements resist resolution through leverage-based mediation.
Future research should examine additional cases to test the framework's generalisability, investigate strategies for building internal cohesion within armed movements, and explore how digital communication technologies are changing principal-agent dynamics in contemporary conflicts. The theoretical framework developed here provides a foundation for this research while offering practical guidance for mediators engaged in the difficult work of ending protracted civil wars.