One of the great political problems in Africa is the scourge of coups. According to Patrick McGowan, 80 successful coups, 108 failed coup attempts, and 139 reported coup plots took place between 1956 and 2001.1P. McGowan, “African Military Coups d’État, 1956–2001: Frequency, Trends, and Distribution,” Journal of Modern African Studies 41, no. 3 (2003): 339–70. The problem has not diminished over the past decade. Since 2001, coups have toppled governments in the Central African Republic (2003), Egypt (2013), Guinea (2008), Guinea-Bissau (2003 and 2012), Madagascar (2009), Mali (2012), Mauritania (2005 and 2008), Niger (2010), Sao Tome and Principe (2003), and Togo (2005).On the positive side, the African Union (AU), unlike the Organization of African Unity (OAU), has a firm policy commitment to reject coups and other unconstitutional changes of government. The AU’s Constitutive Act of 2000 states that governments that gain power by unconstitutional means shall be suspended from the organization, and sanctions shall be imposed on them. The African Charter on Democracy, Elections, and Governance, adopted by the AU heads of state in 2007, prohibits coup legitimization through elections. It stipulates that the perpetrators of unconstitutional actions may not stand in elections held to restore democracy or occupy any positions of responsibility in the political institutions of their states.

At first glance, the ban on coup legitimization is a decisive rejection of the military overthrow of governments and, thus, a compelling deterrent; without it, coup-makers can seize power, withstand suspension and sanctions for a brief period, and then legitimize their capture of power by organizing and winning elections. Upon further investigation, however, the ban turns out to be incompatible with mediation, which the AU invariably undertakes to restore constitutional order.

This incompatibility was strongly evident in the Madagascar coup of 2009, which led to mediation by the United Nations, the AU, and the Southern African Development Community (SADC). SADC’s demand that the coup-makers relinquish power permanently was a major impediment to peacemaking. It generated friction among the mediating bodies, provoked antagonism between the mediators and the Malagasy regime, and intensified the regime’s resistance to a negotiated solution. The ban on coup legitimization was ignored from the outset by the AU, and abandoned at a later stage by SADC, because it retarded the attainment of an inclusive and stable settlement.

The Madagascar Mediation

In March 2009, Madagascar was plunged into crisis when Andry Rajoelina, mayor of the capital city Antananarivo, ousted President Marc Ravalomanana by unconstitutional means. With the backing of the military, Rajoelina formed the High Transitional Authority (Haute Autorité de la Transition, hereafter “the HAT”), which became the de facto government.

The AU and SADC immediately condemned the coup and suspended Madagascar. The SADC Summit demanded that constitutional rule be restored at once, that Rajoelina vacate the presidency as a matter of urgency, and that Ravalomanana be reinstated unconditionally, threatening to use force if these demands were not met.

Over the next four years a number of international and domestic mediation initiatives were launched. The peace process was wracked by unseemly competition among the UN, the AU, and SADC over the leadership of the mediation and by their rancorous normative and strategic disagreements. Both impaired the quality, coherence, and credibility of external peacemaking in Madagascar.2D. Lanz and R. Gasser, “A Crowded Field: Competition and Coordination in International Peace Mediation,” Mediation Arguments working paper series, no. 2, Centre for Mediation in Africa, University of Pretoria, 2013.

The UN, the first of the international actors to embark on mediation, considered this strategy incompatible with a condemnatory attitude. A member of the UN mediation team observed that SADC’s threat of force and demand for Ravalomanana’s reinstatement were not conducive to mediation, as they left Rajoelina and the HAT convinced that SADC favored the ousted president; the UN, on the other hand, “called for a negotiated return to a democratic order, a process-centred approach which helped conserve the organization’s credibility as a mediator.”3P. Gavigan, The 2009 Madagascar Crisis and International Mediation (New York: Social Science Research Council, 2010), 11.

Like the UN, the AU did not regard mediation as compatible with a demand that Rajoelina step down or refrain from participating as a candidate in future elections. During the negotiations in 2009, the chairperson of the AU Commission, Jean Ping, told the Malagasy leaders that the ban on coup legitimization could be overridden by an agreement among them. The AU proposed that Rajoelina, Ravalomanana, and the other former presidents of Madagascar all be free to participate as candidates in future elections.4Agence France-Presse, “Former Madagascar Leader Quits Crisis Talks” (May 24, 2009), (accessed 5 January 2013).

The AU ignored the ban on coup legitimization in the interests of peace and stability. The Malagasy army had indicated it would resist attempts to replace Rajoelina as the transitional president, and the AU feared Ravalomanana’s return to Madagascar from exile in South Africa would provoke violence. In this precarious situation, the AU believed the short-term priority was to prevent a slide into civil war, and the medium-term priority was to forge a smooth passage to constitutional normalcy through credible elections. The ban on coup legitimization was deemed inimical to these priorities and to reaching a mediated solution.

In contrast to the AU, SADC responded to the coup by asserting the AU policy on unconstitutional change of government. It demanded Rajoelina step down and Ravalomanana be reinstated unconditionally. At the outset it did not foresee these outcomes being realized through mediation. Instead, its strategy was to threaten force.

SADC’s hard-line stance damaged the UN and AU mediation endeavors. It gave false assurance to Ravalomanana, emboldening him and discouraging him from engaging in negotiations with Rajoelina.5Lanz and Gasser, “A Crowded Field.” It also reinforced Rajoelina’s intransigence, enabling him to strengthen his domestic status by mobilizing popular support against external intervention,6G. Cawthra, The Role of SADC in Managing Political Crisis and Conflict: The Cases of Madagascar and Zimbabwe, FES Peace and Security Series (Maputo, Mozambique: Friedrich Ebert Stiftung, 2010), 20. and it impeded the international community’s efforts to tackle the crisis in a unified manner, with SADC pressing for an unqualified return to the status quo ante while the UN and AU mediators promoted an inclusive transition to elections.

When SADC appointed a mediator in June 2009, it failed to appreciate the incompatibility between this form of peacemaking and the demand that one of the parties surrender power permanently. The demand led Rajoelina and the HAT to view SADC and its mediator, former president Joaquim Chissano of Mozambique, as biased against them,7 Gavigan, “The 2009 Madagascar Crisis”; International Crisis Group, “Madagascar: Crisis Heating Up?” Africa Report 166 (2010). which made it difficult for Chissano to win the regime’s trust and cooperation. He also struggled to build public support for the mediation, since the pro-HAT media in Madagascar regularly vilified him.

In June 2011, the SADC Summit abandoned the ban on coup legitimization as unattainable and counterproductive. It approved the Roadmap for Ending the Crisis in Madagascar, which envisaged the formation of an inclusive transitional government and the holding of elections under UN supervision. The roadmap entitled Rajoelina to remain the transitional president, permitted members of the interim government to run in the elections if they left office 60 days before the vote, and urged the HAT to allow the unconditional return of Ravalomanana.

At the end of 2012, the SADC heads of state changed tack again and approved the so-called ni-ni solution, under the terms of which Ravalomanana and Rajoelina would agree that neither of them would enter the presidential race. Having failed to persuade the HAT and the military to allow Ravalomanana to participate as a candidate in the election, the summit hoped a deal that also kept Rajoelina from running for office would save face for both Ravalomanana and SADC. The two rivals initially consented to the deal, but Rajoelina subsequently reneged on his promise and announced he would run for office in the election. As of July 2013, the situation was tense and fluid, and when the elections would take place remained uncertain.

Contradictions between Mediation and the AU Policy

The Madagascar case highlights a range of contradictions between the AU policy on unconstitutional change of government and the goal, logic, and character of mediation.

International mediation is defined as a nonviolent process of managing or resolving a conflict, whereby a third party helps the disputants, with their consent, to reach mutually satisfactory agreements.8See, for example, R. Fisher, Methods of Third Party Intervention (Berlin: Berghof Research Centre for Constructive Conflict Management, 2001), 4; J. Bercovitch, “Mediation and Conflict Resolution,” in The SAGE Handbook of Conflict Resolution, ed. J. Bercovitch, V. Kremenyuk, and I. W. Zartman, 340–54 (London: SAGE, 2009) 343–6; United Nations, Guidance for Effective Mediation (New York: United Nations, 2012). The purpose is not to enable one of the disputants to win but, rather, to forge a settlement that is endorsed by all sides. To this end, mediation must be made acceptable to the adversaries, who must in turn cooperate with the mediator.9I. W. Zartman and S. Touval, “International Mediation,” in Leashing the Dogs of War: Conflict Management in a Divided World, ed. C. Crocker, F. Hampson, and P. Aall, 437–54 (Washington, DC: U.S. Institute for Peace, 2007), 437–8. Because the process is voluntary, the mediator must strive to win the parties’ trust, build their confidence in negotiations, and shift their relationship from one of belligerence to one of collaborative problem solving.10United Nations, Guidance, 4–9.

By contrast, the objective of the AU policy on unconstitutional change of government is to get the culpable party to surrender power permanently. Since that party may have nothing to gain and much to lose by complying, the policy anticipates a resort to sanctions, thereby engendering an antagonistic relationship between the sanctioning organization and the targeted party. A mediator deployed by that organization will be hard pressed to secure the targeted party’s cooperation.

The contradictions between mediation and the AU policy are thus stark. Whereas mediation aims to broker an accord that satisfies all the protagonists, the AU policy envisages a win-lose outcome, with the culpable party giving up power. Whereas mediation is a consensual venture, the AU policy seeks compliance under duress. Whereas mediation entails a third party’s efforts to assist the disputants, the AU policy calls on third parties to take coercive action. Whereas mediators must build cooperative relations with the parties in conflict, the AU policy induces an adversarial relationship between the peacemaker and the targeted party. And whereas mediators must be flexible and responsive to different situations and actors, the AU policy is peremptory and not meant to be adapted from case to case.

Another imperative of mediation, advocated by the UN, is that the process be inclusive to ensure the sustainability of a negotiated settlement and the durability of peace.11Ibid., 11–15. Excluded parties might seek to thwart the talks and the settlement. Consequently, mediators should design inclusive processes and minimize preconditions for participation.12Ibid., 11–13.

In the Madagascar case, the AU followed this logic, placing a premium on inclusivity in the negotiations, the transitional government, and elections. Yet the organization’s policy on unconstitutional change of government moves in the opposite direction, blocking the culpable individuals and groups from participating in elections to restore the democratic order and from holding any position of responsibility in the political institutions of their state.

The priority for the AU after the coup, and for SADC at a later stage, was to ensure stability in the short and long terms. Adhering to the ban on coup legitimization would have been counterproductive. The route to stability lay in inclusive processes and institutions. This is not to say the mediators sacrificed democracy on the altar of peace; rather, peace was a necessary condition for elections leading to the reinstallation of democracy. Without peace and stability, there could be no legitimate elections and hence no return to constitutional order.

Also relevant is that the mediators were constrained by the internal and external balance of power. At the domestic level, the HAT’s intransigence was bolstered by the armed forces and by the diplomatic support it enjoyed from France, and Ravalomanana was unable to mount a serious challenge to the regime. At the international level, the mediating bodies struggled to maintain a united front, the UN Security Council did not apply pressure, and the leverage of the AU and SADC was limited. Consequently, the ban on coup legitimization was infeasible.

In conclusion, there is a fundamental contradiction between a strategy that is consensual, cooperative, and inclusive and one that is peremptory, coercive, and exclusionary. Mediation and the prohibition on coup legitimization appear to be strategically and normatively incompatible. They are both based on “good” norms, and they are both legitimate courses of action; but for an international organization to follow both courses at the same time, playing simultaneously the roles of judge, “bad cop” enforcer, and “good cop” mediator, can be dysfunctional.

References   [ + ]