Power-sharing has a number of potentially detrimental effects on democratization, including elite-centrism, that are illustrated with examples around Africa, particularly Burundi, Côte d’Ivoire, Kenya, Zimbabwe, and Central African Republic (CAR). It might be surprising that power-sharing is still a recipe for many crisis situations on the continent given its dismal record. However, there are reasons why power-sharing (or rather the sharing of spoils) is still popular with political elites and international mediators aiming at terminating armed conflicts.
Over the past three decades the African continent has seen conflicts ended (at least temporarily) by power-sharing pacts between contending parties. Rarely, however, have these processes been positively linked to democratization. Power-sharing has become an attractive solution to severe political crises because it is credited with avoiding future armed confrontations by minimizing the security concerns of rival groups, particularly in dealing with a “credible commitment” problem when rebels face a “security dilemma” (Hartzell 1999; Walter 2002; Hartzell and Hoddie 2007). More specifically, three broad arguments are made in favor of power-sharing. First, it is argued that by guaranteeing rebels powerful positions, they can afford to lay down their arms without facing the risk of extinction. The second argument is that the costs of power-sharing (in terms of abandoned maximal positions) and time taken to design required institutions is in itself a signal of a commitment to peace, which should help build mutual confidence. The third argument pertains to the (assumed) functionality of new institutions in the management of future conflict. In this view, new institutions in which power is shared between former belligerents are best suited to deal with deep societal conflict (Hartzell and Hoddie 2007, 3f).
It is important to note that the most prominent form of power-sharing, i.e., its consociational sub-type, was created to make democracy work in plural societies (Lijphart 1977)—first for countries in Europe and around the Mediterranean, but quickly applied to other places, including South Africa. One core element of Lijphart’s thinking was that in deeply divided societies, minorities could be permanently excluded from power if an “undemocratic” majority rule is adopted. On closer inspection, however, the relation between power-sharing (at least as practiced in Africa) and democracy is often much more confrontational. Recent academic contributions on power-sharing have largely overlooked this. This is likely due to a tendency to focus on its relationship to conflict management, rather than democracy (Jarstad 2008; Sriram and Zahar 2009) or the distribution of power resources (Hartzell and Mehler 2019).
In this chapter, five potentially detrimental effects of power-sharing on democratization will be assessed and further illustrated with examples. These detrimental effects are: (1) elite-centrism; (2) sidelining of political parties for group representation; (3) loss of legitimacy of pro-democracy actors; (4) unintended consequences of both interim and permanent power-sharing institutions; and (5) the creation of perverse incentives to engage violently in politics. Finally, the chapter will draw some preliminary conclusions and provide some—potentially imprudent—generalizations.
It makes sense to first distinguish between two rather distinct contexts in which power-sharing has been applied in Africa. The first is as a more or less elaborated set of institutional devices to end protracted violent conflict; the second is as a means to stop post-electoral violence. The concrete forms of power-sharing in the first context can vary from simple cooptation of individual rebel leaders in government (as operated repeatedly in Chad in the 1990s), to a broader inclusion of rebel personnel alongside opposition members in government (e.g., in changing forms in Côte d’Ivoire between 2003 and 2007), to more complex schemes. The latter includes Burundi’s experiment with ethnic quotas (as negotiated in 2000–03), a rotating presidency in the Comoros (2000–03), and South Africa’s introduction of a government of national unity, a two-thirds majority in a Constituent Assembly, substantial provincial autonomy, and proportional representation (1990–93). The second context has become more important in the 2000s, notably as an exit option to severe post-electoral crisis in Kenya (2008), Zimbabwe (2009), or Zanzibar (as part of Tanzania, 2010).
The analysis that follows draws on five cases more deeply: Burundi, Côte d’Ivoire, Kenya, Zimbabwe, and Central African Republic (CAR). The first two are examples of postwar power-sharing agreements; the second two of power-sharing agreements that were signed to end a violent post-election crisis; and the final one resulted from an effort to end a mutiny, but set the tone for further conflict settlements. To varying degrees, they illustrate the problems that power-sharing can pose for democracy.
Burundi has witnessed recurrent civil war and episodes of extreme violence since independence, most frequently along the lines of the two main ethnic groups (Hutu are 85 percent of the population, Tutsi 14 percent). The last important effort to settle this conflict included a peace agreement in 2000 and the adoption of a completely new constitution in 2005. Burundi’s main political actors opted for a consociational-inspired power-sharing experiment that combined a grand coalition, minority (Tutsi) overrepresentation, and inclusion of rebel movements. 1 In 2015, “legal loopholes” (Vandeginste 2016) permitted President Nkurunziza to interpret the constitution of 2005 in a way that contradicted the spirit of the 2000 Arusha peace agreement and allowed him to stand a third time in elections. This resulted in a political crisis and widespread bloodshed; but left intact Tutsi overrepresentation as a cornerstone of the agreement.
From 1993 onwards, “Northerners” in Côte d’Ivoire amalgamated with migrants from Mali and Burkina Faso (of similar ethnic identity) were discriminated against by the inner circle of power. The personal ambitions of former Prime Minister Alassane Ouattara, a Northerner, exacerbated xenophobia from parts of the population in the South. Reflecting this, Ouattara was excluded from standing in presidential elections. A military coup in 1999 initiated a chaotic interim period that ended in contested elections in 2000. In 2002, a rebel movement conquered the Northern half of the country, initiating a short civil war that ended in a military stalemate. Côte d’Ivoire witnessed a series of power-sharing agreements between 2003 and 2007, which mostly focused on the composition of the national government, but also contained issues of election preparation and demobilization of pro-government militia and rebel movements (Mehler 2009). When postponed elections were finally held in 2010 (without prior demobilization), they led to a schism between incumbent President Laurent Gbagbo and challenger Ouattara. The latter had won the elections based on the data of the electoral commission; however, the Constitutional Court invalidated some results to allow Gbagbo to claim victory. A combined diplomatic and military effort by France and the rebel forces (cautioned by the United Nations and the Economic Community of West African States) removed Gbagbo from office, thereby ending the power-sharing experience, although the original motivations for it (i.e., a clear split of the country plus widespread violence) were as strongly present as ever before.
Kenya’s political history has seen a strong ethnicization of politics since independence and a continuity of political violence in contested elections after the reintroduction of multiparty politics in the early 1990s. Particularly severe were the confrontations that broke out after the disputed presidential election of late 2007. International negotiators brokered an interim power-sharing agreement that effectively stopped widespread violence. As one main element of the agreement, incumbent President Mwai Kibaki and his challenger Raila Odinga received leadership positions (president and prime minister respectively). This paved the way for a constitutional review process and subsequent adoption of a new constitution following a referendum in 2010, which necessitated significant reform of key political institutions. Arguably, some of those reforms affect the conviviality of ethnic groups, especially as they include a strong dose of political devolution, including elected county governors and assemblies (Cheeseman, Lynch, and Willis 2014). As a “centripetalist” ingredient (see below), the winner of presidential elections is now required to gain over 50 percent of the votes cast and at least 25 percent of the votes in each of more than half of the forty-seven counties to avoid a second round. Kenya therefore now has a system that encourages grand coalitions at the national level, and which involves the devolution of significant powers to county governments, thereby fostering a sense of power-sharing.
The first round of Zimbabwe’s March 2008 elections resulted in the defeat of President Robert Mugabe by his challenger Morgan Tsvangirai. But because the latter failed to obtain an absolute majority, a second round was necessary. The security apparatus subsequently engaged in widespread violence against the opposition and forced Tsvangirai to withdraw before the second round (Cheeseman and Tendi 2010, 210). This left the road open to Mugabe, whose reputation was, however, deeply damaged. Upon international mediation, a unity government between the Zimbabwe African National Union – Patriotic Front (ZANU-PF) and the two rival branches of the Movement for Democratic Change (MDC) was formed that quickly proved to be to the disadvantage of Mugabe’s challengers, mostly because they remained in a “junior partner” position without a strong grip on power.
In CAR (a very weak state according to all criteria; see Glawion, de Vries, and Mehler 2018), the first democratically elected president, Ange-Félix Patassé, faced a series of three mutinies between 1996 and 1997 by the country’s army, which had remained loyal to his military predecessor in office. France, strongly present with two military bases in the country, had at first quelled the unrest, but only reluctantly as Patassé was seen as anti-French. This led to negotiations and the formation of a grand coalition that included rebel representatives. Starting with this episode, individual politico-military entrepreneurs variably tried to repeat this example: showing the capacity to organize rebellion with the aim of being coopted into government. Not all such violent blackmailing succeeded. Patassé was finally dislodged by force by his former chief of staff François Bozizé in 2003. The most severe confrontation escalated in 2012–13, pitting the government of Bozizé against a rebel alliance mostly based in the neglected Muslim east of the country. Rebel leader and Northeasterner Michel Djotodia at first agreed to become member of a power-sharing government, but only briefly later decided to terminate this interlude, conquering the capital by force. The transitional government he formed had very limited support. Djotodia could retain the presidency only for nine months and no permanent political institutions for including the Muslim minority resulted. The rebel movement split and the arenas of violent confrontations fragmented, leading to continued turmoil even after internationally accepted elections held in 2017.
These five cases, but in fact all major recent experiences of power-sharing in Africa, expose a number of common problematic features. Before proceeding to highlight these, some cautionary remarks have to be made. First, blaming power-sharing institutions for the failure of post-conflict democratization may overlook the obvious, which is that many African countries that have or continue to experience widespread armed conflict already had a “crisis of representation” preceding the outbreak of hostilities, potentially aggravating conflict. A failure to effectively represent the political will of the population, or key segments of it, may therefore not be the result of a certain “imported” peace order, but represent root causes of a preceding war. Second, negative effects of power-sharing on democracy could theoretically be counterbalanced by positive ones. Some fervent advocates of grand coalitions, as one of the main ingredients of the consociational model of power-sharing, strongly believe in the beneficial effects of “governments of national unity” as “schools” to learn democratic behavior (applying mostly to supposedly uncivilized rebel leaders). However, the evidence for this effect is—at least in Africa—hard to find. Third, there are, of course, major variations in effects of different forms of power-sharing. Most accepted is the distinction between political, military, economic, and territorial power-sharing (promoted inter alia by Hartzell and Hoddie 2007), where only political and territorial power-sharing directly affects the workings of the political system. Alternatively, Gates et al. (2016) propose a distinction between inclusive, dispersive, and constraining power-sharing, where the last form is rather a power-dividing formula. The variety in political effects of power-sharing and their sub-forms is even more striking when distinguishing between permanent and interim power-sharing institutions. Finally, as always, articles on the entire continent run the risk of oversimplifying diverse African experiences. These caveats will be reflected in the elaboration below.
Power-sharing can be defined as an inclusive elite pact between representatives of political and/or military parties that divides responsibility in different fields of political and economic life. Most elements of the consociational democracy formula are not only negotiated between elites of supposedly ethnic or religious groups, but are also de facto operationalized and animated by elites. Lijphart, one of the foremost academics in the field, advocated a combination of: (1) government by a grand coalition between the political leaders of all significant segments of a plural society; (2) mutual veto in government decision-making; (3) proportionality as the principal standard of political representation, civil service appointments, and allocation of public funds; (4) and a high degree of group autonomy for each segment to run its own internal affairs (Lijphart 1977, 25ff). Beneficiaries of, and actors within, such institutions are mostly members of an elite and they may be tempted to privilege themselves at the expense of ordinary citizens. Gates et al. (2016, 519) consider the protection of entire societal groups as proof of success of power-sharing institutions, but in real-world peace negotiations, this is clearly of secondary importance to negotiators. In contrast to centuries-old conflict between religious groups in many European countries (the empirical starting point of Lijphart), many African conflict constellations are fairly recent and neither intractable nor frozen, which gives self-declared leaders of supposedly neglected groups enormous leeway to negotiate tailor-made solutions not for group, but personal interests, without consulting their power base. Exceptions confirm the rule: the Hutu–Tutsi divide in Burundi and Rwanda, and the white settler supremacy in South Africa (and also former Rhodesia), may account for a somewhat “frozen” situation with a strong alignment between group leaders and ordinary members of a group. This, in conjunction with a (semi-)permanent institutional architecture, may have led to the relative “success” in lowering inter-communal tensions of both the 1993 pact in South Africa and the 1979 Lancaster House agreement for Rhodesia/Zimbabwe (Spears 2013, 40). In contrast, ethno-religious group boundaries in West Africa’s conflict zones or in Kenya are both more fluid and recent. Only in a “frozen” confrontational situation does it look justified to equate largely the interests of a segment’s elite with that of the entire group.
It is hardly conducive to participatory forms of democracy to let a handful of leaders circumvent majority decisions in the name of power-sharing. The most frequently adopted element of this model are in fact grand coalitions, and probably not by chance. After independence, many African governments grew in size as a consequence of the preferred mode of elite accommodation (Arriola 2009). Oversized governments have indeed been common in many African states; they were not an expression of democracy, but of a specific form of patronage politics. At best, this trend could be perceived as an expression of an inclusionary style of politics, in line with the “hegemonial exchange” model of politics as analyzed by Rothchild (1986). One may be tempted therefore to situate most African power-sharing experiences (i.e., those leading to interim political power-sharing) within a master narrative of the dominance of clientelist politics including a sense of “ethnic arithmetic,” as does Spears (2013, 39). It is, however, worth recalling the genesis of such political systems in postindependence Africa. Historically, they resulted from the—mostly forced—inclusion of smaller parties within unified parties in the 1960s serving the interest of authoritarian leaders. This happened in CAR, Kenya, and Burundi; and the merger of the two main parties in Zimbabwe in 1987 can be interpreted similarly. Only the democratic opening of the early 1990s permitted African polities to become both more competitive and truly participatory; this occurred in all five countries under closer inspection in this contribution.
The risks are that grand coalition governments advocated in the name of power-sharing from the mid-1990s onwards have again the same effects as the dominantly authoritarian political systems established in the majority of African countries in the 1960s. Disturbingly, this trend ignores high demands for a liberally defined democracy on the continent. Consecutive Afrobarometer surveys have made it clear that the popular demand for democracy remains high on the continent, while the supply side—the supply of democracy by elites and institutions—is by contrast weak and rather declining (Gyimah-Boadi 2015; Cheeseman 2015).
Interestingly, African presidents would not find it particularly hard to accept power-sharing deals if this is more about sharing spoils instead of real power. In fact, one may see similar mechanisms at work when aspirants to elite positions were promoted to important positions when proving their capacity to mobilize followers in peace times. Seen from above—from the perspective of a chief of state and ultimate head of all patronage networks—grand coalitions allow for the creation of many well-endowed positions under the president’s oversight and, by the same token, the capacity to silence protest. In a nutshell, this means that consociational power-sharing can add new opportunities for leaders to accommodate their interests and elite critics at the expense of others in society and is therefore detrimental to democracy.
The counterproposal of this so-called centripetalist school of power-sharing (with Horowitz and Reilly as main protagonists; Horowitz 1985; Reilly 2011) seeks to deal with this problem. The aim of this approach is to promote cooperation across ethnic or religious groups via an appropriate design of formal institutions. While general principles like “ethnicity-blind” public policies and inclusiveness within central state institutions serve as underpinnings of such a model, the most concrete and prominent proposal is to alter electoral rules in order to create incentives for coalitions across group boundaries, thereby also reducing the polarization of the party system. Examples on the African continent are rare, but some prominent ones exist. The provisions for presidential elections in Nigeria (after 1978) stipulate that the candidate with most votes is not declared winner in the first round if he does not get at least 25 percent of the votes in two-thirds of Nigeria’s thirty-six states. This provision is a strong incentive to build alliances across regions, ethnic groups, and religious communities, which arguably has saved the most populated country in Africa more than once from collapsing (Horowitz 2014). Arguably, Kenya chose the same option after turmoil in 2008. Kenya’s current electoral system is a bit less demanding, but also provides incentives for presidential candidates to find allies from many locally dominant ethnic groups. This did not spare Kenya from another severe political crisis after the 2017 presidential elections, probably pointing at differences within group constellations. Although the merits of such centripetalist devices seem clear, the model is also elite-centered.
It is rarely disputed that political parties are important actors in established democracies and likewise in peaceful processes of transition to democracy. Political parties, at least moderate parties in the middle of a political spectrum, have a role in organizing majorities. With the advent of violent conflict, the prominence of political parties can dramatically decline as rebel leaders and commanders move center stage. In fact, civilian opposition parties are in an uneasy position before, during, and after peace negotiations, as they are forced to take either a pro-government position (and become negligible) or a pro-rebellion position (and stand accused of being their allies) (Tull and Mehler 2005, 390). In peace negotiations and during the implementation of peace agreements, such parties risk being fully sidelined. One old truth in peace processes is that those not allowed to the negotiation table will not be prominent players in the post-conflict peace order. This can be well observed in the aftermath of peace negotiations and forums in Africa.
In CAR, political parties have lost out both in the country’s recurrent violent crises starting from 1996 onwards and in the different phases of the peace process (Mehler 2011). Early on, the inclusion of some “rebel ministers” was the main objective of international negotiators to end mutinies and rebellions. In a series of peace forums, the civilian opposition became less and less well represented. Concomitantly, over time, established parties lost representational weight in CAR’s national assembly (the growth of independent candidates winning mandates being a main indicator). The parties themselves clearly bear part of the responsibility, as they have not had a convincing record on the “supply” side of democracy (Mehler 2011). However, their sidelining in both peace negotiations and dialogue processes has arguably contributed to this negative trend. The amount of attention that civilian and military heads of state, but also diplomats, have given to the country’s political parties has declined rapidly since 1997, when a series of mutinies ended in a first power-sharing government. To some extent, political parties—all of them with some regional strongholds—were replaced by rebel organizations claiming at times outright ethnic or religious allegiance. This was most obvious with the Séléka rebel alliance that put the defense of Muslim interests on its agenda and later split into mostly ethnic fractions.
The storyline is slightly modified when investigating the fate of Burundi’s first successful Hutu party Front pour la Démocratie au Burundi (FRODEBU). The party had managed to concentrate Hutu votes in the 1993 elections and won both the legislative and the presidential contest by a landslide. When the first freely elected president, Melchior Ndadaye, was killed in a coup attempt only a few months later—together with other high-placed FRODEBU members—this was the start of a new civil war in which the party was quickly outpaced by more radical pro-Hutu organizations. Particularly, Burundi’s Conseil National pour la Défense de la Démocratie – Forces pour la Défense de la Démocratie (CNDD-FDD), a new, even dominant, Hutu party emerged out of a former rebel movement. This effectively sidelined FRODEBU, which clearly had more democratic credentials (see below).
Here the question arises as to why some violent conflicts are treated with temporary, others with permanent, power-sharing institutions (and others by none of them). The vast majority of Africa’s power-sharing experiences are of a temporary nature, mostly ending with the next scheduled national elections. Liberia (2003) and Zimbabwe (2010) are prominent examples. In such cases, one would not expect a long-drawn-out negotiation process with legal experts discussing the eventuality of a specific element within a new constitution having a specific peace-enhancing effect. Rather, the emphasis is on immediate power (contained in positions) and on the spoils coming with such positions. However, there are good reasons why such temporary forms of power-sharing can have problematic effects on democratic principles. First, interim governments permit incumbents to stay in power without facing imminent or regular elections. Second, interim governments can change the rules of the electoral game to a substantial degree while they are in power, thereby facilitating their chances to be elected after an interim phase. Third, being in the limelight might also create an advantage for the main actors within interim governments in the first postwar elections. Advantages of incumbency are therefore extended to actors who may never have acquired popular legitimacy.
Permanent forms of power-sharing come with other challenges to democracy. The excessive focus of negotiations on deactivating the danger of intergroup violence comes with a neglect of other political problems, like the absence of democratic procedures within political parties. Also problematic is the necessity to frame policies in constant respect of fixed group quotas that frequently are a key ingredient of consociational power-sharing. Applying ethnic quotas can affirm group boundaries and generate a considerable obstacle to political change via elections.
Burundi is the outstanding case of a nearly complete consociational power-sharing model enshrined in its constitution of 2005, which in itself mirrored most of an elaborate peace agreement finalized in 2000. Many observers would argue that the salient Hutu–Tutsi juxtaposition responsible for grave political crises in 1972, 1988, 1991, 1994–96, and beyond has subsided as a consequence of the new order. The initial positive assessment by Lemarchand (2007) focuses on many scrupulously observed rules associated with Lijphart’s consociational model. However, it is clear that (Hutu) President Nkurunziza (since 2005) is no less autocratic than (Tutsi) military rulers serving as Burundi’s heads of state before him. It was certainly not the intention of outside mediators that a new dominant party with authoritarian habits would install itself within this framework. Yet this is exactly what happened. Until a May 2018 referendum that altered the constitution substantially, one could have claimed that all the quotas and hurdles to amend the constitution 2 would have precluded the ruling CNDD-FDD from more openly changing the rules of the game. However, the institutions were not a barrier to the personalization of power to the point of approaching an earlier level of autocracy. From a more recent perspective, even the inter-ethnic conviviality appears at risk. New fears that the Tutsi minority could become the scapegoat for antagonistic conflicts within the Hutu camp were also voiced. Horowitz (2014, 10–11) had profiled in more general terms the “degradation problem” and recalls that “majorities prefer majority rule to consociational guarantees.” For Burundi, Curtis (2012, 91) found that power-sharing did not bring about “a fundamental break from the patterns of politics in the past.” One is tempted to conclude that at least Nkurunziza was never convinced of the value to find compromise with his political opponents, regardless of which ethnicity.
As I argued with Tull, power-sharing can also create perverse incentives to engage in violent conflict, instead of engaging in a peaceful competition (Tull and Mehler 2005). Power-sharing issues a bonus on violent behavior; it can be perceived as a shortcut to political power and wealth as long as some quite prominent political entrepreneurs enjoy the rents of power acquired on this pathway. Among the most prominent cases is former prime minister and National Assembly speaker of Côte d’Ivoire Guillaume Soro. Once a student leader 3 sympathizing with Laurent Gbagbo, he ended up in the opposing camp organizing one faction of the Northern-based rebels in 2002, later becoming the Forces Nouvelles leader in 2003. Soro’s political ascension is indeed a source of concern from a democracy theory perspective: his political career has rested on his rhetorical talent as well as his capacity to mobilize violent, not peaceful, support.
Soro, who quickly managed to eliminate his rivals within the rebel sphere, was part of negotiations in the period 2003 to 2005. His oratory skills and mere presence at the negotiation table, often in direct contact with many African heads of states, provided him an aura of importance. He also obtained ministerial positions in the many short-lived cabinets under two compromise prime ministers. In 2007, after a month of intense negotiations in what was called “direct dialogue” between Gbagbo and Soro, the Agreement of Ouagadougou was signed. This represented a much more viable power-sharing arrangement than those preceding. Interestingly, the elite power-sharing deal between President Gbagbo and rebel leader—and by this point prime minister—Soro was not made part of the agreement itself, although it was clearly related and a precondition for securing Soro’s signature. This created a situation whereby Soro was explicitly rewarded for taking up arms. Soro’s career continued after the violent episode in 2011.
Non-permanent forms of power-sharing are especially prone to such effects. Interim institutions may foster the appetite for quick enrichment by a few rebel leaders without any positive repercussions on a potentially neglected minority that continues to be marginalized. Africa offers many examples. In Chad, the two rebel movements Comité de Sursaut National pour la Paix et la Démocratie (CSNPD under Moise Ketté) and its offspring Forces Armées pour la République Fédérale (FARF) were recognized as political parties (in 1994 and 1998 respectively) as a result of only a few days of negotiation. Arguably more important for Ketté was that he enjoyed ministerial rank for about one year before being dismissed, after which he founded a new rebel movement. Ketté was ultimately murdered in 2001. Besides ministerial sinecures for rebel leaders, there was no tangible reward for an entire group that was supposedly represented by the armed movement, in this case southerners in Chad. The formula to end a series of three mutinies in CAR (1996–97) was equally interpreted as largely responding to rent-seeking goals by armed actors (Tull and Mehler 2005, 390–91). Two handpicked mutiny leaders were given ministerial positions in an enlarged government (February 18, 1997); and substantial pensions were handed out to former presidents Dacko and Kolingba, who were at least sympathizing with the mutineers (Mehler 2005, 138). This did not prevent Kolingba from backing a bloody coup attempt in May 2001. Blackmailing sinecures with such military means is certainly not a recipe for peacebuilding. More generally in such cases, a willingness to provide payoffs for insurgent violence has created incentive structures with an appealing option for politico-military entrepreneurs to pursue otherwise blocked political aspirations (Tull and Mehler 2005, 376).
The same mechanism may also hold for post-electoral violence once power-sharing has been established as a mode of conflict “resolution.” At least in media reports it was speculated whether the Kenyan post-electoral crisis in late 2017/early 2018 was provoked by opposition leader Raila Odinga as a ploy to come to a new power-sharing deal that would provide him with similar opportunities to those he had obtained in 2008. Odinga at one point asked for a six-month interim arrangement of governance involving representatives of both parties (Gathara 2018).
By now, few would claim that power-sharing as practiced since the 1990s was successful in promoting peace and democracy in Africa’s plural societies. However, many authors still find merit in the principles of power-sharing and provide general advice on how one should go about it. Those principles might be recapitulated in this way: Only permanent power-sharing may avoid the pitfalls of elite-centrism. A broader effect on group or minority representation can in fact be achieved by permanent power-sharing when more than a handful of political entrepreneurs are part of a deal.
With this said, it is problematic if the merits of majority rule are unable to materialize at all. Power-sharing devices, and particularly permanent forms of power-sharing, prove dangerous. It could be counterproductive for peace when majorities feel abused by overrepresented minorities, which can happen when the group frontiers are cemented and accentuated by power-sharing institutions. It is important to counterbalance the effects of different ingredients of a constitutional order if democracy is to be preserved or even strengthened via constitutional engineering. This means, for territorial aspects, that elements of “self rule” and “shared rule” have to be combined. The centrifugal tendencies inscribed in the principle of segmental autonomy within consociational democracy could be balanced by centripetal effects of other devices. This might include deliberate efforts regarding the electoral system (as described above) or in national lawmaking, for example via a second chamber composed of representatives of provinces, but discussing national interests.
The effectiveness of stopping violence via a power-sharing deal on short notice is rarely disputed. Its pitfalls are recognized only subsequently, although by now the mechanisms are fairly well-known: sidelining of political parties, civil society organizations, and established forms of group representation; and, in parallel, the promotion of violent political entrepreneurs. Therefore not all negotiated rules should be sacrosanct: it makes sense to critically review the effects of power-sharing after two, five, and ten years, and to secure the political will of main stakeholders to modify the institutional devices.
Finally, there is little merit in experimenting with interim power-sharing without an “exit strategy”; potentially a “soft exit formula” needs to be found. The shock effect of terminating power-sharing in favor of a strongly winner-takes-all form of democracy on those now (again) excluded has not only detrimental effects on peace, but also on democracy.
Faced with the rather dismal record of power-sharing in Africa, one may wonder why power-sharing is still a recipe for many crisis situations on the continent. The answer might be twofold: (1) lack of alternatives and a tendency within international organizations to apply blueprint solutions to rather complex and sometimes unique actor constellations; and (2) familiarity of African elites with the key ingredient of power-sharing, i.e., the sharing of spoils within a clientelistic logic of political organization (see also Mehler 2013, 40). There is every reason to believe that power-sharing will therefore play an important role on the African continent in the foreseeable future—even if, admittedly, the 2010s have seen fewer applications of the model in Africa.
An important research agenda can be drawn from here. We have seen that power-sharing has not only created unintended effects, but also that some of them materialized relatively late. Research on power-sharing should therefore focus on the medium- to long-term effects of individual elements of a specific ingredient or institution of the broader formula. Beyond this, the interplay of such different elements of power-sharing is under-researched. How do modifications of the electoral system on a national level interact with decentralization reforms? What are the effects of ethnic quotas within the police on the protection of group rights even at a local level? Does military power-sharing preclude the prosecution of war crimes and is it therefore a disincentive for fighting impunity? There is also a need for both in-depth single-case studies and small-N comparisons to draw out specificities and commonalities. The trajectories of political history differs from country to country and this must be taken into account when judging power-sharing, which is always an ongoing, multidimensional process. On the other hand, as political scientists we have an interest in generalization, but potentially this is more fertile on a lower level of abstraction than has often occurred—with advocates and opponents of consociationalism overstating the advantages of their preferred model. The very concrete effects of institutional choices depend strongly on a given environment, popularity of earlier experiments, and, not least, on the presence or absence of reinforcing or balancing effects from other reforms.
The constitution of 2005 prescribes inter alia that the two vice-presidents are from distinct ethnic backgrounds (and party affiliations), and that parliament consist of 60 percent Hutu and 40 percent Tutsi. Laws have to be voted by a two-thirds majority. In the Senate, a Hutu and a Tutsi must represent each of the seventeen provinces. All parties obtaining more than 5 percent of the votes hold ministerial positions in government; legislative elections are by proportional representation; and the ethnic quota for security forces were fixed at 50:50 (Hutu, Tutsi).
A proposal to amend the constitution could be adopted only with a majority of four-fifths of the members of the National Assembly and two-thirds of the members of the Senate. In the referendum 79.1 percent of the participants officially voted for an amendment that removed such limitations and reduced minority rights.
Soro was President of the Féderation Estudiantine et Scolaire de la Côte d’Ivoire (FESCI) 1995–98. His successor was the equally notorious Charles Blé Goudé who similarly engaged into violent forms of politics, but at the side of Gbagbo.