Power-sharing agreements have become a blueprint for efforts to end violent conflicts in many parts of the world, particularly in Africa. Such agreements, however, rarely include territorial power sharing – at least, not according to the formal, rather unhelpful narrow definition that includes federalism and decentralization. This article argues that the concept of territorial power sharing needs to be broadened in order to account for the manifold informal or indirect manifestations of such arrangements. Drawing on extensive fieldwork data from the DRC, Liberia, and Kenya, the article analyses the history of spatiality and power in Africa in order to explain why formal mechanisms of territorial power sharing are rare and why more subtle types of informal territorial power sharing are much more common. Based on this analysis, we conclude that territorial power sharing is present in many African states, but that typically it is overlooked because of its informal nature.

Power sharing is usually defined as an elite pact between representatives of political or military parties on the division of responsibility in different fields of political and economic life.1 A number of scholars have argued that among the different forms of power sharing – political, territorial, military, and economic2 – territorial power sharing is one of the more powerful mechanisms to secure peace.3 The reasons for and importance of territorial power sharing are clear from a theoretical perspective, as it can reduce the security dilemma of rebels signing a peace agreement (providing a recognized zone of influence outside central government control). Territorial arrangements may also generate ‘group autonomy’, which from a Lijphartian perspective is one of the more effective ways of containing societal divisions.4 These theoretical findings are at odds with current practice, as explicit forms of territorial power sharing are rarely part of peace agreements in Africa. Usually, the parties to conflicts have been more concerned with political power.5 Of the 48 peace agreements concluded in Africa between 1989 and 2006, 20 contained elements of political power sharing, 32 included military power sharing, and just 15 embraced territorial power sharing. By contrast, territorial arrangements were by far the most frequent type of power sharing in both Europe and the dominant type in Asia over the same period.6

Although a lot of research on power sharing has highlighted its effects on the likelihood of conflict, generally this focus has dwelt on the national elite level. Little work has been done on the forms and impact of power sharing at the so-called local level or the periphery of a country. Yet wars originate somewhere, and are fought in very specific places: since violence is rarely ubiquitous, it is just as important to understand the local and the particular as it is to understand the national and the general. Consequently, the focus on elite settlements – in both academia and practice – has obscured many of the mechanisms at play in war and peace processes, especially at the sub-national level. Elements of territorial power sharing included within elite settlements represent a rare example of cases in which places and actors outside the capital are taken seriously and – as we will show throughout this article – directly touch the periphery or local level, whether formally or informally, directly or indirectly.

We have argued elsewhere that the importance of the local level in power sharing tends to be romanticized or overestimated in terms of the impact this can have on national peace.7 Nevertheless, an analysis of local-level power sharing must be a key element in any understanding of the scope of shared power within a post-conflict agreement. The aim of this article is to broaden our conception of territorial power sharing in order to account for the manifold informal or indirect manifestations of such arrangements that current definitions fail to encapsulate. While the ultimate goal of such an undertaking is to pursue a better understanding of the effects of these mechanisms on conflict mitigation, the aim of this article is to tackle the necessary prior conceptual work.

The puzzle this article therefore seeks to explain is the gap between the importance accorded to territorial power sharing in theoretical discussions of conflict mitigation and the seeming absence of such arrangements in African peace agreements. Our argument is that while forms of territorial power sharing are in fact present in Africa, they are typically overlooked because they are informal. We demonstrate this in three ways. First, we describe cases of informal local power sharing – or what we call local power balancing, whereby decision making or political power is relatively evenly distributed between local actors. Second, we consider the indirect territorial effects from the other types of power sharing. For example, as a result of military power division parties can gain informal control over a territory due to their leadership roles in the new army. Finally, we identify several arrangements that allow de facto group autonomy – for example, where a territory is handed over or simply left to local strongmen where the central government is unable to control the area, or needs to appease local actors. This results in different outcomes at different levels of the political system: power is in a sense shared by actors at the national level, but local power is monopolized by one party. Because these local monopolies fail to balance power between the important groups in an area, they often stand in direct contradiction to the spirit of national power sharing, in which certain areas are shared amongst competing governing and rebel elites.8

In making our case, we draw on extensive fieldwork in the Democratic Republic of Congo (DRC), Liberia, and Kenya. Between June 2011 and February 2014 around 200 semi-structured interviews as well as 29 focus group discussions with market women, youth, and teachers were conducted. These groups were chosen in order to understand the everyday effects of power sharing on those not part of the local and national elite we interviewed.

The article proceeds as follows. First, the theoretical framework of territorial power sharing and its conflict mitigation capacity is outlined, primarily with regard to the formal dimensions of decentralization and federalism. In this light we then consider the first case study of the DRC, where we locate a myriad of territorial power-sharing elements that current framings of territorial power sharing fail to encapsulate. Using this example as the basis for further reflection, we look at the history of spatiality and power in Africa in order to explain why outright forms of territorial power sharing are rare and why more subtle forms are much more common. We argue that this is not because territorial power sharing is insignificant in Africa; rather, we point to its presence in manifold different forms that have so far not been identified in these terms.

The conflict mitigation effects of territorial power sharing

Territorial power sharing is generally defined as an elite settlement relating to the division of responsibility between different geographical sub-entities of a national territory. Structurally, territorial power sharing is supposed to foster an inclusive state system or allow for territorial integrity whilst permitting greater autonomy to minorities.9 The forms of territorial power sharing most widely discussed are formal political and administrative decentralization and federalism.10 These arrangements promise – but may not always deliver – two major conflict management outcomes. On the one hand, the grievances of groups towards the state can be appeased through the very costly introduction of federalism. On the other hand, competition between groups for access to resources can be reduced through decentralization.

Let us start with federalism, which regulates group representation by offering a far-reaching concession of self-rule to a set of sub-national states. The introduction of such arrangements represents the overhaul of a political system, enshrining both self-rule and shared rule in a constitution. When federalism is introduced as a conflict mitigation strategy, it is assumed that at least one ethno-linguistic group will profit from the regulation of their own affairs. Federalism is thus said to help with cleavage management (primarily between the state and different groups) and as a tool to prevent separatist conflict.11 However, this presumes ethno-linguistic identity markers that are spatially identifiable – in other words, that certain groups only live in certain places. This is not always the case. There is also the danger that the introduction of federalism creates new marginalized minorities within freshly created units, creating a ‘paradox of federalism’.12 At worst, critics fear federalism might end in secession or partition, which was partly the reason for South Sudan's independence in 2011.13 Moreover, from a conflict theory perspective, federalism is a particularly costly arrangement for leaders to commit to in peace negotiations, as constitutions are much more difficult to abrogate than simple legislative acts.

Decentralization does not usually come at such a high price (or bear the same symbolic importance) for ruling elites as the introduction of federalism. Decentralization is commonly portrayed as the transfer of powers (political, fiscal, or administrative) from central government to lower levels in a political-administrative and territorial hierarchy.14 Decentralization is usually differentiated as devolution, delegation, and deconcentration. Devolution comprises the most expansive form of decentralization, with the creation or increased reliance on sub-national levels of elected government or political autonomy outside direct control of the central government. Delegation is the transfer of managerial responsibility for a specifically defined function away from the central government. Lastly, deconcentration refers to the transfer of administrative authority from central government to regional offices.15 Decentralization – mostly in the form of devolution – has been widely discussed as a tool to empower the poor and marginalized and to improve political legitimacy, accountability, and public sector efficacy.16 Initially promoted as an instrument for development, it is increasingly seen as having a positive impact on political stability and conflict mitigation.

Stability can be enhanced through greater responsiveness and accountability of the state, including a fairer distribution of the state's resources, more citizen participation, and the entry of rebel groups into formal bargaining processes with the government. Nevertheless, only devolution generates power sharing in the true sense of the term. Deconcentration is compatible with the centralization of political power and may just result in better control over the periphery. In any case, decentralization is seldom properly institutionalized and often remains superficial. Ndegwa finds that while nearly all of the 30 African states he surveyed claimed to be decentralized, objectively only a third have functioning decentralized structures.17 A good example of this is Burundi, where decentralization was for a long time rhetorically promoted but, in the beginning, hardly implemented. Today, formally speaking, the state is comprehensively decentralized. However, decentralization is understood mainly in terms of service delivery and development, and not in terms of political emancipation from the central government. Furthermore, decentralized state structures are increasingly used by the ruling party as a control mechanism at the local level.18

Even when decentralization is implemented it is no panacea. Decentralization changes the shape and composition of actor constellations within local arenas (or the spaces accorded to agents, ideas, and structures). This could mean that more room is given to individuals to follow local customs without fear of persecution and that the competition and fear that a winner-takes-all political system can create is mitigated.19 However, decentralization can also increase the likelihood of conflict when more competition is created between local and national power holders, especially when the division of responsibility is unclear and resources are channelled into unaccountable patronage networks.20

In short, both of the formal versions of territorial power sharing – decentralization and federalism – have their own downsides but also have the potential to generate conflict mitigation benefits. These have been widely discussed in the literature. Why then do they so seldom appear in practice when it comes to African power-sharing agreements – at least formally? The case of the DRC serves to illustrate why formal territorial arrangements did not get included in the peace deal, and what happened instead.

Territorial power sharing in the Democratic Republic of Congo (DRC)

The second Congo war (1998–2003) pitted the government in Kinshasa against three main rebel groups. At the end of the war, the country was divided into three main zones of influence, with the government controlling the west and south of the country, the Mouvement pour la Libération du Congo (MLC) controlling the north and two different factions of the Rassemblement Congolais pour la Démocratie (RCD) controlling large parts of the East, notably the province of North Kivu. The Sun City peace agreement of 2003 introduced an extensive power-sharing formula, which mainly focused on political and military power sharing.21

The RCD's proposal of a federal structure met with fierce resistance and, as a result, was not included in the 2003 power-sharing agreement. Opponents, including the government faction and civil society groups, considered the proposition to be the first step in a plan to separate parts of North Kivu from the DRC, referring to it as the ‘balkanization’ of the Congo.22 From the government's point of view, agreeing to the RCD's request would have meant officially handing over control of the periphery to the rebels, thus endangering territorial integrity and potentially the legitimacy of the central government. Because of this, federalism was never a serious consideration and decentralization – though promised throughout the ensuing period – has never materialized.

In the place of federalism, power sharing was supposed to be replicated at the local level, with key positions to be shared amongst the three main actors, thus producing formal local arrangements similar to the “government of national unity” seen at the national level. In reality, however, more often than not one group monopolized power. In North Kivu, the RCD was dominant. Officials sent from President Kabila's faction in Kinshasa were not able to work effectively, as they were denied access to offices and information, or killed. Some of them returned to Kinshasa; others became involved in power struggles with the local RCD elite.23 For the most part, Kinshasa simply accepted the de facto power configuration that emerged in North Kivu.

What evolved in North Kivu was thus a series of local power monopolies – the de facto control of a territory by one group. In this sense, a degree of power sharing emerged between elites at the centre and elites at the local level, even if power sharing at the local level remained ephemeral. This process of informal group autonomy has been repeated several times, with the central government officially holding to the principle of a unified Congolese state while unofficially allowing large parts of the periphery to be controlled by the RCD and the splinter group that emerged later, Congrès national pour la défense du people (CNDP). Anomalies in administrative titles and function demonstrate the informality of this arrangement.

During the war several new administrative positions were created by the rebellion, including positions of chef de poste (or chef d'encadrement administratif) and administrateur du territoire assistant resident (ATAR). After the 2003 agreement, these posts were kept, even though they did not feature in any legal documents.24 Officials retained their positions, earning their living in different ways, including through informal taxation. Most of these office holders are linked to the RCD or the CNDP, and are generally seen as guarantors of rebel power at the local level.25 In addition, only those traditional authorities that sided with the rebellion continued to exercise their power in North Kivu. Therefore, we see a de facto arrangement of territorial power sharing, quite different to the detailed administrative and military power sharing under tripartite leadership agreed upon at the Sun City talks in 2003. The informal arrangement survived and was arguably just as important for the rebels, alongside formal promises to decentralize the country (which never happened).

The Congolese case also illustrates the indirect territorial effect that other forms of power sharing can have in local arenas. One example is military power sharing. When an important post in the army is awarded to a regional leader, it can have the same effect at the local level as territorial power sharing. In the case of the DRC, both the formal plans and the informal implementation of military power sharing eventually had territorial implications.

The Forces Armées de la République Démocratique du Congo (FARDC) was designed to include all the different rebel groups in the DRC. However, it has been fraught with problems since its emergence in 2004. Parallel hierarchies were immediately visible and the loyalty of many troops remained with former rebel commanders.26 In a series of peace agreements and security sector reforms, rebel groups that were integrated into the army were supposed to be sent to regions outside their home areas. But this was never applied to either the RCD or the CNDP. Thus, the former rebels remained present – now in government uniforms – in their stronghold areas, maintaining similar organizational and command structures as well as control of the rich mining areas. As a result, military power sharing had discernible territorial effects, effectively granting the RCD and CNDP a military monopoly over large parts of Eastern DRC, mirroring the local power monopolies described above.

Economic power sharing, of course, also has territorial effects when control of resource-rich areas is shared, which is highly relevant in the DRC. With regard to economic power sharing, the de facto monopolistic control by the RCD (and later CNDP) in parts of North Kivu also gave these groups access to mines and other resources through border control and taxation.27 However, given the fact that these rebel movements by no means represent the majority of people in North Kivu, their power over resources has always been contested. Smaller rebel groups continued their armed fight for control of the area, even after agreements were signed.

In summary, the case of the DRC clearly shows some important informal and indirect dimensions of territorial power sharing that are missed by a strict formal definition. In order to relate this case back to the theoretical literature, we now turn to a brief outline of the historical relationship between space and power in the DRC and elsewhere on the continent.

Spatiality and power in Africa

Long-term demographic trends play an important role in the relationship between the projection of power and spatiality in Africa. Most pre-colonial states in Africa up to the nineteenth century (exceptions include Burundi and Rwanda) were not densely populated. On the contrary, vast stretches of earth were uninhabited. This had enormous consequences for governance in those states, on both ends of the equation. For the population it was comparatively easy to escape from domination.28 For rulers, it meant concentrating their power at the centre. In the absence of good infrastructure (and modern weaponry), power could barely be projected to a distant periphery. Remote zones were only superficially integrated by less intense exchanges, tribute payment, or loose alliances with smaller, “proto-state” structures. In such a context, political and economic power was mostly about controlling people (or securing their allegiance) and not about controlling territory – in stark contrast to other continents during that same phase of world history.

The colonial state brought demarcated borders and notions of territorially bounded spheres of power. This created a pretension by the state to claim control over its entire territorial periphery. In practice, the colonial economy only brought a valorization of land in fertile zones of “useful Africa” (l'Afrique utile), mostly close to the coast or in a few administrative centres, which meant that landlocked and sparsely populated peripheries were only administered by indirect rule. Nevertheless, the control of land in addition to people became increasingly important.

With independence, the claim of territorial control (and being “sovereign”) was upheld by new state elites, even though frequently it was a hollow rhetoric.29 Some differentiation between spatially bound models of authority began to set in, with smaller states facing lesser challenges in controlling their territory than bigger ones. Governments in countries such as the Central African Republic, Sudan, and the DRC were never able to control their entire territory, but this had no major consequences for their official recognition as sovereign states by the international community. Nor did it affect the capacity of elites to syphon off the rents of sovereignty – income generated from taxes, licences, royalties, development assistance, and so on. Overall, a contradiction can be seen here. On the one hand, the legitimacy of central governments – both internationally and domestically – is based on the claim to territorial control and sovereignty. On the other hand, central governments often disregard the periphery, which results in considerable leeway for local actors and thus undercuts the claim to territorial control made by the state in question.

More recently, new patterns of spatiality and power have emerged in Africa, shaped by several factors such as tremendous population growth throughout the continent and large-scale investments in land by national and foreign private investors.30 Moreover, discourses around the “politics of belonging” and the “land of the ancestors”31 have become prominent. Controlling land has become a source of power and conflict.32

Several conclusions can be derived from this review and applied to the case of the DRC.33 First, despite the growing trend towards the expansion of territorial authority, the periphery is only superficially and indirectly integrated, giving sub-state authorities considerable leeway in local power relations. Second, this leeway undermines the efforts of central government to broadcast sovereignty over an entire territory, and to use this as the way to legitimize the state. As a result, there is a tendency for African states to allow the periphery as much leeway as possible as long as this does not question their sovereignty.

Empirically, both central governments and peripheral elites tend to have more to gain from the preservation of the existing state than the disintegration of the national territory.34 The RCD may have argued for federalism in the DRC, but never secession. While the demand for federalism was never met, informal de facto monopoly arrangements unfolded after the 2003 Sun City agreement. The state was willing to “give up” peripheral parts of North Kivu, an area the central government had never managed to provide with public services or to control (or been interested in controlling). This informal arrangement ensures that territorial integrity is maintained.

It is also important to note that the eastern parts of DRC are of immense interest to the various rebel movements – RCD, CNDP, and most recently M23 – that are backed by the DRC's smaller and resource-poor neighbour, Rwanda. It is thus no surprise that informal de facto arrangements have been made, benefiting the rebels while allowing Kinshasa to save face. Reflecting the broader trend across Africa described above, the central government was never really able to control the east. As a result, the government did not so much choose these informal arrangements as come to terms with an existing reality. The fact that federalism was not formally included in the power-sharing agreement shows that Kinshasa had some leverage during the peace talks, and considered de facto autonomy to be a preferable compromise because it enabled the government to maintain its claim to sovereignty.

Nevertheless, recent military intervention shows that the Congolese state remains interested in the periphery. The reasons for this relate to the valuable resources located in the area, and the fact that increased population density makes the region an increasingly important voter pool. Thus while territorial control is not necessarily a goal of central government, it may become one depending on the resources and population numbers in the periphery. In response to the M23 rebellion, the centre has become increasingly interested in controlling the periphery. It remains to be seen whether this new approach from the Congolese government will bear fruit or whether at some point in the future another set of de facto autonomy arrangements with strong rebel groups in the Kivus will be established.

But while the DRC reflects a number of features common to other African states, it is also a particularly challenging environment when it comes to political control and governability. Jeffrey Herbst has identified geo-demographic factors through which one can assess “objective” difficulties facing the government. In essence, these relate to population distribution and the sheer size of the territory. From this perspective, the DRC stands out as ‘perhaps the classic case of an extremely challenging population distribution’.35 Whilst Herbst argues for such a differentiation in terms of ‘governability’, we argue that instances of territorial power sharing are better explained by historic variations in centre–periphery relations. In order to demonstrate this point, we turn to the cases of Liberia and Kenya.

Territorial power sharing in comparative perspective

Tiny Liberia is at the opposite end of the spectrum to the DRC in terms of Herbst's classification of governability. Yet if we compare their historical power-spatial relations there are surprising similarities. Liberia, small and relatively sparsely populated, is a state in which until very recently the periphery or the “hinterland” (except some well-defined mining enclaves) never counted in the eyes of the capital's elites. Instead, a culturally fairly distinct coastal elite (the Americo-Liberians descending from resettled slaves) kept its distance from the hinterlands and limited exchanges between capital and periphery to a strict minimum – the revolt of the periphery occurred very late.36 The one exception to this rule has been patronage ties, which (according to Amos Sawyer) qualify modern Liberia as a ‘system of indirect rule’ in which citizens are consistently differentiated from subjects.37 The Liberian case suggests that historical patterns of centre–periphery relations can trump the influence of territorial shape and demographic size.

Kenya's story is very different. A medium-sized country, it is more difficult in terms of governability than Liberia, but nowhere near as difficult as the DRC. Nonetheless, over the last century comparatively large parts of the country were of economic interest and saw the extension of state infrastructure, and hence Daniel Branch and Nic Cheeseman qualify post-colonial Kenya ‘as an outlier among sub-Saharan African states because of its strength in both urban and rural areas’.38 The colonial state and elites in the independent state aimed at maximizing control over the periphery; they engaged in ‘assimilation réciproque’ in the sense described by Jean-François Bayart.39 Kenya was not alone. States such as Burundi and Rwanda had a highly centralized political system in pre-colonial times, leaving little autonomy for local strongmen and for cultural difference in the periphery.

It is therefore clear that the context for centre–periphery relations, including the opportunity and incentives for the exertion of central control and the legacy of patronage and other linkages, varies considerably between our cases. But how does this impact on power-sharing practices? In the small country of Liberia, power has been concentrated in the capital for most of its existence. Controlling the periphery was never important, despite the widespread availability of natural resources in the so-called “hinterland”. Nevertheless, in contrast to the DRC, Liberia has always been relatively easy to govern, even without significant government presence outside of Monrovia. Unsurprisingly, therefore, no formal elements of territorial power sharing were foreseen in the Comprehensive Peace Agreement that was signed in 2003 by three conflict parties–the (then) Taylor government of Liberia, Liberians United for Reconciliation and Democracy (LURD), and the Movement for Reconciliation and Democracy (MODEL).

The power-sharing agreement was the fifteenth agreement signed to end a civil war that took place in two stages; the first began in 1989 and continued until 1997, while the second started in 1999 when LURD, and later MODEL, tried to oust Taylor from government. Although the figures are disputed, during the fourteen years of intermittent war an estimated 200,000 people were killed and 1.5 million people were displaced, either internally or as refugees.40 The formation of a ‘Governance Reform Commission’ (Article XVI) was included in the peace agreement, though a decade later proposed legislation on decentralization (the Local Government Bill) – which required significant changes to the constitution – has still not been debated in Parliament, let alone put into law. Power sharing in Liberia was never supposed to be replicated at the sub-national level, but was reserved for central government. The absence of formal mechanisms of territorial power sharing, whether direct or indirect, raises the question of whether or not it can be assumed that the local elites of warring groups in former hotspots of violent conflict have agreed to share power when a national settlement has been signed. In the Liberian case, the answer has often been no.

The purpose and reasoning behind centralized peace negotiations in a small country like Liberia was to foster peace in both national and sub-national arenas. However, wars involve reconfiguring power relations in a number of different spaces, not least because forced displacement has brought a different ethnic composition to a number of localities. As a result, informal power-sharing arrangements at the local level that allow for a degree of inclusion based on the status quo will risk excluding the losers of the conflict. In turn, this would cement a distribution of power positions that rewards those who have taken territory and ignores others, and is consequently regarded as unfair by at least one party. The effects of such exclusion, which can so easily result in the creation of an informal power monopoly, can be observed in the Liberian case, where a lack of power sharing has arguably exacerbated inter-group tensions since the end of the war.

In the border town of Ganta, situated in north-eastern Liberia, land conflict between the minority group of Mandingos and the majority groups of Manos and Gios continues to simmer.41 From the perspective of the majority groups, the Mandingos have never gained rightful land titles to currently disputed land; they are said to have become “enemies” when the rebel group LURD (predominantly Mandingo) destroyed the town in 2003. Conversely, Mandingo leaders claim that they are not represented on any local government bodies, and as a result feel threatened. This stands in contrast to other towns like Gbarnga, in central Liberia, where informal efforts have been made to ensure Mandingo representation – at least partly because there are fewer disputed properties, making for a less polarized situation. More recently, government intervention has reduced land conflict in Ganta through compensation payments, but in the immediate period after the war the situation was far more difficult.

The first mayor in post-war Ganta, appointed by the notorious General Peanut Butter who later became Senator Adolphus Dolo, allegedly exacerbated land conflict by distributing properties to Mano and Gio civilians and ex-combatants.42 Thus, members of the Mandingo community not only felt that they were not properly represented, but also that their former enemies held a power monopoly in the town, unchallenged by the central government in Monrovia. In this sense, an emerging local monopoly of power led to an increase (or continuation) of conflict in post-war Liberia, though one that was by no means as violent as during civil war.

The absence of local power sharing in the context of national power sharing highlights two key points. The first is a path dependency in centre–periphery relations: despite being small and therefore more easily governable, Liberian elites at the centre have not sought to extend effective control over the hinterland. The second is the significance of informal agency in the periphery.

A different process of historical development can be seen in Kenya, which, as a medium-sized country, has demonstrated a tendency towards slightly more formalized elements of territorial power sharing. In late 2007, after disputed election results were announced in favour of the Kikuyu candidate Mwai Kibaki of the Party of National Unity (PNU) – as opposed to earlier predictions that suggested victory for the Orange Democratic Movement (ODM) candidate Raila Odinga (a Luo) – a wave of violence erupted across the country, most notably in parts of the Rift Valley. Over 1,000 people died and up to 600,000 were displaced, bringing Kenya to the ‘brink of civil war’.43 Whilst formal territorial power sharing was not included in the Kenya National Dialogue and Reconciliation, which was signed in 2008 by Kibaki and Odinga, it played an indirect part through what became known as ‘Agenda 4’. This Agenda was part of the National Dialogue and specifically sought to address long-term issues and solutions. A new constitution followed, promulgated in 2010, which introduced new elements of administrative and political devolution. In March 2013, 47 counties came into formal existence, as governors, senators, and county assemblies were elected for the first time.

Whilst these changes are just starting to be implemented, their mere existence has provoked mixed reactions, some very emotionally charged. On the one hand, in parts of the Rift Valley it is believed that devolution will benefit the local majority groups, although official discourse puts this in a slightly different way, simply claiming that devolution will lead to a fairer distribution of access to security, justice, and wealth. On the other hand, however, devolution is linked to the notorious idea of majimboism (stemming from the word for regionalism in Swahili), which is now perceived to have connotations of ethnic cleansing among some Kenyan communities as a result of President Daniel Arap Moi's efforts to use the notion of majimbo to legitimize a campaign of intimidation and state-sponsored violence in the 1990s. As a result, Kikuyus in places like the Rift Valley fear that they may have to leave the area as a result of devolution, although they have nowhere to go.

Fears that certain groups will lose out in this process have resulted in some interesting conflict management techniques. In order to avoid further conflict, the Council of Elders in the Rift Valley began brokering ‘power-sharing’ agreements in 2011, helping to decide who should run for which seat during elections in order to keep local peace. More generally, the National Commission on Cohesion and Integration (NCIC) led peace dialogues all over the country. In the Rift Valley this culminated in a peace agreement signed by the Council of Elders in August 2012, promising to refrain from violence in the foreseeable future.44 Nevertheless, in many places across Kenya such arrangements were made not primarily for conflict management purposes, but rather to ensure that presidential coalition partners do not undermine each other. Unsurprisingly therefore, most governors are members of the existing political elite, having used their networks and patronage to win office.45

Kenya's experiment with devolution clearly has consequences for conflict mitigation, but was not specifically mentioned in the power-sharing part of the 2008 peace deal and so does not count as territorial power sharing in most academic analyses. Nonetheless, devolution as a consequence of constitutional reform and restructuring, directly flowing out of Agenda 4, shapes centre–periphery relations just as much as the local distribution of power.

Kenya has also seen the emergence of other informal power-sharing arrangements that range beyond most current definitions of territorial power sharing. In the heterogeneous city of Nakuru in the Rift Valley in Kenya, one of the hotspots during the post-election violence, local politicians took it upon themselves to replicate the national power sharing in the town council. The mayor is elected by the councillors, and the most recent mayor (prior to the 2013 elections) was from the ODM, while his deputy mayor (whom the mayor appoints) was from the opposing PNU. Prior mayors from the PNU party were likewise sure to appoint deputy-mayors from the ODM. Nakuru was the first town to have such a PNU-ODM mayoral team after the post-election violence. The eight committees in the town council are set up in a similar manner. As such there was an informal replication of power balancing at the local level. It would be too much, however, to assume that this resulted in conflict mitigation. Appointments are made along lines of personal interest, not so much political power. In addition, the Nakuru town council is infamous for infighting across and within all political divides.46

The Kenyan case shows that even a state with a history of greater control over the periphery may take on forms of political devolution, both formal and informal. With few immediate tangible effects of formal devolution as prescribed by the new constitution, it was significant that informally instances of power balancing emerged quickly in places like Nakuru. It is also significant that they had the opposite effect to the kind of local power monopolies that developed in the Kivus. While this is just one example, and others will beg to differ, it supports the notion that Kenyan centre–periphery relations remain as strong as they have been historically: national politics continues to affect the local level, a point that is well illustrated by the way in which local elites emulate national solutions.47

Given that decentralization was never implemented in the DRC and Liberia, the implementation of formal devolution in Kenya is equally striking. At a recent meeting of the Governance Commission in Liberia, a civil society expert mentioned a research trip to Kenya in order ‘to find out how to do decentralization’.48

Conclusion: towards a new understanding of territorial power sharing

The preceding historical contextualization provides a better basis for understanding how relations of space, territory, and power have developed in Africa over the years. Considering the fact that the peripheries of many countries, especially in vast counties such as the DRC, have long been ignored, it comes as no surprise that formal territorial power sharing is rarely incorporated into African agreements. The juxtaposition of national sovereignty, on the one hand, and a lack of territorial control on the other hand also underlines the contradictory nature of formal and de facto reality. Thus our cases (see Table 1 below), all have varied histories of control over space, and feature different forms of informal or indirect territorial power sharing.

Table 1

Historical power–space nexuses and territorial power sharing

 Liberia Kenya DR Congo 
‘Governability’ Easy Medium Difficult 
History of centre–periphery relations Most parts of periphery not controlled Strong control over the periphery (except northern margins) Most parts of periphery not controlled 
Indirect or informal territorial power sharing after peace agreement Examples of informal power sharing resulting in power monopolies at the local level Examples of informal power sharing resulting in power balancing at the local level Examples of informal power sharing resulting in power monopolies at the local level; indirect effect on territorial power sharing from military and administrative power sharing 
Relation of informal and indirect dimensions to the national arrangements In contrast to the national agreement Largely mirrors the national agreement In contrast to the national agreement 
Decentralization Decentralization still being planned Decentralization currently being implemented Federalism proposed but rejected, decentralization partly reversed 
 Liberia Kenya DR Congo 
‘Governability’ Easy Medium Difficult 
History of centre–periphery relations Most parts of periphery not controlled Strong control over the periphery (except northern margins) Most parts of periphery not controlled 
Indirect or informal territorial power sharing after peace agreement Examples of informal power sharing resulting in power monopolies at the local level Examples of informal power sharing resulting in power balancing at the local level Examples of informal power sharing resulting in power monopolies at the local level; indirect effect on territorial power sharing from military and administrative power sharing 
Relation of informal and indirect dimensions to the national arrangements In contrast to the national agreement Largely mirrors the national agreement In contrast to the national agreement 
Decentralization Decentralization still being planned Decentralization currently being implemented Federalism proposed but rejected, decentralization partly reversed 

Herbst's classification of states in terms of their ‘governability’ – size and population density – does not help us to explain these informal or indirect dimensions. Rather, the pattern suggested by our cases is that the more the periphery is ignored, or has been historically, the more informal. Elements of territorial power sharing emerge. This shows a form of path dependency, with the most extreme example being the DRC. While historical power relations may be shaped by questions of ‘governability’, they are also shaped by the political choices that are made. As a result, we find unlikely similarities between the DRC and Liberia, which turn out to be places where monopolies evolved locally despite the national power-sharing agreement. In parts of North Kivu a de facto autonomy arrangement that represented a local power monopoly for the RDC emerged, and was at best accepted by the central government for a very long time, until the most recent interventions. Liberia, similarly, is a country where the periphery has never been fully controlled, and equally the post-agreement period resulted in an informal power monopoly. In this instance, however, the monopoly was much weaker, more sporadic, and short-lived.

Though the central government of Liberia has little control over the periphery, this is not only because of governance factors but also by choice. This of course underscores the dialectical nature of structure and agency, and the way in which they impact on one another over time.49 It would have been easy to intervene in the local power constellations in Ganta (and indeed the president replaced the mayor in 2008, albeit with a former NPFL leader) but for various reasons central elites have not chosen to do this. It is in these two countries that the post-power-sharing projects of decentralization remain the furthest away from implementation.

In the case of Kenya, which has a history of much stronger control over the periphery, decentralization projects are much further advanced and the repercussions at the local level are mostly the result of national policies such as centrally imposed ethnic power sharing. Thus, arrangements, whether informal or indirect, largely mirror centralized decisions or agreements. As the case of Nakuru reveals, the informal dimension of territorial power sharing may take the form of power balancing, rather than the sort of de facto autonomy arrangements or local power monopolies that have been seen in the other cases. Again, this makes it clear that “governability” considerations did not determine outcomes when it came to territorial power sharing.

It is true that the larger countries, the DRC and Kenya, made more concrete moves towards decentralization, but the lack of implementation in the former case reveals that this was more political theatre than reality. In cases where national sovereignty tends to be bound up with territorial control, it is likely that more formalized elements of territorial power sharing, or stronger implementation of decentralization, can be found. We expect this to be mostly the case, however, in countries with historically high degrees of integration, and where politically relevant identity groups are rather evenly spread geographically. In other cases where territorially small states have less control over the periphery, such as Liberia, more informal and contradictory arrangements have emerged. Nevertheless, we suggest that size does not matter nearly as much as historical patterns of state control.

These cases of power-sharing peace agreements illustrate that although decentralization and federalism are rarely even proposed, let alone implemented, in African countries, this does not mean that there are no elements of territorial power sharing. On the contrary, informal and indirect cases of territorial power sharing can be found in all of the countries discussed in this article. Furthermore, the historical link between space and power offers a reason why this should be so. While governments have long concentrated on power over people, control over land – particularly in resource-rich zones – has steadily become more important in Africa. The current paradox of formal national sovereignty and a lack of real territorial control in many African countries looks ever more untenable and suggests that we need to search for more effective, and locally sensitive, solutions. Meeting this challenge will require us to look harder at the informal dimensions of territorial power sharing.

Andreas Mehler, ‘Power sharing’, in Nic Cheeseman, David Anderson, and Andrea Scheibler (eds), Routledge handbook of African politics (Routledge, Oxford, 2013), pp. 189–201. Whilst the discussion on power sharing started out as a debate on democratic system building in ‘divided societies’ – for example, Arendt Lijphart, Democracy in plural societies (Yale University Press, New Haven, CT, 1977) – more recently it has focused on peace building, where power sharing is used as a tool for the resolution of conflict or its prevention in the future. However, the verdict on its usefulness is still to be delivered. See Rene Lemarchand, ‘Consociationalism and power sharing in Africa: Rwanda, Burundi, and the Democratic Republic of the Congo’, African Affairs 106, 422 (2007), pp. 1–20; Denis M. Tull and Andreas Mehler, ‘The hidden costs of power sharing: reproducing insurgent violence in Africa’, African Affairs 104, 416 (2005), pp. 375–98; Kathrin Heitz, ‘Power sharing in the local arena: Man – a rebel-held town in western Côte d'Ivoire’, Africa Spectrum 44, 3 (2009), pp. 109–31.
Caroline Hartzell and Matthew Hoddie, ‘Institutionalizing peace: Power sharing and post civil war conflict management’, American Journal of Political Science 47, 2 (2003), pp. 318–32; Caroline Hartzell and Matthew Hoddie, Crafting peace: Power-sharing institutions and the negotiated settlement of civil wars (Pennsylvania State University Press, Pennsylvania, PA, 2007).
Matthew Hoddie and Caroline Hartzell ‘Power sharing in peace settlements: Initiating the transition from civil war’, in Philip G. Roeder and Donald Rothchild (eds), Sustainable peace: Power and democracy after civil wars (Cornell University Press, Ithaca, NY, 2005), pp. 83–106. See also Anna K. Jarstad and Desirée Nilsson, ‘From words to deeds: The implementation of power-sharing pacts in peace accords’, Conflict Management and Peace Science 25, 3 (2008), pp. 206–23.
See, for example, Barbara F. Walter, Committing to peace: The successful settlement of civil wars (Princeton University Press, Princeton, NJ, 2002); Lijphart, Democracy in plural societies.
Walter, Committing to Peace, pp. 30–1; see also Donald Rothchild, ‘Power dividing as an alternative to power sharing’, in Philip G. Roeder and Donald Rothchild (eds), Sustainable peace: Power and democracy after civil wars (Cornell University Press, Ithaca, NY, 2005), pp. 51–82; Hoddie and Hartzell, ‘Power sharing in peace settlements’; see also Christof Hartmann, ‘Territorial power sharing and the regulation of conflict in Africa’, Civil Wars 15, 1 (2013), pp. 123–43.
Martin Ottman and Johannes Vüllers, ‘The power-sharing event dataset (PSED): A new dataset on the promises and practices of power sharing in post-conflict countries’, Conflict Management and Peace Science, forthcoming.
Claudia Simons, Franzisca Zanker, Andreas Mehler, and Denis Tull, ‘Power sharing in Africa's war zones: how important is the local level?’, Journal of Modern African Studies 51, 4 (2013), pp. 681–706.
See also Simons et al., ‘Power sharing in Africa's war zones’, p. 686.
David A. Lake and Donald Rothchild, ‘Territorial decentralization and civil war settlements’, in Philip G. Roeder and Donald Rothchild (eds), Sustainable peace: Power and democracy after civil wars (Cornell University Press, Ithaca, NY, 2005), pp. 109–32, p. 120; Hartzell and Hoddie, Crafting peace, p. 34.
See, for example, Lake and Rothchild, Territorial decentralization, or Hartmann, ‘Territorial power sharing’.
Nancy Bermeo, ‘Conclusion: The merits of federalism’, in Ugo M. Amoretti and Nancy Bermeo (eds), Federalism and territorial cleavages (Johns Hopkins University Press, Baltimore, MD, 2004), pp. 457–82.
Jan Erk and Lawrence Anderson, ‘The paradox of federalism: does self-rule accommodate or exacerbate ethnic divisions?’, Regional and Federal Studies 19, 2 (2009), pp. 191–202.
Ibid., p.195; Donald L. Horowitz, Ethnic groups in conflict (University of California Press, Berkeley, CA, 2001), p. 638; see also Roeder's ‘segmental institution thesis’ in Where nation-states come from: institutional change in the age of nationalism (Princeton University Press, Princeton, NJ, 2007).
Richard C. Crook and James Manor, Democracy and decentralisation in South Asia and West Africa: Participation, accountability and performance (Cambridge University Press, Cambridge,1998); see also Hartzell and Hoddie, Crafting peace, p. 33; Pippa Norris, Driving democracy: Do power-sharing institutions work? (Cambridge University Press, Cambridge, 2008), pp. 164–6.
See Dele Olowu and James S. Wunsch, Local governance in Africa: The challenges of democratic decentralization (Lynne Rienner Publishers, Boulder, CO, 2004).
Gordon Crawford and Christof Hartmann, Decentralisation in Africa: A pathway out of poverty and conflict? (Amsterdam University Press, Amsterdam, 2008); Joseph Siegle and Patrick O'Mahony, Assessing the merits of decentralization as a conflict mitigation strategy (DAI Publications, Bethesda, 2009); Lake and Rothchild, Territorial decentralization.
Stephen Ndegwa, ‘Decentralization in Africa: A stocktaking survey’ (Africa Region Working Paper Series No. 40, World Bank, Washington DC, 2002); see also Horowitz, Ethnic groups in conflict, p. 623 on aborted plans for decentralization.
For a general argument that liberal peace building has not changed the character of control-oriented politics in Burundi significantly, see Devon Curtis, ‘The international peacebuilding paradox: Power sharing and post-conflict governance in Burundi’, African Affairs 112, 446 (2013), pp. 72–91.
Siegle and O'Mahony, Assessing the merits of decentralization, pp. 1–5.
Crawford and Hartmann, Decentralisation in Africa, pp. 14–16.
For details of the Congolese civil wars and the negotiation process see Gerard Prunier, Africa's World War: Congo, the Rwandan genocide, and the making of a continental catastrophe (Oxford University Press, Oxford, 2011).
Interview, founding member of the RCD and MP in the transition parliament, Goma, DRC, 29 August 2011.
Interview, radio journalist, Goma, DRC, 14 July 2011.
Interview, civil society member, Sake, DRC, 23 July 2011.
Interview, former Mai Mai fighter and local politician, Sake, DRC, 24 August 2011.
This notoriously came to light when several ex-RCD officers who shared military power with officers from the Kabila faction in North and South Kivu attacked South Kivu's capital Bukavu in 2004 and a war between the different factions within the “integrated” army broke out. For a short description of the case see, for example, Denis Tull, ‘Peacekeeping in the Democratic Republic of Congo: Waging peace and fighting war’, International Peacekeeping 16, 2 (2009), pp. 215–30. See also Maria Eriksson Baaz and Judith Verweijen, ‘The volatility of a half-cooked bouillabaisse: Rebel–military integration and conflict dynamics in the eastern DRC’, African Affairs 112, 449 (2013), pp. 563–82.
Other resource-rich areas like Katanga were in turn under the exclusive control of the pro-Kabila factions.
Albert O. Hirschman, ‘“Exit, voice, and loyalty”: Further reflections and a survey of recent contributions’, Social Science Information 13, 1 (1974), pp. 7–26; see also Igor Kopytoff, The African frontier: The reproduction of traditional African societies (Indiana University Press, Bloomington, IN, 1987); Jeffrey Herbst, States and power in Africa: Comparative lessons in authority and control (Princeton University Press, Princeton, NJ, 2000).
Robert H. Jackson and Carl G. Rosberg, Personal rule in black Africa: Prince, autocrat, prophet, tyrant (Cambridge University Press, Cambridge, 1982).
See Lorenzo Cotula, The great African land grab? Agricultural investments and the global food system (Zed Books, London, 2013).
Peter Geschiere and Francis B. Nyamnjoh, ‘Capitalism and autochthony: the seesaw of mobility and belonging’, Public Culture 12, 2 (2000), pp. 423–52.
Pierre Englebert, Africa: Unity, sovereignty, and sorrow (Lynne Rienner Publishers, Boulder, CO, 2009).
Our focus is on geographical space, but for more on social space and the spatial turn see Ulf Engel and Paul Nugent, Respacing Africa (Brill, Leiden, 2010).
Englebert, Africa: Unity, sovereignty, and sorrow.
Herbst, States and power in Africa, p. 146.
Angola's domination by Luanda and the Mestiços with minimalist Portuguese penetration of the hinterlands may be a similar case.
Amos Sawyer, Beyond plunder: Toward democratic governance in Liberia (Lynne Rienner Publishers, Boulder, CO, 2005), p. 185.
Daniel Branch and Nic Cheeseman, ‘The politics of control in Kenya: Understanding the bureaucratic-executive state, 1952–78’, Review of African Political Economy 33, 107 (2006) pp. 11–31, p. 12.
Jean-François Bayart, L'etat en Afrique: La politique du ventre (Fayard, Paris, 1993), p. 200.
For details on the Liberian civil wars and the negotiation process see Morten Bøås, ‘The Liberian civil war: New war/old war?’, Global Society 19, 1 (2005), pp. 73–88; Desirée Nilsson, Crafting a secure peace: Evaluating Liberia's Comprehensive Peace Agreement 2003 (Uppsala University, Uppsala, 2009); Sawyer, Beyond plunder.
Bøås, ‘The Liberian civil war’, p. 83.
See also Jairo Munive, ‘Ex-combatants, returnees, land and conflict in Liberia’ (Working Paper, Danish Institute for International Studies, 2010), p. 17.
For more details on the post-election violence in 2007 and the negotiations and agreement that followed see David Anderson and Emma Lochery, ‘Violence and exodus in Kenya's Rift Valley, 2008: Predictable and preventable?’, Journal of Eastern African Studies 2, 2 (2008), pp. 328–43; Nic Cheeseman, ‘The Kenyan elections of 2007: an introduction’, Journal of Eastern African Studies 2, 2 (2008), pp. 166–84; Karuti Kanyinga, ‘The legacy of the White Highlands: land rights, ethnicity and the post-2007 election violence in Kenya’, Journal of Contemporary African Studies 27, 3 (2009), pp. 325–44.
Centre for Humanitarian Dialogue, ‘Kenya Rift Valley’, August 2012, <http://www.hdcentre.org/en/our-work/peacemaking/past-activities/kenya-rift-valley/> (23 July 2014).
Agnes Cornell and Michele D'Arcy, ‘Plus ça change? County-level politics in Kenya after devolution’, Journal of Eastern African Studies 8, 1 (2014), pp. 173–91.
NTV, ‘Nakuru councilors in shocking fistfight’, August 2012, <http://ntv.nation.co.ke/news2/topheadlines/nakuru-councilors-in-shocking-fistfight> (23 July 2014).
For the reiteration of this pattern after the gubernatorial elections, see Cornell and D'Arcy, ‘Plus ça change’.
Workshop at the Governance Commission, ‘Briefing Session with CSOs on Local Government’, see below, Monrovia, 15 January 2014.
For a similar argument regarding local-level effects of Gacaca Courts see Phil Clark, ‘Bringing the peasants back in again: State power and local agency in Rwanda's gacaca courts’, Journal of Eastern African Studies 8, 12 (2014), pp. 193–213.

Author notes

Franzisca Zanker is a Research Fellow at the GIGA Institute of African Affairs; Claudia Simons (Claudia.Simons@swp-berlin.org) is a Research Fellow at the German Institute for International and Security Affairs; Andreas Mehler (mehler@giga-hamburg.de) is the Director of the GIGA Institute of African Affairs. This research was funded by the German Research Foundation (DFG) within the Priority Programme 1448 ‘Adaptation and Creativity in Africa’. We thank Anders Sjörgen, Martin Ottman, Johannes Vüllers, Denis Tull, the editors of African Affairs, and two anonymous reviewers for their comments on earlier versions. All errors remain our own.