Federalism and Decentralization in Sub-Saharan Africa: Five Patterns of Evolution

Abstract

The 1990s were marked by democratic reforms throughout Sub-Saharan Africa. This went in tandem with decentralization reforms which either created or strengthened subnational levels of government. More than twenty years later it seems everywhere to the south of the Sahara there is a gap between the institutional/constitutional blueprints introducing the reforms and the facts on the ground. Understanding and explaining this gap in the workings of federalism and decentralization is important to both theorists and practitioners. This article proposes five benchmarks in order to map out the evolutionary patterns of the last two decades: a) symmetrical recentralization; b) differentiated performance; c) legitimizing traditional authority structures and indigenous conflict resolution; d) politicization of local conflicts over land, water, and other natural resources; and e) federal extinction.

Key Words:

• Federalism
• decentralization
• democratic reform
• institutions
• constitutions
• Africa

Ex Africa semper aliquid novi

(Always something new out of Africa)

In his Naturalis Historia, Roman scholar Pliny the Elder (23-79 AD) writes about the new animal and plant species that are constantly being discovered in Africa.¹ His dictum above has since become an oft-used phrase for natural scientists finding new things in Africa. There is now a social science validation of Pliny's statement. The new species coming out of Africa is the suddenly-and-massively-decentralizing-state. Its first appearance is with the large-scale post-conflict federal reforms in South Africa, Ethiopia, and Nigeria two decades ago, but this species is also found throughout the continent; all states south of the Sahara have adopted various political reforms toward decentralization. Some of the first territorial decentralization programs on the continent in fact precede the post-conflict federal reforms, and can be traced back to the late 1980s.² In a decade or so, the entire Sub-Saharan Africa almost in unison has become decentralized.

The suddenly-and-massively-decentralizing-state is defined by its size and scope: decentralization is far-reaching and comprehensive; it has been designed top-down at the centre, has arrived with a big-bang, and was put in place at unprecedented speed; it is often carried out with strong international (both political and financial) support; the new species promises better democracy, policy efficiency, development, growth; but twenty years after its introduction, it loses some of its sheen, and mutates in unexpected directions. The Sub-Saharan species of the decentralized state is certainly there to stay, but not in the form codified in constitutions. It is therefore important to try to map out how and why it has evolved since its introduction to the political eco-system.

While carrying with it inevitable signs of fine-tuning in order to better fit with the specific political, social, demographic and geographical context, not every mutation of the Sub-Saharan decentralized state is sui generis. There is indeed a pattern of evolution that applies to the mutation of all of the sub-species of the suddenly-and-massively-decentralizing-state. Long-term structural uncodified factors – be it societal, economic, geographic, demographic – interact with a handful of constitutional clauses introduced during the sudden moments of top-down large-scale institutional change. Put simply, on its own the original institutional/constitutional blueprint is not helpful to understand and explain how the decentralized state works in practice two decades later. Despite sharing the core characteristic of the species whereby uncodified structural factors influence the workings of constitutions, the pattern of evolution reveals itself under slightly different guises. In one way or another, all chapters to the collection share the desire to expose what happens once decentralization reforms are adopted and implemented.

The contributors follow up in different directions and examine to various degrees the distance between what was in the constitutional/institutional blueprints introducing large-scale top-down decentralist reforms, and the facts on the ground twenty years later. Some look at how the sudden decentralization reforms slowly gave way to recentralization because none of the new subnational entities were politically and economically strong enough to balance off the centre; some look at how inherent deficiencies in infrastructure and personnel at the subnational level brought the central government back in; some look at how different subnational units ended up working differently in practice due to differences in demographic and social factors; some look at how uncodified factors – particularly the party system – came to determine how national politics functioned; some look at how decentralization created new conflicts between ethnic groups competing for the control of the new subnational entities; some look at how decentralization blew new life into traditional forms of authority at the expense of modern political parties, movements, and causes. But in sum, the big lesson is nothing looks like it was intended; and the variations in the ‘how’ and ‘why’ hopefully helps us better understand and explain the workings of federalism and decentralization in Africa – and beyond.

Federalism and decentralization are conceptually distinct from one another. In federalism, the existence of the constituent entities and their autonomous powers are constitutionally entrenched; that is, they cannot be created, merged, or abolished by ordinary legislation passed by the central government. Decentralization, on the other hand, is by definition an act of the centre as it devolves some of its powers to regional and local governments. The existence and autonomy of the subnational units are not enshrined in the constitutions; their numbers, borders, and powers can be determined and changed by the central government; what they do and how they do it can be set by the centre. In some countries like Ghana and Kenya, decentralization reforms have indeed been codified into the constitutions, but these clauses refer to decentralization in comprehensive terms as a part of the state architecture. Individual subunits lack the direct constitutional recognition and protection American states, German Länder, Swiss cantons, and Canadian provinces have.

In principle, decentralization can exist both in federal and unitary states. Most Scandinavian countries are constitutionally unitary but they practice strong forms of decentralized governance for example. The United States has both powerful constituent states whose powers and autonomy are constitutionally enshrined, as well as strongly decentralized local governments within these states implementing both national and state legislation. Centralized governance, on the other hand, is often associated with unitary states, but so can federal systems exercise such governance within their constituent entities. The existence of autonomous and strong subnational entities enshrined in the constitution does not mean that the level of government below has to be equally strong. Federalization in Belgium has gone hand-in-hand with the gradual erosion in the powers of the provinces which operate below the constituent units of the federation.

The practical distinction between federalism and decentralization is a little blurred for the African species of the suddenly-and-massively-decentralizing-state however. For the three big federations of Sub-Saharan Africa (South Africa, Ethiopia, and Nigeria), federal constitutions were part of post-conflict democratization settlements which also created newer forms of local governments.³ Federalization was not an act of formerly separate units coming together in a federal union; it was a top-down creation of new constituent entities – some with cultural, ethnic, historical, social, and geographic roots, some without. Some subnational units have seen their borders and powers change through legislation passed at the centre (albeit through special majority requirements). The fact that federalism and decentralized governance belong to the same constitutional settlements, and have been put in place at the same time, have made their functioning inseparable from one another.

Decentralization in the unitary states of Sub-Saharan Africa, on the other hand, has been part of similar large-scale top-down institutional change reform projects, often with strong international involvement and support. Most importantly, in all cases, decentralization was presented a way to bring in better democracy, better public policy formulation and delivery, economic development and growth. According to Paul Smoke the potential advantages of decentralization were presented in four categories: it promised improved efficiency, improved governance, improved equity, improved development and poverty reduction (Smoke 2003: 9-10). These promises were particularly prominent in the agendas of international organizations like the United Nations, World Bank and the International Monetary Fund (IMF) who encouraged, supported, and often financed decentralist reforms in Sub-Saharan Africa. Consequently, there is much to be learned from studying the federal-decentralized and unitary-decentralized sub-species together under the category of the suddenly-and-massively-decentralizing-state.

The pattern of evolution that applies to the mutation of both the federal-decentralized and the unitary-decentralized sub-species is the long-term and diffuse impact of structural uncodified factors. Looking at the institutional/constitutional blueprint will not help one understand the workings of decentralized governance. What is more, the formal blueprint might in fact paint a picture quite removed from reality in some cases, and thereby mislead both theorists and practitioners. But this is not a defeatist conclusion claiming everything is sui generis and context-dependent; there are indeed some generalizable patterns explaining the distance between the formal design and the facts on the ground two decades later.

The demise of the Soviet Union and the end of the Cold War ended most proxy wars in Africa. The changing geopolitical context subsequently increased the effectiveness of western aid conditionality in western development assistance (Dunning 2004). The early 1990s were marked by democratic reforms throughout Sub-Saharan Africa (Bratton and van de Walle 1997). Multiparty politics were the defining feature of this wave if democratization, but this went in tandem with decentralization reforms which either created or strengthened subnational levels of government. Strong backing from international institutions and donors was a constant feature of the continent-wide wave of democratization. According to Shabbir Cheema and and Dennis Rondinelli, “the International Monetary Fund, the World Bank, and other international development organizations prescribed decentralization as part of structural adjustments needed to restore markets, create or strengthen democracy, and promote good governance” (Cheema and Rondinelli 2007: 4). United Nations Development Program's African Futures Program subsequently became one of the key players in decentralization reforms in many African states (Fessha and Kirkby 2008: 265).

The World Bank and IMF have also lent financial and institutional support to decentralization reforms throughout Africa. That is, they have not only provided various types of funding but also institutional know-how in the design of decentralized institutions. But this is not a pattern limited to Sub-Saharan Africa of course. The contribution by Janet I. Lewis reports that between 1990 and 2007 the World Bank spent 7.4 billion dollars on decentralization programs throughout the developing world (Lewis, this volume). It is also imperative to note that international donors do not confine themselves to advocating decentralization, “they formulate the implementation of such policies as a prerequisite for receiving monetary support” (Erk and Swenden 2009: 3). According Gordon Crawford and Christof Hartmann: “[decentralization's] current popularity in the developing world is unparalleled, with 80 per cent of all developing and transition countries undertaking some form of decentralization over the past two decades” (Crawford and Hartmann 2008: 7).

Looking back at Africa's last two decades, it seems a little puzzling to see all international actors sharing a consensus around the promises of a concept that had not yet developed a track-record justifying this gushing endorsement. In his masterful 2007 book The Architecture of Government: Rethinking Political Decentralization, Daniel Treisman uses a rather low-key statement to make a very important point: “almost no robust empirical findings have been reported about the consequences of decentralization” (2007:5). Treisman is not alone. Paul Smoke questions the unqualified endorsement: “despite the limited empirical evidence to support decentralization and clarify how to reap its potential benefits, policy makers seem to be willing to push it forward in many countries” (Smoke 2003: 7). In fact, most development economists and students of public administration working on decentralization in the developing world generally acknowledged that decentralization did not always bring improvements in effective governance (Cheema and Rondinelli 2007: 8). Most of these scholars highlighted the need for an additional set of structural factors in order to help decentralization deliver. Dennis Rondinelli, James McCullogh and Ronald Johnson sum it up as follows: “ultimately, the success of decentralization policies hinges on institutional capacity-building” (Rondinelli, McCullogh, and Johnson 1989: 78).

Despite lack of clear evidence on the consequences decentralization, facile institutional recommendations based on promises derived from deductive hypotheses seemed to win over international donors and policy-makers. Empirical research on the results of decentralization failed to leave a dent on the international consensus around its promises. 10 years after decentralization in Benin, Thomas Bierschenk and Jean-Pierre Olivier de Sardan looked to see if reforms introduced in the early 1990s lived up to their promise. Their research shows that decentralization failed to deliver on its promises of incorporating the grass-roots, better governance, more accountability, and better governance (Bierschenk and Sardan 2003: 166-67). Instead, what the authors found is that the new decentralized institutions only added to the pre-existing complexity and fluidity of local politics. The picture is not very different in most other parts of Africa. Catherine Boone's recent work on territorial politics in Africa highlights the co-existence of formal institutions and informal politics (Boone 2013). The question in fact is not one about federalism and decentralization in Africa, but a broader one about the comparative study of institutions, and the premium most social science theories put on the causal centrality of institutions. In a piece reflecting on the attempts to engineer federalism and decentralization in Afghanistan and Iraq, Erik Wibbels argued that: “As a matter of social science, the lack of widespread attention to institutional endogeneity represents the most serious limitation on further understanding decentralized governments and processes around the world” (Wibbels 2006: 167). That is, uncodified background conditions, rather than the institutions themselves, could be playing more of a causal role in the outcomes.

Not only in Africa, but also in the established federations of the developed world factors exogenous to institutional design often influence and determine the workings of federalism and decentralization. By itself neither federalism nor decentralization represents a magic formula. What makes the combination of self-rule and shared-rule deliver on democratic representation and accountability, on efficiency and effectiveness in public policy performance, and on better economic performance – in, say Canada or Switzerland – relies on a number of uncodified structural factors. Without an infrastructure that allows the subnational units to exercise their newly-found powers; without a trained bureaucracy and personnel who can devise and administer policies within the new jurisdictions assigned to them; without sufficient financial resources – either locally generated or transferred from the centre – to pay for the new policies; and without the democratic autonomy to deviate from the political priorities of the centre; neither federalism nor decentralization can over-night turn countries into prosperous and democratic entities where minority rights are respected and dissension is considered a legitimate part of running diverse societies. Even in older established federations like Canada and Switzerland, the record of federalism and decentralization is a mixed one. The duplication of budgets and personnel, the premium political dynamics place on maintaining confrontational intergovernmental relations, and the way minorities can maintain and indeed cherish illiberal practices under the guise of federalism (the paramount example being 15,000 inhabitants of the Swiss canton of Appenzell-Inner Rhoden denying women the right to vote on local issues until 1991) show the need for a balanced conceptualization of both federalism and decentralization.

The reasons behind the readiness of international organizations and donors to prescribe decentralization as a cure for all ills under the sun perhaps needs its own separate collection; what we want to do here is to map out the evolutionary patterns the suddenly-and-massively-decentralizing-state goes through once this species is introduced to the political eco-system in Sub-Saharan Africa. Identifying potential patterns of evolution can help both theorists and practitioners better understand how things really work decades after the introduction of decentralization reforms. This chapter proposes five such categories: a) symmetrical recentralization; b) differentiated performance; c) legitimizing traditional authority structures and indigenous conflict resolution; d) politicization of local conflicts over land, water, and other natural resources; and e) federal extinction. Of course, several of these patterns can coexist and indeed influence each other, but using them as benchmarks can help bring a sense of order to what otherwise might look like contextual complexity and sui generis mutation.

a) Symmetrical Recentralization

The first evolutionary pattern that defines the state of affairs after the reforms is across-the-board recourse to centralization. Twenty years after the reforms, decentralized institutions remain in pro forma terms, but the workings of politics have reverted to back to nation-wide terms. One reflection of this phenomenon is the recentralization that comes with large scale national infrastructural projects. Patrimonial development programs and grand infrastructural projects more often target national goals as part of nation-wide policies, subsequently bringing the central government into policy-areas which formally belong to regional and local governments. This is something that applies symmetrically across the board to all subunits.

Symmetrical recentralization also happens to be the picture where the new subnational units have been numerous. This phenomenon is examined in the Ugandan context in the contribution ‘When Decentralization Leads to Recentralization: Subnational State Transformation in Uganda’. Janet I. Lewis shows the ultimate centralizing consequences of multiplying the number of the subnational administrative units (Lewis, this volume). Decentralization reforms, as formalized with the 1995 constitution, had initially produced 34 local government districts, but the numbers steadily grew. As a consequence, subnational entities which lacked the administrative capacity to carry out their new competences, became dependent on patronage networks to manage the day-to-day affairs, and subsequently witnessed the erosion of their ability to balance off the centre in intergovernmental relations.

This state of affairs is not uniquely Ugandan and applies to both unitary-decentralized states like Malawi and Burkina-Faso. In their comparative survey of subnational autonomy across the continent, Yonatan Fessha and Coel Kirkby highlight the role of ‘administrative and financial backdoors’ play in recentralization (Fessha and Kirkby 2008: 266). Essentially, these are shortcuts back to nation-wide politics that have come about through the political, administrative, and financial might of the central government. In his comparative piece on the three large federations of Africa, J. Tyler Dickovick argues that “federalism in Africa has the legal trappings of decentralized governance, set against the backdrop of dominant and overarching centralizing institutions” (Dickovick, this volume). Dickovick highlights the distance between the pro forma constitutional/institutional blueprints and the facts on the ground, labeling this “a sort of unitarism within formal federalism”. The contribution by Thomas A. Koelble and Andrew Siddle examines similar dynamics at the local level below the provinces and regional-states. The particular context of their analysis is post-apartheid decentralization reforms in South Africa. The authors add a nuance to this line of reasoning highlighting the weakness of an over-burdened local government, combined with a “distracted, inertia-bound, and overwhelmed central government” (Koelble and Siddle, this volume).

An interesting question concerns the reasons behind the process of recentralization. Is this state of affairs the result of the force of circumstances? In other words, is recentralization the long-term consequence of structural deficiencies in infrastructure, know-how, and personnel limiting the administrative capacity of the new subnational units? This would suggest that the importance of extra-constitutional structural factors were ignored during phases of sudden and massive institutional design. Or were the reasons of recentralization inherent in the contradictions of the original constitutional settlements? The Ugandan case-study shows how institutional design – in particular, the rapid creation of numerous new subnational units – account for the subsequent recentralization. Administrative unit proliferation is also visible in the federal-decentralized of Ethiopia, Nigeria, and South Africa – albeit with more federal room for differentiation among the units which will be dealt with in further detail in the following section. For example, while not being as symmetrically comprehensive as the Ugandan case, there has been a comparable process of recentralization in the Ethiopian case as well. While structural deficiencies in their administrative capacity have led some regional-states to turn to the dominant party in national politics for securing financial and political support, some observers see this not as a long-term political dynamic but one that owes it to the original constitutional blue-print. According to Lovise Aalen:

The institutional framework of the Ethiopian federal system includes two contradictory elements. According to the Constitution, the regional governments have rather wide powers to run their own affairs and implement their own policies. Contrary to this is the meager financial resource-base of the regions and the strong powers of the central executive

(Aalen 2006: 248).

While Aalen's emphasis on the importance of matching constitutional competences with the requisite financial resources required to deliver them is a key constitutional sine qua non for federalism, there is more to the Ethiopian case than just an inherent constitutional contradiction. While the system might display a limited amount of recentralization akin to the process that has unfolded in Uganda (especially taking into account the proliferation in the number of local subunits), the picture is more asymmetrical – partly because above the local government level rest eleven subnational units, which are quite different from one another in size, population, economic development and ethnic composition. Some are small city-states with urban multi-ethnic populations, some are large but underdeveloped peripheral regional-states; some are ethnically homogenous, others are amalgamations of dozens of different ethnic groups some of which are indigenous to the territory while others have settled there in recent generations. Such differences have had a bottom-up influence on the day-to-day workings of decentralization introduced in pro forma symmetrical terms. Here what we see might better be labeled ‘differentiated performance’ rather than across-the-board recentralization.

b) Differentiated Performance

The three large federations of Sub-Saharan Africa (Ethiopia, Nigeria, and South Africa) are defined by ethnic, religious, economic, geographic and demographic diversity – and when such diversity is territorially based, this often reflects itself in de facto asymmetry across the regions even while the constitutional blueprint follows a pro forma symmetrical format. For example, the Ethiopian constitution grants extensive powers of self-rule to ethnic communities. Especially for regional-states which are homelands to ethnic groups constituting the majority of the population of the territory, this is potentially a very important power of cultural self-rule as regional-states also choose their own working language. However, the constitutional right to cultural autonomy means little for poorer, underdeveloped, peripheral regions. Assefa Fiseha reports that the Afar regional-state in Ethiopia could not use its own Afar language until 2011 due to the lack of trained personnel and resources (Fiseha 2012: 449).

In their interim assessment of the South African constitutional reforms five years after their implementation, Richard Simeon and Christina Murray warned that “the benefits of federal or multilevel institutions depend crucially on the ability of the component parts to carry out their assigned roles and responsibilities” (Simeon and Murray 2001: 66). The authors listed various components of what they called the ‘governing capacity’ of the new provincial governments: political capacity, legislative capacity, administrative capacity, fiscal capacity, and intergovernmental capacity (Simeon and Murray 2001: 66-7). If regions differ in their governing capacities, it seems self-evident that the regional-states and provinces will indeed display differentiated performance in the execution of their newly-found competences. It is this very within-country variation that the contribution by Heather Huntington and Erik Wibbels deals with (Huntington and Wibbels, this volume).

The May 2014 in South Africa elections resulted with yet another overwhelming victory for the African National Congress (ANC) and returned the party to run the national government as well as all provincial governments but one. The main opposition party the Democratic Alliance (DA) maintained its control of the province of Western Cape, while the remaining eight are all governed by ANC. One of the main electoral platforms the DA promotes in their electoral campaign is their track-record in provincial governance – despite being left out of the informal networks of the dominant national party. Western Cape is the showcase for DA's potential, so a lot of effort is put in for good governance. Be that as it may, the Cape region (with its metropolis Cape Town) has always been a ‘have’ region regardless of who governs. Differentiated performance all comes down to the make-up of the subnational unit in question. ‘Have’ provinces often tend to make most use of federalism, while ‘have-nots’ end up with competence over new policies that they cannot deliver – unless they have mighty benefactors within the dominant national party of course.

If you are a ‘have-not’ province, and lack infrastructure, trained bureaucracy and personnel, and requisite financial resources – and want things to be done in your province – you look for benefactors within the ranks of the dominant party. It is through the informal networks of the dominant national party that a form of opaque inter-regional equalization takes place. This is a characteristic that has influenced the workings of federalism in South Africa, Ethiopia, and Nigeria at various degrees during different phases of their political history (Suberu 2009: 80-83; Bach 1989). Especially when party membership and bureaucracy are separated by fairly permeable boundaries, we see more of this pattern. This is different from the symmetrical recentralization covered in the previous section. Internal party politics and informal networks come to play a role simply because subnational units differ in their governing capacities. Regional differences play a key role here. The type of centralization covered in the previous section on the other hand, is a pattern that applies across the board to all subnational units.

Despite the short-term challenges, what follows from the pattern of differentiated performance is more relevant to federalism and decentralization in the long-run. These are signs that bottom-up politics have started to take roots. An interesting reflection of this phenomenon is investigated in the contribution entitled ‘Ethnic Decentralization and the Challenges of Inclusive Governance in Multiethnic Cities: The Case of Dire Dawa, Ethiopia’ (Asnake, this volume). Asnake Kefale's field-research has yielded an interesting and little-known insight into how the federal spirit of power-sharing has (informally) permeated the local level. His focus is on the divided and contested city of Dire Dawa – a city-state within the Ethiopian federal system. Kefale's chapter shows the prevalence of informal politics at the local level, reflecting the influences of a blend of contemporary federal ideas and uncodified conflict resolution mechanisms belonging to the indigenous traditions predating federalism.

c) Legitimizing Traditional Authority Structures and Indigenous Conflict Resolution

The third pattern defining the state of affairs is the resurgence of traditional authority structures in some local governments, especially in the peripheries; and the new life decentralization has blown into customary law – in particular to indigenous conflict resolution mechanisms for managing dispute between ethnic groups. However, this has not been a straightforward process. Two decades after the reforms, the pattern is uneven and diffuse, suggesting a complex mix of factors behind the processes of when and how traditional authority and customary law sweep into the workings of regional and local governments. Before looking at the decentralist reforms of the mid-1990s, we would have to go back to the first post-colonial constitutions in order to make sense of the circuitous and informal path towards legal pluralism; that is, the coexistence of modern constitutional law and local customary law.

Influenced by the prevailing modernization theories of the time, most post-colonial constitutions throughout Africa had rejected legal pluralism. The decentralist reform projects of the 1990s did not undo this. Nowhere in Africa did indigenous legal traditions in managing diversity and settling disputes find their way into the formal design of decentralist institutions. The institutional/constitutional blueprint was often carbon-copied from the West. But the processes and mechanisms of adjudicating federalism disputes within the newly adopted constitutional set-up did not receive much attention (Abebe 2013: 55). However, all societies around the world have home-grown mechanisms and social practices to handle conflicts and resolving disputes (Gulliver 1979). These tend to rest on a bottom-up form legitimacy with roots in local communities that top-down institutional/constitutional reform projects lack. Such indigenous conflict mechanisms tend to command widespread social acceptance.

In a piece in the International Journal of Constitutional Law, Kwasi Prempeh laments the way such mechanisms and practices with organic and cultural legitimacy were sidestepped during the reforms of the mid-1990s: “Constitutional change in much of contemporary Africa has failed generally to redress or reconsider … the post-colonial exclusion of Africa's homegrown customary institutions from the formal structures of local representation and governance” (Prempeh 2007: 495). There are, however, a number of factors that complicate the formal recognition of traditional authority structures and customary law. The first is how such structures were used by former colonial powers as a short-cut to establish control over wide swathes if territory with a small contingent of soldiers and officials. As Sandra Joireman Fulllerton points out, “[The British] needed traditional leaders in order to make their policy of indirect rule work.” (Fullerton Joireman 2001: 579). Throughout Africa, chiefs and elders became partners to the British or they were loosely co-opted into colonial administrations (Nkrumah 2000). Despite some marked differences with the British, French colonial administrations also relied on traditional authorities (Crowder 1964).

It is therefore not surprising that following decolonization, these traditional structures and customary laws were left out of the architecture of most of the newly independent African states. New post-colonial constitutions almost always represented modernist and nation-building aspirations and a principled rejection of legal pluralism. And these were put in place while the apartheid regime in South Africa was starting to use a more cynical variation of the colonial practice. Colonial indirect rule was born out of the force of circumstances. The thinness of colonial administrations, both in manpower and in infrastructural reach, had historically necessitated the recognition of traditional chieftaincy structures. The National Party that came to power in 1948 turned what was a diverse collection of ‘native reserves’ and nominally independent chieftaincies into a more comprehensive political system as part of its new policy of ‘apartness’. Unlike the colonial powers of the 19^th century, the apartheid regime did not lack manpower or infrastructure. Giving the outward appearance of autonomy to satellite tribal ‘homelands’, derisively known as the Bantustans, bought a semblance of constitutional legitimacy to the system of apartheid. This was an additional factor discrediting traditional authority and customary law throughout the continent in the post-colonial years.

Add to this the fact that many among the post-colonial African elite viewed traditional authority structures and customary law as illiberal and backward. The influence of modernization theories entrusting the central state with a developmental vanguard role further underpinned the rejection of legal pluralism. Kwasi Prempeh's call to formally integrate such home-grown institutions into constitutional structures will therefore continue to face serious hurdles before it can command widespread acceptance and support among constitutional scholars and politicians. There are signs that such a rejection is weakening however. Janine Ubink's work keeps track of the recent reforms bringing formal recognition to traditional authority structures and customary law. Mozambique, Uganda, Ghana, and Botswana are among African countries where legal pluralism is now formally recognized (Ubink 2008: 11). In informal terms, decentralization has also been blowing new life into traditional authority structures at the local level. Catherine Boone reports that “In many places, formal decentralization programs have worked to augment the social, economic, and political capital – including the land-related powers – of customary or neotraditionalist actors within local jurisdictions” (Boone 2013: 219)

The informal recognition and application of customary law is in fact a lot more widespread than what the formal picture suggests. There is a visible distance between the pro forma letter of the law and the practice on the ground which tends to be more inclusive towards legal pluralism. Especially in more peripheral regions of Ethiopia and Nigeria beyond the reach of much of modern infrastructure, officials often informally consult with traditional chieftains and incorporate customary practices of conflict resolution in order to settle disputes between groups. The 2008-2009 Afrobarometer survey of citizen attitudes displays remarkable levels of influence traditional kings, queens, and paramount chiefs hold over local affairs in many African countries. With decentralization spreading throughout the continent, this is an angle that awaits further research.⁴

In this context, it is worth noting that legal pluralism has found a formal foothold even in the constitutional order of post-apartheid South Africa. A council of traditional chiefs was set up in 2003. It is a largely advisory body, but considering the earlier episode with Bantustans, this is an important constitutional bridgehead nonetheless. In addition to recognition at the constitutional level, traditional authority and customary law has also found itself into some of the policies of South African local government. In the context of land administration, the complex links between democratic decentralization and traditional forms of authority in rural areas were exposed by Lungisile Ntseba (Ntseba 2004). Traditional customary law remains a controversial question however as many women's organizations fear that its formal recognition would undermine the rights of rural women. A recent Traditional Courts Bill even failed to survive the parliamentary committee stage in the South African parliament.

d) Politicizing Local Conflicts over Land, Water, and Other Natural Resources

The constitutional recognition of ethnic identity accompanied post-conflict democratic settlements in South Africa, Ethiopia, and Nigeria. One common criticism of the post-conflict settlements in these three federations is how constitutional recognition of ethnic group rights politicizes the very divisions it aims to manage. But the issue here is much bigger than federalism however and goes to the heart of individual vs. group rights debates in democratic theory. It is a question without an easy answer. Leiden political scientist Frank de Zwart calls this the ‘dilemma of recognition’ (de Zwart 2005). While the pro and cons of formal recognition of ethnic identities continue to be debated in scholarly circles, it is self-evident that in its most basic form it disenfranchises individuals of mixed backgrounds of various group rights, ranging from formal political representation to informal ethnic quotas in employment. This is also the case for individuals who hold other identities – be it regional, political, ideological, or religious – higher than ethnic ones. As Alem Habtu puts it gently, “for de-ethnicized Ethiopians and the offspring of inter-ethnic marriages, the imposition of ascribed ethnic classification, in the wake of multiethnic federalism, raises issues of identity and citizenship” (Habtu 2005: 332). A prolific observer of Ethiopian politics, Jon Abbink provides a more incisive reading of the same phenomenon: “[Ethiopia] ‘has now itself fallen into the trap of ethnicizing social and community relations, this often blighting daily relations and creating new oppositions” (Abbink 2009: 15). These criticisms reflect one angle to the broad individual vs. group rights debate in democratic theory. But what we are more interested in here are two types of ‘identity’ conflicts decentralization has specifically engendered.

One of the broad patterns defining the picture on the ground twenty years after decentralist reforms is the politicization of local conflicts between communities. Competition over farmlands, water, grazing, and other natural resources always existed among different communities. Constitutional recognition of ethnic identities has in some instances filtered down to the local level adding a political character to what were originally local conflicts. This is the first type of conflict associated with decentralization that we sometimes see.

For both federal-decentralized and unitary-decentralized states, the establishment of a new level of local government with new competences has also engendered often ugly conflicts over who is a local, who is a settler, and who is entitled to control the land. Instead of politicizing pre-existing local conflicts, this second type of conflict is rather recent and seems to have direct links to decentralization, the high stakes involved in the competition over the control of local government, and the potential material benefits that accompanying such control. For some, this has been seen as an opportunity to establish exclusive control over land using references to territorial self-rule. The International Crisis Group has been trying to monitor such conflicts. As their 2009 Africa Report puts it, “relations between ethnic groups have become increasingly competitive, as they view for control of administrative boundaries and government budgets in addition to land and natural resources” (International Crisis Group 2009: 22). In addition to the conflict this creates between nomadic pastoralist and farmer communities, a growing conflict has been between indigenous ‘titular’ communities and settlers who had arrived in previous generations. Historically many communities have moved throughout the continent, and their oral traditions have kept the stories of their treks alive. Others have moved fleeing violence or they had been forcefully settled in the territories they now found themselves in. Many regional-state constitutions in Ethiopia now have ‘sons-of-soil’ references, often leading to the political disenfranchisement of settler communities, or even expulsion. In some cases, non-indigenous have been evicted from the areas they had been living in for generations, something the International Crisis Group interprets as an act akin to ethnic-cleansing (International Crisis Group 2009: 24). Nigerian citizens inhabiting regional-states outside their ancestral territories also experience political and economic disenfranchisement.

It would be misleading to portray decentralization as the sole cause of such conflicts of course, but it is nonetheless imperative to expose some of the consequences of simplistic social science hypotheses, and the international experts’ soft-spot for social engineering projects. One of the new policies adopted by the World Bank and the IMF in the mid-1990s was to incorporate local communities into development programs. In particular, they would be entrusted in the management of local resources. It sounds all good in the abstract. But the sudden appearance of the potential of previously unthinkable levels of material gain often unleashed conflicts over who ‘owns’ the land. One example of this has been the 1994 forestry law in the Cameroon adopted in line with the demands of the international donors – particularly the World Bank – who wanted local communities to benefit from the local natural resources (Geschiere 2004). The outcome was indigenous communities with historical and cultural claims to the land terrorizing and expelling migrant settlers who had arrived in the previous generations. Ana Katherina Schelnberger shows similar patterns in Uganda: resources-based conflicts between the indigenous and settlers suddenly acquired political characteristics with the creation of a local government, leading to escalating violence (Schelnberger 2008).

It might be too late for the evicted communities, but the contribution ‘The Geography of Governance in Africa’ examines a number of new methodological tools that can map out internal variation in local politics, and provide the necessary knowledge basis before communities are subjected to naïve social engineering projects (Huntington and Wibbels, this volume). One of the reasons with such unexpected outcomes in development policies is the misleading potential aggregate national indicators carry with them, which often conceal within-country variation and pave the way for one-size-fits-all policies where complicating factors are held constant for the sake of empirical generalizability. Heather Huntington and Erik Wibbels's contribution shows how innovative approaches using new technologies might help prevent the repeat such accidents in social engineering.

e) Federal Extinction

One of the sub-species, it turns out, is not completely new to the Sub-Saharan political eco-system. The federal-decentralized state has had a short-lived ancestor during the transition from colonial to home-rule to eventual independence. The transition phase included a number of federal experiments that are now mostly forgotten. The Central African Federation (1948-53) was a last-minute attempt to unite imperial British colonies in Africa under a confederal union (Hyam 1987). It brought together Britain, South Africa, Northern Rhodesia, and Southern Rhodesia for short while. The French experimented with similar confederal union designs combining continued French presence in Africa with home-rule to its former colonies. Senegal, Guinea, Dahomey, Ivory Coast, Niger, and French Sudan (present-day Mali) and Mauritania would unite under the framework of Afrique Occidental Française with political links to mainland France. A similar union brought Gabon, Chad, Congo Brazzaville and the Central African Republic together under the framework of the Afrique Équatoriale Française. These rather clumsy confederal attempts to stem the tide of decolonization soon floundered, but they were replaced by other short-lived federal experiments.

The Federation of Northern and Southern Rhodesias with Nyasaland (1953-63) fell victim to increasingly politicized racial identities. The Congo was a federation for a short while (1960-1965); Kenya had a short phase of federalism during its early independence (1963-1965); so did Uganda (1962-1966). The so-called Mali Federation between French Sudan and Senegal formed in 1959 lasted a little more than a year. Then there is the Federal Republic of Cameroon (1961-1972), which, due to its convoluted colonial legacy, had separate anglophone and francophone elites (Fegue 2008). Federal Cameroon is an interesting and informative case for the comparative federalism literature since it presents a rare example of a federation abolishing itself through a referendum and adopting a unitary state instead.

As these now defunct experiments indicate, federal solutions were high on the agenda during the early phase of post-colonial politics in Africa as potential ways to reconcile unity and diversity, but they ended up being rather short-lived experiments. Writing during this time-period, Donald Rothchild points out that by the 1960s the enthusiasm for federalism had quickly given way to preference for unitary solutions (Rothchild 1966: 275)⁵. Neither British and French confederal union initiatives nor the post-colonial federation experiments seem to have left much mark on the political eco-system. That being said, this is an understudied angle so we might be overlooking possible links between past federal experiments and the present. Both for the theory and practice of federalism, there is need for more scholarly attention to the short-lived federal ancestors.

A somewhat different, but still extinct, federation is the UN sponsored initiative bringing the unitary Ethiopian state into a federal union with neighboring Eritrea, which had been an Italian colony. Right from its inception 1952, it was clear that the asymmetry between its two component parts would prevent the Ethio-Eritrean federation from functioning as a genuine federal union. Soon after its foundation, Ethiopia's official language Amharic was made the official language of the federation, the Eritrean flag was abolished, and the chief executive of Eritrea came to function as the provincial administrator of the emperor. Not long after, the Eritrean assembly voted in 1962 to be united with Ethiopia (Tekeste 1997: 138). That being said, this extinct federation is somewhat of a different species from the previous two categories discussed above. It is a more of an instance of geo-strategic power politics where a strong state incorporates a smaller neighboring territory with which it has had historical, political, cultural and geographic ties. In any case, the current federal system in Ethiopia has little to do with the post-colonial international initiative from sixty-five years ago. But there is one contemporary federal-centralized state in Africa whose federal past is inseparable from its present.

Nigeria has gone through a number of federal options making it hard to determine where the spotlight should be directed. Nigeria's origins as a country, and its subsequent troubles, can be traced back to the three provinces created by the colonial authorities, and the subsequent federal union of these three in 1954 (Suberu 2009: 71-2). But how does one judge whether federalism is a success or failure; or more precisely, whether federalism is the problem or the answer? And in doing so, what type of federal institutions are we are evaluating: 1954's federalism of three provinces, 1967's federalism of twelve regional-states, the civil war and ethnic divisions, military rule, 1976's federalism of 19 regional-states, 1991's federalism of 30 regional-states, or the 1996's federalism of 36 states? Throughout its turbulent history since independence, Nigeria has been searching for a federal institutional set-up that can best accommodate its various social divisions (Suberu 1993: 39). But all the institutional changes seem to have had little long-term impact on the persisting societal divisions in the country. Understanding and explaining such change and continuity has relevance beyond the Nigerian case.

Federations might go extinct, or their organizing logic and set-up replaced by new institutions. Most of the literature on comparative federalism looks for answers simply by looking at existing federations, and there might be a little risk of a selection bias here. That is, by looking at existing federations, we look at those that managed to survive. These are ‘success’ stories in the most basic sense because they are still around. In order to evaluate success or failure of decentralization and federalism, we not only have to cast the net wider than institutional/constitutional factors but also to make sure that our case-selection does not have an inherent tendency to overplay the promise of federalism. That is, we have to look beyond surviving federations in order to provide a more even-handed evaluation.

Decentralization, both in unitary-decentralized and in federal-decentralized states, is no cure to all ills. Common-sense calls for prudence and sober reflection before one embraces decentralization unconditionally. That being said, it is imperative to remember that the unitary-centralized alternative comes with more of a chequered track-record than its decentralized siblings. The contribution by Thomas A. Koelble and Andrew Siddle shows how both levels of government often fail the citizens. According to Janet I. Lewis, there are indeed deficiencies at the centre that make the local suffer. In other words, failure in governance might very well be a shared responsibility. At the same time, it is hard to imagine a unitary-centralized system bringing better democracy, policy efficiency, development, and growth to ethnically, economically, geographically, and demographically diverse societies. One-size-fits-all policy-making from the centre faces an uphill battle in presenting itself as a credible alternative to federalism and decentralization. The question then becomes how to study federalism and decentralization, and related factors influencing their day-to-day workings. This calls for an even-handed evaluation of the advantages as well as the potential risks of both. Neither federalism nor decentralization can ‘solve’ all the divisions – and they should not be marketed as such; but they seem to be the least-worst mechanisms to ‘manage’ the divisions within diverse societies.

One should also be hesitant to paint a type of federalism or decentralization that is specifically and entirely African. All political systems evolve with time. Established federations with longer histories have all gone through different phases of centralization, decentralization, differentiated performance, and even civil war. The institutional/constitutional blueprints of older federations have also interacted with uncodified social structures over the long-run leading to gradual, and sometimes sudden, change. In this sense, there is nothing unique about unitary-decentralized and the federal-decentralized states of Sub-Saharan Africa. We do not need a separate subfield to study the variations of federalism and decentralization in Africa. In fact, bringing in African countries to our studies can only enrich the field of study. As J. Tyler Dickovick points out, the benefits of adding these cases to our repertoire can be especially important in approaching federalism as a dependent variable that tends to reflect a complex mix of other variables (Dickovick, this volume).

That being said, there is something that is specifically and entirely African here. What makes the species of the suddenly-and-massively-decentralizing-state a new discovery is its top-down promise of an instant fix. Once they fail to deliver everything under the sun, such inflated promises might come back to tarnish the sheen of entire democratic reform packages. Prudence might not grab headlines, but it has a better long-term track-record.

Notes

1 Institute of Political Science, Leiden University, P.O. Box 9555, Leiden, 2300RB, The Netherlands. [email protected]. The special issue was completed while author was visiting professor of the South African Research Chair on Multilevel Government, Law, and Policy at the Community Law Centre, Faculty of Law, the University of Western Cape, in South Africa. He thanks the Community Law Centre staff, in particular Jaap de Visser, Nico Steytler, and Derek Powell, for the warm welcome.

2 It is imperative to note that the literatures on public administration and development economics view territorial decentralization as but one facet of the broader concept of decentralization. Deconcentrating central policies to the field offices of national departments, and delegating national competences to third parties are seen as the two other facets (Cheema and Rondinelli 2007: 3). Our focus is solely on the territorial element in federalism and decentralization.

3 Despite strong federal characteristics, the South African constitution avoids the formal label ‘federation’ for the political system it introduces. As Daniel Elazar put it, the country is “a federation in almost every respect but does not refer to itself as such” (Elazar 1998: 9). Nico Steytler and Johann Mettler see this as a factor of the careful balancing act between various pro- and anti-federal political currents during the transition to democracy (Steytler and Mettler 2001). The Comoros islands off the east coast of Africa are formally a federation as well, but the political patterns on the islands come closer to one what Jan Erk and Wouter Veenendaal identify in island microstates (Erk and Veenendaal 2014).

4 It is usually within the social science approaches to comparative federalism and decentralization where we see such a gap. Lawyers are a little more aware of this angle as some of the recent pieces in the Journal of Legal Pluralism and Unofficial Law show.

5 The renowned historian Frederick Cooper's most recent book deals this theme in more detail (Cooper 2014). A common pattern throughout the continent at that time seemed to be a preference for the nation-state, despite an awareness of the shortcomings and risks. It was both internal inequality within the new states and inequality vis-à-vis the industrialized west that led African leaders to adopt the centralized path.