Introduction
Federalism literature has engaged extensively with why federal systems emerge and how federalism affects governance in diverse polities. In the first sense, federalism is a dependent variable, modelling the historical ‘coming together’ of constituent units (Anderson, Reference Anderson2014; Dickovick, Reference Dickovick2014). In the second sense, it is an independent variable, associated with a range of generally desirable governance outcomes. These include federalism’s positive effects on increasing local self-governance (Burgess, Reference Burgess2006; Candela, Reference Candela, Lemke and Tarko2021; Ostrom, Reference Ostrom1973, Reference Ostrom1991), market preservation (Weingast, Reference Weingast1995), and state stability in multiethnic polities (Maiz, Reference Maiz1999; Majekodunmi, Reference Majekodunmi2013).
In this conceptual paper, we consider how federal institutional rules and designs affect the likelihood of attaining the expected benefits of federation, with particular attention to the expected effects on conflict-alleviation in multiethnic polities. In policy-relevant literature on contemporary state-building, especially in African states, federalism is often evoked as a mechanism for alleviating ethnic conflict and governing diversity; federalism has been proposed as a peaceable solution to alleviate the permanent states of conflict in countries like Somalia and South Sudan (Johnson, Reference Johnson2002; Osaghae, Reference Osaghae2004; Roach, Reference Roach2016, Reference Roach2023). Currently, the three established federal systems in Africa are Nigeria, Ethiopia, and South Africa (Dickovick, Reference Dickovick2014).Footnote 1
With the possible exception of South Africa, none of the aforementioned federal systems have achieved the expected benefits of ethnic conflict alleviation. Nigeria and Ethiopia have experienced cycles of group-based violence and state breakdown, despite adopting federal systems, whereas quasi-federal structures have not brought about stability in Somalia or South Sudan. There is a conceptual gap, then, between the expected benefits of federalism and its actual governance outcomes in the African context. One of the main reasons for this is that federalism is frequently treated synonymously with decentralization.
We identify this conceptual gap when comparing existing literature on federalism deficits in Africa with the conceptual literature on polycentric ideal types. Whereas the need for decentralization is highly emphasized in the African context, studies of ideal-type federations emphasize broader institutional parameters, such as levels of intergovernmental competition and individual choice (Ostrom, Reference Ostrom1973, Reference Ostrom1987; Stephan et al., Reference Stephan, Marshall, McGinnis, Thiel, Blomquist and Garrick2019). As such, the current framing of federalism in African scholarship may overlook critical institutional design factors that influence its effectiveness in managing diversity.
We respond to this gap by asking two questions. First, how can federalism be conceptually reconfigured in alignment with its expected positive outcomes on governance? Second, why does federalism fail to effectively manage identity-based conflict in ethnically diverse states, particularly in Africa? We use two case studies to highlight examples of the failures of federal institutional design that can be addressed through better conceptualization. Although the cases are African states, the lessons from these cases can help inform more sound conceptualizations of federalism, contributing to policy-relevant discussions in other parts of the world. Beyond the contributions to policy, our analysis contributes to institutional economics by illustrating how federal institutional design affects identity-based group dynamics in conflict-ridden multiethnic polities.
We start by discussing our methodology in the section below to illustrate our analytical approach and case selection strategy. Next, we review conceptualizations of federalism to configure an analytical framework of polycentric federalism as a necessary institutional design component for achieving the desirable outcomes associated with federalism. Polycentric federalism entails a federal institutional design that prioritizes autonomy of decision-making at lower levels of government, peaceful evolutionary competition between different governance authorities, and individual choice (Candela, Reference Candela, Lemke and Tarko2021; Ostrom, Reference Ostrom1973, Reference Ostrom1987; Stephan et al., Reference Stephan, Marshall, McGinnis, Thiel, Blomquist and Garrick2019).
Polycentric federalism, we argue, illustrates the importance of accounting for the elements of competitive governance and individual choice, moving policy-relevant discussions of federalism away from a focus on decentralization. In configuring this approach, we develop three analytical parameters for assessing how institutional design in specific systems is oriented toward or further away from the polycentric ideal type, namely, administrative devolution, peaceful competition among governance units, and individual choice of alternating governance structures. We develop these parameters through a paired comparison of two analytical case studies, Nigeria and Ethiopia.
On the parameter of administrative devolution, post-1999 Nigeria performs slightly better than Ethiopia under the Ethiopian People’s Revolutionary Democratic Front (EPRDF). On the parameter of peaceful competition, both cases have similar limitations on the practical possibilities of inter-regional competition. On the parameter of individual choice of alternating governance structures, the Nigerian system, on paper, provides a much more viable structure than that of EPRDF’s Ethiopia. Although we see marked levels of decentralization in Ethiopia after the transition from the EPRDF to Prosperity Party in 2018, the identity-based institutional design in that case undermines peaceful inter-regional competition and individual choice. We end by reflecting on how institutional design and rules in these two countries can be reconfigured in alignment with polycentric federalism.
Methodology and cases
We do not claim to test theory in this paper; rather, we develop a configuration of federalism to address the gap between theoretical expectations of federalism and its actual governance outcomes in the African context. Our methodological approach configures the concept of polycentric federalism, developing clear parameters for assessing whether a given federal system approximates the ideal type of polycentric federalism. In the subsequent section, we detail how this approach is situated within the broad literature on federalism and how the parameters are developed. We also identify examples of federalism that approximate this ideal type before analysing the two cases.
The case selection scheme is guided by the importance of illustrating how federal institutional design precludes the attainment of desired governance outcomes in disparate contexts, through a most-different systems (MDS) design (Tarrow, Reference Tarrow2010). The MDS strategy of case selection guides us to choose two geographically, historically, and culturally divergent cases, wherein similar causal mechanisms may explain the similarity in suboptimal federal governance outcomes.
Accordingly, Nigeria and Ethiopia are selected as case studies because both countries have experienced cycles of identity-based violence and ongoing ethnic tension, despite their contrasting federal designs, divergent political histories, and disparate cultural-geographic contexts. Nigeria’s post-1999 federal system has three tiers with a strong central government, while Ethiopia’s post-1995 ethnofederal experiment has oscillated between centralism and state fracture. The Nigerian federal states are geographically drawn, whereas regional states in the Ethiopian system assume the national sovereignty of each major ethnic group. Federal governance outcomes in these cases, however, have been fairly similar, including administrative monopoly by the central government and exacerbation of ethnic tensions. In line with MDS analysis, despite ostensible differences in institutional design, both cases fall short of managing ethnic difference peacefully through federalism.
Since much of the debate over federalism’s expected effects in alleviating ethnicity-based conflict is concerned with cases of African states, it makes sense to draw upon the experiences of these two largest countries in Africa by population, which are also among the six largest federal systems in the world. By examining these cases, we demonstrate how polycentric federalism elucidates flaws of federal institutional design in ways that other approaches cannot and discuss how these insights inform policy for reconfiguring federal institutional design. The objective, then, is to highlight the most pertinent institutional mechanisms for achieving federalism’s expected benefits in managing ethnic difference.
For case analysis, we use Yin’s (Reference Yin2018) cross-case synthesis and Gerring’s (Reference Gerring2007) paired comparison approach. This involves the proposition of a framework, the description of the cases, and cross-case analysis based on the three analytical parameters we developed for polycentric federalism. The paired comparison approach allows us to structure a side-by-side comparison of the two cases along some common parameters established by the polycentric federalism framework. We rely on constitutional documents as well as on a systematic review of the extant scholarship on federalism deficits in Nigeria and Ethiopia. More specifically, we consider the most oft-cited empirical literature on federalism deficits in these two countries, with a focus on more contemporary critiques (2011–2024). In considering this literature, we centre on main arguments for why federalism fails in these two countries and alternative explanations in the literature. We then map these back to specific provisions in the two countries’ constitutions (for instance, the Federal Character Principle in the Nigerian Constitution and the provision for ethnofederal organization and secession in the Ethiopian Constitution), as well as other policy documents such as political party charters. We present a brief historical background of federalism’s emergence in the two cases below, before developing our framework in the next section.
Nigeria
Federalism in Nigeria emerged under British colonial rule in 1946. The British sought to accommodate divergent administrative traditions in the Northern and Southern regions, but were also intent on preventing the emergence of a strong, centralized state that would threaten British interests (Afigbo, Reference Afigbo1991). A series of constitutional conferences in the 1950s led to self-government for Nigeria’s three regions and eventually independence in 1960, with the 1963 Constitution officially declaring the Federal Republic of Nigeria (Adamolekun, Reference Adamolekun1991; Afigbo, Reference Afigbo1991). In 1966, a military junta, led by Major General Johnson Aguiyi-Ironsi, staged a coup ande reverted the country into a unitary state. A return to civilian rule and a new constitution in 1979 centralized more power at the national level, but this arrangement was again disrupted by military rule from 1983 to 1999. With the return of civilian rule in 1999, the new government established the federal constitutional arrangement in the country that persists to the present day.
This current federal arrangement establishes three tiers of government: national, state, and local governments. The local governments are constitutionally recognized as part of the federation, listed explicitly in Schedule 1 of the Constitution. Many scholars have argued that the 1999 Constitution is still far from guaranteeing the benefits of a genuine federal system in Nigeria (Babalola, Reference Babalola2019; Kalu, Reference Kalu2016; Majekodunmi, Reference Majekodunmi2013). A core component of these debates is the institutionalization of the ‘Federal Character Principle’, which sought to manage the problems of ethnic diversity and regional autonomy that historically animated conflicts in the country (Lanre, Reference Lanre2018; Osaghae, Reference Osaghae1991; Suberu and Diamond, Reference Suberu and Diamond2003). Historically, the federal system in Nigeria nurtured rivalries among ethnic groups, intensifying regionalism and fostering mutual distrust, particularly between the North and South (Rogge, Reference Rogge1978). The culmination of these tensions was the Biafran civil war of 1967–1970.
Subsequently, the centralization of power during the era of military rule, which persisted until 1979, significantly eroded regional autonomy, exacerbating tensions rather than alleviating them. These legacies continue to influence the institutional functioning of the contemporary Nigerian state. Despite the theoretical distribution of power in the 1999 Constitution, the central government retains substantial control, often limiting true administrative autonomy at the regional level. As a result, neither regional autonomy nor ethnic cohesion is achieved.
Ethiopia
Ethiopia was never colonized by European powers, though its current borders were shaped by agreements with Italy, Great Britain, and France (Kornprobst, Reference Kornprobst2002; Matshanda, Reference Matshanda2022; Muhumed and Siraj, Reference Muhumed and Siraj2017). The Ethiopian Empire was initially crafted by Semitic-speaking groups and governed by emperors claiming descent from the Solomonic Dynasty (Clapham Reference Clapham2017; Demerew, Reference Demerew2024; Marzagora, Reference Marzagora2020). Three seminal events in the sixteenth and nineteenth centuries created identity group dynamics that precipitate the move towards the contemporary ethnofederal system.
First, Ahmed Gragn’s invasion of the empire pushed its borders northward; second, concurrently, the Oromo peoples migrated into the southern and central parts of what is now Ethiopia, establishing themselves and gradually assimilating into the empire’s political structure (Pankhurst, Reference Pankhurst2001; Ullendorff, Reference Ullendorff1961). Third, in the nineteenth century, after a failed attempt by Italy to colonize the empire, Menelik II expanded the empire’s southern frontiers into areas occupied by non-Semitic ethnic groups. While these expansions were framed as strategic defences against European threats, some viewed them as attempts at internal colonization.
Later, in the twentieth century, modernization campaigns under Emperor Haile Selassie included the promotion of the Semitic Amharic as a national language, which exacerbated some of the identity-based fault-lines of the empire. In an institutional sense, the empire learn to govern these fault-lines by adopting a quasi-federal structure in the nineteenth century, with regional leaders like the Sultans of Jimma and Afar enjoying autonomy under the Ethiopian monarch (Clapham, Reference Clapham2017). After a second failed colonial attempt by Italy in the twentieth century, Haile Selassie implemented another federal arrangement with Eritrea in the north. However, with an increasingly centralized approach to modernization, Haile Selassie dissolved the federation in 1962, sparking an insurgency that escalated into a civil war by the mid-1970s (Weldehaimanot, Reference Weldehaimanot2010).
Discontent with Haile Selassie’s rule led to a military coup in 1974 and the establishment of a communist state under the Dergue. The Dergue’s policies further fuelled ethnic tensions, culminating in its overthrow by ethnically organized rebel forces in 1991. This also signalled the culmination of the Eritrean insurgency, with Eritrea gaining independence in 1993. Around the same time, in 1995, Ethiopia formally ratified a new constitution that adopted, for the first time, a federal system based on ethnic territories, or ethnofederalism. This constitution created nine ethnically defined regions, and two charter cities, with additional regions established later.
This ethnofederal system aimed to address ethnic tensions by granting significant autonomy to ethnically defined regions. The Ethiopian Constitution recognizes two government tiers – federal and regional; local governments operate under the authority of regional governments. Importantly, the regional state governments are assumed to protect the right to self-determination of the ‘nations, nationalities, and peoples’ of the country, elevating ethnic groups to the stature of nations, with a recognized right of secession. Despite these formal concessions, ethnic tensions and secessionist movements continued to rise due to the dominance of the Tigrayan ethnic party, the Tigray People’s Liberation Front (TPLF), within the ruling coalition, the EPRDF (Clapham, Reference Clapham2017; Demerew, Reference Demerew2024).
In 2018, the accession of Abiy Ahmed as Prime Minister marked a significant shift, with the subsequent establishment of a new growing authority, the Prosperity Party (Prosperity). However, subsequent tensions between the new elites and the old guard escalated into full-scale military confrontation in November 2020 with the Tigray war. From 2020 to 2024, the Ethiopian state has been engulfed in conflicts with ethnonationalist movements on multiple fronts, leading many to question the viability of the country’s federal system. Despite the appointment of a National Dialogue Commission, no meaningful progress has been made toward constitutional reform of the country’s federal system.
Configuring federalism: towards a polycentricity framework
In this section, we review perspectives on what federalism is, what it can achieve, and how it comes about. We then develop an Ostromian framework of polycentric federalism, centred on the analytical parameters of devolution in administrative decision-making, peaceful competition among governance units, and individual choice of alternating governance structures. We illustrate how each of these parameters operates within the two case studies selected for this paper. To our knowledge, this is the first body of work that assesses African cases of federalism through a framework rooted in polycentricity.
Polycentricity is a system of governance with four components: i) multiplicity of overlapping autonomous decision-making units; ii) whose interactions are defined through constitutional rules; iii) whose functioning results in spontaneous orders generated by evolutionary competition; and iv) with the individual as the fundamental unit of analysis (Candela, Reference Candela, Lemke and Tarko2021; Ostrom, Reference Ostrom1973, Reference Ostrom1987; Stephan et al., Reference Stephan, Marshall, McGinnis, Thiel, Blomquist and Garrick2019). In this sense, polycentricity is a broader concept than federalism, since it accounts for concurrent administration at different levels of government. Federalism, in the ideal form, can be conceptualized as a specific manifestation of a polycentric order and one that encourages institutional learning (Candela, Reference Candela, Lemke and Tarko2021).
In Riker and Follesdal (Reference Riker, Follesdal, Goodin, Pettit and Pogge2012) classical definition, federalism is a Hobbesian political arrangement for increasing the mutual security of units. Such federal systems were promulgated in the nineteenth and early twentieth centuries, particularly to bring different constituencies together under one protective political umbrella (Mengie, Reference Mengie2016; Osaghae, Reference Osaghae2004; Riker and Follesdal, Reference Riker, Follesdal, Goodin, Pettit and Pogge2012; Sluga, Reference Sluga2013). Later, a liberal-Tocquevillian conception of federalism emerged to challenge this realist-Hobbesian explanation for federalism emergence (Dosenrode, Reference Dosenrode2010). In this liberal tradition, federalism is conceptualized as any social compact encompassing self-rule and shared rule, with a focus on the actual equilibrium of relations between different political forces (Elazar, Reference Elazar1991, Reference Elazar1996). The formal adoption of a federal system was dubbed as a ‘federation’, but federalism could operate as either a political-legal (formal) phenomenon, as a social phenomen, or as both (Elazar, Reference Elazar1985, Reference Elazar1991). This conception of federalism is somewhat aligned with the polycentric federalism framework we develop in this paper. However, because we are illustrating how federal institutional design precludes the purported benefits of federalism, we find it important to define federalism, akin to federation, as a formal constitutional choice (see Wagner, Reference Wagner2005). Polycentric federalism then allows us to assess institutional design in countries that have formally adopted federal systems.
Elazar’s (Reference Elazar1985, Reference Elazar1991) liberal approach to federalism emphasized the multi-centred, interdependent networks that emerge in the late-twentieth century as a contrast to the statist approach to state building in previous centuries: a novel analogy is drawn to the cultural exposition of the American cowboy, who willingly sells his labour but not his sense of autonomy or personal sovereignty to the ranch. Polycentric federalism is also aligned with this liberal tradition. Here, federalism is conceptualized as the existence of at least two government units with concurrent and limited authority over the same territory, on the basis of preset constitutional rules for exclusive and shared jurisdiction (Candela, Reference Candela, Lemke and Tarko2021; Hueglin, Reference Hueglin2003; Ostrom, Reference Ostrom1973, Reference Ostrom1991).
Much of the early literature on this Ostromian conception of federalism is centred on the case study of the American experience. In the constitutional underpinnings of US federalism, Ostrom (Reference Ostrom1973, Reference Ostrom1987, Reference Ostrom1991) perceives the makings of a Tocquevillian compact among different units of social organization, with the individual citizen at the centre of the choice matrix. Unlike previous federal constitutions, the US Constitution emerged through true ‘constitution-level thinking’, meaning that institutional design deliberately introduced constraints designed to address the fallibility of human nature as well as the weakness of democratic systems (Ostrom, Reference Ostrom1987). The US system brings together the concepts of federalism and polycentricity in ways that reinforce liberty, at least compared to other alternative constitutional choices.
Still, some scholars, particularly those influenced by the realist tradition, highlight the drawbacks of the US federal experiment. For instance, Riker (Reference Riker, Greenstein and Polsby1975) notes how US federalism empowered a minority of southern states at the expense of the national majority, in keeping with the political bargain that brought the structure into place. For Hueglin (Reference Hueglin2003), the Framers’ abandonment of Montesquieu’s ‘small republic’ federalism amounted to championing the federal government’s supremacy over the idea of a social compact among equal partners. Among those within the liberal tradition of federalism, Candela (Reference Candela, Lemke and Tarko2021) notes that the US federal system failed to extend economic rights to large groups of people and was thus unable to address this problem without bloodshed at a national level. For these reasons, some scholars (Dickovick, Reference Dickovick2014; Stepan, Reference Stepan1999) argue that the United States is an ill-fitting model for countries that are considering adopting federalism to alleviate group-based conflict.
We disagree at least partly with this assessment. The United States provides neither a perfect nor the only blueprint for approaching the polycentric ideal type. However, it provides a good starting point precisely because its polycentric orientation provides for more individual liberty compared to other alternative systems. In this vein, US federalism goes beyond the Hobbesian compact and introduces two constraints to make federalism an effective political instrument: a ‘compounding effect’ by means of requiring deliberative decision-making at multiple levels of government, and a ‘concurrent regimes’ effect, by means of empowering multiple territorially coterminous units of sovereign authority (Ostrom, Reference Ostrom1987). In effect, the citizen is at the centre of the federal system, ultimately expressing sovereign authority, through representatives at various levels of government that are both hierarchically and horizontally differentiated. Figure 1 presents a modelling of this ideal type of polycentric federalism, centering on the individual.
Figure 1. Federalism and the individual.
There are other federal systems that have approximated this ideal type. For instance, India’s federal system has been lauded for incentivizing peaceful competition, wherein the various states and territories have experimented with varying economic priorities during the transition to a market-led economy (Jha, Reference Jha2019; Marquardt et al., Reference Marquardt, Dasgupta, Höhne, Lederer and Sankhyayan2024). Administrative devolution and regional state autonomy allowed for the emergence of multiple economic hubs in the country, especially evident in sectors like technology, agriculture, and energy policy, helping to alleviate various social tensions and conflicts in the 1990s (Arora et al., Reference Arora, Kailash, Saxena, Suan, Suri and Vanaik2013; Jha, Reference Jha2019; Marquardt et al., Reference Marquardt, Dasgupta, Höhne, Lederer and Sankhyayan2024; Sharma and Swenden, Reference Sharma and Swenden2018). Though further away from the ideal type than India, South Africa has also instituted structures for competitive and collaborative governance through three tiers of government, national, provincial, and municipal, creating the institutional infrastructure for polycentric federalism to emerge (Plaatjies, Reference Plaatjies2021; Lieberman and Lekalake, Reference Lieberman and Lekalake2022). This is likely one reason why ethnic rivalries are generally mediated peacefully in the contemporary South African state.
In this sense, polycentricity can be thought of as a necessary condition for federalism to result in the purported positive institutional effect of conflict alleviation. In so far as it may alleviate group conflict, polycentric federalism does so by maximizing individual choice, and creating multiple channels of authority. Importantly, the design of the federal arrangement, not whether federalism is formally adopted, is the primary determinant of identity-based conflict alleviation. This focus on design goes beyond the breadth of literature that proposes formal federalism as a peaceful solution for African countries destabilized by ethnicity-based conflicts (Johnson, Reference Johnson2002; Osaghae, Reference Osaghae2004; Roach, Reference Roach2016, Reference Roach2023).
Where federalism scholars have engaged with institutional design, they have focused primarily on distinctions between centralized and decentralized forms (Deiwiks et al., Reference Deiwiks, Cederman and Gleditsch2012; Lijphart, Reference Lijphart2012; Plestina, Reference Plestina2019; Rodden, Reference Rodden2004; Tavares de Almeida, Reference Tavares de Almeida2006). However, decentralization does not necessarily lead to positive-sum outcomes, unless peaceful inter-regional competition is also prevalent. Indeed, as one of the cases we examine illustrates, decentralization could easily lead to balkanization rather than peace and stability (Christensen, Reference Christensen2021). Balkanization in this context refers to the violent break-up of a republic into several smaller, hostile, nation-based states, as in the Balkans (Danopoulos and Messas, Reference Danopoulos and Messas2019). Ethnofederal designs, in which each minority receives a territorial designation, are most prone to balkanization, as they have led to secession in 33 jurisdictions in the past 108 years (Roeder, Reference Roeder2009; Roeder and Rothchild, Reference Roeder and Rothchild2005).
In short, the institutional choice of particular interest is not necessarily the degree of decentralization but rather the terms of engagement among the different spheres of authority. It is unhelpful to view federalism synonymously with ‘decentralization’, or to assume that it sits at the opposite end of centralization or empire. Instead, federalism is to be located on a continuum predicated on ‘the effective combination of unity and diversity’ (Elazar, Reference Elazar1985: 23). We contend that the Ostromian conception of polycentric federalism is well aligned with this notion of combining unity with diversity, with the individual at the centre of the decision matrix. Indeed, federalism may exist at any point of a continuum between centralized governance on one end and decentralization on the other; but, more importantly, decentralized federal systems may also function alternatively along the typologies of balkanization or polycentricity, as illustrated in Fig. 2.
Figure 2. Federalism beyond decentralization.
Compared to the alternatives that result in balkanization or centralization, polycentric federalism is far better positioned to effect positive outcomes in governing diversity.
In configuring polycentric federalism to analyse federalism gaps in African states and beyond, we draw from the Ostromian tradition emphasizing the maximization of individual liberty and the importance of peaceful, evolutionary competition. The three parameters we develop for assessing comparative levels of polycentric federalism are devolution in administrative decision-making, peaceful competition among governance units, and individual choice of alternating governance structures. We develop each of these parameters in this section, configuring them to analyse federal institutional design in the Nigerian and Ethiopian cases. In doing so, we show how federal instutitonal design can preclude the purported benefits of alleviating ethnic conflict.
Administrative devolution
Devolution in administrative decision-making, with multiple centres of authoritative governance, is arguably a minimum requirement for polycentric federal design (Elazar, Reference Elazar1985).To forestall abuse of power, such design implements a system wherein ‘certain kinds of coercion require the joint and co-ordinated use of different powers or the employment of several means’ (Hayek, Reference Hayek and (Hamowy2011: 275). Because such devolution forces both division and sharing of power, it also approximates actual social equilibria, allowing for the market-preserving benefits of federalism (Weingast, Reference Weingast1995). With the other two parameters held constant, greater levels of administrative devolution result in higher levels of polycentric federalism.
Despite constitutional provisions, in both Ethiopia (before 2018) and Nigeria, party-based, clientelist networks dominate policymaking at all levels. In Nigeria, in particular, governance at regional and local levels is transformed primarily into an interest-based alignment with central government objectives (Onimisi and Osasona, Reference Onimisi and Osasona2021). Regional elites are incentivized to further their own careers by promoting the will of a dominant centre. In Ethiopia, the clientelist networks are less fluid and more dependent upon personal ties between ethnic elites (Yimenu, Reference Yimenu2022, Reference Yimenu2023).
EPRDF’s Ethiopia practiced top-down governance through a party culture of ‘democratic centralism’ (Fiseha, Reference Fiseha2024; Yimenu, Reference Yimenu2022, Reference Yimenu2023). This allowed the ethnic coalition party to maintain tight control over regional governments, ensuring that regional policies aligned with the party-state agenda. This centralization created regional disparities in economic development and administrative capacity and alienated various ethnic groups despite the constitution’s ostensible commitment to balanced ethnic representation. For instance, federal government intervention in the Oromia, Benishangul, and Somali regions directly contributed to the reanimation of ethnonationalist movements, such as the Oromo Liberation Front (OLF) (Yimenu, Reference Yimenu2023).
Although Nigeria’s three-tier federal design allows for more devolution on paper, the practical limitations mirror those of the EPRDF’s highly centralized system. Unlike the latter case, however, the Nigerian polity did not fall under one dominant party-state regime. Rather, centralized governance was facilitated through the country’s peculiar adoption of parliamentary procedures and party discipline, a carryover of British colonial influence, within a presidential framework. The combination of these institutional rules under the country’s federal system has incentivized elite political bargains that reinforce centralized control (Ifesinachi, Reference Ifesinachi and Onah2006; Nwabueze, Reference Nwabueze1983; Osaghae, Reference Osaghae1991; Suberu and Diamond, Reference Suberu and Diamond2003). More specifically, patronage between the president and the president’s party in parliament diminish the legislature’s independent ability to check executive power. The president has exclusive control over appointments, budgetary allocations, and the security forces and can use these levers to influence party affiliates in the legislature, undermining regional and local autonomy in parliamentary representation.
The lack of administrative devolution in each case is especially evident in fiscal policy. Politicization of resource distribution and development policies has reinforced centralized governance in both cases, undermining merit-based allotments (Endale, Reference Endale2019; Presbey, Reference Presbey2003). This is more visible in Nigeria, where the national government retains significant control over oil and gas resources, doling out funds to mineral-hosting communities, but keeping the larger share of revenues. Oil and gas constitute 85% of government revenue in Nigeria, with regional and local governments heavily dependent on fiscal allocations from the centre, where the executive branch exercises oversized influence. This reliance stifles the potential for local governments to develop independent, competitive economic policies.
In Ethiopia, the EPRDF lacked a mineral resource that could easily be allocated as rents to regional party affiliates. Instead, it announced a grand economic program of state-led development and established powerful state-led organizations that dominated banking, cement-production, and pharmaceuticals, and monopolized receipts of foreign aid (Clapham, Reference Clapham2017). Lack of specialized economic activity or policy experimentation across regions meant that regions were fiscally dependent on the centre. There was a level of unpredictability to fiscal allocations that could be instrumentalized to reinforce democratic centralism. Intra-party bargains were far more important in terms of determining fiscal allocations. For instance, the new ethnic bargains within the ruling coalition after 2007 caused significant changes in budgetary allocations to regions, whereas previous introductions of new formal allocation rules did not (Yimenu, Reference Yimenu2024). Far from an informal federal arrangement within the ruling coalition, the TPLF’s status as a first among equals, along with its centralized military and economic control, created a clear hierarchy in administrative decision-making (Demerew, Reference Demerew2024; Fiseha, Reference Fiseha2024).
During the transition from the EPRDF to Prosperity after 2018, Ethiopia’s federal system experienced higher levels of administrative devolution. This is because the party effectively abandoned democratic centralism, at least when it comes to administrative and economic policy. Although the EPRDF’s formula for fiscal grant allocations did not change, there was a marked increase in the transparency of allocations as well as in the inclusion of peripheral regions in the process (Yimenu, Reference Yimenu2024). Further, the transition saw the elevation of peripheral ethnic parties from regions like from external coalition partners to the centre of the party. One of the most striking outcomes was the increasing implementation of multiple administrative languages in border regions, as indicated in the Prosperity party bylaws. As such, if administrative devolution was the only parameter by which polycentric federalism would be assessed, Prosperity’s Ethiopia would be considered a marked improvement over its predecessor and over the Nigerian system.
However, Ethiopia’s federal system experienced two challenges during the Prosperity-led transition. First, increasing levels of personal rule by the new Prime Minister, Abiy Ahmed, contravened established and new institutional rules. Although the party-state was weakened and formal power was dispersed towards regions, the new multiethnic party faced increasing levels of personality-based centralization. It became necessary to further separate party organization from formal federal state structuring in order to secure the benefits of administrative devolution, a dynamic that is yet to be achieved. Second, increased administrative devolution coincided with a marked decrease in the other two parameters of polycentric federalism, as we discuss below.
Peaceful competition
Evolutionary competition or competitive bargaining between different authority structures is central to polycentricity (Candela, Reference Candela, Lemke and Tarko2021; Eusepi and Wagner, Reference Eusepi and Wagner2010; Ostrom, Reference Ostrom1973, Reference Ostrom1991; Wagner and Yokoyama, Reference Wagner and Yokoyama2013). Accordingly, competition becomes an important criterion for achieving polycentric federalism. More importantly, such competition must be rules based and predictable to remain peaceful. Violent competition is the clearest indication that the second criterion has not been established through institutional design.
In both Nigeria and Ethiopia, centralization of power has stifled the type of inter-regional competition that fosters innovation and more responsive governance (Fiseha, Reference Fiseha2024). In Nigeria, the Federal Character Principle was introduced into the country’s federal structure to balance power between the country’s North and South regions. While intended to promote national unity and equitable representation, the principle often prioritizes ethnic and regional balance over local autonomy, leading to governance practices that can overshadow merit and efficiency (Adeosun, Reference Adeosun2011; Ezeibe, Reference Ezeibe2010). For instance, in the educational sector, the application of ethnic quotas in admissions policies has hindered regional authorities from making autonomous decisions that reflect local needs and priorities (Joshua et al., Reference Joshua, Loromeke and Olanrewaju2014).
In effect, rather than encouraging healthy rivalry and cooperation, the Federal Character Principle often solidifies ethnic divisions (Adamolekun et al., Reference Adamolekun, Erero and Oshionebo1991; Mustapha, Reference Mustapha2007). At worst, its implementation has frequently been perceived as discriminatory (Asaju and Egberi, Reference Asaju and Egberi2015; Bello, Reference Bello2012; Majekodunmi, Reference Majekodunmi2013; Okolo, Reference Okolo2014). These dynamics have entrenched stratified levels of Nigerian citizenship, making distinctions between ‘indigenes and non-indigenes’ in each region; consequently, ‘non-indigenes’ to a region cannot hold national appointments based on their place of residence (Faboye, Reference Faboye2022). The Federal Character Principle, in effect, serves to diminish the spirit of open competition among different authority structures.
The central government’s monopoly over oil revenue has also created incentives for more violent contention. Although Section 162(2) of the 1999 Constitution mandates at least 13 percent of oil revenue to be allocated to producing states under the derivation principle, institutional rules privileging the central government have caused inconsistencies. As a result, the South-South region has argued for resource control as a means to ensure fair compensation, while non-oil-producing states have resisted changes in institutional rules (Okolo and Akpokighe, Reference Okolo and Akpokighe2014). Much of this competition has thus far been institutionally contained, perhaps demonstrating slightly higher levels of the peaceful competition parameter in the Nigerian case. However, the grievance has also reanimated secessionism movements in the country’s Southeast and Southwest regions, prompting military intervention from the Nigerian government, which is intent on crashing down on any potential re-enactments of the 1967–1970 Biafran War.
In Ethiopia, the EPRDF failed to promote peaceful competition due to the dominance of the TPLF, which marginalized other ethnic groups and centralized power, a dynamic we discussed above. This stifled the capacity of regions to compete or innovate and instead led to heightened ethnic tensions and separatist movements (Fiseha, Reference Fiseha2024). After the TPLF was ejected from federal power in 2018, the transition towards the Prosperity-led government provided a new opportunity for healthy inter-regional competition, creating an institutional opening for transition towards polycentric federalism.
Due to the incentives of ethnofederal institutional design and provisions for territorial secession in Article 39 of Ethiopia’s 1995 Constitution, the outcome of the Prosperity transition was an increase in inter-regional competition, but with a complete abandonment of the peaceful component. During the 2018 transition, the collapse of personal elite networks, the removal of key elites from federal power, and the splintering of the country’s centralized security apparatus gave rise to competitive violence, with clear ethnic undertones. By 2018, almost every ethnic region had established and trained special regional police forces that functioned more like militias; in some cases, such as Oromia, the regional states resembled ‘nations in waiting’ (Demerew, Reference Demerew2024). The ethnic regions turned to these militias, as well as to other informally organized violent groups, in a bid to increase their bargaining power and positioning in the new informal hierarchy. The inability of the federal system to manage peaceful competition was most clearly demonstrated in the 2020 conflict between the TPLF and the new Prosperity-led government, which escalated into a civil war rooted in ethnic and political grievances. As such, Ethiopia arguably experienced an even further decline in the prospect for peaceful inter-regional competition after 2018, mostly as a result of the incentives of ethnofederal institutional design.
Individual choice of alternating governance structures
Polycentric federalism emphasizes individual choice of alternating governance structures; implicitly, this framework relies on methodological individualism and rejects methodological collectivism. Methodological individualism is central to our framework because it is only by taking individual choice as the primary unit of analysis that evolutionary competition between constituent units can align institutional incentives of power with polycentricity. Without the centring of individual choice, any federal region may choose to exercise unilateral authority to ‘control the movement of men and goods across the frontiers of its territory’ (Hayek Reference Hayek and (Hamowy2011: 275). The only way to limit the ability of constituent units to pursue such policies would be by prioritizing individual choice over group preferences within the federal design.
Critics of methodological individualism would argue that it reduces deep cultural complexities to mere elements of individual choice, negating the group experiences of grief or mistrust that inform contemporary bargains (Anderson, Reference Anderson2014; Maiz, Reference Maiz1999). We would respond to these critiques by specifying, as Hodgson (Reference Hodgson2007) does, that methodological individualism need not entail a complete rejection of the structured social relations that inform individual choices and decisions; rather, it assumes that the ‘individual is a social being, enmeshed in relations with others’ (ibid, p. 221). Consequently, individual choice serves as a parameter that incentivizes various concurrent governance authorities within a wider polity to compete. The result, as Hayek (Reference Hayek1979) envisages, is ‘the transformation of local and even regional governments into quasicommercial corporations competing for citizens’ (p. 146). This entails the removal of formal barriers to exiting or entering diverging realms of governance authority. In this sense, the best operative measure of individual choice is arguably the ‘right of exit’.
The right of exit criterion, often called ‘foot voting’, incentivizes individuals to seek out information and use it wisely; in this sense, it can be superior to the ballot box as a form of democratic participation as it requires far more thoughtful decision-making (Somin, Reference Somin2020). As such, we find that the right of exit fits squarely within the individual choice parameter of polycentric federalism. Of course, individual choice sometimes goes hand in hand with the preceding criterion of peaceful competition. In the absence of peaceful competition between different governance units, it is difficult to imagine much innovation, differentiation, or specialization in governance practices or economic policy. So, to the extent that the Nigerian and Ethiopian federal systems (the latter, especially under the EPRDF) significantly limited competition between regions, the prospects for individual choice were also very limited for each case.
As a result, the variance between the two cases in levels of individual choice can be based, at a primary level, on comparative consideration of legal provisions rather than the availability of multiple options for public service delivery or governance structure. At the legal-political level, individual Nigerian citizens have far more individual choice of alternating governance structures, if only because there is no formal restriction on the right to exit. Further, because the country formally recognizes local government authority, the system provides a viable infrastructure for facilitating freedom of choice and even the right of exit.
By contrast, Ethiopia’s ethnofederal system makes it nearly impossible for citizens to practice individual choice of alternating governance structures. By setting ethnicity as the basis of citizenship, the country’s constitution recognized citizens only in terms of their group rights. In some cases, the regional constitutions also expressly granted ownership of regional territories to a particular ethnic group (Van der Beken and Dessalegn, Reference Van der Beken and Dessalegn2020). The rigidity of this institutional design locked individuals into specific regional governance models that may not align with their preferences or needs, especially in areas where ethnic minorities are present (Abbink, Reference Abbink2012; Fiseha, Reference Fiseha2024). In effect, the system made it difficult for people to move freely between different regional governance structures. Ironically, Prosperity’s dismantling of the party-state and the previous culture of democratic centralism emboldened ethnonationalist armed movements and further diminished prospects for individuals to exercise their right to live under a governing authority of their choosing. Indeed, as the Prosperity-led government finds itself engaged in armed conflict with ethnonational armed movements in Tigray, Amhara, and Oromia, the country’s three main ethnic regions, millions of individuals have found themselves internally displaced (Demerew, Reference Demerew2024; Gichamo, Reference Gichamo2023). Often, there is no legal recourse for displacements because institutional design assigns ethnic tiers for participation and membership in regional governance structures.
Discussion
As we have shown, institutional designs and rules in Nigeria and Ethiopia are in many ways opposed to the parameters of polycentric federalism. This, in turn, makes it impossible to achieve the peace-making benefits of federalism. On the parameter of administrative devolution, Nigeria performs slightly better than EPRDF’s Ethiopia. On the parameter of peaceful competition, both cases have similar limitations on the practical possibilities of inter-regional competition. On the parameter of individual choice of alternating governance structures, the Nigerian system, on paper, provides a much more viable structure than that of Ethiopia, under the EPRDF.
We see marked levels of decentralization in Ethiopia, amidst the transition to the Prosperity-led government in 2018. While this transition slightly increases administrative devolution in the federal system, the constitutionally bounded ethnofederal design continues to undermine peaceful inter-regional competition and individual choice. With the collapse of the previously dominant party-state regime, the prevailing institutional design has incentivized violent rather than peaceful competition between identity-based groups. The prospects of individual choice and the freedom of exit, always feeble under ethnofederalism, were further weakened, with violent threats to ethnic out-groups in increasingly hostile ethnic communities and regions. As a result, the outcome of decentralization under the Prosperity-led transition in Ethiopia more closely resembles the balkanization ideal type rather than that of polycentric federalism.
A number of policy interventions may be proposed to help reorient institutional design in each of the two cases. In Nigeria, pre-colonial governance models may provide some informal institutional codes for bolstering polycentricity (Oyerinde, Reference Oyerinde2022). The Nigerian state has made some recent progress in granting greater fiscal autonomy to local tiers of government. In 2024, the Nigerian Supreme Court issued a ruling allowing the country’s 774 local councils to manage their finances and resources without state interference, promoting better accountability and local service delivery (Ogubike, Reference Ogubike2024). We suggest the full implementation of this ruling and further provisions for increasing the autonomy and professionalism of local tiers of governance. Additionally, a return to a full parliamentary system and an abandonment of the Federal Character Principle may be difficult but necessary undertakings over the long term.
In Ethiopia, it is necessary to move the basis of federal institutional design away from group-based factors. While allowing for more devolution, ethnofederalism directly undermines two of the other parameters of polycentric federalism. To reimagine ethnic representation, it is useful to begin by categorizing ethnic groups as informal, localized institutions that play a role in transactional processes (Amoako and Lyon, Reference Amoako and Lyon2014). Ethnic representation can be achieved without formalizing ethnic institutions as a second tier of government. For instance, non-territorial sovereignty would allow individuals of a certain ethnic group to practice self-governance in local or sub-zonal contexts, without necessarily increasing the national political salience of ethnic difference (Van der Beken and Dessalegn, Reference Van der Beken and Dessalegn2020).
Lessons can also be drawn from Belgium and Switzerland, which implemented group-based self-governance within federal systems without entrenching territorial and linguistic divisions (Burgess, Reference Burgess2006; Elazar, Reference Elazar1991). In these cases, institutional rules have allowed for urban zones with non-territorial arrangements (e.g., Brussels) and ensured coordination across community, regional, and federal levels. A final policy proposal is to further separate party organization from formal federal state structuring. All of these changes would require significant constitutional reform.
The three parameters of polycentric federalism are not necessarily in competition with one another; in fact, they may reinforce each other. For instance, high levels of individual choice will likely cause more competition between constituent governance units, whereas high levels of devolution provide a wider realm of options upon which individual choice can be reflected. It is possible that a federal system can be considered polycentric, despite lower levels of devolution, as long as it has very high levels of peaceful competition among governance units and no restrictions on individual choice or the right of exit. Presumably, in this hypothetical scenario, there may be one fundamental policy area over which regional governments can compete, and individuals are free to choose between different policy packages with regard to that one area. The parameters operate similarly to how regular elections, civil society, and minority rights may operate, for example, as parameters of democracy.
Still, there will be trade-offs for increasing any one of the three parameters. For instance, high levels of inter-regional competition over economic policy could lead to a short-term contraction in economic growth at the national level. Similarly, high levels of devolution in education policy could diminish the federal government’s ability to develop coordinated education strategies, reducing the global competitiveness of the country’s labour. Devolution could also increase the transaction costs of inter-regional exchange. For instance, if education policy devolution creates enclaves where only certain languages are taught, then the likelihood of inter-state migration or trade may decrease; this could also make the country less attractive to foreign investment.
This is why we do not argue that polycentric federalism would necessarily increase economic growth, the quality of education, or even democracy at a national level. Rather, in line with the liberal tradition, we argue that polycentric federalism as a formal institutional design is best placed to maximize liberty compared to alternative federal designs. In increasing liberty, polycentric federalism also makes peace more achievable, particularly in societies with identity-based conflicts. This is highly relevant in the context of African states, where scholars have often proposed federalism as a policy prescription, without necessarily conceptualizing the type of federalism that may help alleviate conflict, as we have done here.
Open access
