Ethiopia's Contested Federalism: How to Deal with Cleavages?

Introduction and Background

Whether it is possible to ensure democracy, stability, peace, and social cohesion in countries with deep divisions and, if so, through what institutional arrangements is one of the central political issues of our time.¹ This is particularly the case in many diverse countries in Africa where nation-building is linked with coercive and arbitrary processes by which the same communities are subdivided into different countries by artificial colonial borders, with ethnonational- and regionally based cleavages contesting such borders.²

In some African countries with deep cleavages, the state continues to suffer from structural problems where the central government is often accused of centralization of power, promoting narrow and factional interests, corruption, abuse of rights and marginalization of the bulk of society. Thus, claims for accommodation, ethnic conflicts, civil war, threats of secession, and state fragmentation remain major challenges. Some postcolonial African countries attempted to address these challenges by resorting to some form of federation and autonomy, but with the exception of Nigeria, all such efforts collapsed within a decade of their establishment.³ The failure of the federal experiment resulted in centralized unitary governments, imperial presidents and one-party rule.⁴ The respective federations failed because they were confronted by strong one-man leadership that took the perspective that federalism would lead to state fragmentation and was thus opposed to their own vision of centralized nation building.⁵ Many political leaders across a diverse continent thought federalism in the context of artificially drawn borders might lead to polarization and may in the end put territorial integrity at stake.⁶

After the end of the Cold War, however, there was resurgence in the use of federalism and devolution in Africa. For some it became a means to “domesticate the Leviathan”⁷ by transferring power from the all-powerful center to subunits, diffusing power into many centers. In others, federalism and devolution go beyond diffusing power and aim to manage territorially based and politically-mobilized cleavages. Four main African countries (Nigeria, Ethiopia, South Africa, and Kenya) have used federalism and devolution to achieve either or both of the above objectives.

Ethiopia is one of the most diverse countries in the Horn of Africa with an estimated 110 million people and many politically mobilized ethnonational groups. Although it had centuries of experience with “devolved autocracy”⁸ it remained a centralized authoritarian state for the most part of the last century and went through a civil war that came to an end with the defeat of the military junta in 1991. Ethiopia went through a transitional process (1991-1995) dominated by the Ethiopian People’s Revolutionary Democratic Front (EPRDF 1991-2018), a coalition of four ethnonationally based parties,⁹ and adopted a federal system, with the regional states mainly based on language. The principal objective of the federal system is ensuring self-government for ethnonational groups in response to the “nationalities question” as framed by the Ethiopian Student Movement in the 1970s.¹⁰ Self-government has previously been a victim of democratic centralism (as the centralized party controls all affairs of the state-making self-government farce) and a developmental state where the federal government focused more on economic growth, pushing democracy and self-government to the sidelines.¹¹

The federal system under EPRDF operated under three major unwritten frameworks: a big man (Meles Zenawi, nicknamed by Clapham “the philosopher-king of the EPRDF”),¹² democratic centralism (his main tool), and the vanguard party. Frustration with centralization and an increased sense of alienation led to violent protests for three years (2015-2018). particularly in the Oromia and Amahra regions, forcing former Prime Minister Hailemariam Desalegn from the Southern Coalition of the EPRDF to resign and precipitating the coming to power of Abiy Ahmed from the Oromo People’s Democratic Organization (later renamed as Oromo Democratic Party or ODP). To its credit, the new government responded to demands from the South. Two new states, The Sidama and South West Ethiopia People’s State, emerged. The Sidama’s quest for self-government is an age-old demand that was suppressed for decades. The southwest is too far to be managed from Hawassa, the Southern Nations Nationalities Regional state’s capital. There was thus hope that the new leadership that came to power in 2018 would reverse the centralization as Oromia, the regional state where Abiy comes from, called for end to federal intervention and centralization.¹³ Yet, centralization and authoritarianism have continued in a new form and, as will be shown later (in section two), the federation has informally turned into a decentralized unitary system, triggering sub-state nationalism to a new level and leading to a high risk of falling apart.¹⁴

Despite a promising start, the Prosperity Party’s (PP, the ruling party that succeeded the EPRDF following the withdrawal of the Tigray People’s Liberation Front or TPLF) increasingly authoritarian and centralizing policy emerged, threatening self-government. Although Ethiopia has been at war since the coming to power of the new leadership in 2018, with increasing interethnic and interregional conflicts, the war in Tigray remains a major factor behind Ethiopia’s current instability.¹⁵

The claim on the part of the federal government is that the TPLF “attacked the Ethiopian army base (the Northern Command) that was based in Tigray” on November 4, 2020 and triggered in response what it calls “law enforcement operation”¹⁶ to bring the culprits to justice. However, it is now clear that the TPLF and the Federal government have been at war since Abiy assumed power, and that what happened on November 4, 2020 is the culmination of what was already happening since 2018.¹⁷ At the core of this is an ideological crisis within EPRDF, a power struggle among the coalition members and a lack of competent leadership within the ruling party. The EPRDF was dissolved prematurely and rebranded as the PP and turned the member coalitions into branches, centralizing the party structure without enough consultation and consensus. The TPLF saw that as a step towards centralization that would lead to dissolving the federal system and declined to join as the process liquidated the autonomous position of the members of the coalition. The federal government continued to target the TPLF, accusing it of human right violations while the TPLF responded by characterizing the PP as a centrist regime. As illustrated later, there is a clash between state-led nationalism and its new narrative of nation-building that considers the EPRDF (1991-2018) era as a disruption to both the centralized nation-building project that commenced in the early 20^th^ century on one hand,¹⁸ and the Tigray and Oromo-based ethnonationalism on the other, demanding genuine self-government or loose confederation. The political elite’s visions for the country thus significantly vary and in the absence of political settlement, this leads to conflict. The nature of the political cleavage in Ethiopia, as elaborated in the next section, remains very distinct.

This article discusses the distinct nature of the political cleavages in Ethiopia and shows that an integrationist presidential federation, a choice the centralist elite seem to be pushing,¹⁹ is problematic, and considers the consociational parliamentary federation to be more relevant. The latter certainly is not risk-free, but there is little that one can do when one faces deeply entrenched cleavages, as is the case in Ethiopia. A key entry point is an understanding of the nature of the cleavage that has continued to challenge the nation-state and occupied political and expert attention to a new form of accommodation.²⁰ Thus, the first section provides a brief account of the nature of cleavages. The second section demonstrates the rise of a new authoritarian elite in Ethiopia which partly explains the widespread discontent and the raging civil war in parts of the country. The following two sections explain the integration and accommodation features as developed in the comparative institutional design literature. The final section draws some comparative conclusions.

1. Territorially Based Cleavages

Politically mobilized cleavages continue to threaten the nation state. Following the end of the Cold War, such cleavages caused what Arend Lijphart dubs a “wave of ethnic conflicts,"²¹ instead of the promised “third wave of democracy."²² Mobilized ethnonational minorities are “regionally concentrated ethnic groups who once enjoyed or aim to enjoy political autonomy and have become part of states in which they constitute an ethnic minority through conquest, annexation, colonization or incorporation during the coercive process of nation building."²³ They mobilize politically around assertions of national identity and self-determination. The goal of such mobilization is to recover the extensive self-government that they claim to have enjoyed historically or they aspire to have now. The degree of self-government they seek ranges from autonomy, to national self-government, to independent statehood. Countries that have politically mobilized ethnonational groups cannot assume to have stable territory. The demands of such groups are framed and entrenched in relation to a particular territory and the very unity and territorial integrity of the state is put to the test.²⁴

Ethnonationally based mobilization is a very potent force that, if not managed carefully, can result in fragmentation. It has resulted in the formation of some 27 states that have joined the United Nations following the end of the Cold War.²⁵ In the 1960s and 70s, nearly all major schools of thought (liberalism, socialism, globalization, modernization), regardless of their different viewpoints predicted that ethnonational minorities would wither away through liberalism, socialism, “melting pots,” and assimilation. Some have called this the “post national illusion”²⁶ and counselled actors to properly understand this force and design appropriate institutions and policies to manage it. Territorially based and politically-mobilized cleavages continue to challenge the process of nation building both in the developed (Canada, Spain, Belgium, United Kingdom) and the developing world. Within Africa, Nigeria, Ethiopia, South Africa, and Kenya, despite differences in the degree of mobilization, all continue to face challenges related to group-based cleavages.

According to Horowitz, a “deeply divided society” refers to cases in which identity-based politics gains a high degree of salience, exceeding that accorded to alternative forms of political mobilization such as ideology, class, and gender, and the relationship between groups is affected by deep levels of mistrust and antagonism, making it less cooperative.²⁷ As argued by Sujit Choudhry, “a divided society is not merely a society which is ethnically, linguistically, religiously or culturally diverse … it is hard to imagine a state today that is not diverse."²⁸ What makes a society divided is when the differences are politically salient and an identity-based distinct group uses them as a basis for political mobilization. As such, identity becomes the prime source of political mobilization around which political claims for recognition, resource control, accommodation, and self-government are framed, political parties are formed, elections are contested and governments are composed.²⁹ These forces affect the process of constitution making and constitutional design. A distinct identity-based political mobilization demands for a more autonomous self-government, while less mobilized groups could be managed through integration and other, softer options.³⁰

As a result of cleavage, Ethiopia continues to face the threat of fragmentation. Ethiopia lost Eritrea and the threat of secession is still a problem as it is a goal harbored by many national liberation movements such as the Oromo, the Ogaden/Somali, and the Tigray, among others.³¹ Some ten ethnonational groups that used to administer themselves at local government level in the South have, following the winds of change in 2018 demanded constituent unit status and only two have succeeded. The war between the federal government and Tigray, although it has multiple causes, is very much related to Tigray’s age-old demand for genuine self-government³² being triggered by the rise of a new authoritarian centrist elite. Ethnonational groups are thus demanding more, not less, even after two and half decades of federal practice.

Nigeria’s splits—between north (Hausa Fulani, predominantly Muslim) and south (the oil-rich Niger Delta), west (Yoruba) and east (Igbo), and Muslim and Christian—remain visible, particularly during presidential elections. The split is kept in delicate balance based on an unwritten convention that guides presidential elections to commonly rotate the office of the president between the north and south.³³ Nigeria faced a secession threat from Biafra towards the end of the 1960s and Igbo nationalism has not withered away. There is also an insurgent group in the Niger Delta (Ijaw) that has the potential to cause trouble. In reaction to an extremely centralized federation and suppressive marginalization during the military era, a demand for “true federalism,” that grants genuine political autonomy and resource control is now fully expressed by ethnoregional groups (Yoruba, Igbo, Niger Delta region) in Nigeria.³⁴

Kenya has its northeast frontier Somali problem and the Rift Valley region where ethnic conflicts have erupted on several occasions, in addition to the Indian Ocean coastal region where there is strong demand for self-rule. Ethnic rivalry for controlling the overbearing presidency in a winner-take-all politics resulted in ethnic conflicts in 2007 and subsequent elections, and continues to divide Kenyans along ethnic lines.³⁵

A precaution is in order here. First, cleavages do not automatically translate into a political project. Political and economic injustice that is reinforced by deep cleavages, elites that frame the issues, and a state response or the lack thereof all play their own roles in the process of transformation. As argued by Anderson and Choudhry, cleavage or diversity is not destiny. It requires agency—political entrepreneurs that read into the political dynamics of the country and frame the issues in a way that appeals to their audience. These political entrepreneurs are critical in “framing of narratives,"³⁶ articulating perceived or real exclusion and subjugation. “Political entrepreneurs are critical to the success of political mobilization by framing the case (of ethno nationalism), developing strategies, and marshalling resources."³⁷ They articulate alternative narratives that seek to deconstruct any centrist narrative about the past, present, and future (fears and possible hopes) addressing the grievances and the entitlements of the population they claim to represent with appeasements including territorial entitlements. Through this, an ideology is framed and a plan set for concrete action.

Second, Gurr argues that conflict between competing nationalisms typically escalates in stages, and it is here that one finds the link with state policy. The secessionist demand of Eritrea’s elites in the mid-1980s was limited to a restoration of the federation (1952-1962) that was unilaterally abolished by Emperor Haile Selassie in 1962. With the military regime’s failure to respond and its resort to violence, radical proposals such as secession emerged. Thus, cleavages often start with nonviolent modest demands, and when regimes fail to respond, evolve into violent protest and finally rebellion. This escalation occurs through a pattern of demands and responses: nonviolent protest is met with a lack of political responsiveness, which in turn leads to violent protest, which is met with a violent reaction, and which then leads to rebellion and an armed conflict and civil war. State policy and action or inaction is thus a major factor that can escalate or moderate ethnonationalism. Territory also remains key aspect of mobilization. Ethnonational-based cleavages aspire for self-government over a defined territory at times claimed as homeland, and even wish it to be territory over which they exercise statehood if they secede. On the contrary, the state insists on territorial integrity, unity, and full control over each territory. Contestation over territory is thus an integral part of the political debate. Thus, countries with deep cleavages cannot presume to have stable territory, a condition normally assumed in countries that do not have such politically-mobilized cleavages.

Third, the process of transformation from a diversity that is not mobilized to a political project is heavily associated with the nation-building project pursued by the central government, including the forceful annexation of previously semi-autonomous territories, the imposition of a common national identity including language, and the centralization of power and resources. Left outs from the process design a defensive response to the central state-led project of nation building.³⁸ In other words, ethnonational-based cleavage and political mobilization is often a reaction to centrist elites' project and a search for finding a space for political power and identity.³⁹ It is a substate nationalism framed in reaction to the central government’s nationalism. There are thus competing nationalisms within the nation-state that, if not addressed, could lead to violence, civil war, and state fragmentation. Both are pursued in the name of nationalism and have the potential to fuel the passions and emotions on both sides of the political spectrum to cause political instability and state collapse. One may call this a clash of nationalisms, one pursued by the central government in the name of nation building, patriotism, civic nationalism, unity, and ensuring territorial integrity, and the other by the ethnonationalist elites' aspiration for self-government and autonomy.

The civil war in Ethiopia (1974-1991) fits well into the above framework. The centrist military regime fought the ethnonational-based liberation fronts in the name of ኢትዮጵያ ትቅደም/ īteyop̣eyā teqedame (Ethiopia First) and ensuring the territorial integrity of the country while branding them as secessionists. The ongoing war in Ethiopia is framed by the ruling elite as “Ethiopia shall prevail.” The ethnonationalist elite accuse the centralized nation-state as a mask whereby the centrist elites' culture, language, religion can become the national culture, language, or religion.⁴⁰ In other words, the group—however narrow its base—which controls the state uses it not only to marginalize others from power and resources but also uses state institutions and policies to “promote, consolidate and create a privileged position with respect to its identity and its manifestations. The state is defined as the expression of the group’s nationhood."⁴¹ The non-material aspects of conflict, such as the search for dignity and collective self-esteem,⁴² a rich history (“we were great and we want to be great again”), and claims by ethnonationalist groups to regaining lost social status (dignity) in response to historical traumas fuel the political and resource conflict.⁴³ As Connor points out, “Men do not allow themselves to be killed for their interests; they allow themselves to be killed for their passions."⁴⁴ It is not surprising then that the legitimacy of the government, its institutions, and the values upon which it is established remain one of the sources of tension and at times the cause of its terminal crisis.

Fourth, majoritarian-based democracy in deeply divided societies could generate problems, necessitating the need for other inclusive options. This is an area where institutional design matters a lot. This is particularly acute if the drivers of the central government’s version of nationalism are not a majority. In many cases, as in the four African case studies, there is no dominant ethnonational group that enjoys a demographic majority and could claim to have democratic majority to pursue its goals. This becomes a clash between a minority that has state resources at its disposal and attempts to impose its will on other minorities. Yet even if the central government’s project of nationalism enjoys a majority, it pits a permanent majority against a permanent minority with no hope of becoming a majority. Arend Lijphart brought this tension to the fore in post-Saddam Iraq. Majoritarian democracy applied to Iraq would mean “a national government mainly or exclusively Shi’ite majority that excludes Sunnis and Kurds … and it will be naïve to expect such minorities condemned to permanent minority to remain loyal or constructive."⁴⁵ With this comes the question of why ethnonational minorities under perpetual rule of a majority could be expected to be loyal to such regime and stay in the union. The general assumption of majoritarian democracy, that the rulers alternate such that today’s political majority will become tomorrow’s political minority, does not hold true in deeply divided societies. In this context, majoritarian institutions may suffer from a legitimacy crisis in which the decisions of the majority are not accepted by the minority. The values of the centrist state such as the flag, the national anthem, public policy, the media, and language and cultural policies that are instruments of centralized nation building are deeply resented and rejected by ethnonational minorities. Thus, alternative theories such as consociational democracy have been recommended when the political system faces deeply divisive cleavages. Instead of having winners and losers, consociational democracy brings major political actors together either on equal footing or through proportional arrangements to power and insists on consensus decision-making on fundamental issues. Those left out in the majoritarian democracy become decision makers through power sharing, reducing the potential for conflict. Distinct institutional arrangements thus matter in managing cleavages and reducing conflicts.

2. The Rise of New Authoritarian Centrist Elite and Its Impact on Self-rule

It is now time to examine the policies and decisions of the rising, new, centralizing elite. The first indicator of authoritarian centralization in Ethiopia is the federal government’s action and the removal of regional state heads. Since the new government came to power in 2018, many of the leaders of the regional states have been removed by the ruling party—in disregard of regional states' mandate to self-rule and the people’s right to elect their own leaders⁴⁶ —imposing its will by force including waging civil war in regional states. This is happening despite popular support for democracy and federalism as indicated by Afrobarometer.⁴⁷ Instead of providing political solutions to major issues (e.g., a more inclusive political system, more working federal languages, transition to democracy, fair sharing of resources and genuine self-rule have been, among others, the most popular slogans of the popular protests since 2015),⁴⁸ the new trend is to use force to suppress popular demands. From Somali region⁴⁹ to Sidama,⁵⁰ Oromia,⁵¹ Tigray, and Wolayta,⁵² the actions taken by the federal government speak for themselves: a one-party show, military rule (including the use of emergency decree to remove regional state leaders and suppress public demands), violence, excessive use of force,⁵³ massive abuse of human rights,⁵⁴ political killings and imprisonment of key opposition political leaders (including Jawar Mohammed,⁵⁵ Lidetu Ayalew, Bekele Gerba, several senior Oromo leaders, Eskinder Nega—many of them released later—and thousands of junior opposition party members) and restricting media outlets and internet.⁵⁶ The federal government has issued several emergency rules imposing military rule (otherwise called “command posts”) in regional states such as Oromia,⁵⁷ Tigray, Benishangul-Gumuz,⁵⁸ parts of the Amhara⁵⁹ and the South⁶⁰ but only a few of them have been submitted to parliament for approval.⁶¹ Command posts imply that civilian rule is being suspended and replaced by military rule and, owing to its frequency and the wide geographic coverage, it has become the new normal. It is as if militarism has replaced federalism.

More serious political parties that pose electoral challenge to the ruling party, such as the Oromo Federalist Congress (OFC) faction of the Oromo Liberation Front (OLF) and TPLF are either marginalized or branded as terrorists. These are the major parties that are calling for more inclusion at the center and genuine self-government or even confederation. The marginalization has thus a peculiar dimension: it pushes the ethnonational-based parties and creates favorable ground for parties that opt for a more centralized form of government.⁶²

The second indicator is the ideological framework of the ruling party. The government’s close advisor stated “we will continue to implement Menlik’s (1889-1913) nation building project that was disrupted by EPRDF in 1991."⁶³ This was a centralized Ethiopia that liquidated quasi-autonomous kingdoms that existed for a long time through brutal and coercive processes that became the source of Ethiopia’s political agony in the last century. The paradoxes inherent in this process are well known, ranging from those who think it was a normal process of nation building⁶⁴ to those who think it was empire building and a part of colonialism.⁶⁵ In between the two extremes, one finds those who describe that era as “prison house of nationalities”⁶⁶ and advocate for a new social contract that allows for a more inclusive center and genuine self-rule, be it in the form of a federation or confederation.⁶⁷

Reinforcing this development is the transformation of the EPRDF to PP under Abiy that turned former relatively autonomous coalition and affiliate members of the EPRDF into branches. The party’s internal rule provides that members of PP in regional states are branches of the central leadership in Addis Ababa, no longer autonomous units as they were before.⁶⁸ Applied in a federal context this means that regional states are branches, not autonomous bodies. Thus, since the establishment of PP, Ethiopia is effectively now a unitary decentralized state, not a federation. Besides, the new equivalent that advocates for the return to the old nation-building era is Abiy’s “መደመር/madamare.” The ongoing war in Tigray is sloganized by its supporters and media activists as “Ethiopia shall prevail,” the new version of Derg’s (1974-1991) ኢትዮጵያ ትቅደም/ īteyop̣eyā teqedame (Ethiopia First). Prime Minister Abiy recently rejected the concept of Ethiopia as a pluri-nation state as outlined in the preamble of the 1995 Constitution and said “we are one people.” Hence the critique, as aptly explained by Trueman, that the “imperial narrative is being recycled” in a new form through Abiy.⁶⁹ This new trend was partly tempered following the intra-party crisis and power struggle between the ADP and ODP wings of PP, shifting the balance in favor of narrow circles of ODP after deposing senior members of ADP during mid-2022. It is yet to be seen what this shift means, but so far, the new regime is fully backing the pre-1991 narrative. This is not surprising. Abiy’s induction speech to government officials in his early days as prime minster included the line “when I was seven years old my mother told me that I will be Ethiopia’s seventh king”.⁷⁰ One cannot expect democratization and genuine self-government under a leader who thinks he is Ethiopia’s seventh king. A centralized kingdom is antithesis to federalism and self-government.

The third indicator is a new policy that regulates regional state police. The Constitution empowers regional states to establish their own police to ensure peace in their territory (Article 52).⁷¹ However, a recent policy document of the federal government shows that the federal government has proposed to dissolve the regional state police alleging that it has been heavily militarized beyond the requirements of its mission and is becoming a threat to peace and security.⁷² The document further states that the “special police” forces of the regional states are becoming tool for extremist ethnic and religious groups.⁷³ It centralizes the recruitment process of regional state police by subjecting it to federal control.⁷⁴ Regional state police, according to this document, must be made accountable⁷⁵ to the federal police while reserving the administrative accountability to the regional states.⁷⁶ Reversing previous trends and violating the regional state mandate, it requires the promotion and appointment of deputy commissioners and the commissioner of regional states to be made by the federal government (Ministry of Peace).⁷⁷

Fourth, there are signs that evince a move towards a centrally-designed making and unmaking of constituent unit boundaries as part of the shift to “geographic federalism”, doing away with the current federation. Under an apparent rationale that addressing interregional state disputes has become difficult to manage and requires an “expert-based” solution, the federal government established an “Identity and Administrative Boundary Commission”⁷⁸ with an allegedly country-wide mandate and perhaps to dismantle the federal system identified by the centrist elite as the “main source of Ethiopia’s political crisis."⁷⁹

As per the federal Constitution Article 48, disputes and misunderstandings between two or more regional states are expected to be resolved amicably by the parties in question. If that fails, the matter has to be submitted to the House of Federation (HoF)—a non-legislative second chamber composed of representatives of the different ethnonational groups. Yet, despite such a clear constitutional principle, the federal government issued a proclamation establishing the Identity and Administrative Boundary Commission that is accountable to the Prime Minister. The Commission is required to bring studies addressing identity-based and boundary disputes to the federal executive, parliament, and the HoF. As per the Constitution, identity-based demands are decided by the local and regional state elected councils once the demand is submitted by the concerned community. The new body takes away the mandate of the community and the regional state councils and ultimately the substantive mandate of the HoF. What has been bestowed by the Constitution to the regional states and the HoF is now given to the Commission by law. Self-government as stipulated in the Constitution is mandate of the “nations and nationalities” (Articles 8 and 39) and any controversy related to self-government is adjudicated by the HoF: the house that is supposed to guard their interests. As it takes away the mandate of the HoF and introduces a centralized decision-making process disregarding regional states, experts argue it is an unconstitutional body. Indeed, some of the members are well-known political activists and leaders of political parties—partisans, not independent adjudicators as claimed by the federal government and the law itself.⁸⁰

The nomenclature of the Commission is also controversial. In Amharic the word “የአስተዳደር ወሰን/yaʼasetadādare wasane” is centrist in its connotation. Regional states in Ethiopia are autonomous units, not provincial boundaries, as in a unitary state, and it is indeed very odd to refer to የአስተዳደር ወሰን/yaʼasetadādare wasane as if the boundaries of the states can easily be made and unmade by the federal government. Any adjustment to the boundaries of regional states can only be made with the participation of both the federal government and the regional states, as it amounts to a constitutional change. The normative assumption of the law thus speaks for itself as very centrist and thus is rightly criticized as designed with a “ኣሃዳዊ(ʼahādāwi)/centrist mindset."⁸¹ Ironically, the decision of the Council of Constitutional Inquiry (CCI)⁸²—made in response to Tigray’s challenge of the constitutionality of the law establishing the Commission—cites examples from Nigeria and India to justify its decisions by explaining the constitutionality of the law to reject the application from Tigray that questioned its constitutionality. In both federations, particularly India, the federal parliament is mandated by the Indian constitution to make and unmake provincial boundaries as if India is a unitary state (Articles 3 and 4). On the contrary, the Ethiopian Constitution does not give this kind of mandate to the federal government or the federal parliament. Indeed, under Articles 8 and 39, constituent units are given a wide range of autonomy and this autonomy covers both political and territorial autonomy. In a genuine federation, the federal government cannot unilaterally make or unmake boundaries of constituent units. It requires the consent of the units.

Arguably, the law could have established an independent body of experts, not political activists, under the HoF that could provide expert opinions on the same matters to the HoF, helping it make sound decisions. Abiy’s speech in July 2021 that defined the contested territories (in western Tigray) as belonging to Begemidir (the old name of Gondar of the Amhara region) also prejudges an outcome that may be made by the Commission.⁸³ Federal institutions are supposed to mediate impartially in disputes between regional states. The Amhara and Tigray regional states have claims and counter-claims over a disputed territory. The federal army is thus given a green light to side with the Amhara region and thereby reduced to serving as an agent of the Amhara elite, no longer the impartial Ethiopian defense force. The outcome of the Commission is now a forgone conclusion.⁸⁴ By doing this, Abiy has made the federal institutions subservient to the interests of the centrist elite. This approach is unlikely to bring lasting solutions both to the inhabitants of the contested territories and to the regional states (Amhara and Tigray).

Thus, since the establishment of PP, Ethiopia is effectively a unitary decentralized state, not a federation anymore. This remains a puzzle because the Constitution has not been amended. One should examine the forces that brought Abiy to power to understand the risks of this development. He came from Oromia, and it is well known that the core content of the protests in Oromia was “we need a genuine federation, more self-rule, and ceased federal government intervention in states.” In other words, it was a demand for more, not less autonomy. It is an irony then that Abiy is keen to centralize power contrary to the wishes of his own social base. This is a paradox but clearly reveals his plan. As mentioned already, the Constitution has not been amended and the new ruling elite introduced these policies in violation of the political autonomy of the states.

One has to note as well that the centralization drive is limited by deinstitutionalization of the public and security sectors limiting the capacity of the state. Indeed, deinstitutionalization characterizes the new regime. The federal government has lost monopoly over the use of force and has not been able to ensure law and order throughout the country, the bare minimum role of any government.⁸⁵ This development has affected public trust in authorities and public institutions.⁸⁶ Ethiopia has more than four million internally displaced people as a result of horizontal and vertical conflicts that the government was not able to handle, one of the highest in the world.⁸⁷ Following an interethnic conflict between Amhara and Benishangul Gumuz, Deputy Prime Minister Demeke Mekonen said “there is no other option for residents in Metekel zone other than organizing, arming and defending themselves."⁸⁸ The federal government has let the proliferation of informal forces such as Fano, Qeerro and regional state special forces threaten it, and as a result the federal government has lost physical control of parts of Oromia, Benishangul Gumuz, Amhara, and Tigray. The Ethiopian state is thus contracting and shrinking. This fact speaks volumes about the state of affairs in Ethiopia. A former official, Dawit W. Giorgis, wrote that Ethiopia under Abiy is “officially a failed state by all indicators."⁸⁹ Thus, alongside the drive for centralization, there is also fragmentation.

It should be mentioned that, in addition to the sham nature of the federations, it was the effort to centralize by the ruling elite that triggered the failure of the USSR and the Yugoslav federations.⁹⁰ There is thus a worrying parallel in Ethiopia at present: the more PP tries to centralize and impose military rule, the more it triggers ethnonationalism and the threats of fragmentation that comes with it. It is this worrying trend that we have labelled as the rise of a new centrist authoritarianism. Thus, return to centralization and any effort to break down already-mobilized ethnonational groups into pieces in an attempt to end group-based rights could therefore risk accelerating the country’s collapse.

Bargaining and negotiation between the two levels of governments is an integral element of federalism; unilateral action is not. Federalism requires the will and commitment to implement the pact as outlined in the supreme constitution by both levels of governments. Resort to force or unilateral action goes against the federal will. The Ethio-Eritrea federation (1952-1962) was unilaterally dissolved by the Emperor and the result was civil war and Eritrea’s secession in 1993. The Addis Ababa Agreement that ensured autonomy to South Sudan (1972-1983) and that brought relative peace to the Sudan was unilaterally abrogated by the regime in Khartoum, resulting in a devastating civil war and secession of South Sudan in 2011. Withdrawal of federal promises unilaterally has similar consequences.

3. Weighing the Integrationist and Accommodationist Approaches for Ethiopia

The tension between those seeking a more centralized system and the ethnonational cleavages has reached a new level. Yet, given that centralization is carried out in the name of promoting unity, questioning the post-1991 federal design, and preferring the integrationist presidential federal system, it is vital to present its essential features and its limitations.

3.1. Integration

At a global level, one softer option advocated by many experts and international organizations, such as the European Union, World Bank, and IMF,⁹¹ as a means to manage diversity is the integrationist federal system advocated by Donald Horowitz⁹² and practiced in Nigeria, South Africa,⁹³ and Kenya. Ethiopia’s centrist elite has long either resisted federalism or, when it thought unitary system would make it unpopular in the context of highly mobilized ethnonational groups, resorted under pressure and half-heartedly towards “geographic federalism”⁹⁴—the Ethiopian version of integration. The post-1991 federal design is often identified by the centrist elite as the “main source of Ethiopia’s political crisis."⁹⁵ Abiy Ahmed, since coming to power in 2018, has indicated this model to be his preferred system and hinted at amending the Constitution along those lines.⁹⁶ Lately, Ethiopian Citizens for Social Justice (better known by its Amharic acronym, ኢዜማ/ʼizémā),⁹⁷ a party that works closely with the regime in power, has issued its policy that propounds a presidential system, a strong federal government, and a Nigerian type of weaker regional states wherein the major ethnonational groups broken down into several smaller states. Given that this option has emerged as a new narrative from the centrist elite in Ethiopia, the following sections elaborate on it in detail, showing its pillars and weaknesses.

The intellectual and ideological roots of integration are often linked to American federalism. American experts on federalism, compared to their counterparts in Europe, prefer integrationist federation with the sole aim of diffusing power to many centers. In the process of adopting the United States federation (1789), “it was decided that no territory would receive statehood unless minorities were outnumbered by white Anglo-Saxon Protestants”⁹⁸ and, hence, there is little overlap between ethnic groups and territory. Indeed, Donald Horowitz,⁹⁹ the main architect of centripetal/integrationist federation, who has influenced the design process of most African federations, argued federations should aim to prevent ethnic minorities from becoming majorities at constituent unit level to weaken competing ethnonationalism and prevent a group from becoming a coherent unit, thereby frustrating the potential of becoming a nation state.¹⁰⁰ Following Horowitz’s approach, in South Africa, Nigeria, and Kenya the subunit boundaries do not generally match the territorial distribution of major cleavages or are drawn deliberately to cut across the main cleavages. The subunits are not meant to empower politically mobilized cleavages. The main goal of the federation is to disperse power, and as stated in the 2010 Kenyan Constitution the aim is to ensure development and service delivery at the local level.¹⁰¹

The system does recognize the right to culture and language as is the case in South Africa, but one can exercise these rights only as an individual. Kenya and Nigeria recognize that they are diverse, but they do not seek to politically empower diversity. The system recognizes diversity without empowering it. Both federations do not ensure the right to self-rule to ethnonational groups. Rather the main objective of the federal system is to diffuse power into many centers to reduce the risks of abuse of power at the center. Besides, the model does not allow political organizations based on ethnicity.¹⁰²

The underlying assumption of the presidential integrationist federal system is that a charismatic and selfless president who wins a popular majority and with wide support from a significant number of constituent units will unify the cleavages.¹⁰³ A strong federal government run by such a president and weaker states where the major ethnonational groups are broken down into many smaller units will be less a threat to the unity and integrity of the country. As already indicated, the model supports partition of constituent units belonging to ethnonational groups even against their will, as was the case in Nigeria¹⁰⁴ under a military rule meant to prevent emergence of substate nationalism.¹⁰⁵ Integration aims more toward building in (enhancing representation at national level) than building out (ethnonational-based subunits with political autonomy).¹⁰⁶ Both Nigeria and Kenya have constitutional rules that encourage political candidates to seek votes beyond their ethnic and social base. For example, in Nigeria, through “the principle of the federal character” (Article 14), a principle that aims to ensure fair representation in national institutions, the federal executive must include a member from each state. Parties that compete in national politics must have a national character, with a certain level of membership and number of candidates across the country. The president is also required to win a majority of votes as well as securing at least one-fourth of the vote in two-thirds of the states (Article 133).¹⁰⁷ Kenya has similar provisions in its 2010 constitution. A presidential candidate must win 50%+1 of the total votes cast in an election and at least 25% of the votes cast in more than half of the 47 counties. The appointment of the cabinet should also reflect the regional and ethnic diversity of the people of Kenya.¹⁰⁸ Political parties that provide presidential candidates are required to have national character and the constitution expressly prohibits the formation of political parties on the basis of ethnicity, region, or religion.¹⁰⁹

Yet, in societies with deep divisions, where there is no clear demographic majority (a common concern in many federations in Africa), getting such a selfless, unifying president is a rare possibility. The presidency as a one-person institution associated with a winner-takes-all outcome, along with the mandate to make and break the executive and the resources that it commands both formally and informally, is rather a divisive institution.¹¹⁰ The Nigerian and Kenyan presidential elections demonstrate this point. Instead of unifying different communities, the presidential institution remains very divisive, contentious and cause of electoral violence. The record is simply dismal. The Building Bridges Initiative (BBI) in Kenya, though constitutionally challenged by the Court, is an effort to bring back the position of Prime Minister and reintroduce a shared executive, as happened following the 2007 election crisis.¹¹¹ Thus, some have indicated a parliamentary system to be the preferred option.¹¹²

Second, when societies are deeply divided as in Ethiopia today, there is no guarantee that the presidential candidate will win across ethnonational divides. On the contrary, communities seem to be keen to choose their own candidates. Thus, a presidential system may perpetuate a contest between a permanent majority and a permanent minority, resulting in political frustration and electoral violence as it did in Kenya in 2007.¹¹³ Where there is no clear dominant group, it may even result in a struggle between permanent minorities. “Politicians from minority groups that occupy a few states/single state or share one with other groups have their guaranteed support confined to that state or a fraction thereof … In a divided society where ethnic categories are the primary lens for viewing conflicts, ethnic minority politicians are hardly accepted outside their home base."¹¹⁴ Thus, the presidential system has little to offer to those left out, leaving them to be under perpetual rule of others, and it is this sense of isolation that fuels ethnonationalism: why stay in a system that is not theirs?

The integrationist presidential federal model works where there exist softer, shallow cleavages in which voters are open to support a candidate even outside of their ethnic group, cutting across cleavages. As argued by Richard Simeon, federations that adopt integration work better in societies with a single dominant culture, softer cleavages, and a broad variety of minorities in particular when such minorities are dispersed throughout the country rather than territorially concentrated.¹¹⁵ Integration has little to offer to politically mobilized, territorially concentrated ethnonational groups.¹¹⁶ The current Nigerian federation is in many ways the making of the military rulers that envisioned a very centralized federal system and many small size constituent units that frustrated the demands of bigger ethnonational groups' ambitions for self-government.¹¹⁷ True federalism by the Yoruba, confederation by the Igbos, a high sense of marginalization and demand for resource control from the oil producing Niger Delta (home of the Ijaw) appear frequently in constitutional reform debates.¹¹⁸ Thus, cries for reform in the Nigerian federation, leading to rotating the presidency among the six bigger geoethnic regions, demands for “true federalism,” and more power and resource devolution to states have all been on the agenda since the return to civilian rule in 1999, although little progress has been made.¹¹⁹ These are some of the major forces in Nigeria pushing towards federal accommodation and it is thus obvious that integration, resulting in weak states and strong federal governments, has not succeeded in containing cleavages. Thus, Suberu, the Nigerian expert on federalism, concludes that Nigeria is on an unending search for democratic true federalism.¹²⁰ These are clear signs that integration of the sort found in Nigeria needs to be flexible to have some accommodative features.

Integration assumes a single people (demos),¹²¹ as it is a federation of citizens, not a federation among many mobilized ethnonational groups (demoi). Democratic legitimacy rests on the “people” but defining the people is central to the debate in divided societies. Integration assumes that democratic legitimacy comes from a “single common people.” As argued by Erk, “the center and the constituent units are orders of governments of the same nation."¹²² Experts argue that “with the sole exception of the state’s citizenship, [integrationists] are against the public institutional recognition of group identities”,¹²³ though they accept diversity in the private sphere (often called soft multiculturalism) and in some respects is not assimilation. In other words, it aims at public homogenization through common citizenship.¹²⁴ The preambles of the constitutions of Kenya and South Africa recognize that they are diverse, but focus is on unity. The preamble in the Kenyan constitution states: “Proud of our … diversity, and determined to live in peace and unity as one indivisible nation.”

Public policies, such as education, history, language, media, culture, national symbols, and the legal system are used as tools to promote a single public identity. Public policy also plays a key role in making substate identities publicly invisible.¹²⁵ As it is a product of the nation state, it promotes one (national) identity and discourages substate identities. It pretends to turn a blind eye to difference, but the culture and identity of the dominant elite implicitly and sometimes explicitly becomes the “national” identity. It promotes and nurtures the chosen identity of the political elite that controls power and is thus not inclusive. Identity then becomes a means for inclusion (a source of pride) or exclusion (a source of subordination and stigma) and a tool for mobilization by the left-outs to end their subordinate position. It is “rooted in the old liberal principle that the individual is the only atom from which to construct and analyze society."¹²⁶

The limitations of this model in dealing with groups that are mobilized against the center became clear towards the middle of the twentieth century, and thus the nation state was forced to reconfigure itself to provide space for substate entities.¹²⁷ Thus the UK, Spain, Canada and Belgium had to open up through a renegotiated social contract for the Scottish, Catalan, Quebec and Flemish autonomous subunits.¹²⁸ One could argue in this sense that it was integration applied in the context of deep cleavages that produced political mobilization. Failed integration thus led to accommodation systems where mobilized ethnonational groups in the above countries are by design made majorities at subunit level to ensure the right to self-government to groups as entities. The old liberal model based on the individual had to give in and be reformed to create space for left-outs. Deeply divided societies are pluri-nations and demand political recognition and empowerment of many nations, not one nation in the country. As argued by Erk, “Democratic legitimacy is based on the union between multiple demoi. And not every constituent nation sees the union in identical terms. For English speakers, Canada is a federation of ten provinces; for French speakers, it is a union of two nations."¹²⁹ The center and the constituent units are orders of governments of different nations. There are thus competing views on the nature of the union and the source of democratic legitimacy owing to multiple demoi. In the presidential federal model, the goal is to construct one people out of many and by design prevent ethnonational groups from becoming constituent unit majorities.¹³⁰ In other words, it does not ensure group self-government at substate levels.

In Ethiopia, owing to the coerced state formation process towards the end of the 19th and early 20th centuries, and the subsequent failure by the centrist elite to address ethnic-based marginalization, the bigger ethnonational groups established national liberation fronts and brought the central government to its knees in 1991. Since 1991, they have exercised limited political autonomy and continue to demand more, not less autonomy. It was centralization and undue interference in subunit affairs that partly triggered the 2015 protests, resulting in regime change in 2018. One could state that the level of ethnonational-based mobilization in Ethiopia is a record within the African context. The new elite has since 2018 intensified its centralization measures, triggering reaction from ethnonational groups.

Furthermore, in the case of Nigeria, integration and territorial manipulation worked because the military imposed it by force, but the issue is: Why would a strongly mobilized, territorially concentrated group that feels that the central government is repressive and exclusive, as was the case for the most part of 20^th^ Ethiopia, consent to such a divide and rule strategy? The effort by the central government to impose centrist policies in Ethiopia resulted in its liquidation and the regime’s demise in 1991. The effort to do the same by the current government has provoked widespread discontent and civil war in Oromia, the South, and Tigray.

Given the distinct nature of the cleavages in Ethiopia, it is hardly possible to address the demands of territorially based and politically mobilized groups within the integrationist presidential federal system as outlined already. Given the peculiar nature of the political cleavages, the next section demonstrates a more relevant consociational parliamentary federation as a preferred alternative.

3.2 Accommodation and power sharing

For countries with deep divisions where ethnonational groups are politically mobilized and identity is politically salient, the two well-recognized alternatives, depending on whether the cleavage is territorially grouped or dispersed, are power sharing and accommodation within a federation or a combination of both.¹³¹ As the cleavages in Ethiopia are for the most part found territorially grouped, power sharing will remain relevant at the federal level.

First, a brief discussion of power sharing as it could exist separately in a polity that is not a federation using the Netherlands as an example. A federation could also exist without embracing power sharing and espousing majoritarian democracy. The works of Arend Lijpart¹³² and the late O’Leary and McGarry¹³³ have articulated the various elements of consociational, as opposed to majoritarian, democracy. The essence of this approach has been Lijphart’s long-held observation: stability despite deep divides in some countries such as Switzerland, Belgium, and Netherlands are associated with consociational democracy. Lijphart writes, “Power sharing denotes the participation of all significant communal groups in political decision-making especially at the executive level."¹³⁴ Unlike integration that prefers a presidential system, power sharing is associated with a parliamentary system that is suitable for a power sharing arrangement in the executive, often in combination with a proportional electoral (PR) system. A parliamentary coalition is certainly more inclusive compared to the office of the presidency. As argued by Lijphart, it is a flexible arrangement that aims to ensure broad-based representation and inclusion of all major actors in decision-making, and could be on equal or proportional basis. In Belgium there is equal representation of the Dutch and French speaking communities in the executive. In South Africa (1994-1999) all parties that won representation above a 5% threshold in the legislature had the right to be represented in the cabinet. The executive power may be divided and shared, as was the case in Kenya between Kibaki and Odinga following the 2007 election crisis, and the rest of the cabinet shared between rival parties. It could also take the Swiss form where all communal groups are conventionally represented in the collegial executive, with the presidency rotating among the seven members annually.¹³⁵ The inclusion of the major political actors in the political institutions, either through a proportional electoral system or on an equal basis, along with representation of the different groups in public institutions is thus the core feature of power sharing. Power sharing, unlike federalism, often takes a non-territorial form.¹³⁶ In other words, it is often recommended for societies that are deeply divided by identity differences among groups that are found to be geographically intermixed. Through a combination of power sharing that enables them to influence policymaking at the center, and autonomy that entitles them to decide issues related to language, education, and culture, the groups are believed to contribute to political stability. Power sharing could thus remain short of federalism so long us there is no territorially based cleavage demanding political autonomy and self-government.

One could not rule out the relevance of power sharing in most African federations that lack a clear dominant group that, taken alone, constitutes a “50 plus 1” majority.¹³⁷ O’Leary has argued that in a context where there is no staatsvolk (a dominant group that enjoys absolute majority and hence has the demographic advantage), political instability will prevail unless there is an inclusive political system that brings the major political actors to power.¹³⁸ The solution he proposed is to have executive power sharing among the major political actors. While in the federations mentioned there is an effort to have an inclusive federal government, in reality the federal government is perceived as belonging to one dominant group and thus continues to face a legitimacy crisis. Left outs continue to threaten the central government. For instance, since the establishment of the Ethiopian federation, both as a result of constitutional principle (Article 39[3]) and practice, there has been an attempt to reflect the country’s diversity in the establishment of the executive. The issue, however, is that representation is meant for some only in the nominal and not real sense. The two coalition members of the EPRDF, that led the Oromia and Amhara regional states in particular, faced serious legitimacy crises until recently, and thus a section of the Oromo and Amhara used to think they were not genuinely represented in the federal institutions; hence the narrative of marginalization and protests since 2015. At present, the federation operates without Tigray and that partly explains the ongoing war against the federal government.

In nearly all African federations, and this includes Ethiopia as well, the federal principle of representation has never been genuinely implemented. Representation of subunits in federal institutions is mainly in the second chamber, federal executive, judiciary, army, and security. Yet inclusive and broad-based federal government remains a scarce commodity, explaining what Steytler and de Visser call “the fragile nature”¹³⁹ of the federations as they face threats of fragmentation and secession. Non-inclusive federal government in a divided society means it is perceived as belonging to some and not all, and with this comes the legitimacy crisis explained earlier. To stay in power, the federal government resorts to brute force. The constitutional clauses contain some element of representation. Nigeria has the “federal character” (Article 4[4]), Kenya “institutions should manifest Kenya’s diversity” (Article 91), Ethiopia the right to equitable representation of nationalities (Article 39[3]) in federal institutions. However, this representation becomes more symbolic than genuine. In many cases it is the federal government that handpicks the “representatives” instead of leaving the matter for the subunits to designate their genuine delegates in the federal institutions. This has been a major paradox in Ethiopia for example as the “representatives” do not have legitimacy in the eyes of the people of the states.

Second chambers take a special role in representing states in the law-making process at the federal level and in airing out their voices. This is particularly critical when the federal government designs laws and policies that may affect the interests of the states. In addition to airing their concerns, constituent units can also block the other house when it exceeds its mandate and impacts their autonomy. The second chamber thus safeguards the interests of the states. The basic principle that guides federations is that, irrespective of their population size, the states are represented on an equal basis. Thus, in Nigeria, there are two senators from each state (Article 48) while in Kenya, each county has one representative in the Senate (Article 98). In South Africa, each province has ten representatives in the National Council of Provinces (Article 60). The Ethiopian House of Federation is distinct in two respects: it has no law-making function but has wide powers in resolving intergroup conflicts (although in reality it is less-used than it might be because its members are not full time). It also enjoys wide powers in the allocation of subsidies that the federal government allocates to the states—the main source of state revenue. Second, the representation is not based on equality. It is a majoritarian house where each nationality has one seat, but for every additional one million, there is one more seat (Articles 61 and 62). The fact that states in Ethiopia have no law-making functions means that the federal government can easily enact laws that affect the autonomy of states, and in the absence of a constitutional court that impartially umpires intergovernmental disputes, the system leaves states at the mercy of the federal government. The experiences in Nigeria, Kenya and South Africa are relevant in this regard. The states have a role in the law-making process at federal level and they all have either a constitutional court (South Africa) or Supreme Court (Kenya and Nigeria) to address intergovernmental disputes in an impartial manner. Institutions that enforce the supremacy of the constitution and the rule of law and that umpire intergovernmental disputes impartially are indeed vital preconditions for the operation and development of federations.¹⁴⁰ These institutions give life to the powers allocated to the two levels of government and thereby allow the federation to evolve within bounds. Such institutions keep the federation balanced and give concrete meaning to the division of powers. In Africa, political power is yet to be tamed and institutionalized. The “big man” or the dominant party at the center is a major obstacle to federalism, power sharing, and the separation of powers. Independent institutions such as supreme or constitutional courts play vital roles in limiting power and institutionalizing it. South Africa and Kenya have made some steps, but Ethiopia is far behind in this respect and the federation has largely been open to political manipulations.

It is vital to mention that successful federations with politically mobilized cleavages often combine federalism with power sharing. Groups often are found at times to be intermixed and in other places geographically concentrated. Thus, combining the two offers an ideal solution. It is for this reason that the late Elazar argued that consociational federations are the best fit for divided societies.¹⁴¹ The Swiss and Belgian federations are in many ways consociational. Compared to presidential centripetal federations, consociational parliamentary federations bring the major political actors into the executive and parliament and minimize the risks of winner-takes-all conditions associated with presidential systems. As it often leads to coalition government, the system requires a consensus-based process of decision-making to avoid government collapse, and that is an incentive for elite bargain.¹⁴²

The highest form for accommodating ethnonationalist groups particularly when they are territorially concentrated and politically mobilized is through consociational federalism.¹⁴³ The accommodation approach takes “divisions” seriously and does not aim to abolish or weaken them but instead recognizes them and turns them into constitutive elements of democracy and empowerment. When combined with federalism, it treats “the segmented elements” as building blocks of political engagement and aims to make the ethnonational group become a majority at constituent unit level.¹⁴⁴ While Switzerland and Canada were pioneers in this respect, Kymlicka argues, the idea of ensuring self-government to territorially concentrated national minorities is now universal in the West.¹⁴⁵

The geographic distribution of diversity and its political mobilization remains a vital factor. Unlike integrationist presidential federations, ethnonational parliamentary federations aim to empower such groups by redrawing territories to ensure that they become a majority at substate level. Ethnonational minorities challenge the coercive process of state formation and the subordinate relationship they have with the center. They assert a national identity whose goal is to ensure self-government within a defined territory, and thus the relationship between groups and territory becomes critical.¹⁴⁶ Self-government is intrinsically linked to territory. At the core of the mobilization is the aim to address political and economic deprivation and regain collective self-esteem by ending their subordinate relation to the elite that controls the center.¹⁴⁷ Their claims may extend to include right to exit, while the center wants to ensure the unity and territorial integrity of the state. Political autonomy within a defined territory as part of the political system is thus a middle way solution to the competing claims of the center and the subunits.

Unlike adherents to the presidential federal model, ethnonational-based federations do not aim to nurture one identity but recognize more than one public identity. Public policy thus promotes multiple identities, not a single identity. Managing mobilized ethnonational diversity and ensuring the right to self-government remains the major task of the federal design. Through its combination of constitutionally entrenched division of power as well as the principles of shared rule and self-rule,¹⁴⁸ it allows mobilized groups to enjoy political autonomy at the subunit level while ensuring representation at the federal level. Through the institutions of shared rule and representation in federal institutions, ethnonationalist groups are given the opportunity to influence decision-making at the center. The logic of this form of federalism is that ethnonationalist groups can only respect the institutions of governance and thus contribute to stable federation when they are granted a satisfactory combination of influence at the center and meaningful autonomy at the substate level with regard to their own affairs.¹⁴⁹ As practiced in the Swiss collective presidency of the federal executive and the Canadian and Belgian executive by convention or constitutional principle, the different identities have guaranteed representation in key decision-making institutions at federal level.¹⁵⁰ This somehow determines whether ethnonationalist groups will remain loyal to the overarching federation or will prefer to fight or walk away from it. As it recognizes more than one public identity (at the federal and substate levels) and aims to secure coexistence among them, it remains a key instrument for polity-building and managing diversity.

The limitations of the accommodationist approach are well known. It is argued by experts that “it institutionalizes divisions and deepens the fault line that it is meant to address."¹⁵¹ Experts warn that such arrangements reinforce, not alleviate the cleavages; they provide ethnonational groups with the resources needed to mobilize and challenge the territorial integrity of the state, facilitating fragmentation. However, the alternatives are also very limited. In the first place, in deeply divided societies it is because the cleavages are already deeply entrenched and mutually reinforcing that accommodative approaches are introduced to prevent the next step: protracted conflict, secession, and state fragmentation. Power sharing and ethnonational-based federations are “put in place where other institutional options have failed”¹⁵² or are not available. As argued by Liam Anderson, the choice is between ethnonational federation and nothing. In the menu of options, it is the last resort before break-up or state failure. Indeed, the strength of ethnonational-based accommodation rests on the absence of viable alternatives.

However, it is rare that a subunit will become a homogenous unit, and the mobilized ethnonational groups that constitute a majority at subunit level could tend to be local tyrants that abuse the rights of minorities and thus require institutional guarantees to ensure citizenship and minority rights throughout the federation.¹⁵³ One should note, in the countries that recognize group rights (such as Canada, the U.K., and Belgium) by adapting their respective constitutions to provide space to hitherto marginalized groups, group rights and autonomy came to broaden the scope of human rights protection, not limit them. They already had a well-protected individual rights system long before their constitutions were adjusted to grant autonomy to deeply mobilized cleavages. Yet the individual rights system was found not enough. Representation and autonomy were added to the already well-entrenched individual rights system; it came to reinforce, not weaken it. In most African federations, this is a major concern as the record on citizenship rights is dismal and group rights continue to threaten already weak individual rights. Ethiopia’s post-1991 federal experience is a textbook example of the above limitations. Under EPRDF’s tight rule and state-led economic growth, Ethiopia was able to ensure relative stability and fast economic growth with massive investment in basic services such as education, roads, and health. With the change of leadership in 2018, a lack of leadership, fragmented ideology, and power struggle made the component parts collapse along fault lines. Critics have aired their concerns, stating that the Ethiopian federal system, by empowering ethnonational groups to self-rule and placing resources, security forces, and the media at their disposal, may weaken common ties.¹⁵⁴ Federalism is about, in the words of David Miller, “nesting identity”¹⁵⁵—an identity that is aware of itself is unleashed by self-rule but it is also about becoming aware of the identity of others and respecting, recognizing, and having positive interactions with them as well. Federalism assumes intergroup interaction to build cohesion among groups.¹⁵⁶ It does not promote non-interactive existence of parallel and segregated identities. It also entails the coexistence of multiple identities whereby two or more identities feel belonging both to their smaller (substate) communities and to a larger, overarching political community and they do not think that both are mutually exclusive. Yet the Ethiopian federal system lacks a comprehensive policy for nesting diversity and for creating cohesion among different groups. Strong institutional protection of individual rights and ensuring fair and genuine representation of minorities at different levels of government mitigates the weaknesses of accommodation and enhances social cohesion by encouraging free movement of people across the federation. Accommodation, thus, needs to manifest some element of integration. Otherwise, emboldened ethnonational groups and their elites may, by focusing on the politics of difference, ultimately target independent nationhood as their objective. It is thus a high-risk strategy that may eventually bring about the country’s disintegration.¹⁵⁷ Some of the above symptoms were clearly manifested in the pre-election debates in 2005 and in the ongoing war where some leaders from the opposition fomented ethnic hatred.¹⁵⁸ State institutions such as the police, whose main role is to ensure peace and order and prevent crimes, often took sides in intergroup conflicts and aggravated the problem.¹⁵⁹

The federal system has also shaped and transformed the nature of intergroup relations and tensions over the years. The fact that disputed areas coincided with the administrative boundaries between regional governments seems to have transformed the conflict between local communities into con­flicts between regional states. The Oromia-(Ethio)-Somali Conflict (2017),¹⁶⁰ Amhara-Benishangul, and Amhara-Tigray boundary disputes are examples in which local political elites, the media, and the police apparatus took sides in the conflict, causing death and displacement for hundreds of thousands from both sides and threatening the peace and stability of the country. Federal institutions are also falling into that trap of late. Citizens in different parts of the country accuse federal institutions such as the police and the army of failing to protect them in times of conflict.¹⁶¹ Disagreements are expected to be resolved through compromise and dialogue using the existing political institutions such legislative bodies, intergovernmental platforms, party-level negotiations, and, if that fails, legal means: the supreme/constitutional court that serves as the ultimate guardian of the constitution. Ethiopia failed to build these institutions and the EPRDF, now rebranded as the PP (after the TPLF declined to join) relied on its own party machinery, and this worked to some extent, at least partly due to the effects of democratic centralism, combined with elitist leaders such as Meles Zenawi (1991-2012).¹⁶² In the absence of such leaders, and with the weakening or fragmentation of the party system, there is nothing left that can serve as a platform to sort out normal political business. Amid failing public institutions, the different actors thus race over to the special police and informal forces as a means to safeguard their interests.¹⁶³ The multitudes of vertical and horizontal conflicts demonstrate the death of political institutions. Federal institutions are supposed to mediate impartially in disputes between regional states, but it is the army and security forces that are often the main actors and taking sides. The war in Tigray is perhaps the clearest manifestation of the level of political crisis the country faces amid the rise of the centrist authoritarian elite since 2018, where military intervention and violence, not political solutions, have become the new normal. Federalism has become a victim of militarism and the command posts that ended subunit autonomy.

Conclusion

This article has demonstrated that the nature of a cleavage and its level of mobilization matter to and impact institutional design. Though a softer and preferred option, integration has not been able to respond to strongly mobilized groups in Nigeria and Kenya, and failed integration policy is slowly fueling ethnonationalism, resulting in more radical demands in both countries, albeit to varying degrees.

The nature of the political cleavage in Ethiopia is very unique compared to many countries in Africa. Ethiopia is a textbook example of a deeply divided society where rival nationalisms one led by the state and others led by countless ethnonational liberation fronts have brought about a clash of nationalisms. As a result of a coercive and narrowly based nation building process that liquidated quasi-autonomous kingdoms in the early 20^th^ century and a failure by successive regimes to respond to demands made by ethnonational groups for accommodation, the number of national liberation fronts in Ethiopia is a record within the African continent. The left-outs from the nation building process have continued to challenge the center. The debate between the centrist elite and the left-outs is Ethiopia’s major political paradox and has made the country very fragile. In comparison with the other federations under discussion, two points further complicate Ethiopia’s political context. The extreme centralization of power and the brute use of force to deal with political issues was an incubator for the ethnonational-based liberation fronts. In addition, the failure to dialogue and compromise in order to resolve political issues characterizes modern Ethiopian politics. It is hard to find something comparable to the political settlement in post-apartheid South Africa or Kenya’s post-2007 political deal following the election crisis. Entrenched, territorially based, and politically mobilized cleavage is the outcome. The distinct and radical nature of Ethiopia’s post-1991 federal system, in its determination to address the nationality question (albeit with limited implementation), was an effort to address this issue, though it remained a victim of democratic centralism and the developmental state under EPRDF, leading to wide spread protests since 2015 and a change of government in 2018.

There was much hope that, when the new leadership came to power in 2018, the demands for genuine federalism—in which regional states would exercise political autonomy free from central interference—would be realized. What transpired was however a new form of centralization that has changed the narrative to centralize nation building, issued policies that promote unitarism, removed regional state heads by force and continued to impose “command posts” in the states. Given its geographic coverage affecting most regional states and the frequency of its use, this has become the new normal. The current state of affairs shows that militarism has replaced federalism in Ethiopia. The reaction is vivid: an all-out war against the center, particularly in Tigray and Oromia—two of the states that harbor age-old demands for self-government and are now pushing for a loose confederation. Centralization and marginalization are currently two major challenges, and both are unfavorable to deeply divided Ethiopia.

The fact that the nature of the cleavage has been politically mobilized for decades has an effect on institutional design and on the type of democracy practiced. Ethiopia is home to many armed ethnonational liberation movements that have survived for no less than five decades despite a heavy onslaught from the central government. The goal of their mobilization is to recover the extensive self-government they claim to have enjoyed historically or that they aspire to have now. The current ruling party has chosen to recycle the imperial narrative instead of addressing the cleavages. The more the elite in power tries to centralize, the more it provokes ethnonational elites to push to the extreme. This is a paradox because the new elite claims to hail from Oromia, the very region that abhorred centralization and federal intervention during the protests between 2015 and 2018. In Tigray, Oromia, and the South, demands for more autonomy and self-government have been on the rise, not in decline. Given the level of mobilization and the current polarization, a loose political arrangement in the form of confederation in which most powers are devolved to the states, with the center only symbolizing unity, can possibly mitigate the tension between the center and subunits. By devolving more power to the states, the stakes at the center are lowered and this could lead to less disagreement. Given that the center is increasingly becoming source of insecurity (at least in the modern era), constituent units will assume more responsibility and reduce the insecurity. The fact that Ethiopia is a country of minorities in which, at central level, none of the groups taken alone constitute a “50 plus 1” majority means that a more innovative political arrangement needs to be crafted to fit the reality. Whatever powers the center assumes must be consensus-based and should engage regional states possibly on equal basis.

As already noted, Ethiopia fits well into the literature on divided societies and needs a peculiarly designed federal and democratic package. Even if Ethiopia were to be democratic and exercise majoritarian-based democracy, its general assumption that the rulers alternate every possible election (that today’s majority will become tomorrow’s minority) does not hold true. In Ethiopia, there is no dominant ethnonational group that enjoys a demographic majority and could claim to have democratic majority to pursue its goals. This leads to clashes between a minority that has state resources at its disposal trying to impose its will on another minority, and thus is very destructive. Yet, even if the central government’s project of nationalism enjoys a majority, it pits a permanent majority against a permanent minority and the latter has no hope of becoming a majority. It would be naïve to expect groups labelled as permanent minorities to remain loyal to the system, and that explains the major source of political crisis. Thus, alternative theories such as consociational and parliamentary democracy need to be adopted. Instead of having winners and losers, consociation democracy brings major political actors together either on equal footing or through proportional arrangements of power sharing and insists on consensus decision-making on fundamental issues. In this manner, those left out of majoritarian democracy become decision makers through power sharing, reducing the potential for conflict. Distinct institutional arrangements thus matter when managing cleavages and reducing conflicts. At heart, it is about political inclusion and having a just political order. Given the polarization and mistrust that characterizes the relationships between the federal government and the regional states, in the short run a more accommodative system in which decision-making at federal level is more consultative and inclusive of all regional state actors and adopts consensus-based decision-making could serve as a trust-building process. Federal government should use intergovernmental platforms that are inclusive of all regional states as a means to mitigate existing tension and mistrust. Consensus-based decision-making on fundamental political and economic issues between the federal government and regional states builds more trust and reduces the sense of alienation. Without addressing this major issue, it is unlikely that Ethiopia will democratize or be able to respond to deep cleavages, and ignoring this issue may indeed be accelerating the country’s fragmentation.

If Ethiopia is to remain a federation, the other distinct institutional feature that fits deeply divided societies with territorial cleavages is political autonomy and representation. Unlike the integrationist presidential federations that disperse power into many centers, ethnonational parliamentary federations aim to empower such groups by redrawing territories to ensure that they become a majority at substate level and so will exercise meaningful political autonomy and self-government while at the same time ensuring representation in the federal political process. As noted already, centralization has affected autonomy. Representation has also remained more symbolic than real. Regional states need to have the autonomy to elect or select their representatives in federal institutions. The fact that the center continues to handpick representatives of the states has made the principle of representation a farce. With genuine representation, consociational parliamentary federations bring the major political actors into the executive and parliament and minimize the risks of winner-takes-all conditions associated with presidential systems. As it often leads to coalition government, the system needs a consensus-based process of decision-making to avoid government collapse, and that is an incentive for elite bargain. Mobilized ethnonational groups that feel less represented in federal institutions have little incentive to stay in the union unless they are assured of some level of influence or even a veto at the center. To minimize the growing mistrust between the federal government and regional states and build trust, indeed the key decisions that affect the country and the regional states need to be decided by a consensus between federal government and regional state leaders.

Discussions]

Dr. Mohamed Dejen

As we know, EPRDF was criticized for being heavily centralized and disregarding the federal state structure, and now you say that the PP is even moving towards a unitary system. I have difficulty accepting this assertion since, in the era of identity-based polarized politics, it is very difficult to judge whether it is moving toward a unitary system or that of authoritarian system. Because, although it claims to be one party, PP is more of ethnic-oriented and divided party. Rather, regional states are challenging the center, resisting the policies set by the federal government and acting as a sovereign state with their own armed forces. I see rather that we are moving back to The Era of Princes (ዘመነ መሳፍንት—Zamana masāfenete), than moving to a centralized governance system. If the current regime is centralizing, why is it not effectively governing the country? If you say at the same time that there is centralization and the state is falling apart, is it not a paradox? What is your reflection on this?

Dr. Zemelak Ayele

You mentioned confederal arrangement just once and you do not push it as a possible alternative for Ethiopia. It would be nice if you would expand on that. And if a confederal arrangement is opted for, what would keep the country together: security? Economy? Or what else?

Dr. Zelalem Mogessie Teferra

You said that Ethiopia is a country of minorities. I think we should have conceptual clarity on who is a minority (are we talking about numbers, or historical injustice) and regarding which place (geographical area) we define as minority? Is it in a specific region like Oromia, or in cities like Addis? In my view, Ethiopia is a country of majorities, a country of minorities, a country where majorities live as minorities and minorities live as majorities, or minorities which seek to be treated as a majority. In Harari region, Amaharas and Oromos are majorities, but they live as minorities; Amharas in Oromo Special Zone of Amhara Region live as minorities. We need to come up with a new conceptual articulation of who is minority in Ethiopia. I am not sure if we can take the general Ethiopian population as a reference point to define who minority is. Maybe we need also to look at historical injustices. For example, Tigray has been dominant in the political landscape of Ethiopia historically—do we consider them minorities or majorities? So, we need to probably reinvent the wheel here when it comes to the definition of minority.

A fundamental premise of your presentation is that the current conflicts in Tigray and Oromia are the result of a clash of nationalisms/visions. I find this assumption very simplistic; is the Tigray war actually against centralization or the perceived or actual exclusion from the center? Is it for self-autonomy? I don’t think so. Is the problem in Oromia a fight for autonomy? Or is it an intra-ethnic struggle for power dominance?

The last four years also do not offer a clear picture of the current trend in our country as regards the federal exercise. You suggested that everything is being more centralized, but the official narrative of the government has been that it will maintain or even strengthen the multi-national, multi-ethnic federation. But the reality on the ground does not support your assertion. Take for example, the case of Sidama, which has become a new region added to the federation, and the South West Ethiopia Region is also coming into existence; both are evidence of more decentralization than centralization in Ethiopia. So, I do not think there is a clear picture which shows that the whole trend is towards centralization; what I rather see is a mixed picture: There is centralization at party level and decentralization in terms of the federal structure.

Dr. Semir Yusuf

I would like to emphasize the ironic nature of the current situation in Ethiopia. Two processes that are diametrically opposed are happening at the same time. There is the drive towards centralization and the drive to state disintegration. A very important manifestation of centralization is the rhetorical aspect of the regime. It also tends to amass power in the center and at the higher echelons of the ruling party. These are clear indications of the move towards centralization. But at the same time, we also witness very clear indications of state fragmentation. An example is the kind of debate and fragmentation within the PP itself. The breakdown of the command-and-control chain within the security apparatus or the entire security structure is a very interesting move against the drive for centralization. We also see the creation of deep states within the state structure: parallel state structures, especially in the economic and security sectors of the state. This is another very alarming move in the reconstitution of the Ethiopian state. Finally, the loss of the state’s monopoly over violence in different parts of the country is very much contrary to the envisioned idea of further centralization of state power and party structures.

At this juncture I would like to point out the area where accommodation intersects with integration. We need to resurrect the state somehow; we do not need to see a failing state; we need a state to function in the way it should. The state should be revived to provide its basic function in society, while at the same time accommodating diversity and divergent perspectives.

Dr. Solomon Nigussie

My first question concerns the title and content of your paper. I found that there is divergence between the phrasing “from federalism to militarism” in the title and the discussion in the body. Is the title referring to the timeline or the nature of the federal setup? Second, there are some general statements in your analysis which require further details and data. For example, there are concerns around the centralization of the Ethiopian federation but: Who is centralizing and for what end? What are the parameters to gauge the centralizing tendency? You seem to say that the Ethiopian empire of the beginning of the 20^th^ century is coming back in the last four years. Is that what we see on the ground? Is there a clear pattern of this centralization tendency, compared to the clear prospect of fragility?

Third, is the Constitution, as it is, suitable to implement your recommendations? For example, can we realize checks and balance with the current constitution? Which one should be given priority: integration and ensuring rule of law, or addressing the inherent problems of the constitutional federal setup?

Finally, is centralization not the inherent nature of the state when its very existence is challenged? What can be done when the federal structure and the existence of the country is in danger? How do you balance these two exigencies?

Dr. Sisay Alemahu

Would you consider the possibility of de-ethnicizing political organization as a solution for what you call “the politicization of the cleavage,” which would be a mammoth task? Could it be an option? Because the Constitution nowhere requires political organizations to be along ethnic lines, although it is a logical conclusion to the constitutional architecture.

Reply: Prof. Assefa Fiseha

Some of the questions relate to what Dr. Semir aptly articulated as the paradox of “the more you centralize, the more you lose the subunits.” So, there is no contradiction in my presentation in stating that there is centralization of state power and the state is getting weak; the more the center is trying to centralize power it is losing physical control over a huge part of the country. This is a paradox we are facing. There is this clear, deliberate deinstitutionalization of the public sector in general but even more in the army and the security apparatus. The party system is also deinstitutionalized. So, as a result what you see is fragmentation of the state.

Another issue is deinstitutionalization of the public sector in the name of reform. We know public institutions were weak in Ethiopia but at least the Party (EPRDF) was there; you may not like it, but it used to function. Now the party itself is deinstitutionalized; the whole public sector is also deinstitutionalized. And in the end, the state-citizen relationship is completely falling apart. Subunits are reacting against the drive for centralization in a way that is defined in political science and security studies as state contraction; the Ethiopian state is contracting, shrinking. So, the threat of fragmentation is visible; that is why I wanted to talk about it seriously.

About minorities: if you want to know about subunit minorities there is an article I published in Utrecht Law Review which is freely accessible. The focus in my presentation here is on the center. What I am saying is that establishing legitimate, inclusive government in Ethiopia at the center becomes problematic because it is a country of minorities. The statistics are very clear; there is no ethnic nationality with a 50+1 demographic majority that also entitles it to be a democratic majority. At this point Dr. Adem Kassie, one of the conference participants, interjected and asked: If there is no majority can there be minorities? Prof. Assefa answered that all ethnic groups in Ethiopia are minorities. Dr. Adem again asked: How can there be a minority if there is no majority? Prof. Assefa replied saying: Because there are political elites who at one time centrally control the army, the security and the whole public sector in the name of, say, nation building or federalism, and control everything, and the rest of the minorities want to do the same; so, the cycle continues. Dr. Zelalem, another conference participant, interjected and asked whether Tigray was historically a minority or a majority, to which Prof. Assefa responded by saying: I am not going into that issue.

To Zemelak’s question: It all depends on whether you agree with the premises that Ethiopia is deeply divided and what one needs to respond to the deeply mobilized cleavages. The level of mobilization of cleavages is a record within the African continent; I could not find any state in Africa with the level of mobilization of cleavages as chronic as in Ethiopia. So, it needs a peculiar arrangement. The federalist promise could have addressed many of the problems, even with the institutional design gaps it has, but it was never implemented. At this point in time, we have reached a level where the existing constitution is no longer enough to respond to the issues we are now facing. So, we need a renegotiated social contract.

With respect to Dr. Solomon’s comments, the paper develops clear indicators that show the trends of centralization, including the frequent use of “command posts” which suspend civilian rule, and I see no contest on those points. A distinction should be made between a desire for centralization from the center’s side and lack of capacity to implement its desire owing to deinstitutionalization and the resistance to it from the regional states.

Notes