Ethiopia’s National Dialogue Conference closed on 22 August in Addis Ababa, with President Taye Atske Selassie and foreign dignitaries, among them former Kenyan President Uhuru Kenyatta, in attendance. Parliament had established the Ethiopian National Dialogue Commission in January 2022 and charged it with “identifying the root causes of the differences on fundamental national issues” and laying “a firm foundation for national consensus and the building of a state with strong legitimacy.” Four years of consultation were meant to fulfil that mandate. The closing ceremony was presented as their culmination.

The Commission reports that roughly 4,000 participants – drawn from woredas across the country, from regional and federal institutions, and from the diaspora – produced some 400 recommendations. They span eight agenda areas: nation-building; government structure, political representation and the electoral system; the status of Addis Ababa and Dire Dawa; religion; institution-building, the rule of law and human rights; social and economic questions; corruption and governance; and peacebuilding. Nested inside those headings are some of the most profound questions any political community must eventually face: where sovereignty resides; how state power should be organised and distributed; what rights its peoples possess; how national identity and citizenship are to be understood; and what kind of federation Ethiopia is to be.

The Commission has not yet published its final recommendations but one need not wait for that text. The eight agenda areas already reopened the 1995 settlement, putting foundational constitutional questions back on the table. The broader public debate has made the same point unmistakable. Whatever particular formulations the Commission settles on, a process organised around those questions is not preparing a tidy amendment package. It is opening the possibility of recasting the settlement’s foundational terms. A transformation of that magnitude requires a new social contract – adopted through an exercise of constituent authority – not an amendment of the constitution already in force. That is why this argument needs to be heard before the report reaches the government: while there is still time to shape how commissioners, and those who will receive their work, understand the constitutional character of what they are doing.

Amendment or a New Social Contract

The government has not said which course it intends to pursue. The signals point in different directions. The clearest came on 7 July, eight days before the conference opened, when Prime Minister Abiy Ahmed answered questions from MPs. He rejected the pattern of constitutional rupture through which the Derg discarded the imperial constitution and the EPRDF, in turn, discarded the Derg’s – a practice, he said, that had not served Ethiopia well. His government would not “tear up” the existing constitution. It would consult the public, retain what remained useful, add what was missing, and amend what required improvement. Beneficial constitutional ideas would stay. He did not expect the dialogue to produce provisions that would “destroy what exists,” only proposals that would strengthen it. Taken at face value, these are the strongest indications yet that the government contemplates reform within the present constitution rather than its replacement.

In the same intervention, however, Abiy described the dialogue in more expansive terms: to confront problems accumulated over generations and to “sign a new social contract.” At the conference’s opening he spoke of ensuring that future “constitutional changes” rested on broad national consensus – of creating a “social contract that future generations will honour,” or producing “a more stable constitutional order that cannot be easily altered or replaced.” He did not say whether the Commission’s recommendations would travel through Articles 104 and 105, go to the people as part of a constituent process, or be embodied in a new constitutional instrument. Two years earlier he had suggested that questions the dialogue could not resolve by consensus might be referred to Ethiopians by referendum. No subsequent statement has clarified how that proposal relates to the process now contemplated. 

The government speaks, at once, the language of amendment and the language of a new social contract. Those are not two names for the same thing. Amendment presupposes continuity: it permits change within an existing constitutional order while preserving that order’s basic identity. What is now under consideration is of a different kind.

The 1995 constitution’s foundational bargain rests on a set of choices: that Ethiopia is a federation of self-governing nations, nationalities and peoples rather than a unitary state of individual citizens; that those nations, nationalities and peoples hold a constitutionally guaranteed right to self-determination, up to and including secession; that federal boundaries track ethnic and linguistic communities rather than administrative geography; and that sovereignty is expressly vested – in the preamble and in Article 8 – in the nations, nationalities and peoples of Ethiopia rather than in an undifferentiated Ethiopian citizenry. Articles 8, 39, 40(3), 46, 47, 49 and 50–52 are not incidental to that architecture. They are the provisions that give ethnic federalism its distinctive normative shape.

How far can that architecture be altered before it ceases to be the same constitutional settlement? Changes to particular provisions need not cross that line. Amendments that, taken together, redefine who bears constitutional sovereignty, the rights of nations, nationalities and peoples, the meaning of self-determination, the territorial basis of the federation and the structure of government would do more than revise discrete rules. They would alter the organising premises through which the 1995 constitution conceives the Ethiopian state, identifies the political community from which its authority derives, and structures relations among its constituent peoples. Changes of that magnitude can no longer be characterised as amendments within the existing order. Their cumulative effect would amount to constitutional re-founding.

Ethiopia’s constitution contains no eternity clause. The amendment power nevertheless has an inherent limit that does not depend on the text saying so, and does not depend on how comprehensive the political consensus behind a given change happens to be. Courts and scholars working in systems with no explicit eternity clause have reached the same conclusion. Colombia’s Constitutional Court developed a “substitution doctrine”: Congress may modify the constitution’s defining pillars within their existing character; it may not replace them. Turkey’s Constitutional Court, reviewing amendments against the constitution’s unamendable provisions, struck down amendments it judged incompatible with the fundamental assumptions of the constitution, before a change in its composition altered the trajectory of that jurisprudence. Yaniv Roznai’s Unconstitutional Constitutional Amendments (2017) offers the most systematic treatment of the debate. The distinction between amendment and replacement is one of degree and of kind: an amendment power is, by definition, a power to alter a constitution while preserving its identity – not a power to produce a different constitution while keeping the old one’s name and numbering. Nothing in that account turns on how widely a change is agreed, inside Ethiopia or outside it. A unanimous decision to replace a constitution’s foundational premise is still a replacement.

There is also the constitution’s longstanding crisis of legitimacy. Across Ethiopia’s fractured political spectrum there is by now something close to consensus that the process which produced the 1995 text was neither democratic nor genuinely inclusive – drafted and ratified by a Constituent Assembly convened when the EPRDF dominated the field and rival forces were marginalised or excluded outright. Abiy himself hasacknowledged as much, contrasting that non-inclusive constitution-making with the broad national consensus he says the present exercise should achieve. The objection has never concerned only the substance of particular provisions. It has also concerned the absence of genuinely shared authorship. Amendment may change the text. It cannot, by itself, repair that original deficit.

Toward a New Social Contract?

Changing the 1995 constitution is very likely warranted. It is warranted, however, only on the condition that the process fixes what denied that constitution its legitimacy in the first place. Among supporters and detractors alike, the real debate has rarely been whether the constitution should be left untouched. It has been about what political and institutional arrangement should replace it – and whether a new social contract can be built with the participation of as much of the country as it is genuinely possible to include.

This is where the National Dialogue’s limitations, not its achievements, deserve the greater share of attention. A process convened by the state, staffed by commissioners appointed by a parliament dominated by the ruling party, and boycotted by some of the most significant stakeholders it was meant to reconcile is not obviously positioned to have done that work impartially – whatever its scale. That is no reason to disregard what it produced, or what comes next. The 1995 constitution functioned for thirty years as the country’s foundational legal and political document despite the legitimacy crisis that shadowed it from the start. It remained the reference point: the text every government, opposition party and court has had to argue from, and a continuing source of regime legitimacy. That combination – an origin most Ethiopians regard as illegitimate, paired with three decades of practical function as the only constitutional text the country has – is itself the strongest argument for taking the present exercise seriously.

Taking the exercise seriously does not mean treating the dialogue’s flaws as proof that some cleaner alternative was realistically available, or that every process of this kind is equally compromised. Processes convened to build national consensus are, to some degree, generative of the agreement they claim only to record. As my late mentor, Professor Emilios Christodioulids, argued, projects of inclusive constitutionalism that claim to host a plural civil society generate their own irresolution: they either demand of constitutional form what it cannot deliver, or they organize, and thereby reshape, the plurality they purport merely to represent. Which agendas get raised, who gets a seat, which proposals survive successive rounds of supermajority voting – all of this shapes the outcome as much as what people actually believe. South Africa’s CODESA negotiations, which produced the framework for its post-apartheid constitution, brought together a genuinely broad range of parties, collapsed in 1992 amid deadlock and violence, and resumed after a narrower bargain between the African National Congress and the ruling National Party. Tunisia’s National Dialogue Quartet, awarded the Nobel Peace Prize in 2015, drew its authority not from breadth of participation but from the specific credibility of four civil-society institutions stepping into a political vacuum. Kenya’s constitutional process ran through an appointed Committee of Experts and a Parliamentary Select Committee before any Kenyan voter had a say through a referendum.

All of these processes involved, in varying degrees, manufactured consensus — filtered through institutional design choices made by whoever convened them. It does not follow that they were equally exclusionary, or that their outputs were equally valuable or worthless. What distinguished the more durable outcomes was what happened afterwards: whether the resulting text was still subjected to a further test of popular authorship that the initial process could not satisfy.

The question Ethiopians should be asking is not whether the consensus generated by the national dialogue was manufactured. Some degree of that is unavoidable. The question is what further test, if any, its output will be made to pass before it is allowed to stand as Ethiopia’s new constitutional settlement.

If Abiy’s ambition is to create “a social contract that future generations will honour,” Ethiopia needs a new constitution made through an inclusive process – one capable of addressing both the exclusions that attended the making of the 1995 constitution and the limitations of the present National Dialogue. The Commission’s forthcoming report to the government is not the end of that work. It is the moment at which the country must decide whether what follows will be amendment within the old order, or the making of a new one.