The Tigray War as Constitutional Question: Federation, Secession, and Article 39

Thesis

The Tigray war that began on 4 November 2020 and ended formally with the Pretoria Agreement of 2 November 2022 was the most destructive constitutional event in modern Ethiopian history. Estimates of the death toll vary from 385,000 (the lowest credible estimate, from late-2022 academic analysis) to 600,000–800,000 (the range cited by AU lead mediator Olusegun Obasanjo and by post-war assessments), with some assessments approaching one million when famine and excess mortality are included.¹ The war displaced over two million people, destroyed substantial infrastructure across northern Ethiopia, drew in Eritrean forces and Amhara militias, and produced documented atrocities by all parties. As of mid-2026, the constitutional crisis the war embodied has not been resolved: the Pretoria Agreement’s implementation has been incomplete; the TPLF has been deregistered as a political party; the federal government has removed five constituencies from Tigrayan electoral oversight; the TPLF Central Committee voted on 19 April 2026 to reinstate the pre-war regional council, in what the federal-aligned Tigray Interim Administration has called a nullification of the Pretoria Agreement; and the Critical Threats Project assesses renewed conflict as increasingly probable.²

The article’s argument is that the Tigray war was, in its constitutional content, a war over what the 1995 Constitution meant. It was not a war for Tigrayan secession in the Article 39 sense; the TPLF did not invoke Article 39 procedures. It was not a war about boundary disputes in any narrow sense; the boundary questions were occasions for the conflict, not its substance. It was a war about whether the federal government or the regional government had the authority to determine the constitutional status of a regional government — about which body’s constitutional interpretation would prevail in the conflict that the June 2020 House of Federation decision had created. The federal government’s military victory in November 2022 settled the question in its favour by force of arms, but it did not produce a constitutional resolution. The constitutional question remains open in 2026, and the renewed crisis since April 2026 is the consequence.

The constitutional content

The Tigray war can be analysed at several levels — military, humanitarian, regional-geopolitical, diplomatic. The constitutional level is the one this series is concerned with.

At the constitutional level, the war was a confrontation between two opposing interpretations of the 1995 Constitution.

The federal interpretation. The federal government’s position, articulated by Abiy Ahmed and Prosperity Party officials throughout the war and afterwards, was that:

(a) The June 2020 House of Federation decision extending the legislative terms was constitutionally valid, having been made through the constitution’s interpretive procedures (Articles 83–84);

(b) The TPLF’s September 2020 regional election was therefore unconstitutional, as the House of Federation declared on 5 September 2020;

(c) The Tigray regional government after October 2020 was constitutionally illegitimate, lacking the federal recognition that the constitution requires;

(d) The TPLF’s attack on the Northern Command on 3–4 November 2020 was constitutionally insurrectionist — an attack on federal institutions by a non-constitutionally-recognised regional authority;

(e) The federal government’s military response was a “law enforcement operation” rather than a war, conducted under federal authority over a region whose constitutional status was contested.³

This is a coherent constitutional position. It depends on accepting the June 2020 interpretation as constitutionally valid and on accepting the federal government’s authority to determine which regional governments are constitutionally recognised.

The TPLF interpretation. The TPLF’s position, articulated by Debretsion Gebremichael and Getachew Reda throughout the war, was that:

(a) The June 2020 House of Federation decision was unconstitutional, having introduced constitutional language not present in the text and having been issued by a body politically dependent on the federal executive;

(b) The Tigray regional government’s term continued constitutionally until elections, which the TPLF organised in September 2020 in compliance with what the TPLF understood as the constitutional requirement for periodic elections;

(c) The Tigray regional council elected in September 2020 was the constitutionally legitimate regional authority for Tigray, holding its mandate from the Tigrayan electorate under the constitutional framework;

(d) The federal government’s actions against the Tigray regional government — budget suspension, diplomatic isolation, military preparation — were constitutional violations of the federal-state relationship the constitution prescribed;

(e) The Northern Command attack of 3–4 November was a pre-emptive defensive action against an impending federal military operation that the TPLF had documented in advance.⁴

This is also a coherent constitutional position. It depends on rejecting the June 2020 interpretation as constitutionally invalid and on asserting the regional government’s constitutional authority within Tigray.

The two interpretations are mutually incompatible. Both invoke the constitutional text. Neither was adjudicated by any institution capable of independent adjudication. The conflict was, in this sense, constitutional in its content: the parties were disputing what the constitution meant. The conflict was resolved, in November 2022, by military victory rather than by constitutional adjudication.

The Article 39 question

Article 39’s secession clause was invoked rhetorically by various parties during and after the war, but it was not the operational constitutional question.

The TPLF did not invoke Article 39 procedures. It did not call a two-thirds vote of the Tigray regional council to demand secession; it did not request a federal-government-organised referendum; it did not seek a transfer of federal powers to Tigray; it did not invoke the constitutional procedures for division of assets. The TPLF’s claim was constitutional standing within the federation, not secession from it.⁵

The federal government nevertheless framed the conflict in terms that implicated Article 39. The federal narrative was that the TPLF was attempting to secede in fact if not in form — that the September 2020 election, the rejection of federal recognition, the holding of regional military forces, and the attack on the Northern Command together amounted to an extra-constitutional secession that the federal government was constitutionally obligated to prevent. The Eritrean precedent of 1991–93, in which de facto independence preceded the constitutional referendum, was implicitly invoked: the federal government acted to prevent the de facto independence that, in its framing, would have produced a fait accompli that Article 39 procedures could not undo.⁶

The framing matters because it converts the political conflict into an existential constitutional question. A regional government that defies federal authority is, in this framing, not asserting constitutional rights but attempting constitutional dissolution. The federal response, on this framing, is not war but constitutional defence. The framing was used internally to mobilise the federal political base and internationally to justify the military operations.

Whether the framing was accurate is contested. Critics of the federal position have argued that the TPLF’s September 2020 election was a constitutional act, that the TPLF’s claim was for constitutional standing within the federation, and that the federal framing of the conflict as quasi-secession was a strategic narrative rather than an accurate constitutional analysis.⁷ The series shares the critics’ view in significant respects: the TPLF’s documented constitutional claims do not amount to secession in the Article 39 sense, and the federal framing was strategic. But the series also notes that the constitutional question is not fully resolved either way, and that the Schmittian framework — sovereign decides on the exception — applies to both parties’ positions. Each was making a sovereign decision about what the constitution meant in the absence of an institution capable of independent decision.

The war’s conduct

The war’s conduct produced what international observers and human-rights organisations have documented as systematic violations of international humanitarian law by all parties.

The Ethiopian National Defence Force and allied Amhara militias, with substantial support from Eritrean Defence Forces who entered Tigray in November 2020, conducted operations that produced large-scale civilian casualties, sexual violence, destruction of medical facilities, looting of property, ethnic cleansing in western Tigray, and a blockade of humanitarian aid that produced famine conditions for substantial parts of the Tigrayan population.⁸ The most heavily documented atrocities — Aksum, Adwa, Dengelat, Mai Kadra, Adigrat — were conducted by combinations of ENDF, EDF, and Amhara forces.

The Tigray Defence Forces, the military arm of the Tigray regional government, conducted operations in Tigray, Amhara, and Afar regions that also produced documented atrocities — particularly in the 2021 Amhara campaign, when TDF forces pushed south and east toward Addis Ababa. Documented massacres at Chenna and Kobo and reported atrocities in other towns under TDF control during the 2021 offensive add to the war’s accumulated record.⁹

The constitutional significance of the conduct is that the laws of war — the procedural and substantive constraints that operate even during armed conflict — were not observed by any party in ways that the constitutional order’s commitments to human rights would have required. The 1995 Constitution’s rights chapter (Articles 14–28) is among the most extensive in any constitution. Its substantive protections — life, security of person, freedom from torture, fair trial — were among the most extensively violated of the war’s outputs. The constitutional commitment to human rights did not constrain the constitutional government’s military conduct. The constitutional commitment to humanitarian law did not constrain the regional government’s military conduct. The constitution operated, in this respect as in others, as text without substance.

The Pretoria Agreement

The Pretoria Agreement of 2 November 2022 was the African Union-mediated peace deal that ended the active conflict. Its principal terms:

(a) Permanent cessation of hostilities;

(b) Disarmament, demobilisation, and reintegration of TDF forces (DDR);

(c) Restoration of federal authority in Tigray, including reopening of federal institutions and resumption of federal services;

(d) Establishment of a Tigray Interim Administration to govern the region pending elections;

(e) Return of internally displaced persons;

(f) Accountability for atrocities, transitional justice;

(g) Resolution of contested territories.¹⁰

The agreement was, in its terms, a comprehensive framework for post-conflict reconstruction. In implementation, it has been partial.

The disarmament has been partial. The TDF has retained military capacity that, in the federal government’s framing, exceeds what the agreement permits. The Tigray Interim Administration, established with federal endorsement, has overseen partial demobilisation but not full disarmament.

The return of displaced persons has been incomplete. As of 2026, over one million Tigrayans remain displaced, primarily because the contested territories in western Tigray (Welkait, Tselemt, Humera) remain under Amhara administration, where displaced Tigrayans cannot safely return.¹¹

The accountability process has been minimal. The transitional justice process that the agreement contemplated has not been implemented at meaningful scale. The federal government has not produced credible accountability for ENDF or EDF actions; the Tigrayan administration has not produced credible accountability for TDF actions. The most documented atrocities have not been prosecuted.

The territorial questions have not been resolved. Western Tigray remains under Amhara administration; the constitutional question of whether the territory is constitutionally part of Tigray or Amhara has not been adjudicated.

The political question of who governs Tigray has been contested throughout the post-Pretoria period. The Tigray Interim Administration, initially headed by Getachew Reda (April 2023–April 2025) and then by Lt. Gen. Tadesse Worede (April 2025–present), has operated under federal authority. The TPLF has contested the Interim Administration’s legitimacy in various ways; the post-April 2024 expulsion of Getachew Reda from the TPLF, the March 2025 hardliner consolidation, the April 2026 vote to reinstate the pre-war regional council — these are sequential moves by the TPLF to reassert its constitutional claim.¹²

The post-April 2026 crisis

The April 2026 crisis is, in operational terms, the post-Pretoria continuation of the constitutional conflict.

On 8 April 2026, the federal government extended the mandate of Lt. Gen. Tadesse Worede as president of the Tigray Interim Administration for an additional year. The extension was based on Tadesse’s annual report and the federal government’s assessment that conditions were not yet sufficient for regional elections.¹³

On 15–16 April 2026, the TPLF Central Committee held a meeting in Axum. On 19 April, the Committee announced its decision to reinstate the pre-war Tigray regional council — the council elected in September 2020 whose constitutional legitimacy the federal government had rejected. The Committee’s framing was that the council elected by more than 2.8 million Tigrayans was the constitutionally legitimate authority and should resume its functions, and that the Interim Administration was a federal imposition without democratic mandate.¹⁴

The Tigray Interim Administration’s response, articulated by Tadesse Worede on 21 April, was that the TPLF decision violated the Pretoria Agreement, that the reinstatement of the pre-war council would “nullify” the agreement, and that the Interim Administration would not “hand over” power.¹⁵

The post-April 2026 trajectory has been one of escalation:

• The TPLF has reinstated military command over the TDF;
• The federal government has continued military buildup along Tigray’s borders since February 2026;
• The Eritrean military has continued occupation of contested Tigrayan territories;
• The Fano insurgency in Amhara has constrained federal capacity to deploy fully toward Tigray;
• The Tigray Peace Forces, an anti-TPLF rebel group, has clashed with TPLF forces in Tigray;
• Letters from Ethiopia to the UN in June and October 2025 alleged coordination among the TDF, the Eritrean army, and Fano in operations against federal forces;
• The federal government has imposed economic restrictions on Tigray, including fuel restrictions and budget pressures.¹⁶

As of mid-2026, the constitutional question of who governs Tigray, and the broader question of what the federation’s relationship with Tigray will be, remain unresolved. The Critical Threats Project assesses renewed conflict as increasingly probable.

Steelman of opposing readings

The federalist defence of the federal position: the federal government conducted a legitimate constitutional defence of the federation against a regional rebellion that, if successful, would have triggered the federation’s dissolution. The military victory in 2022 prevented the worst outcome — the disintegration of the Ethiopian state along ethnic-territorial lines — and the post-Pretoria framework has produced, with all its imperfections, a path to reintegration. The April 2026 TPLF moves are confirmation that the original constitutional concern was correct: the TPLF, given the opportunity, has reasserted the same claim it made before the war.¹⁷ This has weight. The reply is that the constitutional resolution by force did not address the underlying constitutional question; it deferred it. The April 2026 crisis is the deferred question reappearing.

The TPLF/Tigrayan defence: the federal government conducted a war against the Tigrayan people in defence of an unconstitutional federal interpretation, used the war to dismantle Tigrayan political and military capacity, and is now using the post-Pretoria framework to extend that dismantling through political means. The April 2026 moves are the legitimate reassertion of Tigrayan constitutional rights that the federal interpretation has continued to violate.¹⁸ This also has weight. The reply is that the TPLF’s constitutional claims, while not amounting to secession, were also not free from problematic strategic choices — particularly the September 2020 unilateral election that gave the federal government the constitutional framing it used to justify the war. Both parties operated within a constitutional framework that did not constrain them; both made choices that escalated the conflict.

Multi-causal accounting

The Tigray war reflects (a) the constitutional dispute over the June 2020 interpretation; (b) the political conflict between Abiy and the TPLF leadership over the Prosperity Party reorganisation; (c) the strategic alignment between the federal government and Eritrea against the TPLF; (d) the Amhara nationalist claims to western Tigray and other contested territories; (e) the regional geopolitical context including Sudanese, Egyptian, and Gulf state interests; and (f) the long history of TPLF-Amhara, TPLF-Eritrean, and TPLF-Oromo conflict that pre-dated the constitutional dispute. The constitutional dispute is one cause among several; it is the cause that this series focuses on because it is the one that the constitutional architecture was supposed to address and did not.

What the article establishes

The Tigray war was, in its constitutional content, a war over what the 1995 Constitution meant. The federal interpretation and the TPLF interpretation were both grounded in the constitutional text, both internally coherent, both unmediated by any institution capable of independent adjudication. The military resolution did not produce a constitutional resolution. The post-Pretoria framework has been incompletely implemented. The post-April 2026 crisis is the constitutional dispute reappearing in operational form. The war’s death toll, displacement, atrocities, and continued political consequences are part of the cost of having a constitutional order that does not produce binding adjudication of constitutional disputes.

The next article takes up the broader continuing crisis of the federation in 2026: the Amhara conflict, the OLA insurgency in Oromia, the unresolved SNNPR questions, and the federation’s structural fragility.

---

References

Footnotes

1. On death toll estimates, see “Ethiopia’s Abiy is using the Pretoria accord to strip Tigrayans of their rights,” The Africa Report, April 2026, https://www.theafricareport.com/414003/ethiopia-the-pretoria-agreement-as-a-weapon-of-war-against-tigray/, citing Obasanjo’s 600,000 estimate; “Tigray war,” Wikipedia, citing 385,000–600,000 academic range; The Human Rights Crisis in Ethiopia, May 15, 2026, https://www.genocidewatch.com/single-post/the-human-rights-crisis-in-ethiopia-may-15-2026, citing up to one million. ↩

2. For the post-April 2026 crisis, see “Tplf Undoes Tigray Peace Agreement,” Africa File, April 23, 2026, https://www.criticalthreats.org/analysis/ethiopia-tigray-tplf-tia-mozambique-rwanda-ismp-sudan-rsf-saf-africa-file-april-23-2026. ↩

3. For the federal position, see Abiy’s addresses to the House of Peoples’ Representatives, November 2020 through 2026; Ethiopia’s letters to the UN in June and October 2025. ↩

4. For the TPLF position, see TPLF Central Committee statements 2020–2026; Debretsion Gebremichael’s letters and addresses. ↩

5. This is supported by the TPLF’s failure to invoke Article 39 procedures and by its repeated framing of its claims as constitutional standing within the federation. ↩

6. On the federal framing of the conflict as quasi-secession, see various Abiy speeches and government communications. ↩

7. See “Ethiopia’s Abiy is using the Pretoria accord to strip Tigrayans of their rights,” The Africa Report; “Ethiopia: Tigray’s parallel governance crisis tests limits of the Pretoria agreement,” The Africa Report, May 2026, https://www.theafricareport.com/417854/ethiopia-tigrays-parallel-governance-crisis-tests-limits-of-the-pretoria-agreement/. ↩

8. On war conduct, see Human Rights Watch and Amnesty International reports 2021–2023; UN Human Rights Council Commission of Inquiry reports; the documented MSF report on the killing of three MSF staff in 2021. ↩

9. For TDF conduct, see Amnesty International reports on Chenna, Kobo, and other locations; Human Rights Watch World Report 2022. ↩

10. Pretoria Agreement, 2 November 2022; “Ethiopia–Tigray peace agreement,” Wikipedia, https://en.wikipedia.org/wiki/Ethiopia%E2%80%93Tigray_peace_agreement. ↩

11. On displacement and Western Tigray, see The Africa Report April 2026 analysis. ↩

12. On the Interim Administration leadership transitions, see “TPLF oppose extension of Tigray Interim Administration,” Addis Standard, April 2026, https://addisstandard.com/news-tplf-oppose-extension-of-tigray-interim-administration-says-violates-pretoria-agreement/. ↩

13. Addis Standard, 9 April 2026. ↩

14. On the TPLF Central Committee decision, see “Tigray Interim President says TPLF move to reinstate pre-war council ‘nullifies’ Pretoria Agreement,” Addis Standard, April 2026, https://addisstandard.com/tigray-interim-president-says-tplf-move-to-reinstate-pre-war-council-nullifies-pretoria-agreement/. ↩

15. Tadesse Worede press briefing, 21 April 2026. ↩

16. On the post-April 2026 trajectory, see “Tplf Undoes Tigray Peace Agreement,” Africa File, April 23, 2026; “Military Buildup In Northern Ethiopia,” Africa File, February 12, 2026. ↩

17. For the federalist defence, see various ENA and Government Communication Service statements; Tigray’s Reckoning analysis at The Reporter Ethiopia. ↩

18. For the TPLF/Tigrayan defence, see TPLF statements and Addis Standard coverage 2025–2026. ↩