The 2005 Election Crisis: The Constitution's First Major Failure

Thesis

The May 2005 general election was the first serious test of the 1995 Constitution. It was the first election in which opposition parties were permitted to campaign across the country, granted access to state media, and allowed to organise mass rallies in major cities. It produced an opposition vote share that the constitutional structure had not been designed to absorb: the Coalition for Unity and Democracy (CUD/Kinijit) won, by the official count, 109 seats in the House of Peoples’ Representatives, the United Ethiopian Democratic Forces (UEDF) won 52 seats, and the EPRDF was reduced from a near-total monopoly to a working majority.¹ The official totals are themselves contested; opposition observers and several international observers concluded that the official totals understated the actual opposition performance. The post-election period — protests, mass arrests, the November 2005 crackdown that killed approximately 193 demonstrators in Addis Ababa, the detention of an estimated 30,000 people across the country, the prosecution of 131 CUD leaders, journalists, and civil society leaders on charges of treason and “outrages against the constitutional order” — is the constitution’s first major failure under operational test.²

The article’s argument is that the 2005 crisis was not an aberration; it was the constitutional architecture under load. The EPRDF leadership, faced with the possibility that the constitutional electoral mechanism might produce an outcome they had not expected, made a sovereign decision: they would not accept the constitutional electoral outcome at face value, they would suppress the post-election opposition challenge, and they would use the criminal-law apparatus (treason charges, the post-2005 anti-terrorism framework, the post-2009 Charities and Societies Proclamation) to ensure that no comparable opposition challenge could be organised in subsequent elections. The constitutional protections that should have constrained this response — judicial review of detentions, prosecutorial independence, freedom of expression, freedom of association — did not constrain it. The constitution was, in the May–December 2005 period, the document that the EPRDF interpreted to authorise what the EPRDF had decided to do.

The campaign

The pre-election campaign of 2005 was, by Ethiopian standards, unprecedented in its openness. The EPRDF granted opposition parties access to state-controlled radio and television; allowed them to hold rallies in Addis Ababa and other major cities; permitted relatively unrestricted campaigning across the country. International observers, including the European Union Election Observation Mission led by EU parliamentarian Ana Gomes, described the campaign environment as genuinely competitive in ways that no previous Ethiopian election had been.³

The two major opposition coalitions had distinct platforms.

The Coalition for Unity and Democracy (CUD/Kinijit) was a pan-Ethiopian liberal-democratic platform. Its programme rejected ethnic federalism in its Ethiopian form, called for the strengthening of individual rights against the state, advocated economic liberalisation including the privatisation of land, and proposed constitutional amendments that would remove Article 39’s secession clause. Its leadership included Hailu Shawel (chairman), Birtukan Mideksa (vice-chair), Berhanu Nega (Addis Ababa mayoral candidate and chief economic spokesperson), Mesfin Woldemariam (the founder of the Ethiopian Human Rights Council), and Yacob Hailemariam (a former international prosecutor at the ICTR).⁴

The United Ethiopian Democratic Forces (UEDF) was a coalition of federalist parties that accepted ethnic federalism as the constitutional framework but criticised the EPRDF’s implementation of it. It was led by Beyene Petros and Merera Gudina. Its programme emphasised democratic governance within the federal structure and was less directly confrontational than the CUD’s challenge to the constitutional architecture.⁵

The CUD ran an unusually effective campaign. Its message of pan-Ethiopian unity, its critique of ethnic federalism, and its proposed economic and political reforms resonated with the urban middle class, with the Amhara and Gurage populations who were over-represented in the urban professional and business sectors, and with significant parts of the diaspora that funded the campaign.⁶ The CUD rallies in Addis Ababa in early May 2005 — one massive rally on 7 May for the CUD, a parallel rally for the EPRDF — were among the largest political gatherings in Ethiopian history.

The vote and its aftermath

Polling took place on 15 May 2005. Voter turnout was approximately 90 percent of registered voters. The National Electoral Board of Ethiopia (NEBE), composed entirely of EPRDF appointees, was responsible for vote counting and result announcement. Within days, the NEBE began releasing partial results that suggested an EPRDF victory; the CUD and UEDF rejected the partial results and accused the EPRDF of vote-rigging.⁷

The full official results, released in stages over the next four months:

• EPRDF: 327 seats (59.5 percent vote share by official count)
• CUD: 109 seats (21.9 percent)
• UEDF: 52 seats (10.9 percent)
• Other: 59 seats (7.7 percent)⁸

Crucially, the CUD won 137 of the 138 seats in the Addis Ababa City Council, effectively giving the opposition control of the capital city government. It won 106 of 294 seats in the Amhara regional council. It won significant though minority shares in Oromia, the SNNPR, and other regions.

The post-vote period was politically tense. Both sides claimed the election had been mishandled; the EPRDF accused the opposition of pre-emptively rejecting results, while the opposition accused the EPRDF of rigging. The European Union Election Observation Mission’s final report, released in 2006, identified significant irregularities in the counting process and in the management of complaints; it concluded that the elections “fell short of international standards” while also noting the genuine competitive character of the campaign.⁹

The CUD leadership was internally divided about how to respond. One faction, led by Berhanu Nega, argued for taking the seats won and participating in parliament while challenging the broader results through legal and political means. The other faction, led by Hailu Shawel, argued that participation would legitimise a stolen election and that the CUD should organise mass protests and stay-at-home strikes. The hardliners won the internal argument.¹⁰

The crackdown

In late October and November 2005, the CUD called for stay-at-home strikes and for taxi drivers to honk their horns in protest. Some demonstrations turned violent; buses were burned in Addis Ababa. The federal government responded with mass arrests. Between 7 and 9 November 2005, security forces opened fire on protests in several locations in Addis Ababa, killing dozens. Estimates of the November death toll range from the official 193 to higher numbers cited by opposition sources; in addition, six police officers were killed.¹¹

The arrests escalated rapidly. CUD offices were closed, party membership lists confiscated. An estimated 30,000 people were detained in Addis Ababa and other cities, transported in some cases to detention camps far from urban centres. The detention conditions were poor; reports of torture and inhuman treatment circulated.¹²

The most senior CUD leadership was arrested: Hailu Shawel, Birtukan Mideksa, Berhanu Nega, Mesfin Woldemariam, Yacob Hailemariam, and many others. They were charged with treason, attempt to overthrow the constitutional order, and a range of related offences. The trial began in 2006 and continued for over a year. The defendants refused to recognise the court’s authority or to defend themselves on the merits, arguing that the trial was a political prosecution. They were convicted in mid-2007 and sentenced to long prison terms — some to life imprisonment.¹³

In July 2007, after mediation by a council of elders led by Ephraim Isaac and pressure from international actors, the CUD leaders were released under terms that required them to sign an admission of partial responsibility for the post-election violence. Hailu Shawel and others later said they had signed under duress and had no intention of apologising. The released leaders were granted formal restoration of their constitutional rights.¹⁴ But the CUD as a political organisation had been destroyed: its membership had been intimidated, its organisational infrastructure dismantled, and its leadership scattered into exile (Berhanu Nega and others) or into politically reduced positions within Ethiopia.

The Ethiopian government also took administrative steps to ensure that subsequent elections would not produce similar outcomes. The names of the CUD and the Oromo National Congress (a UEDF member party) were transferred by the National Electoral Board to government-allied splinter groups, depriving the founding leaders of their party names. Local elections scheduled for 2006 were postponed to 2008 and were conducted under conditions in which the EPRDF won effectively all local seats. The Charities and Societies Proclamation of 2009 dramatically restricted civil society. The Anti-Terrorism Proclamation of 2009 (subject of the next article) gave the federal government broad detention and prosecution powers that would be applied repeatedly in subsequent years.¹⁵

The constitutional failure

The 2005 crisis is the constitution’s first major failure under operational test. Several specific constitutional failures should be named.

Article 38 (right to vote) was substantively violated by the post-election manipulation of results and by the prevention of opposition representatives from participating in the post-election process on equal terms.

Article 29 (freedom of expression and opinion) was violated by the closure of opposition media, the harassment of journalists, and the arrest of editors and writers.

Article 30 (right to assembly) was violated by the ban on demonstrations and by the security-force violence against demonstrators in November.

Article 18 (freedom from inhuman treatment) was violated by the conditions of detention and the documented cases of mistreatment in detention.

Article 19 (rights on arrest) was violated by the mass detentions without warrant, without prompt presentation to a judge (the constitutional requirement is 48 hours), and without prompt notification of family members.¹⁶

Article 79 (judicial independence) was effectively violated by the conduct of the treason trial, which the international legal observers — including the International Commission of Jurists — concluded had failed to meet basic fair-trial standards.¹⁷

In none of these violations did any constitutional institution provide effective remedy. The federal judiciary did not strike down the detentions or the trial conduct. The House of Federation did not address the constitutional questions raised. The Federal Police and the Ministry of Justice (the prosecution authority) operated as instruments of the executive without constitutional check. The Ethiopian Human Rights Commission, established in 2000, was not yet operating in a way that produced independent investigation.

The constitutional architecture that the 1995 framers had designed was incapable, in 2005, of protecting opposition participants from a federal executive that had decided to suppress them. The architecture’s hollowness, in this case, was demonstrated comprehensively.

What the crisis established

The 2005 crisis established three things that would shape Ethiopian constitutional politics for the next two decades.

First, that the constitutional electoral mechanism would not produce alternation of power. The EPRDF demonstrated, in May–November 2005, that it would not accept an electoral outcome that produced opposition victory, and that the constitutional protections that should have prevented its post-election response were inoperative. The lesson the opposition drew was that electoral competition within the constitutional framework was, in operational terms, futile.

Second, that the constitutional rights chapter was not, in practice, enforceable against the federal executive. The most extensive bill of rights in Africa had failed, when first seriously tested, to provide actual protection. The lesson for civil society and political organisers was that the constitutional protections existed on paper but could not be relied on in moments of crisis.

Third, that the EPRDF’s tolerance for opposition was conditional on the opposition’s electoral irrelevance. When the opposition became electorally relevant, the EPRDF withdrew the tolerance. The lesson the political class drew was that political organisation outside the EPRDF would always be vulnerable to the EPRDF’s discretionary decisions about how much organising would be permitted.

These lessons shaped the next thirteen years of Ethiopian politics. The 2010 election produced an EPRDF victory of 545 seats out of 547; opposition parties won effectively nothing. The 2015 election produced 547 seats out of 547 for the EPRDF.¹⁸ The constitutional electoral mechanism had been emptied of meaning.

The 2018 transition — the subject of Article 19 — was not an electoral transition; it was an intra-EPRDF leadership change produced by protest mobilisations and elite manoeuvring rather than by constitutional electoral process. The 2021 election under the Prosperity Party produced what the EPRDF had failed to produce in 2005: a dominant ruling party victory in conditions that opposition critics argued were not genuinely competitive. The 2026 election produced a similar outcome with additional constraints from active conflict in Amhara and the absence of voting in much of Tigray.

The 2005 crisis is what set the pattern. Subsequent elections operated within the constraints that 2005 had established.

Steelman of the opposing reading

Two opposing readings of the 2005 crisis deserve engagement.

The first is the EPRDF/federalist reading: that the post-election protests were organised by extremist elements within the opposition who were attempting an extra-constitutional seizure of power, that the November violence was provoked by opposition tactics, and that the government response — though regrettable in its excesses — was substantively defensive of the constitutional order against an attempted overthrow.¹⁹ This is a real argument that has serious advocates. The reply is that the constitutional response to attempted overthrow is constitutional process: prosecution under existing law with fair-trial protections, judicial review of detentions, evidence-based determinations of who was responsible for what. The 2005 response was not constitutional process in this sense. The detentions were mass; the trials were not fair by international standards; the political party suppression went far beyond any specific finding of responsibility. The EPRDF response was an executive decision, not a constitutional one.

The second is the comprehensive-failure reading: that the 2005 crisis demonstrates that the 1995 Constitution was never a serious constitutional document, that the EPRDF’s behaviour was always implicit in the constitutional design, and that the post-2005 erosion was not a deviation from constitutional commitment but its honest expression.²⁰ This is a reasonable critique that this series in significant part endorses. The reply is that the constitution does contain text that, if it had been enforced, would have prevented or constrained the EPRDF’s response. The failure was not that the text was inadequate; it was that the institutional architecture to enforce the text was not in place. This is a difference of degree but it matters for prescription: addressing the institutional architecture is a more tractable problem than rebuilding the text from scratch.

Multi-causal accounting

The 2005 crisis reflects (a) the EPRDF’s decision that opposition electoral success could not be permitted to translate into political power; (b) the CUD leadership’s strategic choice to escalate rather than accept partial gains; (c) the absence of independent constitutional institutions capable of mediating the post-election dispute; (d) the international community’s relatively muted response, which signalled to the EPRDF that severe internal repression would not produce comparable external costs; and (e) the inherited weakness of Ethiopian civil society. Each is a contributing factor. The constitutional failure is the through-line: each of the other factors operated within a constitutional framework that did not constrain.

What the article establishes

The 2005 election crisis is the constitution’s first major failure under operational test. The constitutional architecture proved incapable of protecting opposition participants from federal executive suppression, of producing alternation of power through electoral means, or of constraining the criminal-law apparatus when applied to political opponents. The post-2005 period — examined in the next three articles — operated under the constraints that 2005 had established. The constitution did not produce constitutional democracy; it produced the legal infrastructure within which authoritarian consolidation continued.

---

References

Footnotes

1. For the 2005 results and analysis, see John Ishiyama, “Examining the 2005 Ethiopian Parliamentary Election Results Under Alternative Electoral Rules” (Western Michigan University, 2007); EU Election Observation Mission Final Report, 2006; Human Rights Watch, Suppressing Dissent (2005). ↩

2. Human Rights Watch, “One Hundred Ways of Putting Pressure”: Violations of Freedom of Expression and Association in Ethiopia (2010), https://www.hrw.org/report/2010/03/24/one-hundred-ways-putting-pressure/violations-freedom-expression-and-association. ↩

3. EU Election Observation Mission Final Report, 2006, led by Ana Gomes. ↩

4. On the CUD/Kinijit, see “Coalition for Unity and Democracy,” Wikipedia; Berhanu Nega’s later memoir/account. ↩

5. On the UEDF, see Ishiyama, “Examining the 2005 Ethiopian Parliamentary Election Results.” ↩

6. On the CUD campaign and rallies, see “Ethiopia’s May 2005 elections, It was not your normal election,” Sudan Tribune, January 2006. ↩

7. On the post-election manipulation allegations, see EU Election Observation Mission Final Report; Human Rights Watch, Suppressing Dissent (2005). ↩

8. Official 2005 election results as released by the NEBE in September 2005. ↩

9. EU Election Observation Mission Final Report, 2006. ↩

10. On the CUD internal split, see “For Ethiopia’s political tribes, it’s 2005 redux,” Ethiopia Insight, May 2020. ↩

11. On the November 2005 violence and death toll, see Human Rights Watch, Suppressing Dissent, and the Ethiopian government’s own Inquiry Commission Report (Final Version, October 2006), which acknowledged 193 deaths. ↩

12. On the mass detentions, see Human Rights Watch, “Ethiopia: Crackdown Spreads Beyond Capital” (2005), https://www.hrw.org/news/2005/06/15/ethiopia-crackdown-spreads-beyond-capital. ↩

13. On the 2007 convictions and sentences, see Human Rights Watch, World Report 2008: Ethiopia. ↩

14. On the July 2007 pardons, see “Ethiopian leaders and journalists are out of jail,” Ethiopian Review, July 2007. ↩

15. On the post-2005 administrative measures, see Aalen and Tronvoll, “The 2008 Ethiopian Local Elections: The Return of Electoral Authoritarianism,” African Affairs 108 (2009); Charities and Societies Proclamation No. 621/2009; Anti-Terrorism Proclamation No. 652/2009. ↩

16. 1995 Constitution, Articles 18, 19, 29, 30, 38. ↩

17. On the fair-trial standards in the treason trial, see International Commission of Jurists statements 2006–2007; Human Rights Watch trial observations. ↩

18. On the 2010 and 2015 elections, see Aalen and Tronvoll, “The 2010 Ethiopian Elections: Re-establishing the One-Party State,” African Affairs 110 (2011); various International Crisis Group reports. ↩

19. This is broadly the EPRDF and post-EPRDF government’s position; see Meles Zenawi’s various 2005–2006 statements and subsequent EPRDF apologetics. ↩

20. This is the position of the comprehensive critique; see Aalen, Tronvoll, and the broader human-rights-organisations literature. ↩