Mozambique 2024 General Election

Acórdão 24/CC/2024 places the epicentre of the discrepancies in the intermediate district count, where it says the data were inflated, and partially invalidates Deliberação 105/CNE/2024 of 24 October, in the part concerning the districts it lists, across Maputo City and the provinces of Maputo, Gaza, Inhambane, Tete, Zambézia and Nampula; for Manica, Sofala, Cabo Delgado and Niassa, where the volume of litigation did not warrant requisitioning the records, it assessed the election on the parties' copies alone. Re-tabulating from the polling-station results sheets, the Council reduced Daniel Chapo from 70.67 to 65.17 per cent and FRELIMO from 195 to 171 of 250 seats, while raising PODEMOS from 31 to 43, RENAMO from 20 to 28 and MDM from 4 to 8.

The EU mission reported instances of fraudulent alteration of polling results, deliberate invalidation of opposition votes and several instances of results manipulation conducted at district level. Analysing 121 of the 161 district provisional results, it found seven districts where the number of votes cast in the legislative election ran far above the presidential election - Chibuto by 15 per cent, Inhambane City and Maxixe by 21, Kamubukwana by 24, Quelimane by 30, Pebane by 60 and Massinga by 83 - that practically all the additional votes went to the ruling party, and that in those seven districts alone FRELIMO appeared to have been unduly awarded over 150,000 extra votes. In Manhiça, where observers tallied the polling-station results as they were read aloud, Daniel Chapo was announced with 42,485 votes against the roughly 28,000 the observers had summed. In four districts of Zambézia observers were asked to leave the tabulation centre on the second day after noting differences between polling-station and tabulated figures.

The commission published results through WhatsApp while its own website and Facebook page stayed inactive; district results for 40 of 161 districts were never available to the EU mission or to national actors. The technical secretariat told the EU mission that publication by polling station was technically possible; the commission's president said no law required it. Asked to explain why voter numbers differed between the three elections, the commission attributed it to errors on polling-station forms; the Ministério Público found that answer incongruent and without technical or legal basis, and the Constitutional Council adopted that position in full. EU observers were also prevented from watching tabulation in some districts and provinces and at national level.

In the early hours of 19 October 2024, gunmen blocked a car on Avenida Joaquim Chissano and killed the two front-seat passengers, lawyer Elvino Dias and Paulo Guambe; a third passenger was wounded. The Centro de Integridade Pública describes Dias as the election agent (mandatário) of Mondlane's CAD coalition before that coalition was de-registered, and Guambe as the election agent of PODEMOS; the EU mission describes Dias as Mondlane's legal adviser and Guambe as PODEMOS's spokesperson. Dias was at the time preparing Mondlane's challenge to the Constitutional Council, which was due to be filed that week. The EU mission recorded that at the time of writing its final report the perpetrators had not been brought to justice.

The EU mission states that police used live ammunition and tear gas on unarmed civilians and that the protests left over 300 people dead, citing the civil-society Plataforma Decide for 314 killed by firearm, around 633 injured by firearm and more than 4,236 arrested as of mid-January 2025. Amnesty International, drawing on the same platform's telephone hotline tally and on 28 interviews and more than 100 verified videos and photographs, reports around 315 killed and over 3,000 injured between 21 October 2024 and 16 January 2025, against police commander Bernardino Rafael's acknowledgement of 96 deaths including 17 officers as of 23 January 2025. International IDEA's democracy tracker records over 350 killed as of the end of February 2025 without naming who compiled that figure. Every count above 96 is a civil-society tally collected through hotlines and video verification, not an official death toll.

The EU mission's social-media unit identified 51 inauthentic Facebook accounts and 23 inauthentic Instagram accounts, of which 68 supported Daniel Chapo. On 25 October, the day after Mondlane's Facebook broadcast responding to the commission's announcement drew a national record 115,000 views, mobile internet services of all three operators suffered a national data blackout, which the EU mission characterised as a limitation of the rights to information and expression by the government. Amnesty International's connectivity analysis, drawing on IODA, Cloudflare Radar and OONI, found access to Facebook, Instagram and WhatsApp blocked or reduced during peak protest periods in October and November 2024.

PODEMOS submitted a parallel count covering, by its own account, 59.95 per cent of the total vote, giving Mondlane 4,419,040 votes or 53.38 per cent. The Council pointed out that the submission's total of 8,244,403 valid votes already exceeded the commission's entire national valid-vote total of 6,961,925 by 1,282,478, and exceeded the 7,238,027 people recorded as having voted by 1,006,376; scaled to 100 per cent coverage it would have implied more votes than there were registered voters. Recounting the results sheets PODEMOS itself supplied, the Council obtained 2,229,609 valid votes, with Chapo on 57.79 per cent and Mondlane on 31.67 per cent, and dismissed the request. What this refutes is the parallel count as submitted; it does not establish what an accurate national count would have shown, and no recount of ballots was ever conducted.

The Council held that copies of the polling-station results sheets in the hands of political parties carry the same legal force as those supplied by the commission, and that comparing the two sets was enough to locate where the discrepancies arose, making the second option, a recount of the ballots, unnecessary. It did not publish the criteria by which it accepted one sheet over another, or which individual sheets it accepted or rejected. The Mozambican Bar Association said in a 13 December statement that this comparison of results sheets did not follow the legal criteria for validating documentary proof and was illegal, and recommended either a recount or annulment of the election. The EU mission found that the Council modified the commission's result without providing a sound legal basis and recommended that future validations state the criteria used. Two domestic observer consortiums, Mais Integridade and Decide, refused to hand over their own copies of the results sheets, telling the Council in December that the documents they held would not help establish the electoral truth and would serve to legitimise fraud.

On 11 October the SADC mission concluded that the pre-election and voting phases had been professionally organised and conducted in an orderly, peaceful and free atmosphere, and the African Union mission found the process was held in accordance with the electoral law and fundamental rights and freedoms. Both statements were preliminary, issued two days after polling and before tabulation was complete. The EU mission, which observed through the tabulation phase and into November, found the count and tabulation marked by anomalies, serious flaws and an acute lack of transparency.

During voting, EU observers recorded four cases of voters receiving more than one ballot per election, 30 improperly sealed ballot boxes, and five of 182 stations where seal numbers did not match the polling logbook, describing these as practices suggesting potential stuffing before or during their presence. During the count they found that in 16 of the 51 stations where the ballots in the box were compared against the voters marked off on the voter list, the ballot count exceeded the voter count in every one of the three elections, and they saw stacks of folded ballot papers in ten counts; the mission called both indications of possible stuffing. A late amendment introducing ballot boxes with a narrower slot, proposed by the opposition to stop the practice, was not applied to this election. Neither the Constitutional Council nor any observer mission quantified stuffing or identified it as the mechanism behind the national figures.

The two men were shot by unidentified armed men. Many of the EU mission's interlocutors believed the killings were meant to silence and weaken Mondlane and PODEMOS before the provisional results and to disrupt their legal challenge, but no perpetrator had been identified or prosecuted by the time the mission published its final report in January 2025, and none of the sources cited here records a prosecution or a named suspect since.

A total of 17,169,239 citizens were registered, including 333,839 in the diaspora, a 30 per cent increase on 2019 in a period when the National Statistics Institute projected adult-population growth of 17 per cent. In-country registration reached 103.81 per cent of the projected eligible population, ranging from 86.21 per cent in Niassa to 150.01 per cent in Gaza; provincial commissions reported rates of 174, 204 and 215 per cent in the districts of Chokwe, Ibo and Moatize, and the commission itself put the national surplus at roughly 650,000 registrants above what the projections would support had every eligible citizen enrolled. The election administration attributed the high figure to easier access to registration and the new plastic voter card, while acknowledging that a fuller study was needed; the EU mission noted that the register and the census projections are compiled by different methods, that the commission published no district-level rates for any of the 161 districts, and that no audit of the register was carried out. The gap was therefore never resolved in either direction.