The CNE found 152,000 more votes for parliament (AR) than the district elections commissions (CDEs) reported. How, where? If anyone knows, they are not telling. Are they real, mistakes, or frauds? There is no way to know. Comparing registration figures shows that the CDEs excluded some polling stations which the CNE included, but there is no way of knowing why. Did CDEs exclude some polling stations which had unbelievably high turnout, but the CNE include them?
The CNE does not publish results by district; results sheets (editais) are posted by district and provincial elections commissions, but the CNE does not make them available on the web. Parties and journalists must go to the elections commission offices and copy down the numbers – which our team did. CNE remains above the law. It must publish district results at least from provincial assemblies (APs), but it did not.
We have posted district elections commissions results for president (http://bit.ly/Dis-Pres) and parliament (AR, http://bit.ly/Dis-AR) The Constitutional Council did publish district results for APs. We compared them to the CDE official results from our correspondents; almost every district was different.
The electoral law (2/91) sets up an elaborate cascade process for tabulation – district STAEs add up the polling station results sheets (editais), provinces add the districts, and the CNE adds up the provincial lists. But it apparently never happens that way. A copy of every polling station edital is also sent to national STAE, which apparently does its own count which it gives to the CNE – thus the differences from the district counts.
The secrecy of the second count is compounded by basic arithmetic errors made (in secret) by the CNE and CC – the two senior bodies on elections. The CNE repeatedly left out the diaspora vote from some, but not all, of its totals. It approved these wrong results and gave them to the CC which rubber stamped them. Following Bulletin articles, in secret, it twice corrected the error – the first time for President and the second time, partially, for AR. But the second secretly corrected version still has obvious errors. The first table on page 32 of the CC ruling (acórdão 25) has been corrected but not the second; the table “Eleições para deputados da Assembleia da República” is correct but the table below “Distribução de mandatos” still excludes votes from the diaspora.
In two different tables later in the ruling annex, on page 104 “Eleição dos Deputados da AR” and page 487 “Eleição does Membros das Assembleias Provinciais” the bottom row “votos na urna” (votes in ballot box), is different from “Númerio Total de Votantes” (Total Number of Voters) higher up in the table, yet by definition they must be the same.
Votos na Urna are both incorrect because the totals exclude nulos (invalid ballots), which were obviously in the ballot box. This, in turn, leads to another error. On page 104 the percentages for blank and invalid votes are wrong because they are based on the erroneous Votos na Urna rather than Númerio Total de Votantes. CC and CNE may dismiss these as small errors, but if the highest court in the land and the top electoral authority cannot even add up columns of numbers, how can any of their secretly compiled numbers be trusted?
Sloppiness in secret has proved contagious. Observers reported widespread violations of procedures during the counting at polling stations and in the district tabulation processes, which opened the door to errors and fraud. Results sheets were filled in with no one watching, and some numbers were changed. In key areas, observers could not obtain credentials, so there was no check.
It is hardly surprising that many people do not trust a vote counting process riddled with errors and lack of transparency. A meme circulated widely on the internet during 2018 local elections advertising a master’s degree in electoral fraud taught by Mozambican professors with 24 years practical experience.
Just using the limited official information and the civil society parallel count (PVT), we are able to identify 557,000 votes given to President Filipe Nyusi or taken from opposition candidate Ossufo Momade (detailed below). This is 8% of the total vote and 17% of Nyusi’s margin over Momade. We also find 5 parliament (AR) seats improperly given to Frelimo instead of Renamo. And this is just the tip of the iceberg we can see. Because of the secrecy and lack of information, it is impossible to identify, quantify and pinpoint the large amounts of fraud we cannot see.
