The Electoral Observatory, the largest and most credible group of Mozambican election observers, has announced that for the general elections scheduled for next Wednesday, it will not be publishing the results of its parallel count of the results.
The director of the Observatory, Rev Anastacio Chembeze, made this announcement at a round table organised by the anti-corruption NGO, the Centre for Public Integrity (CIP) last Tuesday.
This is a sudden change in the Observatory's methods of work - at previous general elections, its parallel count, covering almost every district in the country (and eight per cent of the total number of polling stations in 2009), had won wide praise, as a valuable check on the official results.
But this time, although the Observatory will still undertake a parallel count, it will not publish the results, under an agreement reached with the National Elections Commission (CNE). Chembeze made it clear at the round table that this decision would not be reversed.
A press release from CIP says this decision was received with “surprise and concern” by the other participants at the round table. CIP remarks that the parallel count is the only activity which justifies the huge investment involved in deploying the Observatory's 2,500 observers.
“When the parallel count is not followed by publishing its results, then it is emptied of content”, adds CIP. “The parallel count is a fundamental element in avoiding electoral fraud since it makes citizens aware of the results during the count and it allows comparison of the official results with the results gathered by independent observers”.
The round table participants described the Observatory's decision as “a great setback” for the transparency of Mozambican elections.
In 2009, it was the Observatory's parallel count which allowed AIM to estimate the scale of fraud in those elections. This could be done statistically, looking for how many polling stations in the Observatory's sample had serious anomalies. Since the sample was trustworthy, and genuinely random, the results could be extrapolated to all polling stations.
The main type of fraud experienced in Mozambican elections is the addition by dishonest polling station staff of extra ink marks on ballot papers, making it look as if the voters concerned have tried to vote for two or more candidates. This always shows up as a suspiciously large number of invalid votes on the polling station results sheets.
Then there are those polling stations where corrupt staff indicate an impossibly high turnout of 100 per cent or more of the registered voters - a serious problem in parts of Tete province, particularly Changara district which was completely corrupt in 2009.
If there is no longer a random sample of polling stations, such as that which the Observatory used to provide, then the only way of checking the scale of fraud will be to go through all 17,201 polling station results sheets on the data base of the Electoral Administration Technical Secretariat (STAE), which would be a hugely time consuming task.