Document #1452464
Bijlert, Martine van (Author), published by AAN – Afghanistan Analysts Network
What can we expect on Afghanistan’s third election day?
Kandahar’s Provincial Electoral Officer Neamatullah Wardag told AAN that the IEC intends to open 173 polling centres on 27 October 2018, with a total of 1,113 polling stations. This is, on average, a little over six polling stations per centre but the variation in size will be wide. Some centres registered only few voters and will have one or two stations only, while other will have to have at least 15 or more. (1)
Some polling centres, particularly in Kandahar city, would be very crowded if all voters did indeed turn up (which is unlikely, even in the city). Nahiya nine in Kandahar, for instance has almost 29,000 voters registered in three polling centres, while three other urban districts have around registered 45,000 voters for six polling centres. Of course there will be many polling stations, but the crowds could still be large. Alternatively, if the centres are captured by strongmen or partisan security forces, as has happened in the past, the opportunity for fraud in these places would be significant.
So far, there have been no signs of a breakdown in the security situation or in discipline in the security forces, since Abdul Razeq’s death (AAN background here). Security in Kandahar city has been tight and there has been a steady stream of high profile visitors who have come to pay their respects, including President Ghani, Chief Executive Abdullah, former president Karzai and several Jamiat leaders.
It is difficult to predict the overall turnout, other than that it will be very low in the remote districts, where Taliban threats warning people not to take part in the election are expected to keep people away from the vote. In Kandahar city, Abdul Razeq’s supporters believe turnout might be buoyed by the wish to honour his memory. They emphasised that he had been killed only minutes after a meeting that focused on election security and had called on the people of Kandahar to come out and vote.
The death of Police Chief Abdul Razeq has the potential to rearrange Kandahar’s political scene. Different groups, both from Kandahar (such as the Karzai and Gul Agha families) and outside (President Ghani and others) may see new opportunities to try to get their own affiliates elected. His absence will also be felt during this vote, even though so far there seems to have been no security vacuum. But he was a looming figure in the elections, controlling security and, sometimes, the vote.
Unpacking Kandahar’s election numbers
There are a total of 112 candidates (13 are women, more by AAN here) competing for Kandahar’s eleven seats, of which three are reserved for women. Five have registered a political party affiliation; the remaining 107 have registered as independents. 99 of the candidates are men, 13 are women (full candidate list here). According to the 2018 population estimate by the Central Statistics Office (CSO), Kandahar has a total population of around 1,3 million.
According to the IEC’s voter lists Kandahar registered a total of 522,984 voters in this year’s new voter registration: 442,512 men and 67,276 women for the vote in the provincial constituency, and 13,196 Kuchis (gender not specified) for the nationwide election of kuchi representatives.(2)
The rate of registration varied greatly across the province. The district with the most registered voters was Spin Boldak, the birthplace and main power base of General Abdul Razeq. In Spin Boldak 86,082 voters registered to vote in 26 polling centres (11,801 of them were women; 2,015 were Kuchis), representing 15 per cent of the province’s total number of votes.
Four urban districts in Kandahar city (nahiya one, two, four and ten) each registered between 45,000 and 55,000 voters, representing a little less than 10 per cent of the total vote each. Three other urban districts registered a fair number of voters, (almost 29,000 in nahiya nine, and around 13,000 in nahiyas seven and eight). The five rural districts closest to the city – Zherai, Panjway, Dand, Arghandab and Daman – each registered around 20,000 to 30,000 voters in 16 or 17 polling centres each.
After that, it starts petering out: Mianeshin (909 voters), Ghorak (564 voters), Khakrez (409 voters), Shorabak (203 voters) and Reg (161 voters) all registered voters in single polling centres – with no women, except a handful in Khakrez. There was no voter registration in two rural districts – Nesh and Maruf – which also means there will be no vote. There are also four urban districts where no voter registration took place (nahiyas 11, 12, 14 and 15), which include the prosperous, largely-gated community of Aino Mina, Kandahar’s most densely-populated neighbourhood of Loya Wala are located and an insecure area bordering on Dand district. (3) Residents of Kandahar city, however, said that voters from these two areas would have easily been able to register in other neighbourhoods close by. [Update 27 October: According to the IEC there is voting in all nahiyasof Kandahar. They said the centres in question may have been listed under other areas.]
Normally, an election would most probably be won or lost in the areas where most votes are cast (in this case likely to be Spin Boldak and areas of Kandahar city). However, in an election like this, with many candidates and relatively small margins, seemingly insignificant areas can swing a vote, particularly if the votes are concentrated on a certain candidate. This is further compounded by the fact that the final results have so far been largely determined afterelection day, when the IEC had to decide which votes could be counted and which must be invalidated. For this reason, candidates and supporters are often incentivised to try to manipulate the vote wherever they can, in the hope that some of it will slip through.
What was the vote like in Kandahar in previous elections?
In the 2009 presidential election, Kandahar became famous for its “industrial scale” ballot stuffing (for an AAN report that raised the issue of mass fraud days after the election, see here). The fraud was particularly concentrated in the border provinces, where the late Abdul Razeq was then Border Police Chief at the time. Out of the around 250,000 votes that were cast that year, around 180,000 were disqualified in the ECC-led sampled audit, leaving only a little over 70,000 votes to be counted (see details here).
In the 2010 parliamentary elections, every single polling station in the province had reported open on election day, despite serious security challenges. 30 per cent of all polling stations were disqualified in full. After the IEC and ECC disqualifications, there were around 75,000 votes left (see here for more details).
The 2014 presidential election saw 270,000 votes counted in the preliminary count. In the various audits (the results of which were never made public) the province appears to have lost around 50,000 votes (see here). In the 2014 provincial council vote, which took place at the same time and was audited by the IECC rather than the IEC, Kandahar lost around 11 per cent of its votes, including all the votes of its top candidate. (4)
Will the new anti-fraud measures work?
In the past, when voters could cast their vote wherever they wanted, Kandahar provided ample opportunity for unmonitored and/or unopposed ballot stuffing that could be explained away with claims of suddenly improved security or surprise high turnouts. This was particularly the case in the province’s remote districts and in the districts that were under the firm control of local strongmen.
With the new registration system, voters have to cast their vote in the centre where they registered. The last-minute addition of biometric verification aims to weed out multiple registrations, multiple votes and the use of fake documents. If implemented properly, it has now become much more difficult to engage in multiple voting, mass proxy voting, ballot stuffing and manipulation of the count and data entry. But only if procedures are actually followed. The scope for (mass) fraud, and the difficulties the IEC will have to deal with this, will therefore depend on how messy the election becomes.
So far, the IEC in its public statements has downplayed the messiness, saying that only in a minority of the polling centres procedures were not followed and that the main problem had been teachers and other staff turning up late. The IEC also said it hope the vote in Kandahar would be an improvement, compared to last week’s two election days (see here and here and here for observation reports). Provincial IEC head, Wardag, told AAN that atechnical team, dispatched from IEC headquarters, trained the district electoral officers (DEOs) and polling centre managers, who in turn trained the polling station staff. The focus was mainly on the use of the new biometric machines.
The IEC initially made it clear it would invalidate all votes that were not cast using both the voter lists and biometric verification, but it has already been wavering. From what happened on 20 and 21 October 2018, it appears that there will be considerable leeway to get votes counted that were not without properly cast. (5)
This could provide an incentive for all forms of fraud, including the very blatant, in the hope that the results will slip through the cracks. Things to watch will thus include implausible high turnouts in both secure and insecure areas, significant numbers of voters turning up without having registered, and the mass malfunction or loss of biometric devices and voter lists.
Edited by Sari Kouvo and Thomas Ruttig
(1) According to the IEC’s lists voter registration earlier this year took place in 172 polling centres. Wardag could not explain why the number of polling centres had now risen from 172 to 173. He thought it might simply be a repetition on the list somewhere.
(2) The English-language summary, which can be found here has slightly different figures for Kandahar’s registration totals: 567,608 registered all together, with 557,344 registered for the vote in the provincial constituency – 483,749 men and 73,595 women (a little over 13 per cent) and – and 10,262 as part of the nationwide kuchi constituency.
The total number of registered kuchi voters nationwide is 168,015. Kandahar, according to this list, comes in fifth and is dwarfed, by far, by Kabul’s largely settled kuchi constituency (71,506 registered kuchi voters, or 43 per cent of the total). The other provinces with significant Kuchi registration have numbers comparable to Kandahar: Kapisa (13,651 or 8 per cent of the total), Nangarhar (11,535 or 7 per cent of the total), Logar (10,298 or 6 per cent of the total).
(3) Voters registered per rural and urban district, as provided by the IEC:
Kandahar’s rural districts
A total of 132 polling centres in rural districts registered voters. Mianeshin, Shorabak and Reg were among 32 districts that could not be accessed due to security problems at the time of the polling centre assessment in the second half of 2017. These three districts, as well as Ghorak, ended up with just one polling centre. Maruf and Nesh have no polling centres at all, even though are not mentioned in the list of 11 districts countrywide which, according to the Afghan Ministry of Interior, are fully under Taleban control.
Kandahar’s urban districts (with a brief description of the nahiyas):
A total of 40 polling centres in urban districts registered voters. The districts that are not on the list are:
(4) In a letter to the IEC, the IECC said it decided to disqualify the candidate after an audit of over 600 polling stations in Kandahar, when they found that “all his votes had problems.” Such problems included ballots not being removed from their stubs, large numbers of ballots with similar tick marks, made with the same markers, which led them to the conclude that there had been systematic fraud.
(5) As reported here, the IEC will need to deal with three types of election results: a) votes from polling stations that used both the voter lists and the biometric verification, b) votes from polling stations that used only the voter lists, without capturing the biometric verification data, and c) votes from polling stations that used neither of the two systems and made new on-the-spot, handwritten voter lists. (see here for background).