Afghanistan Analysts Network – English

Political Landscape

Afghanistan’s 2019 Election (9): Presidential poll primer

Ali Yawar Adili Jelena Bjelica Thomas Ruttig 15 min

After two delays and the cancelation of a peace deal which might have scuppered the poll altogether, Afghanistan’s presidential election is finally to happen, on Saturday 28 September. It will be the country’s fourth presidential election and seventh election in total since 2001. In this primer, AAN’s Ali Yawar Adili, Jelena Bjelica and Thomas Ruttig (with input from Obaid Ali) list some basic facts about: the number of registered voters, polling centres and stations, the candidates and observers and the security provisions for election day. The authors have had to struggle to gather the baseline data due to contradictory official figures, leading them to conclude that, despite 18 years of efforts to bring Afghanistan closer to democracy, Saturday’s elections are characterised, yet again, by murky data and limited transparency. 

The candidates

What started two months ago as an 18-candidate presidential race (for their biographies, see previous AAN report) has now turned into a final line-up of 15 candidates. The major drop-out is Hanif Atmar, National Security Adviser to President Ashraf Ghani until 2018, and former Minister of Rural Rehabilitation and Development, Education and Interior under President Karzai. Atmar declared his election campaign “halted and suspended” on 8 August after the collapse of his team (see previous AAN report). Since then, two more candidates – Zalmai Rasul, former foreign minister and National Security Advisor under Karzai, and Shaida Abdali, who served as deputy head of the National Security Council under Karzai – withdrew their candidacy and joined president Ghani’s State-Builder team.

Atmar issued a statement on 24 September saying he will stay impartial and not support any specific candidate. His second running-mate, Muhammad Mohaqeq, and both Abdali’s running-mates, Abdul Basir Salangi and Zulfeqar Omid, have joined incumbent Chief Executive Abdullah Abdullah’s Stability and Integration ticket (see here).

On 22 September, Ghani’s campaign manager Muhammad Omar Daudzai said that the running mates of candidate Enayat Hafiz, Jannat Fahim Chakari and Abdul Jamil Shiranai, had announced their support for Ghani’s team and had immediately been appointed as deputies to Daudzai. Hafiz said he would continue to compete with new running-mates. (1)

According to the Independent Election Commission (IEC), the candidates and running-mates who have withdrawn from the race or joined other teams have not followed proper procedure. On 22 September, the IEC said it had only learned through the media that a number of candidates had withdrawn and asked them to notify the IEC in writing by filling the withdrawal form. It also warned that if they did not do so, it would refer them to the Electoral Complaints Commission (ECC).

Shifts, defections and withdrawals may continue until the last minute before polling starts on 28 September. All the candidates who have withdrawn will still appear on the ballots, as these have already been printed, and as long as the IEC has received no formal notifications of withdrawal, it can not even notify voters that these candidates are no longer in the running. As many voters will have missed the withdrawal announcements, these candidates may still receive votes which, according to IEC spokesman Zabihullah Sadat, will be counted as invalid.

Regardless of what is the final number of candidates competing in Saturday’s first round of the election, a successful candidate will need to win 50 per cent of the votes plus one to claim immediate victory. If no candidate wins this majority, a second-round will be held. It is currently scheduled to take place on 23 November 2019. According to the electoral calendar, the release of the preliminary results is planned for 19 October. Election observers have, however, told AAN they doubt this date will be kept. In past elections, the IEC has also been unable to announce the results on, or even close to, the scheduled date.

Polling on Saturday will start at 7 in the morning and is scheduled to end at 3 in the afternoon (with a possible extension of two hours, at the discretion of the IEC). The shorter than usual polling hours are intended to allow polling officials to finish the count before it is too dark. During lunch and prayer times, the polling staff should perform their duties in turns, with the polling station chairperson carrying out the duties of the pausing staff member or assigning other staff from the same polling station to them.

The ballot papers for the presidential poll are blue and have been prepared in pads of 50 and 20 with numbered stubs. Each polling station will receive a number of ballot papers equal to the number of voters registered and listed for that polling station, plus an additional five per cent as a contingency. These are supposed to be used when ballot papers are spoiled by a voter (see IEC procedure here).

Number of voters

A little less than 9.7 million Afghans have registered to vote (the exact number released by the IEC on 18 August 2019 was 9,665,745). The number is based on the new voter registration that was conducted in the run-up to the 2018 parliamentary election and the top-up registration conducted in June 2019. (2) Of this number, the IEC said, 6,331,515 are men and 3,334,230 are women, ie the male/female voters’ ratio is 65 to 35 per cent.

These numbers, however, do not add up. In the run-up to the 2018 parliamentary election, according to data shown on the IEC website, the commission registered 9,072,208 voters (see footnote 2). In the top-up process carried out in June 2019, it registered an additional 329,312 people countrywide, along with 244,409 people in Ghazni province where registration and voting in the parliamentary elections had not taken place. The sum of these three numbers is around 19,000 people short of the number given by the IEC in August 2019. It also does not match the later, lower figure given by the IEC for the 2018 registration (see footnote 2). These inconsistencies cast doubt on the reliability of the registration data, the strength of the verification process and the quality of the voter lists that will be used in polling centres on Saturday.

Number of polling centres and stations

The total number of polling centres in the country in a normal situation would be 7,366 (see this AAN reporting in 2018), but given insecurity and other logistical constraints, the IEC prepared its operational plans for a total of 5,373 polling centres. These 5,373 polling centres altogether accommodate 29,586 polling stations (18,467 for men and 11,119 for women), with each centre containing the minimum of two polling stations – one for female voters and one for male voters. This is an increase in number of polling stations, as compared to the 2018 elections, because the IEC reduced the number of voters per polling station from 600 to 400. This change was made to avoid repetition of the administrative chaos from last year (see AAN reporting here), and to ensure that all registered voters can cast their vote between 7am and 3pm.

When the IEC provided the list of 5,373 polling centres to the security agencies for a security review, they found that 3,360 polling centres were secure, 907 were under a low-security threat, 675 under a high-security threat and 431 were reported to have “collapsed.” In the last days of August in a press statement, acting minister of interior Massud Andarabi announced that the police would not be able to secure these 431 polling centres and they would, therefore, remain closed. He recently told AAN that “it is possible that [even] more than the 431 polling centres will remain closed after the latest security reviews.” IEC told AAN that the security agencies will have a joint press conference on 25 September, when they will explain exactly how many polling centres will be scheduled to open.

Countrywide, the polling stations will be serviced by over 200,000 election staff that have been recruited, trained and paid by the IEC. (3) This is a novelty in comparison to previous elections when schoolteachers were obliged by law to play this role and were often not paid for this service. This obligation was recently removed, perhaps because of specific Taleban threats against education workers. (4) Each polling station will have six election workers: a chairperson, an identification assistant, a biometric operator, a ballot paper issuer and inking officer, a ballot box controller and a queue controller (see IEC procedure here). Additionally, each polling centre will have one overall manager and a queue controller for the whole centre.

The hiring of a designated BBV (Biometric Voter Verification) operator for each polling station is an innovation compared to the 2018 election, when the queue controller was made the biometric operator, and the chairperson of the polling station was supposed to control the queues. This led to a lack of crowd control and great disorder at many polling sites.

Number of observers

The IEC has issued 115 observation accreditation letters to, respectively, 18 candidates, 29 political parties, 26 national observer organisations, 24 media agencies and 18 international organisations. The request of 75 other organisations have been approved, but not yet processed. In total, over 118,000 individual agents and observers have been recorded in the system so far.

The political parties have registered around 20,500 observers, civil society organisations around 13,800 and the media around 7,000 observers. There will also be 174 international observers.

Additionally, the IEC has issued accreditation to more than 70,000 candidates’ agents. Of this number, by Sunday 22 September, the two main candidates had registered more or less equal numbers: 26,875 for Ghani and 26,649 for Abdullah. Hekmatyar’s camp had registered around 8,500 agents, Nabil’s around 4,400 and Ahmad Wali Massud’s around 2,300. Candidate agents are not considered impartial electoral observers because they are there to represent the interests of their respective candidates.

Of the non-election staff spread over the various polling sites, there are far more political agents than independent observers. The number of independent observers, even if evenly spread, would still only cover fewer than half of the polling stations, with an average of three observers per polling centre. As in previous elections, it is likely that polling sites in urban centres and safe areas will be fairly crowded in terms of observers and agents and that the more remote and insecure areas will be very thinly-monitored, if at all.

This year, we may also witness more citizen journalism in action, since Tolonews invited citizens to contribute to reporting on election day by sending pictures from their polling sites via Whatsapp and Viber (see here for Dari and here Pashto).

Vote’s security (biometric voter verification)

This year’s presidential election will be the second time Afghanistan goes to the polls using biometric voter verification. Each polling station has been assigned one biometric device comprising a printer, a power bank and a charger. Each polling centre (not station) also has one spare device in case of malfunctioning, the IEC has said. The use of biometric devices is intended to verify the identity of the voters and thus mitigate fraud. The devices have SIM cards that should provide an internet connection to transmit data to the centre. The devices also have a memory card that contains the information of all voters registered to that particular polling station.

The IEC had earlier said that 90 per cent of the devices would be connected to the internet and only 10 per cent that were used in remote areas would not be connected to the internet at the end of the polling day. On 24 September, however, deputy head of IEC secretariat for strategic communication said that the IEC contracted Etisalat company and that 1,200- 1,300 (24.19 per cent) out of 5,373 polling centres will not have internet coverage. According to IEC Chief Electoral Officer Dr Habibullah Nang, the information would be transmitted in digital format, including a photo of the result sheets, which he said would be “unchangeable and with no possibility of fraud.”

The IEC introduced the Biometric Voter Verification system in 2018 (5) and upgraded it for this election by pre-installing a digital copy of the voter list onto the devices and enabling the digital transmission of result sheets from the polling centres to the national tally centre in Kabul. Each action taken with the device – from switching on to switching off – is said to be registered and time-stamped.

On 25 August, when the IEC and representatives of the producer of the biometric devices, the German company Dermalog, briefed stakeholders on the system, the participants noted that the devices were unable to detect duplicate fingerprints at the polling stations. This meant that the verification and addressing of duplicates has been deferred to the central server. (6) Sources told AAN that the IEC Commissioners acknowledged they had not known about these issues prior to the meeting. Dermalog responded that the software had been configured as per the contract, which had not required such a function.

IEC Chair Hawa Alam Nuristani decided to put the movement of election materials on hold until the issue was resolved, which was announced to have happened on 29 August (see here). According to the IEC, the technical deficiencies in the devices have now been solved (7) and both national and international observer organisations had approved the use of the biometric system.

Yusuf Rashid, executive director of observer organisation FEFA, however, questioned whether the proper installing of new software on at least 39,000 devices could have been done so quickly. He also told Hasht-e Sobh daily on 31 August that the observer organisations had not been given the opportunity to assess the devices again. Hasht-e Sobh also quoted Muhammad Nateqi, a member of the political committee of parties, saying the improvements to the software had not been shared with them.

The voting procedure

According to the IEC procedures for polling and counting the votes, the biometric operator will switch on the biometric device at the beginning of the day, scan the polling station barcode and confirm that the voter list is from the same polling station, after which the list will be activated. (The procedures in English can be found here and Dari here.)

When the voters arrive, the biometric operator will enter their registration number (affixed to the back of his or her paper tazkera or on an empty page of the booklet-type tazkera) into the device after which the voter’s details (name, father’s name, grandfather’s name and serial number) will appear. The operator will then take a photo of the voter, his/her tazkera and the voter registration confirmation, and capture the fingerprint of both of his/her thumbs. If the voter does not have thumbs, the prints of the index, middle, ring and little fingers will be captured. If the voter has no hands, the thumbprint of the polling station chairperson should be captured and the voter allowed to vote. In this case, the details of the voter will be recorded by the chairperson in the polling station journal. This finished, the voter will be allowed to cast her/his ballot.

After voting has ended and the ballots have been counted, the biometric operator will take a photo of the result sheet and send the photo, with the digital results form that has been installed on the device, to the national server. If the device is connected to the internet, the data will be automatically transmitted. In areas without internet coverage, the devices will transmit once they are connected to the internet in the provincial IEC office and this should happen in the presence of agents and observers.

In the 2018 election, many devices malfunctioned or stopped working altogether and there were no spare devices at the stations. As AAN reported, the devices which had been presented as a panacea against fraud thus turned into a widespread problem. The IEC had to scramble to update its procedures and finally decided to allow voters to vote without their biometric data being captured. (8)

The IEC has, for now, said it will invalidate votes that are cast without the proper capturing of the BVV data (registration number, fingerprints, photo). It might, however, be forced to reconsider its procedure, if widespread malfunctioning or inconsistent use of the devices would again lead to many voters being disenfranchised, as was the case in 2018. The IEC decision to also take photos of female voters has given rise to concerns that women in certain areas may choose to abstain from voting. The Afghan Women’s Network has, for instance, recommended that taking photos of female voters should be optional in certain areas or that only the photos of their eyes be taken.

Physical security

The spokesman for the Ministry of Interior, Nasrat Rahimi, said that 72,000 security forces, including army soldiers and intelligence units, would be deployed to secure the poll. He added that an additional 20,000 to 30,000 strong-force would be kept in reserve. While the police will secure the immediate surroundings of the polling centres, the army will provide security in the wider parameter. The army has also been tasked with servicing remote polling centres – transporting election material to them and then and taking ballots back to the tally centre in Kabul.

In a bid to reassure the people, Rahimi said that all sensitive and non-sensitive election material had already been transferred to all 34 provinces without any security problems. This was on15 September. However, the transfer of election material within the provinces is still on-going on the more vulnerable routes from the provincial centres to the districts. Abdul Aziz Ibrahimi, a spokesman for the IEC, said this had started on 13 September and would continue until 27 September, one day before the poll.

From the other side, the Taleban issued a general warning on 8 August that their leadership had instructed its fighters to “prevent” the election from happening. On 18 September, the Taleban’s Education and Higher Education Commission followed this up with a statement calling on “employees, teachers, instructors and principals of the schools… to not allow any campaign and election centres in their education centres.” The Taleban also told teachers and students not to work as polling site staff, including – AAN has heard – through ‘night letters’ in some neighbourhoods in Kabul.

The Taleban has so far conducted two major attacks, one on a heavily-guarded campaign office in Kabul on 28 July, and one on an election rally in Charikar, capital of Parwan province, on 17 September. In the attack on the Kabul office, they killed at least 20 people and injured 50 more, while the casualty figures for the Parwan suicide bombing were reported to be 26 killed and 42 wounded (media report here).

Both attacks targeted the campaign of incumbent President Ashraf Ghani. His first vice presidential candidate Amrullah Saleh had a narrow escape in the Kabul attack, while Ghani and Saleh were reportedly only 500 metres from the suicide bomber in Charikar. Other candidates have reported threats against them (see this AAN reporting).

The powerful 19 September car bomb that destroyed a hospital in Zabul province’s centre Qalat also contributed to an increased feeling of insecurity among voters. The relative calm in the immediate pre-election period may or may not be a sign of the Taleban preparing to mount major attacks on election day, or just before.

Apart from the Taleban, there is also the local Islamic State franchise, IS Khorasan Province (ISKP), and its terrorist structures that have predominantly targeted Shia Muslim citizens with suicide and bomb attacks in Kabul, Jalalabad, Herat and several other areas. It also constitutes a potential threat to voters, election workers and polling sites.

Some improvements, but still concerns

Although the IEC seems technically somewhat better prepared for this election than last year’s – or, for that matter, any of the earlier elections – the usual sloppiness in providing exact and updated data on voters, polling sites, observers etc cannot pass unnoticed (more on this in an upcoming AAN analysis).

There have clearly been improvements to the biometric voter verification process, which was one of the major sources of chaos in the 2018 parliamentary election. However, there are still serious concerns as to whether the devices will function properly, whether the designated staff will be able to operate them as they should, and whether the capture of biometric voter verification data will be consistent enough to properly function as an anti-fraud measure.

AAN is publishing a second pre-election dispatch on what to look out for on election day.

Edited by Martine van Bijlert


(1) Hafiz confirmed that his running-mates had defected to Ghani’s team, a move which he claimed had been mediated by Senate chairman Fazl Hadi Muslimyar, a Ghani ally, and a number of the president’s advisers. It is unclear whether he will be able to find new running-mates in the short time remaining. It is also not clear whether the IEC will approve them.

IEC’s regulation (see here) sets out that if one or two running-mates of a candidate withdraw after publication of the final candidate list, the candidate should introduce new running-mates within 72 hours. It also says that if fewer than 22 days remain to the election (which is the case with Hafiz’s running-mates), the IEC will take an appropriate decision.

(2) In August2018,the IEC said they had registered 9,072,208 voters: 5,783,037 men (63.7 per cent) and 3,114,942 women (34.3 per cent). AAN then reported “rising doubts about the integrity of Afghanistan’s voter lists” which needed to be addressed through vetting and targeted audits. In early October 2018, the IEC published – on Facebook and in Dari and Pashto only – another, now only approximate figure of 8.8 million registered voters. This did not include voters from Ghazni province, where the parliamentary elections, and the registration drive, had been cancelled due to insecurity (see AAN’s reporting). The IEC said that, nationwide, over 600,000 registrations had been cancelled after they were either found to be underage, had dates of birth missing, or were duplicate entries. The IEC did not explain to the public how it had arrived at the latest figures or how it had verified the registration data (AAN’s reporting here). There were no further updates and no exact numbers given before the parliamentary election.

(3) Some election observer groups that had proposed a delay in the election, cited a lack of electoral staff and a slowness in training the staff as one of the reasons the elections could or should not go ahead as planned. AAN was told by an international observer working closely with the IEC that by 22 September, almost 95 per cent of the electoral staff had been recruited and trained.

(4) For the 2018 parliamentary election, the IEC had employed civil servants as temporary electoral staff, in accordance with article 23 of the 2016 electoral law. The article said that “temporary electoral workers at the polling centres and stations shall be assigned from amongst school teachers, lecturers of government institutions for higher education and government administrative personnel, in accordance with procedures adopted by the Commission.” The provision to make use of civil servants as polling day employees was intended to reduce the costs of elections (see AAN’s reporting here). This article was removed when the electoral law was amended after the parliamentary election. The amended law (articles 23.1 and 33.1) now says that the procedure for recruiting permanent staff to both the IEC and ECC should be jointly organised by the IEC and Civil Service Commission and that recruitment and should be conducted through open competition (see AAN’s reporting here).

(5) Afghanistan introduced the Biometric Voter Verification (BVV) technology ahead of the 2018 parliamentary election (AAN’s reporting here). Since then, the electoral law has been amended and now requires the IEC to use BVV technology in all phases of the election (AAN’s reporting here). In May 2019, the IEC decided it would be impossible to implement this provision of the law and that it could not use BVV in the voter registration (AAN’s reporting here).

(6) On 28 August, Nur Rahman Akhlaqi of Jamiat-e Islami claimed in a Facebook post that as a result of such loopholes, one individual could cast 400 duplicate ballots in a single ballot box. He listed the following problems for the software:

  • It fails to identify duplicate fingerprints at the polling stations and defers it to the central server;
  • It fails to identify duplicate photos and defers it to the central server;
  • It is possible to manipulate votes at the central server;
  • Stakeholders cannot monitor the central server and do not have technical permission to access the information in the central server;
  • People who are not on the voter list and have not previously registered can still vote by using the ‘Skip’ option;
  • Printing three QR codes for each voter paves the way for a person to use two additional codes.

Akhlaqi recommended  that only one QR code should be printed; that the list of people who should not be on the voter list should be shared with the candidates; and that the list of security forces and IEC staff who will vote at polling stations where they have not registered should be announced in advance and shared with parties and candidates, so they can be compared with the votes that were cast with ‘Skip’ option.

(7) The IEC also said that as an additional measure, the QR code would no longer be readable by any device other than the BVV devices distributed by IEC. There had been serious concerns during the 2018 parliamentary election that the sticker with the – albeit encrypted – voter data that was affixed to each ballot paper would compromise the secrecy of the vote, as the QR image could be read with unauthorised software.

(8) The IEC has now issued every polling centre with a spare BVV device. If the devices in more than one polling station stop functioning, the polling station chairperson or polling centre manager should take the following steps: stop polling until the device is repaired or a new one is available; try to troubleshoot; provide a device from another station if the polling in those stations has finished; request a spare device from the nearest polling centre and; notify the provincial office of the issue and request a new one.

When a device is used in another polling station, the chairperson should insert the memory card of the previous device in the new one as shown during the training in order to be able to avoid repeated voting. The polling station chairperson should record this in the station’s journal and ask the monitors and observers to confirm the veracity of the incident by signing the journal.

The polling station chairperson and the designated biometric operator are responsible for the proper use and protection of the biometric device and are accountable for any intentional damage or not using of the biometric device.



election Election 2019 polling centers presidential candidate Presidential election


Ali Yawar Adili

More from this author