Political Landscape

Afghanistan’s 2019 Election (24): Disputed recount, threats not to accept results, and some interesting new data


Tent for the audit and recount at the Kabul provincial office located inside the IEC headquarters compound. AAN saw only President Ghani’s State-Builder agents, and observers from FEFA when we visited around midday on 12 November, a day before it was suspended. The audit and recount has now been resumed, but most candidates continue to boycott it. Photo: Ali Yawar Adili, 12 November 2019

Tent for the audit and recount at the Kabul provincial office located inside the IEC headquarters compound. AAN saw only President Ghani’s State-Builder agents, and observers from FEFA when we visited around midday on 12 November, a day before it was suspended. The audit and recount has now been resumed, but most candidates continue to boycott it. Photo: Ali Yawar Adili, 12 November 2019

The Independent Election Commission (IEC) has overruled the demand by most of the presidential candidates to invalidate around 300,000 out of the 1.8 million votes (about 16 per cent) from the 28 September president election. Critics wanted these excluded before the IEC started auditing and recounting the votes from 8,255 ‘problematic’ polling stations. However, the IEC has gone ahead with the audit and recount anyway, causing some candidates to threaten not to recognise the election results. The election process now faces a stalemate, says AAN’s Ali Yawar Adili. He ends this piece with interesting new statistics about the polling stations that were open and closed on election day, turnout and ‘de-duplication’, all broken down province-by-province, here published for the first time.

Audits and recounts: the what and the why

Audits and recounts can be triggered by a wide range of irregularities and discrepancies. According to the IEC’s audit and recount regulation (here in Dari), a recount is, as the name suggests, a recounting of the ballots the result of which is recorded on a new results form. An audit involves a broader range of actions, inspecting a wide range of documentation and processes from the polling stations, such as tamper evident bags (TEBs), results forms, voter lists, journals and other relevant documents and physical checks of ballot boxes. Generally, audits and recounts happen together. In the last 12 days, they have come to be highly controversial in Afghanistan. Exploring why and looking at this issue in depth, this dispatch has the following structure:

  • Problematic Polling Stations: a list of the reasons why 8,255 polling stations were deemed problematic and how many stations fell into each category, and then a table showing a provincial breakdown of where the different types of problems were found.
  • Audit and recount launched, suspended, resumed, ongoing: how the IEC audit and recount of the votes cast in the 8,255 problematic polling stations was launched, suspended and then, despite opposition from most of the presidential tickets, went ahead again.
  • Reactions and threats to not recognise the election result: a look at why most candidates and election watchdogs have opposed the recount and audit, both initially and when it was resumed.
  • Disputed votes: a closer look at the 300,000 disputed votes which most of the candidates, in particular Dr Abdullah’s team, would like to be declared invalid, but which the IEC has now ruled are valid. Critics say these must be excluded before any audit and recount of polling stations is carried out.
  • 102,012 votes cast outside polling hours.
  • 137,630 votes which the IEC initially quarantined due to ten types of discrepancy, for example, results being sent from multiple polling stations using one device, and missing biometric identities for devices that transmitted result data.
  • The votes from 2,423 polling stations (which are a subset of the 8,255 problematic polling stations) where there BVV devices or memory cards were missing and thus there was no biometric data from them.

Finally, the dispatch includes new data not yet published by the IEC which AAN has been given:

  • Provincial breakdown of the polling stations that were open and closed
  • Provincial breakdown voter turnout
  • Provincial ‘de-duplication’ – when duplicated votes, for example, where two ballots have matching fingerprints or tazkeras are reconciled.
  • Conclusion: what this means for what appears to be a very close-run election.

The IEC had initially announced that an audit and recount was required for 8,255 polling stations. IEC chair Hawa Alam Nuristani said in a press conference on 12 November that the IEC had reached this conclusion after comparing “all the data from the National Tally Centre with the Digital Tally Centre and then with the biometric data of voters which Dermalog [the provider of biometric technology] had dissected and analysed [tajziya wa tahlil].” As a result, she said, it had become clear to the IEC that these 8,255 stations had “problems” and the IEC had therefore referred them for an audit and recount. The aim, she said, was to be able “to discern valid from invalid votes” from these stations.

Nuristani said that this was in accordance with decisions 104 and 105 that the IEC had made about the audit and recount. Decision 104, made on 4 November, stated that a total of 8,494 polling stations had problems (here in Dari and unofficial English translation of it here). It then listed nine categories of problems including: a discrepancy of more than five votes between those that were backed up by Biometric Voter Verification (BVV) data and the relevant results forms (also called result sheets); missing results forms and; polling stations that were referred by designated teams (meaning intake teams at the National Tally Centre (NTC)), or polling stations that were known to be closed but from whom biometric information was received. (See the full list in Table 1 below, and footnote 1).

The IEC’s breakdown of the stations by the nine categories of problems did not add up to 8,494.

In a separate, second decision, number 105 from 7 November, the IEC ordered the audit and recount of votes from a further 2,423 polling stations whose biometric devices/memory cards were missing. With these 2,423 polling stations, there was a total of 8,494. The candidates critical of IEC’s actions are not unhappy with the votes from all of the 8,494 polling stations being recounted and audit, only those stations where the BVV devices or memory cards were missing. (See in Dari here and unofficial English translation of it here) (2)

Table 1: Problematic polling stations with a break-down of the types of problems

Total polling stations 29,586
Closed polling stations 3,006
Open polling stations 26,580
   
Total problematic polling stations to be audited and recounted 8,494
    Decision
Problems with missing BVV devices/memory cards 2,423  
·       National Tally Centre (NTC) and Digital Centre received results but no BVV devices, memory cards and biometric data 1,287 Audit and recount
·       Stations reported open but no result form at National Tally Centre and Digital Centre and no BVV device, memory cards or biometric data 1,136 Audit and recount
     
Polling stations with discrepancies 6,071  
·       Discrepancy of more than five votes between result forms and biometric data 3,893 Audit and recount
·       Result form is not at National Tally Centre but is at Digital Centre and there is also biometric data 543 Refer to the copy of the result form inside the ballot box; if the copy is not correct, conduct a recount
·       Result form is available at National Tally Centre, with biometric data available, but result forms are not at Digital Centre 93 76 A discrepancy of five or fewer votes between result forms and biometric data; to be processed
17 A discrepancy of more than five votes between result forms and biometric data; audit and recount
·       Zero votes on result forms from Digital Centre, despite votes registered on the original result forms at the National Tally Centre; BVV data available 1,110 533 A discrepancy of five or fewer votes between result forms and BVV data; to be processed
577 A discrepancy of more than five votes between result forms and biometric data; audit and recount
·       Zero votes on original result forms at National Tally Centre, despite votes in the digital centre and BVV data 23 Audit and recount
·       Referred by designated teams (ie National Tally Centre teams) to commissioners for decision 23 To be investigated and decided
·       Referred by designated teams for audit 24 Audit and, if necessary, recount
·       Referred by designated teams for recount or audit and recount 362 Recount or audit and recount

Source: Table by AAN using data from IEC decisions 104 and 105

It is important to note two additional figures contained in IEC decision number 104. First, it says that 11 (out of 3,006) polling stations which the IEC had reported to be closed had nonetheless uploaded biometric data. It ordered results from these 11 stations to be quarantined and investigated. Second, the Electoral Complaints Commission (EEC) in a letter dated 26 October asked the IEC to recount, or audit and recount 1,709 polling stations. These included 117 stations which had been reported closed. Of the remaining 1,592 stations, it said the IEC would take action as per ECC decisions, regardless of whether or not these stations were included in the IEC list of problematic 8,255 polling stations.

The polling stations ordered by the IEC or ECC or both to be audited and/or recounted are illustrated in the table below, broken down by province and with reasons given:

Table 2: Provincial Breakdown of Problematic Polling Stations, showing the geographic distribution of problems (download the table here).

Source: Election stakeholders

 Audit and recount launched, suspended, resumed, ongoing

The IEC initially started auditing and recounting on 9 November but was forced to suspend this action, four days later, following boycotts by most of the presidential candidates, including Chief Executive Abdullah. In its decision 106 (here in Dari) which it issued on 13 November, the IEC said that, “In order to address the objections and concerns of some political parties, candidates and civil society organisations,” it had decided that the audit and recount “be halted until further instruction.” However, it also said the audit and recount of stations whose ballot boxes were open and already under audit and recount would be exempted from this ruling. (The decision was signed by all seven commissioners except Mawlana Muhammad Abdullah).

Prior to the suspension, IEC chair Nuristani told a press conference on 12 November that audits and recounts had begun in 20 provinces and that the work had already been completed in Badghis province. Nuristani said that, according to reports from IEC provincial offices, in the remaining 14 provinces, election campaigns had protested against the audit and recount – and blocked it.

The IEC’s suspension of the audit and recount was short-lived, however. The IEC resumed the audit and recount on 17 November, as relayed in its decision 110, dated 16 November. The IEC also said it had held several rounds of meetings with representatives of electoral campaigns and political parties where it had “explained and discussed disputed technical issues.”

Before the resumption of the audit and recount and perhaps in a bid to meet at least some of the candidates’ demands, the IEC in its decision 107 dated 14 November (see here in Dari and unofficial English translation here) tasked the IEC secretariat with preparing a list of all election employees who had committed malpractice and violations during the election by 19 November and present it to the ECC. The IEC listed four categories of those employees and barred them from participating in the audit and recount. (See the list of the four categories in footnote 3).

Reactions and threats not to recognise the election result

There were strong political reactions to the resumption of the audit and recount, as there had been when it was first announced.

  1. A) Reaction to the initial launch of the audit and recount

When the IEC first started the recount and audit, Abdullah’s team issued a statement declaring a boycott. They accused the IEC of “breaching the law” by starting the recount “before separating and cleaning invalid and non-biometric votes and without specifying the number of cleaned votes per station.” They also criticised the IEC for including the “2,432 [probably a typo meaning 2,423]” stations [ie those with missing BVV devices or memory cards] in the recount, claiming that “until now no-one was aware of [whether or not] they existed, and there is no biometric data from them.” The campaign team said that it had instructed all its provincial members and candidate agents to avoid participating in “this illegal process” until the issue had been clarified.

Similarly, the Council of Presidential Candidates, a coordination group of initially 13 presidential candidates, see AAN’s reporting here) issued a statement on 11 November announcing a boycott of “any fraudulent election results,” adding that they “will not recognise as legitimate any government formed out of organised fraud.” It called for the identification of “the perpetrators of systematic fraud” and for them to be dealt with by courts. The Council said, “If the current electoral crisis plagues the whole country, the heads of the National Unity Government, who are responsible for the mismanagement and electoral crisis, should resign so that there is a [more] conducive environment and a new government can hold fresh presidential elections as soon as possible.”

The election watchdog organisation, the Transparent Election Foundation of Afghanistan (TEFA) said in a statement released on 10 November that it considered the IEC’s recount decision as contravening article 19 of the electoral law (which concerns the use of biometric technology).TEFA echoed the demand not to validate any votes without biometric data nor those cast outside the designated time on election day. If this demand was not heeded, the statement concluded, this could “pave the way for electoral tickets and for the people of Afghanistan not to accept the election results.”

The deputy head of another watchdog, the Free and Fair Election Forum of Afghanistan (FEFA), Naim Asghari, echoed these concerns, warning that the commissioners’ decisions were not above the law and asking the IEC to reconsider its decision (media report here).

A group of seven other observer organisations including Election and Transparency Watch Organisation of Afghanistan (ETWA) and Free Watch Afghanistan (FWA) released a statement on 13 November, calling on the IEC to suspend the recount for two days to allow for a consultative meeting with the electoral tickets to discuss their objections and “whether or not the recount is right.” It also called on the IEC to publicise the list of employees who were responsible for “damaged or missing biometric devices and memory cards” and refer them to the judicial authorities.

AAN visited the recount stations at the IEC headquarters at around 12 pm on 12 November, a day before it was suspended and saw only agents from President Ashraf Ghani’s State-Builder team and FEFA observers. The author was told there were agents from Sayyed Nurullah Jalili and Faruq Nejrabi’s teams, but they had gone out for lunch at the time of the visit.

  1. B) Reaction to the resumption of the recount

Despite all these objections, after pausing the audit and recount for four days, the IEC resumed this task, prompting another wave of protests. The Council of Presidential Candidates issued a statement saying it had not been consulted about the recount and would not participate in it. It reiterated its “previous principled position” that “until clean and unclean votes are separated,” it would not approve any recount by the IEC.

Chief Executive Abdullah also told a press conference that his team and most of the other electoral tickets would not participate in the recount. “A collective boycott of the recount process by the electoral tickets means that the recount of votes and the outcome of the process in the absence of agents is not legitimate to us.” He called on the IEC to immediately revise its “illegal decision and stop the recount.” He insisted that the IEC should exclude from the count what he called “300,000 fraudulent, ghost and non-biometric votes” before starting the recount. A member of his campaign team claimed on 18 November that they had forced the closure of IEC offices in Badakhshan, Baghlan, Balkh, Faryab, Jawzjan, Kunduz, Samangan, Sar-e Pul and Takhar provinces (media report here). IEC commissioner Awrangzeb confirmed to the media that the recount had been blocked in at least five provinces, Sar-e Pul, Jawzjan, Faryab, Takhar and Panjshir (media report here).

In a press conference on 19 November, Gulbuddin Hekmatyar also called for exclusion of the 300,000 votes, saying that that no one had the right to announce the election results until that happens. He, however, called for a full recount that should be observed by agents of all electoral tickets and that fresh election should be held in those polling stations whose “biometric devices are missing and have been stolen by the ruling team” (media report here) – this refers to the 2,423 polling stations described above.

President Ashraf Ghani’s State-Builder team, alone, has welcomed the IEC’s decision about the recount, saying it would “ensure transparency of the election and the IEC has made the decision based on its authorities enshrined in the electoral law and procedures” (see the statement issued by its political committee on 18 November here)

On 19 November, the IEC announced that the audit and recount had been completed in 15 provinces (Ghor, Daikundi, Samangan, Farah, Kunduz, Herat, Logar, Badghis, Uruzgan, Bamyan, Laghman, Maidan Wardak, Balkh and Nimruz) and it was on-going and other provinces. The results of six provinces have been sent to IEC headquarters and it has started to take in the results of their audit and recount.

The disputed votes 

At the heart of this stand-off are several categories of disputed votes. Chief Executive Abdullah’s Stability and Integration ticket has been calling for the invalidation of up to 300,000 out of the 1,843,107 total turnout figures announced by the IEC so far. These include: 102,012 votes cast outside polling hours; 137,630 initially-quarantined votes and; between 50 and 70,000 votes with invalid photos details about our previous reporting here). (4)

  1. a) Votes cast outside polling hours (102,012)

On the question of votes cast outside official voting hours, Table 3, below, shows that these 102,012 votes have date stamps that were recorded as early as 28 January 2019, eight months before election day, and as late as 28 November 2019, two months after it. AAN put the table together from a document prepared by election stakeholders, called “Blacklist _ Elections Date Timeframe_7am-to_5pm” (which is based on data sent by Dermalog to the IEC). The document notes in a footnote that Dermalog had not taken responsibility for the removal of these votes because it felt this to be outside its mandate. IEC spokesman Ali Eftekhari has also said that IEC staff had configured the time stamps, not Dermalog.

Table 3: A provincial breakdown of the dates when votes from outside election hours were cast (download the table here). 

Table 3: A provincial breakdown of the dates when votes from outside election hours were cast

Source: Election stakeholders

Stakeholders also provided to AAN the following table, Table 4, which divides the votes with problems of when they were cast into 15 categories. It also gives the likely reasons for these votes being recorded outside regular polling hours, reasons which have not been viewed as satisfactory by most candidates: 

Table 4: Categories of votes cast outside voting hours

Classes Categories (explanation) Polling stations Voters before 7am Voters after 5pm Total

 

Unlikely to be caused by irregularities Category 1- Started polling at 7 am and polled past 5 pm on a continuous basis (without breaks outside regular hours)

 

5,912 0 28,819 28,819
Category 2- Started polling slightly before 6 am and did not poll past 5 pm (without breaks outside regular hours) 22 72 131 203
Category 3- Started polling slightly before 6 am and polled past 5 pm on a continuous basis (without breaks outside regular hours) 119 290 0 290
Total 6,053 362 28,950 29,312
Probable abnormal time on devices Category 4 – Wrong dates (before 26 and after 30 September 2019) 19 158 172 330
Category 5 – am/pm misconfiguration (devices started polling at 7 pm instead of 7 am) 516 29,749 10,199 39,948
Category 6 – am/pm misconfiguration (devices started polling at 6 pm instead of 6 am) 13 581 178 759
Category 7 – devices polling before 7 am on 28 but working around 10-11 hours of polling 259 10,472 0 10,472
Category 8 – devices recorded that polling started after 7 am of 28 September, operated for between 2 and 11 hours, with reasonable a gap between voters 43 0 2,394 2,394
Total 850 40,960 12,943 53,903
Possible irregularities Category 9 – started polling at the normal time and stopped polling on time and then restarted polling much later in the night 93 0 5,299 5,299
Category 10 – started polling much earlier in the morning than the scheduled start of polling. Stopped polling. Then started at the scheduled time when polling was to start. Then polled in a regular manner. Then stopped on time 87 3,211 0 3,211
Category 11 – started polling at much earlier than the scheduled start of polling, stopped polling, then started at the expected time of polling and proceeded with polling after time 82 1,562 3,877 5,439
Total 262 4,773 9,176 13,949
Inconclusive Category 12 – voting started after 5 pm and results were captured within 5 hours of final voter being processed 20 0 299 299
Category 13 – voting started later than 5 pm and results were transmitted afterwards in Kabul 9 0 611 611
Category 14 – devices that started polling before 7 am of 28 September, operated for less than 13 hours, and captured results less than 10 hours after the last voter 49 673 0 673
Category 15 – polling stations that could not automatically fall into any of the above categories but because of limitation of time could not be adjudicated on a case-by-case basis 111 766 2,499 3,265
Total 189 1,439 3,409 4,848
  Grand total 7,354 47,534 54,478 102,012

Source: Election stakeholders

Regarding the 102,012 votes with time stamp problems, the IEC in its decision 109 dated 15 November, ruled that there should be an audit and recount of 13,949 votes from 262 polling stations. It said the remaining 88,063 biometric data from 7,092 polling stations had no problems, in terms of the timing of the recording of the biometric data, and it thereby validated these votes. (5)

  1. b) Initially-quarantined votes (137,630)

Concerning the 137,630 votes that were initially quarantined, the document given to AAN by election stakeholders says this number represents the difference between all the processed votes that were uploaded and received from polling stations, and the number of votes with biometric data that was extracted from all BVV devices. There are multiple causes for a discrepancy; these are outlined in Table 5, below.

Table 5: Categories of the causes of differences between the processed votes and the biometric data extracted from BVV devices.  

Classes Categories of causes Polling stations Processed voters Biometrics Difference
More processed voters than biometrics Category 1 – results of multiple polling stations sent using one device 36 3,740 2,438 -1,302
Category 2 – 100% missing biometric identities for devices that transmitted result data 141 10,038 0 -10,038
Category 3 – partially missing biometric identities for devices that transmitted result data 159 14,572 3,163 -11,409
Category 4 – polling stations that have five or less discrepancies between the number of processed voters and the number of extracted biometric identities 859 88,298 86,869 -1,429
Category 5 – polling stations that have more than 5 discrepancies between the number processed votes and the number of extracted biometric identities 271 40,951 34,521 -6,430
Total 1,466 157,599 126,991 -30,608
More extracted biometrics than processed voters Category 6 – results for multiple polling stations sent using one device 23 785 1,480 695
Category 7 – stations where processed votes are 0 but there are biometrics 2,105 0 115,371 115,371
Category 8 – stations that submitted biometric identities but have no result data 94 0 8,320 8,320
Category 9 – stations that had biometrics collected using multiple devices 721 44,515 83,251 38,738
Category 10 – small differences in votes, fewer than 5 (probably due to duplication) 154 15,256 20,372 5,116
Total 3,097 60,556 228,794 168,238
Grand total 4,563 218,155 355,785 137,630

Source: Election stakeholders

The document notes that this report was sent to Dermalog, and that the response from Dermalog largely agreed with these explanations. Dermalog did add an explanation for category seven, which was that the device only transmitted the processed voter count when data entry of some kind was done. It sent a 0 when only a photo of the result sheet was captured.

These 137,630 initially-quarantined votes were validated by the IEC in its decision 108 dated 14 November. It ruled that the accompanying biometric records “do not have any specific problems.” (6) Most of the candidates, however, think these votes are invalid.

  1. c) Audit and recount of 2,423 polling stations

As mentioned above, IEC decision 105 ordered the audit and recount of the votes from an additional 2,423 polling stations where there were missing BVV devices or memory cards. It divided these polling stations into two categories:

  • 1,287 polling stations which were reported open and sent results sheets to the National Tally Centre or the Digital Data Centre and whose ballot boxes reached the provincial office warehouses, but whose biometric devices, memory cards and biometric data are missing.
  • 1,136 polling stations which were reported open but did not send any results form or photos of result forms, or memory cards of the biometric devices or biometric data.

IEC spokesman Ali Eftekhari told the BBC on 9 November that the IEC needed to check these 2,423 stations as “most of the devices might be inside the ballot boxes” which were in the provincial offices. He said,  however, that recounting these votes did not mean the IEC necessarily considered them “clean votes.”

However, most of the presidential candidates, especially Chief Executive Abdullah, believe the votes in these ballot boxes should not be counted at all. Abdullah himself said (video here) in an address to his supporters at the Loya Jirga tent in Kabul on 10 November that there was “no justified reason” for the votes from these stations being counted because, for the last month, the ballot boxes had been “in the control of those who organised widespread fraud and anything could be found inside them.” He suggested that, because there was no biometric data from these stations, “it is possible, they are part of the chain of fraud.”

IEC chair Nuristani, however, stressed in a press conference on 12 November that “only votes which have a biometric basis and are correct and according to regulations and procedures are valid.” Nuristani referred to the “strong concern” that a number of electoral campaigns had expressed about the validation of votes without biometric data. “We understand their concern,” she said “and ask them to pay deep attention to the IEC’s decisions, especially regarding the 2,423 stations which have been referred for a special audit and recount.” She said that the IEC would decide the validation or invalidation of these votes after it had received the audit and recount reports. A letter sent by Nuristani to Abdullah’s team on 18 November (a copy of which was published by Etilaat Roz) insisted that the IEC had enshrined the use of the biometric validation of votes as a principle in all its regulations and procedures and would therefore “decide accordingly.” The letter said that the IEC’s decision to launch a further investigation of these polling stations did not mean their votes from those stations would be deemed valid.

Polling stations, voters and ‘de-duplication’ of data

In a second part of this dispatch, we publish a variety of data about the election. The IEC has not yet published disaggregated statistics about which polling stations were open on election day and which were not, nor data on voter turnout by province, nor a provincial breakdown of the votes that have been removed through ‘de-duplication’ – this is the process of removing duplicate and irregular votes. However, AAN has received much of this data from election stakeholders, which we publish below. It should be noted that, as with all IEC figures up to this point, this data remain subject to change, given the audits and recounts, and the outstanding objections raised by political campaigns about the process. However, given the dearth of official data from the IEC – a major problem with its work this year – it seems important to publish this data to give citizens and others a chance to analyse what might have happened on election day.

Polling stations

  • 29,586 polling stations planned to open
  • 26,714 polling stations reported open on election day
  • 24,177 polling stations with biometric data before de-duplication
  • 24,168 polling stations with biometric data after de-duplication, ie 81.7 per cent of polling stations that were planned to be open

Table 6: Provincial breakdown of polling stations which were planned to open, reported open, and which transmitted biometric data (download the table here).

Source:  Election stakeholders

Turnout

  • 9,665,777 registered voters
  • 1,929,333 votes cast biometrically-verified identities of voters (before de-duplication)
  • 1,843,107 votes with biometrically verified identities of voters (after de-duplication)
  • Total turnout (after de-duplication) is 19.1 per cent of registered voters

Table 7: Provincial breakdown of registered voters, and voters biometrically-verified before and after de-duplication (download the table here).

The polling station data in Table 6 and Table 7 shows how the polling centres initially reported as open hit or almost hit 100 per cent in certain highly insecure provinces, such as Helmand, Farah and Uruzgan and somewhat lower extent, Nimruz and Kunar. Also showing very strong opening figures were the relatively secure provinces of Panjshir (100 per cent open) and Daikundi (also 100 per cent). However, when the biometric data transmitted from the polling stations are taken into account, it seems far fewer polling centres in the insecure provinces actually opened – only 88 per cent of polling centres in Helmand and 90 per cent in Farah. These two provinces also show the highest number of polling stations being audited and recounted – Helmand (63 per cent) and Farah (39 per cent).

The data also shows that Daikundi and Bamyan, two predominantly Hazara provinces, as in earlier elections, had the highest turnout, respectively 58 and 49 per cent of registered voters with, at the other end of the spectrum, Kunduz and Baghlan (both 7 per cent). The rest range from 9 to 37 per cent. The data also reveals that even the provinces with major urban centres, which typically score high turnouts also had very low turnouts – only a just over fifth of registered voters turnout out in Kabul and Herat and ten per cent in Kandahar. This lower than usual turnout trends is almost consistent across the country. This might indicate worsened insecurity was a consistent factor across the country, and/or disillusionment with this election, or with elections in general.

Votes removed through de-duplication:

  • 1,929,333 were cast with the identities of voters biometrically identified (before de-duplication)
  • 47,527 biometric records removed due to facial match duplication
  • 5,822 biometric records removed due to duplication of fingerprints
  • 37,006 biometric records removed due to duplication of voter ID
  • A total of 86,225 biometric records removed
  • 1,843,107 were cast with the identities of voters biometrically identified (post-de-duplication)

Table 8: Provincial breakdown of votes discarded as a result of de-duplication (download the table here).

Source: Election stakeholders

Kabul had the highest de-duplicated votes (17,419) followed by Kandahar (9,115), Paktia (7,058), Khost (5,225) and Farah (5,925). This might be a simple correlation with the number of high number of polling stations in Kabul (4,575), Kandahar (1,251). However, this would not be the case for Farah (205), nor to some extent, Paktia (659) and Khost (760). There are other provinces with a higher number of polling stations and lower number of de-duplicated votes.

Conclusion

The IEC has ruled that the 137,630 initially-quarantined votes and the bulk of the 102,012 votes with questions about their time stamps do not have any problems. It said that those votes with questionable time stamps could be reviewed by the ECC before the announcement of the final results. However, these technical justifications provided by the IEC for accepting these votes have failed to satisfy the majority of the presidential candidates. Demands for the exclusion of these votes continue to be made strongly, along with threats to reject the final results if this is not done. Chief Executive Abdullah’s team is still calling for these two categories of votes to be immediately discarded, along with votes with problematic photos – adding up to approximately 300,000 votes in all.

Meanwhile, it is not known when we can expect the results from the audits and recounts of votes from the 2,423 polling stations which were reported open on election day, but did not send any results, biometric devices or biometric data or only sent the results but no biometric devices/memory cards and biometric data.

The figures for turnout and polling stations at which votes were cast with biometric data and which have since undergone de-duplication, ie votes that the IEC thinks are valid, show an average number of 76.26 ballots cast per station. If this average is the same for these 2,423 polling stations, potentially, there could be more than 184,000 ballots from these stations which the IEC will need to decide whether or not should be validated. The IEC has not yet ruled whether or not these ballots would be valid. However, in a closely-run election – which this has every appearance of being – every ballot is significant. A final decision about the votes from these stations, as well as the final count of the disputed 300,000 votes (which alone amounts to 16.27 per cent of the current total turnout figure) could determine whether there is a first round winner or a need for a runoff.

The IEC has now crashed two dates for the announcement of preliminary results (after missing its first date on 19 October, the IEC said it would announce the results on 14 November, but has again failed to do so). It has not yet set a new date. According to the initial electoral calendar, the IEC should have announced the final results on 7 November and was to hold any possible runoff on 23 November. That is now impossible, pushing any potential runoff further into the harsh winter months, or further on, into spring 2020. It raises the prospect of this electoral spectacle continuing for months to come.

Edited by Thomas Ruttig, Rachel Reid and Kate Clark

 

(1) IEC’s decision 104 ruled (from an unofficial English translation published on the IEC website here):

  1. Following actions shall be taken regarding the 8,494 PS [polling stations] where there are observations due to various reasons:
  • The 3,893 RFs [result forms] where there are discrepancy of more than 5 votes between BVV and RFs data (attached as annex no. 1 to this decision) shall be audited and recounted and the responsible persons shall be referred to ECC for trial according to the law.
  • About 543 PS whose RFs are there in the Digital Center but their original RFs are not there in the NTC (the list of which is attached as annex no. 2 to this decision), following actions shall be taken:

(1) The copy of RF in the box shall be referred to in the presence of monitors and observers, the same form shall be processed.

(2) In case the copy is not there in the box, is illegible or is overwritten against the procedure, they shall be recounted.

(3) In case the RF in the box is not stamped and signed, and there are signs of fraud in it, it shall be audited. Similarly, the voters’ biometric information of such PSs shall be sent to the relevant provincial offices.

  • About the 93 PSs whose RFs are not there in the Digital Center, but their original RFs are there in the NTC (the list of which is attached as annex no. 3 to this decision), following actions shall be taken:
  1. 76 RFs where there is a discrepancy of 5 votes or less between the RFs and biometric information of voters, shall be processed according to the procedure.
  2. 17 RFs where there is a discrepancy of more than 5 votes between the RFs and biometric information of voters, shall be audited and recounted according to the procedure.
  • About the 1,110 PSs whose votes are marked zero on the Digital Center’s RF, but they have vote(s) in the original RFs and BVV information (the list of which is attached as annex no. 4 to this decision), following action shall be taken:
  1. RFs of 533 PSs where there is a discrepancy of 5 votes or less between the RFs and BVV voters, shall be processed.
  2. RFs of 577 PSs where there is a discrepancy of more than 5 votes between the RFs and biometric voter information, shall be audited and recounted according to the procedure.
  • The 23 PSs whose votes are marked zero on the original RFs, but they have vote(s) in the Digital RFs and BVV information (the list of which is attached as annex no. 5 to this decision) shall be audited and recounted according to the procedure.
  • The 23 PSs which were referred by the designated teams to the Commission’s meeting for a decision (the list of which is attached as annex no. 6 to this decision), each shall be investigated and decided upon separately by the Commission based on available documents.
  • The 24 PSs which were referred by the designated teams for audit (the list of which is attached as annex no. 7 to this decision) shall be audited and if required, recounted.
  • The 362 PSs which were referred by the designated teams for recount and recount and audit (the list of which is attached as annex no. 8 to this decision) shall be recounted, audited and recounted.
  • The 11 PSs which were announced by the Commission as closed, but have biometric information (the list of which is attached as annex no. 9 to this decision) shall be quarantined and investigated, and necessary actions shall be taken according to the relevant regulations and procedures, based on investigation findings.
  1. Following action shall be taken about the 1,709 PSs that were sent by ECC via their letter no. 1399 dated October 26, 2019 for recount/audit/audit and recount to the IEC (the list of which is attached to this decision):
  2. Regarding ECC’s decision on 117 PSs which were reported as closed PSs, the issue shall be shared and discussed with ECC.
  3. Action shall be taken according to the ECC’s decision about the remaining 1,592 PSs which were referred by the ECC, irrespective of the fact if they are included or not included in the current decision of IEC.
  4. In case there are both IEC and ECC’s decisions about a particular PS at the same time, it shall be audited and recounted.

(2) In its decision 105, the IEC decided that (from the unofficial English translation published on IEC’s website here):

  • The 1287 polling stations that were open on the Election Day according to the confirmed reports of the PS Chairperson, PC Manager, DEO and DDEO, PEO, Secretariat of the Commission and security forces and the results of which are received at the NTC and DAVE Center, however, the BVV Devices, SD Cards and biometric data have not been received at the HQ and the ballot boxes are there at the provincial warehouse or only the SD Card is lacking, if there is no report or document showing violation and fraud, such ballot boxes shall be audited and recounted in presence of monitors, observers, media and the ECC representative and the ballot papers that have QR Codes attached to it shall be recorded on the specific recount form and a report along with the recount/audit forms shall be presented at the Commission’s meeting. Also, the issue of the biometric data not received, including the BVV devices and the SD Card shall be investigated by the Field Operations, the Services and IT departments and the report shall be presented to the Commission.
  • The 1136 polling stations, if they were open on the Election Day according to the confirmed reports of the PS Chairperson, PC Manager, DEO and DDEO, PEO, Secretariat of the Commission and security forces, however, the result form or its photo has not been received, the SD Card of the BVV device is lacking and biometric data has not been received at the HQ, if there is no report or document showing violation and fraud and if the ballot box is there at the provincial warehouse, the ballot box shall be referred to and actions shall be taken according to paragraph (1) of this, decision.

IEC commissioner Mawlana Abdullah has handwritten on the margin of the decision:

This decision is against paragraph two of article 19 of the amended electoral law; against sections eight and 11 of the regulation on audit, recount and invalidation; against section eight of the procedure for the invalidation of votes, and is a legitimisation of non-biometric votes from which will ensue judicial prosecution due to violation of the law.

(3) It listed four categories of such employees:

  • Those polling station chairpersons who were responsible for signing and stamping the result forms, correctly writing candidates’ votes on respective result forms without alteration but did not do their duties properly, as a result of which the relevant polling stations have been referred to audit and recount.
  • District electoral officer and deputy district electoral officer, polling centre manager, polling station chairperson and BVV operators in whose polling stations voting took place without biometric [data capturing], who have not captured biometric data properly, have not filled the digital result forms or filled them incorrectly or have not taken the photo of result forms.
  • District electoral officer and deputy district electoral officer, polling centre manager, polling station chairpersons and BVV operators who have lost biometric machine or memory card related to their polling station or the devices have not been received at the headquarters or the electoral materials have been taken from them with force and they have not reported it or there is discrepancy between the votes recorded on the result forms and data recorded on the biometric devices of their respective polling stations
  • Provincial electoral officers who have not reported malpractices, violations and electoral crimes within their jurisdiction or have not reported them on timely basis or have somehow been involved in malpractices and violations and other instances.

(4) Abdullah’s team issued a statement on 6 November listing  “four types of fraudulent votes (…) [that] are not acceptable under any circumstance.” They are:

  • 137,630 votes for “ambiguous and unknown statistics which are in the quarantine section of Dermalog’s system”;
  • 102,012 votes cast “outside the [designated opening] time” (i.e. before 7 am and after 5 pm of 28 September);
  • An undefined number of ballots with discrepancies with the photographic verification (such as “a photo of a photo [photo taken of a voter’s photo or tazkera]” or duplicate photos) which have not been cleansed yet;
  • Votes from polling stations whose BVV devices and memory cards which are more than 700 are missing and the biometric data of which is no longer available.

(5) IEC’s decision 109 said (AAN’s working translation):

Given the information, dissection and analysis [tajziya wa tahlil] presented by the secretariat about the 102,012 biometric data recorded outside the polling hours, and the IEC’s international advisors who had a crucial role in the dissection and analysis team, technically, the problem emanates from setting the biometric machine clock, the logical time sequence between data recording is not indicative of violation, fraud and irregularity, Dermalog Company has not taken responsibility for the time of recording biometric data, and considering that the Electoral Complaints Commission, if there are  complaints or not, can review the issue before the announcement of the final results and the IEC will extend the necessary cooperation in this regard, therefore, in order to safeguard the people’s votes and prevent further delay in announcing the preliminary results of the presidential election,

The Independent Election Commission decided that:

Considering the information, dissection and analysis by the secretariat which are attached to this decision, in 262 polling stations where 13,949 biometric data have been recorded outside polling hours and dissection and analysis show potential irregularities, [they] should be audited and recounted and if fraud and violations therein are proven, they will be invalidated.

In the remaining 7,092 polling stations which are in different provinces of the country, the biometric data reported outside the voting time, considering the above-mentioned instances and their inclusion into the whitelist of Dermalog Company, the existence of technical problems in setting the clock of biometric machines as a result of which the effect of the biometric device clocks on validation or invalidation of votes is negated, the existence of the logical time sequence of recording biometric data and ultimately not accepting the responsibility for proper setting of time by Dermalog company, [then] if there is not any other deficiency or constraint, in terms of timing of recording the biometric data, their votes are without problems.

(6) IEC’s decision 108 said (AAN’s working translation):

Considering the information, dissection and analysis presented by the IEC secretariat which are attached to this decision, the 137,630 biometric records which are different from the processed voters due to the technical problems which had been created by Dermalog Company, these data [the 137,630 votes] do not have any specific problems.

 

 

Tagged with: , , , , ,
Thematic Category: Political Landscape