Afghanistan's 2019 Elections (29): A statistical overview of the preliminary results

Original link (please quote from the original source directly):

After four and half months, the final results of the presidential election are still yet to be announced after the Electoral Complaints Commission (ECC) ordered a partial audit of more than 200,000 suspicious votes and a recount of almost 600 polling stations where there were discrepancies or missing biometric data. At the point where we had hoped for a decision, the ECC has instead passed on incomplete information to the Independent Election Commission (IEC) to make the final determination. In this overview of the situation as it stands, AAN’s Ali Yawar Adili scrutinises the statistics of the election: for turnout, polling stations, complaints and appeals, invalidated votes, and voting patterns. He concludes that the 2019 presidential election has been marred by flawed biometric technology, limited transparency, an inability to resolve disputes and questions surrounding its inclusivity to such an extent that it threatens the prospect of the next president receiving a credible, strong mandate.

Statistics on the numbers of votes and polling stations included in the preliminary results

The IEC announced the preliminary results on 22 December, almost three months after the election (see AAN’s reporting here). According to these preliminary results, a total of 1,824,401 votes were cast for 14 candidates.

At the same time, the IEC published the following:

  • results by the polling station in each district and province (see here) and
  • invalidated and closed polling centres and stations (see here).

Initially, 18 candidates had registered to run (see AAN’s previous report here). Four candidates – Zalmai Rasul, Shaida Muhammad Abdali, Nur ul-Haq Ulumi and Haji Muhammad Ibrahim Alekozai – withdrew from the race, all of them in favour of President Ashraf Ghani. However, since they withdrew their candidacies after the deadline specified by the IEC on the electoral calendar, ballots had already been printed and their names remained on them. They may still have received some votes, but the IEC did not count them.

In the preliminary results, 90 per cent (1,644,868) of the votes went to the two top candidates: 923,868 votes for President Ashraf Ghani (50.6 per cent) and 720,990 for Abdullah Abdullah (39.5 per cent). The remaining ten per cent went to the 12 other candidates, with the first runner up, Gulbuddin Hekmatyar, receiving 70,242 votes (3.9 per cent). Nur Rahman Lewal got the lowest number of votes (865), followed by Muhammad Hanif Atmar with 1,565 votes. Atmar was originally seen as one of the top candidates, but his ticket fell apart before the election following a row with his main political allies about the political system (see here). It is not clear why the IEC still calculated his votes. A day after the announcement of the preliminary results, 23 December, the IEC said in a statement (see here in Dari) that Atmar’s name had been removed from the preliminary results after they were published. (1) However, his votes are still counted, for instance on the IEC’s map showing the provincial breakdown of votes. Using the data from this map, we get the provincial breakdown of preliminary results shown in Table 1.

Table 1: Provincial breakdown of preliminary results, with votes per candidate (see the higher resolution here)

Source: Table by AAN, using data from IEC’s map showing the provincial breakdown of votes in the preliminary results

As the table shows, President Ghani won the majority in 16 provinces, including in the largest constituency, that of Kabul, as well as in the south, east and southeast, which are mainly inhabited by Pashtuns, the ethnic group he belongs to: Nangrahar, Khost, Kandahar, Kunar, Paktia, Paktika, Helmand, Laghman, Logar, Nuristan, Nimruz, Zabul, Uruzgan, Farah and Nuristan. Chief Executive Abdullah secured the majority in 18 provinces in the north, northeast and west dominated mainly by Tajiks, Uzbeks and Hazaras, the traditional vote base of his camp: Herat, Balkh, Daikundi, Bamyan, Badakhshan, Takhar, Jawzjan, Samangan, Sar-e Pul, Parwan, Baghlan, Faryab, Panjshir, Maidan Wardak, Kunduz, Badghis, Ghazni and Kapisa.

A blue and green map, published by the IEC on the day preliminary results were announced, 22 December 2019, illustrated this ethno-regional voting behaviour. Later on the same day, the IEC changed the colour of the map to all-blue, perhaps in an attempt to conceal the ethno-regional lines along which people appear to have largely voted.

Map 1: Voting pattern in preliminary results of 2019 presidential election

Source: Roger Helms for AAN

Source: Roger Helms for AAN

While both President Ghani and Abdullah obtained the majority in each other’s sphere of influence, it is interesting to note that Hekmatyar was runner-up (before Abdullah) in three provinces of President Ghani’s vote base: Nangrahar, Khost and Laghman.

Map 2: A Close-Up of voting patterns based on preliminary results in Eastern Afghanistan (for legend see Map 1)

Map 2: A Close-Up of voting patterns based on preliminary results in Eastern Afghanistan (for legend see Map 1)

The votes of Wardak and Ghazni – two provinces ethnically more or less evenly split between the Hazaras and the Pashtuns (with some Tajik enclaves) – have gone to Abdullah. It seems that most votes cast in these provinces came from the Hazara-dominated districts of Behsud One and Two in Wardak as well as Jaghori, Malestan and Nawur of Ghazni and that Pashtuns turned out less in those provinces, as they live mainly in Taleban-controlled areas.

Before the announcement of the preliminary results, in decision number 119 (see here in Dari), the IEC provided the following details on the number of polling stations:

  • a total of 29,586 polling stations had been planned;
  • 3,006 polling stations remained closed on election day;
  • 2,299 polling stations were excluded from the count based on IEC and ECC decisions; and
  • 24,281 polling stations were included in the preliminary results.

The IEC’s figures mean that 10.2 per cent of the polling stations that the IEC planned to open were closed on election day and 8.6 per cent of the polling stations that reported as open were invalidated. In the end, 82.1 per cent of the polling stations the IEC had planned to open were included in the preliminary results. 

Table 2: Polling station statistics

  Status of polling stations Number of polling stations
1 Polling stations planned to open on E-day 29,586
2 Polling stations reported closed on E-day 3,006
3 Polling stations reported open on E-day 26,580
4 Polling stations invalidated 2,299
5 Polling stations including in the preliminary results 24,281

Source: AAN with data from IEC decision number 119 

What we know about the IEC’s validation/invalidation decisions

It is not immediately obvious how the IEC arrived at its preliminary result. Originally the IEC announced that 2,695,890 votes had been cast (less than 30 per cent of 9,665,777 registered voters). This figure was published by IEC commissioner Mawlana Muhammad Abdullah on his private Facebook page on 3 October (he told AAN that the update should be considered official IEC figures, since he shared them in his official capacity; see also AAN’s reporting here). His IEC colleagues seemed to disagree. Later, on 23 October, IEC commissioner Rahima Zarifi told a press conference that the data from all the biometric voter verification (BVV) devices had been transferred to the server and showed 1,932,673 biometric data which she said was “final.” But she said it still needed to go through a de-duplication process. That is, it needed to be cleansed of underage and duplicate votes (see AAN’s reporting here). After the de-duplication, IEC commissioner Awrangzeb announced on 2 November that out of 1,929,333 biometrically verified votes, 1,843,107 had been found to be valid and the rest had “been identified as breaching the criteria.” (An IEC official told AAN on 3 November that the initial figure after transfer of the biometric data to the server had been 1,929,333, not 1,932,673 as announced by commissioner Zarifi, and Zarifi had rushed to announce the figure as soon as it had been emailed from Dermalog – the German company providing the BVV devices. The official said it had been problematic to release ‘raw data’ to the public before it was analysed and finalised; see AAN’s reporting here.)

When the IEC announced the preliminary results on 22 December, it provided a total of 1,824,401 votes. Based on these figures, we can deduce that 18,706 votes (1.02 per cent) were discarded by the IEC from the figure of 1,843,107 that it had reached after de-duplication of 86,225 out of the 1,929,333 votes (although it becomes 1,843,108). At that point, 852,783 votes (31.63 per cent) had already been removed from the initial turnout figure of 2,695,890.

While the IEC had already said that their initial turnout figure was not final, it did not provide any justification or explanation as to how the turnout figure came down so drastically: from 2,695,890 to 1,929,333 votes. It has given no indication whether the votes that ‘disappeared’ had been fraudulent, irregular, not backed up by biometric data or misreported in the first place. 

Table 3: Turnout figures

  Date Source Turnout figures Number of votes % Remarks
1 3 October   Registered voters 9,665,777   There is no census and thus determining the country’s voting population is difficult
2 3 October Commissioner Abdullah Initial turnout announced by IEC 2,695,890 27.89 Of the registered voters
3 23 October Commissioner Zarifi BVV votes before de-duplication 1,932,673 19.99 Of the registered voters
4 2 November Commissioner Awranzeb BVV votes before de-duplication 1,929,333 19.96 Of the registered voters
5 2 November Commissioner Awrangzeb BVV votes after de-duplication 1,843,107 19.10  
5 22 December IEC Preliminary result votes 1,824,401 18.87 Of the registered voters

Table: AAN based on data from various IEC sources

Before the announcement of the preliminary results, the IEC made a number of decisions about which votes to count and which to discard. These decisions were made public only after the announcement of the results. The most important decisions related to

  • the 5,832 polling stations (the number has been changing but IEC’s final list settles on this) found to have one of nine problems (see Table 1 in AAN’s previous reporting here);
  • the 2,423 polling stations where either the biometric device or the memory card had been missing (see Table 1 in AAN’s previous reporting here);
  • 137,630 votes which were the result of a discrepancy between the processed voters and the biometric data; and
  • the 102,012 votes that had been cast outside official polling hours.

Audit and recount of 8,255 polling stations

In its decision 104, dated 4 November 2019, the IEC ordered a set of 5,832 polling stations to be recounted, citing nine problem categories (see AAN’s previous reporting here). In decision number 105, from 7 November 2019, the IEC ordered the audit and recount of votes from a further 2,423 polling stations whose biometric devices or memory cards were missing. The IEC published a list of these 8,255 polling stations on 28 November 2019. See footnote (2) for a provincial breakdown of the polling stations.

 A) Audit and recount of 5,832 problematic polling stations

The 5,832 polling stations referred for recount according to decision 104 were dealt with in decisions 114, 115, 117 and 118 in the following manner:

  • Of these, 4,408 polling stations were recognised as being unproblematic and should thus be processed, ie counted (decision number 114, see here in Dari).
  • According to the secretariat’s report sent to the commissioners, 509 polling stations were found to have been unproblematic and should thus be processed (decision 115, see here in Dari).
  • Out of 962 polling stations, 850 were ordered to be processed, ie counted – 832 had been identified as being without problem, and in 18 the difference between the biometric data and the votes from the result sheet had been more than six. The remaining 112 polling stations had been closed on election day and were to be included on the list of the closed polling stations (decision number 117; see here in Dari) See the table listing the various categories under decisions 114, 115 and 117 in footnote (3).
  • The ECC had earlier invalidated 18 polling centres. These were confirmed as invalid through the IEC’s decision number 118 (see here in Dari). The decision further stipulated that the invalidation would take precedent over any earlier IEC decisions.

Decisions number 114, 115, and 117 thus validated 5,767 polling stations and declared 112 stations as closed (section 3 of decision number 117), while decision 118 confirmed the ECC’s invalidation of 18 polling centres. It did not specify how many polling stations these 18 centres contained.

The decisions were not without internal disagreement. Commissioners Mawlana Muhammad Abdullah and Mosafer Quqandi did not sign decisions 114 and 115. There is no explanation for this in decision 114, but in the case of decision 115, commissioner Abdullah hand-wrote the reasons for disagreement. With regard to the first section of the decision, which was about 276 polling stations, he argued that since out of 35,015 votes only 23,239 were backed by biometric data, “this egregious discrepancy” should have made the polling stations “invalid [or] subject to recount.” Second, he claimed that the 33 polling stations in section two of the decision – which should have been recounted based on a previous decision but had not been recounted –  were invalid. Third, he claimed that the 105 polling stations in section three of paragraph one of the decision, which had been referred for recount, had not been received (this is unclear though) and were thus invalid. He argued that the decision was against the electoral law (articles 2.2, 81, 86 and 19.2), the tallying regulation (articles 9 and 10) and all articles of the audit and invalidation regulations.

Qoqandi and Mawlana Abdullah did sign decision 117, but Abdullah wrote next to his signature that he was against paragraph two of the decision (the 18 polling stations where the difference between biometric data and the result sheets was more than six votes).

B) Special audit and recount of 2,423 polling stations

The 2,423 polling stations that were referred for recount, according to decision 105, were dealt with in decisions 113 and 116 in the following manner:

  • The IEC invalidated the votes of 1,179 polling stations that had lacked biometric data (decision 113, see in Dari here).
  • The IEC invalidated 946 more polling stations (decision number 116, see here in Dari).

This means that 2,125 out of the 2,434 polling stations were invalidated (see the table listing the reasons for invalidation of these stations, according to decisions 113 and 116, in footnote 4).

The remaining 298 polling stations were validated. The IEC argued that the audit and recount reports provided by the electoral affairs department and provincial offices had shown that the ballots of these polling stations had biometric confirmation stickers, no complaints had been registered against them and the biometric devices of 82 of these stations had been checked and showed 1,746 sets of biometric data (the IEC did not explain how and where they had found the biometric devices which it had said in its decision 105 were missing).

Mawlana Muhammad Abdullah and Mosafer Quqandi, again, refused to sign the decision. Abdullah wrote by hand in the margin of the decision that he was against the approval, because it was “against paragraph three of article 19 of the amended electoral law [which regulates the use of biometric technology] and sections of 8 and 11 of article 9 of the regulation for recount and invalidation [which describes what kind of votes should be invalidated].”

On 5 February, the ECC announced (see the video here) its own ruling ordering all 298 out of 2,423 polling stations to be recounted again. If they meet the necessary recount criteria, have biometric confirmation stickers (QR codes), and meet the criteria of valid votes based on the procedures of the two electoral commissions, they are validated and whatever cannot meet the above necessary criteria are invalidated. 

How many votes were invalidated by IEC’s various decisions?

The IEC did not say how many votes were affected by these various decisions. However, following the announcement of the preliminary results, the IEC released three sets of data: (a) a map showing the provincial breakdown of the votes per candidate (details have been given above), (b) the number, names and number of votes of the invalidated and closed polling centres and stations, presented per province (see here) and (c) the results per polling station. The IEC did not provide a consolidated list of the invalidated votes or of the closed polling stations. (The list did not contain grand totals for the 34 provinces.)

AAN went through the list of invalidated and closed polling stations, province by province, and found 102,021 votes that had been invalidated by the IEC, according to these provincial lists. Table 4 shows the provincial breakdown of these invalidated votes, as well as the original recount and audit data. 

Table 4: Invalidated votes

  Province Number of audit and recount polling stations Invalidated votes Number of polling stations invalidated or closed
1 Kabul 1,132 5,230 196
2 Kapisa 189 508 68
3 Parwan 106 349 25
4 Wardak 193 759 289
5 Logar 152 487 59
6 Nangrahar 774 246 95
7 Laghman 60 6 76
8 Panjshir 9 322 6
9 Baghlan 712 1,621 583
10 Bamyan 17 335 6
11 Ghazni 346 3,158 175
12 Paktika 388 2,149 174
13 Paktia 715 44,116 507
14 Khost 327 2,379 75
15 Kunar 125 125 5
16 Nuristan 106 211 125
17 Badakhshan 286 694 138
18 Takhar 77 0 246
19 Kunduz 51 222 213
20 Samangan 28 362 97
21 Balkh 35 272 504
22 Sar-e Pul 43 0 63
23 Ghor 246 1,159 344
24 Daikundi 58 967 8
25 Uruzgan 24 180 18
26 Zabul 154 947 48
27 Kandahar 741 21,149 149
28 Jawzjan 13 0 26
29 Faryab 234 0 305
30 Helmand 590 9,281 85
31 Badghis 5 8 2
32 Herat 177 2,102 227
33 Farah 89 2,568 67
34 Nimruz 53 109 26
  Total 8,255 102,021 4,523

Source: Table by AAN using data from list of invalid and closed polling stations and list of audit and recount polling stations, published on the IEC’s website

Outstanding disputes on IEC decisions awaiting the outcome of ECC’s decisions 

A) How the IEC dealt with the discrepancy of 137,630 ‘suspicious’ votes

In addition to the irregular polling stations that the IEC needed to decide on, there were two other categories of irregular votes. First, there were the 137,630 votes which were the result of a discrepancy between the processed voters and the biometric data. They come from 4,563 polling stations spread across 34 provinces of the country (see Table 5 in this AAN dispatch). Many candidates, including Chief Executive Abdullah’s campaign, called this category of votes suspicious and said it had to be excluded from the count. However, the IEC, in its decision 108, decided that the difference between the processed voters (ie 218,155 processed voters) and those that were backed up by Biometric Voter Verification (BVV) data (ie, 355,785 BVV data) was “due to technical problems which had been created by the Dermalog Company” and that there were “no specific problems” that warranted their exclusion from the count. The 137,630 votes were thus included in the count.

Table 5: 137,630 ‘suspicious’ votes

Processed voters BVV data Difference Number of polling stations Number of provinces IEC’s decision
218,155 355,785 137,630 4,563 34 IEC’s decision 108 dated 14 November 2019 validated all of all these votes

Source: Table by AAN with data from IEC’s presentation

Many candidates, especially Chief Executive Abdullah’s campaign, have refused to accept the IEC’s decision to validate these votes. They have lodged complaints and appeals. The matter awaits a decision by the ECC.

How the ECC dealt with it

As illustrated above, the 137,630 suspicious votes come from 4,563 polling stations, which the ECC in its ruling (see here) said were divided into two categories (1,466 and 3,097). (AAN did not receive any explanation from the commission why that happened and a member of Dr Abdullah’s campaign, Nur Rahman Akhlaqi, said in a Facebook post that this division “did not have any logic”.) The second category (3,097 polling stations) contains 137,630 votes (the ECC has not explained what happened to the 1,466 stations), ten per cent of which (309 stations) will go for a special audit (checking biometric data, the journal and the paper result sheet). If at least 65 per cent of these 309 polling stations meet 65 per cent of the three criteria, all their 137,630 votes will be validated; if not, all these votes will be sent for another special audit, using the above criteria again. In this case, the votes from any polling station that meets the criteria would be valid, and those that do not meet the criteria would be invalidated. 

B) How the IEC dealt with the 102,012 votes that were cast outside polling hours

The second category of irregular votes the IEC needed to deal with was the 102,012 votes that had been cast outside official polling hours. As AAN extensively reported before (see Table 3 in AAN’s previous reporting here), these votes showed date stamps that ranged between 28 January and 28 November 2019 and were cast before 7 am (47,534) or after 4 pm (54,478) in 7,354 polling stations spread across all 34 provinces (see Table 4 in AAN’s previous reporting here).

Many candidates, especially Chief Executive Abdullah, have said that these votes should be invalidated. However, the IEC, in its decision 109, referred 13,949 votes from 262 polling stations for audit and recount (13.7 per cent of the votes that had been cast outside polling hours) and declared the remaining 86.3 per cent (88,063 votes from 7,092 stations) valid. The IEC has not provided details on which votes were deemed valid or how the decision was made. 

Table 6: Votes cast outside polling hours

Votes cast outside polling hours Votes cast before 7 am of E-day Votes cast after 4 pm of E-day Number of polling stations Number of provinces
102,012 47,534 54,478 7,354 34
IEC’s decision 109 dated 15 November 2019
Valid 88,063 votes from 7,092 stations
To be audited and recounted 13,949 votes from 262 polling stations

Source: By AAN with data from various IEC sources

It is not clear from the IEC’s audit and recount results whether the 13,949 votes from the 262 polling stations that were cast outside polling hours and sent for recount were finally validated or invalidated. Meanwhile, the candidates are not convinced by IEC’s decision 109 and continue to call for the invalidation of all votes cast outside polling hours.

Clearly both the 137,630 ‘suspicious’ votes (discrepancy between processed voters and biometric data) and the 88,063 votes that were cast outside polling hours that the IEC deemed valid have been included into the preliminary results. For this reason, many candidates, especially Chief Executive Abdullah, called the preliminary results as announced by the IEC “fraudulent” (see AAN’s reporting here). They took their case to the ECC by filing thousands of complaints.

How the ECC dealt with the 102,012 votes outside polling hours

The ECC’s statement showed the following adjudication (see it here) of the complaints about the 102,012 “out of time votes” which come from 7,354 polling stations: 15 per cent of these which come from 1,103 polling stations (the commission did not specify the number of votes) should go through a special audit (criteria: biometric data, journal and paper result sheet). If 65 per cent of the audited (1,103) polling stations meet the above criteria, ie have valid biometric data, journal and result sheet of the election day, all 102,012 votes and 7,354 stations will be validated; but if not, all of them would be invalidated. 

Other decisions by the ECC

There are three other categories of adjudications by the ECC which are described below: 

  • Appeals against the discrepancy of votes between the paper result sheet and biometric data: The IEC had validated the discrepancy of 1 to 5 votes at the polling stations as a matter of flexibility and according to its relevant procedure and the ECC decided to reject complaints/appeals against these decisions (thus considering them valid) and if the provincial ECCs have decided differently, they will be cancelled and only the decisions by those provincial ECCs which rejected this category of complaints are emphatically approved. This decision applies to appeals related to similar cases.
  • Appeals against the discrepancy of more than six votes between the paper result sheet and biometric data: They are referred for audit and recount and this decision applies to appeals about similar cases.
  • Furthermore, the ECC confirmed the Faryab provincial ECC’s decision to invalidate four polling stations: male station 012 of Dong-e Qala centre, male station 01 of Yaldoz Primary School, male station 02 of Yangi Zarghon Grand Mosque and station 017 of Dong-e Qala (the ECC did not provide the details as to how many votes these stations had contained but AAN’s count shows that 345 votes were affected which were mainly cast for Chief Executive Abdullah and President Ghani).

The ECC did not specify however which polling stations or votes were in categories of more than six and fewer than five discrepancies. However, AAN’s count of this type of polling station in IEC’s decisions 115 and 117 shows that 308 polling stations had a discrepancy of more than six votes and 40 other stations had a discrepancy of between one and six votes. (5) As a result, AAN’s counts show that a total of 1,720 polling stations are referred for audit or recount (see Table 7).

Table 7: ECC’s decisions about different categories of disputed votes

  Categories Total number of polling stations by category Number of polling stations audited or recounted ECC’s decisions
1 137,630 suspicious votes 4,563 309 Special audit
2 102,012 7,354 1,103 Special audit
3 Discrepancy of more than six votes 308 308 Recount
4 Discrepancy of fewer than five votes 40 40 Validated
Total   12,265 1,760  

Source: Table by AAN with data from various IEC and ECC decisions


By 13 January 2020, the provincial ECCs adjudicated 16,545 complaints filed by the four major campaigns (of President Ghani, Chief Executive Abdullah, Hekmatyar and Nabil) that were related to the preliminary results (see Table 6 of AAN’s previous reporting about the results here). During the extended appeals period, which ended on 20 January 2020, another 6,292 appeals were lodged with the ECC. The central ECC had 15 working days to adjudicate these appeals, and it presented the outcome during a press conference held on 5 February 2020. However, the ECC still did not present the exact numbers of validated or invalidated votes, but decisions about procedures used to adjudicate several categories of disputed votes. This presentation ended with the statement that the ECC’s work was “finalised,” adding that now the IEC needed to take the remaining decisions and announce the final result of the 28 September 2019 election. This would include the announcement of whether any of the contenders had crossed the 50 per cent threshold of votes or if a run-off would be necessary.

With this, the ECC continued a trend that shaped the entire election process: information and statistics have never been complete and clear. Figures that are presented are very difficult to use. Readers need to go through them polling station by polling station and district by district to find how many votes have been cast in each district and do calculations of the totals themselves. Moreover, as shown above, numbers often do not add up.

Some general trends, however, have emerged. The turnout figure of 2,695,890 that was given after the election day is now down to 1,824,401 voters in the preliminary results. This means the electoral commissions have so far ruled almost 30 per cent of the turnout figure invalid. With this, the turnout figure dropped from 28 to 19 per cent of the registered voters.

The remaining question is whether the ECC’s decisions for more special audits (of the 137,630 ‘suspicious’ votes and the 102,012 ‘out of time’ votes) as well as another recount (of 298 polling stations) and another audit plus recount (of more than 300 polling stations) will lead to the invalidation of any of these categories of votes. It is important to watch how these audits and recounts will be carried out as well as how the IEC and the ECC will coordinate with each other and with the electoral campaigns. Finally, given Ghani’s narrow margin over the 50 per cent threshold after the preliminary results, any single ruling could potentially change the outcome of the election, and decide whether there will be a run-off or not.

On the question that is on the minds of everyone observing this election – When will we know? – one ECC member, replying to the query of a journalist at the 5 February press conference, ventured to predict that the final result might be out in “another week.” That sounds optimistic.

Edited by Martine van Bijlert and Thomas Ruttig


(1) The IEC’s statement said that Atmar had withdrawn from the race after the deadline enshrined in the law, as well as after the printing of the ballots. The statement added that, like other candidates who had withdrawn, Atmar’s votes were not counted and thus were not valid. If Atmar’s votes are subtracted from the preliminary results as announced by the IEC, the total number of valid votes cast for 13 candidates comes down from 1,824,401 to 1,822,836. This also leads to slight changes in the percentage of each candidate’s votes. President Ghani’s percentage will go up slightly, from 50.64 to 50.68 per cent and Abdullah’s percentage will also go up slightly, from 39.52 to 39.55 per cent.

(2) Table 8: Provincial breakdown of polling stations referred for audit and recount

  Province Number of polling stations Links
1 Kabul 1,132 (See here)
2 Kapisa 189 (See here)
3 Parwan 106 (See here)
4 Wardak 193 (See here)
5 Logar 152 (See here)
6 Nangrahar 774 (See here)
7 Laghman 60 (See here)
8 Panjshir 9 (See here)
9 Baghlan 712 (See here)
10 Bamyan 17 (See here)
11 Ghazni 346 (See here)
12 Paktika 388 (See here)
13 Paktia 715 (See here)
14 Khost 327 (See here)
15 Kunar 125 (See here)
16 Nuristan 106 (See here)
17 Badakhshan 286 (See here)
18 Akhar 77 (See here)
19 Kunduz 51 (See here)
20 Samangan 28 (See here)
21 Balkh 35 (See here)
22 Sar-e Pul 43 (See here)
23 Ghor 246 (See here)
24 Daikundi 58 (See here)
25 Uruzgan 24 (See here)
26 Zabul 154 (See here)
27 Kandahar 741 (See here)
28 Jawzjan 13 (See here)
29 Faryab 234 (See here)
30 Helmand 590 (See here)
31 Badghis 5 (See here)
32 Herat 177 (See here)
33 Farah 89 (See here)
34 Nimruz 53 (See here)

Source: Table by AAN with data from the list of audit and recount polling stations on the IEC website

(3) Table 9: Categories of polling stations validated by the IEC through its decision 114, 115 and 117

A) Validation of polling stations based on decision 114 after audit and recount
  Reasons for processing (ie counting) Number of polling stations
1 Identified as unproblematic 4,320
2 Referred for audit and recount; only audited and found to be unproblematic 18
3 Information inside the box (such as polling station number and QR code reader software information) do not match; but to be processed according to the secretariat’s report 5
4 Referred for audit and recount; according to the secretariat’s report, they have biometric data and correct votes 6
5 Only audited and did not need recount according to the secretariat’s report 8
6 Kabul provincial office report is not clear, but a copy has been confirmed by the Kabul provincial office 4
7 Copy of result sheet is available inside the boxes and has been validated according to the secretariat’s report 9
8 Scanned copy of the result form was received through email and the original result form or a copy of it was obtained from the box and kept in the respective polling centre in a tamper-evident bag 28
9 Repetitive result sheets but the result sheet is recorded in the EDSS system and confirmed by the provincial office; to be processed and the repetitive result sheet to be separately archived 10
Total 4,408
 B) Validation of votes based on decision 115
  Reasons for processing Number of polling stations
1 Biometric confirmation sticker, based on report from the provincial offices, was read by QR code reader and the discrepancy between the biometric data and votes from the result was more than six. 276
2 The discrepancy between biometric data and votes was less than five (31 stations) or more than six (2 stations) according to the secretariat’s report. The result sheets from election day at the National Tally Centre or the digital tally centre closest to the biometric data should be processed. 33
3 Out of 105, 21 stations have result sheets at the digital tally centre and the discrepancy between the biometric data and the votes from the result sheet is less than five (9 stations) or more than six (12 stations). 105
4 Copy of results forms have been received; the copy should be transferred and kept in the data centre. 30
5 Results of the audit and recount conducted by the provincial ECC and provincial IEC offices. 30
Total 509
 C) Validation of votes based on decision 117
  Reasons for validation Number of polling stations
1 Found unproblematic 832
2 Polling stations whose biometric confirmation stickers were read during audit and recount and the discrepancy of votes between biometric data and result sheet is more than six 18
Total 850
Grand total 5,767

Source: By AAN with data from IEC’s decisions 114 and 115

(4) Table 10: Categories of votes invalidated by the IEC through its decisions 113 and 116 

A)    Invalidation of polling stations based on decision 113 issued on 24 November
  Reasons for invalidation Number of polling stations
1 Open on election day, but no voters 812
2 Reported closed on election day 1
3 Ballot papers have biometric confirmation, but the biometric data are not available 94
4 Ballot papers inside the ballot boxes are with or without biometric confirmation, but the biometric data is not available 59
5 Ballot papers have no biometric confirmation sticker 187
6 No information is available 26
Total 1,179
B)    Invalidation of polling stations based on decision 116
  Reasons for invalidation Number of polling stations
1 Reported closed but not discerned 3
2 Ballot papers inside the box lacked biometric confirmation 63
3 Report is not clear enough 47
4 Ballot papers inside the box with or without biometric confirmation sticker 280
5 Invalidated by the ECC and confirmed by the IEC 2
6 Open but no votes 551
Total 946
Grand total 2,125

Source: AAN with data from IEC’s decisions 113 and 116

(5) Following the ECC’s conference on 5 February, a member of Abdullah’s campaign, Nur Rahman Akhlaqi, provided his own details (which AAN cannot confirm):

  • 298 polling stations contained 27,641 votes without biometric data.
  • There are 336 polling stations with a discrepancy of six or more than six votes which contain 14,462 votes.
  • There are a total of 42,103 votes in these two categories of polling stations (which he already called as invalid), 36,497 of which favour President Ghani and the rest favour other candidates.
  • There are 9,335 non-BVVed votes involved in the case of discrepancy of 1 to 5 votes between the result sheet and the biometric data and that the ECC’s decision to validate them is against the law (In a statement, TEFA also called the ECC’s validation of the votes with discrepancy of 1 to 5 between the result sheets and biometric data as against the constitution, without providing further explanation)

Akhlaqi also said that the ECC had not clarified the criteria for sampling and extrapolation for auditing ten per cent of the 137,630 votes. He called the ECC’s decision to audit 15 per cent of 102,012 votes a “compromise” and finding “average,” not “a legal decision.”