Afghanistan Analysts Network – English

Political Landscape

Elections 2014 (7): An emerging mixed picture

Martine van Bijlert 8 min

Women voters in front of a Kunduz polling centre. Photo: AAN (Obaid Ali)

It is four days after the election. The country is generally still in a good mood after a poll that went much better than expected. At the same time, a more complete picture is emerging and it is, unsurprisingly, mixed. Alongside a robust genuine and determined vote, there are indications of significant irregularities: old patterns of intimidation, ballot-stuffing, and ‘ghost polling stations’ in remote and insecure areas, and possible new patterns of ‘early’ and ‘late voting’ in areas where people came out to vote. The latter seems to have been, at least partially, responsible for the remarkable ballot shortages that were reported in many parts of the country. The IEC and IECC now seem faced with the unattractive task of dealing with polling stations where both real and fraudulent votes were cast. AAN’s Martine van Bijlert takes a closer look.

The Independent Election Commission (IEC) is busy retrieving the ballot boxes and results forms and will, by now, have formed a more complete picture of how the vote went in the various parts of the country. The Independent Election Complaints Commission (IECC), in turn, has been collating the election day complaints. In the meantime, journalists, observers and campaign teams have been trying to extrapolate conclusions from ragged samples and anecdotes; this in the absence of solid information, which will still take quite a while to emerge.

Initial, fluctuating figures

Four days after polling, information is still fluid, anecdotal and scarce – which was to be expected. On the evening of the vote the IEC released the widely reported turnout estimate of 7 million voters; which it took to be 58% of an estimated 12 million eligible voters. The figures match observations of a high turnout, particularly in the urban areas and surrounding districts, but It remains anyone’s guess as to how solid they really are, as nobody knows how many voters there are, how many of them hold a voter card, or how many of the ballots cast will turn out to have really been linked to voters.(1)

The IEC has released figures on the number of polling stations that did not open on polling day and these figures have fluctuated: the IEC gave a figure of 205 polling stations immediately after the conclusion of the vote, and respectively 211 and 251 at press conferences in the following days. (The IEC had planned to open 6423 polling centres containing 20,795 polling stations  out of a total list of 7168, after a last round of security-related removals). The number of polling stations that were added after sites said they started running out of ballots, and the details of where they were added, has not yet been released. (Instead of sending spare ballots to polling centres that ran out, the IEC had decided to add whole new polling stations. In this way they could keep better control of the process and uphold the principle that no polling station could report more than 600 ballots cast).

The IEC has just updated its website to reflect the fact that election day has taken place and that results may soon be posted, but so far no information of substance has been posted. A fact sheet with electoral facts and figures was sent around to an election-related mailing list on the eve of the elections and a Dari press release on the night of election day, but other than that written reporting has been scarce. The fact that almost all communications – from both the IEC and IECC – have been oral (press conferences and answers to journalists’ queries), means that information is passed on through news articles and social media, with or without proper sourcing or dating. So far, no collated information – other than an initial, incomplete overview of registered complaints by the IECC – can be found in one place.

The IECC’s complaints statistics as posted on the website (Dari only) are still in flux. Updated figures were released at a press conference on 8 April 2014. However when contacting several provincial IECC offices, several of them appeared to have higher figures (not counting complaints by telephone, which the IECC is treating as information rather than formal complaints). It is also not clear whether complaints registered with IEC staff at the polling stations have also been counted.

Finally, government reporting on the number of security incidents has fluctuated. Initial reporting by the Ministry of Interior (MoI) and the National Directorate of Security (NDS) on the evening of election day, found that the number of ‘security incidents’ had been significantly less than in 2009, citing a figure of around 160 incidents. The Ministry of Defence (MoD) announced on the next day that it had counted 690 security incidents. The difference can probably be explained by a combination of different ways of counting incidents, a time-lag in gathering the full picture and the different messages that the authorities were trying to give at different times.(2) Something similar happened after the 2009 election.

An emerging picture of irregularities

As information is comingin, the picture that is emerging from anecdotal and provincial reporting suggests significant irregularities. Some of it has followed old patterns: Low turnout in remote and insecure districts, polling centres that cannot be found in the indicated locations, and reports of ballot boxes having been taken away – before, during or after polling day – to be stuffed or tampered with. For reporting in the media on cases in Wardak province, see here and here. There are also incoming reports of fraud attempts having been thwarted by observers, voters, representatives of rival candidates and local strongmen. Observers and campaign teams have taken note of these areas, and where possible the affected polling centres (we have done the same), and will be watching the IEC results closely to see whether irregularly cast votes will be counted.

But there also seem to hae been new patterns. One of the themes of this election may turn out to be the ‘ballot shortages’, which eventually caused the IEC to release the full scope of its contingency ballots (500 extra polling stations or 300,000 ballots). The reported shortages have been alternately attributed to a high turnout, bad planning and intentional discrimination, but none of these reasons seem to provide a full explanation. Although voter turnout really was higher than expected, it is unlikely this can account for the scale of the shortages, or the speed with which some polling centres reported to have run out of ballots. According to the National Democratic Institute, in its preliminary statement, the IEC had decided not to respond to early requests for more ballots, as according to their calculations it would normally take approximately 600 minutes, or ten hours, for 600 voters to actually vote.(3) The first polling centres, however, started to report ballot shortages well before noon. The IEC is now looking into the possibility that a considerable proportion of the shortages may have been caused by ‘early voting’, ie the partial stuffing of ballot boxes before polling stations opened, or ‘late voting’, ie the hiding of papers to be subsequently filled in, after hours. The fact that turnout was higher than expected – at least in the cities and in areas which were safe or where the determination to vote could outweigh the risks – may have not only taken IEC planners and observers by surprise, but also the partial ballot stuffers. The full extent of the problem is not yet clear, but the scale of the shortages, the fact that many of them were fairly localised, and the fact that in some areas it coincided with early refusals to let in independent observers, suggest a considerable level of organisation.

If the ballot shortages indeed turn out to be caused by attempts to defraud the vote in those localities, the IEC is faced with the unattractive  technically difficult and politically awkward  task of deciding how to deal with polling stations that contain both real and fraudulent votes.

A lesser degree of tumult – so far

The reporting on fraud has, so far, been less tumultuous than in previous elections. This could be because there simply was less of it, or because it was less widespread. It could also be because it was more evenly spread between the candidates, at least in the presidential race, making it less obvious who would be hardest hit if irregularities were to be investigated rigorously. For the moment, the main candidates seem to be keeping the issue of possible fraud alive, albeit on the backburner, in case they need it in the future. But they are at the same time not raising it too much, as complaints could backfire and unearth their own irregularities, or alternatively taint a competition that they may have a hope of winning.

Additionally, there seems to be quite some hesitation among local journalists, analysts and even observer organisations to ‘spoil the party.’ It is so rare for Afghanistan to have a good news story, especially one that actually resonates with many Afghans (so far this has mainly been the case with sports events) that there seems to be a real reluctance to taint that experience with stories about interference that may or may not turn out to be relevant to the final outcome.

Home-made results

Finally, in terms of determining who people voted for and who may be ahead in the presidential election, there has been a great deal of incomplete and premature extrapolation. An example in the international media is this Reuters article that suggests Abdullah may be in the lead across the country based on a small sample of polling stations (‘dozens’) in the north and west of Kabul only. Members of presidential campaign teams have been posting home-made election results on social media – often dressed up with official-looking pie-charts, diagrams and error margin estimates – invariably claiming a majority win for their candidate. The candidates themselves tend to take a moderate stance, acknowledging the very slim chance anyone would win a first round and reiterating that only the results announced by the IEC are ‘real’ results. But they do not seem to mind that parts of their following are in essence being primed to be upset once the real results start coming in.

Now what?

The IEC has called a press conference for this morning. It will probably release further information on what happened on election day, although the full picture may not yet be clear. A large proportion of the electoral material is still in transit, as it is making its way to the provincial warehouses (the ballot boxes and ballots) and the national tally centre in Kabul (the results forms). As of last night the material from 37 districts (out of some 400) had not yet reached the provincial centres, let alone the tally centre in Kabul. The IEC still needs to release full information on which polling stations were open and which were not, where ballot shortages were reported and where extra polling stations were opened (ie extra ballots delivered). It says it is working on this.

As of 8 April morning, the electoral material from Kabul, Kapisa, Panjshir and Parwan had been received, while staff were waiting for helicopters to arrive with material from other parts of the country. Data entry did not appear to have started yet. It will probably take until Saturday, at least, before the first preliminary results – of the first 5% of the count – will be released. These results will tell us very little about who may have won, but it will provide observers with results that can be analysed for indications as to how robust the IEC audits are.

The process is still very much in its early stages; details are only slowly firming up. But the picture emerging is that the mood of elation on polling day will not be the whole story. Despite the more muted fraud reporting and complaints so far, both the IEC and the IECC are probably looking at a difficult task ahead. Not only are they dealing with what has still been a messy vote, but also with a tension between their duty to act against the fraud they find, and the seeming desire of many Afghans to keep this election untainted by exposures, again, of possible widespread vote-rigging.


(1) As discussed extensively here, nobody knows how many Afghans over 18 there are or how many of them have registered and obtained a voter card. Population estimates suggest that the total number of voting-age Afghans may be around 13.5 million, but by far not all of them have registered as voters. The total number of distributed voters cards, meanwhile, has reached 21 million over the years and thus far exceeds the number of eligible voters. The IEC is now saying it had prepared for 12 million voters. The total number of printed ballots per election (presidential and provincial council) was a little over 15 million.

(2) During election day, the security organs probably downplayed the number of incidents and there is anecdotal information that journalists have been asked not to report on security incidents so as not to dismay the voters (as had been the case in the past). Shortly after election day, they may have mainly wanted to emphasise the failure of the Taleban to launch major attacks, which was then possibly overtaken by the wish to project the image of a security apparatus that had battled valiantly and to add to a mood of defiance. Moreover, ways of counting differ: a ‘security incident’ may range from anything from an armed clash to an unexploded IED. The numbers have to be interrogated further.

(3) Our own observers, as well as others, indicated that they believe voters could vote more quickly than that; each polling station had three or four booths, so voters did not have to go through the whole process before the next one could start. Still, even at the most efficient stations, it seemed to take at least 30 seconds before the next voter could step up. Taking into account delays, hesitating voters, glitches, small upheavals, as well as an hour lunchbreak, it is implausible that even the most efficient polling stations would run out of 600 ballots early in the day.


counting election security Elections fraud IEC IECC


Martine van Bijlert

More from this author