Foreign Policy Research Institute A Nation Must Think Before it Acts The Afghan Presidential Election: a Prediction and a Warning

The Afghan Presidential Election: a Prediction and a Warning

A Report on Surveys Conducted in Afghanistan

Andrew Garfield recently led a joint Afghan-American team that was commissioned by the U.S. Department of State to hold three pre-presidential election polls in Afghanistan between November 2013 and March 2014.  The results of his work make for interesting reading and he has predicted the likely outcome of the first round of the election. Having earlier predicted accurately the outcome of the 2009 election, his findings are worthy of consideration.  He also warns against complacency about the fairness of the likely run-off election.


My team ran two polls prior to the 2009 Afghan presidential election and accurately predicted the voter share of each candidate, including President Karzai and the current front runners Dr. Abdullah Abdullah and  Mohammad Ashraf Ghani Ahmadzay.  We predicted that President Karzai would secure far less than the 50percent needed to avoid a second ballot. However, amid numerous accusations of widespread and rampant election fraud, on September 16, 2009, the Independent Election Commission (IEC) released its final uncertified results for the presidential election, indicating that Karzai had won the first round of election outright with 54.6 percent of the vote, with his closest rival Abdullah getting 28 percent.  As evidence of fraud mounted and independent reviews, including my own analysis, concluding that Karzai had secured less than the 50 percent needed to avoid a second ballot, Karzai announced on October 20, 2009, under significant international community pressure, his agreement to a second ballot and a runoff with Abdullah. The next day, on October 21, 2009, the UN-supported Electoral Complaints Commission (ECC) released its final certified results, giving Karzai 49.67 percent of the vote and Abdullah 30.56 percent.

According to my findings from a second unpublished poll (due to the media embargo imposed close to the election), Karzai may in reality have gotten significantly less than 49.67 percent. Of significance, though, is the fact that Karzai was far more popular with Pashtun voters than Abdullah, whose support was underpinned by Tajik voters.  My assessment therefore was that in a second ballot Karzai would have secured around 60 percent of the popular vote while Abdullah would have great difficulty getting above 35 percent.  Ergo, Karzai could have secured a significant and legitimate victory had he ensured that the first round of the election was run more or less fairly and had he himself had the self-confidence and maturity to accept that a second ballot was necessary.  Given that the results of these polls were available to him and the international community prior to the election, it is disappointing that none of his backers was willing or able to persuade him to run fair elections with the near certainty that he would win a second ballot.  A wonderful opportunity lost.

On November 1, 2009, Abdullah announced that he was withdrawing from the run-off vote, saying “I will not participate in the November 7 election,” primarily because his demands for changes in the electoral commission had not been met, and a “transparent election is not possible.” On November 2, the Independent Election Commission announced the cancellation of the November 7 run-off and declared Karzai the winner by default.  Abdullah said the appointment had “no legal basis” and Afghans deserved a better government.  This was a political master stroke by Abdullah, given that he surely knew that he could not secure sufficient support among the Pashtun to win a second ballot and that he had just about maxed out his support among the Tajiks and the other ethnicities.  By withdrawing, he was able to secure the moral high ground and present himself as the wronged candidate, despite having little chance of winning a second round of elections. He has used this position successfully over the last five years to position himself for a return run in the 2014 elections.


It is against this background of rampant election fraud in 2009, that my company, Glevum Associates, was given a grant by the Department of State to run up to three pre-election surveys.  Two other organizations, including Democracy International, were given similar grants.  The aim of these surveys was to learn more about likely voter preferences, voter participation trends, and attitudes toward the candidates and to then make that information widely available to the Afghan public. It was felt that these surveys might also help inform the candidates regarding the trends, issues and concerns that matter most to likely voters and provide them with statistical evidence of changes in momentum during the campaign. Those involved knew that such systematic polling was a new concept for Afghanistan but it was felt that it is one that both voters and candidates would soon come to value, as they better understood the process and saw that the information being provided was of value to them.

The first survey we conducted was undertaken in all 34 provinces of Afghanistan, with around 2,500 Afghans questioned about the upcoming elections. Of this total, the responses of 2,148 likely voters were included in the survey findings.  The survey was fielded between November 27 and December 3, 2013 using face-to-face interviews. The sampling margin of error is + 2.11 with a 95 percent level of confidence. With regard to respondents’ favored candidates, while it was still early in the election cycle, among likely voters surveyed, Ghani had a 4-point lead over his nearest rival Abdullah.  He had 29 percent compared to 25 percent for Abdullah. Abdul Qayoom Karzai was the next most popular candidate with 8 percent; Professor Sayaf at 6 percent; Zalmay Rasool 6 percent; and Rahim Wardak 5 percent. All other candidates were below 5 percent. About one in ten likely voters (11 percent) remained undecided. We also asked respondents whom else they would vote for, if they could not vote for their preferred candidate. Again, Ghani and Abdullah were neck and neck (21 percent and 20 percent, respectively) as the second choice candidate. Among all candidates, Ghani’s support was potentially the most firm with 70 percent of those who say they would vote for him also saying that they would not change their minds before the April election.  In some elections high-profile endorsements can help the candidate who gets such an endorsement.  However, we found that fully 85 percent said they would not be swayed by Karzai’s endorsement of a particular candidate or that it would not matter to them. 

Despite our hope that the polling process would be welcomed by most Afghans, from the outset this effort was highly controversial. Democracy International (DI) published the findings of their Wave One poll first and their offices were subsequently raided by members of the Afghan security services. Some reports suggest they were raided by supporters of Ghani, who placed second in the DI poll.  Our team published next and while their press conference was well received and there was extensive coverage of the findings both in Afghanistan and internationally (including in the NY Times, December 29, 2013) by now there was considerable hostility toward this polling effort from Karzai’s office and from many of the candidates.  Bowing to this pressure, the grants were eventually cancelled and no further polls were either completed and/or published.  My team did, however, complete the fieldwork for a second wave of polling. This second face-to-face national survey was conducted between 14 and 25 January 2014 with 2,497 Afghans who are likely to vote responding. When one combines the results of both waves of polling and using the same assessment techniques employed so reliably ahead of the 2009 elections, which were subsequently validated by the certified election results, we have again assessed the likely results of the first and, if held, the second round of the 2014 election.

Our assessment is that Abdullah and Ghani will be the clear winners of the first round of the election but neither of them will come close to an absolute majority.  Both candidates will secure between 39 and 44 percent of the vote each.  The other supposed frontrunner (at least according to the BBC) is Zalmay Rassol. He is likley to finish with less than 10 percent of the vote, despite the endorsement of Karzai’s brother, Abdul Qayoom Karzai, who withdrew from the campaign, and the reputed support of the President himself. Neither Ghani nor Abdullah will secure an outright majority, and they will therefore have to contest a second round.  Anything other than a runoff between Ghani and Abdullah would be highly suspicious and would call into question the integrity of the elections.

In a prospective second ballot, Ghani has a slight advantage in that he is more likely to secure majority Pashtun support, which is the largest ethnic group at around 39-40 percent of the population.  He already has significant support amongst urban Pashtun voters and while not easy to do, he should secure the bulk of rural Pashtun voters, especially given that he would be running against a candidate who is closely associated not only with the Tajiks but also the Northern Alliance.  To be sure however, he will need to pick up support from other ethnic groups and the support of some Tajiks.  He may therefore need to add an important Tajik leader to his ticket.  For Abdullah to stand a chance in a second ballot, he needs to be endorsed by some leading Pashtun leaders and secure the majority of votes from the other ethnic groups, including the Hazara and Uzbeks.  He will most likely secure majority support among the Tajiks, who represent about 34 percent of the population.   Were he to secure the support of Zalmay Rassol, perhaps inviting him to join a “dream ticket” for the second ballot, that would really open up the race.  However, on balance, a second ballot is Ghani’s to lose.

Early indications are that the first round has been run more or less fairly, despite some obvious examples of election fraud. The stakes will be far greater in a second ballot and the choices far starker.  All Provincial and District Governors and all police chiefs down to the lowest levels are directly or indirectly appointed by the President.  These key presidential appointees, who have the ability to influence the outcome of the ballot in their locales, will mostly likely favor and/or owe allegiance to one of these two candidates and there will be significant temptation for them to try and protect their positions and their ethnic groups position by rigging the elections within the areas they control.  Inside Afghanistan, it is further suggested that Ghani has significant support from within the Afghan security services, who would also have a vested interest in backing their favored candidate.  The international community must therefore be extra vigilant during a second ballot and not become complacent simply because the first round went well.   The supporters of both candidates and especially Ghani may believe they have much to lose if the other candidate wins and therefore they may feel compelled to influence the outcome of the second ballot by extra-legal means.