Election season Afghanistan has been the source of no shortage of intrigue, violence, negotiations and maneuvering. In recent days, despite pressure from the United States, UN and other international players, the ongoing audit of the Afghan Presidential election is stalled once again. Both of the remaining presidential candidates, Dr. Abdullah and Ashraf Ghani, believe that they have won the election, and will do their best to steer the process in their respective favor.
But how did we get here? A deadlocked electoral result, with seemingly contradictory stories from the candidates, observers, journalists and international actors who are clearly anxious to disengage as soon as possible. The first round of the this year's election (held back in April 2014) included eleven total candidates on the ballot, ranging from the outgoing president's brother Quayum Karzai (once considered a frontrunner) and Gul Aga Sherzai (also briefly considered a contender) to economist Hedayat Amin Arsala (who never had a chance).
Only three candidates won at least one of Afghanistan's 34 provinces in the first round of the election:
- Dr. Abdullah: 19 provinces, 45.0 percent of national vote
- Ashraf Ghani: 14 provinces, 31.6 percent of national vote
- Zalmai Rassoul: 1 province, 11.4 percent of national vote
The top two finishers then continued to a second round vote on 14 June, which resulted in an extraordinary reversal of fortunes. Winning 18 provinces to Abdullah's 16, Ghani earned a whopping 56.4 percent to Abdullah's 43.4; Abdullah had the distinction of winning indeed a lower percentage of the vote in the second round than the first, even with nearly a quarter of first round voters now able to vote for him with their original candidate out.
The following graphic illustrates the geography of the first round, looking just at Abdullah and Ghani. Blue areas are where Abdullah won the first round vote and red areas where Ghani prevailed. We notice that Ghani's support is concentrated in the Pashtun east, along the Pakistani border. Ghani also does well in the Uzbek-dominated provinces of Jawzjan and Faryab, where his alliance with General Dostum comes into play. Abdullah does well just about everywhere else, earning the Hazara and Tajik vote, along with a substantial number of Pashtuns.
The second round, illustrated below, follows the same general outline, but with far greater support for Ghani in the southern Pashtun areas and a major swing in Wardak province (just west of Kabul), which went for Abdullah by 20 points in the first round and Ghani by nearly 60 in the second round.
What do we observe here? Well, there is a generalized swing around the country towards Ghani, however, there are several hotspots that would indicate a whole scale realignment of politics if they were seen in a "normal" election. The following graphic shows the swing towards Ghani from the first to the second round.
Wardak province is by far the most shocking result; somehow Ghani was able to completely turn the tables on Abdullah. Kandahar and Zabul also see humongous shifts towards Ghani, but this was in part a question of turning out voters rather than shifting their preferences.
This massive shift in margins was possible only because of a massive increase in voter turnout between the two electoral rounds, one reason why observers and the Abdullah camp are crying foul. More than 1.3 million new votes appeared in the second round, a nearly 20 percent increase. Many (possibly most) were likely fraudulent or coerced. Wardak is a smoking gun, but there are many other instances.
The following shows the surge in voter turnout in Ghani-dominated areas, including a massive increase in Wardak.
When we look at the places where Ghani achieved big increases in his vote margin, and where the change in turnout was highest we find a strong correlation (the naive stats: running OLS vote we get a t-stat of 2.92 and R-squared of 0.21 ). Following is a scatter plot of the Ghani swing over percent change in voter turnout.
All in all, Ghani's second round numerical victory was brought about by finding 1.3 million new votes, and depositing them in key spots, mainly in his base areas. What happened in Wardak is up for debate, but seems to provide strong evidence of tampering at minimum. Real elections never see those kinds of swings between rounds.
Going forward, it will be a political negotiation between the Abdullah and Ghani camps, perhaps in the form of a explicit or implicit power sharing arrangement. Regardless of the outcome of the formal audit process, whoever walks away with the presidency will have to accomodate his rival in some way to avoid serious unrest and potential violence.