A good national turnout: between 40 and 45 percent

*Difference in turnout between 2004 Presidential and 2005 Wolesi Jirga elections

Source: Compiled from JEMB certified elections results for Presidential, WJ and PC elections and provincial voter registration figures as given on the JEMB website.

By Andrew Wilder

As voter turnout is clearly going to be one of the key issues, I’m adding a post providing a more detailed analysis of turnout for the 2004 and 2005 elections to inform 2009 analyses.

Above is a table extracted from my 2005 AREU paper, “A House Divided? Analyzing the 2005 Afghan Elections,” which provides a breakdown of voter turnout for the 2004 and 2005 elections by region. This shows that turnout had already declined from 70 percent to 50 percent between 2004 and 2005. Already in 2005 many of the voters I interviewed blamed that decrease on growing frustration and disillusionment with the poor performance of the Karzai government.

Given the 50 percent turnout rates of 2005, and the much worse security environment and disillusionment in 2009, any turnout rate between 40 and 45 percent in 2009 would seem to be quite good, and much higher than that bordering on suspicious. If there isn’t massive ballot box stuffing, and voters only vote once, turnout rates should also be lower simply due to the implausibly high voter registration figure of 17 million voters.

In discussions about turnout rates in the south, it’s important to note that turnout for the 2004 presidential election was only 53 percent, which declined sharply to 29 percent for the 2005 parliamentary election. Zabul and Uruzgan had the lowest 2005 turnout rates in the country of 20 and 23 percent respectively.

Given the much worse situation today, I would be surprised if overall turnout in the south was much higher than 20 percent in 2009. The 55 percent turnout rates in 2005 in the southeast were suspiciously high given the security situation, and support the reports of massive ballot box stuffing (or “proxy voting”) in the female polling stations (there are already reports that this happened again yesterday, although the extent is not clear).

Based on the 2005 results I would expect Kabul to have some of the lowest turnout rates. The central region dominated by Kabul had a 60 percent turnout rate in 2004, which declined to 36 percent in 2005. The turnout rate in Kabul province in 2005 was only 33 percent, not much higher than the rates in the insecure south.

In general, turnout rates in urban areas were reported to be lower than rural areas in 2004 and 2005, and I would expect a similar trend in 2009.

This trend is visible in other countries in the region, and probably reflects the fact that there are fewer social and political pressures to vote in urban areas (and possibly greater opportunities to stuff ballot boxes in rural areas!). The very low Kabul figures could also partly reflect a better educated and more cynical and disillusioned electorate.

In both 2004 and 2005 turnout rates were highest in the west, north, northeast and central highlands (and the somewhat anomalous southeast), which should bode well for Abdullah in 2009.

The male/female divide

And here is a table that provides turnout figures by province for the 2005 Wolesi Jirga and Provincial Council elections. It also shows the total percentage of votes cast by women and men by province (note: the gender breakdown is different than voter turnout figures, which would have to compare male and female votes with male and female voter registration figures).

Most notable, of course, are the high numbers of female voters in Daikundi, Faryab, Ghazni, Nuristan, Paktia, and Panjshir. Paktika also had suspiciously high female turnout rates, so much so that in 2005 the JEMB did not post the figures on its website (which is why those figures are missing from the table).

In some provinces high female percentages can at least partially be explained by the high percentages of males who have left to find work in neighboring countries or major cities. Reports of large-scale ballot box stuffing in female polling stations, is probably an even more important factor (especially in the southeast).

The much more plausible 3.7 percent and 14 percent figures for percentages of female voters in Zabul and Uruzgan respectively highlight just how implausible the nearly 55 to 60 percent figures are in neighboring provinces like Paktia, Paktika and Ghazni.

I was in Kandahar just after the 2005 elections doing research for the AREU elections paper and several of the candidates I interviewed highlighted their concerns about ballot box stuffing in the female polling stations.

They noted that outside of Kandahar city they and their polling agents hardly saw any female voters on election day, and in some stations none at all, but then the ballot boxes came back to the counting center in Kandahar full.

The figures from the 2008-9 registration of new voters highlights that use of female voter registration cards to rig the elections is a high risk (and as noted above has already been reported to be a problem in some areas). Particularly suspicious were the percentages of new female voters in Logar (72 percent), Nuristan (71 percent), Khost (65 percent) and Paktia (64 percent) — compared to only six percent in Zabul.

Andrew Wilder is a research director at the Feinstein International Center at Tufts University and the founder of Afghanistan’s first independent policy research institution, the Afghanistan Research and Evaluation Unit (AREU).