Saturday, June 20, 2009

Iran Election: The Numbers

My earlier post was not factual (well, the US map was), but the joke did find its roots in the uncanny coincidence of a victory with nearly two thirds of the vote. This post will discuss the official election results from three rounds of Iranian elections, the first round of the 2005 elections, the runoff round of the 2005 elections, and the first and only round of the 2009 elections (there is only a runoff election if no candidate wins a majority in the first round - in 2009 Ahmadinejad was declared the winner with more than a majority of the votes).

You can view the full spreadsheet with election results broken down by province here. The numbers originally came from the Iranian Ministry of Information and were translated into English by the Guardian newspaper. I have taken the liberty of reformatting it (specifically to make it less annoyingly British - what sort of depraved society adds a period after every single number in a table?) and performing additional calculations (specifically adding a "Differences" section to highlight interesting changes between the 2005 and 2009 elections).

I am not a statistician, I simply work with spreadsheets a lot. This is not a proper analysis, just a basis for discussion amongst laymen.

After looking over the numbers the feature that stands out the most to me is the extreme similarity between the percentage of votes for the different "blocs" in each round of voting:

2005 election, first round
All conservative candidates: 63.0%
All reformist candidates: 37.0%

2005 election, second round
Ahmadinejad (conservative, placed second in first round): 63.2%
Rafsanjani (conservative-leaning moderate, placed first in first round): 36.8%

2009 election, first round
Ahmadinejad (conservative): 63.3%
All other candidates: 36.7%

These similarities led me to consider the following scenario:

Both elections were rigged. Back in 2005 the Iranian leadership thought to itself "63% is a good number to win with. It's almost a two-to-one margin, but it's not so high as to be unbelievabe." In 2005 all the conservative candidates were allocated 63% of the votes during the first round, but no one candidate won a simple majority, just to make it exciting. In the second round they gave 63.2% of the votes to Ahmadinejad, their choice for president, a slight increase of the magic 63% to show his increasing support amongst "the people." In 2009, for whatever reason, the Iranian leadership was too afraid to allow a second round of voting; maybe Ahmadinejad did win, but not with a simple majority, and they were afraid that Mousavi's status as a symbol of reform would only continue to gain momentum if given another week of campaigning. Or maybe they actually saw Mousavi winning as the returns started to roll in and they panicked. Whatever the case, they gave Ahmadinejad 63.3% of the vote (Yet another high! His support grows ever greater!) and hastily declared the election over.

This is pure speculation on my part, with only the barest amount of information inspiring the idea for it. It makes a plausible story, but just because a story fits together neatly doesn't mean that it's true. However, a story can make a good starting point for a discussion.

The fact that the 2005 runoff results broke along almost exactly the same percentage points as the first round of voting broke along conservative/reformist lines is very odd considering both candidates in the runoff were conservative. Most observers of the 2005 election considered Rafsanjani to be something of a reformist candidate compared to Ahmadinejad, but it doesn't make sense that the 22% of conservative voters who supported him during the first round of voting would completely abandon him when once again given the opportunity to vote for Ahmadinejad instead, nor that his only support would come from reformist voters who flocked to him en masse. It does make sense, however, that the Iranian leadership would reclassify Rafsanjani as a reformist when placed in a runoff with Ahmadinejad, and decide to give him the 37% of the votes they had decided to allow the candidates and ideas they wanted to marginalize.

The spread of results is again nearly identical in the 2009 election, despite the fact that there were 10.6 million more votes in 2009 (an increase of 37.8% over the 2005 election) and individual provinces recorded ideological shifts of up to 35.6% increased support for conservative candidates and up to 19.4% increased support for reformist candidates between the two elections (see the spreadsheet for the breakdown). Also, many writers have already expressed extreme skepticism at the official results' claim that two provinces (Mazandaran and Yazd) reported voter turnouts of over 99% of eligible voters in the areas.