« PPBW: Ansolabehere | Main | Electoral politics and FEMA's poor performance in the aftermath of Hurricane Katrina »

20 September 2005

The Afghan Election: Is Voting Really Not a Paradox?

The results of Sunday's Afghan election won't be in for a few more days. No doubt, it will take quite a bit longer to find out whether the country's experiment with democracy will be successful or whether Afghanistan will once again plunge into chaos and civil war. One thing we do know, however, is that about 50% of the 12 million registered voters showed up to the polls. Some officials expressed disappointment in the lower-than-expected turnout; still, 50% is a pretty respectable number, even by the standards of the world's most successful democracies (by comparison, turnout in this weekend's German election was around 40%). Possibly the best-known solution proposed by rational choice theorists for the age-old "paradox of voting" - as stated in Aldrich's classic AJPS article - is that voting is neither a “good� nor is it a particularly “problematic� example of the collective action problem. Turnout is a low cost, low benefit action; the decision is made “at the margins,� with small changes in costs and benefits often proving critical. Given that voting is not a good example of a CA problem - so the argument goes - the fact that millions routinely show up to the polls does not kill rational choice. In most settings, it is quite realistic to think of voting as a low cost activity. Elections held in the past few years in places like Afghanistan and Iraq, however, may be seen as pretty glaring anomalies in this respect. The benefits of voting remain quite low; the costs, however, appear to be immeasurably higher than those we generally face in American or European elections. Is it reasonable to say that the decision made by millions of purple-fingered Iraqis and Afghanis to go out and vote is not much of a puzzle because costs and benefits are low? If not, what does this say about the standard rational choice defense?

Posted by Federico Ferrara at September 20, 2005 8:59 PM

Comments

Voting can be very beneficial, in the sense that if your preferred candidate wins, it can benefit millions, and to the extent that you want others to be happy, this is a benefit to you. I do not think it is at all irrational to vote at some cost if the potential benefits are large enough.

See here for a more formal development of this argument, including various evidence and a feedback mechanism that keeps turnout at reasonable levels.

Posted by: Andrew Gelman at September 22, 2005 8:46 PM

Actually, 50% is an abysmal level by the standards of most democracies, but especially compared to other founding legislative elections in new democracies.

Don't overlook the role of the electoral system and constitutional structure chosen for this election in making turning out to vote seeming less than important to Afghans.

Posted by: Matthew Shugart at September 27, 2005 3:04 PM