Effects of Election Monitoring via Webcams
Evgeny Morozov's commentary in Foreign Policy about Azerbaijan's use of webcams to monitor elections has prompted some pessimistic online commentary about the conduct of elections in the region. Indeed, Azerbaijan's election quality has been rated poorly by international organizations that monitor campaigns (with the notable exception of the CIS).
Morozov argues the following about the use of webcams:
I don't know how it would hurt, but I don't see how it would help either: unless there are enough cameras to cover the entire voting/counting areas, rigging the results would still not be too hard. If anything, the presence of a Web cam would probably give yet another excuse to validate the results...
His note focuses on the installation of webcams for upcoming municipal elections. However, webcams were in place for the last presidential election and referendum. I have analyzed turnout and outcome data, differentiating among polling stations that installed webcams and those that did not. Officially reported turnout was consistently lower in polling stations that had webcams. The presence of webcams had a less consistent effect on pro-regime outcomes.
What could produce these results? I speculate that election administrators have a reduced incentive to inflate outcomes - especially turnout - when they are being monitored. Some officials were punished for election violations in Azerbaijan's 2000 and 2005 elections. While these officials were likely scapegoated, their punishment potentially sends a message to administrators that they could share the same fate. Placement of a webcam in a polling station raises the possibility of being punished if the footage shows clear evidence of inappropriate behavior. With static cameras that show one view of the polling station, viewers could identify egregious inflation of turnout if the footage were observed consistently over time.
I address other possibilities, and present the statistical analysis, in a research paper available to anyone who is interested. Please contact me at: eherron at ku.edu and I will forward a copy.
Morozov argues the following about the use of webcams:
I don't know how it would hurt, but I don't see how it would help either: unless there are enough cameras to cover the entire voting/counting areas, rigging the results would still not be too hard. If anything, the presence of a Web cam would probably give yet another excuse to validate the results...
His note focuses on the installation of webcams for upcoming municipal elections. However, webcams were in place for the last presidential election and referendum. I have analyzed turnout and outcome data, differentiating among polling stations that installed webcams and those that did not. Officially reported turnout was consistently lower in polling stations that had webcams. The presence of webcams had a less consistent effect on pro-regime outcomes.
What could produce these results? I speculate that election administrators have a reduced incentive to inflate outcomes - especially turnout - when they are being monitored. Some officials were punished for election violations in Azerbaijan's 2000 and 2005 elections. While these officials were likely scapegoated, their punishment potentially sends a message to administrators that they could share the same fate. Placement of a webcam in a polling station raises the possibility of being punished if the footage shows clear evidence of inappropriate behavior. With static cameras that show one view of the polling station, viewers could identify egregious inflation of turnout if the footage were observed consistently over time.
I address other possibilities, and present the statistical analysis, in a research paper available to anyone who is interested. Please contact me at: eherron at ku.edu and I will forward a copy.