Thursday, May 16, 2013

Tanzania: Biometric voter registration without biometric verification at the polls

Tanzania: BVR Is for Voter Registration, Not Voting, Says NEC (All Africa)
NEC Vice-Chairman Judge (retired) Hamid Mahmoud Hamid clarified that people should take note of the fact that the system will only be used for registering voters and not for voting purposes. The commission's Head of PNVR and ICT, Dr Sisti Cariah, said NEC will collaborate with the National Identification Authority (NIDA) to reduce costs since the latter is currently doing the same in its national identification project.


Here's a piece, slightly edited, that we posted when initially it was reported that Ghana would forego biometric voter verification. Ultimately, Ghana decided to go for biometric voter verification, and despite some imperfections and a simmering dispute among political parties, they seem to have pulled it off. The same issues apply to the Tanzania voting infrastructure.

Originally posted May 15, 2012:
Without biometric verification, the whole enrollment exercise turns on the ID document. A document-dependent electoral system can be successful if three conditions are met: The process whereby legitimate documents are issued is very rigorous; The document is extremely difficult to counterfeit; And there is no significant corruption of the ballot-stuffing or ballot destroying variety.

Rigor in the document creation would include such measures as a real-time biometric query against the database of registered voters before issuing a new registration card in order to prevent duplicate registrations. Making a document difficult to forge involves high tech printing techniques or embedded biometrics for later verification. The corruption part is a function of culture and institutional controls.

Avoiding over-reliance on the physical ID document is perhaps the greatest benefit of using biometrics in elections. If there is no biometric voter verification, the only voting requirement is to have a more-or-less convincing registration card with a more-or-less convincing photo on it.

Biometric verification, by making the finger rather than the paper the overriding criterion for receiving a blank ballot, confers two tremendous advantages. Multiple voting can be made extremely difficult even for people who have multiple government issued registration cards. Second, ballot stuffing can be curbed because an audit of the total number of votes recorded can be compared to the number of fingerprints verified on election day as legitimate voters.

By creating the perception that the electoral apparatus is more effective than it really is, implementing a biometric voter enrollment system without biometric voter verification could even lead to more electoral uncertainty than the system being replaced.

A well-thought-out biometric voting system can reduce fraudulent voting to very low levels but it's also possible to spend a lot of money on a leaky system that involves biometrics without accomplishing much in the way improving the integrity of the vote.
The same sort of analysis can, and should be applied in Tanzania.