Biometric voter registration and verification technology is rapidly becoming a staple of elections in the developing world, with over 41 low and middle income countries now using fingerprints in their voter registration process. Donors—including the UN, the EU, and many bilateral aid agencies—have supported biometric elections generously despite several concerns. First, technology does not necessarily increase the credibility of elections and could be used to legitimize an otherwise crooked process. Second, one-off electoral registrations leave little or nothing behind in terms of permanent identification assets and may even divert attention and funding away from civil registration and financially sustainable national identification systems. Why do such wasteful exercises continue to be supported despite uncertainty over their effectiveness? And how can voter registration be harnessed to support a more sustainable system of registration and identification?
In our recent CGD Working Paper, we consider the social and economic costs of violently disputed elections in Sub-Saharan Africa and compare them with the cost of biometric technology. While biometric technology can be costly—usually from $15 million to $100 million per election (Table 1)—its price tag seems relatively minor compared to the potential costs of post-election violence. This can run into the billions of dollars as economic growth stalls, in addition to less readily quantifiable human losses. If biometrics can make even a modest contribution to delivering more credible elections—and thus reducing the likelihood of violence—their use could be a worthwhile bet.
This table draws on a range of studies and reports. For a full list of sources, see our working paper.
When might biometric technology contribute to credibility?
Biometrics, most commonly fingerprints, can be deployed at registration to eliminate multiple registration and thus help ensure a clean voter roll. Biometrics can also be used to verify a voter’s identity at the polls—as Nigeria did in 2015, for example—but this is less common. However, technology provides no guarantee for free and fair elections. Biometrics have no direct impact on voter intimidation or vote buying or on the rules regarding who is allowed to run, what financial and other support candidates may receive, or what campaigning tools and platforms they can access. Even with a clean voter roll, votes can be intentionally miscounted, vote tallies altered or simply ignored.
Whether biometric technology can make a contribution to credible elections depends on whether the most egregious threats to electoral integrity coincide with the types of fraud it can address (voter impersonation, inflated voter rolls) and thus on the political context. Where important political actors are not prepared to accept the possibility of an adverse outcome, technology will be ineffective and a waste of resources. Biometrics have the best chance at contributing to credible and peaceful elections in cases where political candidates value the legitimacy conferred by an election that is widely recognized as largely free and fair and where, even though there may be efforts to subvert the election, evidence of overt fraud is seen as a political liability.
Maximizing the potential of biometric voter registration
Even if the use of biometric technology can improve credibility, holding one-off voter registration exercises that have to be repeated ahead of each election and that fail to build a more permanent broader civil registration and identification system is very wasteful. Voter registration drives can be astonishingly successful in registering a large number of the population and have often provided a first official ID to millions living in countries with weak civil registration systems. In Tanzania, 23.2 million people were registered for the 2015 election in only four months compared with 2.4 million for the national ID program over four years. How can the momentum of voter registration be harnessed to support a robust and sustainable ID system that can, in turn, provide the foundation for the voter roll?
“Backward integration” offers the most promising solution. This would require the election authority to transfer data to the civil registry (or population registry) after the election, to enable interoperability between the registers so that the data can be integrated, and to collect adequate quality voter data in the first place. The technology used for voter registration should also be compatible with that used for national registration, so that the costly biometric voter registration kits can be reemployed. Tanzania is currently implementing such a process. The aim is to integrate most of the 22.3 million voter records into the National Identification Authority database and produce—it is hoped—a unified registry. Subsequent voter rolls could then be based on the integrated ID system. This is not a simple process, but if successful it would help maximize the potential of biometric voter registration and turn it into a worthwhile long-term investment.