Elections have emerged as a leading area for the application of biometric technology in developing countries, despite its high costs and uncertainty over its effectiveness. One-off voter registrations, as practiced in many countries and supported by donors, also often leave nothing behind in terms of permanent, sustainable, identification assets. Why then do donors support such programs? The paper considers the costs and benefits of technology, where the latter involves its potential to reduce the probability of seriously disputed elections that escalate into violence. Based on the limited data available, it finds that a reduction in the probability of post-election violence by only a few percentage points could offset the cost of the technology. However, this is possible only in particular situations where political parties value the legitimacy conferred by elections that are sufficiently credible to provide an acceptable basis for governing, but where democracy is not yet well-institutionalized. One priority is therefore to screen potential cases carefully. Another step towards using technology more effectively would be to build on the powerful momentum created by voter registration drives to strengthen permanent identity assets such as civil registration and national ID programs, so that they can provide a more sustainable foundation for subsequent voter rolls. This is possible with careful planning but may require some reconsideration of institutional mandates.