The use of ICT in voter registration can become controversial for several reasons. The accuracy of voter registers is often a contentious issue, and technology upgrades may be seen as a solution to that. However, voter registration is a large and costly exercise, and increased ICT use comes at a high cost. The sustainability of high-tech approaches can be questionable, especially when substantial donor funding is required or when registration mainly focuses on establishing a single register for an upcoming election rather than a permanent system for continuously maintaining a register.
The operational complexity and cost of voter registration primarily depends on how voter registers are built and maintained. If an EMB can share data with other administrative bodies and build a voter register based on existing population registers (e.g. national identification card systems, civil registry data), this will require significantly less technical effort than if it has to compile the voter data itself. For use in elections, population registers must be maintained by a widely trusted institution, use high-quality data, have a high registration rate (with many incentives and few disincentives for registering) and contain all information about eligible citizens required for the electoral process. If these conditions are met, an inexpensive and highly accurate voter register can be extracted automatically from such a system.
If data sharing with other administrative bodies is not possible and the EMB needs to conduct voter registration on its own, costs will be significantly higher, as the EMB must operate a system that can reach all citizens and capture or update the personal details of all eligible voters. The simplest solution in such a case is to only conduct a paper-based registration exercise. However, paper-only voter registration data may often be insufficient: the data quality may be poor, data cross-checks and efficient updates of the register are not possible, and it is difficult or impossible to establish or publish accumulated registers for larger regions or an entire country. Any claims about inaccuracies of the register will be hard to verify.
One step up the technology ladder, and usually the cheapest way to capture registration data electronically, is the use of machine-readable paper registration forms. Data can be collected by manually filling in optical mark recognition (OMR) forms, from which registration data can later be scanned into an electronic database. The disadvantages of such a system include a lack of data integrity checks when the data is captured, a lack of feedback about any problems at the point of registration, the difficulty of correcting or completing any wrong or missing data discovered during the scanning process, as well as mistakes due to incorrectly filled in forms.
Further technical upgrades include the use of electronic registration systems at the point of registration. In their simplest form, electronic registration systems are offline and only capture registrants’ personal details. Such a system may be sufficient if voters’ details can be confirmed by widespread availability of reliable personal identification documents. In such a case, local duplicate checks and data validation must be possible. Data back-up plans at the point of registration need to be extensive, as any lost data cannot be restored.
Online registration systems require data connections to all registration points. They have the potential to immediately check a registrant’s data against the entire voter register, and to detect and resolve duplicate registrations. Such systems can upload data directly to central registration databases and reduce the risk of data loss at the point of registration.
The most comprehensive registration data collection will include capturing registrants’ biometric details such as fingerprints and facial or iris recognition data. The costs of such a system will increase due to the need for biometric data-capturing equipment for every registration system as well as more comprehensive data-processing requirements for storing and comparing biometric data. The additional cost of setting up biometric systems is usually justified by the difficulty of establishing registrants’ identity due to a lack of reliable identification documents.
