Political and election administration leaders on both continents seem to at a minimum express a desire for biometrics and in some cases have infused, or are in the process of infusing, their registration and documentation systems with biometrics.

Although biometrics can help address the problem of multiple registrations and thus in that respect increase trust in the legitimacy of the process, it is unclear what precisely is driving such a strong and pervasive move towards biometric technologies in registration and ID cards across Africa and Latin America at this moment. Biometric technology is expensive, requires training and sufficient staff to implement, and necessitates that all the citizens at some point re-register and/or get the new card. At the same time it only addresses one aspect of potential fraud: multiple registrations. Multiple registrations may or may not be the gravest problem confronting a country’s electoral system.  Moreover, multiple voting, which is the true concern, can be addressed through low tech alternatives such as use of indelible finger ink on Election Day and greater transparency and wider spread publication of the registration list in advance of elections.

Moreover, launching a new biometric system, because it can be so complex and vulnerable to so many different variables, such as environmental and geographical challenges, lack of training and skills, mechanical breakdowns, delays, and usually necessitates to re-registration of every citizen in the country within a finite period of time, can cause disenfranchisement. As Michael Yard has pointed out, “A new way of doing things should not be judged either as good or bad simply because it is new, but should instead be evaluated upon the basis of whether it helps to make elections more or less democratic given the resources, risks, and alternative solutions available.”[i]

Are politicians truly moved by the desire to address duplicate registration and voting or is there some other political interest that drives them? What outside pressures exist? Perhaps technology companies or the international donor community have taken action that is facilitating or accelerating the race to biometrics?  There does not seem to be any strategy, cost/benefit analysis or long term thinking when it comes to this subject among most of the parties involved.  It is still unclear whether biometric technologies are the most efficient use of resources given the particular problems individual countries in Africa and Latin America confront. In some cases increased use of biometrics may make sense; in others it may address a less than urgent problem and only make the process more complicated. These issues require further exploration. Although there have been a small handful of technical studies of biometric systems, little research has been done on the advantages and disadvantages of deploying such a system in the first place.  It appears that the only work that has addressed this issue head on is the report Voter Registration in Africa, an edited volume published by the Electoral Institute for the Sustainability of Democracy in Africa (EISA).

Richard Atwood, writing for the European Parliament, has stated, “EMBs and donors like the EU should… take care that new technologies adopted to prevent registration fraud – like biometric technology – do not themselves lead to citizens’ disenfranchisement...For example, due to the cost of biometric kits, countries may purchase fewer and rotate them round the country, which can reduce the time available for communities to vote and may increase the distance they have to travel to do so. Note also that biometrics – if they work -only resolve multiple registration. They can’t resolve issues of identity…Any erosion in the EMB’s integrity through its procurement of costly equipment will increase risks of conflict and outweigh any of technology’s potential benefits.” [ii]

Moreover, no matter how sophisticated a system is used, if there is lack of faith in the management body responsible for implementing it, it will not help increase trust in the integrity of the elections.

There are countries that have implemented biometric systems relatively successfully, but at the same time one must ask at what financial cost in the immediate and as important the long term in maintaining and upgrading the systems going forward, at the sacrifice of other needs.

There are countries in both regions where biometric technology is being used to varying degrees. Countries that have or are moving toward biometrics include Kenya, Bolivia, Cameroon, Colombia, Cote d’Ivoire, Ghana, Guatemala, Venezuela, and Uganda.

a. Foreign Vendors

Procuring biometric registration technology from foreign vendors can be fraught with potential difficulties, particularly if they then hold the technological expertise and capacity to maintain and improve the systems to the exclusion of the election commission and local technologists. “If it is not possible to retain adequate in-house staff, the EMB should carefully consider whether there will be adequate funds to pay for continued support from external vendors, and whether fostering continuous dependency on external vendors is in the best interest of the EMB."[iii] If a government contracts with a foreign vendor, depending on the state’s internal capacity, “the system may be difficult to use in future registration exercises when donor participation in the electoral process is likely to dwindle.” [iv]

According to the EISA report, in the DRC for example, external stakeholder influence on the process was very high.

This was clearly manifested in the following activities:

• The tender process was handled by IAPSO Copenhagen.

• The number of foreign technical assistants used was high, turning the process into a veritable international affair.

• The decision on the choice of the system used was not mooted by the CEI alone. There were reports that some of the major donors at the time (like the European Commission) had clear preferences when it came to technological choices…[v]

EISA further points out that

It may be difficult for some EMBs to analyse and fully understand study results

presented by vendors who are eager to sell their products. Simple studies are

good for obtaining a rough overview of a system’s performance; however,

more in-depth studies are crucial before a system is deployed to ensure that real

performance will meet EMB requirements and that the system is fully operational

in specific country conditions...Case studies conducted in an air-conditioned office using college students are not a realistic representation of how the voter registration system would work in extreme African conditions.[vi]

b. Implementation Problems

As Astrid Evrensel writes in the EISA publication’s introduction,

Even international donors and technical electoral advisors seem to support the use of biometric-based voters’ rolls and other high-tech solutions to deal with integrity issues and possible political challenges to elections. A registration process that uses sensitive high-tech equipment not only adds significant ‘integrity’ costs to the core costs but also increases organisational and logistical challenges. These include the increased need for technical training as well as continuous supervision and support for registration staff in the field to ensure that the data is captured, collected and processed to the highest possible standard. If EMBs lack organisational and logistical resources while attempting to organise such a complex task, the resulting voters’ roll can be replete with errors. If the voters’ roll is too faulty, polling staff may not even use it on election day. Following generally accepted principles, the voters’ roll should be as accurate and complete as possible to provide for maximum inclusion of all groups. Voter inclusion and the enfranchisement of disadvantaged groups have to be balanced against the security features of a system, which prevent double registration by technological means…[vii]

Echoing Atwood, Evrensel also points out that “Travelling long distances to registration centres disenfranchises thousands of voters especially women, people with disabilities and others who cannot journey easily for whatever reason.”[viii]

There is also the problem of the false positive and the false negative. “A person is falsely rejected, for example, when s/he is already enrolled in the database but the system fails to positively identify her/him. The false reject rate depends on the situation under which the system is used, such as operating conditions and user cooperation. A slight change in one’s fingerprint due to dust, moisture or other environmental conditions can increase the false reject rate. A person could be falsely accepted, for example, when one person’s biometric data is similar to another’s and a match is made erroneously.”[ix]

There are also those citizens for whom recording a fingerprint is not possible and mechanical failures that lead to an inaccurate or unclear recording. “In Zambia in 2010, for example, a ‘failure to acquire’ rate of over 10% was observed at field level in the system’s attempt to capture the digital fingerprints of voters. In other words, even before any further processing of voters’ data, already 10% of voters are excluded from the system’s attempts to detect double registrations because of technological registration issues…”[x]

In a sad example in Kenya recently, an 82 year-old man was turned away after a BVR kit failed to detect his fingerprints. IEBC clerks at the station tried their best several times but failed to register Mr Ole Nzuka, even after he washed his hands. Mr Nzuka who had trekked for more than 10 kilometers from Kibiku area to register in Ngong as a voter told The Standard, that he was under medication yet he failed to register.”[xi]

EISA further argues that

[T]here are numerous ways in which an election can be stolen, and devoting massive resources to advanced registration techniques – even if they are well implemented and work perfectly – may simply displace fraudulent activities into other areas of the process. Electoral authorities and donors are attracted by the opportunity to make improvements to their systems rather than to address the more sensitive and difficult problems of misconduct by politicians, political parties and voters. These actors may in fact be happy to see the EMB and donors pursue a course of action which, in reality, would not disrupt their entrenched activities…[xii]

c. Examples

One example of biometrics gone awry is Fiji. Michael Yard writes for IFES that

In 2009 there was a strong push in Fiji to include automated fingerprint identification as a requirement for every person on the voter list. Responses to the question of how this would improve voter registration were wide-ranging, with expectations that it could eliminate impersonation at the polling station, remove deceased persons from the voter register, and prevent registration by non-citizens and underage voters. None of these issues were addressed in any way by the proposed technology. A survey of political party members and field election staff indicated that there was very little concern about persons registering more than once, the one issue that the proposed system could help to address. … Identifying the problem is not always a simple exercise. Surprisingly, the aforementioned survey also indicated there was minimal concern about impersonation, deceased persons, underage voters or non-citizens. Even the largest city on the islands was small enough that poll workers knew almost every voter who came to cast a ballot. The biggest problem raised in stakeholder discussions was suspicion that some parties were bussing voters from polling station to polling station, allowing multiple voting – a suspicion that, according to most observers, was completely unfounded and was only used by losing parties to cast doubts on the legitimacy of the ruling party. So, the main problem was really a lack of trust, which could have been more effectively addressed by voter information, publicizing the controls that were already in place to prevent bussing in voters.[xiii]

In the Democratic Republic of Congo, the biometric registration exercise reportedly cost $58 million.  In a case study of the Congo, it was found that “US$40,160,000 was needed to buy the 10,000 biometric registration kits and to have them transported by air from Brussels to Kinshasa. Apart from these initial costs, the CEI would have to raise a yearly or second yearly budget of some US$2.5 million1 for equipment maintenance and replacement of lost or damaged equipment to ensure the continued use of this technology.”[xiv] And yet it still resulted in disenfranchisement. “[T]he system was not able to capture most people because the number of units acquired was not sufficient to cover the entire country (especially the outskirts), and people who had to travel several kilometres to register either missed the chance or found it too difficult to make the effort.[xv]

In 2007 the election commission in Guatemala (RENAP) decided to completely overhaul the system using biometrics. This overhaul required all Guatemalans to re-register and get new identity cards to replace the old ones. In a report published in advance of the 2011 election, the International Crisis Group (ICG) described this process as having been fairly disastrous:

The exercise was overambitious from the start. Planners misjudged the time required to retrieve and sort old data from municipalities and then collect and verify new data from around fourteen million Guatemalans. The bad design was compounded by corruption. Deputies in Congress and RENAP’s managers, who have since been replaced, reportedly awarded contracts inappropriately, or at least without sufficient transparency. Ill-qualified cronies were appointed to key positions. The combination of poor design, nepotism and incompetence led, unsurprisingly, to a deficient registration. The exercise has cost far more than originally projected but remains incomplete. Politicians, especially from opposition parties, and some in civil society have accused RENAP of inflating registration numbers so as to benefit the ruling party.[xvi]

Sierra Leone used a biometric system for the first time in their 2012 election.  Yet “the integrity of the electoral register has still been questioned by opposition parties who accuse the government of covertly registering individuals under the legal voting age and foreign nationals…In Sierra Leone, there is no historical evidence of a deliberate strategy by any political party to rig elections through multiple registrations. All previous electoral registers have erroneously contained names of the deceased, the under-age and foreign nationals. But the most significant type of electoral misdemeanor has been physical stuffing of ballots and false recording of results by temporary election workers. Both the All People's Congress and Sierra Leone People's Party, when in power, have at times used their position to fund political campaigns and buy voters. This practice remains widespread. Political parties continue to organize and condone the intimidation of voters, often perpetrated by their youth wings. Biometric technology offers little scope to tackle these transgressions.”[xvii]

Kenya started its 30 day process of biometric voter registration exercise on November 19, making it an interesting current case study.  Reviews so far have been mixed.  The Star newspaper says that , “According to the contract, the price quoted by the Canadian Commercial Corporation for the 15,000 [Biometric Voter Registration] kits was Sh6.1 billion (€56.21 million).[xviii]

Some reports have focused on the number of Kenyans who do not have the identity cards necessary to register. One press report stated that four million Kenyans lacked ID cards and that only a small percentage of them had even applied for the card.[xix]

Early on in the process there were problems with the passwords on the registration kits expiring, the lack of sunlight causing solar power failures in the kits’ batteries, and problems with transportation and logistics.[xx]