The Question
This question is posted on behalf of an election official from Ghana.

My EMB is compiling a new voters list using biometric technology for elections later this year. The registration process requires the scanning of all 10 finger prints and capturing of pictures of applicants. A voter ID card will be issued in-situ.

The voter list will include pictures of voters. Stakeholders have requested that, in addition to the biometric voter registration, the Commission should procure biometric voter verification equipment for the election day. Their expectations are that this will:
i) prevent multiple voting
ii) prevent voter impersonation
iii) prevent ballot stuffing

Related to this I have the following questions:
* Are any countries using this or similar equipment?
* Can the stakeholders' expectations be met by using the equipment?
* Are there any additional benefits using this equipment?
* What are the risks of implementing biometric voter registration and biometric verification in the same electoral cycle?
 

Summary of responses
There are many countries that use, or have used, biometric voter registration technology, but if the conditions are not right, it can be far from successful. Several experts stress the need to carry out a thorough needs assessment and feasibility study before deciding to introduce biometric technology in either voter registration or verification. The various challenges and the time required to implement such technology should not be underestimated. A country’s legal framework must also be taken into account.

Some of the particular challenges associated with the introduction of voter registration biometric technology are highlighted from Uganda’s experience: time required to register voters; the calibration of the facial recognition software; and difficulty of obtaining finger prints from some people.

On the specific issue of biometric voter verification technology, one expert stresses that there is no room for error or setbacks with this, as voting takes place on a single day, unlike the months’ long process of registration. In addition, it is would be a huge challenge to introduce the two new technologies in the same electoral cycle and much that could go wrong simultaneously.  It is warned that this could risk “undoing all the good work of previous decades,” in Ghana and the expert advises to limit the target for 2012 to “limited pilot trials” with “strong fall-back and contingency plans.”

Examples of related ACE Articles and Resources
Encyclopaedia:
• Voter Registration Technologies and Materials 
• Voter Identification

External Resources
• Cooling Ethnic Conflict Over a Heated Election: Guyana, 2001-2006, Innovations for Successful Societies 
• Interview with Humayun Kabir, Secretary of the Election Commission Secretariat, Bangladesh Election Commission,  Innovations for Successful Societies 
• Interview with Peter Eicher, Independent Elections Consultant, Innovations for Successful Societies 
• Cleaning the Civil Service Payroll: Post-Conflict Liberia, 2008-2011,Jonathan Friedman, Innovations for Successful Societies
• Interview with Alfred Drosaye,  Principal Administrative Officer, Liberian Civil Service, Innovations for Successful Societies 
• Interview with Anup Mukerji, Development Commissioner, Government of Bihar, India, Innovations for Successful Societies  
• Voter Registration in Africa: A Comparative Analysis, edited by Astrid Evrensel, EISA, 2010

Names of contributors
1. Jennifer Widner
2. Rudi Elbling
3. John Maphephe
4. Matiar Rahman
5. Debashis Sen
6. Khalid Waheed
7. Charles Benjamin Nsimbi-Kabugujjo

Innovations for Successful Societies cases and interviews help share experience with biometric id in several countries.  Not all of our materials on this subject are currently posted and we can supply more information on request.  (Contact Laura Bacon, associate director, at [email protected]).

Some of the more informative discussions are about biometric id for the purpose of civil service reform, but we also have some cases that mention biometric id for elections:

Nigeria…Last week of January we will publish a case on Nigeria’s use of biometric id (stay posted; the case study is interesting and inspiring)

Guyana Election: Cooling a Heated…  (weighed biometric id and backed off)

https://www.princeton.edu/successfulsocieties/content/focusareas/EL/policynotes/view.xml?id=167

Bangladesh… interviews (we can supply more on the case itself if requested.  Case not posted)

Humayun Kabir

https://www.princeton.edu/successfulsocieties/content/focusareas/EL/oralhistories/view.xml?id=91

Peter Eicher

https://www.princeton.edu/successfulsocieties/content/focusareas/EL/oralhistories/view.xml?id=90

In the civil service (good on execution)

Liberia: Cleaning the Civil Service Payroll (discusses biometric id)

https://www.princeton.edu/successfulsocieties/content/data/policy_note/PN_id180/Policy_Note_ID180.pdf

Alfred Drosaye, Liberia, interview that discusses biometric id in civil service reform

https://www.princeton.edu/successfulsocieties/content/focusareas/CS/oralhistories/view.xml?id=339

Anup Mukerji, Bihar State, India, on biometric id

https://www.princeton.edu/successfulsocieties/content/superfocusareas/traps/CL/oralhistories/view.xml?id=159

Best wishes,

Jennifer Widner

Many countries especially in Africa are opting for biometric VR technology (Sierra Leone, Guinea, DR Congo, Ivory Coast, Nigeria) however mainly end up disillusioned since the use of this technology does by far not necessarily deliver what it promises.

None of these countries is actually known for the high quality of its voter’s list; quite the opposite all of them are of real low quality and have created lots of problems on election days.

Most of the biometric VR using countries have never performed indebt feasibility studies and don’t even know if the technology is a solution for the problem. One has to ask the question first what the problem is and thereafter find out of biometric VR is the answer to it. Often the legal framework doesn’t even exist to use biometrics. Often EMB’s who use it don’t have the capacity to maintain it after an election cycle is over.

Biometric VR is difficult to implement, specification and procurement processes are lengthy, training of operators and data processing are complex and lengthy processes putting a lot of pressure on the EMB to produce the list in time. Most countries eventually fail to implement the biometric verification of the voter which renders the entire exercise questionable plus the shaky legal grounds on which these exercises often are often taken place open ample room for complaints.

I would suggest for your EMB to do its homework and run a proper feasibility study and bit to get carried away by the perception of biometric VR being the silver bullet to all voter’s list problems because it can be quite the opposite; creating more problems than one had before.

There will be a conference on the use of ICT in election processes form 5-9 March in Mombasa Kenya, maybe you can attend and exchange your thoughts with EMBs who have done it,

cheers

Rudi/ UNDP elections Liberia

The most important step is to allow the stakeholders and management of EMB to analyse exactly, what is required and then make a decision. The technology will facilitate and support the execution of the plan which follow legal framework and country political will. The following countries may share the experiences Lesotho, Zambia, Somaliland and DRC. It is very time consuming and resource hanger to implement the solution, careful planning with alternatives is required, very expensive but it work well when supported by the stakeholders. However it cannot solve all the problems as anticipated in many boardrooms discussion, strict supervision and tight management policies needs to be monitored from time to time. Not every one IT industry can support this, therefore require an adequate budgeting to cater for highly skilled IT.

In 2008 Bangladesh Election Commission prepared Voter List with photograph. It was Biometric voter registration system. Now Bangladesh Election Commission captured about 8.6 million voters data base with biometric features, its operate by a sophisticated system. BEC using  world best server system with DRS (Des ester recovery system) for its central database.

Its essay to find and use full features biometric registration equipments.I know Ghana has a story of successful voter registration & election in Africa.

There are some important benefits like, preventing registration duplication& people ? political parties, voters trust , online verification & identification. AFIS (Automated Finger Print Identification System) is still essayist  system in the world. By this system its easy to  identify voter in polling station.

& people support.

Photo ID Cards and photo-electoral rolls together make an effective practical solution. One-to-many matching in biometric data-base is a real problem on an isolated rural polling booth on the voting day

Biometric technology is very effective in electoral processes and offcourse produce good results,

specially in democracies where the voter's lists are not fair and error free.

We, in Pakistan are preparing fresh voter's on the basis of biometrics, where picture of a voter and thumb ipmression will be the prime requirment, because in existing lists there are many duplications and triplications which made them suspicious.This is for the first time that voter's lists are being prepared on the basis of National Identity Card which is again a step towards credible,fair,error free and acceptable voter's list because this is a primary and basic document for democracy and  governance.

In developing democracies the necessity of comprehensive and fair voter lists always play a important role, as in some countries these lists are not prepared according to international standars resulting to have manipulated elections with voilence afterwards.

Biometric technology is not dificult or expensive as is considered rather it is workable and can be adopted by any EMB having independance and powers to do so. This is a practical solution of problem solving in countries where voilence errupts after elections, disturbing the stability and peace.

Election Commission of Pakistan is also working on the use of Electronic Voting Machines in forthcoming General elections and we expect fair ,free and credible Elections after Biometric voters lists and EVMs.

It may not be out of place to mention here that all this is being done and carried out in consultation with all stake holders specially political parties ,therefore, it can be model for other countries as well ,as Democracy is said to be the best form of governance in this DEMOCRATIC erra which is based on popular will and free choice and all steps for achievment of this very importanr objective ,including credible/error free voter lists can gaurantee Good Governace.

In Uganda, we have been using voters' rolls with photographs since 2001. The Commission was using Visage facial recognition software to perform duplicate analysis. Unfortunately, this has not made them particularly accurate or acceptable to all stakeholders (see Election Reports 2006 and 2011).  in preparation for the 2011 general elections, the EMB introduced full biometric registration (photo, 10 fingerprints and signature) for all new registrations done by a kit at the registration point.

The 1st challenge was time with the kits only able to enroll just 8 persons per hour (i.e just over 7 min per enrollment) or 80 persons per 10 hour day.  The average waiting time per applicant was over three hours or 23 people already in the queue.  This discouraged people who did not have 3 hours to wait.  After a 40 day exercise some 5 million new voters had been enrolled.

The next challenge was the facial recognition software that requires atleast 6 months to analyses a 1: total number match if total number is in tens on millions.  This time was not available prior to polling day.  Furthermore, the system has to be set at a lower percentage of similarity because it tends to see all brothers, sisters, sons, cousins and certainly twins as duplicate registrations.  this the requires human intervention to correct further slowing the process.

Finally, fingerprinting often fails particularly with persons working with their hands, e.g farmers, mechanics, carpenters, metal workers and other artisans.  this is due to either the ridges wearing down or dirt/oil filling the troughs.  consequently through washing of hands may be required prior to enrollment.

As said above, a pilot study is of at-most importance prior to commencement of the exercise to ascertain the feasibility and/or usefulness of the biometric enrollment both in time and resources expenditure.

Good luck

Charles

Posted on behalf of Krishnamurthy T.S., former Election Commissioner of India.

In India, there is a school of thought wanting biometric registration of voters to prevent bogus impersonation. There are many illegal immigrants, running already into millions. It is better to have a system of registering citizens and non-citizens to ensure that only citizens vote. I too agree with this school of thought. It is a mammoth task. I believe it is desirable especially to prevent illegal immigrants turning into citizens on the basis of residence/ration card/collusion.
While in office, I came across a company from abroad canvassing a method of verifying the voters after biometric registration by using something like a pen torch which will identify whether a voter carrying  biometric card is genuine or not. Electronically it is possible (Swiss has some technology) although it has to be tested.

I am of the view that with the population increasing everywhere and with electoral fraud gradually increasing technology needs to be increasingly deployed to achieve "Free and Fair Elections"

Best wishes

TSK

.... offers a comprehensive introduction to the single most complex process within the electoral cycle. It critically analysis the efficacy and sustainability of different voter registration systems across the African continent. The first part of the book provides an overview of different voter registration methodologies, including combined civil and voter registration, periodic versus continuous registration and active versus passive registration. It identifies guiding principles for voter registration and introduces the reader to the latest technological developments in the industry, such as fingerprint and face or iris recognition. The challenges of using biometric technology in harsh African conditions are highlighted, and the responsibilities of national election management bodies, international donors and other decision makers in this million dollar business are critically examined.

The second part of the book offers detailed descriptions of the voter registration systems used in eight countries, namely – the Democratic Republic of Congo, Liberia, Malawi, Mozambique, Ghana, Rwanda, Senegal and South Africa. Some have established sustainable and effective systems that provide numerous best practice recommendations, while the lessons learned from those countries with less successful registration exercises are invaluable. A free download is available at  http://www.eisa.org.za/PDF/vrafrica.pdf.

I hope it will provide you with answers to many of the named questions above.

kind regards,

Astrid

First, let me endorse Astrid's suggestion to read the EISA VR Africa document. What is said about technology and voter registration translates very well to the proposed biometric voter verification concept. Remember that you have only one election day - VR can take months and takes place when the political temperature is lower. VR efforts can be extended or repeated. You get only one shot at polling-day technology.

I believe Ghana is in danger of over-extending itself. Ghana has s solid record in the effective use of technology in election management. However, there are major risks associated with the introduction of not one but two new technologies right across the entire country in the same election cycle. If the elections in question were 24-36 months away, your chances of success might make the risk worth taking. But, to judge by your questions, you are only at the conceptual stage -  analysing comparative experiences, seeking the advice and opinions of others in the field - all proper actions to take, but not within 12 months of the election in question. I believe the ambition level for 2012 should be, at best, limited pilot trials, preferably from several potential vendors, with strong fallback (i.e. paper) and contingency plans. To roll out biometric digital registration kits and polling-station level biometric voter verification systems in the same year across the entire country runs the risk of undoing all the good work of previous decades.

I don't agree with Khalid that biometric technologies are "not difficult or expensive". The Pakistani experience is at the national identity level - not in the voter registration centre or polling station. Putting technology in the field is a whole different ball game than putting it in the data centre.

Nor do I agree with former Commissioner Krishnamurthy of India. It is a common misconception that biometrics can weed out non-citizens from citizens. If a non-citizen is in possession of an identity document and the official conducting voter (or, for that matter, national identity) registration accepts that document and enrolls the non-citizen, there is nothing that biometrics can do. All biometrics bring to the table is uniqueness - i.e. they can ensure that the non-citizen registers only once. 

Charles raises valid points about throughput and the problems of failure to enrol fingerprints - remember that this can reach 20% in some cases. He echoes others in suggesting that more careful feasibility and pilot studies are needed. I endorse this.

If you decide to proceed, I wish you luck - you will have a very, very busy 2012!

Ronan

In some Muslim countries may not want to use biometric technology and do not like applying the indelible ink because they want their bodies clean for daily pray.  The reason that Thailand gave up this process because they did not really check the print mark when they found irregularity, cheating problem or multiple votings.  The quality of the ink was not good, difficult to see the mark. We had this process 10 years ago but now we just ask the voters to just signed their name at the counterfoil of ballot papers for prevention multiple votings.

Somsri

ANFREL