Biometric verification: not a silver bullet
With a year to general elections, federal lawmakers are in a race to decide the fate of the new Elections Bill
The writer is a researcher on parliamentary and electoral affairs and a Chevening scholar. He tweets @dnananjum
With barely a year to go for general elections, federal lawmakers are in a race against time to decide the fate of the new Elections Bill 2017. Unless this is handled properly Pakistan could soon find itself meandering into another political impasse in 2018. Any delay in the passage of the new law is likely to have an impact on political parties as well as election authorities and voters. Even though a multi-party committee comprising 34 legislators from both sides of the aisle has been deliberating upon these electoral reforms since 2014, the draft legislation seems to be unpopular with certain political quarters.
Several opposition parties have raised serious concerns about the inclusion of their demands in the recent bill finalised by the Parliamentary Committee on Electoral Reforms. The Pakistan Tehreek-e-Insaf (PTI), for one, remains adamant on the use of electronic biometric verification machines during the upcoming elections.
The system is already being used in Pakistan for electoral purposes in the wake of the 2010 amendments to the legal framework using NADRA’s CNIC records. These contain an electronic photograph and fingerprint, which are exported to the electoral rolls allowing NADRA and the Election Commission of Pakistan (ECP) to prevent duplication and conduct stringent authentication of eligible voters. However, the system is not fully in practice as proper verification of voters is carried out through electronic fingerprint scanners at polling stations. In the 2013 elections, election officials carried out manual capture and analysis of fingerprints but it proved ineffective. Ever since, there have been demands of electronic technology for biometric verification of voters.
As the preeminent supporter of electronic biometric verification (EBV), the PTI has advocated the elimination of the ‘human factor’ from election operations. In an effort to modernise, such expectations can trigger more complex election issues. It is often a mistake to expect technological solutions to curb corrupt practices. Advocate for it tend to overstate the possible benefits. Technology can perhaps change the nature and methodology of corrupt practices, but not corruption itself. Considering Pakistan’s existing electoral infrastructure and the inherent electoral challenges, it is difficult to properly assess the benefits and costs of these machines. They are dependent on an array of factors including electoral infrastructure, type of databases already being used, readily available voter identity system, the voter registration process as defined in the elections laws, logistics, communication and energy infrastructure, security affairs, and human resources and its capacity. For a cost-analysis of EBV, it is important to include the affiliated costs with these machines, eg, the cost of training staff on the use of this technology, pilot-testing, required modifications to existing IT systems, security management, maintenance, storage, etc.
According to the recent voter registration information available on ECP, as many as 97 million citizens are registered. In order to make EBV fully operational and the biometric databases to detect and avoid duplication of voters, the IT solution must have the capacity to process information of each of the fingerprints of 97 million citizens. This implies that these registered voters have to be entered in the database and the system will have to check each print against each finger making 96,999,999 permutations. This number exponentially increases if the system has to check multiple fingers. This process will only be effective, if conducted, using complex and sound computer machines. For the effective and efficient use of such complex technologies, extensive training of polling staff will be required. However, it is important to mention that vendors of these machines give presentations in a controlled environment, which creates a perception that such devices are easy to operate. On the contrary, the actual situation may differ and may not be as smooth on election day.
In the majority of countries where EBV machines were introduced, either the devices crashed or the polling staff failed to operate them. Countries such as Venezuela and Kenya stopped using them as voters were convinced that their vote was not kept secret, and in the case where the machines altogether stopped functioning infrastructure and training issues were at fault. In Brazil, these EBVs were gradually introduced, and it took them more than 10 years of pilot-testing and trials to install them during elections.
Another major issue is that these machines are often difficult to use by differently abled people and by those whose biometric features are impaired, such as the elderly whose finger ridges are often damaged or not recognised. However, the use of such technologies also restricts the opportunities for observation. Digital records cannot be observed and the audit options are limited. Data transfers are even more complex, which again cannot be observed by polling agents, media, election watchdogs, etc.
In a nutshell, if the electoral stakeholders confirm the need for EBV after successful assessment of costs and pilot-testing, the ECP could gradually implement the system. However, keeping in view its pros and cons, parliamentarians must take an informed decision before approving such technologies as they may add to Pakistan’s existing problems. Policymakers should carefully examine the affiliated costs and other potential issues as they certainly outweigh the benefits. These machines are definitely not a silver bullet. Should these machines be put to use nonethless, the commission must not hurriedly install and use them for the 2018 general elections. Proper tests and gradual introduction of these machines would be the key. Otherwise, another disaster may unfold.
Published in The Express Tribune, September 1st, 2017.
Several opposition parties have raised serious concerns about the inclusion of their demands in the recent bill finalised by the Parliamentary Committee on Electoral Reforms. The Pakistan Tehreek-e-Insaf (PTI), for one, remains adamant on the use of electronic biometric verification machines during the upcoming elections.
The system is already being used in Pakistan for electoral purposes in the wake of the 2010 amendments to the legal framework using NADRA’s CNIC records. These contain an electronic photograph and fingerprint, which are exported to the electoral rolls allowing NADRA and the Election Commission of Pakistan (ECP) to prevent duplication and conduct stringent authentication of eligible voters. However, the system is not fully in practice as proper verification of voters is carried out through electronic fingerprint scanners at polling stations. In the 2013 elections, election officials carried out manual capture and analysis of fingerprints but it proved ineffective. Ever since, there have been demands of electronic technology for biometric verification of voters.
As the preeminent supporter of electronic biometric verification (EBV), the PTI has advocated the elimination of the ‘human factor’ from election operations. In an effort to modernise, such expectations can trigger more complex election issues. It is often a mistake to expect technological solutions to curb corrupt practices. Advocate for it tend to overstate the possible benefits. Technology can perhaps change the nature and methodology of corrupt practices, but not corruption itself. Considering Pakistan’s existing electoral infrastructure and the inherent electoral challenges, it is difficult to properly assess the benefits and costs of these machines. They are dependent on an array of factors including electoral infrastructure, type of databases already being used, readily available voter identity system, the voter registration process as defined in the elections laws, logistics, communication and energy infrastructure, security affairs, and human resources and its capacity. For a cost-analysis of EBV, it is important to include the affiliated costs with these machines, eg, the cost of training staff on the use of this technology, pilot-testing, required modifications to existing IT systems, security management, maintenance, storage, etc.
According to the recent voter registration information available on ECP, as many as 97 million citizens are registered. In order to make EBV fully operational and the biometric databases to detect and avoid duplication of voters, the IT solution must have the capacity to process information of each of the fingerprints of 97 million citizens. This implies that these registered voters have to be entered in the database and the system will have to check each print against each finger making 96,999,999 permutations. This number exponentially increases if the system has to check multiple fingers. This process will only be effective, if conducted, using complex and sound computer machines. For the effective and efficient use of such complex technologies, extensive training of polling staff will be required. However, it is important to mention that vendors of these machines give presentations in a controlled environment, which creates a perception that such devices are easy to operate. On the contrary, the actual situation may differ and may not be as smooth on election day.
In the majority of countries where EBV machines were introduced, either the devices crashed or the polling staff failed to operate them. Countries such as Venezuela and Kenya stopped using them as voters were convinced that their vote was not kept secret, and in the case where the machines altogether stopped functioning infrastructure and training issues were at fault. In Brazil, these EBVs were gradually introduced, and it took them more than 10 years of pilot-testing and trials to install them during elections.
Another major issue is that these machines are often difficult to use by differently abled people and by those whose biometric features are impaired, such as the elderly whose finger ridges are often damaged or not recognised. However, the use of such technologies also restricts the opportunities for observation. Digital records cannot be observed and the audit options are limited. Data transfers are even more complex, which again cannot be observed by polling agents, media, election watchdogs, etc.
In a nutshell, if the electoral stakeholders confirm the need for EBV after successful assessment of costs and pilot-testing, the ECP could gradually implement the system. However, keeping in view its pros and cons, parliamentarians must take an informed decision before approving such technologies as they may add to Pakistan’s existing problems. Policymakers should carefully examine the affiliated costs and other potential issues as they certainly outweigh the benefits. These machines are definitely not a silver bullet. Should these machines be put to use nonethless, the commission must not hurriedly install and use them for the 2018 general elections. Proper tests and gradual introduction of these machines would be the key. Otherwise, another disaster may unfold.
Published in The Express Tribune, September 1st, 2017.