Alert Sign Dear reader, online ads enable us to deliver the journalism you value. Please support us by taking a moment to turn off Adblock on

Alert Sign Dear reader, please upgrade to the latest version of IE to have a better reading experience


Cost-benefit ratio of voter verification exercise

December 18, 2012

Objectionable votes will be dispatched to Nadra for verification through thumbprints. – File photo
Apparently in absence of the over-riding command, computers assigned 38,200 voters to the constituencies of their permanent address, claims Nadra.– File photo

Shrugging off allegations that their decision was Karachi-specific and strikes at the heart of Election Commission of Pakistan’s autonomy enshrined in the Constitution and therefore unlawful on both counts, the Supreme Court of Pakistan has directed the Commission (ECP) for 100 per cent verification of the Final Electoral Rolls-2012 through a door-to-door exercise across Karachi. How history will judge the trigons of this judgement only time will tell. However, taking advantage of the decision to highlight some pertinent aspects of the debate needs urgent scientific analyses of the issue.

ECP started drafting the subject electoral rolls through nationwide physical verification of ‘Final Electoral Roll-2007’. The verified data was later cross-checked and augmented by the ‘National Database and Registration Authority’ (Nadra) using records of computerised national identity cards (CNIC) that they started 12 years ago.

As per FER-2007 the number of eligible voters in Karachi was 6.634 million. Out of these, ECP was able to physically verify 59 per cent or 3.913 million voters, leaving 2.721 million records as ‘unverified’. Nadra cross-checked and verified part of unverified records, as well as further added voters through their CNIC records, thus augmenting verified voters by 2.890 million. This brought the total number of voters in Karachi to 6.852 million; registering an increase of 217,710 voters over the registered voters for 2008 elections. While taking care of ‘ghost voters’ this may have been the root cause of what is known as ‘undercount’ by statisticians – a problem that seems to be racking the whole system. Let me explain how:

The first five digits of a CNIC number represent the district of our permanent residence. The last ‘odd number’ stands for male cardholder and ‘even number’ for females. Seven numbers in between are randomly generated for individual records. Apparently in absence of the over-riding command, computers assigned 38,200 voters to the constituencies of their permanent address, claims Nadra.

However ECP insists that for the above reason alone they rigorously advertised Draft Electoral Rolls-2011 offering all eligible voters an opportunity to register at the constituency of their choice. A massive publicity campaign was carried-out inviting citizens to use ‘SMS-8300 Service’ for confirming their vote with hopes to minimise error and help rectification.

The 1998 nationwide census is generally considered non-controversial and statistically a sound exercise. Let us compare 6.852 million votes to the projected population based on last census. Children above the age of four in 1998 have become eligible voters. On that basis factored with annual mortality rate for last fourteen years should yield 7.716 million eligible voters indicating a gap of 864,000 for Karachi. However 1998 population of 132 million if projected at the nation’s growth rate of 2.69 per cent per annum for the last 14 years, the number of people should be 196 million but the Population Census Organisation’s population counter show only 181 million suggesting a certain decline in the population explosion of the country.

The ECP was able to verify only 54 per cent of voters against FER-2007 across Pakistan or 59 per cent of voters against FER-2007 at the country’s second largest metropolis Lahore. The number of votes augmented by Nadra for most other cities remains proportionately much higher then Karachi or Lahore.

In pursuance of the Supreme Court directive, ECP is now considering using 13,000 teams for the verification exercise in Karachi, each with one enumerator and one soldier. ECP is also proposing to conduct door-to-door verification after office-hours hoping to ensure the presence of the “man-of-the-house”.

It is an open secret that once known as ‘city of lights’ Karachi is very hostile at present. Men in uniform feel jittery while performing their daylight duties together even in respectable numbers. With violence at its peak it would be condemning them to gallows, forced to do rounds in one-plus-one configuration in most violence-prone areas of Karachi.

According to Yale PhD, Dr. Farooq Naseer currently teaching economics and statistics at LUMS “a sampling rate of 5-10 per cent may be enough (for voters’ verification in Karachi) but the statistical design of such an exercise would require careful work with expert input”

This scribe tends to agree when Dr. Farooq further points out that “Complete enumeration is not without problems! Unfortunately, the error is unlikely to drop to zero even if one were to conduct a complete census enumeration. There is a phenomenon called census undercount in the decennial US Census whereby the US Census ends up systematically under-counting some individuals/groups. The matter has led to litigation there and provoked much serious debate among the professional statistical community on the best statistical “adjustment” to fix the undercounting error”.

The benefits of over 40 per cent sample are considered to be none or at best negligible. This debate could also be viewed here.

One does however wonder if the wise-men actually demanded ECP to send verification teams to every house or the ECP was required to carry out door-to-door exercise to ensure 100 per cent verification of the eligible voters in Karachi. Both routes may not be the same but could yield the same result if carried out scientifically utilising established statistical norms.

Probably a mindset from Musharaf era, the ECP is not even the actual custodian of ‘final electoral rolls’, Nadra maintains that data close to its chest. To ensure transparency and inclusion demanded by the Supreme Court, the ECP would be well advised to try and make FER-2012 electronically available on its website as well as widest possible dissemination of electoral rolls through electronic distribution over CDs. ECP could also camp at 18 Town offices of Karachi to make ‘voters inclusion process’ transparent and as easy as possible and may be the same could be done in other parts of the country - an exercise Supreme Court has placed no bar over it.

The author is a social activist, a member of Citizens for Democracy and the former Administrator of Karachi.