The most pestering question about the general election 2018 for the last one year has been whether it will be held at all, rather than who will win it. Among the many reasons behind this uncertainty were constitutional compulsions arisen out of the population census conducted in early 2017, after 19 long years. Redrawing constituency boundaries – known as delimitation in election terminology – is a constitutional requirement that needs to be fulfilled once the census results have been publicised. But the timeline that the Pakistan Bureau of Statistics announced this time round for publicising those results left no time to complete the protracted exercise of redrawing constituency boundaries and to hold elections – as per another constitutional requirement – before the end of July 2018. The bureau stated that it would only publish provisional census results by August last year and would require almost another year to finalise them.
Even from a purely electoral perspective, it was obvious after the census that delimitation was urgently required in order to take into account the population concentration map of the country that had significantly changed since the last population count. The provisional census results showed that Pakistan’s population had jumped from 132 million in 1998, when the previous census was carried out, to 208 million. The data revealed that population growth rates were higher in some provinces than in others. Even within provinces, different districts had different population growth rates. Urban areas across all provinces had grown at a higher pace than rural ones.
These differences were important electorally since provinces and districts are allotted seats in the national and provincial assemblies proportionate to their shares in population. Holding an election with the existing constituency limits would not only have lacked legitimacy but could have easily become controversial for not being representative of the current demographic profile of the country.
The catch-22 that the Election Commission of Pakistan faced was that there was no provision in the Constitution to carry out the delimitation exercise on the basis of provisional census results, but waiting for the final results meant that the new constituencies could not be drawn in time for the 2018 elections. The problem was only solved when, after some initial prevarication, Parliament amended the Constitution, allowing a one-time exemption for redrawing constituency boundaries using provisional census results.
However, the question remained as to whether the Election Commission of Pakistan would be able to complete the delimitation process in time for the 2018 elections. An additional factor that made the delimitation exercise more drawn out than before was that it was to be carried out for the first time under election rules duly devised by the election commission and approved by the government in strict accordance with the provisions of the Election Act, 2017. In the past, election authorities had total freedom in interpreting legal provisions regarding delimitation as they pleased.
General Yahya Khan’s Legal Framework Order divided the four provinces, of what was then West Pakistan, into 138 National Assembly constituencies. The 1973 Constitution, enacted after the separation of East Pakistan, raised the number to 200 and decided that these would be distributed among the federating units in proportion to their population. The constituencies were accordingly demarcated in 1975-77 after the passage of the Delimitation of Constituencies Act, 1974.
Then came General Ziaul Haq. Soon after taking over power, he arbitrarily decided that some provinces needed to have more representation. He raised the total number of National Assembly seats to 207, giving four more seats to Balochistan and three more to Sindh — including two to Karachi. The new seats were added by a presidential order — in disregard of laws and without following the processes given in the Constitution.
This capricious act by the general set the pattern for redrawing constituency boundaries over the next 40 years. Election authorities started regarding delimitation as their prerogative rather than following a rule-based process open to public scrutiny.
This new approach was embodied in subsection 10-A — added by Zia to the Delimitation Act and given constitutional cover through the 8th Constitutional Amendment in 1985. The subsection read: “the [Election] Commission may, at any time, of its own motion, make such amendments, alterations or modifications in the final list of constituencies … or in the areas included in a constituency, as it thinks necessary.”
In practice, the military government itself hurt the delimitation process more than what the subsection could do. Zia took some random decisions when the first delimitation exercise was carried out before the party-less general elections in 1985. For one, he declared that the constituency boundaries were not to be redrawn in accordance with the latest census that was conducted in 1981. Through a decree, he replaced the word ‘census’ with ‘1972 census’ in the Constitution in order for delimitation to happen on the basis of 13-year-old population data. For some odd reason, he did not want the provincial shares of the National Assembly seats changed in accordance with demographic changes.
When Ghulam Ishaq Khan took over as President of Pakistan after Zia’s death in an air crash, he removed this aberration through an ordinance before the next delimitation exercise was carried out in the wake of the 1988 elections. But just before the new constituency boundaries were to be made public, he realised that the share of the Federally Administered Tribal Areas (Fata) had gone down from eight to five and that Punjab, Sindh and Khyber Pakhtunkhwa (then known as North West Frontier Province) was given one extra seat each.
Members of the National Assembly elected from Fata were important for the political chessboard Ghulam Ishaq Khan was laying in the country at the time. They would join or ditch any government on the first signal from the right quarters throughout the 1990s. They, therefore, held high regard for those who wanted to keep the democratically elected governments on a tight leash. So Ghulam Ishaq Khan issued an ordinance and froze the number of National Assembly seats for Fata at eight. This effectively meant that the number of seats for each of the four province remained the same as it was in the 1985 election. Ghulam Ishaq Khan would later also use subsection 10-A to favour his hand-picked caretaker prime minister Ghulam Mustafa Jatoi by jerrymandering his home constituency in central Sindh’s district of Naushahro Feroze before the 1990 elections.
The pattern of non-elected governments overseeing the politically fraught process of delimitation continued for at least one more round.
Before the results of the census conducted in 1998 could be used to redraw constituencies, the army overthrew the elected government of the Pakistan Muslim League-Nawaz (PMLN) in 1997. Pervez Musharraf, who replaced Nawaz Sharif as the chief executive of Pakistan, was subsequently given three years by the Supreme Court to hold elections. The judges also gave him the power to change the Constitution, including the power to change the delimitation process.
Musharraf used that power to raise the total number of National Assembly seats from 207 to 272. Seats in all provincial assemblies were increased likewise. This gave the general the opportunity to carve out 65 new National Assembly constituencies by combining, dividing and subdividing the earlier constituencies as he pleased.
Critics pointed out the elections would be too expansive, and also expensive, in larger constituencies. On the other hand, voters in smaller constituencies would become more ‘representative’.
Since the Constitution was not in force, the delimitation was carried out under the Conduct of General Elections Order, 2002 by an election commission appointed by Musharraf himself through the discretionary powers he enjoyed under another decree, the Election Commission Order, 2002.
Though the election authorities maintained a semblance of transparency while carrying out the delimitation, there was no gainsaying the fact that Musharraf had the final say in it. To cite one instance of his power to manipulate the process, he fixed the number of National Assembly seats for Fata at 12, almost double of what the region would have had according to the census results.
It is in this background that the Election Act, 2017, passed by Parliament in October last year, becomes a seminally important piece of legislation. Besides bringing some other improvements in the conduct of elections, it has ensured the return of a rule-based delimitation process. The Election Commission of Pakistan, too, has tried its best to implement the new law both in letter and spirit and devised elaborate mechanisms for delimitation for the first time in our history.
The Parliament was also diligent in removing constitutional obstacles that arose later in the path of delimitation. One of these was related to the Constitution’s Article 50(3) which allocates seats to constituent parts of the federation. The provisional figures thrown up by the 2017 census demanded that Punjab’s share in the National Assembly seats be reduced by seven, and four of these be given to Khyber Pakhtunkhwa, two to Balochistan and one to the federal capital.
The seats set to disappear were concentrated in Punjab’s central districts that are a power bastion of PMLN which was in government at the time. This implied that the net impact of the delimitation would be negative for the party and positive for Khyber Pakhtunkhwa ruled by its arch rival Pakistan Tehreek-e-Insaf (PTI). Showing political maturity, however, PMLN opted to become a part of the solution and not that of the problem. It led efforts to amend the Constitution quickly to facilitate a timely delimitation which – after the one done before the 1977 elections – was to be the first carried out under a democratic dispensation.
These noticeably positive developments would not guarantee that there were no controversies later on. The reason was simple: the lacunae of a law become known only after it is implemented; it was difficult at the time of drafting and the passage of the amendments to envision what those problems could be.
Secondly, any changes in electoral space always create winners and losers since they take some space from one political stakeholder to give it to another. Some negative reaction to them is unavoidable.
One of the early problems that the delimitation process faced was the number of Fata seats. A parallel legislative initiative to bring Fata into the national mainstream was halfway through Parliament when the delimitation-related amendments were being passed. Any change in the number of those seats could have resulted in stalling consensus on Fata’s merger with Khyber Pakhtunkhwa. So the number of Fata seats was again fixed at 12 — double the region’s share in the population and contrary to the provisions of subsection 5 of the Constitution’s Article 50 that allocates the National Assembly seats to every region/province commensurate with its population.
Fata was merged into Khyber Pakhtunkhwa with the passage of the 25th Constitutional Amendment on May 31, 2018, the last day of the tenure of the previous government. This amendment has reduced Fata’s share in the National Assembly to six seats but this change will not come into effect for this election.
The Election Commission of Pakistan started its delimitation work early in January this year and published the preliminary delimitation proposals in March, inviting objections from prospective candidates, voters, political parties and civil society. One of the major criticism that emerged was that the delimitation had created constituencies of unequal sizes. The population, and by extension the number of voters, in new constituencies was not equal despite the fact that the new law had bound the election authorities to create constituencies that did not have more than 10 per cent population variation.
The critics pointed out the elections would be too expansive, and also expensive, in larger constituencies. On the other hand, they argued, voters in smaller constituencies would become more ‘representative’ than those in others.
Some constituencies were more unequal than the others. Take, for example, the neighbouring districts of Bannu and Tank in Khyber Pakhtunkhwa. Both were allocated one seat each in the National Assembly even though Bannu has a population of 1.2 million and Tank only has 400,000 people. This meant that running an election campaign in the former district would require more effort and resources than in the latter. What made this comparison even more skewed was the fact that one voter in Tank would be equal to three in Bannu.
These two were not the only unequal constituencies.
According to the latest population figures, the 260 National Assembly constituencies (excluding 12 for Fata from the total of 272) should have an average of 780,000 inhabitants. The legally allowed 10 per cent variation set this number between 741,000 and 819,000. But the preliminary delimitations showed that the population of seven constituencies was in excess of one million while that of five others was below even half a million. This was by no means an acceptable situation as it violates the equality of vote as given in the one-person, one-vote rule that sits at the heart of an electoral democracy.
The problem was mainly caused by another rule that made a district the biggest unit of a constituency: no constituency can cross the boundaries of a district. In other words, no district can share some part of a constituency with an adjoining district. The problem that arose from this rule was that districts rarely have populations that are an exact multiple of the average number of inhabitants per seat as determined by the election commission. Bannu’s share of National Assembly seats, for instance, was calculated to be 1.49 as per its population while that of Tank came to be 0.50. By following the rule that gives whole seat/seats to each district, the fractional shares of the two districts had to be rounded off to the nearest whole number in both cases even though their populations varied from the average population per seat by way more than 10 per cent.
Strictly abiding by the 10 per cent population variation rule would have resulted in multi-district constituencies which need to be avoided because there is evidence to show that the quality of governance suffers in a constituency if its administrative and electoral boundaries do not coincide.
In a few exceptional cases, though, even the whole seat/seats for a district rule had to be modified. The three districts of Torghar, Lower Kohistan and Kolai-Palas Kohistan in Khyber Pakhtuhkhwa have such tiny populations that they could not be given a whole seat each without violating the average population per seat rule by a huge margin. They, therefore, have been clubbed with their adjoining districts as far as delimitation is concerned.
The same is true for most of the 34 districts in Balochistan. Quetta is the only district in the province that has three seats of its own. Only five other districts have a population that is close to the stipulated average per seat. These five, thus, have been given one seat each. The rest of the districts have been grouped together in such a way that each group forms a National Assembly constituency.
What the election commission has ensured is that variations in constituency populations within each district do not exceed the 10 per cent limit. Lahore district’s 14 National Assembly constituencies are only five per cent smaller or larger than the average. Same is the case in other districts that have more than one seat.
The only outlier is Karachi. Its 21 seats do not seem to follow the average population rule. This is mainly because the city is divided administratively into six districts, requiring the election authorities to allocate whole, not fractional, seats to each of these districts regardless of population variations among them.
To sum it up, the two rules – 10 per cent variation in population and whole seat/seats for a district – in many, if not all, cases worked against each other. The only possible solution to address this paradoxical situation was to redraw the boundaries of both districts and electoral constituencies in a coordinated manner. That was definitely not possible in order for delimitation to be completed in time for the polls, especially as the authority to redraw district boundaries lies with the provincial governments and not with the election commission.
The problems with the new delimitation were not limited to inequality among constituencies and districts. The actual drawing of lines within each district had its own challenges. The election commission tried to solve it by employing a unique rule which required delimitation officials to start from the northernmost smallest unit – revenue circle in rural areas and census circle in urban areas – within a district and then take a clockwise direction in a zigzag manner to include as many smallest units as were required to make a constituency’s population equal to the set average.
This was certainly not the only rule that officials had to follow. They also had to consider geographical features such as rivers, mountains, communication links and other similar things to decide the boundaries of a constituency. It was not always easy to come up with constituencies that abided by all the laid down rules and procedures.
To address the problems arising out of this complicated process, the election commission made the proposed delimitations public and invited objections about them — as it was required to do by law. It received around 1,285 objections from across Pakistan. The hearings on objections were open for the public to attend.
Many objections were subsequently addressed, though, it is difficult to know how many. It is also difficult to find out what changes were eventually made and by which yardstick in cases where the objections were accepted.
Publicising documents related to the proceedings and the verdicts over objections would have offered an important peep into the adjudication process. These would have allowed observers and analysts to understand how the election commission had weighed different delimitation parameters to finalise the constituency boundaries.
This could not be done perhaps because of the limited time allotted for adjudication. As per legal provisions, all the objections received by the election authorities had to be addressed within 30 days. In contrast, preliminary proposals for the first delimitation under the 1973 Constitution were publicised in November 1975 and were finalised in January 1977. Other countries also allow a much longer time to make the process thorough and credible. In the United Kingdom, the Constituencies Act, 2011 has introduced two stages of inviting objections. It gives the public 12 weeks to raise objections during the first stage and four more weeks to do so at the second stage. That the British constituencies have an average population of 70,000 – a good 11 times lower than that in our National Assembly constituencies – makes the time available there even more sizeable.
But no matter how much time could be given and how much effort and energy the drafters would have invested, a redrawing of electoral boundaries would always create winners and losers. There was, thus, a lot of hue and cry in the National Assembly immediately after the publication of the preliminary proposals. To allay the apprehensions of the legislators, speaker Ayaz Sadiq formed a parliamentary committee to examine the proposed delimitations and make recommendations. The election commission, however, pointed out to the parliamentarians that a mechanism to raise objections and have them addressed already existed and that the parliamentary committee had no legal locus standi. The committee still went ahead with its proceedings and recommended that delimitation be abandoned and old constituencies be revived for the 2018 elections.
There were no takers for this proposal so it fizzled out quickly.
Daniyal Aziz, then working as a federal minister, was one of the most agitated and vocal opponents of the new delimitation that had reduced the number of seats in his home district of Narowal from three to two. In the 2013 elections, he was elected on a PMLN ticket from one of the three constituencies. The other two constituencies were also won by the same party. The reduction in seats meant that the PMLN could nominate only two of its associates for the National Assembly in the district for the 2018 elections — with Aziz fearing that the axe might fall on him.
His anxiety was shared by many others in his party. The reason being that Punjab’s loss of seven constituencies and redrawing of the boundaries of many other constituencies had pushed them into electoral no-man’s-land.
A look at the redrawn constituency boundaries shows that 11 districts in northern and central Punjab have lost one National Assembly seat each — seven to other provinces, one to the province’s own capital, Lahore, and three to southern Punjab districts. So, essentially, National Assembly constituencies in 15 districts of the province have been readjusted. Another five districts in northern and central Punjab have lost one provincial assembly seat each – though they have not lost any National Assembly seats – and another district in the same region has gained one provincial assembly seat. On the whole, 21 of Punjab’s 35 districts have been affected by ‘seat adjustment’ during the delimitation process.
Sindh’s share in the National Assembly seats has remained the same as before – at 61 – but the number of districts in the province has substantially increased since the last delimitation. In 2002, the province had 16 districts but since then it has created 13 new ones. So, the same number of seats that were once spread over only 16 districts had to be divided between 29 districts this time round — with each of them having whole seat/seats rather than fractions of them as was the case in 2008 and 2013. This has had a similar impact on Sindh’s electoral map as witnessed in Punjab.
Consider the example of Larkana district which was given four National Assembly seats in the previous delimitation. Qambar Shahdadkot district was carved out of it in 2004 and in such a manner that the boundaries of two of the four constituencies in both districts cut across them. Under the new delimitation process, each district has got two seats but their readjustment means that the boundaries of all four constituencies had to be redrawn.
The administrative boundaries of only five districts in Sindh – Khairpur, Ghotki, Tharparkar, Naushahro Feroze and Shaheed Benazirabad (old Nawabshah) – have remained unchanged since the last delimitation in 2002. But Naushahro Feroze and Shaheed Benazirabad have lost one provincial assembly seat each while Khairpur has gained one. So, in essence, only two of Sindh’s 29 districts – Ghotki and Tharparkar – have the same seat shares in the assemblies as they did under the previous delimitation.
Khyber Pakhtunkwa has largely been a ‘winner’ in the delimitation process. The province has created three new districts since the last delimitation (Torghar, Lower Kohistan and Kolai-Palas Kohistan) but none of them qualifies to have a National Assembly seat of its own. Similarly, 14 of the province’s 39 National Assembly constituencies are already stretched over a single district and thus required no redrawing.
More importantly, Khyber Pakhtunkhwa has gained four National Assembly seats due to its high population growth rate — one each having gone to Peshawar, Swat and Lower Dir districts and the fourth one allocated to Tank which earlier shared its single seat with one of the two seats that Dera Ismail Khan had. The addition of these seats has created opportunities for new political players and, perhaps, it may also help some older ones to extend their electoral outreach even further. Either way, it is a positive development for the residents and the politicians of these districts.
In Balochistan, the provincial capital, Quetta, had only one whole seat earlier and shared another with its neighbouring districts but this time round it has qualified for three whole seats — garnering both additional seats the province has received from the national pool. The delimitation, thus, has redrawn almost the entire electoral landscape of the country — especially that of the more populous provinces of Sindh and Punjab. This change will certainly have an impact on electoral outcomes and it is this impact that has made many political actors nervous.
More intriguingly, the impact will be the strongest in the provinces where provincially leading parties – PMLN and PPP – are already under pressure from various quarters. The newly drawn constituency boundaries have given their detractors among the powerful military establishment yet another tool to browbeat them.
Correction: In a previous version of this story, we stated that General Zia wanted delimitation to happen on the basis of 15-year-old data. It is, in fact, 13-year-old data. We apologise for this error.
The writer works with Punjab Lok Sujag, a research and advocacy group focused on understanding governance and democracy.
This article was originally published in the July 2018 issue of the Herald. To read more, subscribe to the Herald in print.