Volume 21 Issue 4 April 2017


One of the wheels of democracy which helps a country’s smooth functioning is knowing the number of people in different parts of the country, so that they could choose their representatives on the basis of one man-one vote. Besides, a census also provides relevant data at national, provincial and local levels for proper planning. Accordingly, almost all countries of the world conduct a census every 10 years and some even conduct it every 5 years (e.g. Iran and even the Maldives did so). During the past 15 years, a census has been conducted in almost all UN member countries, however, only a handful among major countries have not done so. These include Lebanon, Somalia, Afghanistan and Pakistan. The 1973 Constitution of Pakistan postulates conducting a census every ten years, which could become the basis of demarcating constituencies of the National and Provincial Assemblies, help in allocation of national resources to provinces through National Finance Commission (NFC) and determine job quota to provinces, on the basis of the results of the most recent census.

The United Nations defines census as, “the total process of collecting, compiling and pub-lishing demographic, economic and social data pertaining, at a specified time of all persons in a country.” In practice, this does not mean that every person actually is seen and interviewed but only one adult in a household provides information about all other residents (relatives and non-relatives) to questions asked by the census enumerator.

Population Growth in Pakistan 1951-1998
Since the first census conducted in 1951 and the last in 1998, Pakistan’s population grew four-fold. During 1951-81 period the annual growth rate was 3.1 percent which declined to 2.6 percent during 1981-98. Although a census is supposed to count everybody, but many people are missed which could also happen in a developing country, as well as in more developed countries. For example, in Pakistan after the 1961 cenus when a post-enumeration survey was conducted, about 6 percent under-count in the population was discovered. Therefore, the population was later adjusted upwards. In 1981 a post-enumeration survey was conducted but results were not released. After the 1998 census a post-enumeration survey was not undertaken, therefore the extent of under-reporting could not be determined.

However, it is largely believed that the 1981 census was the most accurate census conducted in Pakistan. After the 1991 census was postponed, it was decided to conduct the census six years later. Apparently, the Federal Government was not sure that the Pakistan Census Organization (which was a part of the Ministry of Finance and Planning & Development) could conduct a fair census on its own. Therefore, for the first time, assistance of armed forces was sought. It was assumed that in the presence of military personnel who would accompany the civilian enumerator, the household member will not exaggerate the number, as was experienced in some areas during the housing census conducted in 1991. However, since a post-enumeration survey was not conducted, therefore results could not be validated, particularly the possible undercount in the census. Subsequently, in 2000 a steering committee of experts was constituted by the Government of Pakistan with support from the United Nations Population Fund. It was found that the 1998 census suffered from various inconsistencies, one of which was a possible undercount or under reporting. This author, who served as a member of the steering committee, by applying annual growth rates derived from Pakistan Demographic surveys, conducted during 1982-97, estimated that Pakistan’s population in 1998 should have been about 142 million. However, after accounting for about 3 million emigration from Pakistan to the Gulf countries, Europe and North America, during 1981-98 inter-censal period, it was suggested that about 5 percent population missed in the 1998 census. Further analysis of 1998 census indicated that the undercount was mostly in Sindh province, which could be to the tune of about 6 million people. Of these, about 2 million, who although were counted living in Karachi, but were not included in the census report as they were considered alien or illegal immigrants from Afghanistan and Bangladesh. In a subsequent report published in 2008 by the Pakistan Institute of Development Economics it was pointed out that, in the 1998 census the population of urban areas and migration to large cities were under-estimated, data on education was inadequate and there were issues pertaining to data collected on economic activity.

Since census is a gigantic undertaking, an in countries where the majority of people are not educated, census results suffer from coverage and under-reporting of population. Besides, since the 1973 constitution of Pakistan (as of many other countries) census results determine allocation of seats in the National Assembly and 85 percent weightage is given to census results for the purpose of resource allocation though the National Finance Commission, the possibility of census results being doctored always exists. Therefore, those who would want a status quo are usually in favour of postponing the census. Perhaps the results of 1998 census could have been affected because, as shown in the figures, the share of Sindh had increased in each subsequent census during 1951 and 1981, when it increased from about 18 percent in 1951 to 23 percent in 1981. Whereas, there was over 3 percentage point increase in the share of Sindh between 1961-81 (in 20 years), during 1981-98 (in 19 years), Sindh gained only 0.4 percentage points . It is a well recognized fact that Sindh has been on the receiving end of migrants from other three provinces which has resulted in the increasing share of Sindh’s population, which was evident till 1981. However, a negligible increase in Sindh’s population can be only explained if there was only negligible migration to Sindh during 1981-98 or the rate of natural increase (difference between birth and death rates) had suddenly declined there.

Role of Inter-Provincial Migration
Further investigation reveals that migration to Sindh had continued between 1981 and 1998 as shown in Table 2.
Thus, due to migration Sindh had a gain of over 1.1 million people, of which over 0.6 million originated in Punjab and another over 0.4 million originated in former NWFP (renamed Khyber Pakhtunkhwa). Another interesting finding of the 1998 census was that Karachi was the main point of destination of most migrants to Sindh. For example, the 1998 census reported that while Karachi then contained about 8% of Pakistan’s population, of all those people who migrated internally within Pakistan during 10 years prior to the census 22 percent chose to settle in Karachi. In comparison, the four largest districts of Punjab - Lahore, Faisalabad, Rawalpindi and Multan - which combined together contained about 21% of the country’s population, received about the same number of internal migrants as did Karachi. Furthermore, demographic surveys conducted by Pakistan’s Federal Bureau of Statistics showed that rates of natural Increase (difference between birth and death rates) during 1980s and 90s in Sindh and other provinces were about the same. There are thus indications that since Sindh’s population was under-enumerated in the 1998 census, therefore its share in the country’s population has remained at the same level as in 1981. This is indicated in Table 3, where it has remained unexplained that why the average number of persons per household in Sindh declined substantially from 7 in 1981 to 5.8 in 1998, while a subsequent household sample survey conducted in 2000 by the Federal Bureau of Statistics showed an average 7.5 persons per household in Sindh. The discrepancy is also noted in Balochistan.

Challenges in the 2017 census
The 2017 census currently underway has already become disputed and various political parties, particularly in smaller provinces, have raised many questions. The results will be subject to challenges by a vibrant media. Besides, due to the concerns shown by many groups, results may be challenged in courts as well. Even prior to the start of the census, Chief Minister of Sindh had requested the Federal Finance Minister to make public the number of people who have been counted at sub-divisional level, fearing that the numbers may be changed in Islamabad. After not getting a positive response from the Finance Minster, Pakistan People’s Party has already filed a petition in Sindh High Court, against the “lack of transparency” praying to carry out the entire exercise in a transparent manner and to provide complete access of data to the provinces, before releasing the results. The petition points out that under Article 19-A of the Constitution for the basic right of access to information, assistant commissioners should maintain the record of data in their offices which should be accessible to the public.

Information being collected in the 2017 census
The 2017 census is collecting information about each household member such as Name, Relationship to the head, Sex, Age, Marital Status, Religion, Mother tongue, Nationality, Literacy status, School attendance, Level of education completed, whether during the past 12 months the person is working, looking for a job, doing house work or any other work. Besides information about the living quarters is also being collected such as number of rooms - tenure status, period since constructed, construction material used in outer walls and roofs, source of drinking water, source of lighting – fuel used for cooking, availability of kitchen, bathroom and type of latrine, availability of TV, Radio, Newspaper, Telephone. Mobile Phone and Computer and if any household member is living abroad and their number. One issue which has remained a bone of contention is obtaining the number of Computerised National Identity Card (CNIC) of the household head, which is being verified by the representative of the armed forces by sending a SMS to NADRA. The expert committee constituted by the Governing Council of Pakistan Bureau of Statistics had taken the view that since obtaining CNIC is not a part of the census it will cause unnecessary delay and verification from NADRA will be against the spirit of confidentiality of the census. Besides, some political parties have raised the issue that since in rural areas many people do not have CINC they may be left out of the census exercise. However, CNIC number is being collected.

Previous census also collected information through a long term Form from every 10th household which collected information pertaining to migration such as district of birth, duration of continuous residence in the present district, district or country of previous residence and the reason of migration and nature and type of any disability. Those 5 years and older, their occupation and from married females 15-49 years old; number of children ever born, number of children still living, number of children born during last 12 months and of these how many are alive. This is crucial information which will help in understanding important demographic, economic and health indicators for planning purposes.

A mission from the United Nations in its report submitted to Pakistan Bureau of Statistics in October 2016, made the following recommendations for smooth conduct of the census, which were not adopted. Theses included:

1. The census questionnaire was printed in 2008, however it was not pilot tested recently. Therefore, to increase reliability of the census, it was strongly advised to organize a pilot census.

2. Since a majority of the population did not participate in a census in almost 20 years, time must be allowed for the population to understand, accept and “own” the results.

3. The communication strategy was to be organized around three “classical” axes: education material (towards teachers), advocacy material (towards politicians, head of villages and other stakeholders) and publicity campaign (towards the larger audience).

4. The advocacy and publicity campaign should be launched no later than three months before the census; one has to keep in mind.



If counting the correct number of people living in Pakistan is a cumbersome task, estimating Pakistan’s current population is even more difficult. Estimating a country’s population is dependent on several key factors which include accurate counting of the previous population and accounting for the natural increase (births minus deaths) each year and adding the number of people who have migrated from other countries and substracting those who left the country. Such information is only collected in countries which have an advanced level of registration of births, deaths and migration.

Most countries in North America, Europe and several countries in South America and Asia have developed a system of vital registration which records births and deaths. In countries where no such system is in place, annual sample registration of vital events (such as in India) serves the purpose. In the 1960s, Pakistan started a sample registration system and in the 1970s switched to annual demographic surveys, which provided estimates of births and deaths. We were then aware at what rate Pakistan’s population was increasing each year. These sample surveys were conducted by the Federal Bureau of Statistics (FBS), which was a sister department of the Pakistan Census Organization and both were part of the Statistics Division of the Ministry of Finance. However, as many good things which used to be a part of the routine data collection, FBS stopped conducting the Pakistan Demographic Survey after the last one conducted in 2008. These surveys did not only provide demographic data for the country but for all its provinces and rural and urban areas. Consequently, the annual rates of population growth in Pakistan as quoted in the Pakistan Statistical Year Book (last published by Pakistan Bureau of Statistics) on the basis of Pakistan Demographic Surveys showed an annual population growth rate of 2.62 percent during 1995-97 which declined to 2.12 percent during 1999-2001. However, since no such surveys have been conducted after 2007, the Planning Commission comes out with “guestimates” of population growth rates, whereby the population is assumed to be declining every year, from 2.10 percent in 2008 to 1.95 percent in 2014. Indeed, according to Pakistan Demographic and Health surveys conducted in 2005-06 and 2012-13, the total fertility rate (average number of children a woman bears by the end of her reproductive period) has remained about 4 while infant mortality rate has declined somewhat, suggesting that perhaps there has been no decline in the population growth rate. Indeed, a better way to estimate or project a country’s population is through a sophisticated computer software developed by the United Nations which requires four important indicators or inputs: namely; population at a previous time, total fertility rate (TFR); expectation of life at birth (ELB) and age distribution of the population. Unfortunately for Pakistan the only input available is the population as recorded in the 1998 census, which is itself questionable as there were indications of underreporting of about 6 million people. Using the UN software, the National Institute of Population Studies (NIPS), which is a part of the Ministry of National Health and Coordination, Government of Pakistan, last year released a paper projecting the population of Pakistan and its provinces, as 198.8 million in 2016 and 202.9 million in 2017. Interestingly, the experts who produced the NIPS report were senior demographers. They assumed that TFR in Pakistan has declined from an average of 4.2 children per woman in 2000-05 to 3.2 in 2010-15 and LEB during the same period has increased from 64.4 years to 68.5 years. To the contrary, the Demographic and Health Survey conducted by NIPS in 2012-13 reported a TFR of 3.8 and United Nations Human Development Report of 2016 indicates that LEB in Pakistan has increased from 62.7 years in 2000 to 66.4 in 2015. If NIPS would have used the correct inputs of TFR and LEB, perhaps Pakistan’s population would stand at 210 million in 2017, and even more if 1998 census would not have undercounted the country’s population. Still NIPS’ estimate of Pakistan’s population in 2017 is on the higher side, as the Government of Pakistan’s official estimates are 195 million. The only proper indication of the population of Pakistan will come out of the 2017 census, when it is correctly done. This will be if the Government of Pakistan again does not panic as they did in 2011, when the preliminary results of the Housing Census indicated that the population had already reached 195.9 million, which was way beyond their expectations. The share of Sindh in the national population was reported to have increased to 28.2 percent from 23 percent in 1998 and that of Punjab had declined from 55.8 percent to 46.6 percent, while it remained 13.7 percent in KPK and increased in Balochistan.
Consequently, the detailed Population Census was immediately postponed and the results of the Housing Census were cancelled.

– Dr. Mehtab S. Karim


After results of the 1998 census were announced, not only the Sindh Assembly unanimously rejected it, which was then ruled by a coalition government led by Muslim League, but Ms. Benazir Bhutto, who was then leader of opposition, in her speech in the National Assembly was quite candid, saying “When things come to Islamabad, they are cooked up and somebody else takes the broth”. Indeed, even today the census exercise being conducted is entirely being controlled from Islamabad. So much so, that data processing centre at Karachi office of Pakistan Bureau of Statistics was shifted to Islamabad prior to the census. This data processing centre was established a couple of years ago with financial support from the United Nations Population Fund, to scan the complete census forms from Sindh and Balochistan. Why a need arose to centralize even scanning of forms, has created suspicions in Sindh, that results may be doctored in Islamabad. Due to this apprehension, Chief Minister Syed Murad Ali Shah, had requested Federal Finance Minister Mr. Ishaque Dar, that a photo copy of the summary REN form - which contains the population counted in each census block and then submitted by the census enumerators to the sub-divisional magistrate for onward transmission to Islamabad - may be retained by the SDM for checking at a later stage. However, this request was rejected by the Federal Government. Sindh Government’s concern is quite genuine, given the fact that even though millions of migrants who had moved to the province in-between the two censuses conducted in 1981 and 1998, the province did not get its fair share, due to the undercount of population of Sindh.

Unfortunately, information on demographic parameters are not available at provincial level to estimate their population, specially for a province like Sindh which has been at the receiving end of migrants since Pakistan became an independent country. In fact in the population estimates prepared by the National Institute of Population Studies (NIPS) for 1999 through 2035, the share of Sindh in the country’s population has remained at 23 percent – the same as was reported both in the 1981 and 1998 census. Thus, the NIPS’ report indicates that the population of Sindh in 2017 is 45 million. They arrived at this estimate without taking into account hundreds of thousands of people who have been migrating to Sindh every year over the past 18 years, mostly from other provinces. On the other hand, the Housing census conducted in 2011 had already counted about 56 million people living in Sindh, which had put Sindh’s share in Pakistan’s population at 28 percent. Perhaps, there was some exaggeration then, but it is likely that at present, Sindh contains minimum of 56 million or even close to 60 million people, which would mean that its share in the national pool will increase substantially from currently 23 to 28 or even 30 percent.

Karachi’s case is more pronounced as most migrants from other provinces end up in the city. MQM has already raised several objections to the way census is being conducted and has even petitioned the Supreme Court with a list of complaints. This time around since everybody is being included in the census count (including illegal persons such as Afghanis, Bengalis and Burmese), one has to see how close the figures will be to the various estimates of Karachi’s population. In answer to a question raised in the Senate, NADRA compiled population living in each of the 20 constituencies of Karachi and presented a figure of 21,347,871 (21.35 million) people, who were residing in Karachi in 2013. If this population figure was correct, then with an annual growth rate of 4 percent, by the middle of 2017, Karachi’s population would reach 25 million. In the event 2017 census grossly underreports Karachi’s population below this number, it would not be accepted by Karachi based political parties as well as the civil society, and most likely will be challenged in the Supreme court.

Another source of Karachi’s current population comes out of various international sources. Thus, according to the 2015 edition of Demographia's World Atlas (see table above), Karachi with a population of 22.1 million people was ranked 7th, having a population density of 23,400 people per square kilometer, making it the second most densely populated, among the world’s 20 mega-cities.

Zahid Husain, a veteran journalist writing in Dawn, rightly points out that “While the census may give a more realistic socioeconomic picture, it could also open a Pandora’s Box. While a more populated Sindh may demand a greater share of resources and representation in parliament, the province is likely to face the same demand from its own urban centres”.

©2017 SouthAsia.com.pk. All Rights Reserved