Ever since the provisional results of the 2017 census were released, they have attracted controversy. It said that there were about 16 million people in Karachi and about 47 million in the province. All of Sindh’s political parties raised objections as did the media, civil society and scholars about what they said was a gross undercount. Even the Chief Justice of Pakistan is on record suggesting that Karachi’s population is perhaps twice than what has been reported in the provisional results. According to my calculations, about 8 million people who were under-reported in Sindh, perhaps 6 million were from Karachi and 2 million were from other urban areas.
It appears that this controversy is why the Council of Common Interests has yet to reach consensus to validate the results.
Before the 2018 elections, the Supreme Court had taken the stand that since the census results were provisional (and could change later), a constitutional amendment was needed so that elections could not be held without the results being made official (or notified). Therefore, prior to the 25th Amendment, an agreement was reached in the Senate, signed by the leaders of all political parties, for a 5% recount to be supervised by a neutral Census Commission, consisting of qualified and experienced demographers. The government notified this in January 2018 but since then the Commission has met only once, when several demographers were invited, and they cast their doubts on the census results.
The Sindh High Court recently indicated that local bodies elections cannot be held until the official census results are released. Unfortunately, a 5% recount is no more a viable option as about three-and-a-half years have lapsed. The figures would have changed due to births, deaths and in- and out-migration. Therefore, a fresh census may be conducted in urban areas of Sindh.
A 10-year plan
One of the wheels of democracy which helps a country function smoothly is counting the appropriate number of people all over the country, so that they can choose their representatives on the basis of one person, one vote. The census also provides national, provincial and local data for planning.
In fact, the 1973 Constitution says that the census must be held every ten years. If this cycle were followed, we could also demarcate constituencies and determine the share of each province through the National Finance Commission award on the census every decade. Since the 1973 Constitution, however, only the 1981 census was conducted on time, with proper planning and under the supervision of experts. The next census was held after a 17-year delay and the most recent one, after a delay of 19 years.
Why the numbers could be lower
Census results are never perfect, and thus under- or over-counting of people does happen and is prevalent in both developed and developing countries. According to a report prepared by the United Nations Observer Mission which was present during our census exercise, the CNICs of each household member was checked, verified and recorded. Enumerators mostly relied on information on the CNIC. This is the ‘de jure method’ of holding a census, in which people are counted at their usual place of residence.
Therefore it is likely that household members with a different address on the CNIC were later allocated to their original place of residence. Perhaps this is why, the reported population of Sindh, particularly Karachi, is much below expectation.
Loopholes in the method
The Statistics Division asked senior officials of the Pakistan Bureau of Statistics (PBS) to evaluate all aspects of the 2017 census. This internal committee published a report in which they pointed out several loopholes while the census was being conducted. The report reveals that the sub-committee formed by the PBS governing council had shown concerns.
“The census is not being planned the way it was envisaged,” the sub-committee noted. None of its recommendations were followed in true letter and spirit. The authorities concerned of the PBS had adopted whatever they liked and rejected that was not according to their “sweet will.” The PBS was left with a number of loopholes because the recommendations by national and international experts were not complied with. There was an acute shortage of expert staff and a lack of technical expertise.
The transparency of data collected by the PBS is further in doubt, since instead of conducting a technology-based population census (as was done in Bangladesh and Egypt), in Pakistan, it was conducted manually on forms printed over 10 years earlier. Interestingly, just prior to the census, the entire computerized data entry setup at the PBS Karachi office was shifted to Islamabad. Therefore, the possibility of manipulation at a later stage cannot be ruled out.
Since the 2017 census was conducted without proper planning and by a team which was neither professionally trained nor had the experience of conducting a census, its results will remain controversial. Therefore, in the national interest, it may be appropriate to scrap the results of the 2017 census and conduct a fresh one in 2021. It may be noted, that while in the earlier censuses, field work was completed in about two weeks, the 2017 census took two months, resulting in a substantial additional cost. Since all the household listings are available along with detailed GIS maps, however, a fresh census could be easily conducted in two weeks in Sindh using modern technology, at a substantially lower cost. Since the PBS is short of trained staff, the task of planning and monitoring a fresh census should be given to the Census Commission, consisting of trained persons and representatives of civil society, with support from the field staff of PBS to get the job done.
The writer is a professor of Demography at universities in Pakistan and the US. He worked at the UN and on its behalf, was an advisor to the Pakistan Census Organization. In 2018, he was notified as a member of the Census Commission by the Government of Pakistan. He has co-authored three books on demographic issues. He can be reached at firstname.lastname@example.org