Population Census 2017 - An insight to the Results

Introduction

Censuses are conducted across the world to assess the population head count and inter-census growth rates of countries, as this is critical information for all policy measures. The efficacy of this exercise is dependent on its frequency; unfortunately the Census of 2017 has been the first one since 1998, which leaves us with a substantial gap of 19 years. The recent census has placed the total population count at 207 Million and a national growth rate of 2.4 %. The results also place the claim that 36.38 percent of the population is Urban.

Given these findings and the context of a delayed census, voices refuting the results have been raised from different corners of the intelligentsia via print and electronic media.

This report aims to examine the perceived anomalies in regards to the census: by providing a comprehensive analysis of the census findings, via cross validation through spatial and non-spatial data. There are two objectives were achieved. • To map and compare 2017 census results with the previous census; at National, Provincial and Cities Level • To triangulate the census results with spatial and non-spatial data to validate census results The spatial datasets; high resolution satellite imagery of different years ,LandScan population data, administrative boundaries up to mouza level were used and non-spatial datasets; census results; 1981, 1998 and 2017, electricity and water connections, Property Tax data and numerous information (extracted from reports of Pakistan Bureau of statistics) were integrated with the geographical features.

Satellite images were processed to extract the builtup of the major cities of Pakistan; including Lahore, Karachi, Multan and Peshawar. Urban extent was calculated by developing a complex python based algorithm which incorporated rules defined in globally acceptable methodology to study urban expansion. The rules were defined by Urban Expansion Program at New York University, UN-Habitat and Lincoln Institute of Land Policy). The urban extent of all major cities of Pakistan was extracted by using same methodology. In addition to the natural urban boundaries, the LG notified urban boundaries were delineated by translating the notifications in consultation with the LG department officials. The maps of the Metropolitan and Municipal Corporation were cross verified and signed by the LG representatives. LandScan, a 1 sq. km. resolution data was used to estimate the population count at various levels i.e. province, district, city extent and also with in the notified LG boundaries including metropolitan and municipal corporations.

Results and Discussion

In a nut shell, total population of Pakistan is 207774 million, which has an annual growth rate of 2.4%. National urbanization rate is 36.38%, Sindh is the most urbanized province. The comparison of Population increase of top twenty districts of Pakistan, along with percentage increase and compound annual growth rate are estimated (see attachment) which only include one district of Punjab i.e. Rajanpur. Whereas Quetta has the maximum gain of 194 percent over the last 19 years. Similarly, the ranking of the districts of Punjab Province has been done on the basis of annual population growth rate. Rajanpur has the highest growth rate, while Nankana Sahib is at the bottom of the ranking. Population data of 1998 and 2017 has been mapped to present the population change with spatial lens.

Karachi vs Lahore

A major discussion surrounding the census results has been about the population count of cities of Lahore and Karachi. The point of contention is essentially driven from how an “urban” area or city is defined. As per the PLGO 2013, the entire district of Lahore has been redefined as a Municipal Corporation. This is the reason that the entirety of the population of the district has been placed under “urban” and the rural population is depicted as “zero”. Under this new definition around 257 villages have been wrongly added to the city of Lahore; this has resulted in the addition of 1.5 to 2.0 million persons to the city counts. When calculated through satellite imagery, the Lahore Metropolitan population was estimated to be 10,591,636, which as per Census 2017 is 11,126,285. The closeness of the values indicate the accuracy of the population estimates, which were done an year before the actual Census took place. In contrast, the previous census (1998) only included the actual urban areas under the MCL boundary and had a substantial rural population. If we compare this with Karachi, only the municipal boundaries have accounted for what constitutes Karachi city. In order to cross check the validity of the census results for Karachi we have used the imagery for the year 2000 and 2015 for Karachi city.

The figures were adjusted for the annual growth rates estimation to come to the comparison of Census with imagery of census years 1998 and 2017; Census 1998=1,051,521, Imagery 1998 = 15,074,333 while census 2017 =9,856,318, imagery=10,220,727.

It is worth to note that the growth rate of Karachi Division is 2.66 percent, whereas of Lahore Division it is 2.55%. Thus the confusion has arisen from the lack of a definition of “urban” and what is a city limit. We cannot compare apples with oranges, and a uniform criterion would resolve much of the muddle. Similar issue seems to be with Islamabad, where decline in urbanization from 65.72% to 55.78% has been observed in Census. On the other hand, Islamabad rural population has been increased by 6%. This contradiction clearly depicts that the definition of urban rural divide is not limited to Karachi or Lahore, but at national level. Bara Kahu, Bani Gala, satellite town, Burma town, DHA Phase 2 To take the discussion to a further level, data of electricity connection of K electric was assessed through National Electric Power Regulatory Authority (Annual Report 2015). The report defines the total number of consumers as 2.5 million, whereas domestic consumers are 1.96 million. It is also to be noted that the K Electirc has jurisdiction in Dhabeji, Gharo and Hub. Thus the number in Karachi City is less than 1.9m. Moreover, MICS and PSLM data show that 98.9 % households are provided electricity in urban Karachi. When analyzed the Karachi Water and Sewerage Board (KWSB) consumer data of Karachi, it came out to be 1.5 million approximately. On the other hand, the total property units as per DG Office, Excise & Taxation Department are 0.9 million.

The purpose of quoting all the data from service delivery department is to build a statement of the validity of the data; as if multiplied with the average household size, it come into the population brackets defined by Pakistan Bureau of Statistics after the survey.

Recommendations

  • According to the rapid assessment of the census, data is fairly accurate. However, for in depth analysis and cross validation, PBS should be asked to share block level shapefile in with GoPb.
  • Urban definition issue should be resolved. Preferably, it should be the mandate of PBS and not left to Local Government, so that we have unified National data
  • Random Third Party Validation (TPV) should be conducted to cross verify census results.

File:census-draft revised.pdf