By Hina Shaikh
Many have observed that Pakistan’s cities are growing fast, but until now that change has not been captured with the exacting data of a census. They would be surprised to find that according to provisional results of the new census, Pakistan is now only 36 percent urban despite a 30 percent increase in the urban growth rate since 1998.
The results have sparked debate around the integrity of this data and its implications on policymaking, political representation, and resource allocation in cities. Social scientists, economists, and urban experts strongly endorse revisiting the definition of the term “urban” to enable policy decisions that are grounded in reality. They also believe definitional anomalies remain the predominant reason for why urban population estimates appear out of sync with expectations.
If the provisional results related to the urban count are accurate, what do they tell us? Here is a table with the official results:
Figure 1: Urban/Rural population figures, Census 2017
Some key takeaways:
Pakistan’s overall urban growth has been slower than expected
While the overall population count surpasses all previous estimates, the provisional results confirm only a marginal increase in the share of urban population, up to 36.7 from 32.5 percent in 1998. In fact, most of the population is still rural in Pakistan, with results showing a tapering off in the rate of urbanization from an annual 3.5 to 2.7 percent since the last inter-censal (1981-1998) period. This is surprising as even the 1998 census is believed by researchers to underestimate the urban population by at least 10 percentage points.
Figure 2: Population of major cities, Census 2017
The highest estimate for 2017 was by National Institute of Population Sciences that had forecast a population of 200 million; Economic Survey 2017 had the figure of 199 million whereas UN’s population division’s 2017 forecast for Pakistan were 196.7 million.
But urban centers have grown explosively – with one exception
A fifth of all Pakistanis now live in just 10 cities, with Quetta, Lahore, and Faisalabad showing the largest percentage increases in population amongst these cities. The changing population distribution toward larger cities versus smaller, intermediary cities reflects an increase in the number, size, and sprawl of citiesacross provinces but also highlights some interesting findings and anomalies.
Contrary to some projections that Karachi’s population would be over 20 million, its ratio vis-à-vis the rest of the province remained unchanged. Urban Karachi shows an annual average growth rate of 2.4 percent, lower than the national average at 2.7 percent. Given that Karachi is the biggest urban economic hub of Pakistan it may be difficult to grasp that growth has not just levelled off but in fact slowed down in the city compared to a much faster average growth for urban Pakistan.
What is more surprising is that Islamabad’s rural population has grown at 6.95 percent per year since the last census. Anyone who is familiar with Islamabad knows there are barely any “rural” localities in the city.
Quetta’s population has also expanded rapidly – its urban population has grown by almost 3.5 percent. This could not have happened without the first-ever counting of Afghan refugees, which is a welcome step by the census administration because policymakers should be interested in the total number of residents regardless of their legal status. However, this practice should be consistent for all regions across Pakistan.
Urban and rural household sizes are nearly equal
Another interesting trend is the parity in urban and rural household size. In all four provinces, urban and rural household size (measured as number of people per house) is almost the same. In fact, in a total reversal from the 1998 census, Sindh’s urban household size (5.6) is larger than rural (5.4). What is usually expected is that urban household sizes are significantly smaller than rural household sizes. A lower mean size of households in rural areas can be partly due to out-migration of household members to urban areas. This trend seems to have intensified in Punjab as well: the lower rural population growth rate has slowed down (1.8 percent) compared to urban growth (2.7 percent).
Provincial shares of the overall population remain constant
Figure 3: Urban share of provincial population, Censuses 1981, 1998, 2017
Despite massive in-migration in the inter-censal period, provincial population shares see little change. Experts find this surprising especially as it leads to minimal adjustments in the division of parliamentary seats across provinces. The total population share for Sindh is unchanged, while the share of its rural population increased by more than 50 percent. This is in sharp contrast to Punjab, where its rural population grew by far less than the national average.
So what is going on here? These figures seem off – either inflated or underestimated – largely due to definitional issues.
One of the reasons for the discrepancies observed is the varying understanding by provincial governments of what constitutes as “urban”.
The definition of an urban area followed by the Pakistan Bureau of Statistics (PBS) in 2017, as in 1998, is a municipal corporation, a town committee, or a cantonment board are administratively classified as urban. This is actually a step back from the definitions used in the 1951 and 1961 censuses where even areas with certain “urban characteristics” were classified as urban.
Characteristics of rural regions change over time. Hence, classifying any region that does not fall neatly within an urban center as rural may not be correct. For instance, access to basic utilities such as water, electricity, mobile phones, and internet connectivity are no longer purely urban features.
Work by several urban experts, such as Reza Ali, provides ample evidence to support this thinking. His path breaking research estimates the scale of urbanization in Pakistan since the 1990s. Ali shows how binary, official definitions of ‘urban’ and ‘rural’ remain insufficient and he suggests that the distinction between the two should be considered more like a transition along a gradient.
Since the last census, researchers have also documented how “ribbons of development” have evolved. These are pockets of urbanization along highways, between major cities, and industrial zones that are still classified as rural. The World Bank has also reported this phenomenon, describing Pakistan’s urbanization as “messy and hidden” due to the proliferation of a low density urban sprawl, slums, and growth of cities beyond administrative boundaries that define urban centers. The United Nations also reports that by 2015, 22 percent of the Pakistani population had been living in urban agglomerations of more than a million people.
How definitional concerns affected the census
In some cases, the suburban/peri-urban outgrowth of cities are marked as rural. As city limits are not frequently revised to reflect this growth, most of it is considered outside of cities. For example, the 1998 census marked Lahore’s Defence Housing Authority, the Lahore Development Authority, and private housing schemes as rural. In the 2017 census, Islamabad’s Bahria Town is marked as rural.
Sindh’s elevation into a majority urban province – 52 percent compared to 48.9 percent (as per the 1998 census) of its population is urban – can most likely be attributed to the designation of areas previously considered rural as urban (apart from the wave of migrants from KP due to conflict).
Also, the far lower than expected difference between the populations of Lahore (11.3 million) and Karachi (14.9 million) warrants a closer look. A more than 100 percent increase in Lahore’s population could be due to the expansion of the city’s borders – reclassification by the provincial government – to include areas formerly counted as rural. However, Sindh has been most vocal about its reservation to accept these results and feels the population of the entire province, not just Karachi, has been undercounted. A big question mark is that almost five million residents of Karachi are declared rural residents as two major districts of the city – Malir and Korangi – are marked as rural.
There is also no standard definition of urban vis-a-vis rural in Pakistan. Some cities in Punjab have an urban-rural divide, while others are entirely urban.
While underestimating the scale of urbanization through definitional interventions may benefit political parties with a strong rural base, the worrying fact remains that Pakistan has added at least 30 million people to its urban population in the last 19 years. Without a appropriate policies, another 100 million will be added to this by 2050 if current growth rates persist.
Rapid urbanization is challenging the flimsy infrastructure of cities. Rising population density also has implications for pollution, waste management and climate change that policymakers need to start focusing on. According to the Global Liveability Report 2017, Pakistan’s largest city, Karachi is also its least livable, ranked at 134 out of 140 cities.
After a delay of 19 long years, even basic changes may appear monumental once accumulated. However, without census data, policy planning is not much better than guesswork. The research community will be excited about using new data, however it will take some time before these results find a way into result-oriented strategies and policies.