Karachi: What the Census Tells Us

Karachi: What the Census Tells Us
By Arif Hasan, Amal Hashim and Dhuha Alvi

The census of a country gives us details of its population structure, its socio-economic wellbeing, and the conditions of its human settlements. By themselves, the census figures do not establish trends – that can only be done by comparing previous census figures with the current ones. However, to understand these statistics in the larger context, it is necessary to compare them with statistics of a city of a similar importance and size. In this article, we compare the results of the 1981, 1998, and 2017 censuses and try to determine how Karachi has fared over time. For certain categories, we have compared Karachi’s statistics to those of Lahore, and for literacy, educational attainment and marital rates, we have also compared them to those of Delhi. In the conclusions we have tried to reflect the main findings of our work^[1].

1. Demography

Between the 1981 and 1998 census, the population of Karachi increased from 5,437,984 to 9,856,318. This means an average increase of 3.56% (or 259,902 persons) per year. Between 1998 and 2017, on the other hand, Karachi’s population increased by 6,168,576 to 16,024,894 or at an average of 2.59% (324,661.89 persons) per year. The city’s density also increased from 2,794.53 persons per sq. km to 4,543.49 persons per sq. km in the intercensal period, creating severe environmental problems. It goes without saying that these figures make Karachi the largest and, in population terms, the fastest growing city in the country. However, Karachi is also unique in other ways. One, it is the only place in Pakistan where local and provincial leaders and citizens have questioned the census results and insisted that Karachi’s population is above 25 million. Credence is lent to this by the figure quoted in NewGeography.com, an authoritative organisation that monitors the population growth of cities around the world, which estimated Karachi’s population at 22.8 million as of April 2016^[2]. Among political circles, the reason for not giving Karachi’s proper population in the census is believed to be the maintenance of a political status quo in the federation. If Karachi’s population is taken at 20 or 25 million, its seats in both the provincial and national assembly will increase, and so will Sindh’s share in the National Finance Commission. Sindh’s politicians, especially those from Karachi, see this as a conspiracy against the city and the province.

The other reason why Karachi is unique is that 59.94% of Sindh’s urban population lives in the city. In contrast to this, Lahore is Pakistan’s second largest city, and it contains only 27.42% of Punjab’s urban population, whereas Peshawar and Quetta contain only 34.34% and 29.34% of Khyber Pakhtunkhwa’s and Balochistan’s urban populations, respectively. Another important factor is that between 1998 and 2017, the population increase in Karachi (6,168,576 persons) alone was more than the total increase in the other urban areas of the province (4,447,393 persons).

Issues related to ethnicity also make Karachi unique in contrast to other Pakistani cities. 42.30% of its population is Urdu speaking and 10.67% is Sindhi speaking, whereas Lahore is 80.94% Punjabi speaking and Peshawar is 90.17% Pashto and 5.33% Hindko speaking. However, Sindh as a whole is 61.60% Sindhi speaking. This makes Karachi the non-Sindhi speaking capital of a predominantly Sindhi speaking province. In addition, Karachi is a rich city generating considerable revenue for the federal and provincial governments. This can be judged by the fact that Karachi contributes 20% of Pakistan’s GDP, 50% of the country’s revenues, 46.75% of direct taxes, 33.65% of federal excise tax, 23.38% of domestic sales tax, 75.14% of customs duty, and 79% of sales tax on imports. It contains the best universities and medical facilities in Sindh, and is the centre of the Pakistan media industry. In addition, Karachi’s large scale industrial sector employs 71.6% of Sindh’s total industrial labour force; the city produces 74.8% of the province’s total industrial output, and it is generally believed that it contains 78% of its formal private sector jobs^[3]. This enormous wealth of Karachi can only be controlled by political parties representing the Sindhi speaking population of the province through a highly centralised form of government while it can only be controlled by Karachi-based politicians through a highly decentralised form of government. This conflict has been at the centre of the discussion on and failure of evolving a consensus-based local government system.

Table 1.1 (Population by Mother Tongues – Karachi):

Table 1.1 gives the changes in populations of different language speakers from the ‘81 to the 2017 census in Karachi. From the table, it will be seen that there has been a decline in the Urdu speaking population (by 12.04%) and a substantial increase in the Pashto speaking population (by 6.30%) because of continued migration from Khyber Pakhtunkhwa. However, there has also been a major influx of migrants from the rest of Sindh; this influx can be seen by the increase in the Sindhi-speaking population from 711,724 persons in 1998 to 1,709,977 persons in 2017. This is because of a shift from a subsistence to a market economy in the province’s rural areas, resulting in an increasing need for cash for survival and difficulty in generating that cash by landless labour and artisans. Moreover, Karachi is the only major city in Pakistan to which the Sindhi-speaking population can easily migrate. Similarly, the Siraiki speaking population, which was miniscule before the 1998 census, has also increased by 589,665 between 1998 and 2017 due to the decline of feudal control and its socioeconomic repercussions in Southern Punjab. These migrations are likely to continue.

Table 1.2 (District-wise Mother Tongues – Karachi – 2017):

It is also important to see where the different language groups live in the city (see Table 1.2). The Urdu speakers are concentrated mainly in the Central district, where they are 70.77% of its population, and in Korangi, where they are 61.34% of its population. The Pashto speaking population is mainly concentrated in district West, which contains a large number of Pashto speaking colonies such as Pathan Colony, Frontier Colony, etc. The Sindhi speaking population is mainly concentrated in the Malir district, much of which is a rural area. The Sindhi and Pashto speaking populations will increase their political strength considerably in the next 10 years, at the cost of the Urdu speaking population, if the present trend of an increase in their populations continues.

2. Literacy and Marital Status

The most important age group in a census is between 15 and 24. This is because it is both the present and the future. Therefore, literacy and marriage rates, especially in this age group, and of women in particular, are important social indicators, and so are their increase and decrease overtime. In the case of Karachi, the number of married women in this age group was 66.71% in 1961, 37.92% in 1981, 28.54% in 1998, and 30.87% in 2017. That of married men was 13.39% in ‘81, 10.08% in ‘98, and 11.24% in 2017. These figures show that today, Karachi has an overwhelming majority of unmarried adolescents in the age group of 15 to 24. Any sociologist would tell you that this is enough to change gender relations and, hence, family structures. Studies and newspaper articles tell us that this is already happening, not only in Karachi, but also in other urban areas of Pakistan – though to a lesser extent. However, between 1998 and 2017, this trend changed for the first time since 1961, and the number of married women and men in this age group increased by 2.33% and 1.16%, respectively. In contrast, the number of married women in this age group in Lahore have continued to decrease from 27.55% to 27.17% during the intercensal years. Here, a comparison with Delhi, a city of a size similar to Karachi in India, would not be out of place. The number of married women aged between 15 to 24 in 2001 was 38.23%, which has fallen dramatically to 31.25% in 2011. These percentages tell us that Delhi was, perhaps, a more conservative society than Karachi but change since 2001 has been far more rapid than Karachi, and this is perhaps also because of Delhi’s much higher literacy rates. It is interesting to note, however, that divorce rates for women are lower in Delhi (0.17% in 2011) than in Lahore (0.70% in 2017) or Karachi (0.73% in 2017).

Table 2.1 (Marital Rates – Karachi and Lahore):

 Table 2.2 (Literacy Rates – Karachi and Lahore):

Literacy determines a number of things. According to the tables above, literacy in Karachi, both for males and females, and especially in the age group 15 to 24, was much higher in 1981 than in Lahore. In 1998, this trend continued. However, in the 2017 census, Karachi’s overall literacy is 78.06%, while Lahore’s is 84.25%. Female literacy in Karachi is 77.79% while that in Lahore is higher than the city’s male literacy, and stands at 84.75%. Karachi’s performance between 1998 and 2017 has been much slower than Lahore’s and of its own previous intercensal figures. The reasons for this need to be investigated.

Delhi’s performance, though, has been exceptional as compared to Karachi and Lahore. As per its last census of 2011, its overall literacy rate stands at 86.21%, whereby that of women is 80.76% and of men is an impressive 90.41%. In 2001, it was 81.67% for the overall population, 74.71% for women, and 87.33% for men. The age group of 15-24 takes literacy a step further with 93.17% of the total population, 94.19% for men, and 91.93% for women. In 2001, by contrast, 87.78% of the total population, 89.68% of men, and 85.24% of women were literate within this age bracket.

Literacy also determines a number of other factors. For instance, the highest literacy in Karachi is in district Central (at 81.52%), which also has the highest employment rate, i.e., 34.78% (taken as a percentage of the district’s total population above 10 years of age). Additionally, it has the second highest divorce rate (0.60%) and lowest marital rate (60.25%). The lowest literacy rate is in district Malir (63.69%). Along with this, it has the highest marital rate (63.61%) and lowest divorce rate (0.31%).

3. Educational Attainment^[4]

Table 3.1 (Educational Attainment – Karachi):

Educational attainment is gauged by the percentage of people with a primary education, matriculates, graduates, and postgraduates. Their increase and decrease overtime shows whether the number of ‘educated’ people is increasing or not. As per Karachi’s intercensal comparisons, the higher the educational attainment, the greater women’s chances of being employed would be; however, the same doesn’t hold true for men.

Educational attainment for all categories has increased in Karachi between 1981 and 2017. However, it has increased substantially more in the case of Lahore between 1998 and 2017, unlike in the previous census. The reason for this needs to be understood. In the case of Delhi, the figures, especially for women, are substantially high. For example, female graduates in Karachi are 12.01%, in Lahore 14.63%, and in Delhi they are 21.46%. We need to know what the Indians do right to achieve such statistics.

In Karachi, the best social indicators related to literacy, educational attainment, and marital figures are the highest in district Central, where employment figures are also the highest. The lowest figures for these are in district Malir. Comparing the three censuses district-wise shows that development takes place on a larger scale in already developed areas, although in percentage terms, statistics for the less developed areas may look impressive.

The population which has completed primary education by 2017 in Karachi is considerably higher at 65.96% than in 1998 (57.93%). However, the latest figures for Lahore (68.88%) are higher than those for Karachi, and those for Delhi are even higher (76.35%). The highest figures in all educational attainment categories are for district Central. Given these figures and trends, it is unlikely that Karachi’s future projections will be better than those of Lahore and Delhi. The reasons for this need to be understood by Karachi’s policymakers.

4 Housing  

Table 4.1 (Housing Units By Tenure – Karachi):

Table 4.2 (Intercensal Changes in Housing-related Stats – Karachi): 

Karachi’s census figures for housing are really problematic and do not give us the real housing figures in the city. There are 2.73 million households and 2.73 million housing units. This means that there is no housing shortage in the city, unless we decide that semi-pakka and kacha houses are not considered as houses. Even they amount to only about 5% of total houses, calling into question the often given figures for Karachi’s annual housing demand at 120,000 houses per year. Between the last two censuses, Karachi’s population increased by 6,168,576 – which, at the rate of six persons per household, works out to about 54,000 houses, i.e. half of the officially estimated demand.

There is no differentiation in census data between houses in katchi abadis and those in planned areas. Without katchi abadi figures, it is difficult to judge housing conditions in the city. However, it is clear that rentals have increased by 8.27% between 1981 and 2017, and that today, over 15% of houses are owned by women – the highest being in district Central (21.12%). This is in contrast to 7.65% for the whole of Sindh (including Karachi), while that of Pakistan is 6.99%. Rented housing is 34.67% for Karachi as opposed to 14.05% for the whole of Sindh and 11.35% for Pakistan.

Congestion in housing is best understood by persons per housing unit, persons per room, and houses with 1 room. Between the 1998 and 2017 censuses, there has been a minor decrease in the figures for persons per housing unit and persons per room. However, the change in houses with 1 room has been substantial, going down from 30.09% to 25.74% in Karachi. But these figures can be deceptive because while certain higher-income areas have improved, studies tell us that conditions in the old katchi abadis have become unbearably congested.

5. Utilities 

Table 5.1 (Potable Water Facilities in Housing Units – Karachi):

Table 5.2 (Electricity Connections in Housing Units – Karachi):

 Between the 1981 and 1998 censuses, the number of shared bathrooms and kitchens decreased – the former by 19.13% and the latter by 12.92%. This means that a larger number of Karachi families had these facilities within their homes. However, access to potable water has been continuously decreasing for Karachi between 1981 and 2017, falling from 94.91% to 83.37%, whereas the opposite trend has emerged in Lahore, with the figures increasing from 54.97% to 88.99% during the same period of time. On the whole, the figures related to utility connections for Karachi are similar to those of Lahore. In % terms, they are in the 90s – barring potable water, which is in the 80s. However, we know that water supply is not available to the citizens, connections or no connections, and that electricity (especially in the summers) is subject to loadshedding from anything between 3 to 14 hours per day in various parts of the city. As such the figures on utilities do not give us the right picture.

6 Employment 

It is interesting to compare employment rates in Lahore and Karachi so as to get a real understanding of the Karachi situation. Employment rates in 1981 were higher in Lahore (at 35.23%), but for women they were higher in Karachi (at 3.71%). In 1998, things changed: they were higher in Karachi for all categories. In 2017, the Lahore employment rates increased to 34.45% while in Karachi they were 33.18%. A detailed study is required to identify trends and their causes and repercussions. It is also important to note that in both cities, employment figures fell between 1981 and 1998, and then rose again in 2017, but not to the level of 1981. These differences between the two cities tell us that Karachi has not been able to maintain its rate of economic growth and social upward mobility. For women, though, 1981 stats fell in 1998, and then rose again in 2017 to levels higher than those of 1981.

It’s also important to note that these stats include figures on the formally employed population, as well as unpaid family work – of which the latter is largely performed by women in most households. The census data, however, doesn’t reflect this as only about 7% of women are shown as employed in both Karachi and Lahore.

7. Takeaways:

1. Between 1981 and 1998, Karachi’s social indicators improved considerably. However, between 1998 and 2017, the rate of improvement slowed down. Meanwhile, between 1981 and 1998, Lahore’s social indicator figures were less than those of Karachi, but between 1998 and 2017, Lahore’s social indicators – especially literacy and educational attainment – not only improved significantly but are now well ahead of Karachi. The same holds true for housing and amenities such as water supply. The social indicators of both Lahore and Karachi, and their rate of improvement, are not even a patch on those of Delhi (especially those on literacy and educational attainment), except for Delhi’s low rate of divorce and more female marriages in % terms. Perhaps, it is time that we try and understand the reasons for these trends. Where have we gone wrong and why?
2. The positive change that had come by 1998 in women’s social indicators (such as decreasing rates of marriage) in Karachi couldn’t be continued. While education-related stats improved, they did so at a slower pace between 1998 and 2017 than they did between 1981 and 1998.
3. The categories in the census are not defined, due to which an analysis is difficult to make. For instance, there were issues with categories such as ‘housing units’ and ‘households’. A ‘household’ is defined as “a group of persons living together who have collective arrangements for cooking/eating”. While the 2017 census data reports 97.7% housing units with a kitchen, the number of housing units is equal to that of households (implying that there is no shortage of “households” in Karachi).
4. As mentioned above, the census is without specified definitions. When looking at the census, we are not told what comprises housekeeping as compared to “unpaid family labour”. In the previous censuses, a number of categories were listed under “employment”, and an “economically active” population was defined. Female employment figures are incorrect because large numbers of females working on contract basis (with raw materials provided by middle men or otherwise) consider themselves to be housewives and do not classify themselves as employed.
5. The number of people living in informal settlements needs to be stated categorically, especially since according to census figures, utilities like gas, electricity, and water have been provided to nearly 95% of housing units in Karachi.
6. Generally the statistics given for utility provision seem improbable, as their percentages are in the 90s. Yet, according to the print and electronic media, there are long periods of electricity outages, absence of gas in the pipelines, absence of water in certain localities for weeks on end, and overflowing sewage. Is there any way in which the census can take count of this? It may help us understand the gravity and scale of the situation.
7. A major change has taken place in the registration of demographic data in the 2017 census. This is that for the first time in Pakistan’s census history, the transgender population has been recognized and figures for them provided. However, because of an absence of knowledge and social acceptance, the community’s representatives feel that it is grossly understated.

Arif Hasan is an architect.

Email: arifhasan37@gmail.com

Number: 03062691736

Web: Arif Hasan | Architecture & Research | Pakistan

Amal Hashim is a Social Development & Policy graduate whose primary interests lie in researching the cultural and tangible heritage of Karachi.

Email: amal97.hashim@gmail.com

Dhuha Alvi is a Social Development & Policy student who enjoys researching about the intersections of gender and class with politics.

Email: zohaalvi.21@gmail.com

^[1] All statistics have been derived from the population censuses of Karachi, Lahore and Delhi

^[2] Largest Cities in the World: 2016 | Newgeography.com

^[3]

^[4] These figures show the number of people who have completed a certain level of education

^[5] ‘Overall’ for primary education is the population aged 10 and above, for matric and undergrad it’s 15 and above, and for master’s it’s 20 and above