Alert Sign Dear reader, online ads enable us to deliver the journalism you value. Please support us by taking a moment to turn off Adblock on

Alert Sign Dear reader, please upgrade to the latest version of IE to have a better reading experience


Census and consensus

January 04, 2014


THE prime minister recently announced that the population census will be held this year. The last population census was organised in 1998 by his previous government. The colonial tradition was to hold the decennial exercise in the year ending in one, such as 1951, 1961, etc. Pakistan kept pace till 1981 but the census due in 1991 was delayed for seven years and the next one has been pending for five.

On both occasions, it was deferred partly due to administrative challenges but mainly because the governments were not confident they could handle the political fallout.

A census generates numbers which serve as the basis for deciding the share of various groups in financial resources and political power. The shares for the four federating units are now almost decided and instituted in bodies like the National Finance Commission. But distribution within the provinces is now proving to be more contentious.

Sindh’s politics rests on maintaining a delicate balance of power between ethnic Sindhis and Mohajirs. It might need readjustments if the new census throws up somewhat different percentages. Similarly, groups in other provinces pitching for independent political identities, such as Seraikis in Punjab, Hazaras in Khyber Pakhtunkhwa and Pakhtuns in Balochistan, will also be keen to know the findings of the census which could have an impact on their claims and causes.

All groups aspiring for shares in state power and resources have been wary of census figures and complain of under-enumeration which they see as part of a conspiracy to deny them their due rights. While some of these misgivings can be blamed on the larger-than-life ambitions of these groups, a major chunk owes to the mistrust we all have in our governments’ ability to conduct any exercise that requires universal outreach. How can we trust the state machinery that has not been able to administer polio drops to all its toddlers in over two decades?

The upcoming exercise cannot escape such scepticism but there certainly are ways through which this could be addressed to a large extent. The General Statistics Act passed in May 2011 has merged three institutions, the Federal Bureau of Statistics, the Population Census Organisation and the Agriculture Census Organisation, into one body named the Pakistan Bureau of Statistics. This has put to rest the age-old problem of duplication of efforts.

An institutional stumbling block has been the definition of a basic unit for enumeration. An electoral block delineated by the Election Commission while preparing electoral rolls did not collate with the census block mapped by the Census Organisation. Then Nadra followed its own coding while issuing identity cards. As a result, the three could not benefit from each other’s data resources. Thankfully, there is agreement now on unifying the basic enumeration units and equally importantly, it is being done by the Bureau on Google’s satellite map.

These measures will make data collection more efficient but there is more to census than collection efficiency. The enumeration blocks are primarily defined for enumeration ease, whether or not various departments have unanimity on the limits of the block. They become the basic building blocks of electoral constituencies. It is not possible for a constituency to cut across an enumeration unit. So enumeration ease actually dictates the boundaries of political units. Shouldn’t it be the other way round?

It is important, from the point of view of democratisation of governance, that the administrative boundaries overlap perfectly with the political/electoral limits and the same should then be followed by enumerators.

Similarly, there are issues related to inclusion and formulation of other questions in the census questionnaire. For example, information regarding marital status is gathered in such a format that it is not possible to extract the average age of marriage, which for many purposes is vital. There are other gender-specific anomalies as well that need to be addressed. The 1998 census included, for the first time, questions that could lead to calculation of infant mortality rates and maternal mortality rates for districts and areas. The figures generated by this set, however, were totally contradictory to those collected by other institutions.

Even more important are politically sensitive questions. The census in 1981 enumerated Hindko speakers separately while Seraikis were put in ‘others’ column. The situation was reversed in the next census. These two and many other linguistic groups will certainly not ignore any such aberration now.

The General Statistics Act provides for the formation of user councils at national and provincial levels. Comprising representatives from the private sector, these bodies are assigned an advisory role. These provide an excellent opportunity to make the census exercise inclusive and comprehensive besides building consensus on all potentially controversial matters. If the government constitutes these in time and also ensures wide participation of political parties and civil society, it can lend the census greater legitimacy and pull it out of the morass it has been in since the past quarter of a century.

The first synchronous census in the subcontinent was conducted in 1881. This was the first occasion when our people were quantified in terms of religion, caste, creed, etc. These numbers were instrumental in helping the colonisers accentuate their divide and rule policies. If the governance under a fully functional democracy is to undergo a paradigm shift, it must start by reforming the way these numbers are collected.

The writer works with Punjab Lok Sujag, a research and advocacy group. He recently compiled a compendium on Pakistan’s electoral history since 1970.