Remember census?

Until a few weeks ago, Pakistan Bureau of Statistics (PBS) was all set to release the preliminary results of the much-awaited findings of the much-delayed population census finally held earlier this year. The preliminary results, BR Research is told, are ready; so is its report; and the plan was to release it publicly after a nod from the Council of Common Interests (CCI) that was supposed to meet at the end of July 2017.

But then Panama Judgment happened; and the cabinet stood dissolved. Under the new cabinet, statistics division has been taken away from Ishaq Dar – remember Dar was previously the minister for Finance, Statistics and Privatization – and made into a separate ministry under Kamran Michael, who had held the portfolio of Ports and Shipping until recently.

The disaggregating of finance and statistics ten months before the election seems to be a classic example of patronage politics. But that’s a separate matter. Meanwhile, Kamran held a meeting at the PBS yesterday, and sources say he discussed Nawaz Shariff’s rally more than he discussed the census or the bureau’s other plans.

The good news is that Kamran has asked for a briefing on preliminary census results and other related affairs next week, after which the matter will be put up with the CCI, if it is indeed convened in such a politically hostile environment. Ergo; keeping in mind the political environment, and the upcoming Eid holidays it would be no less than a pleasant surprise if the results are released anytime before mid-September 2017.

If things go as per the plans, the complete results will be released by April or May 2018, PBS sources say. But before that another pleasant surprise would be the holding of post-evaluation survey (PES) or census evaluation survey as it is also often called.

According to good census practices globally advocated by the likes of UN stats, a PES is an independent survey that replicates a census. The survey results are compared with census results, permitting estimates to be made of coverage and content errors. The former refer to people missed in the census or erroneously included, whereas the latter evaluate response quality of selected questions.

However, as was the case in 1998, this time around as well, the PES is not on the cards. Sources in the PBS say that the logic of 1998 census was applied this time as well. That logic being: because the army was involved in both 1998 and this year census, the risks of errors is very minimal if not none. And because the PES will not use the army’s services, the PES may in fact distort the picture.

There are three reasons to be skeptical of this view. First, why must the society take a leap of faith that the presence of the army during the census has ensured zero (or globally acceptable normal level of) errors in the census? The PES in fact should be done to assess whether or not the presence of army did in fact reduce these technical errors that have nothing to do with the politicisation of field work that army’s presence was supposed to check.

Second, even if the presence of the army did mitigate the errors, it is likely to have only reduced coverage errors (see definition above); the presence of army is no guarantee that ‘content errors’ did not occur. Third, if indeed the army-logic was applied in 1998, as PBS sources say, why is that the most level headed people criticized the lack of PES in 1998 census?

Many months after the detailed results of the 98-census were released Pakistan Institute of Development Economics, the think tank of the country’s Planning Commission, wrote a widely well respected publication. In that publication, Muhammad Aslam Chaudhry, the then Census Commissioner of Population Census Organization, made the following remarks on the absence of PES:

“The censuses of 1972 and 1981 had their follow on post evaluation surveys whereas no such survey was conducted for the 1998 census. Therefore, it is no longer possible to ascertain their degree of reliability on the basis of such evaluation surveys. Census being a very large operation is subject to many errors resulting from numerous factors some of which are beyond the control of the census experts. No rigorous and in-depth analysis of census data has ever been carried out which could create confidence in data users. Without detailed analysis, it is very difficult to identify the weaknesses, their implications and to give concrete proposals for improvement in future census taking.”

According to international best practices, the PES should normally take place near enough to the census to ensure that people remember who was in the household on census day. That opportunity is slipping by the day.