Views
6 years ago

Collecting national statistics inexpensively

Published :

Updated :

Data, and its refined form information, are the lifeblood of sound decisions. Governments collect information constantly on different aspects of national life, mainly to administer laws and regulations.

Censuses - a full counting of a particular population - are expensive and difficult to mount. Besides, they are time consuming. For this reason, we mostly get to hear about population censuses. In contrast, researchers reach valid conclusions by surveying a tiny, but methodically selected, sample. Bear in mind how data are collected is as important as which information is collected.

Bangladesh carries out population censuses at ten-year intervals, a universal norm.

The Government of Bangladesh (GoB) carried out a census of commercial establishments in 2013. I do not know whether benefits outweighed the costs of such an elaborate and costly exercise. We should have thought of a less costly, albeit less perfect, alternative. A database of businesses in the formal sector is not difficult to construct from records maintained by income tax, VAT, customs authorities and Bangladesh Bank. While putting together such a database may be challenging at first, it is not impossible. Because information has to be culled from different sources, errors and omissions will tend to creep in. ICT (information and communications technology) should come to our rescue.

Looking beyond, it is also possible to mine data from numerous trade organisations. We note that moves are afoot to regulate these bodies more tightly. The above sources can only get us information from the 'formal' sector. The vast and shadowy 'informal' sector does not afford itself to be captured easily. Guesstimate is a better option.

The government carried out a census of slum areas and floating population in 2014. How can we and why should we take a census of a group that is in a constant state of flux? What are we trying to achieve - supplying clean water, sanitation or health services? I say this ruefully because local authorities in Bangladesh are not known for proactively engaging with citizens, less so who are powerless and disenfranchised. 

I do not wish to denigrate the Bangladesh Bureau of Statistics (BBS) which is, by and large, doing a fine job. Evidence-based policy decisions depend on accurate, timely and relevant data. What we have to bear in mind though is cost-effectiveness. So, what am I driving at?

As a beginning, let us choose between fifteen to eighteen union parishads including their city-based counterparts, wards, all over Bangladesh. In light of the rapid pace urbanisation, we may pick about six or seven wards from city corporations. The rest eleven units (UPs) are to be chosen from zilla parishads and upazillas. I do not claim this a scientific method. The idea is to collect needed data from a limited number of areas for up to five years able to reflect ground realities for most of Bangladesh.

With trained and experienced enumerators, it is possible to collect data on agricultural, cultural, economic, social and many other aspects of the community. Let us take an example. From a population of say, 60,000 people living in a union parishad a sample of 150 (0.25 per cent) can be drawn to find out average monthly expenses of these individuals. This is a vicarious way of knowing their incomes!

The reliability and validity of such data will depend on whether the sample was random. A   scientific sample should be relatively easy to compile from a population numbering 60,000 only. The last population census should be taken as the base (or frame in statistical lingo!).

It won't be unrealistic to assume that results from a random sample taken from a UP in Daudkandi applies equally well to a UP in Barguna, of course with a margin of error. An itinerant team of enumerators should be able to collect data from the above administrative units by turn. A localised problem, say fish resources in haor areas, will need a separate study however. The time has come to make the most of an excellent alternative - drones.

In an overwhelming number of cases, governments don't need comprehensive databases or to pay a king's ransom to make optimal decisions. Policies and decisions can be changed in light of new experiences and better information.    

Raihan Amin is a Visiting Faculty, International University of Business Agriculture & Technology

[email protected]

 

 

 

Share this news