about us
  areas of expertise
  our projects
  ideas & resources


  Index of Articles          Index of Perspectives            Next Article


Why Decennial Censuses Won’t Do

Omkar Goswami


Regardless of what the BJP and the Left may tell you, the British left us many great administrative legacies. One of these was creating machineries for regularly gathering facts. Often this may have had more to do with maintaining control and strengthening the iron frame. Nevertheless, the fact is that from the mid-19th century, India was systematically and scientific ‘mapped’ and ‘quantified’ more than ever before.


One of the greatest legacies of this information collecting drive was the decennial Census of India. What began modestly as the 1871-72 Census soon became a regular once-in-ten-years affair. Myriads of babus and natives, carefully coached and tutored by the sahibs, fanned out across malarial, tiger prowling and snake-infested India to diligently count the number of people, their dwellings and, over time, their economic and social conditions.


The first serious census enumeration was carried out in 1881 and became a three-volume report comprising 751 closely printed pages and 33 tables. Thereafter, the census has been religiously conducted every ten years — 1891, 1901, 1911, 1921, 1931, an abridged version in 1941 because of World War II, 1951, 1961, 1971, 1981, 1991 and the most recent in 2001. Some of these reports are absolute classics such as those of 1901, 1921, 1931, as are the old District Gazetteers and Survey and Settlement Reports of the various districts of India.


The 2001 Census is by far the most exhaustive and, given the times, most comprehensively digitised. Since, in theory and substantively in fact, every household is enumerated, it represents the largest single database of various attributes of India’s population. It is also the largest database on rural India.


What makes the 2001 Census particularly useful is that over and above basic data such as number of men, women, children, households, sex ratio, literacy, religion, caste, SC and ST and the like, it has several fascinating economic ‘add-ons’. For instance, it has household level, tehsil-level, district-level and state-level data on the types of dwellings and the number of rooms, their water sources, whether these have kitchens, toilets and bathrooms; how many households have bank or post office accounts, own TVs, scooters and motorcycles, cars, have telephones, use LPG or other forms of energy for cooking, have electricity connections, and the like. It also has village directories which give extremely interesting data on every single village of India.


These fabulous goodies are being released every month and are making data junkies like me salivate with anticipation. My friend Rama Bijapurkar, three other research colleagues and I have been gleefully going through the data and buying everything that is coming out of the office of the Registrar General and Census Commissioner of India. And let me tell you, it is the most fascinating, elaborate and comprehensive data set that you could ever get hold of about this country. Of course, you would need to know how to organise and use such vast amounts of data.


That brings me to the main point of this article. It was all very good up to the pre-liberalisation era to conduct a nationwide census once every ten years. That won’t do any longer. India is changing so rapidly — and rural India in such unimaginable ways — that we need detailed census data on various economic and social indicators at least once every five years. For instance, the data on telephone penetration in the 2001 Census is already woefully outdated; so, too, in all likelihood, the data on two-wheeler and TV ownership. Today, economists, policy-makers, politicians, market researchers, analysts, corporate planners all need exhaustive, household-level data at least once every five years.


It doesn’t need an excessive increase in budgetary outlay to conduct a full fledged census once every five years. In any case, it is vitally necessary. Moreover, the Census Commissioner can, as he is trying to do, recoup a large part of his costs by selling customised data products to various constituents. It would be a great example of using commerce to cover the cost of information gathering.


Mr. Prime Minister, don’t even think about it. Just order a quinquennial census, beginning 2006. For the sake of India, it can, and must be done.                     


Published: Business Standard, October 2005


                 Index of Articles          Index of Perspectives            Next Article