Home / Opinion / Columns /  India’s caste census has turned into a political football

Counting people should be the easiest statistical exercise in the world. But as the history of failed censuses across the world show, many countries find it very challenging to accomplish this task. Consider Nigeria, whose true population remains a matter of uncertainty till this day.

Independent Nigeria’s first census in 1962 was rejected because it showed that the numerical strength of the northern region, populated by its most dominant tribes, was lower than expected. A fresh census was ordered in 1963 and the north’s hegemony was restored after 8.5 million people were ‘discovered’ there. Censuses in the country have been irregular and controversial since then. Nigerian analyst Fewi Fawehinmi noted that all subsequent censuses (1973, 1991, 2006) have followed the ‘1963 formula’ in allocating people across regions.

Nigeria is not an isolated example. In several fragmented societies, a census is almost like an election to be ‘won’ by a particular ethnic group, wrote the political scientist Donald Horowitz in his oft-cited 1985 book, Ethnic Groups in Conflict. “It comes as no surprise then, that in Tanzania, ethnic questions have been left off the census in 1978; that Kurdish demands for a new census in Iraq were ignored; that the Belgians cannot agree on how to phrase the ethnic question; and that the Lebanese, fearful that changed ethnic ratios would upset the quotas on which state institutions were founded, have conducted no census since 1932," he wrote.

If these examples seem too foreign to you, consider the case of Nagaland, the only Indian state to report a decline in population in India’s 2011 census. The decline was untrue, but the actual census count was perhaps closer to the truth than previous census counts in the state, economists Ankush Agrawal and Vikas Kumar wrote in their 2020 book, Numbers in India’s Periphery: The Political Economy of Government Statistics.

As in the case of Nigeria, the state government of Nagaland allocated public funds based on district headcounts. Since most districts were dominated by a single Naga tribe, each tribe (and district) had a strong urge to overstate its numbers. New concerns came up ahead of the 2001 census. An impending delimitation exercise to map new constituencies and a ceasefire agreement militants signed with the government threatened to upset the balance of power among various Naga tribes.

The 2001 census results, showing a 65% jump in Nagaland’s population, were so bizarre that the state government felt compelled to reject them. The miscount prompted the government to stop using district headcounts as a criteria for allotting funds. It also led to wide-ranging deliberations with civil society groups to ensure an honest headcount in the next census. The 2011 census headcount was indeed more realistic; it led to a reported drop in the population.

These examples from different parts of the world offer a common lesson: when the economic stakes are high, it is hard to prevent data corruption. That’s the challenge a caste census in the country faces today. Given that a caste census would be used to decide future reservation and welfare benefits, respondents may lie about their jati (or sub-caste) or their material possessions. Enumerators too may have strong incentives to portray their own community as especially deprived. In regions dominated by a particular caste group, there may well be attempts to inflate that group’s headcount at the expense of others. Statistically speaking, ‘non-sampling errors’ are likely to be very high.

While the Indian census does record counts of Scheduled Castes and Tribes (SCs and STs), it does not count Other Backward Classes (OBCs) and their constituent caste groups. Given that OBCs are the dominant social group across several states, and several dominant communities want to be classified as OBCs in other states, the contestation around caste and deprivation in India primarily centres around who should be included in the OBC list.

The Socio-Economic Caste Census (SECC) 2011 dataset was supposed to provide detailed data, but the data was in such poor shape that its individual caste tallies were never published. The SECC dataset was an improvement over earlier BPL (below poverty line) censuses in identifying deprivation levels. But as a Plain Facts column by Arjun Srinivas (‘The targeting challenge in India’s welfare programmes’’, 18 May 2019) showed, the SECC may have overestimated deprivation levels in several parts of India.

Politicians in Bihar and elsewhere who have been demanding a caste census have given no indication of how they expect these challenges to be met in a fresh caste census. Conducting a caste census without addressing the fundamental problem of mis-representation will lead to spurious results. Such data would then be challenged by aggrieved social groups and perhaps become the subject of long-drawn court battles.

The idea that welfare and reservation benefits should be delivered to those who need them most is unobjectionable. The challenge lies in putting this idea into practice. No Indian politician or bureaucrat in any part of the country has been able to come up with a foolproof answer to the problem.

A caste census may seem like an alluring answer to recurrent agitations for caste-based reservations in the country. But as the history of ethnic counts suggests, a census is likely to be repudiated or compromised if it is viewed primarily as a contest for state patronage among social groups.

Pramit Bhattacharya is a Chennai-based journalist. His Twitter handle is pramit_b

Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.
Recommended For You
Edit Profile
Get alerts on WhatsApp
Set Preferences My ReadsFeedbackRedeem a Gift CardLogout