The last caste census was in 1931. A look back at its findings

While the then Congress-led UPA government had in 2011 initiated the Socio Economic and Caste Census (SECC), its specific data on caste was never released.

Before the SECC, the last time India conducted a Caste Census was in 1931. The nearly century-old exercise that last counted castes in a Census in India gives a good idea of the challenges the enumerators can face in any fresh effort, plus the complexities of the exercise.

The 1931 Census that counted castes was conducted by the colonial British government, and was the first such exercise after the 1901 Census. The caste section put the numbers of Other Backward Classes (OBC) at 52% of the then total 271 million population of the country. This figure became the basis of the Mandal Commission’s recommendation in 1980 to grant 27% reservations to OBCs in education and government jobs, which was implemented only in 1990.

Chart 1: Largest communities overall

J H Hutton, the Census Commissioner at the time, countered those who argued against adding caste to the Census exercise saying that “the mere act of labelling persons belonging to a caste tends to perpetuate the system”. Hutton’s logic was that “it is impossible to get rid of any institution by ignoring its existence like the proverbial ostrich”.

“It is difficult to see why the record of a fact that actually exists should tend to stabilise that existence,” Hutton wrote in the 1931 Census. “Caste is still of vital consideration in the structure of Indian society… It impinges in innumerable ways on questions not only of race and religion but also of economics, since it still goes far to determine the occupation, society and conjugaI life of every individual born into its sphere.”

Chart 2: Largest communities by province

But Hutton and his team ran into a series of problems while enumerating caste. Hutton enumerated some of them – from “a wave of non-cooperation, and the (salt) march of Mr Gandhi and his contrabandistas” to the Congress observing a “Census Boycott Sunday”, to numerous local-level movements that hampered efforts.

Besides, over the course of previous Census exercises, the methodology on caste had undergone several changes. For instance, in 1881, only groups with more than a lakh population were counted. In 1901, Census Commissioner H H Risley decided to use the “varna hierarchy” system, sparking numerous movements by caste groups who viewed the Census as a vehicle to move up the social order.

In the book “Caste, Politics, and the Raj”, historian Sekhar Bandyopahdyay writes that at the time, the Census was viewed not merely as a population count for each caste, but as a way “to fix the relative status of different castes and to deal with question of social superiority… (which) gave rise to a considerable agitation both at organised and unorganised levels”.

R B Bhagat, a professor at the International Institute for Population Sciences in Mumbai, wrote for the Economic and Political Weekly: “… Many lower caste people represented themselves as higher castes in order to raise their social status. In the Census, the underprivileged found an opportunity to express their aspiration and if possible to acquire new identity through enumeration”.

Hutton was openly critical of using the varna hierarchy in the Census. “All the subsequent Census officers in India must have cursed the day when it occurred to Risley … to attempt to draw up a list of caste according to their rank in the society. He failed, but the result of his attempt is as troublesome as if he has succeeded,” Hutton wrote.

So, in 1931, occupation rather than varna was used to classify castes. But this model had its own pitfalls. For one, it was unable to reconcile the variations in an occupational group’s social standing across regions – for instance, Hutton noted that “cultivation in northern India is a most respectable occupation, whereas in certain parts of southern India it is largely associated with the ‘exterior’ castes’”.

“Admittedly this method… is far from being entirely satisfactory, since it can only recognise traditional occupation… and cannot simultaneously recognise more than one of several traditional occupations for the same caste,” Hutton wrote, but added that grouping castes roughly by occupation “also avoids any semblance of arrangement by order of social precedence”.

While Hutton’s use of the occupational model addressed the question of defining caste, the 1931 Census still was unable to fully account for the fluidity of caste identity and the variations in the names of groups across regions, says Ayan Guha, a research fellow at the University of Sussex who has written on the history of caste.

“The first problem was to define caste. You have to agree on the traits that make certain groups a caste and certain groups not a caste… The second problem was that caste has a lot of fluidity… The third problem was the standardisation of names – same castes with different names in different regions,” Guha says.

Guha adds that Hutton had also flagged the problem of the dynamic nature of identity claims with groups changing caste identities from Census to Census. “A caste group that was Rajput in the last Census is now (in 1931) claiming to be Brahmin. This also happens at the provincial level – a particular caste group can claim different identities in different provinces,” he says.

In fact, an official in Madras noted in the 1931 Census report, “Sorting for caste is really worthless unless nomenclature is sufficiently fixed to render the resulting totals close and reliable approximations. Had caste terminology the stability of religious returns, caste sorting might be worthwhile. With the fluidity of present appellations it is certainly not…”

Guha also points out “there was no uniform approach towards classification criteria for ordering of caste groups”. “It depended on provincial census commissioners and their subjective assessment,” he says, adding that if the objective for counting caste is affirmative action, “you need to agree upon clear criteria to ascertain the social position of caste groups, but such an agreement is likely to be elusive”.

Guha says that if a Caste Census is to be conducted today, it is likely to run into the same issues as the 1931 Census. In particular, the fluidity of caste identity – from one group seeking to be identified as a tribe, like the Meiteis in Manipur, to some groups fusing over time, like herder communities coming together as Yadavs, to the fission of some castes – will likely pose a challenge.

“Those who are saying do (a Caste Census), they should think about how to do it. We already know how the colonial Census had issues with caste. That debate on the methodological dimensions of caste enumeration, its fluidity and the boundaries between groups is missing,” Guha says.