Does caste indicate deprivation?
  Pradipta Chaudhury

back to issue

IN much of scholarly discourse the institution of caste is commonly taken to be the embodiment of fundamental socio-economic inequalities. In particular, during the last decade a consensus has emerged across the entire gamut of political, legal and intellectual opinion regarding the use of caste as an appropriate criterion in public policy oriented towards positive discrimination. Every now and then, particularly before elections, caste groups demand their inclusion in the lists of the ‘backwards’. At the same time, political parties dangle the carrot of the benefits of reservation to more groups and promise to extend the policy to the private sector.

Caste remains a major force in Indian politics. More caste organizations are coming into existence. Overt and covert appeals are being made to caste sentiments in all spheres of public life. The increasing mobilisation of castes is being approvingly described as the emergence of caste ‘identity’. The politics of caste identity has been unanimously hailed in the media, among intellectuals and politicians of all varieties, as a move towards true equality. It is argued that the politics of caste is secular and serves a bulwark against the politics of religion. It is suggested that the recent rise of political leaders belonging to lower castes in northern India and the implementation of reservations for backward castes actually amounts to a silent revolution.

Does the ritual rank of a caste indicate the degree of deprivation actually suffered by its members? Neither the Kalelkar Report, nor the Mandal Report, nor the reports of the backward classes commissions at the state level have so far demonstrated that deprivation was related to the ritual status of a caste. These reports generally emphasized that the higher castes have cornered a disproportionately large share in the bureaucracy and that the presence of low castes in class one and class two jobs was nominal before the implementation of reservations schemes. Not only is caste virtually used as the sole criterion in public policy oriented towards positive discrimination, but categories like OBCs (Other Backward Castes) and SCs (Scheduled Castes) are treated as essentially homogeneous. It is simply assumed that the great bulk of the population of each of the categories suffers from a uniformly high degree of deprivation without answering the question raised at the beginning of this paragraph.

The relation between caste and deprivation is an issue that should be addressed at the macro-level; village studies will not suffice. The results of village surveys typically contradict each other, as is to be expected in a large country characterised by great diversity. The only information available at the macro-level, where castes are treated separately, was collected through the census operations during British rule.1 While scholars are fond of discussing the colonial power’s motives and methods of collection, classification and inadequacies of caste data, the data remains largely unused. Despite all its inadequacies, this data set can be used to throw light on the issue raised here.



If caste is a good indicator of deprivation now, it should have been a better index in the past. Let us consider Uttar Pradesh. It contains some of the most backward regions of India. Here the link between caste and deprivation should be stronger than the more advanced regions of the country. Information on the social, material and educational conditions of castes in U.P. is available only at the beginning of the last century. We get the ritual rank, the literacy rate in Hindi and the work participation rate for 42 castes from the census tables of 1901 and 1911. There are eight high or ‘twice born’ castes, 27 middle ranking or intermediate and ‘shudra’ castes, all of which are now considered OBCs, and seven ‘untouchables’, later designated SCs.2 In 1911 the average literacy rate was about 11% for the high castes. It was only 1% for the OBCs and 0.13% for the SCs. Thus, literacy rate seems to be strongly inversely associated with ritual rank.

Since there is no direct evidence on income or wealth of members of a caste, we have to devise an appropriate index which may enable us to use the available data and compare the average economic status of castes. The work participation rate, defined as the proportion of workers in the population, can be used as an inverse indicator of the economic status of a caste. The commonsense reasoning underlying this is simple: In a traditional economy, with low rates of literacy and industrialisation, the poorer families have to send a greater proportion of their members (namely, women and children) to seek work than the better-off, well-to-do and rich families which send their children to school and confine their women to the home.

This argument is actually built on our empirical observations on U.P. During the early decades of the 20th century the work participation rate was a good index to capture the differences in the economic conditions over regions of U.P. as well as between groups within each region. In 1911 the average work participation rate was 42% for the high castes, 54% for the OBCs and 57.5% for the SCs. Thus, it would again appear that caste is a very good indicator of deprivation. In fact, such aggregate statistics are usually provided in support of caste-based public policy.3



These averages actually conceal the enormous heterogeneity within the OBCs and the SCs. There is a great deal of variation in the literacy as well as work participation rates within each category of castes. The literacy rates for the OBCs vary between 8% and 0.14%. In fact, in the literacy rankings three of the OBCs are placed among the top seven castes while two are placed among the bottom seven. The literacy rates for the SCs vary within a much smaller range, between 0.48 and 0.11%.



The situation with respect to the economic condition of castes is more acute. The work participation rates for the OBCs lie in the range of 40 and 67. On one hand, four of the OBCs figure in the top eight places on the scale of economic status. On the other hand, five of this figure among the bottom-most seven places. Actually the three poorest castes (namely, Bhar, Koeri and Kewat) belong to the OBCs. Likewise, the economic status of the SCs varies a great deal; the work participation rates lie within a range of 44 (for Khatik) and 64 (for Dusadh). Thus, even in a backward region like U.P. at the beginning of the 20th century, there were large variations in the literacy rates and economic conditions of castes that were later pooled together and treated as homogeneous categories.

With respect to literacy rate, three OBCs, namely, Sonar, Halwai and Kalwar, were ahead of four high castes, namely, Rajput, Taga, Bhat and Kandu. Similarly, with respect to economic status, five OBCs, namely, Sonar, Jat, Gujar, Kisan and Mali, were better off than Brahman and Rajput, the two most numerous high castes, which accounted for one-fifth of the Hindu population. Two SCs, namely, Khatik and Dusadh, had higher literacy rates than many OBCs.

In the economic hierarchy two SCs, namely, Khatik and Dhanuk, were placed in the top half. Conversely, five OBCs, namely, Luniya, Barai, Bhar, Koeri and Kewat, figured among the seven lowest placed castes in the economic scale, along with two SCs. None of the large SCs, namely, Chamar, Pasi and Dhobi, figured in the lowest rungs of the economic hierarchy. In fact, the average economic status of Chamar, the most numerous of all castes in U.P., was not lower than that of Ahir, the most numerous among the OBCs. Thus, the high castes, the OBCs, and the SCs were highly heterogeneous in terms of economic status and access to literacy.



It is unambiguous that even in a backward and traditional agrarian society such as U.P. during the early decades of the 20th century, ritual rank of a caste was not a good indicator of its literacy or economic status. The formation of three administrative categories of castes, based on past ritual status only succeeds in hiding the glaring intragroup disparities. High ritual rank could not secure some of the upper castes against low economic status. Similarly, low ritual status did not prevent large sections of Jat, Gujar, Sonar, Kisan and Mali from attaining prosperity. Caste did not preclude the upward economic mobility of a section of the untouchables. Even with ‘5000 years old tradition of learning’, the Brahman population of U.P. could not reach an average of 12% literacy by 1911; they were not the most literate of castes.

Advocates of caste politics argue that the problem will be solved if the OBCs or SCs are arranged according to the degree of backwardness and split into subgroups such as ‘more-backward’ and ‘most-backward’ and sub-quotas created within the total quota. However, the economic status of households varies a great deal within each caste. In a caste, several economic classes exist. Marginal and small peasants, and landless labourers constitute the bulk of the population in each caste. At the same time, every caste contains a section, varying in size, of well-to-do families.

The 1888 Dufferin Enquiry Reports on the condition of the lower classes of the population in India showed that in eastern U.P., the castes of Brahman, Bhuinhar and Rajput contained sections, which though not landless, were worse off than day labourers, were in debt and suffered from insufficiency of food and clothing in normal times. This report also showed that in western U.P., many Chamar families cultivated landholdings of 10 acres or more in size while others of this caste were landless labourers.4 In several non-twice-born castes, for example, Jat, Kurmi and Kalwar, the size of the upper class elite was considerable. The existence of relatively prosperous traders, contractors, and manufacturers belonging to the caste of Chamar in the Agra-Aligarh area of western U.P. during the early decades of the 20th century is well known. Improvement in material condition and educational status of sections of Chamar and other untouchables, because of employment in the government sector, in army cantonments, municipalities and so on, in the cities of Agra, Kanpur and Allahabad is also well documented.5



The census tables on ‘occupational distribution of castes’6 in early 20th century demonstrate that each caste contained landless labourers, cultivators as well as landlords. Some castes were sharply split over occupations, for example, Chamar, the largest caste of U.P., which is believed to be traditionally landless. The workers of this caste were about equally reported as labourers and cultivators, between 35 to 40% in each. In contrast, more than 75% of workers belonging to the caste of Bhangi, later known as Balmiki, were scavengers.



The cultivators, the single largest occupational group in most castes, were highly differentiated in terms of size and economic status. A sample of 17,135 agricultural holdings covering 82,176 acres, distributed over size of holding and caste in Bahraich district in 1939 shows that about one-third of the holdings belonging to the upper castes were of 2.5 acres or less in size. The same was the case with the caste of Kurmi, a backward caste. Such holdings accounted for half of the total number of holdings in case of Kachi, Murao and all other agriculturist castes. The holdings of size 2.5 to five acres comprised one-fourth of all holdings in case of upper castes as well as Kurmi, but about 30% in case of the rest of the castes.

In fact, the size distributions were very similar for the upper castes and Kurmi. In each of these castes, six to eight per cent of holdings were more than 20 acres in size. 1.7% of upper caste holdings and 0.6% of Kurmi holdings were more than 50 acres in size. On one hand the small and marginal peasants formed a majority of households in each caste. On the other hand, the lower castes, like the higher ones, contained many rich peasants. Thus there was enormous intra-caste variation in economic condition. Accordingly, the material interests of different classes belonging to the same caste would differ.



Since the visible heterogeneity within a caste cannot be easily brushed aside, the proponents of caste politics argue that it is not economic but social backwardness from which these castes have historically suffered that warrants reservations. Did all the lower castes suffer from an equal degree of ritual handicap? Actually, there was an elaborate gradation and hierarchy among the intermediate or shudra and even the untouchable castes, which governed interaction between them and kept inter-caste socialisation to a minimum. The rich households belonging to a low caste tried to imitate the customs and rituals of the upper castes such as child marriage, prevention of widow remarriage and payment of dowry for marriage.

Occasionally, prosperous sections of castes broke away to form new castes and claimed higher ritual rank. By and large an affluent caste succeeded in raising its position in the ritual hierarchy, for instance, Kayastha and Jat, both of which rose from the ranks of shudras to be near Rajputs. Members of the Jat caste claimed their caste to be of twice-born rank and they wanted to be classified at par with Rajputs in the census of 1901. In some districts the others accepted their claims. But the provincial committee, which was drawing up the ritual hierarchy, accorded them a rank higher than the shudra castes, while not accepting their claim to be twice born.

Similarly, Kalwars claimed Vaishya status. Surreptitiously, many of them got enumerated in the censuses as Vaishya/Bania or even Rajput. Consequently, the share of the Kalwar caste in the population of U.P. declined significantly over the censuses. The Kurmi elite followed a variety of paths. Their caste association demanded high ritual rank, at par with the Rajputs. Some of them got themselves enumerated as Rajputs. As a result, the share of the caste in the total population declined over the censuses. In Gorakhpur district, the landowners and large cultivators broke away from the parent caste (Kurmi) during the second and third decades of the 20th century. They formed a new caste called Sainthwar, which was the name of a sub-caste of the Kurmis.

Before the census of 1931, well-off Chamars in west U.P. broke away to form a new caste named Jatav, which was the name of a sub-caste. They claimed to be Rajputs and demanded to be enumerated in the census as Jatav-Rajputs. In fact, by 1931, barring a few most downtrodden castes, viz. Bhangi, all other non-twice-born castes had formed associations which claimed high ritual rank, at par with twice-born castes. This has given rise to a peculiar situation. The old ritual hierarchy has disintegrated because the elite of no caste concedes a higher ritual rank to any other caste. But there is no sense of equality in the caste system since none accepts the lower as equal on the social-ritual scale. Even now, inter-caste marriages between OBCs or SCs are almost absent in the rural areas. The rare instances in urban areas usually occur within the same economic class.



Evidently, within a caste, the kinds and degrees of deprivation varied. Within a low caste, the upper income groups felt deprived of a high ritual status that would be commensurate with their economic status. They also felt deprived of education and jobs in the government apparatus. The caste associations formed by them articulated the demands of the upper class. At the same time, the caste also contained a large section of marginal and small peasants, who were oppressed by rent and debt obligations, as well as landless labourers lying at the bottom of the economic scale. In the history of the low or backward caste associations, demands for redistribution of land, or demands for minimum wages, or special measures for the benefit of the poor were virtually never put forward.



It is amazing that in the 21st century not only does caste continue to be the sole criterion in public policy oriented towards positive discrimination, but categories like OBCs and SCs continue to be treated as essentially homogeneous, despite the mass of evidence relating to early 20th century pointing to the contrary! The claim that the use of an income limit for identifying the ‘creamy layer’ among the OBCs would bring the benefits of reservation to the deserving is actually a hoax. The lower income limit of Rs 250,000 recently fixed by the central government would not even exclude the entire top ten per cent of the population.

Furthermore, when it is well known that incomes can be easily under-reported, there is no effort to use any other criterion which can be actually used to exclude the privileged. Obviously, the use of caste and caste-based reservations are designed for the absorption of the elite of the lower castes in the ruling classes.7 At the same time the use of caste in the public sphere effectively keeps the deprived masses politically divided and weak.


1. The censuses of 1911, 1921 and 1931 provide data on the occupational distribution of selected castes at the level of province or state.

2. Actually the castes were classified in 12 groups, depending upon their ritual or social practices, the rules and restrictions observed by them and their traditional occupations. The high castes were divided into six groups, the shudras into four and the untouchables into two groups. Castes within each group were ranked in order of precedence. See Census of India, 1901, v XVI, pt I, p 218-34, 248-53.

3. For example, see, S.K. Thorat and R.S. Deshpande, ‘Caste System and Economic Inequality: Economic Theory and Evidence’ in Ghanshyam Shah, ed., Dalit Identity and Politics (New Delhi and London, 2001), particularly p. 57-70.

4. 1888 Dufferin Enquiry Reports on the condition of the lower classes of the population in India, (available in the National Archives of India), Enclosures, letter from Collector, Ghazipur to Commissioner, Benaras Division, 10 April 1888, p. 134-136, letter from Collector, Mathura to Commissioner, Agra Division, 1 May 1888, p. 4-20 and note on Etah district by N. Crooke, Collector, 12 Jan. 1888, p. 31-100.

5. Owen M. Lynch, The Politics of Untouchability: Social Mobility and Social Change in a City of India (New York, 1969), and Nandini Gooptu, ‘Caste and Labour: Untouchable Social Movements in Urban Uttar Pradesh’, in Peter Robb, ed., Dalit Movements and the Meanings of Labor in India (Delhi, 1993), ch. 10.

6. See Census of India, 1911, v XV, pt II, Table XVI and Census of India, 1921, v XVI, pt II, Table XXI.

7. See Pradipta Chaudhury, ‘The "Creamy Layer": Political Economy of Reservations’, Economic and Political Weekly 39(20), 15 May 2004, p. 1989-1991.