ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

The Continuing Practice of Untouchability in India

Patterns and Mitigating Influences

Amit Thorat ( teaches economics at the Centre for the Study of Regional Development, Jawaharlal Nehru University, New Delhi. Omkar Joshi ( is a doctoral scholar at the Department of Sociology, University of Maryland, Washington, DC and is also working with the India Human Development project.


The caste (jati)-based practice of untouchability in India, shifting the focus from the victims of this practice, the ex-untouchables (Dalits), to the perpetrators, the non-Dalits is examined by identifying and disaggregating communities that continue to practise untouchability. The second wave of the India Human Development Survey data has been used to generate a socio-economic profile of those who practise untouchability in India, and check the hypothesis that households with a wider network outside the community than with one within the community are less likely to practise untouchability, and uses a logistic regression model to measure this effect at the all-India level.

The authors are grateful to their colleagues at the National Council of Applied Economic Research, New Delhi, and especially the India Human Development Survey team there, for helpful comments throughout the preparation of this article. An earlier version of this article was presented at the University of Brandeis, Massachusetts, Amherst; Jawaharlal Nehru University, New Delhi; and the Populations Association of America conference at San Diego, where the authors received several useful suggestions. The data have been obtained from the IHDS, 2012. This survey was jointly organised by researchers at the University of Maryland, Washington, DC and NCAER. The data collection was funded by grants R01HD041455, R01HD046166, and R01HD061408 from the National Institutes of Health and by a supplementary grant from the Ford Foundation. The Government of United Kingdom funded data management as part of its Knowledge Partnership Program, while the analysis was carried out with the aid of a grant from the International Development Research Centre, Ottawa, Canada.

India, the largest democracy of the world and the second most populated nation (comprising 17.5% of the world’s population) (Office of the Registrar General and Census Commissioner, India 2011), is also home to the largest concentration of Hindus in the world. According to the 2011 Census of India, the Hindu population is 79.8%. One of the defining features of the Hindu religion is the division of the Hindus into numerous jatis or castes. The Encyclopedia Britannica (2019) gives the following definition of the word jati:

Jati, also spelled “jat,” refers to caste in Hindu society. The term is derived from the Sanskrit jāta, “born” or “brought into existence,” and indicates a form of existence determined by birth. In Indian philosophy, jati (genus) describes any group of things that have generic characteristics in common. Sociologically, jati has come to be used universally to indicate a caste group among Hindus.

The word caste, on the other hand, derives from the Spanish/Portuguese word casta meaning race, lineage or breed. It was used formally for the first time in India by the British to identify and enumerate the various groups in India as part of their census exercises. One should remember that using the word caste to describe the many groups or jatis in India does not connote a racial differentiation between them, as the original meaning of the word caste might imply. Caste is not the same as race.1

Caste and Its Religious Roots

According to Hindu scriptures, all the people are born into a particular jati or caste. Currently, around 3,000 castes exist inIndia. Caste is hierarchical, hereditary and endogamous innature, and has historically been linked to specific occupations. However, some occupations like agriculture have traditionally been caste-neutral. Over time, the strict one-to-one correspondence between a person’s caste and their occupation has been reduced to some extent, and with the spread of education and the advent of urban migration, people can more or less choose their occupations.

A study of occupational concentration by caste groups (Singh and Thorat 2014) indicates that marginalised groups, such as the Dalits and the Adivasis, are disproportionately concentrated in skilled and semi-skilled occupations in the informal sector of the economy. In the light of studies on labour market discrimination (Thorat and Newman 2012), this indicates the existence of job discrimination in hiring in the private sector labour market. Specific types of jobs are therefore still largely linked to caste identities, particularly in the case of the lowest of the castes, the ex-untouchables, who are now known as Dalits—a new empowering term used to signify defiance and change.

These numerous jatis or castes constitute the real structure of the caste system, which varies in its form and nature regionally. The entire system of castes or the “superstructure” ideologically derives from the Chaturvarna system or the fourfold division of society. This religious theory of the origin of society (cosmogony) derives from the 19th-century hymn of the 10th Mandala of the Rig Veda,2the Purusha Shukta. This divides the society into four varnas or classes that are hierarchical in nature. On the top of this ranking are the priests (Brahmins), followed by the warriors and erstwhile rulers (Kshatriyas). The next to come are the farmers and merchants (Vaishyas), while the last in the hierarchy are the workers and craftsmen, among others (Shudras). These four castes thus form the fourfold classification of varnas.

However, as B R Ambedkar (1990) points out in his path-breaking work, Who Were the Shudras, this is not the only cosmogony in the Rig Veda, but there are others,3which are more secular and talk about the origin of man rather than social groups. He believes that the cosmogony for the origin of society was added later in the sacred text, to seek religious sanction for the division of the society into occupation groups for control and hegemony. A fifth group existed outside this fourfoldclassification, that of the non-classified (avarnas) who did work that was, and is still considered, physically and ritually polluting, such as cremation and the handling of dead bodies, removal and skinning of dead animals, removal and cleaning of human bodily fluids and excreta (manual scavenging) and basket weaving. These are similar to groups like the burakumin (hamlet people) in Japan, the baekjeong in Korea, ragyppa of Tibet (Passin 1955),and the cagots (pariah people) in France (Thomas 2008).

Each of these varnas or groups is comprised of numerous sub-castes or jatis. This classification is hierarchical such that social and economic rights vary across varnas. The Brahmins, who are at the top of the caste ladder, enjoy all social and economic rights, and as one moves down the ladder, these socioeconomic rights diminish progressively. However, it is mostly the social rights that decrease up to the third varna, the Vaishyas, as most of the other rights such as the right to education, right to ownership of land and the right to do business are open to them all.

However, the outcastes traditionally had no rights at all. They did not have the right to be educated, to own land, to conduct business, or to borrow and sell in markets, among other rights. Further, since the system was hereditary in nature, it meant that a person’s caste as well as the caste-dictatedoccupations were passed on from generation to generation. The unique feature of the outcastes is that they constitute the only group considered as “untouchables” and who were born “impure” due to the jobs that have been relegated to them and are considered physically and ritually polluting.

Caste and the Reservation Policy

In contrast to the above-mentioned religious classification of the Hindu society, the constitutional classification of the Indian population is, however, somewhat different. Under Article 341 of the Constitution, the ex-untouchables are now part of anofficial list or schedule, and benefit from the reservation of seats in state-run post-schooling educational institutions as well as in public sector jobs. These castes are now called the Scheduled Castes (SCs). While all the ex-untouchables who follow Hinduism are part of this list, those who have converted to Buddhism and Sikhism in order to escape discrimination are also included. However, those who have converted to Christianity and Islam are not part of the schedule. In recent times, these two communities too have been demanding the benefits of reservation, as they perceive that conversion has not helped them to shed their earlier caste identities. They allegedly continue to face exclusion and discrimination at the hands of both Hindus as well as the high-caste converts within their religion, which makes them feel socially and economically marginalised.

Similarly, under Article 342 of the Constitution, all tribes in the country (Adivasis or aborigines),4 now called Scheduled Tribes (STs), are also part of a schedule and can avail of the same benefits as the SCs. The tribals who either follow their own indigenous religion or have over time converted to other religions like Hinduism, Buddhism or Christianity are entitled to reservation benefits, irrespective of their religious leanings. Under the reservation policy, seats are also reserved in thenational and state legislative assemblies and panchayats for the SCs and STs. In addition, 15% and 8% of all job vacancies in the public sector and government-aided educational institutions, respectively, are reserved for the SCs and the STs.

Meanwhile, the Other Backward Classes (OBCs) have recently been recognised constitutionally as being historically excluded and discriminated against and have also secured reservation similar to the SCs and STs, to the tune of 27%. Although there are varied estimates for the OBC population, the Supreme Court has ruled that the overall reservation should not exceed 50%. Given that reservation for the SCs and STs together amounts to 23%, this puts an upper cap of 27% for the reservation for OBCs.

Recent studies have shown that reservation does help marginalised groups to rise economically, while also leading to efficiency gains. Studies have shown that the reservation of jobs has led to an at least five percentage point gain in regular salaried and wage employment for the SCs and STs (Borooah et al 2007). Similarly, a study of the Indian Railways, the largest federal employer in the world, shows that reservation for the SCs and STs does not reduce efficiency, but, in some cases, is seen to actually improve it (Deshpande and Weisskopf 2014).

Post-independence, as a direct consequence of legislative provisions and affirmative action policies, such as reservation, many of the ex-untouchable sub-castes (jatis) no longer pursue traditional occupations. However, many untouchables (Dalits)5 still perform menial jobs that are traditionally considered “polluting” or defiling. Although the practice of untouchability has been constitutionally banned since the passage of the Untouchability (Offences) Act of 1955, it continues in certain forms not only in private social interactions, but also in the public sector. In private social life, a majority of those who are involved in garbage collection and disposal as well as the cleaning of public places belong to these communities. They are also engaged in cleaning dry latrines.

The Indian Railways, with 8,000 railway stations and 1,72,000 toilets, employs more than 3,00,000 manual scavengers (by some estimates), on a contract basisto clean the railway tracks and the toilets, a practice which has been termed illegal under the Indian law.Notwithstanding the order from theSupreme Court directing Indian Railways to stop this practice, the latter is yet to implement their decision to install modern toilets at stations. A sample survey conducted by the SafaiKarmachari Andolan (2019) (Cleaning Workers Movement) across all the states of India claim that they have found 26,00,000 dry toilets, 7,70,000 sewer cleaners and 36,176 railway cleaners. Although the methodology of the survey might be basic and not rigorous, the results indicate the continued prevalence of the practice of manual scavenging.

Over time, with the access to education and the hope of occupational diversification for some, if not for all untouchable sub-castes, the situation has improved for some, and caste-based occupational mobility has become possible. Reservation has also helped these groups to gain access to education and secure government jobs while facilitating intergenerational mobility. However, the economic reforms in the country, which were introduced in the 1990s, have significantly shrunk the breadth and scope of public sector employment, though not as much of educational institutes.

Many of the low-level public sector jobs (held by Class III and IV employees),6 which saw very high levels of participation under the reservation policy, are increasingly being outsourced to private contractors. In 2006, the Confederation of Indian Industries and the Associated Chambers of Commerce and Industry of India brought out a report on the measures that the industry could take to augment the representation of weaker sections, primarily the SCs and STs, in Indian industries (CII and ASSOCHAM 2019). However, the acceptance of andadherence to the code of conduct for affirmative action isvoluntary for companies.

Studies on Caste

Being the quintessential social and individual identifier inIndia, caste has received a lot of attention not only from sociologists, but also from anthropologists and economists. Most of the sociological literature on caste has revolved around a few dominant themes. The chief amongst these have been: (i) an examination of the origins of the system from the Vedic ages around 2000 BC (Ambedkar 1990; Ketkar 1909; Dumont 1980); (ii) the various changes and manifestations that the system underwent over the subsequent centuries (Srinivas 1952; Béteille 1971); and (iii) the modern-day manifestations of caste. This has included an analysis of the ideas of ritual and physical notions of purity and pollution, and the changing nature of the social practice of caste in pre- and post-independence India. We mention only a few studies from a vast pool of research.

Most of the literature in economics addressing the issue of caste-based exclusion and discrimination examines the effect of this social practice on the lower castes and outcastes inparticular, and the society and its economic implications in general. Neoclassical economics has also attempted to theorise the presence and persistence of caste (Arrow 1971; Akerlof 1976; Becker 2010; Scoville 1996). There are almost no known studies that shift the focus on those who practise untouchability and identify them socially and economically. This is probably because not many surveys have asked respondents about their conduct and perceptions with respect to untouchability.

Data, Motivation and Methodology

This article is based on nationally representative data of 42,152 households, surveyed in 2011–12, which, for the first time, asked direct questions on caste and the practice of untouchability. This data is part of the India Human Development Survey (IHDS-II), a panel survey conducted in 2004–05 and 2011–12, which is a joint undertaking of the National Council of Applied Economic Researchand the University of Maryland. This is the only panel that collects data on household incomes and consumption expenditure amongst data on many socio-economic welfare indicators. We use the data from the second wave, which introduced questions on untouchability.

The mindset prevalent amongst the upper castes is that people belonging to the lower castes are physically and/or ritually unclean, and therefore, they should not be allowed to enter the kitchen (a sacred and clean place) or use the utensils that the household members use for consuming food. The domestic workers employed in homes belonging to the lower castes are usually allowed to mop or swipe the floor or clean the bathrooms, but not allowed to cook food or wash the kitchen utensils. It has also been found that many, if not all the workers performing such tasks, belong to the lower castes. This practice is an example of the notion of “purity and pollution” (Dumont 1980).

The IHDS data also provides information on the strength of a household’s social network, both within and outside its own “community” (read caste). The households were asked to quantify the number of people working as certain professionals, such as doctors, teachers, public servants, and police officials, who were well known to any member of the household and the number of these people belonging to their own community and to other communities. The term “well known” implies that these families are close enough to the respondents’ families for them to visit each other’s homes and share food.

Motivation: Using this unique data set and questions on untouchability, the study hopes to create a profile of households that continue to practise untouchability in India. Further, we specifically seek an answer to the following question, “Can a family having a wider social network outside of its own community be linked to a lower practice of untouchability?” We hypothesise that the wider a particular household’s social network outside its own community, the larger is the likelihood of the household not practising untouchability, as opposed to a household that has a smaller out-community spread and a larger network within its own community.

Methodology: The article uses a logistic regression model that takes as the question of whether the household practises untouchability or not as the dependent variable. The primary variable of interest will be the question of whether a household practises untouchability, and if so, the strength of the effect of this practice within and outside the community networks on the household concerned.

Yt = α + βXt + γZt + δYt + et

Yt = whether a household practices untouchability or not

Xt = number of in-community social networks

Zt = number of outside-community social networks

Yt = set of control (social group, education, urban residence, etc)

Et = error term

The Practice of Reported Untouchability

The household schedule of the 2011–12 IHDS posed the following question to the primary respondent of each of the enumerated survey households: “Do some members practise untouchability in your household?” The respondent’s answer was recorded as a “Yes/No” response. In case the response was a “No,” it was followed by the second question: “Would there be a problem if someone who is a Scheduled Caste were to enter your kitchen or share utensils?,” again seeking a “Yes/No” response. Any household that responded in the affirmative to both or either of the two questions was regarded as practising untouchability.

One must bear in mind that these responses are given by the primary respondent in the household. Untouchability is a sensitive issue, and one must realise that it may be difficult to obtain honest responses on it. The prejudices and biases nurtured by people are highly likely to shape their responses. Figure 1 depicts the share of households across India that gave positive responses to the questions on untouchability, by rural and urban areas. In rural areas, 30% of the households reported practising untouchability, while in urban areas, the corresponding figure was lower at 20%. These findings highlight that either the practice of untouchability is truly more prevalent in rural areas as opposed to urban areas, or that the rural respondents are comfortable about sharing their behavioural information with the interviewers, both of which point towards the existence of a more traditional and conservative mindset of the rural population, as compared to their urban counterparts.


Social Groups

Since untouchability is practised against members of the lower-caste communities, it would be appropriate to analyse the breakup of the households practising it by their social groups or caste identities. Figure 2 illustrates the share of households that, in a face-to-face survey, accepted to practising untouchability, and are presented here by their social belongings (groups). These social groups are classified a little differently than the official categories (that is, SC, ST, OBC, and Others). The general category of “others” is split into the Brahmin and forward castes. The “others” here fall into that category of people who did not mention any social group.

In Figure 2, the social categories include the SCs (Dalits), the STs (Adivasis), the OBCs, the forward castes (Kshatriyas and Vaishyas), the Brahmins (priests) and others, who did not identify themselves as falling into any of the four previously-mentioned categories. Strictly speaking, the Brahmins and members of the forward and other castes and groups are clubbed under “others,” both constitutionally and in the literature.

The IHDS, however, also asked the households to provide a breakup of the constituents of the “others” category, which we report here. We find that the largest share of those practising untouchability from amongst these groups belongs to Brahmins, with 52% of them accepting that they follow this practice. Interestingly, we find that it is the OBCs and not the forward castes (24%) who account for the second largest share of those adhering to this practice. Traditionally, since the forward castes (Kshatriyas and Vaishyas) are next in the hierarchy, below the Brahmins but above the remaining groups that are constitutionally protected, one would expect them to indulge more in the practice than the OBCs, who have more recently been awarded reservation in educational institutes and public sector employment.

This could be indicative of what M N Srinivas (1952)termed as “Sanskritisation,” wherein the low or middle caste groups emulate the rituals and practices of the upper castes in order to be accepted by the latter and to be able to rise up in thesocial hierarchy. This could also be reflective of the increasing competition between the OBCs and the SCs for gaining access to land and other resources, which is being observed in the increasing incidence of violence against the SCs in India in recent times (Hindu 2010).

Interestingly, we also find that 22% of the SCs and 15% of the STs too admit to following this practice. As mentioned earlier, there are numerous jatis within each broad varna category, and amongst the SCs, those who constitute the fifth varna or, more precisely, the avarnas (that is, those falling outside the fourfold classification). There are a number of sub-castes which are also placed hierarchically with respect to each other. These could be the higher sub-castes practising untouchability against the lower sub-castes, particularly those working as cleaners, sweepers and others engaged in similar occupations.

An alternate and more credible explanation could be that the SCs, who have been treated as untouchables for probably more than 10,000 years, have internalised the idea to such an extent that when asked if any member of their family practises untouchability, their natural response is to agree, as they are bound by religious law, which stipulates that they must adhere to certain social norms of behaviour. For example, they are clearly told not to draw water from certain wells or ponds, not to walk on certain streets, and not to enter temples and certain public places, among other such tenets. Thus, they do practise untouchability! One way to test this hypothesis is to examine the responses to the two questions on untouchability separately. Table 1 provides the break-up of households that responded in the affirmative to the first question pertaining to discrimination for all India and by broad social groups. It can be seen from the table that 11% of the SCs agree to practising untouchability.

We can determine the extent to which this is indicative of the respondent admitting to the practice of untouchability as a victim or as a perpetrator, by assessing the responses of the SC households to the second question listed in Table 2. We find that only 5% of the SC households responded in the affirmative to this question. Perhaps, these households belong to the somewhat higher sub-castes within the SCs, who discriminate against their counterparts from the lower sub-castes, possibly in an attempt to emulate upper-caste behaviour.

The Adivasis, on the other hand, have always been outside the Hindu varna system and thus do not fall into any caste (jati) hierarchies. The tribes have always owned land and even functioned as independent kingdoms. They would, therefore, understandably consider themselves as being superior to the lowest of the castes, who have no rights whatsoever. The level of education amongst the tribals is the lowest across any group. The only exception is the group of tribes in the north-eastern regions of India, who have converted to Christianity and received convent education from Christian missions during the British colonial rule in India.

Religious Groups

Across religious groups, we surprisingly find that the Jains account for the largest share of discriminators (Figure 3). However, they comprise an extremely small community and a closed one, whose members marry within their religion, and it is hard for outsiders to join their religious fold (N = 107 in IHDS-II). Although Jainism is a breakaway religion from Hindu religion, it has survived, and Jains have prospered enough to become economically and educationally the most progressive group in India, having seemingly settled into an understanding with the Hindu ideology.

Under the Hindu rubric, Lord Mahavira, the 24th Teerthankar and reviver of Jainism, is seen as one of the incarnations of the Hindu god Vishnu, just as Buddha is also considered by the Hindus as an incarnation. However, the difference between the Jains and Buddhists is that the former have accepted this version and thus found favour with the Hindus, whereas the Buddhists have opposed this and consequently have been marginalised in India. This assimilation of Hindu tenets within the Jain religion is apparent from the fact that the religious days and festivals celebrated by Jains and Hindus often coincide and are quite similar to each other. An additional reason for the high incidence of the practice of untouchability among the Jains could also be that they are strictly vegetarian and would not be comfortable entertaining non-vegetarians in their homes and kitchens. In fact, Jains even avoid using onion and garlic in their food and might not even like vegetarians, who consume these two items, entering their kitchen.

Next to the Jains are the Hindus, with one-third of their population practising untouchability. Sikhism, on the other hand, is a monotheistic religion and a relatively new one,which broke out of the Hindu fold precisely because of the unequal inherent structure of Hinduism. However, Sikhs seem to still carry on the practice of untouchability, with 23% of them responding positively to the question of whether they follow this practice. Many of the ex-untouchables are also known to have converted to Sikhism, but they are known and identified as Mazhabi Sikhs and even have separate gurudwaras (Sikh houses of worship) in Punjab. The state of Punjab, in fact, has the highest share of the SC population across all states in India.

The lowest shares of the practice are seen amongst the tribals, Christians and Buddhists. The Buddhists predominantly comprise of three groups, including the tribals in North East India, the Tibetan refugees in the north, and the neo-Buddhists,7 largely consisting of the ex-untouchables, 6,00,000 of whom publicly converted to Buddhism on 14 October 1956, at Deekshabhoomi, Nagpur, in Maharashtra after heeding to a call from Ambedkar.

Social Networks

We began with the hypothesis that households that know more people well, especially outside of their own community, are less likely to practise untouchability. Figure 4 compares the share of households that practise untouchability by the number of their contacts within and outside the community. We find that the incidence of untouchability falls with a rise in the number of contacts. However, the practice of untouchability is seen to be higher in the case of households that have a large number of contacts outside the community. This finding is contrary to our supposition. One would have expected lesser adherence to social norms in households with more outside contact and relations. In the following section, we will explore these associations in detail and see if this also holds true when sufficient controls are added.


Education seems to have a negative effect on the practice of untouchability. Figure 5 shows the incidence of the practice by the highest level of adult education in the household. The percentage of households practising untouchability is seen to fall with a rise in the level of adult education. We observe a 6 percentage point drop in the incidence of the practice of untouchability between households with no education, and those with an adult who has acquired education up to the graduation level or above.

Significantly, education seems to have much more of a negative impact on the practice of untouchability for the Brahmins and the OBCs, the two communities that report the highest incidence of untouchability. Figure 6 shows a sharp fall in the percentage of households indulging in the practice with a rise in the level of education of adults in the household. Thus, we see a fall of 15% and 16% in the practice, among Brahmins and OBCs, respectively, between the illiterates and those witheducation up to the graduate level or above.

However, we also find that across all social groups, the incidence of untouchability seems to rise among households with graduation or higher level of adult education. This is similar to the findings of studies examining the effect of education on racial attitudes with respect to Whites, Asians, Hispanics and Blacks. G T Wodtke (2012) finds that neither the perspectives on racial enlightenment(for details, see Hyman and Sheatsley 1956; McClelland and Linnander 2006), nor those on ideological refinement(Jackman and Muha 1984; Schuman et al 1997) are able to clearly explain the different group conceptualisations of racial hierarchy and group interests. Groups lying in between the dominant groups (like the Whites) and the lowest subordinate groups (like the Blacks) become acutely aware of their position and the accruing advantages and disadvantages, after acquiring a certain level of education.

This may lead groups higher up the social ladder, say the Hispanics and Asians (the Kshatriyas or Vaishyas in the Indian context, for instance), to ideologically align with the dominant groups, in terms of individualism and meritocracy (say the Brahmins) and mimic their practices. We find a similar pattern in the Indian context. Figure 6 clearly highlights the idea of superiority (that is, the concept of a particular caste being purer than a lower-caste group), as is inherent in caste hierarchies. At the same time, we find that beyond the graduate and higher level of education, the incidence of untouchability rises across all groups, except the SCs.


Income seems to have some effect on reducing the incidence of untouchability. Among the poorest households, 33% admitted to following the practice. On the other hand, among the richest percentile of respondents, only 23% of the total respondents reported the practice. Therefore, there is a 10% drop in the incidence of untouchability from the poorest to the richest percentile (Figure 7).

Regional Variations

Since the practice of untouchability is a part of the religious and traditional mindset of the society at large, it also varies with the strength and level of entrenchment of these mindsets, which too vary across different regions of the country. Figure 8 denotes the level of untouchability practised across broad regions of the country. The practice seems to be the mostprevalent in the central plains of India, where nearly half of all the households interviewed (49%) admitted to the practice. Table 3 indicates the break-up of the states, which fall in each of these regions. The central plains are followed by the northern (40%) and hilly regions (38%) of the country. The lowest incidences of untouchability are seen in the southern, eastern and western regions. Clearly, there seems to be a north-central bias in the pursuit of this practice.


Summing Up

Our preliminary analysis throws some light on the nature of the practice of untouchability in India. The first and foremost differentiation that is evident from the study is the rural–urban divide. It is obvious that the urban space is more homogenising, and since it structurally packs migrants into tighter work and residential spaces, it forces them to relax their customary behaviours and accept modern notions of inclusive sharing of both public and personal spaces. Also, as strangers to a new urban area, people can choose to hide their caste identities or take up new ones. The rural areas, on the other hand, are still carrying forward age-old traditions and customs. Age-old norms are passed on for generations, till they seem to be the natural state of being. However, with the villages getting more connected to the small and large towns, and witnessing a greater degree of out-migration, it seems that rural areas too will eventually experience a social transformation, albeit at much slower rates as compared to their urban counterparts.

Across both rural and urban regions, the practice is seen to be more prevalent amongst the Brahmins, who, more than any other social group, probably feel that it is imperative to adhere to the notions of ritual and physical purity as these perceptions are also profoundly linked to their identity, which imbues them with both a sense of pride and social dominance. The next two social groups to exhibit an adherence to this practice are the OBCs and the forward castes. The practice is also seen to exist amongst the STs and SCs, though to a much lesserextent than among the other social groups.

The preliminary analysis points to an inverse relationship between the size of a household’s network and the incidence of untouchability. This is true for households that have established connections both within as well as outside their community. The incidence of the practice is, however, seen to be lower among households with a larger number of contacts within their respective communities, as opposed to outside their community. The role of education seems to be important in fostering mindset changes and a consequent reduction in the practice of untouchability. A 6% reduction is observed in the practice of untouchability between households with no educated adults and those with an adult with graduate level of education or a diploma. Education is seen to have a stronger negative effect on the pursuit of the practice among Brahmins and OBCs, the two communities reporting the highest incidence of practising untouchability. Household income is also seen to be inversely related to the practice. While 33% of the poorest confess to following the practice, only 23% of the richest accepted to it, signifying a 10% fall in the incidence of untouchability.

Regression Analysis

The fact that a significant segment of the population accepts to the practice of untouchability leads us to ask the question: “What kinds of households are more likely to practice untouchability?” Here, we use a logistic regression specification to measure the strength and direction of the coefficients for various household characteristics in their contribution to this practice (Table 4). We proceed stepwise, first calculating a reduced form model to investigate variations across caste and religion, and between urban and rural areas. Then, we add controls for some of the other household features.

Rural–urban differences: The following results can be witnessed. First, urban residents are 24% less likely to practise untouchability than their rural counterparts. The compulsions of a modern urban life clearly compel households to rethink social customs and to cohabit freely with members of the other castes. In the urban areas, there are hardly any social costs of non-compliance with the notions of untouchability, unlike in the rural regions, where the households flouting social norms may face ostracism and penalties from both—members of their own communities as well as the village council.

Social groups: Akin to the rural–urban differences, the variations across caste groups are also significant and telling. With reference to the SCs, the Brahmins are 6.5 times more likely to practise untouchability, recording the highest incidence of the practice for any group. They are followed by the forward castes and the OBCs, who are, respectively, 2.8 and 2.2 times more likely to indulge in the practice.

Religious groups: Although caste is typically a phenomenon associated with the Hindu religion, it is also visible in more muted and varied forms in other religions, which were introduced into the Indian subcontinent, such as Islam and Christianity, as well as among those that emerged at home, such as Jainism and Sikhism. The coefficients for Muslims and Christians are nearly equal and significant, and these two groups are less likely to practise untouchability by 58% and 56%, respectively, as compared to the Hindus. The odds of the Sikhs practising untouchability are 14% lower than those of the Hindus, but not significant. Some of the lowest and most significant coefficients are seen among the STs and other groups who record 78% and 98% lower odds, respectively, of practising untouchability. However, since the samples for these groups are small, the results are not conclusive, but merely indicative.

Adult education: Higher education is seen to negatively affect the practice of untouchability. As compared to a household with no literate members, a household with an adult having acquired education up to the 10th or 11th standard has 14% lesser odds of practising untouchability. As the level of education rises to the higher secondary/college and graduation/diploma levels, the odds of practising fall by 23% and 24%, respectively. Education is thus clearly one of the critical factors that lead to a change in conservative and orthodox mindsets.

Community networks: It is evident that social networks play an important role in determining the level of interaction that a particular community has with members of another community. This would, in turn, dictate if any member of the household practises untouchability or not. One is likely to find many households wherein the younger members do not indulge in the practice, though the older family members or grandparents may continue to harbour traditional mindsets. One would also expect households with larger social networks outside their communities to practise lower levels of untouchability.

On the other hand, one could say that households thatpractise untouchability are likely to have fewer contacts outside the community and more contacts within the community. Our earlier findings (delineated in Figure 4) suggest that households with more community networks outside the community indulge more in the practice. However, our regression coefficient suggests that households with outside networks are 4% less likely to practise untouchability. On the other hand, households with contacts within the community are 1.6% more likely to practise untouchability, but these results are not significant.

Occupation type: Across occupation types, the coefficient for those involved in farming (cultivation/animal husbandry/managing agricultural property) is seen to be positive and significant. Thus, cultivators are 32% more likely to practise untouchability as compared to salaried professionals.

Income: Figure 4 shows that the practice of untouchability falls by 10 percentage points from the poorest to the richest quintile. The regression results indicate that the coefficient is positive for all quintiles except for the middle quintile, which shows a marginal negative relationship. However, the results are significant for the fourth quintile, indicating a 15% higher likelihood of the pursuit of untouchability.


Our results show that the reported incidences of untouchability are trivial. In fact, the under-reporting of the practice is more likely to be the norm. Overall, 27% of the households admitted to practising untouchability. The incidence of untouchability is higher in rural areas at 30%, as compared to a corresponding figure of 20% in urban areas. A breakup of the households that practise untouchability by their social affiliations reveals that within particular groups, the incidence of the practice is highest amongst the Brahmins, with half of the households in this group reportedly practising untouchability. The next two groups comprising the second and third largest shares of the households practising untouchability are the OBCs and the forward castes. The lowest shares of the incidence are seen among the STs and SCs. Thus, group differences vary across broad caste and ethnic identities.

A surprising result is the involvement of entire SC households in the practice. Across religious groups, the Jains record the highest share of households indulging in this practice, though these results are not conclusive as they are based on a small sample of the respondents. They are followed by Hindu, Sikh and Muslim households, recording progressively falling shares of households practising untouchability. Community networks also account for a significant factor in influencing household behaviour. The larger the size of a household’s network outside its own community, the higher are the chances of the household refraining from the practice. Conversely, the higher the spread of their in-community network, the higher are the chances of households practising untouchability.

Education shows up as having a direct and a negative impact on the inclination of the household members to follow the practice. Households with adults having acquired education up to the 10th standard or higher are less likely to follow the practice, and the strength of this negative association increases with a rise in the level of education. In addition, the negative effect of education seems to be more pronounced among the Brahmins and the OBCs, two groups recording the highest shares of households practising untouchability.

The economic standing of households in terms of their income levels seems to have a marginal effect on their inclination, or as Becker (2010) would call it, the “taste” for following untouchability. Higher income levels, in fact, seem to encourage the practice somewhat, though the effect is more or less similar across all quintiles, except the poorest. Lastly, the regional spread of the practice indicates that the incidences are lower in the southern, eastern and western parts of the country, whereas they are higher in the central, northern and hill regions. This finding is also reiterated by a map (Figure 9), which indicates the level of the practice in the districts wherein the survey was conducted.

Further Research and Conclusions

A first look at the household-level practice of untouchability in India would probably begin with the identification of those practising untouchability by posing the question as to who among them are practising untouchability and who are likely to practice it. Further research would involve a closer exploration of the practices by the sub-castes within each caste in the households. Since the SCs and STs are themselves seen to be practising untouchability, it would be interesting to test the assertion that it is the higher sub-castes among the SCs and certain tribes that consider themselves to be high up in the social hierarchy and above the lowest-ranked categories among the SCs.

The notions of “purity and pollution” are ideas that, despite the spread of education and the advent of modern lifestyles, tend to stick and prey on our religious and social insecurities. Social change is invariably slow and necessitates a change in the political, economic, social and cultural environment. This, in turn, requires the social psychology of the masses to transform over time. The economic, social and cultural evolution of the Indian landscape has seen many of its traditions withering away or facing persistent resistance from modernity and rationality. Notwithstanding the likelihood of under-reporting of the practice of untouchability, 70% of the population reported not indulging in this practice. This is an encouraging sign.


1 For similarities and differences between the two terms, see Béteille (1971).

2 The Rig Veda is one of the four vedas considered to be among the most sacred books of the

3 Ambedkar points to the following locations in the Rig Veda, which propound various different, yet more secular, cosmogonies of the origin of man as opposed to a stratified society. These are mentioned in the Rig Veda at the following locations: the 72nd hymn of the 10th Mandala; and the Rig Veda sections i.96.2; i.80.16; i.114.2; ii.33.13; viii.52.1; ii.36; iv.37.1; and i) vi.14.2.

4 The term “adivasis” refers to the “original inhabitants,” akin to the aboriginal communities.

5 The word Dalit in Hindi means the “oppressed,” and has thus become a politically and socially empowering term used by the ex-untouchables to address themselves.

6 The civil service of the central government is organised into four groups, namely Group “A” (which includes all-India services) Group “B”, Group “C” and Group “D”. Such classification broadly corresponds to the rank, status and the degree of the level of responsibility attached to the posts. Group “A” posts carry higher administrative and executive responsibilities and include senior management positions in the ministries/departments and field organisations. The middle and junior levels of Group “A” alongwith Group “B” constitute middle management. Group “C” posts perform supervisory as well as operative tasks and render clerical assistance in ministries and field organisations. Group “D” posts are meant for carrying out routine duties (DOPT nd).

7 According to the 2011 Census, there are currently 8.44 million Buddhists in India, at least 5.81 million of whom are based in Maharashtra (Office of the Registrar General and the Census Commissioner of India 2011).


Akerlof, George (1976): “The Economics of Caste and of the Rat Race and Other Woeful Tales,” Quarterly Journal of Economics, Vol 90, No 4, pp 599–617.

Ambedkar, B R (1990): Who Were the Shudras? Mumbai: Dr Ambedkar Foundation.

Arrow, Kennith (1971): “The Theory of Discrimination,” Working Papers 403, Department of Economics, Industrial Relations Section, Princeton University.

Becker, Gary (2010): The Economics of Discrimination, Chicago: The University of Chicago Press.

Béteille, Andre (1971): “Race, Caste and Ethnic Identity,” International Social Science Journal, Vol 23, No 4, pp 519–35.

Borooah, V K, Amaresh Dubey and Sriya Iyer (2007): “The Effectiveness of Jobs Reservation: Caste, Religion and Economic Status in India,” Development and Change, Vol 38, No 3, pp 423–45.

CII and ASSOCHAM (2019): “Affirmative Action,” Confederation of Indian Industries and the Associated Chambers of Commerce and Industry of India,

Deshpande, Ashwini and Thomas E Weisskopf (2014): “Does Affirmative Action Reduce Productivity? A Case Study of the Indian Railways,” World Development, Vol 64, pp 169–80.

DoPT (nd): “Classification of Central Government Posts,” Department of Personnel and Training, Government of India,

Dumont, Louis (1980): Homo Hierarchicus: The Caste System and Its Implications, Chicago: University of Chicago Press.

Encyclopedia Britannica (2019): “Jati: Hindu Caste,”

Hindu (2010): “Khairlanji: The Crime and Punishment,” 23 August,

Hyman, H H and P B Sheatsley (1956): “Attitudes on Integration,” Scientific American, Vol 195,
pp 35–39.

Jackman, M R and M J Muha (1984): “Education and Intergroup Attitudes: Moral Enlightenment, Superficial Democratic Commitment, or Ideological Refinement?,” American Sociological Review, Vol 49, No 6, pp 751–69.

Ketkar, S V (1909): History of Caste in India: Evidence to the Laws of Manu, Delhi: Low Price Publications.

McClelland, K and E Linnander (2006): “The Role of Contact and Information in Racial Attitude Change among White College Students,” Sociological Inquiry, Vol 76, No 1, pp 81–115.

NCAER and University of Maryland (2012): “India Human Development Survey-II,” National Council of Applied Economic Research,

Office of the Registrar General and Census Commissioner, India (2011): “Census of India,”

Passin, Herbert (1995): “Untouchability in the Far East,” Monumenta Nipponica, Vol 11, No 3, pp 247–67.

Safai Karmachari Andolan (2019): “Manual Scavengers: Bonded to Caste and Humiliation,”

Schuman, H, C Steeh, L Bobo and M Krysan (1997): Racial Attitudes in America: Trends and Interpretations, Cambridge: Harvard University Press.

Scoville, J G (1996): “Labor Market Underpinnings of a Caste Economy,” American Journal of Economics and Sociology, Vol 55, No 4, pp 385–94.

Singh, S and A Thorat (2014): “Caste Restricted Occupational Mobility: Have Reforms Compelled Employers to Be Caste Blind?” Himalayan Journal of Contemporary Research, Vol 3, No 2, pp 2319–3174.

Srinivas, M N (1952): Religion and Society among the Coorgs of South India, London: Oxford University Press.

Thomas, Sean (2008): “The Last Untouchable in Europe,” Independent, 28 July,

Thorat, Sukhedeo and Katherine S Newman (eds) (2010): Economic Discrimination in India, New Delhi: Oxford University Publications.

Wodtke, G T (2012): “The Impact of Education on Intergroup Attitudes: A Multiracial Analysis,” Social Psychology Quarterly, Vol 75, No 1, pp 80–106.

Updated On : 24th Jan, 2020


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top