Relationship between Caste and Occupation

The ahistorical analysis of Desai (1984) and Shah (1996) fails to acknowledge how some castes trace their roots outside the formal village and urban economies and the tradition of literacy. Centuries of socio-physical segregation and illiteracy compromise their position in today’s economy and society, including preventing them from taking advantage of the emerging so-called “caste neutral” occupations and India’s modern economy. The majority of individuals from hundreds of “lower” castes are concentrating in informal sectors outside the formal economy and in the slums of India’s cities.

Many studies show that despite the breakdown of the jajmani system and the dissociation between caste and traditional occupations, large sections of “lower” and artisan castes are concentrating in unskilled or low paid semi-skilled occupations in the informal sectors (Breman 1990, 1993; Jeffery 2001; Kumar 2008). Analysing all-India data on scheduled castes and scheduled tribes (SCs and STs), Thorat (1993) points out that many of them are agricultural labourers. In comparison, the proportion of agricultural labourers among other castes is substantially smaller. Thus, though rural occupations have now been freed of caste restrictions, there is a marked tendency for certain castes to cluster in particular occupations (Panini 1996).

This tendency is not very different in the urban scene. Available evidence suggests that although urban and industrial occupations and professions have attracted members of diverse castes, here too, certain castes tend to be concentrated in particular occupations. To begin with, the Mandal Commission, in its report published in 1980, estimates that the upper and middle castes form nearly 90% of the Class I services, although according to the commission they constitute not more than about 20% of the total population of the country (GoI 1980). Similarly, Deshpande (2003) has recently calculated the poverty-caste relationship on the basis of the National Sample Survey Organisation (NSSO) consumption data, which reinforce the strong relationship between low-caste status and poverty.

The concentration of various castes in particular professions and occupations is well-documented in recent sociological studies. Navlakha’s (1989) study highlights the concentration of upper and middle castes in modern professions such as engineering, medicine, banking and journalism. Jayaram (1977) shows a continuing high concentration of upper and dominant castes in higher education, including engineering and medical colleges. Holmstrom’s studies (1976, 1985) in Bangalore and Bombay show that the upper and middle castes dominate the supervisory and skilled worker category whereas the unskilled workers are drawn from the “lower” castes. He finds that in the unorganised sector, caste is particularly critical for underprivileged workers because they lack resources other than those offered by their caste linkages.

On the whole, there is no dearth of argument suggesting that the underemployed and poor mostly belong to the lower castes. Social seclusion is said to have led to economic deprivation. And hence the essence of the government policy in an attempt to reduce poverty and rehabilitate the poor-rested on the reservation policy since independence. Social integration was to be achieved through availability of education and employment opportunities to the lower castes.1 However, even after pursuing the reservation policy for more than five decades the percentage of population below and marginally above the poverty line is not negligible. Perhaps this would tend to suggest that poor can belong to higher castes also, and that improvement in economic well-being cannot be attained purely on the basis of caste.

The functioning of the rural labour market is largely castebased, but that is expected to get blurred in the context of an urban job market. In other words, as urbanisation follows and results in commercialisation, it is likely to erode the influence of caste factor in the job market although job seekers may initially access information pertaining to the labour market with the help of caste and kinship bonds (Banerjee 1986; Mitra 2003). In such cases what is then the need to introduce reservation policy on the basis of caste? It is done perhaps to compensate for the discrimination shown in the past, from which the existing social inequality is seen to have emerged. However, to answer some of these questions one may examine the caste and economic profile of those located in the low income households, which in turn can bring out the overlaps or differences, if any, between them. And this would have important policy implications in suggesting whom to mitigate the benefits.

There may lie a point of justification in holding the framework of social and economic dualism because in the rural context, as Srinivas (1969: 268) pointed out, “the ban on contact between castes and the solidarity of a sub-caste, express themselves in the spatial segregation of castes”. Extrapolating the similar situation to the urban context – though one would expect the impact of caste factor to get diluted in the process of urbanisation – spatial segregation of residence, education and skill formation, and sector of employment all may be in relationship, each deducing its root from the manifestation of caste, “the prototype of rigid social inequality” (Beteille 1969: 263).

Migration, Caste and Slum Dwellers in Cities

Landlessness and the lack of assets in villages are believed to be the propelling forces of migration to the urban areas. In other words, slum and pavement dwellers are said to be overwhelmingly poor rural migrants primarily from lower castes or disadvantaged communities who migrate to the city through caste, kinship and village networks in search of better economic opportunities (Singh and D’Souza 1980). The character of urban poverty, in fact, according to Dandekar and Rath (1971) is the “consequence of the continuous migration of the rural poor into the urban areas in search of a livelihood, their failure to find adequate means to support themselves there and the resulting growth of pavement and slum life in the cities.” It is, therefore, worth exploring empirically whether slums are mostly dominated by the lower castes, whether lower castes are mostly absorbed in petty

december 12, 2009 vol xliv no 50

EPW

activities with meagre earnings compared to others and whether the incidence of poverty is higher among the lower castes than in other categories in the slums.

To examine the dominance of lower castes in slums, Mitra (1988) based on NSSO’s slum survey (1976-77) data considered the proportion of SC population in the slums to the total slum population and compared it with the same ratio for the city as a whole. Next, he examined the proportion of SC population in the slums to the total SC population in the city. Though the percentage of SC population in the slums turned out to be higher than that in the city as a whole for all the four metropolitan cities of Bombay (now Mumbai), Calcutta (now Kolkata), Delhi and Madras (now Chennai), the incidence of SC population in the slums did not turn out to be exceedingly high, particularly in the slums of Bombay and Calcutta (10.75% and 15.02%, respectively). Even in the slums of Delhi and Madras nearly 65% and 60% of the slum population were non-SC. As regards the proportion of SC population in the slums to the total SC population in the city it was noted that except in Delhi and Madras these figures were not very high. Around 67% of the SC population in the city of Delhi resided in the slums, and in Madras the corresponding figure was more than 90%. However, in Bombay and Calcutta slums their prominent presence was not visible (23% and 34%, respectively). One may still hold that it was actually difficult for them to find an entry to secure accommodation in slums and, hence, they might have been residing in other unhygienic localities – not even suited to be called as slums – and pavements. But even in such a situation the converse of the statement, i e, most of the slum dwellers are from lower castes is certainly not correct. The disadvantaged classes may have migrated to the cities from the rural areas either recently or decades back through caste, kinship and village networks, but the existence of other castes in slums due to the shortage of housing, i e, the phenomenon of downward social mobility, certainly cannot be overlooked.

These issues of social and economic inequality, however, need to be studied in a more comprehensive manner, which the present study addresses itself to. Whether the probability of joining certain specific occupations increases with the prominence of a particular caste group, whether the probability of lying below the poverty line shoots up in response to the presence of certain caste groups are some of the important questions that need a thorough investigation. Needless to add that such an analysis along the lines of caste suffers from major limitations as the status of a particular caste varies considerably across regions/states and the places of origin and destination. Hence, in forming groups such as SCs, Other Backward Classes (OBCs) and “General” only at the place of destination, a great deal of reality might have been ignored.

This paper uses demographic and economic data (2006-07) from slums in four major cities (Jaipur, Ludhiana, Mathura and Ujjain) in India to examine if caste still plays an important role in determining people’s choices among occupations, levels of education, place of residence, and income, and whether caste continues to reproduce inequality in contemporary India. This dataset assesses if social and economic inequalities are closely intertwined, and if so, it would then imply that caste-based reservations can

Economic & Political Weekly

EPW

play an important role in redressing the century old cumulative inequalities of “lower” castes or social groups.

The organisation of the study is as follows. At the end of the present section we discuss the sampling framework used in collecting information on social, demographic and economic attributes of the population from slum households in four cities in India. Section 1 based on bivariate tables looks into some of these attributes along the lines of caste. It also estimates a well-being index and considers the distribution of households from different social categories across various size classes formed on the basis of this well-being index. Section 2 estimates a binomial logit model relating to the probability of falling into higher (or the bottom two) size classes of the well-being index, highlighting the role of caste in explaining such outcomes. Finally, Section 3 summarises the major findings.

The survey of slum households was carried out in 2006-07 in four cities in India, based on a three-stage stratified random sampling technique. It was sponsored by the United Nations Development Programme (UNDP) and the government of India under the Jawaharlal Nehru National Urban Renewal Mission (JNNURM) programme to alleviate urban poverty. Four cities (with a sample size of 500 households from 30 clusters in each city) were selected keeping in view the variability of the cities in terms of population size and other demographic and economic characteristics.

As a first step of the sampling framework each city was divided into several administrative districts or zones, and the slum clusters – the list of which was obtained from the city municipality2

– were distributed across these zones: Z(i)s, i= 1, 2, …..

Second, 30 clusters were selected from these zones on the basis of random draw, using the proportion of the number of clusters in each zone to the total clusters as weight: 30 * C(i)/ΣC(i), where C(i) is the number of clusters in zone Z(i).

Finally, 500 households from each of the four cities were selected from these clusters on the basis of random draw, using the ratio of the number of households in each cluster to the total number of households in 30 clusters as weight: 500 * N(j)/ΣN(j), where N(j) is the number of households in cluster j = 1, 2, …..30.3

In the questionnaire social capital has been conceptualised in terms of social networks from an empirical standpoint. Different channels through which job market information is accessed provide clues to the social capital that an individual possess. Migration status of the household-head and each of the members of the household has been recorded by entering the date of entry of the individual to the place of destination. Finally, information on past and present income and occupation of the workers has been recorded to delineate the inter-temporal changes, if any.

1 Caste and Attributes: Broad Patterns

Under the UNDP-GOI project the survey was undertaken in four cities of Jaipur, Ludhiana, Mathura and Ujjain. The distribution of sample households across various social groups was carried out on the basis of the caste list provided by the respective states.

Table 1 indicates that the relative size of scheduled tribe population except in Jaipur is almost negligible. Similarly, the relative size of those who could not be placed in the specified categories

Table 1: Distribution of Sample Households across Social Categories (%)

Source: Slum Survey under the UNDP-GOI Project (2006-07). Table 2: Distribution of Workers across Social Groups and Nature of Employment (%)

All Cities Combined Daily Wage Regular Self-Employment Wage/Salaried

Each row adds up to 100%. Source: See Table 1.

and have been designated as “others” is nominal. Our analysis, therefore, in the rest of the sections will largely focus on the categories other than the scheduled tribes and “others”. We may note from Table 1 that the general Hindus/Sikhs and general Muslims together constitute a substantive part of the total sample households (combined sample of all the four cities) though the general impression is that the slums are mostly inhabited by the lower castes. In Jaipur and Ludhiana the figure is more than 40% and in Ujjain it is only around 15%, while in Mathura the relative size of general Hindus is the highest (28.5%) among all four.

The nature of employment gauged in terms of daily wage, regular wage and self-employment is structurally different across social groups (Table 2). For example, a relatively higher percentage of st and sc workers are engaged in regular wage employment compared to other social categories. However, at the same time a higher percentage of SC workers along with Muslim (OBC and general both) workers are engaged in daily wage employment. The percentage of general Hindu workers in self-employment is the highest among all if we ignore the category of “others”.

Inter-city variations are substantive. In Jaipur nearly half of the general Muslim, OBC Muslim and OBC Hindus are engaged in daily wage employment. On the other hand, general Hindus and SCs and STs are relatively better off as a comparatively higher percentage of workers from these social categories are employed in regular wage jobs. However, general and OBC Hindus are proportionately more present in self-employment also.

Ludhiana being an industrial city, the percentage of daily wage employment is lowest in most of the social categories. General Hindus, OBC Hindus and SCs are more prominently engaged in regular wage employment than the rest. However, an equally high percentage of general Hindus are also located as selfemployed workers. Nearly half of the Muslims (both general and OBCs) are also employed in this category of employment.

Nearly half of the SC workers are employed in daily wage jobs in Mathura. However at the same time nearly 40% of the Hindu workers (general and OBCs) and around 50 and more than 40% of the general Muslim and OBC Muslim workers respectively are also engaged in self-employment. Ujjain being a purely religious city with a relatively scarce possibility of getting absorbed in high income jobs, a very high percentage of the workers across all social categories are by and large employed in daily wage jobs. Only the percentage of general Hindus is less than 50 in this category of employment while nearly 33% from this social group are located in a relatively better category of employment, i e, regular wage employment. On the whole, evidence across all the four cities seems to be quite mixed and therefore we need to examine many other variables before concluding that lower social categories match the lower economic categories.

The occupational structure varies considerably across caste groups (Table 3, p 59). For example, around 13 and 34% of the general Hindus are engaged as semi-professionals and sales and trade workers respectively whereas among the general Muslims the comparative figures are only 4.2 and 21.2%, respectively. Certain occupations are caste-centric, e g, a very large chunk of the sc workers (around one-fourth) are engaged in one of the most vulnerable categories of occupation – daily wage labour.

december 12, 2009 vol xliv no 50

EPW

Table 5: Level of Education – All Cities Combined (%)

Education General General OBC OBC SC ST Others

Hindu/Sikh Muslim Hindu/Sikh Muslim All Cities Combined General General OBC OBC

Semi-professional 13.3 4.2 7.9 4.8 7.0 23.3 0.0 Primary 22.0 30.1 31.0 31.2 28.3 19.8 27.3

Sales and trade 34.4 21.1 27.3 28.4 18.1 28.8 66.7 Middle 21.9 16.8 21.4 13.9 21.5 31.9 9.1

Personal services 4.1 2.4 5.6 3.3 15.1 5.5 0.0 Secondary 16.2 7.2 10.9 5.4 10.2 17.9 31.8

Manufacturing and repair 11.7 23.5 16.2 18.5 13.1 9.6 11.1 Higher secondary 6.7 2.9 4.7 1.5 3.0 4.8 13.6

Commercial and security 4.8 1.4 2.7 1.5 2.6 1.4 0.0 Graduate and others 9.6 2.7 5.0 1.3 3.2 5.8 0.0

Transport 8.0 10.4 5.6 12.9 5.3 6.8 11.1 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Tailoring 6.1 16.3 6.6 9.1 3.6 4.1 0.0 Source: See Table 1.

Construction 4.6 12.5 7.8 8.3 7.2 6.8 0.0 Labour 7.9 5.2 15.6 9.3 23.8 12.3 11.1 Table 6: Distribution of Workers across Monthly Income Categories (in Rs)

Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: See Table 1. Gen Muslim 5.6 14.6 51.3 25.1 3.4 100

OBC Muslim 7.3 23.9 47.6 19.9 1.3 100 among all the caste groups and across all the four cities they play a SC 7.0 29.6 45.2 17.1 1.2 100 significant role in accessing an entry to the urban job market. More ST 1.4 21.1 35.2 35.2 7.0 100 than 90% of the workers seem to have used networks to get an Others 12.5 12.5 12.5 62.5 0.0 100 urban employment (Table 4). These networks could be caste-based Total 6.3 24.8 45.2 22.3 1.4 100

which means that each social category in the urban context expands due to its own momentum. While from the point of view of an indi-Whether the educational differences translate themselves into vidual, these caste-based networks function as an effective way of income differences is an important question. The income strucproviding accessibility to livelihood, these networks at the same ture at the top level (Rs 5,000 to 10,000 per month) does not time tend to inhibit the process of social assimilation and lead to seem to be much different between Hindus and Muslims, though segmentation and possibilities of conflicts. In fact, urbanisation in OBCs are worse off compared to the general category among both several developing countries has often been accompanied by social the communities (Table 6). On the other hand, the percentage of disturbances of different forms though ideally urbanisation is ex-SCs in the higher income category (5,000 and above) is slightly pected to diffuse the social differences and give rise to a more ho-smaller than that of the OBCs. Though low levels of education and mogeneous set which in turn can facilitate the joint consumption low level incomes are present among higher social categories of public goods and thus help experience upward mobility for all across both the religions, the incidence of low incomes and low sections of the population (Mitra 2006). One way of neutralising levels of education is slightly higher among the lower social catethese caste-based networks is to introduce other effective channels gories. Besides, a religion-based distinction can also be deciof information-flow relating to jobs and shelter.phered – Muslims seem to be worse off compared to the Hindus.

The educational levels however vary substantially across caste On the whole, educational, occupational and income disadgroups (Table 5). The incidence of illiteracy is higher among the vantages are somewhat more pronounced among the socially disadvantaged classes, though it will be

Social Groups All Cities Combined Jaipur Ludhiana Mathura Ujjain equally erroneous to suggest that those Self-Network Self- Network Self-Network Self-Network Self-Network

Initiative Initiative Initiative Initiative Initiative

General Hindu/Sikh 4.8 95.2 6.2 93.8 3.9 96.1 6.1 93.9 1.7 98.3 off. All this would tend to indicate that

EPW

EPW

number of observations on STs in the sample, any conclusion in this regard has to be drawn carefully.

However, these patterns are aggregative in nature. We may therefore like to pursue further experiments. One important line of enquiry could be relating to the factors which explain the overall well-being. Since well-being is measured in terms of a number of variables we need to be careful in selecting the determinants so that they do not include some of the constituents of the well-being index. The right hand side variables include nature of employment, caste categories, duration of migration categories (the long duration migrants of more than 20 years, and the natives constitute the comparison category), nature of construction of the huts and the number of rooms per household, ownership pattern, access to

Table 8: Binomial Logit Model (Maximum Likelihood Estimates)

sanitation and water and ability to remit. As there are mainly three types of employment – daily wage, regular wage and self-employment – two dummies representing the first two categories have been included. In relation to the social groups scheduled tribes and “others” have been taken as the comparison category.

2 Caste-wise Probability of Experiencing Well-being

As mentioned above, the analysis is carried out in a binomial logit framework: households which registered a well-being index of less than 400 are assigned 0 and those lying above the threshold limit are designated as 1. The estimation is carried out for the samples drawn from the four cities separately and also for the pooled sample. Results presented in Table 8 tend to suggest that households in which heads are engaged in regular wage employment correspond to higher levels of well-being while those in daily wage employment are statistically not different from those in self-employment. In individual cities gender dummy does not appear to be significant though in the pooled sample it is significant at 10% level with a negative sign, implying that women headed households have lower well-being index though in terms of marginal effect the probability of having higher levels of wellbeing drops marginally (-0.08) as one moves from male-headed to female-headed households.

EPW

In relation to the caste categories we may again refer to the marginal effects. Relative to the scheduled tribes and “others”, most of the categories seem to have a lower probability of experiencing higher well-being index. However, the extent of fall in the case of OBC Hindus, OBC Muslims and SCs measured in terms of marginal effect is higher than that for general Hindus or Muslims. However, these results hold only when we pool the data from all the four cities. Otherwise, based on individual cities this pattern is not uniformly evident. For example, only OBC Muslims in Ludhiana and Mathura have a different impact on well-being while the rest of the categories are statistically insignificant. However, in Ludhiana they tend to have a higher well-being while in Mathura the marginal effect corresponding to OBC Muslims is -0.30. In Ujjain none of the caste dummies is significant, implying that the variations in well-being across categories are not substantive. In Jaipur, both general and OBC Muslims tend to have a lower well-being index than the rest. On the whole, based on the individual city data it is not possible to conclude that lower castes such as OBCs and SCs show lower well-being though based on the pooled data such a tendency is evident.

Duration of migration does not play a significant role in enhancing the well-being level of the households as seen from the pooled data. However, households which are able to remit and are residing in greater space (measured in terms of the number of rooms) are able to acquire higher levels of well-being. The only surprising result is in relation to the access to piped water and the concrete roof, which have negative coefficients. The access to piped water does not mean individual access; it could be the accessibility of the cluster as a whole. Similarly, the households with concrete roof do not seem to be better off in terms of the well- being index. On the whole, while some of the physical characteristics of shelter and the well-being levels of households move in the positive direction, some others do not show any such pattern. This would imply that even in better dwelling structures one may locate households with lower well-being index in terms of income, consumption, education, health and certain social and demographic variables. Land tenure can improve the quality of shelter but that does not ensure improvement in well-being. Similarly, accessibility to piped water can be a function of the residents’ association with political parties and such political contacts are of mutual help only when well-being levels are poor (Edleman and Mitra 2006).5

3 Conclusions

This paper, in an attempt to revisit the caste issue in the Indian context, analyses a sample of households collected from a relatively deprived section of the society, namely, slums. In order to keep in view the variability of city structure which might have an impact on the socio-economic lifestyle of the slum households, four cities of different nature were selected for the study. Based on a stratified random sampling technique, a sample of 2,000 households, 500 from each of the four cities, was selected. The relative size of upper castes conceptualised in terms of general Hindus/Sikhs and general Muslims is not negligible, though the popular belief is that slums are inhabited primarily by lower castes.

Even within the slums, which are often taken to represent a homogeneous lot, variations across social groups in terms of certain important indicators like education, occupation and incomes are evident. However, vulnerability conceptualised in terms of several socio-economic and demographic indicators exists among most of the social categories despite variations in the relative size of deprivation (measured in terms of the incidence of households located below a threshold limit). By and large, the religion-based distinction rather seems to be more prominent than a caste-based one though there are instances of a somewhat higher incidence of deprivation among the socially backward groups.

The econometric analysis pursued on the pooled sample from all the four cities brings out the caste aspect distinctly. In a binomial logit framework the probability of experiencing well-being beyond a threshold limit turns out to be lower for the socially backward classes than the others. However, in individual cities such a pattern is not so conspicuous. In fact, even in a city like Ujjain, which lacks dynamism and where caste factor is expected to have remained dominant, the dummies representing social categories are not statistically significant, implying that all the social categories are equally vulnerable. These findings have important policy implications, indicating that policy initiatives for deprived areas need to be based on economic criteria rather than simply on the caste factor. In fact, in the deprived areas, particularly in the urban space where vulnerability has political lineages as well, caste-based schemes hold the possibility of igniting caste-war or communal tensions instead of smoothing the contours of inequalities.

The other policy conclusion which follows from the analysis is that simply provision of basic amenities or land tenure to improve the quality of shelter does not necessarily improve the overall well-being of the households measured in terms of economic, social, demographic, health and education-specific variables. Developmental expenditure and employment programmes are some of the important dimensions of government initiatives to reduce urban poverty.

Notes

1 The Mandal Commission (1990) reinforced these issues by including Other Backward Classes as beneficiaries.

2 The list of registered or recognised slum clusters is available with the local governments. Information on the unrecognised or unregistered slum clusters is not available as a result of which these had to be left out.

3 On an average 17 households were taken from each of the 30 clusters in each of the four cities. The sample is representative of the slum population in each city.

4 It excludes health expenditure. 5 The clusters act as political vote banks and in return they are assured of legal tenure and accessibility to basic amenities.

december 12, 2009 vol xliv no 50

EPW

References

Banerjee, B (1986): Rural to Urban Migration and the Urban Labour Market: A Case Study of Delhi (Bombay: Himalaya Publishing House).

Beteille, A (1965): Caste, Class and Power: Changing Patterns of Social Stratification in a Tanjore Village (Berkeley: California University Press).

Breman J (1990): “Even Dogs Are Better Off: The Ongoing Battle between Capital and Labour in the Cane Fields of Gujarat”, The Journal of Peasant Studies, 17: 546-608.

– (1993): Beyond Patronage and Exploitation (Delhi: Oxford University Press). Dandekar, V M and N Rath (1971): Poverty in India (Pune: Indian School of Political Economy).

Desai, I P (1984): “Should Caste Be the Basis for Recognising Backwardness?”, Economic & Political Weekly, Vol 19, No 28.

Deshpande S (2003): “Caste Inequality and Indian Sociology: Notes on Questions of Disciplinary Location” in M Chaudhuri (ed.), The Changing Contours of the Discipline (New Delhi: Orient Longman).

– (2004): Contemporary India: A Sociological View (New Delhi: Penguin Books).

– (1994): Urbanisation, Slums, Informal Sector Employment and Poverty: An Exploratory Study

(Delhi: B R Pub).

Navlakha, S (1989): Elite and Social Change: A Study of Elite Formation in India (New Delhi: Sage Publications).

Panini, M N (1996): “The Political Economy of Caste” in M N Srinivas (ed.), Caste: Its Twentieth Century Avatar (New Delhi: Penguin Books).

Singh, A M and A D’souza (1980): The Urban Poor: Slum and Pavement Dwellers in the Major Cities of India (Delhi: Manohar Publishers).

Shah A M (1996): “The Judicial and the Sociological View of the Other BackwardClasses” in M N Srinivas (ed), Caste, Its Twentieth Century Avatar (New Delhi: Viking)

– (2007): “Caste in the 21st Century: From System to Elements”, Economic & Political Weekly, 42: 109-16

Srinivas, M N (1969): “The Caste System in India” in A Beteille (ed.), Social Inequality (Harmondsworth: Penguin Books).

– (2003): “An Obituary on Caste as a System”, Economic & Political Weekly, 38.

Sundaram, K (2001): “Employment and Poverty in 1990s, Further Results from NSSO 55th Round Employment-Unemployment Survey, 1999-2000”, Economic & Political Weekly, 11 August.

Thorat, S (1993): Economic Development Policies and Change: Emerging Status of Scheduled Castes After Independence, paper presented at the Seminar on Social Composition of Limited Elite, Centre for the Study of Social Systems, School of Social Sciences (New Delhi: Jawaharlal Nehru University).

Economic Political Weekly

EPW