ARCHIVE
Vol. 4, No. 1
FEBRUARY-JUNE, 2014
Research Articles
Research Notes and Statistics
Book Reviews
Referees
Intergenerational Occupational Mobility in Rural
India:
Evidence from Ten Villages
A. Bheemeshwar Reddy* and Madhura Swaminathan†
*Senior Research Fellow, Economic Analysis Unit, Indian Statistical Institute, Bangalore, bheemeshwar@isibang.ac.in.
†Professor, Economic Analysis Unit, Indian Statistical Institute, Bangalore.
Abstract: Given the relatively limited employment opportunities available within villages, the main vehicle for intergenerational occupational mobility for people in rural India is migration to urban or semi-urban areas. At the same time, since 69 per cent of India’s population still lives in villages, it is important to examine and understand the level of intergenerational occupational mobility within villages themselves. This paper examines intergenerational occupational mobility among rural males in India using data from household surveys in ten villages in different agro-ecological regions of the country. The mobility matrix approach is applied to two father-son pairs: heads of households and their fathers and heads of households and co-resident adult sons. A four-fold occupational classification is used: big farmers, small farmers, skilled workers and persons engaged in business or salaried employment, and lastly, rural manual workers. The main finding of the paper is of low intergenerational occupational mobility in all ten villages, particularly among big farmers and rural manual workers. Intergenerational occupational immobility was higher among manual workers from Scheduled Castes than manual workers from Other Castes. Odds ratios showed that downward mobility from any occupation to that of manual worker was higher for Scheduled Caste men than men of Other Castes. The data strongly support the view that Scheduled Caste men who remain in villages are unable to move out of rural manual employment.
Keywords: Intergenerational mobility, occupation, caste, three-generational mobility, village, India, occupational mobility, mobility matrix, immobility.
Introduction
There are very few studies of occupational mobility in India, mainly because there are very few sources of data on the subject. This paper examines intergenerational occupational mobility among rural males in India, using data collected from 10 villages located in different agro-ecological regions in five States of the country. These data were collected by the Foundation for Agrarian Studies as part of the Project on Agrarian Relations in India (PARI).
There are many studies, covering developed and less-developed countries, that have documented the persistence of economic and social inequalities across generations based on outcome indicators such as income, earnings, occupation, and level of education.1 In the literature on social mobility, occupation is considered a good indicator of social status, incomes, and living standards (see Weeden 2002; Goldthorpe and McKnight 2006; Giddens 2009; Kunst and Roskam 2010; and Lambert and Bihagen 2011). A low degree of intergenerational occupational mobility implies that the advantages and disadvantages inherent in the occupational status of one generation are transmitted to the next generation. A situation of low mobility across generations may be favourable for families that are in fortunate socio-economic circumstances, but in the case of families that are less fortunate, low mobility often entails “social exclusion, material and human capital impoverishment, and restrictions on the opportunities and expectations that would otherwise widen their capability to make choices” (Hancock et al. 2007, p. 43).
There are further reasons why intergenerational occupational mobility in rural India should be of particular interest to social scientists and policy makers. Rural India is marked by extreme forms of social and economic inequality, and in particular by a variety of forms of caste discrimination. The study of occupational mobility can help identify the extent to which the process of economic development and modernisation has broken traditional hierarchies and caste and class barriers to occupational choice.
Given the relatively limited employment opportunities available within villages, the main vehicle for intergenerational occupational mobility in India is migration to urban or semi-urban areas.2 At the same time, since 69 per cent of India’s population still lives in villages, it is important to examine and understand the level of intergenerational occupational mobility within villages themselves.
Studies of Intergenerational Occupational Mobility in India
Because of a lack of panel data and of surveys that capture multi-generational information, the evidence on intergenerational occupational mobility is scanty in India.
Prior to the 1980s, sociologists studied intergenerational mobility in villages in order to examine the role of caste in influencing the choice of occupation of individuals. One argument that emerged was that while modernisation allowed everyone in a village to choose his/her occupation, in practice, traditional hierarchies were reinforced with respect to individual occupations (Sharma 1970). Given the differential capacities of various sections of rural society to gain access to modern occupations, “prestigious secular occupations [were] being virtually monopolised by the ex-privileged castes,” while leaving the oppressed strata of the village society once again at the less-privileged end of the occupational hierarchy (ibid., p.1539).
Turning to studies by economists, Ramachandran (1990) analysed occupational mobility in terms of differentiation among peasants and among other socio-economic classes of rural society. In his study of Gokilapuram village in south-western Tamil Nadu, Ramachandran found that “the dispossession of the peasantry, eviction of tenants, erosion of demand for the services of village artisans and the loss of traditional rural non-agricultural occupations” were among the major reasons for the working people becoming agricultural and other manual workers over time (ibid., p 100).3 Swaminathan (1991), in a study based on panel data collected from households in the same village, found that “agricultural modernisation within the existing structural framework has provided restricted opportunities for occupational change [but] has not mitigated the extreme polarisation in the distribution of land” (Swaminathan 1991, p. 261). Evidence from Palanpur village in western Uttar Pradesh showed that the agricultural labourers experienced very little occupational mobility (Dreze, Lanjouw, and Stern 1992).
Kumar et al. (2002a, 2002b) examined intergenerational occupational mobility in India using National Election Study data from 1971 and 1996. They found that a high level of inequality between classes persisted with respect to opportunities for mobility.4 Surveys undertaken for identifying patterns of voter behaviour in elections may not, however, pay detailed attention to socio-economic variables, and hence the quality of information in these surveys on occupation and land ownership may not be reliable.
More recently, Motiram and Singh (2012) used data from the India Human Development Survey, 2005, jointly conducted by the University of Maryland and the National Council for Applied Economic Research (NCAER), to study intergenerational occupational mobility. This study showed that a substantial proportion of sons of low-skilled and low-paid workers remained in the same occupations as their fathers at the all-India level, for urban and rural areas combined.
The major official source of data on employment in India, that is, surveys conducted by the National Sample Survey Organisation (NSSO), does not include any information on fathers or parents of current heads of households. A restricted sample comprising co-resident fathers and (adult) sons can be constructed from various rounds of the NSSO’s Employment and Unemployment Survey. Following this method, Majumder (2010) used data from the 50th (1993–4) and 61st (2004–5) rounds of the Employment and Unemployment Survey to show that intergenerational mobility was significantly lower among the “excluded classes” (Scheduled Castes, Scheduled Tribes, and Other Backward Classes taken together) than among the “advanced” classes. He found that occupational mobility was lower than mobility with respect to educational outcomes, and argued that that could be a sign of discrimination in the labour market. Based on the NSSO’s Employment and Unemployment Surveys from 1983 to 2004–5, Hnatkovska, Lahiri, and Paul (2013) observed that changes in intergenerational mobility rates were similar among Scheduled Castes and Scheduled Tribes, and Other Castes (non-Scheduled Castes and non-Scheduled Tribes). There are, however, methodological problems in the latter study, including that of conflating three generations and of combining farmers with agricultural workers in a single occupational category (see Reddy 2014).
Reddy (2014) used data from six rounds of the Employment and Unemployment Survey to examine changes in intergenerational occupational mobility over the last three decades, that is, from 1983 to 2009–10, among co-resident father–son pairs in rural India. He classified occupations into four groups: white-collar workers, skilled workers, farmers, and unskilled workers. He found, first, that absolute mobility rates were low but rose over the reference period (Table 1). Secondly, in each round of the Employment and Unemployment Survey, absolute intergenerational occupational mobility rates were lower for Scheduled Caste and Scheduled Tribe males than for Other Caste (non-Scheduled Caste and non-Scheduled Tribe) males. Thirdly, sons of unskilled workers and farmers experienced greater immobility than sons of white-collar workers and skilled workers.
Survey year | All | Scheduled Castes/ Scheduled Tribes | Other Castes |
1983 | 24.1 | 23.6 | 24.2 |
1987–88 | 27.4 | 26.4 | 27.8 |
1993–94 | 26.9 | 26.4 | 27.0 |
1999–2000 | 27.7 | 26.5 | 28.1 |
2004–05 | 33.2 | 32.6 | 33.5 |
2009–10 | 35.2 | 34.0 | 35.7 |
Note: The aggregate mobility rate measures the proportion of individuals in the off-diagonal cells of a mobility table.
Source: Reddy (2014), computed from NSSO’s Employment and Unemployment Surveys, 1983, 1987, 1993–94, 1999–2000, 2004–05, 2009–10.
Data and Methodology
Features of the Data Set
The data used in this paper come from 10 villages surveyed between 2005 and 2010 by the Foundation for Agrarian Studies as part of the Project on Agrarian Relations in India (PARI).5 A census-type survey of households was conducted in each village, and the total number of households covered was over 2,500. The villages differ in agro-ecological features as well as in socio-economic characteristics. Table 2 provides a brief description of the 10 study villages. The size and caste composition of the populations of the villages is shown in Table 3.
Village name | Taluk/Mandal/Tehsil | District | State | Agro-ecological zone | |
1 | Ananthavaram | Kollur | Guntur | Andhra Pradesh | East Coast Plains and Hills |
2 | Bukkacherla | Raptadu | Anantapur | Andhra Pradesh | Southern Plateau and Hills |
3 | Kothapalle | Thimmapur | Karimnagar | Telangana | Southern Plateau and Hills |
4 | Alabujanahalli | Maddur | Mandya | Karnataka | Southern Plateau and Hills |
5 | Siresandra | Kolar | Kolar | Karnataka | Southern Plateau and Hills |
6 | Zhapur | Gulbarga | Gulbarga | Karnataka | Southern Plateau and Hills |
7 | Harevli | Najibabad | Bijnor | Uttar Pradesh | Upper Gangetic Plains |
8 | Mahatwar | Rasra | Ballia | Uttar Pradesh | Middle Gangetic Plains |
9 | 25F Gulabewala | Sri Ganganagar | Sri Ganganagar | Rajasthan | Trans-Gangetic Plains |
10 | Warwat Khanderao | Sangrampur | Buldhana | Maharashtra | Western Plateau and Hills |
Source: Data archive, Foundation for Agrarian Studies.
Village name | District | Survey year | Number of households | Total population | Scheduled Castes as per cent of total population | |
1 | Ananthavaram | Guntur | 2005 | 669 | 2216 | 44.2 |
2 | Bukkacherla | Anantapur | 2005 | 292 | 1130 | 18.8 |
3 | Kothapalle | Karimnagar | 2005 | 373 | 1299 | 29.6 |
4 | Alabujanahalli | Mandya | 2009 | 249 | 1081 | 14.1 |
5 | Siresandra | Kolar | 2009 | 84 | 477 | 25.6 |
6 | Zhapur | Gulbarga | 2009 | 115 | 729 | 39.2 |
7 | Harevli | Bijnor | 2006 | 115 | 674 | 38.0 |
8 | Mahatwar | Ballia | 2006 | 159 | 1114 | 60.0 |
9 | 25F Gulabewala | Sri Ganganagar | 2007 | 204 | 1132 | 44.0 |
10 | Warwat Khanderao | Buldhana | 2007 | 250 | 1142 | 8.7 |
Source: PARI (Project on Agrarian Relations in India) survey data.
There are two distinguishing features of this data set that are relevant to this study. First, the PARI survey data provide detailed information on the occupation and a range of other items of information about the current head of household, as well as on the occupation and extent of land owned by the father of the head of the household. This enables us to examine aspects of occupational differences (if any) between current male heads of households and their fathers. As most women leave their natal villages at marriage, we have only examined data on males in this paper.
The paper uses data on (a) current heads of households and their fathers; (b) heads of households and co-resident adult sons; (c) both groups combined (all men); and (d) a three-generational data set comprising heads of households, their fathers and co-resident sons. For each study village, we have two sets of father-son pairs and one three-generational set. The first set of father–son pairs consisted of all male heads of households and their fathers (P1). The second set of father–son pairs consisted of all heads of households and their co-resident adult sons (P2).6 The two groups represented broadly two different age cohorts. The mean age of sons (in effect, the head of household in the first group) was around 45 years, whereas the mean age of sons in the second group was around 23 years (Table 4). The first set of father–son pairs (P1) covered all households in the village (as it is based on census surveys), while the second group (P2) consisted only of those households with male heads of households and their adult sons living together. As can be seen in Table 4, P1 (heads of households and fathers) comprised 82 to 97 per cent of the surveyed households (the excluded numbers were female-headed households); P2 (heads of households and co-resident sons), however, comprised only around 35 to 50 per cent of the surveyed households.7 The exceptions were the two villages in Andhra Pradesh and one village in Telangana, where the number of heads of households and co-resident sons was much lower than elsewhere.
Village | Number of households | Number of father and HoHpairs (P1) | 3 as per cent of 2 | Average age of HoH | Number of HoH and co-resident son pairs (P2) | 6 as per cent of 2 | Average age of co-resident son |
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
Ananthavaram | 669 | 546 | 81.6 | 48 | 172 | 25.7 | 24 |
Bukkacherla | 292 | 250 | 85.6 | 45 | 75 | 25.7 | 24 |
Kothapalle | 373 | 305 | 81.8 | 45 | 71 | 19.0 | 23 |
Alabujanahalli | 248 | 231 | 93.1 | 47 | 107 | 43.1 | 29 |
Siresandra | 81 | 79 | 97.5 | 47 | 41 | 50.6 | 27 |
Zhapur | 113 | 104 | 92.0 | 47 | 55 | 48.7 | 25 |
Harevli | 115 | 112 | 97.4 | 49 | 50 | 43.5 | 25 |
Mahatwar | 159 | 141 | 88.7 | 48 | 57 | 35.8 | 27 |
25F Gulabewala | 204 | 201 | 98.5 | 46 | 90 | 44.1 | 25 |
Warwat Khanderao | 250 | 230 | 92.0 | 48 | 88 | 35.2 | 25 |
Note: HoH=head of household.
Source: PARI survey data.
Interpreting data on the P2 group is clearly more difficult, as it is a restricted sample with no information on sons who have left the household altogether. The bias arising from focusing on co-resident sons and excluding sons who have left is not known.8 Some sons who left the household may have experienced upward mobility. On the other hand, some sons who left their natal households may, for example, have formed nuclear households within the same village and may have become worse off (in terms of land ownership, for example) than when part of the parent household.
In short, the pattern of mobility observed among co-resident sons and fathers may be biased if co-habitation itself affects mobility, but the direction of bias is unclear. We shall return to this point when we discuss villages where the two groups (P1 and P2) show distinctly different patterns of mobility. Despite the limited coverage of the P2 group, we use it for our analysis as (a) we are including all father–son pairs resident in the village, and (b) we are able to create a unique multigenerational data set comprising heads of households, their fathers, and their co-resident sons.
The second distinguishing feature of our study is that it clearly demarcates two categories of farmers or cultivators. We have classified rural occupations into four broad categories: (1) big/rich farmers; (2) small/poor farmers; (3) skilled workers/salaried persons/persons engaged in business (henceforth skilled workers); and (4) rural manual workers. Other studies of occupational mobility in rural India (particularly studies based on NSS data) do not differentiate between classes of cultivators and rural workers.9 Given the size of the cultivator population in rural India and the fact that cultivators are a heterogeneous group, combining wealthy capitalist farmers and penurious poor peasants in one category is clearly problematic.
A detailed and calibrated classification of cultivators belonging to the current generation is possible, but given that the information we have on the fathers of heads of households is less detailed than on the head of household himself, we have used a two-fold classification, namely rich/big farmers and small/poor farmers. This categorisation is based on the extent of land ownership. The exact cut-off for extent of land owned is based on the average land owned by capitalist farmers/rich peasants in each village; the criterion for division of farmers into big and small is the same.10 The exact cut-off is not uniform across the villages, but varies according to the local agro-economic context.11 Since each village is a separate unit of analysis, this non-uniformity in categorisation does not invalidate the analysis.
In this paper, we have classified all individuals who are engaged either in skilled work or in regular salaried work - such as public sector jobs or relatively remunerative private sector jobs, or are self-employed or in other modern occupations - as skilled workers. We brought this somewhat heterogeneous group into a single category mainly because of the small numbers in each component occupational group.
The category of rural manual workers consists of individuals who are engaged in manual work, agricultural and non-agricultural. Manual workers in rural India today are engaged both in agricultural and non-agricultural tasks, and it is difficult to separate those who are exclusively agricultural workers from those who combine agricultural and non-agricultural work (see, for instance, Ramachandran 1990 and Ramachandran, Rawal, and Swaminathan 2010).12
Lastly, each village had some people engaged in small-scale self-employment and artisanal activity. As the numbers of such persons were small, we classified them on the basis of their secondary occupations in one of the four categories above. In most villages, the secondary occupation of such workers was reported to be agricultural work or manual work.
We have not attempted to order the four occupational categories in a socio-economic hierarchy. The two categories of big farmers and small farmers, based on our class analysis, are ordered. Of the four occupational categories, manual labour, which is viewed as the occupation of last resort in rural India, can be placed at the bottom of the socio-economic hierarchy. Even though the income of small farmers may be lower than that of manual workers in some cases, this categorisation indicates that manual labour is viewed as the least desirable of all occupations in rural areas. In our analysis of upward and downward mobility, therefore, we focus on the upward mobility of manual workers and downward mobility of persons belonging to other occupational categories to the category of manual workers.
Caveat
With the mechanisation of agriculture and better connections between rural and urban areas (and the general development of market relations), there has been a decline in the demand for products and services supplied by village artisans (Ramachandran 1990; Sharda 2005). In all 10 study villages, the number of artisans as a proportion of all workers in each generation was very small. For example, in Harevli village in western Uttar Pradesh, there were only two households engaged in artisanal or traditional service-related occupations. These included the households of a carpenter and a priest, but only in the former were both father and son in the same occupation. Further, in all the villages, the share of male workers in artisanal occupations was much lower among sons in father-son pairs than among fathers. For example, in Ananthavaram village in Guntur district in coastal Andhra Pradesh, in all 759 father-son pairs, 32 fathers were engaged in different artisanal occupations, but only 18 sons were also artisans. More importantly, in all 10 villages, only rarely did men take up artisanal work exclusively. In most cases, artisanal work or work at traditional caste callings was supplemented by agricultural or non-agricultural manual labour, either as a secondary or the main occupation. In short, our village data support the observation of a decline in village artisan and traditional occupations. However, given their small numbers and the fact that most sons in our father-son pairs were engaged in more than one occupation, we have not included artisans as a separate category in our mobility tables.
Methodology
As occupation is a categorical variable, we have used a matrix method to measure the extent of mobility in each village in aggregate, and by social group and age cohort.13 A mobility matrix cross-classifies fathers’ and sons’ occupations in the rows and columns of the matrix. A father’s occupational category is referred to as “origin,” and a son’s occupational category is referred to as “destination.” The diagonal cells represent immobility across successive generations, that is, the percentage of sons who remained in the same occupation as their fathers. The off-diagonal cells represent mobility, either upward or downward.
To obtain an aggregate measure, we have calculated the absolute or gross mobility rate. Following Xie and Killewald (2013), let us denote fij as the observed frequency in the ith row (i = 1,…, N) and in the jth column (j = 1,…, N) of a mobility table with N rows and N columns. The simple measure of aggregate mobility, or the absolute mobility rate, measures the proportion of individuals who fall in the off-diagonal cells of a mobility table.
The absolute mobility rate, M, is defined as:
Where, is the sum of diagonal cells of the mobility table and is the grand total of cells of the mobility table.
Row percentages in a mobility matrix indicate the outflow rates from each origin to different destinations. The outflow rates show how men of particular occupational origins (as defined by their fathers’ occupations) were distributed across occupational destinations.
In order to examine variations by caste, we disaggregated the data by broad social group into two categories: Scheduled Castes and all Other Castes.14 We used odds ratios to compare the upward and downward mobility of Scheduled Caste men with men of Other Castes.15 An odds ratio is a relative measure of mobility. In this paper, we used odds ratios as a measure of the relative chances of downward mobility among sons of Scheduled Caste fathers as compared to sons of Other Caste fathers. To illustrate, the odds ratio is given by the ratio of the odds of sons of non-rural manual workers becoming rural manual workers for Scheduled Caste relative to Other Castes. More formally, for the above example, an odds ratio (Ө) for the 2x2 square matrix with two rows (i, i*) and two columns (j, j*), and f as the observed frequency in respective cells, can be written in the following form:
where, fij= number of Scheduled Caste sons in rural manual work whose fathers were in other occupations.
fij*= number of Scheduled Caste sons in occupations of their fathers whose fathers were in other (non-manual worker) occupations.
fi*j= number of Other Caste sons in rural manual worker occupations whose fathers were in other occupations.
fi*j*= number of Other Caste sons in occupations of their fathers whose fathers were in other (non-manual worker) occupations.
In this paper, we examine absolute intergenerational occupational mobility rates over three generations by using matrix-based partial father-son mobility tables, categorised by grandfathers’ occupation. We have also plotted changes in occupational distribution over three generations.
Overall Patterns of Mobility across the Study Villages
We bring together some generalisations based on the patterns of occupational mobility observed in the 10 villages. Village-wise mobility tables for the 10 villages are shown in Appendix Tables A1.1 to A1.10.
Once again, we request readers to note that results are discussed with respect to three groups: (1) current heads of households and their fathers, (2) heads of households and co-resident adult sons, and, (3) the two groups combined (all men). In the case of group (3), the term “father” may refer to fathers in head-of-household and father pairs (P1) or heads of households in head-of-household and co-resident son pairs (P2). Similarly, “son” refers both to heads of households in P1 and co-resident sons in P2. Most of the tables in this paper show results for all men; results for the two sub-groups (P1and P2) are available on request.
Occupational Structure Across Generations
Before we proceed to analyse patterns of intergenerational occupation mobility in the study villages, we briefly present the occupational distribution of fathers and sons by caste group in Tables 5 and 6. Table 5 can be read as follows. In Bukkacherla village, among fathers belonging to Other Castes, 28 per cent were big farmers, 52 per cent were small farmers, 3 per cent were skilled workers, and 17 per cent were rural manual workers. The occupational distribution for fathers among Scheduled Caste men in the same village was as follows: 62 per cent were small farmers and 38 per cent were rural manual workers. There were no big farmers or skilled workers among Scheduled Caste fathers (in all father-son pairs), although Scheduled Castes comprised 19 per cent of the total village population. Similarly, Table 6 shows that, in Mahatwar village, 64 per cent of Scheduled Caste sons (in all father-son pairs) were rural manual workers. Among Other Caste sons, 20 per cent were rural manual workers, 57 per cent were small farmers, 9 per cent were big farmers, and 14 per cent were skilled workers.
Village | Caste | Big farmers | Small farmer | Skilled workers | Rural manual workers | Total |
Ananthavaram | Other Caste | 8 | 55 | 12 | 25 | 100 |
Scheduled Caste | 0 | 32 | 3 | 65 | 100 | |
Bukkacherla | Other Caste | 28 | 52 | 3 | 17 | 100 |
Scheduled Caste | 0 | 61 | 0 | 38 | 100 | |
Kothapalle | Other Caste | 16 | 39 | 9 | 36 | 100 |
Scheduled Caste | 0 | 30 | 3 | 67 | 100 | |
Alabujanahalli | Other Caste | 12 | 75 | 1 | 11 | 100 |
Scheduled Caste | 4 | 54 | 0 | 42 | 100 | |
Siresandra | Other Caste | 38 | 57 | 0 | 5 | 100 |
Scheduled Caste | 3 | 66 | 6 | 25 | 100 | |
Zhapur | Other Caste | 32 | 46 | 3 | 19 | 100 |
Scheduled Caste | 0 | 46 | 0 | 54 | 100 | |
Harevli | Other Caste | 31 | 38 | 14 | 17 | 100 |
Scheduled Caste | 0 | 28 | 3 | 70 | 100 | |
Mahatwar | Other Caste | 19 | 58 | 8 | 15 | 100 |
Scheduled Caste | 0 | 50 | 3 | 46 | 100 | |
25F Gulabewala | Other Caste | 52 | 12 | 8 | 27 | 100 |
Scheduled Caste | 0 | 5 | 1 | 94 | 100 | |
Warwat Khanderao | Other Caste | 22 | 63 | 4 | 11 | 100 |
Scheduled Caste | 7 | 65 | 0 | 28 | 100 |
Note: “Other Castes” comprise all castes other than Scheduled Castes and Scheduled Tribes.
Source: PARI survey data.
Village | Caste | Big farmers | Small farmer | Skilled workers | Rural manual workers | Total |
Ananthavaram | Other Caste | 3 | 53 | 22 | 22 | 100 |
Scheduled Caste | 0 | 37 | 8 | 55 | 100 | |
Bukkacherla | Other Caste | 19 | 38 | 11 | 31 | 100 |
Scheduled Caste | 0 | 23 | 6 | 71 | 100 | |
Kothapalle | Other Caste | 5 | 39 | 20 | 36 | 100 |
Scheduled Caste | 0 | 21 | 12 | 67 | 100 | |
Alabujanahalli | Other Caste | 10 | 62 | 12 | 17 | 100 |
Scheduled Caste | 4 | 33 | 6 | 56 | 100 | |
Siresandra | Other Caste | 24 | 50 | 13 | 14 | 100 |
Scheduled Caste | 0 | 53 | 6 | 41 | 100 | |
Zhapur | Other Caste | 20 | 28 | 13 | 39 | 100 |
Scheduled Caste | 1 | 21 | 5 | 74 | 100 | |
Harevli | Other Caste | 25 | 42 | 9 | 25 | 100 |
Scheduled Caste | 0 | 12 | 7 | 81 | 100 | |
Mahatwar | Other Caste | 9 | 57 | 14 | 20 | 100 |
Scheduled Caste | 0 | 27 | 9 | 64 | 100 | |
25F Gulabewala | Other Caste | 43 | 18 | 13 | 27 | 100 |
Scheduled Caste | 0 | 1 | 7 | 93 | 100 | |
Warwat Khanderao | Other Caste | 10 | 65 | 10 | 15 | 100 |
Scheduled Caste | 0 | 52 | 7 | 41 | 100 |
Note: “Other Castes” comprise all castes other than Scheduled Castes and Scheduled Tribes.
Source: PARI survey data.
A salient feature of the occupational distribution of populations in the study villages was the disproportionately high percentage of Scheduled Caste men (compared to men of Other Castes), both fathers and sons, engaged in rural manual work.
Absolute Mobility Rates
Across the 10 villages, absolute intergenerational mobility rates for all men ranged from 14.8 per cent to 43.8 per cent (see Table 7, column 3). In eight study villages, the absolute mobility rate was 30 to 40 per cent, indicating that about three-fifths of sons’ occupations remained the same as the occupations of their fathers.
The lowest intergenerational mobility was observed in the village of 25F Gulabewala (in the Gang Canal region of Rajasthan). In this village, there was high immobility across generations in two occupations: rich farmers and manual workers. In 25F Gulabewala, 81 per cent of big farmers’ sons remained in the same occupation, the highest proportion among the 10 villages, and 92 per cent of rural manual workers’ sons remained in the same occupation as their fathers.16 This indicates an almost perfect transmission of advantage and disadvantage from one generation to the next among big/rich farmers and rural manual workers. The high occupational segregation prevalent in the fathers’ generation was perpetuated in the next generation.
By contrast, Bukkacherla, a village in the dry and drought-prone district of Anantapur, Andhra Pradesh, showed the highest intergenerational occupational mobility among the 10 study villages. The higher mobility rates in Bukkacherla were mainly on account of downward mobility: from small farmers to rural manual workers and from big farmers to small farmers.
In most villages, the absolute mobility rate among Scheduled Caste men was higher or similar to the absolute mobility rate among Other Caste men, with the exception of 25F Gulabewala (see Table 7, columns 4 and 5). In Gulabewala, Scheduled Caste males were predominantly manual workers in both generations (94 per cent and 93 per cent of fathers and sons, respectively) and there was very little mobility across generations. This village stands out in terms of the low mobility rate among Dalits (10 per cent). Rawal and Swaminathan (2011) have noted elsewhere that Gulabewala was characterised by an extremely high level of income inequality (with a Gini coefficient of 0.6 for per capita income) in aggregate, as well as high inequality between Dalits and Others. The data on occupational mobility show that the close correlation between caste and occupation (class) is being perpetuated over generations.
In Mahatwar, a Dalit-majority village in eastern Uttar Pradesh, 50 per cent of fathers among Dalit males were small farmers but the corresponding proportion among sons was only 26 per cent (Tables 5 and 6). A distinct shift from small farmers to rural manual workers and less so to skilled workers explains the relatively high mobility rate among Dalit men in Mahatwar.17
Village | District | All men | Other Castes | Scheduled Castes |
1 | 2 | 3 | 4 | 5 |
Ananthavaram | Guntur | 39.0 | 40.0 | 40.6 |
Bukkacherla | Anantapur | 43.8 | 41.4 | 52.3 |
Kothapalle | Karimnagar | 40.1 | 40.7 | 40.9 |
Alabujanahalli | Mandya | 26.2 | 23.0 | 27.0 |
Siresandra | Kolar | 39.9 | 40.0 | 44.0 |
Zhapur | Gulbarga | 35.3 | 37.0 | 29.0 |
Harevli | Bijnor | 32.8 | 37.0 | 30.5 |
Mahatwar | Ballia | 36.0 | 29.0 | 42.0 |
25F Gulabewala | Sri Ganganagar | 14.8 | 18.8 | 10.4 |
Warwat Khanderao | Buldhana | 32.4 | 27 | 34.5 |
Note: “Other Castes” comprise all castes other than Scheduled Castes and Scheduled Tribes.
Source: PARI survey data.
With a few exceptions, the absolute mobility rate was not very different across the two father–son pairs: heads of households and fathers (P1), and heads of households and co-resident sons (P2) (Table 8).18 For example, the absolute mobility rate was 31 in Harevli among P1 group men and 35 among P2 group men.
In Zhapur, the mobility rate was higher for P2 than for P1 (41 and 29 respectively). There was higher downward mobility among sons in the head-of-household and co-resident son group (P2) than in the head-of-household and father group (P1): downward mobility from small farmers to rural manual workers for co-resident sons was 59 per cent, whereas the same for heads of households was 34 per cent. By contrast, in Warwat Khanderao, we found higher mobility among P1 as compared to P2 (40 and 20 respectively). The difference in intergenerational mobility between P1 and P2 is mainly on account of higher downward mobility from big farmers to small farmers in the previous generation: 75 per cent of heads of households among big farmers owned less land than their fathers. This could be due to partition of landed property or loss of land for other reasons.19
Village | District | P1 father–HoH | P2 HoH and co-resident sons |
Ananthavaram | Guntur | 38.5 | 40.4 |
Bukkacherla | Anantapur | 44.7 | 41.6 |
Kothapalle | Karimnagar | 40.2 | 39.7 |
Alabujanahalli | Mandya | 25.5 | 28.6 |
Siresandra | Kolar | 48.1 | 29.7 |
Zhapur | Gulbarga | 29.1 | 41.3 |
Harevli | Bijnor | 31.4 | 35.1 |
Mahatwar | Ballia | 32.8 | 41.7 |
25F Gulabewala | Sri Ganganagar | 16.0 | 13.1 |
Warwat Khanderao | Buldhana | 39.8 | 20.2 |
Source: PAI survey data.
Note: P1 refers to a father-HoH pair and P2 refers to a HoH-co-resident son pair, where HoH = head of household.
We have also divided fathers in all father-son pairs into two age cohorts: fathers below 40 years and fathers above 40 years (see Table 9). The mobility indices are not very different across age cohorts for most of the villages, the exceptions again being Zhapur in Gulbarga district, Karnataka, and Warwat Khanderao in Buldhana district, Maharashtra. To put it differently, for men who have not migrated and remained in the village, there is no general pattern of higher occupational mobility among sons of the younger age cohort of fathers than the older age cohort of fathers.
Mobility among Rural Manual Workers
We now turn to the category of rural manual workers, the single largest occupational category in most of the study villages. First, we examine the mobility rate for rural manual workers by village, household group, and caste group. Across the 10 villages, immobility ranged from 63 per cent to 94 per cent (see Table 10, column 3). In five villages, more than 80 per cent of sons of manual workers continued to be manual workers. In other words, the overwhelming majority of sons of rural manual workers remained in the same occupation as their fathers. Nevertheless, there was some upward mobility for sons of rural manual workers, mainly into the category of skilled workers or small farmers.
We have also reported the immobility rate for the two father–son pairs, P1 and P2 (Table 10, columns 4 and 5). In almost all the villages, the immobility rate among manual workers was higher among heads of households and their fathers than among heads of households and their co-resident sons, indicating higher mobility among the younger generation.
Further, sons of rural manual workers belonging to Scheduled Castes had fewer chances of moving out of their fathers’ occupational origins than Other Caste men (Table 11, columns 6 and 3). To illustrate, in Harevli village, 84 per cent of sons of rural manual workers were rural manual workers among Other Caste men, whereas the proportion was 90 per cent among Scheduled Caste men. In Siresandra, the only exception, the absolute number of rural manual workers was very small among Other Caste men. Note also that the immobility rate was similar across the two father–son pairs (Table 11).
Village | Age<=40 years | Age>40 years |
Ananthavaram | 37.6 | 40.6 |
Bukkacherla | 44.9 | 42.0 |
Kothapalle | 36.4 | 41.8 |
Alabujanahalli | 29.9 | 22.4 |
Siresandra | 37.6 | 44.0 |
Zhapur | 43.8 | 15.9 |
Harevli | 32.2 | 33.8 |
Mahatwar | 37.1 | 34.8 |
25F Gulabewala | 11.7 | 20.3 |
Warwat Khanderao | 27.8 | 39.2 |
Source: PARI survey data.
Village | District | All | P1 | P2 |
Ananthavaram | Guntur | 63 | 84 | 57 |
Bukkacherla | Anantapur | 75 | 91 | 68 |
Kothapalle | Karimnagar | 76 | 82 | 75 |
Alabujanahalli | Mandya | 93 | 96 | 91 |
Siresandra | Kolar | 71 | 100 | 56 |
Zhapur | Gulbarga | 95 | 97 | 92 |
Harevli | Bijnor | 90 | 90 | 91 |
Mahatwar | Ballia | 84 | 88 | 80 |
25F Gulabewala | Sri Ganganagar | 92 | 90 | 94 |
Warwat Khanderao | Buldhana | 64 | 100 | 59 |
Note: P1 refers to a father-HoH pair and P2 refers to a HoH-co-resident son pair, where HoH = head of household.
Village | District | Other Castes | Scheduled Castes | ||||
All | P1 | P2 | All | P1 | P2 | ||
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
Ananthavaram | Guntur | 54 | 62 | 51 | 63 | 83 | 56 |
Bukkacherla | Anantapur | 72 | 86 | 66 | 76 | 86 | 72 |
Kothapalle | Karimnagar | 74 | 82 | 72 | 75 | 82 | 74 |
Alabujanahalli | Mandya | 92 | 92 | 88 | 95 | 92 | 100 |
Siresandra | Kolar | 100 | 100* | 100* | 50 | 100* | 33* |
Zhapur | Gulbarga | 89 | 100 | 80 | 96 | 96 | 96 |
Harevli | Bijnor | 84 | 88 | 82 | 90 | 89 | 91 |
Mahatwar | Ballia | 80 | 80 | 80 | 85 | 95 | 79 |
25F Gulabewala | Sri Ganganagar | 89 | 89 | 93 | 94 | 92 | 95 |
Warwat Khanderao | Buldhana | 75 | 78 | 70 | 88 | 43* | 88 |
Notes: (i) “Other Castes” comprise all castes
other than Scheduled Castes and Scheduled Tribes.
(ii) *The absolute numbers in these cells are small.
(iii) P1 refers to a father-HoH pair
and P2 refers to a HoH-co-resident
son pair, where HoH = head of household.
Source: PARI survey data.
Relative Mobility Using Odds Ratios
We use odds ratios to examine the shift between two occupational categories. For a 4x4 mobility matrix, we can calculate 36 odds ratios.20 Here we focus only on a few odds ratios in order to compare the relative chances of upward and downward mobility among Scheduled Caste men and Other Caste men. We have taken two cases of downward mobility (Tables 14 and 15) and one case of upward mobility (Table 13).21
In nine out of 10 villages, immobility was higher among rural manual workers than any other occupational category (see Table 12). In the case of Scheduled Castes, this was true in all 10 villages. For example, in Mahatwar village (in western Uttar Pradesh), the immobility rates for big farmers, small farmers, skilled workers, and rural manual workers were 47, 56, 56, and 84, respectively (see diagonal terms of Appendix Table A1.8).
Village | Big farmers | Small farmers | Skilled workers | Rural manual workers |
Ananthavaram | 29.6 | 61.3 | 59.3 | 63.3 |
Bukkacherla | 60.3 | 46.2 | 70.0 | 75.3 |
Kothapalle | 22.0 | 50.4 | 56.0 | 76.3 |
Alabujanahalli | 48.9 | 77.1 | 40.0 | 93.1 |
Siresandra | 51.2 | 61.9 | 100.0* | 71.4 |
Zhapur | 56.3 | 46.5 | 33.3 | 94.5 |
Harevli | 68.6 | 55.9 | 15.8 | 90.4 |
Mahatwar | 47.4 | 56.3 | 56.3 | 83.6 |
25F Gulabewala | 81.3 | 65.5 | 58.8 | 92.3 |
Warwat Khanderao | 34.4 | 77.2 | 71.4 | 63.5 |
Note: *The number of observations in this case was small.
Source: PARI survey data.
Village | District | All men | Other Caste men | Scheduled Caste men | Odds ratio |
1 | 2 | 3 | 4 | 5 | 6 |
Ananthavaram | Guntur | 36.3 | 44.9 | 37.2 | 1.5 |
Bukkacherla | Anantapur | 24.7 | 27.9 | 24.0 | 1.2 |
Kothapalle | Karimnagar | 23.7 | 26.2 | 24.7 | 1.1 |
Alabujanahalli | Mandya | 6.9 | 7.9 | 5.0 | 1.6 |
Siresandra | Kolar | 28.6 | 0.0 | 50.0 | # |
Zhapur | Gulbarga | 4.6 | 11.1 | 4.3 | 1.3 |
Harevli | Bijnor | 9.6 | 15.8 | 9.8 | 1.7 |
Mahatwar | Ballia | 16.4 | 20.0 | 16.4 | 1.4 |
25F Gulabewala | Sri Ganganagar | 7.7 | 10.6 | 6.2 | 2.1 |
Warwat Khanderao | Buldhana | 36.5 | 25.0 | 12.5 | 7.0 |
Notes: (i) Odds ratio is a measure of relative chances of upward mobility from rural manual workers to any other occupation for Other Castes relative to Scheduled Castes.
(ii) # In
Siresandra, as cell frequency for Other Castes is zero, the odds ratio is not
valid.
(iii)
“Other Castes” comprise all castes other than Scheduled Castes and Scheduled
Tribes.
Source: PARI survey data.
Village | District | All men | Other Caste men | Scheduled Caste men | Odds ratio |
1 | 2 | 3 | 4 | 5 | 6 |
Ananthavaram | Guntur | 22.8 | 12.5 | 41.8 | 5.1 |
Bukkacherla | Anantapur | 29.9 | 23.2 | 67.5 | 6.9 |
Kothapalle | Karimnagar | 21.7 | 13.6 | 50.0 | 6.4 |
Alabujanahalli | Mandya | 8.9 | 7.2 | 28.6 | 5.2 |
Siresandra | Kolar | 13.9 | 8.6 | 37.5 | 6.4 |
Zhapur | Gulbarga | 38.1 | 38.1 | 47.5 | 1.5 |
Harevli | Bijnor | 21.3 | 12.6 | 61.1 | 10.9 |
Mahatwar | Ballia | 25.3 | 9.4 | 47.6 | 8.8 |
25F Gulabewala | Sri Ganganagar | 8.4 | 43.8 | 77.8 | 89.6 |
Warwat Khanderao | Buldhana | 11.8 | 10.0 | 23.8 | 2.8 |
Notes: (i) “Other Castes” comprise all castes
other than Scheduled Castes and Scheduled Tribes.
(ii) The odds ratio is a measure of the relative chances of downward mobility from any other occupation to rural manual labour for Scheduled Castes relative to Other Castes.
Source: PARI survey data.
Village | District | All men | Other Caste men | Scheduled Caste men | Odds ratio |
1 | 2 | 3 | 4 | 5 | 6 |
Ananthavaram | Guntur | 25.2 | 14.8 | 42.0 | 3.5 |
Bukkacherla | Anantapur | 40.9 | 33.8 | 67.5 | 3.4 |
Kothapalle | Karimnagar | 28.5 | 18.7 | 47.1 | 4.9 |
Alabujanahalli | Mandya | 10.6 | 8.5 | 30.8 | 4.3 |
Siresandra | Kolar | 21.4 | 14.3 | 42.9 | 3.3 |
Zhapur | Gulbarga | 45.4 | 36.4 | 47.5 | 1.7 |
Harevli | Bijnor | 32.3 | 20.4 | 64.7 | 9.2 |
Mahatwar | Ballia | 31.9 | 13.8 | 49.2 | 7.2 |
25F Gulabewala | Sri Ganganagar | 24.1 | 4.5 | 85.7 | 108.0 |
Warwat Khanderao | Buldhana | 15.3 | 13.8 | 49.2 | 1.8 |
Notes: (i)“Other Castes” comprise all castes other than
Scheduled Castes and Scheduled Tribes.
(ii) The odds ratio is a measure of the relative chances of downward mobility from the small farmer category to rural manual labour for Scheduled Castes relative to Other Castes.
Source: PARI survey data.
Since rural manual work can be viewed as an occupation of last resort, a shift to any other occupation is treated here as upward mobility. Table 13 (column 6) shows that the odds ratio was greater than one in every village, implying that sons of rural manual workers belonging to Other Castes had a higher chance of upward mobility than sons of rural manual workers belonging to Scheduled Castes. Specifically, in Ananthavaram, the odds ratio was 1.5, that is, the chances of upward mobility for sons of rural manual workers were 1.5 times higher among Other Caste men than Scheduled Caste men. Nevertheless, as Table 13 (column 5) shows, in absolute terms, the rate of upward mobility among Dalit men in Ananthavaram was the highest among all the study villages. This was largely on account of upward mobility from rural manual worker to small farmer through the institution of tenancy. Tenancy was widespread in this region, and sons of landless manual workers were able to lease land and cultivate it.22 Even though incomes from such cultivation were meagre, there was a shift in occupation as such households were classified as poor peasants (small farmers). Another case of a relatively high upward mobility rate among Dalits is Kothapalle village in Telangana. In this village, 17 per cent of sons of Dalit rural manual workers became small farmers and another 8 per cent became skilled workers. The location of this village on a highway provided access to non-agricultural skilled jobs in nearby urban areas.
At the same time, the odds ratios for downward mobility from any occupation to that of rural manual worker were very high and above one in all cases (Table 14, column 6). To illustrate, in Harevli village in western Uttar Pradesh, the odds ratio indicates that downward mobility among Scheduled Caste men was 11 times (10.9) higher than among Other Caste men. The odds ratio was 90 in 25F Gulabewala, the village with the lowest rate of absolute intergenerational occupational mobility.
We also specifically examined mobility from the small-farmer category to the rural manual worker category. Again, the odds ratios were greater than one for all the villages (Table 15, column 6). In Mahatwar, a Dalit-majority village in Uttar Pradesh, the relative chances of downward mobility from small farmers to rural manual workers was seven times higher for Scheduled Caste men than for Other Caste men.
Together, these three tables on upward and downward mobility show that in all the study villages, Scheduled Caste men were at a clear disadvantage as compared to men from Other Castes in respect of both upward and downward mobility.
Results: Three-Generational Mobility
We also examined patterns of occupational mobility over three generations for heads of households, their fathers and their co-resident adult sons in each of the 10 villages. By way of illustration, changes in occupational structure over three generations for two villages, Harevli and Alabujanahalli, are plotted in Appendix 2. The mobility tables showing outflow rates for fathers and sons given the occupation of grandfathers for another two villages, Ananthavaram and Bukkacherla, are shown in Appendix 3. Tables and figures for all the villages are available on request.
As expected, across villages, skilled workers as a proportion of all workers increased over time (in terms of the occupational structure of three generations), while the share of big farmers fell. The changing occupational structure among the three generations in Alabujanahalli (southern Karnataka) is shown in Figure A2.3. The figure shows that the proportion of big farmers and small farmers fell steadily as we moved down generations, whereas the proportion of skilled workers and rural manual workers rose. To take another example, in Ananthavaram (coastal Andhra Pradesh), the proportion of skilled workers in all workers rose from 3 per cent among the first generation to 7 per cent among heads of households and 15 per cent among their sons. The proportion of big farmers in total workers (men) declined from 4 to 1 per cent when moving from the first to third generation.
Although the Scheduled Castes were restricted to fewer occupations than Other Castes even among earlier generations, some changes have occurred over time. For example, in Mahatwar, a Dalit-majority village in eastern Uttar Pradesh, the proportion of skilled workers went from zero among the first generation to three among heads of households and 14 among their sons. The occupations of three generations of Dalit males in Harevli (western Uttar Pradesh) are plotted in Figure A2.2. An interesting feature is the change in the small farmer category: the proportion of small farmers in the three generations went up from 25 per cent among the first generation to 38 per cent among heads of households, and down to 13 per cent among their co-resident sons. In Harevli, too, tenancy was prevalent, with rich, upper-caste Tyagi households leasing out paddy land for cultivation to Dalit households (Rawal 2013).
The multigenerational mobility matrices indicate a high degree of immobility over three generations at both ends of the occupational structure, i.e. rural manual workers and big farmers. Immobility was most pronounced in the category of rural manual workers. To take an example, in Bukkacherla village in Anantapur district, Andhra Pradesh, 100 per cent of men whose grandfathers and fathers worked as rural manual workers entered the same occupation. Similarly, 92 per cent of men whose grandfathers and fathers were big farmers became big farmers themselves (Annexure Table A3.2). In other words, if we examine the marginal occupational distribution of sons by fathers’ (head of households’) occupation for each occupation of the grandfather, men with advantaged grandfathers were more likely to have advantaged fathers and men with disadvantaged grandfathers were more likely to have disadvantaged fathers. In Harevli village (western Uttar Pradesh), 85 per cent of men whose grandfathers were big farmers also had fathers who were big farmers. By contrast, 90 per cent of men whose grandfathers were rural manual workers had fathers who were also rural manual workers. 25F Gulabewala village of Rajasthan stands out as the most extreme case of immobility across the villages, where 91 per cent of men whose grandfathers and fathers were big farmers themselves became big farmers, and 92 per cent of men whose grandfathers and fathers were rural manual workers themselves became rural manual workers.
Nevertheless, in some villages, there is a distinct pattern of what has been termed “counter-mobility” in the literature. Hertel and Groh-Samberg (2013) explain the phenomenon of counter-mobility as the presence of a high effect of grandfathers’ occupation on sons’ occupational choice. They argue that “parents who experienced upward mobility may still have a sense of belonging to their lower class origins. It is likely that they do not object or even fear as strongly as immobile parents their children’s return and, in fact, downward mobility to parents’ class origins” (ibid., p. 18). Harevli provides a good illustration of this observation. In this village, we observed a counter-mobility pattern among rural manual workers: 39 per cent of heads of households’ fathers were rural manual workers; the corresponding proportion was 29 per cent for heads of households, and then rose again to 49 per cent for sons (Figure A2.1). The counter-mobility effect was more marked for men from Scheduled Castes than Other Castes (see Figures A2.2 and A2.3).
Conclusions
This paper examined intergenerational occupational mobility among males in 10 villages in different agro-ecological regions of India.23 The data came from detailed village census surveys conducted by the Project on Agrarian Relations in India (PARI) of the Foundation for Agrarian Studies between 2005 and 2010 in the States of Andhra Pradesh, Telangana, Uttar Pradesh, Maharashtra, Rajasthan, and Karnataka.
There are two distinguishing features of this analysis. First, the data permit us to consider two types of father–son pairs: heads of households (HoH) and their fathers, and heads of households and their co-resident adult sons. While the former is based on a census, the latter is a restricted sample.24 Data on co-resident sons allowed us to map occupational change across three generations. As occupation is closely linked to caste in rural India, we have also compared mobility between social groups, using a two-fold categorisation of all men: Scheduled Castes and Other Castes.25
The second distinctive feature of our study is the four-fold occupational classification that we have used. The four occupational categories are big farmers, small farmers, skilled workers/salaried workers/persons engaged in business, and rural manual workers. As data on incomes and assets are not available for the fathers’ generation, it is not possible to fully rank all four occupations. Nevertheless, it is clear that big farmers are better off than small farmers. It is also clear that rural manual work is an occupation of last resort in rural areas.
Our first finding, based on mobility matrices showing fathers’ (origin) and sons’ (destination) occupations is of high immobility. This picture of immobility is observed across all 10 villages located in diverse regions of the country. Immobility was particularly marked among big farmers on the one hand and rural manual workers on the other. For example, in Mahatwar, a Dalit-majority village of eastern Uttar Pradesh, 81 per cent of big farmers’ sons and 92 per cent of rural manual workers’ sons remained in the same occupation as their fathers. While there are few comparable studies, these data show much higher rates of immobility, particularly among rural manual workers, than has been observed in the region (see Asadullah 2006, based on 141 villages in Bangladesh).
Secondly, in every village, aggregate occupational immobility was higher among manual workers from Scheduled Castes than among manual workers from Other Castes. Thirdly, upward mobility out of the category of rural manual work was much lower for men from the Scheduled Castes than for men from Other Castes. At the same time, downward mobility from any other occupation to the category of rural manual work was much higher for Scheduled Caste men than men from Other Castes. To illustrate, in Harevli village of western Uttar Pradesh, downward mobility among Scheduled Caste men was 11 times higher than among men from Other Castes. At the same time, upward mobility of men from Other Castes was twice as high as among men from Scheduled Castes.
These data strongly support the view that Dalit men who remain in their villages are unable to move out of rural manual labour. The few exceptions are villages where skilled work is available in the vicinity (such as Kothapalle in Karimnagar district of Telangana), or where Dalit households can lease in land and become small cultivators (such as in Ananthavaram in Guntur district in Andhra Pradesh).
While the pace of urbanisation in India has risen in the decade of 2001–11, it is still very low in comparison to other developing countries, including China and countries of East and South-East Asia. A large section of India’s population and work force is therefore going to remain rural for the next few decades. Our observations on occupational mobility underline the urgent need for generating opportunities for skilled employment for the mass of rural manual workers, Dalit workers in particular. Such employment generation is critical to improving the well-being of rural populations.
Acknowledgements: The authors are grateful to Aparajita Bakshi, Sripad Motiram, and an anonymous referee of this journal for their comments. An earlier version of this paper was presented at the Tenth Anniversary Conference of the Foundation for Agrarian Studies, “On Agrarian Issues,” Kochi, January 9–12, 2014. The authors are grateful to participants for comments and suggestions.
Notes
1 For reviews of the multi-country literature, see Solon (1999, 2002); Bjorklund and Jantti (2008); Black and Devereux (2011); and Blanden (2013).
2 Some scholars have argued, for instance, that migration to urban areas is the only way that Dalit households can improve their social and economic situation (Kapur et al. 2010).
3 Djurfeldt et al. (2008) studied social mobility over 25 years in six villages in the former Tiruchirapalli district of Tamil Nadu. They attributed the changes in social mobility to local industrialisation and social policy.
4 For other studies based on election surveys, see Nijhawan (1969) and Vaid (2012).
5 For further descriptions and discussion of each study village, see http://www.fas.org.in/pages.asp?menuid=16.
6 The criterion for inclusion in our data set was residence in the village and not place of work. Persons may reside in the village and commute to another rural or urban location for work.
7 In the NSS restricted sample of 2009–10, fathers and co-resident sons comprised around 30 per cent of the total sample (Reddy 2014).
8 The extent of out-migration from these villages remains to be studied, and at present, we are not able to examine the relation between the degree of out-migration and the mobility of those who remained in the village.
9 In fact, as mentioned earlier, a major paper based on NSS data clubbed farmers with agricultural labourers. Even in studies based on village data, “self-employed in agriculture” is taken as a single occupational category and differentiation within this group is ignored (Asadullah 2006).
10 We draw here upon the detailed analysis of classes undertaken for each PARI village See, for example, Ramachandran, Rawal, and Swaminathan (2010), for a detailed discussion of the identification of classes.
11 For example, in a dry village like Bukkacherla in Anantapur, the cut-off for big farmers is 13 acres, whereas in Alabujanahalli in the Kaveri-irrigated region of south Karnataka, the cut-off is 6 acres.
12 The nature and type of work done within the same occupational category may have changed substantially over the course of a generation. Our analysis here does not attempt to capture the changing nature of tasks within the category of rural manual work.
13 In the context of comparing mobility across different groups and over time, some studies have used log-linear models to analyse relative mobility rates (see Hertel and Groh-Samberg 2013 and Chan and Boliver 2013).
14 There were very few Scheduled Tribe observations in the selected villages and therefore we have excluded them from this analysis.
15 When two groups with different occupational distributions are to be compared, a modified mobility measure can be calculated by adjusting for differences in marginal distributions (Long and Ferrie 2013). Such an adjustment is not possible with our data as there are often zero values in certain occupations for particular social groups (for instance, there are no big farmers among Scheduled Caste men).
16 Similar high rates were observed for manual workers in the villages of Zhapur (94) and Alabujanahalli (93) in Karnataka.
17 As shown in a later section, downward mobility from small farmers to rural manual workers was higher among Scheduled Caste men (42 per cent) than among Other Caste men (26 per cent).
18 For one village, Ananthavaram, mobility tables for P1 and P2 are shown separately in Appendix Tables A1.11 and A1.12. Even when aggregate mobility rates are similar, there are differences in the matrices for the two pairs.
19 We can only speculate as our data set does not contain information on reasons for land loss. Sivakumar (1980), using data on two villages in Tamil Nadu, showed that partition of land is only one of the factors that explains downward mobility among peasants.
20 Long and Ferrie (2013) show that for a matrix with r and s columns, we can calculate [r(r–1)/2] [s(s–1)/2] odds ratios.
22 For details of the terms of tenancy, see Ramachandran, Rawal, and Swaminathan (2010, chapter 4).
23 This paper outlines the pattern of intergenerational mobility in 10 villages. The processes and mechanisms through which intergenerational mobility occurs, and its impact on the village economy and society will be explored in future research.
24 Results for the two groups combined are reported here; separate results are available on request.
25 For this analysis, data on persons belonging to religious minorities and Scheduled Tribes were excluded.
References
Asadullah, M. N. (2006), “Intergenerational Economic Mobility in Rural Bangladesh,” paper presented at the Royal Economic Society (RES) Annual Conference, Nottingham, available at http://www.ifw-kiel.de/konfer/2006/preg/asadullah.pdf, viewed on May 10, 2014. | |
Blanden, J. (2013), “Cross-Country Rankings in Intergenerational Mobility: A Comparison of Approaches from Economics and Sociology,” Journal of Economic Surveys, vol. 90, no. 2, pp. 38–73, Jun. | |
Beidelman, T. O. (1959), A Comparative Analysis of the Jajmani System, J. J. Augustin, New York. | |
Björklund, A., and Jäntti, M. (2000), “Intergenerational Mobility of Socio-Economic Status in Comparative Perspective,” Nordic Journal of Political Economy, vol. 26, pp. 3–33. | |
Black, S. E., and Devereux, P. J. (2011), “Recent Developments in Intergenerational Mobility,” Handbook of Labor Economics, vol. 4, part B, Elsevier, Amsterdam, pp. 1487–541, Apr. | |
Chan, T. A., and Boliver, V. (2013), “The Grandparents Effect in Social Mobility: Evidence from British Birth Cohort Studies,” American Sociological Review, vol. 78, no. 6, pp. 62–78, Jun. | |
Dreze, J., Lanjouw, P. F., and Stern, N. H. (1992), “Economic Mobility and Agricultural Labour in Rural India: A Case Study,” Indian Economic Review, vol. 27, Special Issue in Memory of Sukhamoy Chakravarty, pp. 25–54. | |
Djurfeldt, G., Athreya, V., Jayakumar, N., Lindberg, S., Rajagopal, A., and Vidyasagar, R. (2008), “Agrarian Change and Social Mobility in Tamil Nadu,” Economic and Political Weekly, vol. 43, no. 45, pp. 50–61, Nov. | |
Giddens, A. (2009), Sociology, Polity Press, Cambridge, sixth edn. | |
Gist, N. (1954), “Caste Differentials in South India,” American Sociological Review, vol. 19, no. 2, pp. 126–37, Apr. | |
Goldthorpe, J. H., and McKnight, A. (2006), “The Economic Basis of Social Class,” in Morgan, S. L., Grusky, D. B., and Fields, G. S. (eds.), Mobility and Inequality: Frontiers of Research in Sociology and Economics, Stanford University Press, Stanford. | |
Hertel, F. R., and Groh-Samberg, O. (2013), “Class Mobility Across Three Generations in the U.S. and Germany,” Research in Social Stratification and Mobility, Mar., available at http://dx.doi.org/10.1016/j.rssm.2013.09.007, viewed on February 20, 2014. | |
Hancock, K. J., Edwards, B., and Zubrick, S. R. (2013), “Echoes of Disadvantage across the Generations? The Influence of Unemployment and Separation of Grandparents on their Grandchildren,” LSAC Annual Statistical Report 2012, Australian Institute of Family Studies, Melbourne. | |
Hnatkovska, V., Lahiri, A., and Paul, S. B. (2013), “Breaking the Caste Barrier: Intergenerational Mobility in india,” Journal of Human Resources, vol. 48, no. 2, pp. 435-73, May. | |
Kapur, D., Prasad, C. B., Pritchett, L., and Babu, D. S. (2010), “Rethinking Inequality: Dalits in Uttar Pradesh in the Market Reform Era,” Economic and Political Weekly, vol. 45, no. 35, pp. 39–49, Aug. | |
Kumar, S., Heath, A., and Heath, O. (2002a), “Determinants of Social Mobility in India,” Economic and Political Weekly, vol. 37, no. 29, pp. 83–87, Jul. | |
Kumar, S., Heath, A., and Heath, O. (2002b), “Changing Patterns of Social Mobility,” Economic and Political Weekly, vol. 37, no. 40, pp. 91–96, Jul. | |
Kunst, A. E., and Roskam, A. J. (2010), “Using the ESeC to Describe Socio-Economic Inequalities in Health in Europe,” in Rose, D., and Harrison, E. (eds.), Social Class in Europe: An Introduction to the European Socio-Economic Classification, Routledge/ESA Studies in European Societies, London. | |
Lambert, P., and Bihagen, E. (2011), “Stratification Research and Occupation-Based Social Classifications,” Technical Papers, DAMES Node, Apr, available at https://www.dames.org.uk/docs/tech_papers/DAMES_tp2011-1.pdf, viewed on February 20, 2014. | |
Long, J., and Ferrie, J. (2013), “Intergenerational Occupational Mobility in Britain and the U.S. since 1850,” American Economic Review, vol. 103, no. 4, pp. 1109–37, Jan. | |
Majumder, R. (2010), “Intergenerational Mobility in Educational and Occupational Attainment: A Comparative Study of Social Classes in India,” Margin – The Journal of Applied Economic Research, vol. 4, no. 4, pp. 463–94, Nov. | |
Motiram, S., and Singh, A. (2012), “How Close Does the Apple Fall to the Tree? Some Evidence from India on Intergenerational Occupational Mobility,” Economic and Political Weekly, vol. 47, no. 40, pp. 56–65, Oct. | |
Nijhawan, N. K. (1969), “Intergenerational Occupational Mobility,” Economic and Political Weekly, vol. 4, no. 15, pp. 1553–7, Oct. | |
Ramachandran, V. K. (1990), Wage Labour and Unfreedom in Agriculture: An Indian Case Study, Clarendon Press, Oxford. | |
Ramachandran, V. K., Rawal, Vikas, and Swaminathan, Madhura (eds.) (2010), Socio-Economic Surveys of Three Villages of Andhra Pradesh: A Study of Agrarian Relations, Tulika Books, New Delhi. | |
Rawal, Vikas, and Swaminathan, Madhura (2011), “Income Inequality and Caste in Village India,” Review of Agrarian Studies, vol. 1, no. 2, available at http://ras.org.in/income_inequality_and_caste_in_village_india, viewed on June 10, 2014. | |
Rawal, Vikas (2013), “Cost of Cultivation and Farm Business Incomes in India,” Working Paper No. 2012–15, Institute of Economic Research, Hitotsubashi University, Apr. | |
Reddy, A. Bheemeshwar (2014), “Intergenerational Occupational Mobility in India: Evidence from National Sample Surveys,” ms., Economic Analysis Unit, Indian Statistical Institute, Bangalore, Jan. | |
Sharda, B. D. (2004), “Status Allocation in Village India,” Research in Social Stratification and Mobility, vol. 22, no. 2, pp. 19–54. | |
Sharma, K. L. (1970), “Modernisation and Rural Stratification: An Application at the Micro-Level,” Economic and Political Weekly, vol. 5, no. 37, pp. 1537–43, Sep. | |
Sivakumar, S. S. (1980), “Efficient Small Farm, Demographic Differentiation and Iron Chest Accounting System: Some Aspects of Peasantist Economics,” Economic and Political Weekly, vol. 15, no. 13, pp. 1337–43, Mar. | |
Solon, G. (1999), “Intergenerational Mobility in the Labor Market,” in Ashenfelter, O., and Card, D. (eds.), Handbook of Labour Economics, vol 3, part A, Elsevier, Amsterdam, pp. 1761–1800. | |
Solon, G. (2002), “Cross-Country Differences in Intergenerational Earnings Mobility,” Journal of Economic Perspectives, vol. 16, no. 3, pp. 59–66, Nov. | |
Swaminathan, Madhura (1991), “Gainers and Losers: A Note on Land and Occupational Mobility in a South Indian Village, 1977–85,” Development and Change, vol. 2, no. 2, pp. 261–78, Oct. | |
Thorat, S., and Newman, K. (eds.) (2009), Blocked by Caste: Economic Discrimination in Modern India, Oxford University Press, New Delhi. | |
Vaid, D. (2012), “The Caste–Class Association: An Empirical Analysis,” Asian Survey, vol. 52, no. 2, pp. 395–422, Mar–Apr. | |
Weeden, K. A. (2002), “Why Do Some Occupations Pay More than Others? Social Closure and Earnings Inequality in the United States,” American Journal of Sociology, vol. 108, no. 1, pp. 55–101, Jul. | |
Xie, Y., and Killewald, A. (2013), “Intergenerational Occupational Mobility in Great Britain and the United States Since 1850: Comment,”American Economic Review, vol. 103, no. 5, pp. 2003–20, Aug. |
Appendix 1
Occupational Mobility Tables
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 8 | 15 | 4 | 0 | 27 |
(30) | (56) | (15) | (0) | ||
Small farmers | 2 | 187 | 39 | 77 | 305 |
(1) | (61) | (13) | (25) | ||
Skilled workers | 0 | 12 | 35 | 12 | 59 |
(0) | (20) | (59) | (20) | ||
Rural manual workers | 0 | 103 | 32 | 233 | 368 |
(0) | (27.9) | (8.7) | (63) | ||
Total | 10 | 317 | 110 | 322 | 759 |
(1) | (42) | (14) | (42) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 47 | 22 | 5 | 4 | 78 |
(60) | (28) | (6) | (5) | ||
Small farmers | 5 | 86 | 19 | 76 | 186 |
(3) | (46) | (10) | (41) | ||
Skilled workers | 0 | 1 | 7 | 2 | 10 |
(0) | (10) | (70) | (20) | ||
Rural manual workers | 0 | 12 | 6 | 55 | 73 |
(0) | (16) | (8) | (75) | ||
Total | 52 | 121 | 37 | 137 | 347 |
15 | (35) | (11) | (39) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 9 | 24 | 8 | 0 | 41 |
(22) | (58) | (19) | (0) | ||
Small farmers | 3 | 62 | 23 | 35 | 123 |
(2) | (50) | (19) | (28) | ||
Skilled workers | 0 | 5 | 14 | 6 | 25 |
(0) | (20) | (56) | (24) | ||
Rural manual workers | 0 | 25 | 16 | 132 | 173 |
(0) | (14) | (9) | (76) | ||
Total | 12 | 116 | 61 | 173 | 362 |
(3) | (32) | (17) | (48) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 22 | 17 | 6 | 0 | 45 |
(49) | (38) | (13) | (0) | ||
Small farmers | 0 | 219 | 35 | 30 | 284 |
(0) | (77) | (12) | (11) | ||
Skilled workers | 0 | 3 | 2 | 0 | 5 |
(0) | (60) | (40) | (0) | ||
Rural manual workers | 0 | 3 | 1 | 54 | 58 |
(00) | (5) | (2) | (93) | ||
Total | 22 | 242 | 44 | 84 | 392 |
(6) | (62) | (11) | (21) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 22 | 16 | 5 | 0 | 43 |
(51) | (37) | (12) | (0) | ||
Small farmers | 5 | 52 | 9 | 18 | 84 |
(6) | (62) | (11) | (21) | ||
Skilled workers | 0 | 0 | 2 | 0 | 2 |
(0) | (0) | (100) | (00) | ||
Rural manual workers | 0 | 4 | 0 | 10 | 14 |
(0) | (29) | (0) | (71) | ||
Total | 27 | 72 | 16 | 28 | 143 |
(19) | (50) | (11) | (20) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 18 | 2 | 6 | 6 | 32 |
(56) | (6) | (19) | (19) | ||
Small farmers | 1 | 46 | 7 | 45 | 99 |
(1) | (46) | (7) | (45) | ||
Skilled workers | 1 | 1 | 1 | 0 | 3 |
(33) | (33 | (33) | (0) | ||
Rural manual workers | 0 | 1 | 3 | 69 | 73 |
(0) | (1) | (4) | (94) | ||
Total | 20 | 50 | 17 | 120 | 207 |
(9) | (24) | (8) | (58) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 24 | 8 | 3 | 0 | 35 |
(69) | (23) | (9) | (0) | ||
Small farmers | 0 | 38 | 8 | 22 | 68 |
(0) | (56) | (12) | (32) | ||
Skilled workers | 4 | 8 | 3 | 4 | 19 |
(21) | (42) | (16) | (21) | ||
Rural manual workers | 0 | 5 | 2 | 66 | 73 |
(0) | (7) | (3) | (90) | ||
Total | 28 | 59 | 16 | 92 | 195 |
(14) | (30) | (8) | (47) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 9 | 8 | 2 | 0 | 19 |
(47) | (42) | (10) | (0) | ||
Small farmers | 0 | 67 | 14 | 38 | 119 |
(0) | (56) | (12) | (32) | ||
Skilled workers | 0 | 6 | 9 | 1 | 16 |
(0) | (37) | (56) | (6) | ||
Rural manual workers | 0 | 9 | 2 | 56 | 67 |
(0) | (13) | (3) | (84) | ||
Total | 9 | 90 | 27 | 95 | 221 |
(4) | (41) | (12 | (43) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 78 | 12 | 6 | 0 | 96 |
(81) | (12) | (6) | (0) | ||
Small farmers | 1 | 19 | 2 | 7 | 29 |
(3) | (65) | (7) | (24) | ||
Skilled workers | 0 | 2 | 10 | 5 | 17 |
(0) | (12) | (59) | (29) | ||
Rural manual workers | 0 | 0 | 15 | 180 | 195 |
(0) | (0) | (8) | (92) | ||
Total | 79 | 33 | 33 | 192 | 337 |
(23) | (10) | (10) | (57) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 21 | 34 | 5 | 1 | 61 |
(34) | (56) | (8) | (2) | ||
Small farmers | 4 | 176 | 13 | 35 | 228 |
(2) | (77) | (6) | (15) | ||
Skilled workers | 0 | 4 | 10 | 0 | 14 |
(0) | (29) | (71) | (0) | ||
Rural manual workers | 1 | 12 | 6 | 33 | 52 |
(2) | (23) | (11 | (63) | ||
Total | 26 | 226 | 34 | 69 | 355 |
(7) | (64) | (10) | (19) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 6 | 15 | 3 | 0 | 24 |
(25) | (63) | (13) | (0) | ||
Small farmers | 2 | 144 | 26 | 28 | 200 |
(1) | (72) | (13) | (14) | ||
Skilled workers | 0 | 10 | 26 | 5 | 41 |
(0) | (24) | (63) | (12) | ||
Rural manual workers | 0 | 98 | 23 | 160 | 281 |
(0) | (35) | (8) | (57) | ||
Total | 8 | 267 | 78 | 193 | 546 |
(2) | (49) | (14) | (35) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Father’s occupation | Son’s occupation | ||||
Big farmers | Small farmers | Skilled workers | Rural manual workers | Total | |
Big farmers | 2 | 0 | 1 | 0 | 3 |
(67) | (0) | (33) | (0) | ||
Small farmers | 0 | 43 | 13 | 49 | 105 |
(0) | (41) | (12) | (47) | ||
Skilled workers | 0 | 2 | 9 | 7 | 18 |
(0) | (11) | (50) | (39) | ||
Rural manual workers | 0 | 5 | 9 | 73 | 87 |
(0) | (6) | (10) | (84) | ||
Total | 2 | 50 | 32 | 129 | 213 |
(1) | (24) | (15) | (61) |
Note: Figures in parentheses represent cell values as a percentage of the row total.
Source: PARI survey data.
Appendix 2
Occupational Structure by Generation for Selected Villages
Appendix 3
Big farmer (BF) |
Small farmer (SF) |
Skilled worker (SW) |
Rural manual worker (RMW) |
||
F: BF | HOH: BF | 67 | 0 | 33 | 0 |
HOH: SF | 0 | 75 | 25 | 0 | |
HOH:SW | 0 | 0 | 0 | 100* | |
HOH: RMW | 0 | 0 | 0 | 0 | |
F: SF | HOH: BF | 0 | 0 | 0 | 0 |
HOH: SF | 0 | 53 | 14 | 33 | |
HOH:SW | 0 | 17 | 50 | 33 | |
HOH: RMW | 0 | 0 | 33 | 67 | |
F: SW | HOH: BF | 0 | 0 | 0 | 0 |
HOH: SF | 0 | 100* | 0 | 0 | |
HOH:SW | 0 | 0 | 100* | 0 | |
HOH: RMW | 0 | 0 | 0 | 100* | |
F: RMW | HOH: BF | 0 | 0 | 0 | 0 |
HOH: SF | 0 | 26 | 11 | 64 | |
HOH:SW | 0 | 25* | 25* | 50* | |
HOH: RMW | 0 | 2 | 6 | 92 |
Note: *the absolute numbers in these cells are very small. HOH=head of household.
Source: PARI survey data.
Big farmer (BF) |
Small farmer (SF) |
Skilled worker (SW) |
Rural manual worker (RMW) |
||
F: BF | HOH: BF | 94 | 0 | 6 | 0 |
HOH: SF | 0 | 60 | 0 | 40 | |
HOH:SW | 0 | 0 | 0 | 0 | |
HOH: RMW | 0 | 0 | 0 | 100 | |
F: SF | HOH: BF | 80 | 0 | 20 | 0 |
HOH: SF | 0 | 21 | 32 | 47 | |
HOH:SW | 0 | 0 | 0 | 0 | |
HOH: RMW | 0 | 11 | 11 | 78 | |
F: SW | HOH: BF | 0 | 0 | 0 | 0 |
HOH: SF | 0 | 0 | 0 | 0 | |
HOH:SW | 0 | 100 | 0 | 0 | |
HOH: RMW | 0 | 0 | 0 | 0 | |
F: RMW | HOH: BF | 0 | 0 | 0 | 0 |
HOH: SF | 0 | 14 | 14 | 71 | |
HOH:SW | 0 | 0 | 100 | 0 | |
HOH: RMW | 0 | 0 | 0 | 100 |
Note: *the absolute numbers in these cells are very small. HOH=head of household.
Source: PARI survey data.