Identity Equations and Electoral Politics

UTTAR PRADESH—VORTEX OF CHANGE

Identity Equations and Electoral Politics Investigating Political Economy of Land, Employment and Education Prashant K Trivedi, Srinivas Goli, Fahimuddin, Surinder Kumar

The changes in landownership pattern, educational mobility, and occupational diversification among socio-religious groups in Uttar Pradesh provide crucial insights into the contemporary nature of political mobilisation in UP. Based on a survey of over 7,000 households, representing all socio-religious groups in 14 districts of the state, the article assesses these changes and points to the disparities between the various groups and, more importantly, to the intra-group inequalities that exist within each group. To effectively mobilise support, political parties will have to look beyond the numbers and recognise the acute differences existing within castes.

The authors owe a special intellectual debt to Ravi Srivastava for his comments on previous drafts of this article. Tausif Alam’s support in data collection is thankfully acknowledged. This study was supported by ICSSR, New Delhi. Prashant K Trivedi ([email protected]) and Surinder Kumar ([email protected]) are with the Giri Institute of Development Studies, Lucknow. Srinivas Goli ([email protected]) is with the Centre for the Study of Regional Development, Jawaharlal Nehru University, New Delhi. Fahimuddin ([email protected]) was a former professor at the GIDS. Economic & Political Weekly

EPW

decEMBER 31, 2016

vol lI no 53

A

s assembly elections in Uttar Pradesh (UP), India’s most populous and, perhaps, politically the most important state, are approaching, electoral trends put out by psephologists will be watched closely. The state’s complex religious identity- and caste-based equations are being intensely analysed by political analysts, academics, and social commentators. Some of these analyses appear to be based on the belief that it is sheer numbers that matter most in electoral politics. And often, it is assumed that these numbers are self-explanatory not requiring any reference to the context. Besides the phenomena of identity politics, caste equations and social engineering, recent researches on electoral politics also refer to political parties’ stand on economic liberalisation, globalisation, development strategy, etc (Kumar 2014; Verma 2014; Chhibbar and Verma 2014; Palishkar and Suri 2014; Sridharan 2014). Notwithstanding the importance of these studies in presenting a threadbare analysis of the voting behaviour of different segments of the electorate, these works offer little explanation of long-term political processes due to their rather narrow focus. This is not to argue that there is any dearth of social science investigations into the long-term political processes (Tiwari 2007; Verma 2007; Pai 2007). Though such studies provide valuable insights into political dynamics, most of the recent enquiries appear to conceive politics and economics as autonomous of each other. Political economy enquiries mark a departure from these linear investigations with an emphasis on “studying a phenomenon as an integrated whole of social, political and economic aspects with causal relationship working both ways” (Mosco 2009). For instance, Srivastava (2007: 353) examines the implications of changing landownership pattern, occupational diversification, and educational mobility among Dalits in UP during the 1980s and 1990s. Taking a cue from this approach, this article argues that changing landownership pattern and occupational and educational mobility provide a vantage point for understanding the contemporary nature of political mobilisation in UP. Further, this article questions the popular perception of electoral politics as merely being a number game. Also examined is the paradox that UP’s contemporary politics is witnessing— the rise of Other Backward Classes (OBCs) and Dalit parties (Samajwadi Party [SP] and the Bahujan Samaj Party [BSP]) coinciding with the over-representation of upper castes in the legislative assembly. Not much attention has been paid to the 117


fact that these very parties also perpetuate upper caste preponderance in politics. The Study: Data and Method

The categorisation of population into broad social groups such as Scheduled Castes (SCs), Scheduled Tribes (STs), and OBCs is not adequate enough to understand political mobilisation in India and more so in UP. These wider categories are often used in political rhetoric to make an appeal to the masses. They are also deployed in academic discourses, but it is caste that is the operational category for political mobilisation, and hence forms the basis of our study. But increasing internal differentiation both within social groups and castes has further complicated the picture. This necessitates an intersectional analysis of caste and class. This study uses primary data collected by the Giri Institute of Development Studies (GIDS) under the project1 “Social and Educational Status of OBC and Dalit Muslims” to assess the relative socio-economic conditions of various castes in UP. The household survey was conducted from October 2014 to April 2015. It was administered to a state representative sample of 7,195 households located across 14 districts in the four regions of UP. After a careful consideration of issues, a multi-stage stratified systematic random sampling design for the selection of households was adopted. Considering the focus of this study, wherever required, an oversampling of households belonging to three Muslim groups was done to be able to analyse internal differentiation within the community. Bivariate estimates are used as statistical analyses. Data on caste, religion and the social group was collected from three different sets of questions and matched with each other while classifying socio-religious groups (SRGs). The focus of this article is on the major castes (jatis) from all six SRGs which have sufficient statistical representation in the sample—Yadavs, Kurmis, Jats, Lodhs among OBCs; Brahmins and Thakurs from upper caste Hindus; Jatav–Chamars and Pasis among Dalits; Ansaris among OBC Muslims; besides upper caste Muslims and Dalit Muslims. The article prefers the term “upper caste” over “general category” which gives an illusion of a casteless group (Deshpande 2013). However, this by no means extends legitimacy to caste hierarchy. Caste–Landownership Matrix

The household survey conducted by GIDS reveals the close correspondence between caste and landownership (Table 1). Data from over 7,000 households, spread across the four regions of the state, reveals stark inequalities among various castes and SRGs. According to the survey, while the upper caste Hindu group accounts for 15% of the sampled households, they emerge as the biggest owners of land, controlling close to 30% share of the total cultivable area of the state. Within this SRG, the Thakurs’ share in land is 2.17 times their proportion in the number of households. The corresponding ratio for Brahmins and other upper castes is around 1.84 and 1.85, respectively. Among Hindu OBCs, the ratio is highest for Kurmis at 2.44 followed by 2.20 for the Jats, 1.88 for Yadavs, and 1.16 for 118

Lodhs. This ratio stands at 0.84 for OBCs other than four major OBC castes mentioned above and 1.28 for entire Hindu OBC group. Suffice it to say that this data discloses sharp disparities within the Hindu OBC category. Among Hindu Dalits, variation from group ratio of 0.58 is very limited on either side. Jatav–Chamars own slightly more land than other Dalits. OBC Muslims and Dalit Muslims rank even lower. Land to household ratio for both these groups is abysmally low, indicating that their share in land is close to one–fourth of their proportion in the number of households. Muslim upper castes have a near-proportionate share in land. Estimates based on all rural households reveal average landownership of upper caste Hindus at 2.43 acres followed by Hindu OBCs at 1.59 acres, Muslim upper castes at 1.38 acres, Hindu Dalits at 0.72 acres; and OBC and Dalit Muslims at 0.34 acres. Thakurs, Kurmis, Jats, Yadavs, and Brahmins own much more land on an average than rest of the castes. Estimates based on only landed households also present a similar hierarchy, but the average landownership of all three Muslim groups and Hindu Dalits increases substantially. Probably, it reflects higher landlessness among these SRGs along with the presence of a small landed group. Intergenerational Educational Mobility

The GIDS household survey offers some interesting insights into the dynamics of intergenerational educational mobility, one of the crucial contributors to social change in independent India. While this is cross-sectional data, it also gives intergenerational mobility. Table 1: Mean Household Cultivable Landholding, Caste-wise Percentage Share of Cultivable Land in Total Cultivable Land in Rural Areas of UP, 2014–15 Castes Mean* (n=6,984)

Brahmins Thakurs Other Hindu UCs Hindu UCs Muslim UCs Yadavs Kurmis Jats Lodhs Other Hindu OBCs Hindu OBCs Ansaris Other Muslim OBCs Muslim OBCs Jatav–Chamars Pasis Other Hindu Dalits Hindu Dalits Dalit Muslims All Hindus All Muslims State average

2.30 2.69 2.31 2.43 1.38 2.34 2.34 2.76 1.33 1.03 1.59 0.30 0.36 0.34 0.75 0.63 0.69 0.72 0.34 1.55 0.60 1.23

Rural Area Mean# (n=3,437) Share of Land*

2.66 2.93 2.65 2.76 2.01 2.46 2.97 2.91 1.72 1.44 1.99 1.09 1.08 1.09 1.45 0.95 1.03 1.28 1.06 2.05 1.48 1.93

13.5 11.1 05.0 29.6 9.3 13.0 4.4 7.7 1.4 17.2 43.7 1.3 3.4 4.7 7.5 0.9 2.2 10.6 2.1 83.9 16.1 100.0

Share of Population*

7.3 5.1 2.7 15.0 8.3 6.9 1.8 3.5 1.2 20.5 33.9 5.4 11.8 17.1 12.3 1.8 3.9 18.0 7.6 66.9 33.1 100.0

*: estimates based on all households, #: Estimates based on only landholding households. UCs – Upper castes. Source: Authors’ estimation based on GIDS Survey on Social and Educational Status of OBC/Dalit Muslims.

decEMBER 31, 2016

vol lI no 53

EPW

Economic & Political Weekly


Literacy rate is rightly not considered as an important their respective groups, Jatav–Chamars, Ansaris, and Jats indicator in contemporary discourse on education given its appear to be making efforts at reducing the gap between minimalist nature. Secondary education and graduation are them and the dominant castes. more appropriate levels for assessment of educa- Table 2: Educational Mobility: Percentage of Population† Completing Secondary School and tional attainments of a group. Above by Generation among Different Castes of UP, 2014–15 Oldest Older Younger Average With regard to individuals having completed Castes % n % n % n % n secondary schooling (Table 2), wide disparities Brahmins 18.8 101 44.0 1,023 64.3 824 51.6 2,009 between SRGs are revealed. The majority of Thakurs 10.3 58 31.3 588 62.1 549 45.2 1,227 Hindu upper caste people, 16 years and above, Other Hindu UCs 15.3 59 43.6 597 70.8 517 54.7 1,201 15.6 218 40.5 2,208 65.7 2,011 50.7 4,437 have achieved this educational level as opposed Hindu UCs 8.8 125 9.0 1,242 24.3 1,088 15.9 2,474 to only 9% of Dalit Muslims. The educational Muslim UCs 3.6 84 12.9 853 35.8 734 22.8 1688 attainment of upper caste Muslims at 15.9% is Yadavs Kurmis (9.5) 21 17.2 250 23.6 258 20.0 535 way behind their Hindu counterparts which Jats 6.1 49 19.8 439 37.7 390 27.2 902 stands at 50.7%. Figures for Hindu OBCs and Lodhs (4.2) 24 10.4 240 28.6 154 16.9 420 Hindu Dalits were a poor 20.1% and 16.9%, Other Hindu OBCs 3.7 272 9.7 2,845 30.3 2,351 18.4 5,552 respectively. But within this bleak scenario, Hindu OBCs 4.2 450 11.7 4,627 31.6 4,020 20.1 9,097 2.4 82 6.3 971 23.6 823 13.8 1,890 stories of progress are also hidden. For instance, Ansaris 5.9 118 4.7 1,598 21.4 1,339 12.5 3,083 the proportion of educational achievement of Other Muslim OBCs 4.5 200 5.3 2,569 22.7 2,204 13.0 4,973 the oldest generation of Jatav–Chamars was a Muslim OBCs Jatav–Chamars 2.3 128 8.9 1,667 31.1 1,319 18.2 3,141 lowly 2.3%, which increased to 8.9% for the Pasis – 17 2.0 197 11.1 171 6.6 394 next generation before leaping to 31.1% for the Other Hindu Dalits 2.3 44 9.2 590 28.9 429 16.8 1,064 younger generation. Hindu Dalits 2.1 189 8.4 2,454 28.9 1,956 16.9 4,599 The intergenerational progress between the Dalit Muslims 3.1 64 3.2 1201 16.0 1,012 9.0 2,295 6.8 857 17.8 9,289 40.1 7,987 27.1 18,133 older and younger generation of other Hindu All Hindus 5.5 389 5.8 5,012 21.2 4,304 12.6 9,742 Dalits, Yadavs, and Ansaris was also commendable. All Muslims 6.3 1,246 13.5 14,301 32.8 12,291 21.9 27,875 While most of the castes advanced between the Total (): sample size less than 40. older and younger generations, the upper caste Oldest generation (Grandfather), Older generation (father), Younger generation (respondent), Hindus took leaps between all successive genera- † Average percentage of literates in population age 16 years and above across all four generations. The youngest generation is not qualified for the analyses. tions, widening the gap between them and other UC– Upper castes. SRGs. This is particularly pertinent in the case of Source: Authors’ estimation based on GIDS Survey on Social and Educational Status of OBC/Dalit Muslims. Brahmins and other Hindu upper castes. Over- Table 3: Educational Mobility: Percentage of Graduates and Above by Generation among the all, a “generation gap” of sort has been created Population† among Different Castes of UP, 2014–15 Oldest Older Younger Average between the attainment levels of upper caste Castes % n % n % n % n Hindus and others. The level attained by older Brahmins 3.0 101 9.4 1,022 19.4 656 12.7 1,810 generation of all other SRGs and castes except, Thakurs 3.4 58 5.3 587 14.3 455 8.9 1,118 3.5 57 9.1 595 25.6 429 15.4 1,096 Kurmis and Jats, still falls short of that achieved Other Hindu UCs 3.2 216 8.2 2,204 19.3 1,604 12.4 4,024 by the oldest generation of the upper caste Hindu UCs 4.8 125 1.2 1,229 4.3 814 2.6 2,175 Hindus. Further, the younger generation of only Muslim UCs Yadavs 1.2 84 1.9 849 3.2 565 2.3 1,505 Yadavs and Jats could get close to the levels of Kurmis (0.0) 21 2.8 248 4.4 204 3.4 477 the older generation of upper caste Hindus. Jats 0.0 49 2.3 438 10.1 326 5.3 825 All castes appear to have experienced progress Lodhs (0.0) 24 0.8 238 1.8 110 1.1 373 in educational achievements, but disparity Other Hindu OBCs 0.4 272 1.0 2,826 4.0 1,748 2.1 4,885 0.4 450 1.4 4,599 4.5 3,016 2.5 8,065 between groups further widens as one ascends Hindu OBCs 2.4 82 0.6 964 4.8 568 2.2 1,619 the ladder of educational achievement. At the Ansari Muslims 5.1 118 0.6 1,586 1.9 967 1.3 2,683 graduation level and above, it assumes the mag- Other Muslim OBCs Muslim OBCs 4.5 200 0.6 2,550 3.0 1,552 1.6 4,302 nitude of two generations between certain castes Jatav–Chamars 0.0 127 1.1 1656 4.5 948 2.3 2,741 and groups (Table 3). For instance, the educa- Pasis – 17 0.5 197 – 131 0.3 347 tional attainment of the oldest generation of Other Hindu Dalits 0.0 44 1.5 588 3.5 314 2.1 946 Hindu upper caste population is comparable to Hindu Dalits 0.0 188 1.2 2,441 3.9 1,405 2.1 4,034 3.1 64 0.6 1,190 2.5 722 1.4 1,986 that of the younger generation of Muslim OBCs, Dalit Muslims 1.1 854 3.0 9,244 11.3 6,025 5.3 16,123 Hindu Dalits, and Dalit Muslims. Similar com- All Hindus 4.0 389 0.8 4969 4.3 3088 1.8 8,463 parisons can be drawn between Brahmins, All Muslims Total 2.0 1,243 2.2 14,213 6.5 8,957 3.8 24,586 Thakurs, and other Hindu upper castes on the Figures are not showed due to insufficient or Zero sample size; (): sample size less than 25; one side and Yadavs and other Hindu Dalits on †: Average percentage of literates in population age 21 years and above across all four generations. generation (Grandfather), Older generation (father), Younger generation (respondent). the other. Numerous other castes were further Oldest UC– Upper castes. below on this scale. Placed ahead of others in Source: Authors’ estimation based on GIDS Survey on Social and Educational Status of OBC/Dalit Muslims. Economic & Political Weekly

EPW

decEMBER 31, 2016

vol lI no 53

119


The relative educational progress of groups cannot be seen in isolation. Social science investigations have identified its contribution in altering power relations between castes and groups. It seems to have triggered social churning by opening doors for occupational mobility, particularly in the case of marginalised groups. It is one of the most profound forms of modern capital in which the caste capital is converted. It is not just a coincidence that castes who are better performers on the educational front also intervene effectively in politics. Studies have persuasively documented the role of educated middle class as political campaigners and organisers (Pai 2002; Srivastava 2007; Deshpande 2013). Intergenerational Occupational Mobility

One of the most interesting data collected by this survey relates to intergenerational occupational mobility. The respondents were directly asked about the traditional occupation pursued by three generations of their family. A meticulous analysis of data from over 7,000 households, belonging to numerous castes, yielded revealing results. The significance of this analysis lies in its ability to capture “mobility” from cross-sectional data. The data reveals a gradual but steady transformation of the population’s occupational pattern manifested in a movement from agrarian to non-agrarian sources of employment (Table 4). At the aggregate level, a drop of eight percentage point in the proportion of persons engaged as farmers is captured among three successive generations, but higher variations between

SRGs and castes make this exercise more interesting. For instance, this drop is 12.5 percentage points in the case of upper caste Hindus. For Brahmins, it stands at 19 percentage points—the highest for any caste. The proportion of persons engaged as farmers belonging to Muslim upper castes has come down over three generations by 11.7 percentage points, while comparable figures for Hindu OBCs and Hindu Dalits are 8.7 percentage points and 7.3 percentage points, respectively. Two castes, Thakurs and Kurmis, experienced smaller diversification. It is no coincidence that these two were the topmost landowners. However, the real difference does not lie in the quantum of intergenerational mobility out of agriculture but the sources of employment that engage later generations. Moving away from farming, upper caste Hindus increasingly shifted to grade A and grade B jobs.2 In the first generation, 6.6% of the upper caste Hindus surveyed took these jobs with the proportion rising to 12.9% in the second generation, finally reaching 21.3% in the third. The figure for Brahmins is even higher at 26.3%. Analysing the intergenerational occupational mobility of Muslim upper castes revealed their preference for selfemployment, business and trade; skilled non-agricultural labour; and unskilled agricultural labour. More Hindu OBCs opted to be self-employed or be in business and trade, with the proportion rising from 2.9% to 8.6% over three generations; their preference for skilled non-agricultural labour increased from 4.3% to 11.0%; and that for unskilled

Table 4: Occupational Mobility of the Sample Population among the Three Successive Generations of the Family across Castes in UP, 2014–15 Castes

Brahmins Thakurs Other Hindu UCs Hindu UCs Muslim UCs Yadavs Kurmis Jats Lodhs Other Hindu OBCs Hindu OBCs Ansaris Other Muslim OBCs Muslim OBCs Jatav–Chamars Pasis Other Hindu Dalits Dalit Muslims Hindu Dalits Muslim Dalits All Hindus All Muslims Total

Farmers

Labourers in Agricultural and Allied Activities

G1

G2

G3

G1

G2

63.1 64.6 42.5 57.8 40.8 68.8 58.5 78.2 40.7 43.0 51.6 10.4 11.3 11.0 28.1 48.3 30.1 11.7 30.1 11.7 47.3 20.6 37.2

51.5 56.8 34.8 48.3 33.8 62.2 56.8 76.1 34.7 39.3 47.5 7.8 9.9 9.1 24.0 44.8 26.6 10.3 26.2 10.3 42.1 17.1 32.7

44.2 66.4 27.2 45.3 29.1 59.0 56.8 70.6 26.3 34.4 42.9 6.3 7.3 6.9 21.3 33.3 24.1 9.0 22.8 9.0 38.5 14.1 29.3

0.4 1.5 0.3 0.7 1.5 2.9 0.0 0.0 1.7 3.0 2.5 2.8 3.2 3.1 7.3 4.6 5.6 1.6 6.6 1.6 1.9 2.6 2.9

0.6 0.7 1.0 0.8 1.2 5.2 0.0 0.0 0.0 3.0 2.8 3.0 2.7 2.8 5.5 2.3 4.2 1.7 4.9 1.7 2.1 2.3 2.6

Self-employed, Business and Traders

G3

G1

G2

G3

Self-employed, Petty Business and Skilled Non-agricultural Labourers G1 G2 G3

Unskilled Non-agricultural Labourers

G1

G2

Grade A Service Jobs and Grade B Service Jobs

G3

G1

G2

G3

Grade C Service Jobs

G1

G2

G3

Others

G1

G2

1.2 2.7 3.1 7.7 1.2 2.1 7.9 2.5 2.5 6.0 7.3 14.9 26.3 0.0 0.4 0.6 22.8 24.9 2.6 0.7 1.8 6.3 0.0 0.7 3.0 2.6 2.6 5.5 2.6 6.6 12.2 0.7 0.7 0.4 27.3 29.9 1.4 23.7 25.4 31.7 1.0 2.8 7.3 2.8 2.1 4.5 9.4 15.3 21.6 1.7 0.3 0.7 18.5 18.1 1.6 8.0 8.9 13.9 0.9 1.9 6.4 2.6 2.4 5.5 6.6 12.9 21.3 0.7 0.5 0.6 22.8 24.3 2.4 2.4 5.1 12.2 3.4 6.8 18.3 17.8 16.6 28.1 3.0 4.9 5.9 0.3 0.0 0.0 30.8 31.6 9.3 1.0 1.7 3.4 0.2 1.0 8.1 4.2 4.4 9.8 1.0 3.4 8.1 0.0 0.0 0.7 21.9 22.1 0.0 4.2 6.8 10.2 0.8 0.8 7.6 9.3 5.9 12.7 0.0 1.7 9.3 0.0 0.0 0.0 27.1 28.0 1.0 1.0 1.0 4.1 1.0 1.0 3.0 6.1 5.6 8.6 0.5 4.1 9.1 0.0 0.0 0.0 13.2 12.2 3.4 3.4 4.2 6.8 0.0 1.7 9.3 28.0 31.4 41.5 0.0 1.7 7.6 0.0 0.0 2.5 26.3 26.3 3.8 3.6 4.9 10.2 6.7 7.1 13.4 16.8 16.9 26.5 2.3 3.6 6.8 0.4 0.3 1.0 24.4 25.0 4.4 2.9 4.0 8.3 4.3 4.8 11.0 13.7 13.8 21.9 1.6 3.4 7.4 0.2 0.2 0.9 23.2 23.5 2.8 3.2 5.6 10.2 26.6 27.7 41.3 31.8 28.4 30.3 1.7 3.0 5.2 0.2 0.2 0.2 23.2 24.2 3.3 2.7 3.6 7.7 16.2 18.1 31.7 35.3 34.8 40.1 3.3 3.6 4.1 0.5 0.4 0.9 27.4 26.9 3.1 2.9 4.4 8.6 20.1 21.7 35.3 34.0 32.4 36.5 2.7 3.4 4.5 0.4 0.3 0.6 25.8 25.9 5.0 0.6 1.3 3.8 2.1 3.0 10.4 29.5 31.8 47.9 0.9 2.8 6.0 1.1 1.4 2.5 30.4 30.4 5.7 0.0 0.0 1.1 1.1 1.1 2.3 23.0 28.7 51.7 0.0 1.1 3.4 0.0 0.0 0.0 23.0 21.8 4.2 1.0 2.4 4.9 7.3 7.7 16.8 19.9 19.9 35.7 3.5 6.3 10.5 0.7 1.0 1.4 31.8 31.8 3.8 1.6 3.7 6.6 11.5 13.6 23.8 41.0 37.9 46.4 0.9 2.1 4.3 0.9 0.5 1.2 31.0 30.3 4.9 0.7 1.4 3.8 3.3 4.0 11.3 26.7 28.6 45.2 1.4 3.5 6.9 0.9 1.2 2.0 30.2 30.1 3.8 1.6 3.7 6.6 11.5 13.6 23.8 41.0 37.9 46.4 0.9 2.1 4.3 0.9 0.5 1.2 31.0 30.3 3.5 4.1 5.3 9.6 4.5 5.3 11.7 14.8 14.3 21.1 2.9 5.8 10.8 2.4 0.3 0.9 24.0 24.8 2.9 2.7 4.6 9.8 14.7 16.9 29.8 28.8 27.3 33.8 2.8 3.9 5.0 0.4 0.2 0.4 27.4 27.7 3.6 3.1 4.4 8.6 7.0 8.2 16.5 20.6 20.3 28.6 2.6 4.9 8.5 0.5 0.4 1.0 26.1 26.5

n†

G3

6.0 482 3.7 271 5.6 287 5.3 1,040 4.1 591 1.5 407 3.4 118 3.6 197 2.5 118 3.8 1,377 3.2 2,217 3.7 462 4.8 778 4.4 1,240 3.3 800 2.3 87 2.4 286 4.7 575 3.0 1,173 4.7 575 4.0 4,430 4.3 2,406 3.9 6,836

G1—Oldest generation (Grandfather’s generation), G2— Older generation (father’s generation), G3—Younger generation (respondent’s generation). n† is number of respondents who reported the occupation of all three generation: self, father and grandfather. UC—Upper castes. Source: Authors’ estimation based on GIDS Survey on Social and Educational Status of OBC/Dalit Muslims.

120

decEMBER 31, 2016

vol lI no 53

EPW



agricultural labour jumped from 13.7% to 21.9%. However, the pattern is slightly different for the dominant OBC castes. For instance, the data for successive generations of Yadavs shows a greater tendency to move towards labour in agriculture and allied activities and grade A and grade B jobs, indicating increasing internal differentiation—an observation applicable to all the castes with varying degrees. The Kurmis and Jats never worked as agricultural labourers and opted for self-employment, business and trade, and grade A and grade B jobs in later generations. The third generation of Lodhs has 41.5% of people engaged in unskilled non-agricultural labour as compared to 28% two generations back. It appears that Muslim OBCs and Dalit Muslims never had a significant presence in agriculture. The largest share of their population continues to be occupied as unskilled nonagricultural labourers. In fact, this data shows an increase from 34% to 36.5% for OBC Muslims, and 41% to 46.4% for Dalit Muslims in this occupation. Over three generations, both these groups have also shown a preference for selfemployment, petty business, and skilled non-agricultural labour. A few members of these two SRGs have also gone for business and trade, and grade A and grade B jobs, but the overall occupational diversification of these groups looks constrained. Instead, they have expanded their preponderance in “traditional” occupations. Hindu Dalits are, perhaps, the only SRG that appear to be withdrawing from agriculture, both as farmers and agricultural labourers. Most of the other SRGs while decreasing their participation in farming have shown a slight tendency to move towards agricultural labour. This is not to belittle the fact that regardless of this change the proportion of population working as agricultural labourers was the highest among Dalits. But the proportion of all three generations of Hindu Dalits engaged as farmers was lower than comparable generations of other SRGs, except Muslim OBCs and Muslim Dalits. Over three generations, Dalits have shown an 18.5 percentage point increase in a shift to unskilled non-agricultural labour. They have also opted for self-employment, petty business, skilled non-agricultural labour, and grade A and grade B jobs. Also, Hindu Dalits appear to have moved away from agriculture faster than others. The withdrawal of upper castes and Dalits from the agrarian sector for their respective reasons has been a recognised phenomenon. It has been pointed out that upper castes diversify for better employment opportunities, but Dalits migrate not just for economic reasons such as unsustainability of small farms but also to reclaim their dignity and political freedom. For them, mobility out of agriculture means liberation from the dominant landed classes (Jodhka 2014). Though landownership might not be the only “determinant” of occupational diversification, there is no doubt that it is a crucial one. The landed classes have managed to acquire the sought-after grade A and grade B jobs as well as get into business and trade. Their strong position in the agrarian economy has provided them a launching pad to make the leap into a modern economy. This class of people has “encashed its traditional Economic & Political Weekly

EPW

decEMBER 31, 2016

vol lI no 53

caste–capital and converted it into modern forms of capital like property, higher educational credentials and strongholds in lucrative professions” (Deshpande 2013: 32). At the same time, the landless masses have either persisted in the same occupation or have moved into similar occupations, and their vertical mobility has been limited. Within this overall scenario, there are castes and within each caste a section of households who have appreciably dealt with structural barriers. Further, occupational diversification among SRGs and internal differentiation within them appear to have proceeded simultaneously in the state. Affirmative action also seems to have contributed to this process by elevating a small section among each marginalised group (Pai 2002). Some of the castes like Jatav–Chamars among SCs, Ansaris among Muslim OBCs, and the four major castes among Hindu OBCs appear more equipped to respond to the policies of positive discrimination. These upward mobile segments possess the potential to unsettle political equations and make the politics in the state more unpredictable. Attempts were made by these segments to articulate an agenda for each group by portraying the homogeneous identity of their respective communities. The massive political changes that followed in UP unleashed aspirations of each group to acquire a fair share in the state power. These developments marked a new era in state politics that was different from the politics of caste patronage of Congress era (qtd in Pai 2002, 2007; Chandra 2007). The expansion of rural non-farm sector in the 1970s and the 1980s stimulated occupational diversification (Lieten and Srivastava 1999). But the massive rupturing of existing occupational patterns took place only in the last two decades, when the construction industry drew a large chunk of workers from rural areas who moved from farming and agriculture labour to unskilled non-agriculture labour. The pattern of growth in this period retained a large informal sector that engaged the majority of Muslim OBC and Dalit Muslim households. It was generally believed that migration to urban centres would loosen the grip of “primordial” identities over the migrants, but empirical investigations are yet to substantiate these claims (Oommen 2011). On the contrary, perhaps, community bonding and related identity would have survived among these footloose workers who had no other support network to cling on to. Inter-group and Intra-group Inequality in Economic Status

The ownership of cultivable land is a crucial indicator, but it mainly allows us to appraise the situation in rural areas, whereas wealth index is an indicator transcending the rural– urban divide. The household data collected by the GIDS survey enables us to assess the relative economic status of households belonging to different castes and SRGs by estimating wealth quintiles from wealth values (Table 5, p 122). In a nutshell, it represents the relative position of SRGs based on the total value of durable assets owned by households. Besides inter-group inequalities, this analysis also estimates intragroup disparities. This analysis distributes the total 121


population into different Table 5: Lowest and Highest 20% by Economic Status across stratums—top 20%, bot- Population Castes and SRGs in UP, 2014–15 tom 20%, and the middle Population Share in different Wealth Classes 60%. Table 5 shows underCastes Bottom Top Sample representation and over20% 20% (n) 6.8 37.3 2,822 representation of castes Brahmins 6.5 46.7 1,736 and SRGs in the top and bot- Thakurs Other Hindu UCs 2.2 36.9 1,621 tom stratum. Hindu UCs 5.5 39.8 6,179 Around 5% to 7% of the Muslim UCs 21.9 17.0 3,970 population belonging to Yadavs 7.7 31.3 2,643 Thakurs, Jats, Brahmins, and Kurmis 8.9 19.6 813 Yadavs fall in the lowest Jats 5.1 57.0 1,342 stratum, whereas only 2.2% Lodhs 25.0 11.7 716 population of the other Other Hindu OBCs 16.9 18.0 8,511 14.0 24.0 14,025 Hindu upper castes falls in Hindu OBCs Ansaris 21.2 20.8 3,018 this category. Around 57% Other Muslim OBCs 31.0 4.8 5,082 of the Jat population finds Muslim OBCs 27.3 10.7 8,100 a place in the highest straJatav–Chamars 24.9 11.9 4,851 tum followed by Rajputs Pasis 39.3 22.9 667 (46.7%), Brahmins (37.3%) Other Hindu Dalits 26.9 11.2 1,750 and Yadavs (31.3%), reveal- Hindu Dalits 26.7 12.8 7,268 ing the high economic status Dalit Muslims 35.1 9.8 3,838 of these groups. On the other All Hindus 15.2 25.8 27,472 25.5 12.8 15,908 hand, Muslim OBCs (other All Muslims Upper castes. than Ansaris) and Hindu UC– Source: Authors’ estimation based on GIDS Dalits have approximately Survey on Social and Educational Status of 27% representation in the OBC/Dalit Muslims. lowest stratum. Both these groups have 10% to 12% representation in the top stratum, indicating the presence of a tiny welloff section among them. Dalit Muslims too have 9.8% of their population in the top stratum alongside 35% population in the bottom stratum. Complex Caste Politics

A caste-wise analysis3 of the UP legislature might surprise someone who assumes ascendency of the OBC and Dalit parties to mean the end of upper caste preponderance in state politics. This feeling might get accentuated by the fact that in the last two assembly elections more than three-fourths of the members who returned to the Vidhan Sabha were from the BSP, primarily representing the Dalits, and the SP, which stood for the backward castes. Hindu upper castes won close to one-third of the seats4 in assembly elections held in 2007 and 2012, though their population5 in the state would be less than 20%. Similarly, upper caste Muslims also won a disproportionately higher number of seats in the last two assembly elections. Their representation in 2007 and 2012 state assembly elections was around 10% and 13%, respectively, against a population share of around 6% to 7%.6 Surprisingly, the number of members of the legislative assembly (MLAs) belonging to castes that are seen as holding the reins of power is not so high. Rather, it is roughly proportionate to their population share. The share of Yadavs—who constitute the core of SP—in the UP legislative assembly in 2012 was below 10%, whereas they formed 10.4%7 of the state’s population. Similarly, 11.76% Jatav–Chamars—the core of BSP’s 122

support base—could secure just proportionate representation in the legislative assembly in 2007 when the BSP formed the government in UP. On the contrary, the Hindu upper castes, continue to be over-represented in the state legislature in relation to the proportion of their population. For instance, it has been observed that when the SP wins elections in UP, the Thakurs emerge as the largest group in the state assembly, and in the scenario of BSP’s victory, none other than Brahmins occupy the maximum number of seats. Together these two castes do not constitute more than 15% of the population of the state, but in each election they have held more than 25% seats in the assembly. Other prominent castes who have a significantly higher share among MLAs as compared to their population proportion are the Kurmis and Pasis, both largely concentrated in the Avadh region of the state, with some pockets in Eastern UP and Bundelkhand. Each of these castes won close to 6% seats in both the elections against a population share of around 4% each. The Jats could only achieve proportionate representation in 2012, but in the 2007 assembly, they were close to double their population share. On the other hand, the other Hindu OBCs have a huge representation deficit. Forming around one-third of the total population, they have been unable to win more than 10% to 11% of the assembly seats. This group includes castes such as Gujjar, Kushwaha, Maurya, Nishad, Pal, Kalvaar, Bind, Chaurasiya, Prajapati, Madhesiya, etc. Of the 79 castes included in the OBC list of UP, a majority of them do not have any MLA s. The Muslim OBCs are also grossly under-represented. They have not been able to win more than 3% to 4% seats even with around 11% population share. In the 16th assembly elections, acquiring nearly 17% of the assembly seats, Muslims achieved an almost proportionate representation. This share was around 14% in the 15th assembly. In both 2007 and 2012, the share of Muslim OBCs and Dalit Muslims taken together was around 27% of all the Muslim MLA s. These two groups constitute 70% of the Muslim population in the state. One thing that comes out very clearly from the data on landownership and the political representation is close association between the two. Most of the castes and SRGs who have relatively higher share in land also appear over-represented among the MLAs. Thakurs, Brahmins, other upper caste Hindus, Yadavs, Kurmis, Jats, upper caste Muslims, etc, who own substantial land are powerful politically as well. However, Jatav–Chamars and Pasis also have a strong presence among MLAs without owning sufficient land. One can easily attribute this to reservations for SCs in assembly seats, but it does not explain the varied performance in different spheres of the various Dalit sub-castes. The outperformance of Kurmis and Pasis might also be a consequence of their concentration in the Avadh region of the state. For instance, Pasis constitute around 49% of the SC population in Lucknow district. Their vote to seat conversion raito seems to be higher than other castes. Similarly, Jats are also concentrated in the Western UP, the more developed region of the state. decEMBER 31, 2016

vol lI no 53

EPW



Further, a closer examination of the household data also reveals that each caste is subject to increasing internal differentiation, consequently making it challenging to articulate a politics that reflects the aspirations of the entire community. The loosening grip of the BSP over Jatav–Chamar voters is a case in point. This by no means should lead us to an outright rejection of caste considerations, but the point is that identity equations alone do not explain the electoral politics of the state. The over-representation of Muslim upper castes among MLAs persists at the cost of the OBC and Dalit Muslims. The Muslim elite have always portrayed the Muslim community as homogeneous one and articulated a common agenda for the entire community. The near monopoly of upper castes over Muslim leadership finds its sustenance in the wide inequality between Muslim SRGs on all indicators that is much wider than that found among Hindu groups. With a narrow resource base and limited opportunities for educational mobility and occupational diversification, Muslim OBCs and Muslim Dalits find it difficult to challenge the upper caste hegemony. But as the household data shows that some Muslim OBC castes like Ansaris have moved faster than others to access educational and occupational opportunities. Summing Up

Even a cursory glance at the multilayered nature of electoral politics exhorts social science research to look beyond the obvious identity equations. This article attempts to portray Notes 1 2

3

4

5

6

7

CSpro software for data entry and cleaning and SPSS 22 for data analyses was used. Grade A refers to administrative and executive officials, architects, engineers, doctors, working proprietors, directors, managers, social, biological, physical scientists, lawyers, jurists and such jobs, both in private and public sector. Grade B includes teachers, clerical stenographers, nursing and medical health technicians, accountants, bookkeepers, telephone and computer operators, and similar jobs, again in both sectors. This data was obtained from “Samanya Nirvachan Mein Nirvachit Sadasyon ka Parichay” (Introduction of Honorable Members Elected in General Election) accessed from the UP assembly website. However, some of the members have not given their caste/social group in this document. Their information was collected from local sources. This method involves risk of misreporting in some cases, but the overall picture is unlikely to change. The number of MLAs belonging to a certain caste alone might not represent the relative political power in the hands of those group. Extrapolation based on 1931 Census figures place it around 21% for undivided Uttar Pradesh. NSSO 61st round, 2004–05. A point to note is that data given by the census, NSSO, and estimates based on 1931 Census do not match completely. They should be treated as rough proportions rather than precise estimates. Statistic from the Samajik Nyay Samiti Report (Report of the Social Justice Committee, 2001), Government of Uttar Pradesh, Lucknow. The Committee used “parivar register” to collect caste-wise population. In this way, caste-wise proportion of the rural population was taken


EPW

decEMBER 31, 2016

a more nuanced picture of the state politics by referring to landownership, educational mobility and occupational diversification among certain SRGs. It further seeks to demonstrate that the apparent homogeneity of social groups hides considerable heterogeneity between castes. For instance, the household data reveals wider disparities among OBC castes than among Hindu upper castes in terms of landownership. In addition to social identity issues, these divergences provide political parties the rationale for approaching caste groups which are not the core constituency of another political party in the state. The heterogeneity within a specific caste also provides political parties maneuverability of approaching sections of various castes for support. But if internal differentiation necessitates invoking of the larger social identity to reinforce caste unity, the same intra-group inequalities might pose a threat to elite leadership and agenda. This tension is quite visible in the state’s politics. In the negotiation between parties and castes, powerful groups appear to augment their bargaining capacity by their economic clout. Political parties give undue weightage to these groups which represent a minuscule portion of the electorate. In the case of other groups also, it is the relatively well-off sections among them that play a crucial role in this kind of bargaining. All these insights into the dynamics of electoral politics only push us to investigate deeper into these issues.

for the entire population of the state. However, the committee has rightly noted that in 2001, around 84% of the state’s population was residing in rural areas, and the data on this population could provide an assessment of caste-wise and social groupwise proportions of the state population.

References Ahmad, Irfan (2003): “’A Different Jihad: Dalit Muslims’ Challenge to Ashraf Hegemony,” Economic & Political Weekly, Vol 38, No 46, pp 4886–91. Chandra, Kanchan (2007): “A Benign Rupture in Uttar Pradesh,” Indian Express, 12 May, http:// www.indianexpress.com, accessed on 10 December 2009. Chhibbar, Pradeep and Rahul Verma (2014): “The BJP’s 2014 ‘Modi Wave’: An Ideological Consolidation of the Right,” Economic & Political Weekly, Vol 49, No 39, pp 50–56. Deshpande, Satish (2013): “Caste and Castelessness: Towards a Biography of the ‘General Category,’” Economic & Political Weekly, Vol 48, No 15, pp 32–39. EPW (2012): “Sixteenth Assembly Elections in Uttar Pradesh,” Economic & Political Weekly, Vol 47, No 14, pp 80–86. Jodhka, Surinder (2014): “Emergent Ruralities: Revisiting Village Life and Agrarian Change in Haryana,” Economic & Political Weekly, Vol 49, Nos 26 & 27, pp 5–17. Kumar, Sanjay (2014): “BJP Crafts a New Social Coalition in Bihar,” Economic & Political Weekly, Vol 49, No 39, pp 95–98. Lieten, G K and Ravi S Srivastava (1999): Unequal Partners: Power Relations, Devolution and Development in Uttar Pradesh, New Delhi: Sage Publications. vol lI no 53

Mosco, Vincent (2009): The Political Economy of Communication, Second ed, Sage Publications. Oommen, T K (2011): “Structural Differentiation and Developmental Politics in Post-liberlisation India,” The Globalisation Turbulence-Emerging Tensions in Indian Society,” Prashant K Trivedi (ed), New Delhi: Rawat Publications. Pai, Sudha (2002): Dalit Assertion and the Unfinished Democratic Revolution—The Bahujan Samaj Party in Uttar Pradesh, New Delhi: Sage Publications. — (ed) (2007): “Political Process in Uttar Pradesh: Identity, Economic Reforms and Governance,” Delhi: Pearson Longman, pp 136–56. Palishkar, Suhas and K C Suri (2014): “India’s 2014 Lok Sabha Elections: Critical Shifts in the Long Term, Caution in the Short Term,” Economic & Political Weekly, Vol 49, No 39, pp 39–49 Sridharan, E (2014): Class Voting in the 2014 Lok Sabha Elections: The Growing Size and Importance of the Middle Classes, Economic & Political Weekly, Vol 49, No 39, pp 72–76. Srivastava, Ravi S (2007): “Economic Change among Social Groups in Uttar Pradesh, 1983–2000,” Political Process in Uttar Pradesh–Identity, Economic Reforms and Governance, Sudha Pai (ed), Pearson Longman: Delhi, pp 345–66. Tiwari, Badri Narayan (2007): “BJP’s Electoral Strategies: Development, Caste and Electoral Discourse,” Political Process in Uttar Pradesh– Identity, Economic Reforms and Governance, Sudha Pai (ed), Delhi: Pearson Longman, pp 136–156. Verma, Anil K (2007): “Backward-Caste Politics in Uttar Pradesh: An Analysis of Samajwadi Party,” Political Process in Uttar Pradesh –Identity, Economic Reforms and Governance, Sudha Pai (ed), Delhi: Pearson Longman, pp 157–90. Verma, A K (2014): “Development and Governance Trump Caste Identities in Uttar Pradesh,” Economic & Political Weekly, Vol 49, No 39, pp 89–94.

123