How Caste Continues to Determine Asset and Landholding Structure in Rural India

By highlighting the inadequate representation of Scheduled Tribes (ST), Scheduled Castes (SC), and Other Backward Castes (OBC) in upper echelons of government and industry, Rahul Gandhi has reinvigorated the demand for a caste census in the country. Subsequently, several leaders across political formations have supported this demand.

However, many prominent commentators and journalists have argued that a caste-census will rekindle the waning shadow of caste and create deep fissures in society. So far, the debate has not included any analysis of economic disparity between different castes.

This article addresses this lacuna by presenting evidence of caste-based disparities in incomes and landownership in the agricultural sector. The evidence suggests that a caste census will not only allow us to assess the extent of ‘general’ caste domination in the ownership of the means of production but will also provide evidence on intra-caste heterogeneity in asset and land ownership within various marginalised caste groups.

The Situation Assessment Survey (SAS) of Agricultural Households of 2018-19 is the most recent nationally representative sample survey which records details of landownership by farming households across the country. So far, only three rounds of the SAS data have been collected.

The first round was conducted in 2003, and the second one was almost a decade later in 2013, and finally the third and last wave of data was collected in 2018-19. Further, no information about future SAS data collection is available on the Ministry of Statistics and Programme Implementation website.

The other major source of landholding data is the quinquennial Agricultural Census. The most recent publicly available data from this Census is for 2015-16 as findings from the current round which began in 2021-2022 is yet to be made public. Therefore, akin to the absence of periodic consumption and employment data in the last decade, more recent information on farm incomes and landholding is also scant.

This makes it harder to assess government claims about rural incomes and asset ownership during the BJP regime. However, since rural land markets in India are thin, information from the 2018-19 SAS data can be useful in understanding the caste-based landowning patterns today.

In this article we use the SAS data to show how landownership and consumption expenditure vary across castes in different states. For this analysis, we divide states into six groups.

The first group consists of northern states including Jammu and Kashmir, Himachal Pradesh, Punjab, and Haryana.

The second group consists of central states – Uttar Pradesh, Uttarakhand, Madhya Pradesh and Chhattisgarh while the third group consists of the northeastern states – Arunachal Pradesh, Assam, Manipur, Meghalaya, Nagaland, Mizoram, and Tripura and Sikkim.

The fourth group consists of the eastern states of West Bengal, Jharkhand, Odisha and Bihar.

The fifth group includes the western states of Rajasthan, Gujarat and Maharashtra and the sixth group comprises of the Southern states including Andhra Pradesh, Telangana, Karnataka, Kerala and Tamil Nadu.

ST households have the lowest levels of monthly consumer expenditure

Table 1 shows the average monthly consumer expenditure (MCE) incurred by a household for each social group – SC, ST, OBC and general category in each of the six regions of the country.

Since the MCE comprises of expenditure on consumption goods and durable assets, it is considered an indicator for a household income. Northeastern states have the most equal MCE distribution across caste groups in the country.

However, in all other regions, general category households have the highest levels of MCE (10,611 INR) while ST households have the lowest levels of monthly consumer expenditure (8,204 INR). In the Central and Southern states, the average MCE of general category households was almost twice that of ST households.

Further, in all regions, MCE by SC households was comparable to that of ST households and the MCE of OBC households is closer to that of the General category households. The distribution of MCE by caste groups suggest that the SC and ST households are still lagging other caste groups across all regions except the Northeast.

General category households are most likely to live in pucca houses

Next, we analyse the nature of dwelling structures in the countryside. Table 2 shows the proportion of rural households owning permanent (pucca) houses in each of the six regions of the country. Once again, in all regions, except the northeast, general category households are most likely to live in pucca houses in our villages (57 per cent).

The corresponding figure for ST and SC households is only 27 per cent and 36 per cent respectively. The disparity is most stark in Central states where only 18 per cent of ST and 33 per cent of SC households have pucca houses while the corresponding figure for general caste households was 64 per cent. Both Northern and Southern states have relatively lower dispersion among castes in the proportion of households living in pucca houses across castes.

Land Access Index lowest for SC households

Finally, we discuss caste-wise distribution of the most important asset in the countryside – agricultural land. For this analysis we construct a Land Access Index (LAI) which is defined as the percentage share of agricultural land owned by a caste in total agricultural land divided by the percentage share of that caste in the rural population.

For example, if a state had 50 per cent OBC households, and they own 50 per cent of the agricultural land, then the LAI for OBC households would be 1. In other words, OBC households’ land ownership is proportional to their population.

Similarly, if a caste’s LAI is less than 1, it means it’s landownership by that caste is less than its population proportion and an LAI greater than 1, implies that the caste disproportionately owns more land than its population proportion.

Table 3 reports the LAI for each caste across the six regions of India. Four important points emerge from the analysis of caste-wise landownership. First, the average LAI for ST households across the country is 1.2. This suggests that ST households who lagged other castes in MCE and pucca housing, are not the worse-off group in terms of land ownership.

However, there may be significant differences in the quality of land owned by ST households. Second, the LAI is lowest (0.5) for SC households. Therefore, the proportion of land owned by SC households is almost half of their population proportion. This implies that average landholding of SC farmers would be smaller than that of other caste groups.

Third, the average LAI for OBC households is 0.9 which suggests that landownership by OBCs, the largest caste group in the country is roughly proportional to their population shares in most regions of the country. However, we cannot ascertain intra-OBC disparity in landownership as that data does not exist.

Finally, the average LAI for general caste households is 1.5 which means that this group’s landownership is 1.5 times its share in the population. Moreover, there are significant regional variation the LAI for this caste. In Southern states, general caste households’ landownership is almost twice their proportion in the population. The corresponding figure for Eastern and Central states, is 1.8 and 1.7 respectively.

Our analysis shows that monthly consumer expenditure, and pucca home ownership by SC and ST farmers continue to lag the general and OBC households. Further, in terms of landownership, general category households disproportionately own more land than any other social group. Data from farmer households contributes to the debate around caste census in two ways.

First, the disproportionate landownership by general castes across the country shows their dominance in ownership of productive assets which may form the basis of their socioeconomic prominence in the countryside.

Therefore, a caste-based census may illuminate differences in ownership and employment patterns across sectors which may be necessary for more targeted welfare and redistribution policies. Second, the aggregated indices for OBCs suggest that their landownership is proportional to their population proportions. However, this may conceal intra-caste disparities between various sub-castes within the OBCs. Therefore, a caste census could better inform public expenditure and policy design.

Kartik Misra teaches economics at Sewanee: The University of the South.