Introduction

This article aims to conceptualise and identify the socio-economic classes in two study villages located in north Bihar. A socio-economic class comprises a group of people defined by their relationship to the means of production and with one another in the process of production. The structure of agrarian classes, and their interrelations with caste, gender, and other forms of sectional deprivation, continues to be a fundamental determinant of social and economic inequality in rural India (Ramachandran 2011). Studying class structure is vitally important for understanding the mode of production, process of surplus generation, and forms of exploitation in the agrarian system of the study villages. For this purpose, we have stratified the households in the villages by socio-economic class, drawing on the analytical framework developed in Ramachandran (2011).

The article is arranged as follows. Debates over the formation and characterisation of agrarian classes in Bihar are long-standing, and a brief account of the literature is provided in the next section. A modified version of the framework developed in Ramachandran (2011) is provided thereafter. We then proceed “to establish certain general theoretical categories and criteria in order to distinguish classes . . . and, on the other hand, to identify classes in situ,”¹ in the two study villages. We describe aspects of economic and social power relations through three case studies. The final section summarises some of our observations.

Conceptualisation of Class in Studies of Bihar

Caste and agrarian class relations have long been debated in Bihar, perhaps because the State has been, and continues to be, a largely agricultural economy. Most early studies of Bihar conceptualised agrarian relations in terms of class, while also studying the interface between caste and class.

In his study of agrarian transformation in Bihar, Pradhan Prasad (1987, 1979, 1975) writes of the foundational importance of the economic roots of the peasant movements in the State. In terms of class differentiation, he divides agrarian classes into two broad categories: a rural oligarchy, and direct producers (which include agricultural labourers and poor peasants). According to him, caste plays a stimulating role in intensifying emerging contradictions in rural Bihar, but not a constitutive role in social and agrarian unrest (Prasad 1975). Chaudhry (1988), in his case study of agrarian unrest in Patna district of Bihar, concludes similarly that class contradictions masquerade as caste conflict and argues that caste does not constitute a fundamental category of agrarian change (see similar analyses by Chandramohan 1998; Louis 2000; and Sinha 1977, 1978a, 1978b and 1978c). Das (1983, 1984) also argues that it is essentially class-based mobilisations that are articulated in the form of caste in Bihar.

Chakravarti (2001) has a contrasting view which posits that caste relations are embedded in agrarian class, at least in the State of Bihar. Caste and agrarian class should not be seen as two separate and analytically exclusive categories because, “whether a person controls land or not is conditioned by that person’s caste status” (ibid., p. 1449). Taking issue with the pronouncements of ‘class/caste war’ across central Bihar, Chakravarti points out that this arbitrary division of factors behind agrarian unrest does not provide a comprehensive analysis of the actual form that class struggle takes following the classical Marxist understanding. He explicitly rejects the caste–class binary, and argues in favour of studying both agrarian class and caste within a single framework. In his framework, which distinguishes between the principal exploiters of labour on the one hand and an underclass on the other, the data he has obtained from his study village clearly indicate that there is a “strong connection between belonging to an upper caste and being an exploiter of labour”; and that “the strong connection between social status and membership of the underclass” is found among the Dalits.

Chakravarti (2001) utilises this understanding in his conceptualisation of agrarian classes, whereby he outlines five classes, namely, malik,² grihast,³ tenant cultivator, petty cultivator, and landless labourer. In his study village, maliks and grihasts were the “principal exploiters of labour,” and together constituted only 12 per cent of total households but commanded over 79 per cent of owned land. Grihasts were about 8 per cent of total households and owned over 14 per cent of the land. Tenant cultivators (comprising about six per cent of total households) and petty cultivators (comprising about 19 per cent of total households) owned two and four per cent, respectively, of the total land. Landless labourers constituted the largest group in terms of their share in total households at 57 per cent. Chakravarti classifies petty cultivators and landless labourers as the “underclass” of the village.

Sharma and Rodgers (2015) assert that there has been a shift in production relations in Bihar since the early 1990s, and that semi-feudal relations⁴ have “virtually disappeared.” They arrive at this conclusion by looking at three broad tendencies that appeared in their study villages.⁵

They noted, first, a decline in the numbers of attached labourers and pure landlords; secondly, increasing proletarianisation and pauperisation of the peasantry; and thirdly, a shrinking of agrarian classes alongside an increase in the number of non-agricultural households, which includes an increase in migration. This has impacted the class structure of the village too, and consequently they identify the following classes in their study villages: a) agricultural labourers, tied or casual; b) poor peasants involved in subsistence farming; c) middle peasants who hire in labour as well as hire out their labour; d) big peasants who are primarily users of non-family labour; e) landlords who primarily rent out land and do not engage in manual labour; and f) non-agricultural households not working in agriculture at all. However, the caste profile of these six classes, in their own analysis, shows a “degree of stability in their relationship.” For instance, the upper castes continue to make up about 65 per cent of all big peasant/landlord households. Their representation in agricultural labour is negligible whereas their share in non-agricultural occupations has increased, from about 14 per cent in 1983 to 42 per cent in 2009. At the other extreme, the share of the Scheduled Castes in agricultural labour seems to have stagnated at about 39 per cent in the 1998–99 and 2009 rounds. Their share in non-agricultural labour is miniscule, at seven per cent, compared to 11 per cent in 1983. The shares in the overall population of the upper castes and the Scheduled Castes have remained almost the same throughout the period of analysis.

Thakur et al. (2000) and, more recently, Sahay (2020) have done away with the category of landlords. Thakur et al. (2000) do not use the term “class” even once in their paper. They divide the village population based on landholding and type of land tenure into four categories: “owner cultivator,” “owner-tenant,” “tenant-owner,” and “pure tenant”. Sahay (2020) argues that there are no landlords in his study villages, and divides the village society into five classes based on landholding: landless and near-landless, small peasants, lower-middle peasants, upper-middle peasants, and large peasants. According to Sahay, the share of landless households in his study villages continues to be very high, at about 60 per cent. The large peasant category forms only about two per cent of total households but controls over 25 per cent of all cultivable land. Similarly, the upper-middle peasants are about three per cent of total households but own about 15 per cent of all land. In Sahay’s view, none of these features constitutes landlordism in the study villages, because many of the landed households engage in manual labour and reverse tenancy.

In conclusion, let us look at some pointers for the identification of socio-economic classes in the study villages. First, the relationship between caste status and class position needs to be explored in greater detail, using qualitative evidence wherever possible. This is because, following Chakravarti (2001), we have learnt from our village studies that caste and class relations are intertwined. While Sharma and Rodgers’ (2015) data point to the persistence of intertwined class–caste relations, they do not pursue this point in their analysis. Secondly, there is a debate about whether the recent transformation of agriculture in Bihar is leading to the end of landlordism (see, for instance, Sahay 2020). We need to investigate this further through our data and qualitative evidence. Finally, most of the earlier studies have taken landholding as the primary criterion for the identification of socio-economic classes. We need to develop a methodology that brings more factors into the process of identification of various classes in the study villages.

Framework for Socio-economic Classification of Households

In this article we try to identify different socio-economic classes and their characteristics in the study villages by drawing on the framework developed in Ramachandran (2011) and CPI(M) (2016).⁶

Based on household-level data, we use broadly three criteria for the classification of households: ownership and control over land and other means of production; labour ratio;⁷ and surplus generation in the production year. Within this framework and using the above three criteria, we classify rural households into four classes. The first is that of landlords and big capitalist farmers. Landlord households are historically the dominant households in the village, owning the most and the best land, and that do not engage in manual operations (and so have a labour ratio of zero). With the penetration of capital into agriculture a new class has emerged from among households belonging to the class of rich peasants and upper-middle peasants, known as the class of capitalist farmers. They too do not engage in any manual operations but appropriate surplus by organising farm work using hired-in manual labour. Households in this class are usually from the dominant castes.

At the other end of the class spectrum in the village is the class of manual workers, mostly devoid of ownership or access to means of production. The major source of income for this class is wage employment in the farm or non-farm sector. They engage in almost all kinds of labouring activities available in and around the village. However, they do not work exclusively in the wage labour market and may have multiple sources of income. This is the most caste-heterogeneous class. Households from Dalit and Adivasi communities, however, are usually numerically dominant in this class.

Between the class of capitalist farmers and that of manual workers comes the peasantry. The peasantry is a highly heterogeneous class, and households in this class engage at least in some manual work on the land. For classification of the peasantry and their sub-categorisation, we use five criteria:

1. Ownership of means of production and other assets;
2. Labour ratio;
3. Rent exploitation or rent received or paid by the household;
4. Net agricultural income; and
5. Sources of income of the household.

The economic condition of peasant households fluctuates from year to year due to the interplay of many different factors. Data gathered at a point in time cannot adequately capture these interrelationships. Of the above-mentioned five criteria, ownership of means of production and other assets is least susceptible to year-to-year variations even with a functional land and assets market. The variable labour ratio is complicated by the presence of peasant households in the wage labour market. Furthermore, differences in cropping pattern, use of machinery, agro-ecology, and, most importantly, labour processes, impact the labour ratio greatly.⁸ Nevertheless it remains an important criterion to distinguish the peasantry from the classes of landlords and manual workers. Rent received or paid by households is a crucial criterion with which to identify the location of peasant households within the broad array of the peasantry. For lessors, who belong to the upper section of the peasantry, rent exploitation might continue uninterruptedly due to the persistence of land hunger among the poorer sections of the peasantry and the landless manual worker class. For lessees, however, who usually belong to the poor sections of the peasantry, operational rights over land might change even within a given production year due to the insecure nature of tenancy. In a situation of eviction, landless and land-poor peasants lose operational rights over land, and hence change their class position from peasant farmers to that of manual workers. So, the rent exploitation criterion may identify different sections of the peasantry, but does not address the boundary problems that arise within the poorer sections of the peasantry, and between the poor peasant and manual worker classes.

Another important marker of peasant households is the number and variety of sources of income. Determination of the sources of earnings and their relative contribution to total household income serves two important purposes in socio-economic class differentiation. First, it helps to locate the peasant households within the broad frame of the differentiated peasantry, as well as in solving boundary problems that appear among different strata of the peasantry, and between peasants and manual workers. Secondly, the relationship between ownership of productive assets and sources of income, along with other criteria, informs us about the forms in which differentiation and class formation process take place in a given socio-economic system. The net agricultural income manifests extreme year-to-year variations due to exogenous and endogenous factors. So, net agricultural income collected for one year may not be sufficient to differentiate the peasantry. In our class differentiation scheme, to circumvent the shortcomings that arise when relying on data from one point of time, we use labour ratio, rent exploitation, and net agricultural income to distinguish the peasantry from the classes of landlords and manual workers, and ultimately asset ownership to draw the boundaries among the peasant classes.

There are also households in the countryside that do not fall under the three agrarian classes. These households are classified based on their major source of income, as follows: artisan work and work at traditional caste calling, business and self-employed, salaried, rent-receiving and moneylending class, and pension and remittance receivers.

Socio-economic Classes in the Two Study Villages

Class, caste, and landholding are the three most significant variables in defining the agrarian structure of Bihar. All three are highly correlated, such that the impact of one or two cannot be studied in isolation. Landholding is historically linked to caste, and caste identity is deeply entwined with the agrarian class structure. This is the primary reason for considering landholding and caste as the main variables in house-listing.

The Project on Agrarian Relations in India (PARI) survey in Bihar was conducted in 2012, and the reference year of the data is the agriculture year 2011–12 (May 2011 to April 2012). This study is based on surveys of two villages in the State: Katkuian in West Champaran district and Nayanagar in Samastipur district. In Katkuian, a census-type enumeration was conducted. In Nayanagar, a sample survey was conducted using stratified random sampling, where the stratification was based on caste and landholding. Some follow-up studies were conducted in both villages in 2018.

Based on the framework discussed in the previous section, the following were identified as the determining factors for socio-economic class differentiation: ownership and operational holdings of land, labour ratio, asset ownership, ownership of other means of production, different sources of earnings and their relative contribution in total household income, and last but not least, engagement in traditional caste activities.

We have classified the class of “landlords/ Capitalist farmers” based on the criteria of land ownership and operational landholding, and labour ratio. This class is also at the top of the agrarian class hierarchy. Then we separated the class of peasantry in general from other classes on the basis of household income, operational landholding and ownership of means of production criteria. We used an asset-value criterion to differentiate classes among the peasantry.

The class of manual workers was defined based on the share of income from wage labour. Here, wage income from short-term manual labour migration was also considered as part of household income. Hence, the number of labour days employed in short-term migration is considered in the calculation of the labour ratio.

Finally, based on the major source of income criterion, the major non-agrarian classes are identified as below.

Artisan work and work at traditional caste calling: Households whose major source of income comes from engagement in their socially/historically determined caste activities or artisanal work (carpenters, barbers, etc.).

Business activity/Self-employed: Households whose major source of income comes from business activities or self-employment in the non-farm sector. It includes businesses of all scales.

Remittances/Pensions: These are households dependent on remittances or pension.

Rents/Moneylending: Households whose major source of income is rent from agricultural land and commercial establishments. Households with large rental earnings from agricultural land are not included here; they are classified under ‘landlords/capitalist farmers.’ This class includes households with small rental earnings from land and moneylending. Earning through moneylending is quite prevalent in the villages. Although it is a significant class in the village economy, it is numerically very small in the study villages.

Salaried/salaries: Households whose major source of income is salaried earnings.

Katkuian

Katkuian is located in Bagaha II block of West Champaran district, in north Bihar. The agro-climatic region is that of the North Alluvial Gangetic Plains. The total ownership holding of the village is 510 acres and operational holding is 714 acres.⁹ This means that the village households between them own 510 acres of land and operate 714 acres, indicating the significance of absentee land ownership. Sugarcane, paddy, and wheat are the principal crops of the village, covering more than 85 per cent of gross cropped area. Cultivation in the village is done primarily through hired labour, with operations such as harvesting and threshing remunerated more or less fully on piece-rates. The classes of manual workers survive primarily by doing manual labour in and around the village, and by means of seasonal work migration – sometimes to as many as four or five different destinations. They also lease land on a very small scale – usually less than an acre – and engage in subsistence cultivation of rice in the kharif season and mixed cropping of pulses in the rabi season. The level of ownership in the village of means of production was low and that has created an active market for renting machines at high rates; the rental cost of irrigating one acre just once was around Rs 1,000 in 2012.

A large share of agricultural land in Katkuian is owned by non-resident Bhumihar and Brahman households.¹⁰ Their managers have emerged as dominant households in the village over time. One such manager, who is categorised among landlord/capitalist farmers, was brought to the village in 1965 by a big Bhumihar landlord. The manager is Brahman by caste, from the Mithila region of Bihar. He now owns around 5 acres of land in the village, but operates more than 80 acres. The additional 75 acres of land are in a holding that he has managed for decades. After the Bhumihar landlord left the village, the land was leased out to the Brahman manager for a fixed cash rent, which has not altered much over time. The rent to gross value of output (GVO) for this person is one-third of the prevailing rent in the village. In fact, during the PARI re-survey in 2018, he reported that he had stopped paying rent and filed litigation making a claim over the land. The absence of Bhumihar and Brahman households from the village production system led to a discrepancy between the sum totals of ownership and of operational agricultural landholdings in the village.¹¹

In Katkuian, the five households that own and operate the most extensive landholdings belong to three different castes: three to the Yadav/Ahir caste, and one each to the Brahman and Kushwaha castes. Their average operational land holding was 56 acres, ranging between 15.5 acres and 138 acres. The members of these households did not participate in manual work, so there was negligible use of family labour as compared to use of hired labour. Three out of the five households earned rental income from agricultural land and machinery. Among the five households, one household – of Brahman caste – was part of the historical landlord elites of Bihar, while the other four emerged from Other Backward Classes (OBC). All the households directly or indirectly dominate the social, political, and economic life of the residents of the village, of whom 86 per cent were OBCs.¹²

The demarcation of the peasantry from the class of landlords/capitalist farmers on the one hand, and from the class of manual workers on the other, is based on the extent of ownership and operational holdings, the value of other owned means of production, a labour ratio of greater than one, and share of income from various sources. The peasantry cultivates either their own land or leased-in land, or both. However, the upper section of the peasantry leases out land to secure rent. In this village, 10 of the identified peasant households leased out some part of their ownership holding for rent. On the other hand, among 66 peasant households, 58 per cent leased in land either on share rent or fixed rent. Thus, a majority of the peasant households were tenants and subject to rent exploitation. The labour ratio of the peasantry is generally greater than one, with the exception of those from the upper section who do participate in manual work but whose hiring of labour outweighs the use of family labour in crop production. Moreover, members of these households do not participate in the agricultural and non-agricultural wage labour markets. As we move towards the lower section of the peasantry, the use of family labour increases as also the number of days of work in the wage labour market, and less wage labour is hired in on operational holdings. The households fall between the classes of landlords/capitalist farmers and manual workers, and have labour ratio varying between 0.01 and 150.89 with the average labour ratio being 6.6.

Regarding the number and variety of sources of income, three patterns can be observed. First, the share of crop income is highest among all the sources of income availed by the peasant households. Secondly, the importance of income from wage employment in total household income increases as we move from the richer to the poorer section of the peasantry. Thirdly, apart from crop income, for the richer section of the peasantry, the important sources of income were rent from land and machinery, business, remittances, etc. On the other hand, for the poorer section of the peasantry, the sources of income after crop production were wage earnings from agricultural and non-agricultural work, salary from low remunerative jobs, petty business, rental income, etc. Diversification of income was more prominent among the lower section of the peasantry, primarily in order to cope with the uncertainties prevailing in crop production. The sources of income for the richer and poorer sections of the peasantry were markedly different. Those from the upper section enter the job market with a certain level of education and technical knowledge, while the poorer section of the peasantry is engaged mostly in manual work and has low returns from employment.

The peasant households were ranked based on the following criteria: extent of ownership and operational holdings, the value of other owned means of production, labour ratio, and share of income from various sources. The ranking of peasant households based on the above-mentioned criteria was correlated with asset ownership. The ranked variables along with the following asset size classes clearly demarcate the boundary lines of the peasant classes.

Peasant 1: Total asset value more than Rs one crore.

Peasant 2: Total asset value less than Rs one crore but more than Rs 15 lakh.

Peasant 3: Total asset value less than Rs 15 lakh.

The class of manual workers is identified based on wage-earnings from manual labour, even though they may have diversified sources of income such as operation of agricultural land, income from animal resources, and transfer payments (old age pension, widow pension, etc.). Based on the operation of agricultural land and extent of income diversification, the class of manual workers is differentiated between “manual workers: with operational holding and diversified income sources” and “manual workers: without operational holding.” The characteristics of both are made more complex by the presence of migrants in almost all the manual worker households.

The migrants are engaged in different forms of wage employment in crop production in the States of Punjab and Haryana, and mostly in the informal sector in the destination of migration. The status of such a large number of migrants within the family structure and in the household economy requires further qualification.

Based on three criteria –the likelihood of migrants to return, the frequency of return, and the duration of migration – Gonzalez (1960) classified migrants into three types: temporary-recurrent, indefinite, and permanent relocation (cited in Wiest 1973). The temporary-recurrent type of migration (sometimes referred to as “circular migration”) is of interest to us, in order to locate the wage-labour migrant within the class hierarchy of Katkuian. As defined by Wiest,

temporary-recurrent migration applies to individuals who are absent for a period of time usually not exceeding one year, but who indicate their intention to return, so the resident household members describe the absence in temporary terms

where “temporary migration emphasises the intent of the migrant as perceived by himself and/or household members, rather than the length of his absence” (ibid.)

Apart from the nature of migration, the location of wage-labour migrants in a household – in terms of economic cooperation, co-residence, and domestic function – needs to be established (ibid.). In Katkuian, the wage-labour migrants perform the function of economic cooperation, as their wage-earnings in the destination in some cases supplement household income in the village, and in others, change the structure of income generation through the acquisition of certain means of production. Furthermore, the wage-labour migrants influence household affairs, even in absentia, through different modes of communication, and they remain attached to the household as “non-localised family,” in the words of Gonzalez (1960).

Based on this, we have included the entire labour and wage-income of migrant workers as part of the household labour use and household income in the village, rather than considering only remittances that the household received from the migrant workers. Taking into account only remittances in the household income categorises a large number of such households as part of the class of remittances/pension-holders instead of the class of manual workers. This would have resulted in misclassification of a large number of households whose main source of income comes from the wage employment of migrants, even though the migrants earned their wages outside the village production system. In Katkuian, at least one member from each of 189 households (altogether 346 individuals) migrated in order to obtain wage employment in agricultural and non-agricultural work. The migrants went to multiple destinations in a single year. The wage-labour migrants were mostly from the class of manual workers (both those with and those without diversified sources of earning), and some of them (10 per cent of the total) were from the Peasant 3 class.¹³ These households can be considered “households with two nodes.”

The incorporation of the income and labour of migrant workers in distinguishing classes is perhaps a good example of identification of classes in situ, which is an important task in the identification of the nature of classes (see Ramachandran 2011). It should also be noted that differences in the nature and significance of migration between groups play a part in reinforcing the class hierarchy. There is migration among the landed classes but it is likely to involve more permanent residence – for employment as security guards as well as white-collar employment.

Based on the above framework, Tables 1 to 4 show different aspects of the socio-economic classes in Katkuian. In this village, 82.6 per cent of all households belonged to the agrarian classes. Of all the households, only 19 per cent constituted different classes of peasantry. The classes of manual workers constituted 62 per cent of all households. This shows the concentration of means of production in the hands of just 20 per cent of households (including landlords/capitalist farmers and the peasantry). Among 61 households belonging to non-agrarian classes, 36 households were categorised under business activity/self-employed. The classes of salaried persons and rents/moneylending were very small in the village.

As mentioned, only 20 per cent of all households owned a major share of means of production, specifically agricultural land. For instance, 38.7 per cent of cultivable land was owned by the top one per cent of households belonging to the class of landlords/capitalist farmers. The land ownership of the entire peasantry (19 per cent of all households) was 35.5 per cent. The remaining 25.8 per cent of land was owned by non-agrarian classes and manual worker households. Given the caste composition of the village, among all households, 13 per cent of BC-2/BC households (mostly Yadavs, followed by Kushwahas) belonging to agrarian classes owned 67 per cent of all agricultural land. More generally, BC-2/BC constituted 34.6 per cent of all households but controlled around 79 per cent of agricultural land. On the other hand, BC-1/EBC was numerically the largest caste group, constituting 46.6 per cent of all households, but controlled just 16.5 per cent of agricultural land. This caste group was over-represented in the classes of manual workers in comparison with their proportion in the population. The shares of BC-1/EBC in the classes of manual workers with operational holding and diversified income sources, and of manual workers without operational holding, were 57 per cent and 50 per cent, respectively.

The means of production, primarily for agricultural purposes, were concentrated among four agrarian classes (landlords/capitalist farmers and the three classes of the peasantry), and accounted for 72 per cent of the total value of means of production of the households. Among all the households in the village, ownership of means of production per household was Rs 5,51,191 (the highest) for the class of landlords/capitalist farmers, while the classes of artisan work and work at traditional caste calling owned no means of production. It is important to mention here that the ratio of the value of means of production to the value of all assets was very low in the village, only 2.2 per cent, signifying that capital investment in agriculture was very low.

Nayanagar

Nayanagar is situated in Samastipur district of northern Bihar, and also lies in the Northwest Alluvial Gangetic region. It is a big village consisting of 1,205 households, of which 348 households were surveyed by PARI in a stratified random sample. Nayanagar is a village where the social and economic hierarchies are fully enmeshed. It presents a coexistence of modern machinery, mechanisation techniques, and cropping practices (controlled and executed by Bhumihar households), along with continuity of subsistence-based manual cropping practices (carried on by Dalit households). Households from the top socio-economic classes control all the means of production including, but not limited to, land. Several forms of labour control practices are still prevalent in the village through land and credit. The landed households lease out small parcels of land on a sharecropping basis to attached or non-attached labourers, in exchange for assured labour during the peak season.

Total ownership holding in the village is 2,040 acres and total operational holding, 1,814 acres.¹⁴ A large chunk of the crop area is low-lying land, where, due to water-logging, kharif crops cannot be grown; therefore, most of the cultivation happens during the rabi season. The principal crops grown in the village are maize, wheat, sugarcane, and rice. Bhumihars are the dominant caste in Nayanagar. The proportion of the population made up by Bhumihars and other forward castes was 25 per cent, while 34 per cent of all households were Dalit, 31 per cent of households belonged to Extremely Backward Classes (EBC), and 10 per cent of households were from the Backward Classes (BC).

Traditional land and caste-based relations have a strong hold on Nayanagar (see case studies below). The population share of the Bhumihars is around 20 per cent, but they own more than 95 per cent of the agricultural land in the village with a few households claiming the lion’s share. The extreme inequality in land ownership is the primary factor that governs socio-political and economic relations between the landed and the landless.

In this village, the class of “big landlords” is identified by a labour ratio of zero and ownership holding of 40 acres and above. All seven big landlords are Bhumihars. Their historical claim of being Brahmans keeps them away from manual labour operations. They combine tenancy and self-cultivation on their ownership holdings. Self-cultivation was done using hired labour or long-term labour, or both. All the households in this class extracted rents from various strata of cultivators and manual workers who leased in parts of their lands. Four of the seven households have hired long-term labour either to perform agricultural tasks or to operate machinery. Even though long-term hiring of labour was common among these households, a major share of the labour for crop production came from the wage-labour market.

To identify the peasantry, an important indicator is the labour ratio. The members of peasant households engage in their own cultivation, participate in the wage-labour market (more prevalent among the lower sections of the peasantry), and also hire in labour during the peak agricultural season. In Nayanagar, this pattern of labour does not hold, however, as these agrarian households cannot be classified as either big landlords or manual workers. In the case of these households, it is imperative to understand their cultural and economic position in society. Most of them are Bhumihars or Brahmans, the ritually dominant castes in Bihar. Because of brahmanical notions of purity, Bhumihars and Brahmans, particularly from economically better-off households, do not engage in manual operations. They neither do manual work for their own cultivation nor sell their labour power in the wage-labour market. Recent field studies have reiterated and highlighted the strong relationship between caste status and agricultural work, especially in the case of Bhumihars in Bihar. The self-perception of Bhumihars is of being owners of land who do not perform any manual agricultural work (Nandan and Santhosh 2019). In fact, a recent field study demonstrates that they do not view agriculture as kaam (work), or as a productive vocation or occupation; they describe themselves as “unemployed” even when they supervise agricultural work on their own land (kaam karana) – which is considered appropriate for their caste status (Kantor 2020).

In Nayanagar, use of the peasant class category hardly seems appropriate, given that the labour ratio is close to zero for all households that might be so described. As mentioned above, land ownership in this village historically has been exclusively with the Bhumihar caste, and they do not participate in any form of manual work in crop production. So, technically, they cannot be categorised as peasants. For purposes of class definition with regard to this village, therefore, we have used the term “cultivator.” A working definition of such “cultivators” can be as follows: they own and operate land, they have means of production and derive a major share of household income from crop production, but they do not engage in any manual work in crop production. Household members of this class of cultivators perform the roles of managers and supervisors in crop production. As discussed earlier, this class of cultivators manifest the characteristics of the class of grihast as proposed by Chakravarti (2001). Within the broad category of “cultivators,” the criterion of total value of assets is used to differentiate their positions. Based on the village-specific distribution of value of asset ownership, cultivators are divided into four classes:

Cultivator 1: Total asset value more than Rs one crore.

Cultivator 2: Total asset value less than Rs one crore but more than Rs 50 lakh.

Cultivator 3: Total asset value less than Rs 50 lakh but more than Rs 25 lakh.

Cultivator 4: Total asset value less than Rs 25 lakh.

About 80 per cent of cultivator households in Nayanagar belonged to the Bhumihar caste. The average size of operational holdings was 12.8 acres. The average labour ratio of all the cultivator classes was 0.78. All 64 households classified in the top three cultivator classes were Bhumihar. The average labour ratio of the Bhumihar households in the top three cultivator classes was merely 0.07. On operational holdings, the average number of family labour days was just 15 as compared to 366 hired labour days for the top three cultivator classes. Moreover, apart from using daily-wage labour, they hired long-term workers to do agricultural labour and operate agricultural machinery. For instance, eight out of 14 Cultivator 1 households and 18 out of 26 Cultivator 2 households hired long-term workers (see Case Study 3 to understand the nature of attachment of long-term workers). Moreover, nobody from these households participated in agricultural and non-agricultural wage-labour markets. In terms of rent exploitation, 36 per cent of Cultivator 1 households and 38 per cent of Cultivator 2 households had earned rental income from land and machinery.

We have classified manual workers in the village into two classes.

Manual workers: with operational holding and diversified income sources: Households whose major source of income is manual wage-earning, but who also have small operational holdings and/or earn small incomes from other sources (such as animal husbandry, old age pension, etc.).

Manual workers: without operational holding: These are households whose major source of income is manual wages from agricultural or non-agricultural labour.

In the analysis of data for Nayanagar, as for Katkuian, households with migrants were considered as households with two nodes, and the income and labour of the migrant workers were incorporated into the household economy. In this village, at least one member from 540 households (531 manual worker households and nine Cultivator 4 households) migrated to obtain informal employment. Migration was multi-directional, with workers having migrated as far away as Gujarat to the west, Assam to the east, Jammu to the north, and Bangalore to the south. The most important destination, however, was Delhi – the site of work for 300 workers – followed by Kolkata, Asansol and Assam.

All other households were classified into other categories based on their major source of earning, as discussed previously.

Based on the above framework, about 76 per cent of all households in the village belonged to the agrarian classes (Table 5). Among these agrarian classes, 61 per cent belonged to the classes of manual workers with operational holding and without operational holding. Only 16 per cent of households constituted the classes of landlords and cultivators. In the case of non-agrarian classes, out of 288 households, 122 households or 42 per cent households constituted the class of remittance/pension-holders. The size of the class of salaried person/s was very small.

The axes of socio-economic class and caste overlapped quite strongly in the village. There was no representation of Scheduled Caste (SC), BC-1/EBC, and BC-2/BC among big landlords and the top three classes of cultivators (Table 6). On the other hand, SC, BC-1/EBC, and BC-2/BC were over-represented in the classes of manual workers, with 79 per cent of SC households, 78 per cent of BC-1/EBC households, and 75 per cent of BC-2/BC households belonging to these classes. The representation of other castes in the classes of manual workers was just 4.4 per cent of all manual worker households, and the majority of them belonged to the Bhumihar caste. The participation of Bhumihars in the wage-labour market can be considered as an aberration, brought about by poverty or due to the subdivision of landed property over generations that made crop production economically unviable.

One of the important markers of class power in the village was control over agricultural land (Table 7). Only 16 per cent of all households controlled 80 per cent of ownership holdings. More specifically, 21 households (belonging to big landlords and Cultivator 1 classes) out of 1205 households (constituting less than 2 per cent of all households) controlled 49 per cent of ownership holdings. Furthermore, the top four surplus appropriator classes (namely, big landlords, Cultivator 1, Cultivator 2, and Cultivator 3) controlled around 70 per cent of ownership holdings of land, and they all belonged to the Bhumihar caste. The SC, BC-1/BC, and BC-2/BC households, who constituted 75 per cent of the entire population of the village, owned just 3 per cent of agricultural land. Such a high degree of inequality in landholding implies immense social and economic power for the Bhumihar caste in the village society.

Similar to the concentration of land ownership, the means of production were also controlled by four agrarian classes (big landlords, Cultivator 1, Cultivator 2, and Cultivator 3), with 86 per cent of the total value of means of production controlled by 16 per cent of all households (Table 8). The big landlords owned around 27 per cent of the total value of means of production, averaging Rs 6,97,053 per household. As we move down the class ladder, ownership of means of production per household declines. In aggregate terms, the value of means of production was just 1.3 per cent of the total value of all kinds of assets, showing that land was by far the most valuable asset owned by the households.

Discussion

This paper explores the agrarian structure in two study villages of Bihar through caste, class, and landholding, and shows that the interrelationship among the three variables is high. The findings of our field survey of the two villages depict, on the one hand, a continuity of surplus extraction from tenancy relations and exploitation of different forms of labour, and on the other hand, significant penetration of markets and the introduction of modern machinery.

Sharma and Rodgers (2015), based on a longitudinal study of the villages of Bihar over three decades, concluded that feudal relations -- of the interlocking of labour, land, and indebtedness, and relations such as attached labour -- had largely disappeared. Sahay (2020) too concluded that landlordism has disappeared from rural Bihar based on two factors: first, a heavy decline in surplus extraction from land in the form of rent through leasing out; secondly, near-disappearance of institutions of attached labour.

Landlordism, however, “is not just an economic category, but social and political as well” (E. M. S. Namboodiripad, cited in Ramachandran 2011). We should not conclude that a shift away from the practice of large-scale leasing out of land means that landlordism has disappeared. An agrarian structure persists in which a few households derive their power from land, and control the social, political and economic affairs of the village. Evidence from our study shows stark inequality in terms of land ownership and other means of production in the two villages. In both villages, the top class in the socio-economic class hierarchy owns 42 per cent of village assets. In Nayanagar, the biggest landlord reported having leased out 197 acres to innumerable tenants. In Katkuian also, absentee landlords owned a large chunk of land mostly controlled by persons who had worked as their managers before.

As the data show, more than 60 per cent of households belonged to classes of manual workers, most of them engaged primarily in wage work in agricultural and non-agricultural activities within the village or outside. Our survey also shows the prevalence of three kinds of long-term labour contracts, and confirms the persistence of economic and non-economic coercion in labour relations. First, there were long-term labour contracts to operate machines or manage the farms of big landowners. These were usually seasonal or annual contracts, and the mode of payment was monetary.

Secondly, in another arrangement, the big landowner households employed workers to guard cropped land. In exchange a landowner granted the worker a piece of homestead land and a small parcel of agricultural land. A minuscule portion of the total production from the agricultural land guarded by the worker was also given to the worker (the modal share was 1:40). In this arrangement, the length of the contract and the kinds of work expected of the worker were not defined. The worker had to provide all kinds of labour services and had to be available all the time for work. This also entails long term bondage for the worker households over generations as refusal to provide labour services can lead to eviction from their homestead land. We found this arrangement in Nayanagar.

Thirdly, if a household was in urgent need of money and asked a landed household for money, it had to repay a loan in the form of labour. Credit was given mainly to ensure labour availability in the peak agricultural season. Working males from indebted households could not go out of the village without permission till the debt was repaid in full. In this arrangement, the daily wage – set at Rs 50 less than the prevailing wage rate – was adjusted as a deduction from the debt. The contract continued to be valid till the debt was fully repaid. This kind of contract was prevalent mainly in Katkuian, where migration of agricultural labour to Punjab during the peak agricultural season is common and has been going on for almost 50 years.

Although the first kind of arrangement could be deemed a monetised capitalist long-term labour contract, it is clear that it involves extra-economic pressure and non-economic and unpaid work obligations (see Case Study 3). The second and third kind of long-term contracts surely represent forms of attached labour. With single-point data we cannot arrive at a conclusion regarding the direction of landlordism in the study villages; however, what we observed was its persistence.