In 2011–12, the Foundation for Agrarian Studies, under its long running Project on Agrarian Relations in India (PARI), undertook research in two villages of Bihar. These were Katkuian in West Champaran district, in the far northwest of the State, and Nayanagar in Samastipur district, in central Bihar. The papers in this special section of the Review of Agrarian Studies report on these village studies. The purpose of this Introduction is to draw attention to key findings of the PARI surveys in the context of a brief review of literature on agrarian production relations in Bihar. The papers include studies of land relations and the class and caste structures of the villages; a detailed analysis of wealth inequality; a note on the commonly neglected subject of landlessness and inequality with respect to homestead land; and a note on schooling and literacy which reflects on the still poor state of public education in rural Bihar.

In the late twentieth century, Bihar had an unhappy reputation for appalling agrarian violence. Journalists referred to the “killing fields” of Bihar and under this title described what were usually called “caste wars” between dominant upper-caste or sometimes Other Backward Class (OBC) groups with their organised private armies on the one hand, and (mainly) Dalits, often led by Marxist-Leninists, on the other. Caste and class are deeply intertwined in rural Bihar, as the anthropologist Anand Chakravarti explains, drawing on his ethnographic research carried out in 1978-80 in a village of Purnia (then “Purnea”) district in east Bihar. He writes of a “culture of exploitation . . . determined in crucial ways by the dominant caste, which is also the dominant class. . . . ‘Caste power’ underlies . . . class relations” (Chakravarti 2001, p. 279). The “caste wars” of Bihar were also class struggles.

In his introduction to a special issue of the Journal of Peasant Studies on “Agrarian Movements in Twentieth Century Bihar” (1982), the scholar of rural Bihar, Arvind Narayan Das notes that whereas in the 1930s and 1940s the main lines of agrarian unrest saw substantial tenants pitted against zamindars, by the 1960s and 1970s, agrarian conflict was about the struggles of poor peasants against landlords, rich peasants, and the State. Indeed, in the village he studied in Purnia, Chakravarti found “stark polarisation between the dominant class (comprising maliks) and the underclass (comprising landless labourers and petty cultivators),” with evidence of extensive unfree labour in the labour process (Chakravarti 2001, p. 278). The maliks made up 11.6 per cent of village households while the “underclass” made up 75 per cent. Chakravarti also noted “…. the nexus between class power in the community and various arms of the state” (p. 278). The historian Stephen Henningham, commenting on the papers that Das had brought together, observed the use of the term “landlord” rather than “rich peasant” by the authors of reports concerning contemporary agrarian struggles. This, he thought, underlined the fact that the rich peasants, through their control over land and labour and their capacity to act exploitatively, are comparable to the landlords of the colonial period (Henningham 1984, p. 235). He described those whom he thought should be considered to be “rich peasants’ as ‘the heirs of the generally high-caste village elite of rich peasants and small landlords of the colonial era, supplanted by some lower-caste, former middle peasants who have been able to improve their position through good fortune and the skilled cultivation of cash crops.” (p. 235)

Henningham argued, in line with Chakravarti’s observations, that it was commonly the case that the police sided with the landlords/rich peasants, and that the discontent of poor peasants was treated as a law-and-order problem. It was, he wrote, a “sad commentary on contemporary Bihar that downtrodden peasants are portrayed as irresponsible subversives merely for seeking to attain their legitimate and constitutionally guaranteed rights. And it is because of the apparent inability of the current social order to fulfil their just and often modest demands that they have begun looking towards more radical alternatives.” (p. 235)

It appears therefore, from the observations of these scholars, that in the late twentieth century, the principal class contradiction in rural Bihar was that between landlords/rich peasants (or maliks) and those whom Chakravarti refers to as the “underclass” of landless labourers and poor peasants. What, then, was the character of the agrarian economy that gave rise to this contradiction? An influential answer to this question was given by the late Pradhan Prasad who, in a series of articles (Prasad 1973, 1987), argued that agrarian production relations in Bihar might be described as “semi-feudal”:

The big landowning classes approach the whole process of production and distribution with a view to strengthening their control over land and their hold on the rural masses, resulting in a set-up (which we may call semi-feudal) where an indissoluble bond between the direct producer and his overlord is maintained by “production relations” essentially characterised by the two modes of appropriation, namely share-cropping and usury. (Prasad 1973, p. 872).

He went on to comment on the ruthless coercion exercised by the rural rich, so that the direct producers from the “underclass” rarely sought to secure their legal rights; and in subsequent writing, he referred specifically to the significance of unfree labour in the labour process, as well as sharecropping and usury in semi-feudal relations of production. Semi-feudalism was held to explain agricultural stagnation and persistent poverty. Prasad argued, as did Amit Bhaduri (1973) in his more widely known theory of semi-feudalism, that the production relations he described established an equilibrium with negative implications for investment and for productivity, because development might undermine dependency and the control over labour that the rural rich exercised, and on which their wealth depended. The two economists analysed agrarian production relations in a manner that recalls Daniel Thorner’s earlier argument about the significance of what he called the “depressor” or “the complex of legal, economic and social relations uniquely typical of the Indian countryside” that he held responsible for the low capital intensity that characterised Indian agriculture (Thorner 1976, [1956] p. 16).

The theorisation of agrarian production relations as “semi-feudal” gave rise to vigorous debate, with some scholars (Anand Chakravarti for one) arguing – contra Prasad – that unfree labour may well be reproduced within a capitalist economy. Others questioned whether such relations must necessarily constitute a barrier to the development of capitalist agriculture, and the overcoming of agricultural stagnation and rural poverty associated with it. This was the context in the late 1970s in which Prasad, with the economist Gerry Rodgers from the International Labour Organisation, set out to study patterns of change in rural Bihar more systematically. They ended up, as Rodgers, Mishra and Sharma (2016) explained, with a sample of 36 villages, selected to be representative of the diverse conditions of rural Bihar as a whole (excluding that part of Bihar which became the separate State of Jharkhand in 2000). The villages were first surveyed over the period 1981–83, then again in 1998–99, and most recently in 2009–10. Rodgers and his co-authors have discussed their findings from this remarkable corpus of data in a series of publications (see, in particular, Rodgers et al. 2013; and Rodgers, Mishra, and Sharma 2016).

In 1981–83, according to the data from the village surveys, rural Bihar did have characteristics of semi-feudalism, with 16.6 per cent of village households being made up by those depending on attached labour, though many of them also cultivated some land, often as sharecropping tenants. “Landlords renting out land and landowners who only supervised cultivation on their land” (as the class category was defined) made up 8.8 per cent of households, while the underclass of labourers and poor peasants (though the latter constituted only 7.6 per cent of households) together made up almost 60 per cent. Twenty-five per cent of households were those of middle or more frequently of “big” cultivating peasant households (a much higher proportion than in the village in Purnia district studied by Chakravarti). By 2009–10, according to Rodgers and his co-authors, attached agricultural labour had virtually disappeared from the villages (they accounted for only 1.6 per cent of households), and the share of landlords/supervising landowners had declined to a little over just one per cent of village households. Landless agricultural labour households (35.4 per cent of all households), together with cultivating labourers (16.8 per cent) and poor peasants (now 11.2 per cent), made up 65 per cent of village households; and the households of those depending on non-agricultural work of some kind made up 12.7 per cent (up from 5.6 per cent of village households in 1981). The historically better-off village households, however, had maintained their economic position through long-term employment of family members outside the village, and though the labouring and peasant classes had become somewhat better off, the richer classes had more than compensated with high incomes obtained through migration outside. Rodgers and his co-authors sum up:

Overall, we can see both continuity and change. . . The end of semi-feudal relations did not have a noticeably adverse impact at the top of the village hierarchy as new non-agricultural opportunities (largely outside the village) provided an economic alternative for previously dominant groups. Agricultural labour households also had access to non-agricultural occupations through migration, (though) few abandoned agricultural work altogether. (Rodgers et al. 2016, p. 132)

The “big story,” therefore, of change in agrarian social relations and in the agrarian economy of Bihar over the last half-century, has to do with migration outside the villages for work or sometimes for investment outside agriculture. Almost all have benefited, and the kind of dependence of the vast underclass on the dominant landowners that was at the heart of agrarian production relations forty years ago has been greatly reduced – but returns from migration are much greater for the rich, who have turned increasingly outside the villages, as other scholars have also reported (Singh 2005). There is still a peasantry in Bihar villages that has more or less maintained itself – now constituting around one-fifth of all village households – with the big peasant category having shown the least movement out of agriculture.

Gerry and Janine Rodgers told much the same story – though with an additional gloss – in an account of their findings in villages in Purnia district that they had studied in the 1970s and 1980s, when they went back briefly in 1999. Their paper is entitled “A Leap Across Time: When Semi-Feudalism Met the Market in Rural Purnia.” They were particularly struck by a dramatic rise in real wages since 1981, that had come about very largely because of the circular migration from the villages for work outside, that was “occurring on a massive scale” (Rodgers and Rodgers 2001, p. 1979). Whereas wages had been determined in the 1970s by what the principal landowners considered to be necessary to maintain the labour force (what Gerry Rodgers referred to as the “efficiency wage”), wages were now determined by market forces. “In essence,” the two authors argued,

the control of local landlords over their workers and over the village labour market has been undermined. The improving distribution of income [that they observed in this case] has no doubt added to the forces reducing the political power of the larger landowners. (Ibid., p. 1981)

In some parts of Bihar, perhaps more than in Purnia, the political mobilisations of agricultural workers and poor peasants in struggles against the dominant landholders were another part of these forces. “The net outcome has been a decline in the mechanisms of semi-feudalism.” (p.1981)

In some parts of Bihar, no doubt, the opportunities for profitable investment in agriculture as a result of such factors as the availability of new technology, irrigation, mechanisation, and of improved infrastructure of rural roads and markets have also played a part in breaking down semi-feudal production relations. This had happened already in the 1970s in the village in Purnia studied by Chakravarti, where the combination of social and technological factors (here the use of tractors and the scope for multiple cropping generated by canal irrigation) had brought about a steep decline in the significance of sharecropping – though in this case, at the time of the study, the reliance of the maliks on sharecroppers had been replaced by mostly unfree labour (as noted earlier, as many as 75 per cent of manual workers were unfree). Labourers were settled by maliks on land that they owned, provided by them with loans in cash and kind, and in many cases further bound by the provision of a small plot of land for paddy cultivation under a tekdi arrangement (which gave the farm servant a one-third share of the harvest from the plot allotted to him for the season). Chakravarti saw this as a kind of partial proletarianisation of labour, but one in which the polarised nature of the relationship between maliks and the underclass was the defining feature (Chakravarti 2001, p. 167).

How do the PARI village studies in Bihar reflect upon these arguments in the literature? Overall, the kind of polarisation that Chakravarti identified, and that is clearly evident in the data from the surveys conducted by Rodgers and his co-researchers, is also observed in Katkuian and Nayanagar – where it remains as much interlinked with caste relations as it was in the past, according to the earlier studies. Manual workers made up 62 per cent of village households in Katkuian and 61 per cent in Nayanagar – somewhat higher shares than in the Rodgers villages (53 per cent). Using Chakravarti’s conception of the “underclass” (of landless labourers and poor cultivators, or those classified as manual workers and poor peasants in the PARI studies), this accounted for 75 per cent of households in Katkuian and 71 per cent in Nayanagar – again a somewhat higher share than in the villages studied by Gerry and Janine Rodgers (65 per cent), but closely comparable with Chakravarti’s findings in the Purnia village in 1978–80. On the other hand, the landlords in both Katkuian and Nayanagar constituted just about one per cent of village households, exactly as in the villages studied by Gerry and Janine Rodgers in 2009–10; and the share of peasant households in both PARI villages (19 per cent in Katkuian and 16 per cent in Nayanagar) was much the same as in the villages studied by them (20 per cent). In general, it is only among the big and middle peasants (in Katkuian six per cent of village households, and in Nayanagar five per cent) that agriculture remains the most important source of income, and it is only among them that migration is of small significance.

There is extensive landlessness in both villages (51 per cent in Katkuian and as much as 70 per cent in Nayanagar), and households owning less than five acres (that is, landless households plus small and marginal cultivators) make up more than 90 per cent of all households in both villages (94 per cent in Katkuian, 93 per cent in Nayanagar). In Nayanagar one landlord household, owning more than 600 acres, accounts for one-third of all the land owned by households in the village. So much for redistributive land reform. The sharpest indication of polarisation of all, however, comes from the measures of inequality in the two PARI villages that are reported by Swaminathan and Nagbhushan, who comment on “the absurdly high concentration of wealth among the top few per cent of households.” Indeed, they calculate the net worth of a landlord household in Nayanagar as being roughly 600 times that of a household of landless manual workers, who make up nearly 50 per cent of village households. In Katkuian the same figures are that landlords have about 500 times the net worth of one of the landless manual worker households making up 37 per cent of all village households. The figures mean that the landless manual workers essentially have no assets at all.

Rodgers and his co-authors, speaking of “continuity and change,” refer to evidence on the one hand of the sharp decline in attached and presumptively unfree labour, and on the other of the significance of new non-agricultural opportunities outside the village in boosting the wealth of the dominant land holders, as well as that of migration for work for agricultural workers and poor peasants – which may have helped to raise the level of real wages in agriculture. These two trends mean that the extent of polarisation between the dominant class, and agricultural workers and poor peasants has not changed a great deal, though there has been some change in the distribution of incomes in rural society. This is because the new opportunities have been concentrated at the top and the bottom of society, with less in the middle – for the peasants (Datta et al. 2014). While the story from the two PARI villages might also be described as one of “continuity and change” – and with evidence of the same trends in non-agricultural opportunities and migration for work – there are very striking differences between the PARI village stories and that from the villages studied by Gerry and Janine Rodgers. One is that tenancy remains significant, with share-rent contracts being predominant in cases when manual workers, mainly from the oppressed castes, lease in small plots, and fixed-rent contracts predominating in tenurial relations between members of upper castes. But the most significant difference is that both in Katkuian and especially in Nayanagar, landlordism with the employment of attached labour remains significant. An indication of this, as Baksi and Mahato report in their note on “Homestead Landlessness,” is that in Katkuian 17 per cent of households and in Nayanagar 29 per cent do not own the land on which their homes are sited. Those who are “homestead landless” are liable to become dependent on those who own the land on which they live. In these circumstances it is unsurprising that the state of production relations in both PARI villages resembles that which Chakravarti described forty years ago in Purnia.

In both villages there has been substantial investment in agriculture – and in Nayanagar, in technically and commercially sophisticated horticulture – which has been combined with the continuing employment of attached/unfree labour. In Nayanagar, there is apparently extensive leasing out of small parcels of land to manual workers in exchange for assured labour in the peak season (comparable with the practice of tekdi described by Chakravarti). The three case studies in the paper by Dhar and his co-authors provide rich evidence of the persistence of relations of dependence between landlords and manual workers, in which the old terms of raja and praja continue to be used. The practice of leasing out small plots to labour also obtains in Katkuian (116 cases were recorded), along with the contracting of manual workers through cash advances that must be repaid with labour employed at less than the going wage rate, in the peak season. Given the evidence from both the PARI villages of the significance of labour migration from them, the observation of the persistence of labour attachment is very striking, and in sharp contrast with the findings of Rodgers and his co-researchers.

The analysis of Utsa Patnaik’s (cited by Chakravarti) seems entirely apposite:

The superimposition of commercialisation of production for the market upon traditional exploitative social relations, tends to lead to a strengthening of these relations on a servile basis rather than heralding a transition to greater freedom of the worker. (Patnaik 1986, p. 22)

But is this situation appropriately described as one of “semi-feudalism?” The authors of the PARI studies are broadly in agreement with Chakravarti in seeing this rather as a case in which unfree labour is reproduced within a capitalist economy. Their discussion recalls the work of Philip Corrigan in the 1970s, on the persistence of unfree labour in “advanced” capitalist economies, in a paper entitled “Feudal Relics or Capitalist Monuments?” (1977). The big difference from the agrarian production relations that were described as “semi-feudal” by Pradhan Prasad is that the big landowners are not locked into relations of dependence with labour, and they are clearly not constrained from investment in agriculture when it is profitable to do so – even when their principal investments may be outside agriculture, or their incomes come from employment in non-agricultural activities.