Introduction

In December 2021, Bangladesh celebrated 50 years of its liberation from its post-colonial status as East Pakistan. This paper, in two parts, considers the country’s agrarian transition over this period. Bangladesh remains dominated by agriculture despite the rapid growth in urbanisation and some diversification of the economy. It continues to face the challenge of food security that has arisen from its environmental vulnerability as a deltaic floodplain, from the political economy of colonial exploitation, and from its history as a nation of famine on account of a genocidal war. However, much has changed and indeed improved.

The paper aims to advance a thesis about the disarticulation of the Bangladesh farm along with earlier agrarian class relations, and the rearticulation of the agrarian system through services and capital investment (entailing both “rurbanisation” and new forms of class relations and participation). The pursuit of this argument, in Part 2 of the paper, builds on an analysis of post-liberation agrarian history, which is the focus of Part 1 (Adnan 1999). Overall, the paper can be considered as applying a depeasantisation thesis to Bangladesh (Araghi 1995; McMichael 2012).

The paper across both parts pursues a particular thesis. It does so keeping in view three possible agrarian trajectories. One is that there will be the continuation of a nation of small, petty commodity farmers, with some shift from subsistence motivation towards increasing net marketable surplus. I suggest that most commentators continue to maintain this working assumption of the small-scale family farm as the object of any reforms.¹ The second possibility is that Bangladesh agriculture will be dominated increasingly by large-scale commercialised farming, as corporate interests dispossess the peasantry. A third possibility is that we will see a more contextualised hybrid in which the present family farms become a “rentier”² class of small landholders, having in effect handed over their scattered plots to agriculture service contractors acting commercially and capturing most of the gains from the enhanced productivity of such “consolidated” land use. This third scenario is my hypothesis, one which I have been tracking since the 1990s (Wood 1991 and 1999a).

The context for this argument is a growing rural non-farm sector³ comprising expanding agricultural services connected to technical innovations in cultivation alongside more diversification away from field crops and towards horticultural crops, major (meat and dairy) and minor (poultry/eggs) livestock, and, especially, inland culture fisheries.⁴ The latter accounted for 21.8 per cent of total fish production in the late 1980s, and for 54.5 per cent by 2013 (World Bank 2013). The earlier Green Revolution initiated a demand for inputs (variable and fixed – especially irrigation) and services, entailing new supply chains and rural interventions via dealerships for hybrid seeds, machinery (irrigation pump sets,⁵ power tillers and post-harvest equipment), chemical fertilizers and pesticides, and repair and maintenance services. As explored further in Part 2, in agricultural “system” terms, the farm is losing its pre-eminence to services and other activities, which are also increasingly shifting from public to private hands.⁶ The returns to agriculture are therefore shifting within the overall system from “farmers” to other players. These trends have also heralded new external flows of liquidity: initially through public subsidies⁷, but later via private credit (including micro-credit for smaller operators in minor livestock and handicrafts), and strongly from remittances both internally and from overseas.⁸

These trends strengthened from the late 1980s, as indicated by a rise in domestic savings as a percentage of GDP (20 per cent by 2000) and a decline in the consumption share. And, while agriculture accounted for over 50 per cent of GDP in the 1970s, this had declined to 13 per cent by 2017 (Bangladesh Economic Review 2020) while retaining significant employment (47 per cent in 2010, declining to 38 per cent by 2020). Over a similar period, industry rose from 15 per cent to 30 per cent of GDP (though it has declined to 25 per cent more recently), and employment in services had climbed to 41 per cent by 2020.

It must be obvious to any observer from these headlines that Bangladesh has been on a rapid and intensive journey of agrarian change during 50 years of its liberation. This recognition encourages reflection on the important earlier debate about historical structural “depressors” on agricultural productivity in the subcontinent (originally, Thorner 1956 and later, specifically for Bengal as a whole, Boyce 1987) and on what processes have been responsible for lifting them (Harriss 1992 and Leiten 1996 for West Bengal; Wood 1991 and 1999a for Bangladesh). But the rapidity of change also prompts an assessment – to which we return in the conclusion to Part 2 – of what these changes have done for poverty reduction, inequality, and a deeper liberation of the economy and society from the grip of unaccountable forms of power expressed through continuing patronage and rent-seeking. This calls for looking beneath the present headlines about the transforming economy of Bangladesh, to consider both the “feudal” legacy of East Bengal,⁹ and the politicised and essentialist liberation discourse that was designed to reinforce the unity of agrarian East Bengal as a nation of small farmers, in contrast to the then hierarchical feudalism of Punjab in West Pakistan. The question must be asked as to whether the “feudal” legacy is at all compatible with the nationalist small farmer narrative about Bangladesh. The next two sections of Part 1 engage with this problem.

Agrarian Legacy: Quasi-Feudalism and Antediluvian Capital

Rather than the liberation discourse of homogeneity (the claim to there being a nation of small farmers having a common identity, in contrast to “feudal” West Pakistan), the clue to appreciating agrarian heterogeneity in Bangladesh lay in a puzzle: why, in the period 1940–80, did different Marxist parties advance contrasting agendas around the rural class struggle, and thereby inform the activist post-liberation NGOs in their respective regional strategies? Some reforming activists focused primarily on the landholding question, seeking reform either in terms of land redistribution, or improved security and economic conditions for sharecropping tenants. Others emphasised the issue of usurious moneylending as the key instrument of class exploitation, leading to support for institutionalised microcredit with social collateral and low information costs. Yet others were more concerned about the landless, seeking variously to improve their participation in agriculture through wage and remuneration struggles, access to tenancy opportunities, or access to untitled (khas) land; and to support them in non-agricultural activities in the rural economy, such as handicrafts, small-scale production, post-harvesting processing and trading, transportation, and fishing (alternative views discussed in Wood 1981¹⁰).

In my early work I sought to associate these different analyses and applied approaches with geographical variations in local conditions of the agrarian structure (Wood 1981). These showed complex heterogeneity in the patterns of clientelism and personalised dependencies through which absolute surplus value was extracted. In the preceding Mughal and colonial periods, and up to the 1980s, opportunities outside the rural, predominantly agrarian economy were extremely limited. Local patrons exercised monopoly power over key resources and opportunities, thereby determining the livelihoods of everyone else in their localities. But there were important variations in the relations of production prevailing locally, in which recognition of the distinctions between revenue and rent, and accumulation via usury, is crucial for our understanding. “North Bengal,” with its jotedars and toujidars, showed strong similarities with areas over the border in West Bengal and into Bihar, emanating from the sub-infeudation of the zamindari system, following from the Permanent Settlement of 1793. The formal abolition of zamindari prebendary landholding (1952) left erstwhile landholders (at different levels of sub-infeudation) with large tracts of better-quality khudkasht land (considered to be in their personal possession rather than held under prebendary status). Thus, a significant transition occurred from 1940 to 1980 in landholder rights and conditions of tenure. But sub-infeudation then extended below the last, most “junior” tenure holder (a toujidar) into the classes of occupancy tenants (raiyats)¹¹ and their tenants (i.e. under-raiyats and bataidars or bargadars – sharecroppers). Raiyats, as occupancy tenants, had secure title in return for being the primary source of revenue, and emerged, after zamindari reforms, as the significant class of landholders and petty landlords (jotedars), in parallel with those erstwhile revenue holders with their khudkasht holdings. However, given the prevailing technological conditions (before the Green Revolution and mechanisation), various forms of sharecropping (involving different types of bargadars¹²) were required below the raiyat/jotedar/ex-toujidar level in order to manage the land, due to the challenges of direct labour supervision across scattered plots (Rogaly 1994). Class relations were therefore articulated through rents, frequently acting as a “depressor” (Thorner 1956) upon investment and innovation in higher productivity of land and labour.

Outside “North Bengal” the picture across Bangladesh was mixed.¹³ Certainly, there were common features, but other climatic conditions intruded to alter the “classic” North Indian agrarian structure. Uncertainties in coastal areas owing to levels of downstream flood water and saline tidal intrusion made for lower population densities with reduced land productivity and more single cropping. In these areas, while rents continued to be a means of exploitation, patronage through control over employment and debt also featured strongly. The monsoon-inundated North East was similar to the coastal areas, partly due to the single-cropped haor¹⁴ area that was reliant upon in-migrant seasonal sharecroppers in the dry season, but also because of outward rural-to-rural migration for agricultural work and remittances, as well as migration further afield in the flood season. Uncertain volatile cropping seasons also seemed to correlate with higher rural indebtedness.

It should also be noted that East Bengal/East Pakistan was variously disturbed by Partition. Hindu landlords, especially concentrated in the upper end of the zamindari system, deserted their estates (in the Barisal region, for example) for India, and lower-level Muslim landholders quickly moved into the social vacuum, supported by their musclemen. In other areas, where Muslims were more established as toujidars, their power remained intact even as they were soon forced to “retreat” to their khudkasht lands, adjusting their patronage to rents (rather than revenue collection), moneylending, and more direct control over agricultural employment.

In all these versions of the “quasi-feudal” story,¹⁵ class relations were barely monetised and labour hardly commoditised, in the sense that the relationship between patron and client was rarely a single-stranded transaction in a single-period game. The simple reproduction of class relations through the extraction of absolute rather than relative surplus value under static technological conditions was a complex appropriation of labour, characterised by multiple, long-standing ties across the interlocked domains of rent, labour reward, and debt. It also extended into other aspects of behavioural control – over religious observance, morals, honour, and marriage, for example. The intersection between patron and mosque was strong. Multiple-stranded intrusion into the daily life of subaltern households was normal.

The principle of interlocking is key to understanding how these relationships worked. Supplicant labour was rewarded daily with food (cooked or raw) distributed by the employing household. Cash was paid seasonally at harvest times to enable purchase of essential non-food items. Other household members might also be employed for post-harvesting and domestic work, usually for food alone. But it is also important to reflect upon how households became landless and were thus compelled to enter these multi-dependency relations. They and their ancestors might have been landholders of some kind, but subsistence pressures – involving debts incurred to cope with morbidity, shelter, clothing, or shocks – required mortgaging out land, and so started the downward spiral, ending with pauperisation and clientelism for survival.¹⁶ Interlocking works in other ways too. For example, advance payments (effectively, loans) to supplicant labour in lean seasons secure their labour services for below-market rates in high demand seasons.

These agrarian conditions, applying variously to most of East Bengal and heavily summarised above, remained well into the 1980s and even the 1990s. Their analysis was part of a well-established debate across the Indian subcontinent about the relationship between these relations of production and land/labour productivity in the context of fragile food security. In short, quasi-feudal agrarian relations were regarded as a depressor upon risk-taking and entrepreneurship in agriculture, and thus a constraint on the possibilities of expanded output and yields arising from the introduction of Green Revolution technologies. Numerous papers in Economic and Political Weekly during the 1970s and 1980s attest to this point.¹⁷

What was thought of as “underdevelopment” during the British and “Punjabi” colonial periods needs to be connected to the problem of the depressor. Herein lies the key to understanding the agrarian transition. As I argued earlier (Wood 1981), up to the early 1980s, the region’s rural class relations were significantly articulated in the generation of unproductive capital accruing from rents and usury.¹⁸ The options to expand productive capital in East Bengal/East Pakistan were restricted by successive eras of colonial discrimination in which Bengalis experienced the classic conditions of “underdeveloping” regions.¹⁹ Control over the main cash crop of jute shifted from the East India Company monopoly, first to British mill owners and traders, and then, post-1947, to Punjabi control over exports, directing foreign exchange earnings to investment in the western wing of Pakistan. Other sectors of commercial expansion (energy, infrastructure, consumer goods) were also dominated by these external interests and their comprador counterparts in the country. Landlord classes (jotedars) thus accumulated capital through different means – of intensifying the rate of absolute surplus value appropriation, involving a combination of rents, usury, and control over wages and other local prices.

For these rent-receiving classes, their capital did not need to be re-invested productively with accompanying risk-taking. Instead, further land could be acquired from which further rents could be accrued. Residences could be expanded, including city dwellings; sons could be educated in elite schools; wives and retainers maintained; status-seeking pursuits and lifestyles indulged. While it can be true that ostentatious spending generates employment (professional and artisanal), it remains classified as unproductive or antediluvian capital, in the sense of not being re-invested in upgrading the major means of production through which the capital is itself accumulated. The relations of class-based accumulation do not change, so that labour is not really subsumed under capital to enable a rise in its productivity, and thereby shifting from the extraction of absolute to relative surplus value. This static process of formal subsumption (or subjugation) of labour under colonial capital (Brenner 1977) is commonly described as one of simple rather than extended (or expanded) social reproduction, and thereby underpins the depressor argument.

However, while simple reproduction may sustain landowning classes for lengthy periods, there are endogenous limits to this mode of appropriating unproductive capital due to the parallel monopolies over land and labour by neighbouring landlords, as well as the pauperisation of dependent clients from whom surplus value is derived.²⁰ Nothing moves forward, and lives remain fragile. In the case of emerging, liberated Bangladesh, such reproduction of antediluvian capital was a disaster, because it removed the prospect of increasing the net marketable surplus in food to support the growth of other sectors, as well as a growing population – the food security imperative.

The exogenous innovations associated with Green Revolution technologies, however, offered a determining opportunity to address these major challenges of famine and food security, though the changes were not simply linear. Across much of the Indian subcontinent, these core challenges required public or state intervention, and changes in incentives for how capital and wealth might be accumulated. There was more market integration for inputs, including equipment and fuel. But surface water irrigation had virtually collapsed due to bureaucratic corruption and incompetence, and land redistributed to the poor via land reform measures was frequently “reverse leased” as the poor had no access to inputs and groundwater irrigation.²¹ In North Bihar and West Bengal for example, landless Dalit families remained in debt to fund outmigration of family members to labouring opportunities in areas such as Punjab. The remaining small farmers, sharecroppers, and labourers in the villages continued as clients of their respective patrons who controlled access to new, higher productivity inputs via networks with dealers, police, and other officialdom while diversifying, through family members, into business and educated employment. In Bangladesh, similar conditions prevailed into the 1990s, as there were no strong internal growth spots to attract significant internal migration.

Bangladesh’s Own Liberation Theology: A Nation of Small Farmers?

The history of agrarian East Bengal that I have summarised was however at odds with an emerging nationalist self-image through the 1950s and 1960s, as the territory’s regional colonial status within Pakistan was challenged. Nested within the language movements and student protests was an existentialist proposition contrasting a small peasantry exploited at the hands of “Punjabi” landed and commercial interests. The nationalist case was to be advanced through a solidaristic liberation ideology of a nation of oppressed small farmers.

This “nation of small farmers” proposition, very strongly and understandably associated with the framing of struggle between Bengali primary producers and a Punjab-dominated West Pakistan, was a powerful mobilising construct – an alliance in East Pakistan between an oppressed peasantry and a frustrated petty bourgeoisie (students, civil servants, and urban trading classes) experiencing low, colonially imposed ceilings on aspirations (Jahan 1973).

However, with its implied homogeneity, this framing did ultimately get in the way of a proper understanding of agrarian Bangladesh. The distortion was unwittingly stimulated by the otherwise progressive “Cumilla” initiative from 1959 onwards.²² The research and action-research embedded in the Cumilla programme of farmer cooperatives, the Krishi Samabaya Samity (KSS), advanced the small peasant or minifundist image that was projected on to the country as a whole and underpinned the nationalist stance. The flaw lay in the failure to appreciate the uniqueness of the Cumilla–Dhaka belt in terms of its topography and crop-carrying capacity, with double-cropping as a norm and triple-cropping also widespread in the area. This rich natural resource base for agriculture and horticulture over many decades had stimulated a greater population density than elsewhere in the country, leading to what I refer to as “squared fragmentation,”²³ and then more intensive cultivation. A “large” farmer in the belt in the mid-1970s would have no more than five acres, and there were not many such farmers.

There were four problems in projecting this image of the independent, small, virtually subsistence-level farmer on the rest of the country (see Abdullah and Nations 1974). First, it implied, incorrectly, that there were no significant class divisions in the belt itself, thus constituting a strong social base for inclusive cooperatives (the KSS). Secondly, the Cumilla domination of agrarian research in political discourse meant that evidence of variation in this image was insufficiently revealed at the national level, offering a false looking-glass. Thirdly, the socio-economic conditions in other parts of the country, especially “North Bengal,” were almost the opposite of those in the Cumilla–Dhaka belt, with larger holdings, tenants, and landless agricultural labouring classes. It was more proximate to the fading zamindari structures, as described above. Fourthly, the extent of landlessness was only revealed in 1977 by the Januzzi and Peach study (GoB 1977). The earlier understanding underwent a paradigm shift. Within the rural population of almost 90 per cent, approximately 50 per cent were functionally landless in having less than 10 decimals of land, and therefore living significantly below subsistence unless employed. Neither the Cumilla experience nor the liberation image of a nation of small farmer families with shared interests captured the third and fourth points above.

We are confronted, therefore, with the “limitations of the special case” in understanding agrarian Bangladesh well into the 1990s. The founding assumptions of what became known as the “Cumilla model” were overprivileged in the formation of an agricultural and rural development strategy for the whole country, which continued to construct the idea of the small and marginal Bangladeshi farmer as the primary agricultural actor in the country.

Taking the first problem noted above, the evidence from the Cumilla–Dhaka belt was itself misleading in assuming insignificant class divisions in the belt. Anthropologists like Bertocci (1972 and 1977) working alongside the Cumilla Academy during the 1960s had pointed out some divisions through the notions of structural fragmentation and “cyclical kulakism”²⁴ – essentially a churning of fortunes as the land of richer peasants became divided through multiple inheritance, thus establishing a process of rotating inequality which might enable the KSS to be adequately inclusive. Akter Hameed Khan, the founder of the Academy in 1959, himself detected power divisions within the KSS and embarked upon a “Tour of 20 Thanas” in 1970, just before hostilities and the genocide forced him to leave (Khan 1971). His “belt-based” report indicated that all was not well with the homogenous small peasant image. These concerns prompted the commissioning of a study entitled Exploitation and the Rural Poor (Huq 1976). In that volume, several village studies (mine was on “Bandakgram”) revealed the unequal class relationships even within this minifundist setting which were not cyclical, but rather reflected “kulakism” through the rich family’s acquisition of land (usually via bandak²⁵) and consolidation of such holdings by diverting non-land-inheriting sons into small town businesses and government employment. The KSS was clearly being captured, and the best land and loans (Blair 1978) were increasingly concentrated in few hands. But these processes were not apparent to a casual outsider, and were acted out slowly over time, with landlessness effectively disguised by insecure tenancy often on mortgaged-out land which would eventually be lost.

The significance of these findings, and the emergence of further data and village studies elsewhere in the country, then engaged with the other three problems noted above. If such deepening inequality could occur within this apparently minifundist context, then what was happening in the rest of the country? If agrarian relations elsewhere in the country were indeed more unequal, quasi-feudal, and antediluvian, as argued by Abdullah and Nations (1974), then a policy of rolling out the KSS model across the entire country where the socio-economic assumptions about homogeneity were demonstrably different would be misplaced. This conclusion had led to the Land Occupancy Survey conducted by Januzzi and Peach (GoB 1977; Januzzi and Peach 1980). The public “discovery” of the extent of landlessness was clearly a game-changer, reinforced by other field studies. For example, villages were studied in Rangpur (Hartmann and Boyce 1983) and Kustia (Arens and Van Beurden 1977) later in the 1970s. Added to this were both the BIDS (Bangladesh Institute of Development Studies) poverty studies from the 1980s and the applied work of emerging NGOs engaging with landlessness, in the absence of attention from the Government of Bangladesh (GoB), due to its “ideological” attachment to the Cumilla approach through the Integrated Rural Development Programme (IRDP) and/or Bangladesh Rural Development Board (BRDB).²⁶ The idea of target groups among the landless became a focus of emerging national NGOs like the Bangladesh Rural Development Committee (BRAC), Proshika, Grameen Bank, and, a little later, Gono Shahajo Sangsta (GSS), supported by associated studies such as the influential paper, The Net (BRAC 1986).

Food Security and “Irri-boro”

The events leading up to the liberation of the country from its status as a dependent Bengali colony within the artefact that was Pakistan, bequeathed by the British and Partition, were traumatic, combining a devastating cyclone in 1970, a hostile and genocidal military crackdown, and the short war in December 1971. The euphoria of liberation itself was short-lived, with the 1974–5 famine arising from excessive annual flooding, severely damaged infrastructure, and the legacy of depressed agricultural productivity (except in the Cumilla–Dhaka belt). All this made for dependence upon imported staples, with a sudden loss both of effective demand (from collapse in seasonal agricultural employment) and economic entitlements (a combination of rapid change in tradeable values between distressed sales of non-food items and grain value, due to hoarding in local haats or markets). The country was held to ransom by the temporary withdrawal of concessionary food imports (PL 480) from the USA, until the newly “sovereign” Bangladesh ceased exporting jute to Cuba.

Under these conditions, the country experienced a military coup in August 1975²⁷ and the assassination of its founder leader, Sheikh Mujibur Rahman. During his short tenure as Prime Minister, Mujib was acutely aware of the overwhelming challenge of food security – not only to feed a very poor, populous nation, but also for the sake of political sovereignty. As implied above, he saw a nation of below-subsistence small family farms in desperate need of state support in the form of direct subsidies for inputs, alongside price protection delivered through public sector institutions that could bring technological and institutional innovations interacting with the roll-out of cooperatives pioneered by BARD at Cumilla.

Thus, during his short period in power (1972–5), Mujibur Rahman developed the Bangladesh Agricultural Development Corporation (BADC), which exercised a tight grip on agriculture by controlling input supply, subsidies, and credit through public sector dealerships and national banks. A decade later, in the era of structural adjustment, the BADC “behemoth” was seen as a rent-seeking drag on overall economic growth, with agriculture still very dominant in GDP. Mujib’s other institutional creations or upgrades worked through scientific institutions as part of the Green Revolution: the Bangladesh Agricultural Research Council (BARC), the Bangladesh Rice Research Institute (BRRI), and the Bangladesh Agriculture Research Institute (BARI). These institutions were allied to similar institutes in the Philippines and Mexico, and formed part of the Consultative Group on International Agricultural Research (CGIAR) network. Their work was reinforced by the Agricultural University in Mymensingh and is ongoing.

The 1974–5 famine that came so soon after the genocide also prompted deep concern among the international agencies supporting Bangladesh. In December 1979, the World Bank produced a report on Food Policy Issues (WB 1979), which advocated a market-linked procurement and storage strategy with a district network of godowns, enabling seasonal interventions in local grain markets within a 20 per cent band of normal prices to offset hoarding and price hikes. With procurement closer to market prices, resistance to selling to government was low, and therefore the approach was indeed an incentive, given guaranteed forward prices. This policy contributed significantly to reducing seasonal fluctuations in rice prices. However, in the same time frame, the World Bank and the International Monetary Fund (IMF) were promoting “structural adjustment” – essentially a neo-liberal, market-driven institutional conditionality derived from a value position about the corrupt, cumbersome, rent-seeking state acting as the new contemporary depressor on growth, including, crucially, in agriculture. This is a familiar story globally.

Within this strategic shift in the policy environment was a further application of Green Revolution technologies from the early 1980s, led by the public institutions developed by Mujibur Rahman. By far the most significant of these was the introduction of short-stem rice varieties, reliant upon irrigation and managed water levels, as well as other supporting variable inputs (chemical fertilizers and pesticides). This also involved prescribed changes in land management practices (such as taking transplanted aman practices into the earlier, pre-monsoon boro season from late February to late May). Due to its introduction by BRRI, this rice crop is known as irri-boro. It replaced the highly uncertain, early monsoon-reliant, intermediate aus season. There is little doubt that this innovation has been the main game-changer for food security in the country, with Bangladesh thereafter becoming mainly self-sufficient in grain staples. Its principal advantage is predictability when compared to pre-Green Revolution cropping patterns with the monsoon rain-fed aman (broadcast and transplant) as the major annual staple crop. The aman was subject to vagaries of the monsoon, which could result in either excessive flood damage or drought, and lower yields.²⁸ The landscape in Bangladesh has been transformed in the post-rabi season from barren, dry, grey fields to one of emerald-green paddy waving in the breeze prior to its pre-harvest golden hue.²⁹

In a sense, this aspect of the Green Revolution came late to most of Bangladesh, after being pioneered under the favourable conditions of the Cumilla–Dhaka belt as described above. Due to its fertility and more elevated topography (being farther away from the major rivers) triple cropping had been possible, with an early summer rice crop (aus). Elsewhere in the country, this was a more unpredictable, speculative crop, often written off due to early monsoon uncertainties. The initial adoption of irri-boro in the belt was thus aided by cultivators’ familiarity with trying to squeeze a rice crop, aus, between the rabi and aman seasons. Irri-boro thereby became the substitute for aus but under more intensive input conditions, and crucially required irrigation, which brought another technological dimension into play. For the drier boro season, groundwater or low-lift sources were necessary.

There had been limited prior experience during the 1970s with deep tubewells (DTW) across the country, supplied usually with subsidies through the BADC. These were large-scale fixed installations with borings of around 100 feet, capable of supplying about 100 acres of command area, embracing the plots of many different farmers and calling for cooperation between them. They were mainly installed at higher-elevation, drought-prone areas such as the Barind tract in the Northwest, to support transplanted cultivation (t-aman), unreachable by surface water and gravity flow infrastructure. There had been numerous institutional problems with the management of deep tubewells – partly due to BADC inefficiencies and corruption in distribution and installation, alongside technical operating problems. But problems also came about because of plot-holding farmers being unable to cooperate effectively as self-suppliers of water. An amoral familistic (Banfield 1958) distrust of predatory “plot neighbours” fuelled mutual suspicion and cooperative breakdowns, stimulated by inequality and capture, as had been observed in the Cumilla–BARD KSS approach, even under those minifundist conditions (Huq 1976). These institutional problems had resulted in widespread underutilisation of the technology, with only a fraction of the potential command area irrigated (Murshid 1985). By the early 1980s, however, groundwater irrigation options were shifting strongly towards shallow tubewells (STWs) with 30-foot borings, acknowledging high water tables and irrigating around 30 acres of command area. The shallow tubewells were a more disaggregated technology, easier to install and indeed to move around if confronted with institutional barriers, and they reduced – though did not eliminate – problems of cooperation among plot holders in a command area. They also required lower subsidies and could be purchased by means of loans, which was more compatible with structural adjustment pressures from the World Bank, intent on persuading the Bangladesh government to dismantle the BADC as part of privatisation.

These Green Revolution technological and institutional initiatives, intersecting with the structural adjustment policy environment, implied an expansion of agricultural services, and the distribution of returns to the higher productivity of land among cultivators and other stakeholders. The conditions of “squared fragmentation” (as defined above), and of outmigration from villages as people took up other economic opportunities within and outside Bangladesh, “invited” more contracting services from entrepreneurs who were not necessarily farmers themselves. In other words, agrarian relations experienced the entry of capital, changing class-based rural exploitation from the extraction of absolute to relative surplus value as a function of technological change.³⁰

Overcoming the Depressor: The Real Subsumption of Labour under Capital

This route to overcoming food insecurity had to be predicated upon reforming the legacy of agrarian relations which had traditionally relied on intensifying absolute surplus value as the dominant mode of rural class exploitation, not only within the boundaries of Bangladesh but more widely across the northern river plains of the Indian subcontinent. In 1987, James Boyce published his depressor thesis comparing West Bengal and Bangladesh, in Agrarian Impasse in Bengal: Institutional Constraints to Technological Change. This described conditions up to the early 1980s, after which significant transformations were occurring in both regions of Bengal. But were they the same transformations?

By the mid-1990s, there was clear evidence of agricultural growth through rising land and labour productivity (including for tenants, whether sharecroppers or those paying fixed cash rents), and thus increasing returns to both landholding and services. The evidence challenged the continuing relevance of the depressor thesis.³¹ But the challenge, certainly in West Bengal, was how to account for growth. And was the explanation different between West Bengal and Bangladesh?³² How far was the growth experienced in West Bengal due to the institutional reforms implemented by the Left Front government, including Operation Barga and other land reforms, and democratic decentralisation through the panchayats? How far was growth to be explained, as Harriss (1993) and Leiten (1996) both suggested in their field studies, as the outcome of technological changes, bringing about the increased organic composition of capital – or, in other words, by development of the forces of production?

The explanations for Bangladesh seemed clearer. I had earlier argued the case for the significance of agrarian entrepreneurialism in Bangladesh (Wood 1991). Updating this analysis (Wood 1999a), I attributed agricultural growth in Bangladesh to technological innovation, capital expansion, and commercialisation,³³ with a strong implication that “farmers” were not themselves the main sources of that capital but rather derived rentier income from their own land, sharing profits of higher productivity with the contractor-providers of capital inputs and services. One important caveat for Bangladesh, and indeed elsewhere, is that some rich farmers enjoy dual returns from their own land through directly employed labour as well as from their investment in contracting services,³⁴ thus profiting from the enhanced land/labour productivity of their neighbours. An obvious way this occurs in Bangladesh (as well as across different parts of India as far apart as Tamil Nadu³⁵ and Bihar) is through the renting out of tractor power – or via the relocation of owned pump-sets on others’ lands for the pumping of groundwater (Wood 1999b). Another way has been through investing in rice and, in wheat areas, in flour mills, thereby accruing a commission on everyone else’s production in the neighbourhood. Transportation of output of neighbour farmers to markets or of inputs to farm localities is another example.

Conclusion: Agrarian Change Rather Than Land Reform

In reviewing this debate over attributing agricultural growth, I do not mean to dismiss the significance of Operation Barga in West Bengal. There is a heuristic contrast to be made between the notions of agrarian reform and of land reform, which is rooted in the contrast between Bangladesh, a densely populated delta, and the Indus and Gangetic northern plains of the Indian subcontinent. In these northern plains, successive post-Independence governments have sought to overcome agrarian inequalities and productivity depressors through variations of “land reform,” starting with zamindari abolition. These variations have included the imposition of ceilings on individually owned holdings; redistribution of land released by ceilings to poorer families; attempting land consolidation through swapping arrangements to improve efficiency in land, labour, and input management; improved security and rights for tenants (such as in Operation Barga, implemented by the Left Front government of West Bengal); reduction of rural indebtedness through initiatives in rural banking; attempts at establishing cooperatives to pool resources and enhance market leverage either as sellers or buyers; and subsidies to support holders of acquired land to be risk-averse, to invest and produce marketable surpluses as a function of wider input–output market integration.

The objective of land consolidation for efficient management has worked differently in Bangladesh, however. There has been “agrarian reform,” but not land reform as such. Consolidation occurs not in the sense of concentrating land ownership through redistribution of scattered plots, but through changes in its management and operation by irrigation (command areas) and the provision of agricultural services, particularly with regard to fixed, mechanised inputs; in other words, consolidation as a function of capital intrusion and technological imperatives rather than for justice between households and the efficiency of family-based management. Mandal (2021) argues that Bangladesh has bypassed redistributive land reform and cooperatives (despite the Cumilla–BARD legacy) or collectivisation, instead pursuing technology-led agrarian intensification enforced by the decreasing availability of arable land. Why this contrast in agrarian transition? The “natural delta” aspect of land productivity³⁶ attracts higher rural population densities in the delta, and consequently the intensified processes of “squared fragmentation” and smaller per capita holdings, thus reducing the policy options for imposition of ceilings and meaningful redistribution due to the intersection with demographics. Larger holdings in the northern Indian plains also imply that mechanised technologies vary in their lumpiness from the delta. More families can own and efficiently deploy a tractor for their own land with lower spare capacity available for personalised renting out to smaller farm neighbours, whereas that same technology in the Bangladesh delta becomes lumpier relative to a family farm. It cannot therefore be owned in the same way, and has to be deployed across discrete, separately owned and cultivated landholdings, either through sharing (cooperatives) or externally contracted agricultural services. And the cooperative solution for self-supply of inputs and services has not really worked if we consider the underutilisation of deep tubewells in Bangladesh as a case in point (Murshid 1985), thereby reinforcing the rise of agricultural services through contractors and other suppliers of capital and technology.

Thus, in Bangladesh, beginning in the 1980s but accelerating across the 1990s and thereafter, we have been witnessing a steady endogenous replacement of antediluvian capital by productive capital in agriculture. This is not to say that all “rents” have been cleared out of the system and market imperfections remain. However, the expansion of contracted agricultural services is evidence of several trends. First, the problem of achieving consolidation to increase the productivity of the tiny family holdings that have resulted from multiple inheritance – and therefore from “squared fragmentation” – is being solved through contractor operation of command areas (irrigation-led, with other concomitant services and inputs) rather than by regrouping land. Secondly, this “operational consolidation of fragmented holdings” (Mandal 2021) is also addressing the problem of lumpy mechanised inputs by achieving economies of scale. Thirdly, it partially overcomes the free-rider problem among mutually suspicious landholders or tenants by converting them to consumers of externally provided services. Fourthly, it bypasses the barriers for small family farmers in accessing credit, as they are not taking any upfront crop investment risk, thereby reducing loan requirements. Fifthly, and perhaps structurally the most significant, agricultural labour is increasingly de-personalised and untied or “freed” from multi-stranded, multi-period dependencies upon employers, and so is increasingly commodified. Thus, sixthly, combined with the use of productivity raising technologies, there is appropriation of relative surplus value. In other words, appropriation increasingly moves from formal to the real subsumption of labour under capital (Brenner 1977). The persistence of patron clientelism in other spheres, however, indicates that such labour is only partially commodified, with dependencies transferred to contractors who may, for example, still also be larger farmers and landlords retaining patron status in local communities. They may also be local mastaans³⁷ (rural musclemen) and possibly politically connected. In Part 2 of this paper, the implications of this distinctive Bangladesh model of agrarian change will be further explored within the disarticulation/rearticulation framework.