Dried up Bt cotton narratives: climate, debt and distressed livelihoods in semi-arid smallholder India

ABSTRACT

This paper interrogates technological fix narratives on the genetically modified crop Bt cotton in India that claim to address poverty under climate change. Furthermore, I focus on the political economy of input markets as the mechanism for technology adoption and theoretically maximizing profitability for farmers. I compare these narratives to empirical reality by drawing on 94 interviews and 151 household surveys conducted in the south Indian state of Telangana, the second biggest Bt cotton producing state in India. I show how Bt cotton has been propagated as a technological climate fix crop through its technical traits of thriving in higher temperatures and adapting to particular pests in rainfed conditions relative to non-GM cotton varieties. Yet, I show how Bt cotton increases economic risks for farmers due to higher input costs, which are financed by debt relations with market intermediaries. These debt relations, which I term ‘indebtedness treadmills’, are intrinsically linked to droughts and rainfall climatic variability in Telangana, owing to the increasingly unreliable agroecological rainfed and semi-arid Bt cotton growing conditions. The paper therefore highlights unintended consequences of technological fix climate narratives that arise from siloing technologies from their contextual conditions of adoption to the detriment of real-world outcomes.

KEYWORDS:

• Bt cotton
• India
• climate change
• debt
• smallholder agriculture

1. Introduction

Over the last two decades, genetically modified (GM) crops have been positioned by private sector agribusiness, philanthropic organizations, and policy makers as a solution for smallholder agriculture in tackling poverty and climate change (Asafu-Adjaye et al., 2015). Since 1996, GM crops such as cotton, soyabeans, and maize have been adopted in 24 global South countries by ∼15 million smallholder farmers (ISAAA, 2019).

In this paper, I argue that GM crops constitute a technological fix in responding to adverse environmental change. I extend scholarship on technological fixes and GM crops in the context of climate adaptation, building on Nightingale et al.’s (2019) critique of technological fixes, which they argue separates the biophysical climate problem from its political economy drivers. They point out that these fixes prioritize a narrow set of knowledges in addressing climate change, exacerbating inequalities of power and creating climate injustices in the process. I apply their theoretical lens to empirically examine the failures of a key narrative for GM proponents, the technical pest resistant traits GM crops are bred with compared to non-GM crops, which supposedly address poverty under climate change through better adaptation to environmental stress (see in Gras & Cáceres, 2020). I also situate this narrative within the political economy of markets which deliver the credit, inputs and output markets to supposedly increase incomes for smallholder farmers from technology adoption (Pal, 2015).

I locate my paper by analysing Bt cotton, a pest-resistant GM cotton variety, in semi-arid India. Agroecological conditions often include poor soils, erratic rainfall and drought extremes and a short four-month growing season (Gutierrez et al., 2020). In 2019, India cultivated half the global Bt cotton area at 11.6 million hectares, spread across 10 states (ISAAA, 2019). Monsanto (now owned by Bayer), an agribusiness giant and key producer of GM seeds worldwide, paint a technological imaginary for development and climate change in India via Bt cotton and markets, stating:

Our efforts are aimed at enhancing crop productivity, contributing towards doubling farmer incomes and making Indian agriculture sustainable and globally competitive. (Business Today, 2022, p. 1)

Bt cotton is designed to be resistant to lepidoptera (primarily American and pink bollworm) pests. Under particular agroecological conditions, local/global political economy factors and genetic lines Bt cotton was bred into, it may (but not always) reduce pesticide use and increase yields by preventing crop loss under higher temperatures and rainfed conditions compared to non-Bt cotton (Huang & Hao, 2020; Narayanamoorthy et al., 2020). Critical agrarian studies literature on the empirical reality of Bt cotton adoption in the global South is extensive. Authors have highlighted problems from the crop such as deskilling of farmer knowledge through reliance on input dealers (Flachs, 2019; Stone & Flachs, 2018); the economic treadmill of adopting Bt cotton for short term cash to repay existing debts (Luna, 2020); high fertilizer use rather than Bt cotton seeds driving higher productivity (Kranthi & Stone, 2020); the crops failure to address new pests (Najork et al., 2022) and the problems with relying on yield maximization, not yield stability as a response to climate variability (Schnurr et al., 2022).

This paper examines the prospect of Bt cotton as a rainfed technical fix to complex climate problems in smallholder India. It builds on existing literature by Gutierrez et al. (2020) who have explored the ecological and socio-economic conditions of rainfed smallholder farmers in India adopting Bt cotton that has been bred into hybrid cotton varieties. I examine the processes of how climate uncertainty of drought, rainfall extremes and monsoon timing in rainfed areas in India is braided through and exacerbated by economic consequences of crop failure (Matthan, 2022) due to the reliance on high interest informal credit for input investments for Bt cotton cultivation. Often lacking cash or collateral to grow the crop, farmers often rely on interlocked factor markets to buy inputs on credit (charged at 24%–36% annual interest), seek extension advice and sell their crop harvest to repay their input costs through a highly decentralized network of input dealers in rural village peripheries (Bownas, 2016; Ramprasad, 2019).

To frame my argument for this paper, I draw on mixed methods ethnographic research consisting of 94 interviews and 151 household surveys conducted over eight months in 2019 in rural Telangana, a semi-arid and smallholder farmer state in south India and the second largest Bt cotton producer in India (Directorate of Economics and Statistics Telangana, 2019, 2022). I find that in the case study village of Kavarampur, Bt cotton was adopted as a rainfed crop by all farmers since its entry into the village. Short term, high interest credit financed Bt cotton production goals promoted by input dealers intersected with existing and rising climatic vulnerability to droughts and monsoons. This worsened long-term harvest losses for households across class, caste and entrenched indebtedness treadmills within interlocked factor markets. The most severe effects from Bt cotton were borne by marginal SC farmers with the smallest landholding sizes, least collateral and reliance on informal credit, facing regular threats and humiliation for their inability to repay debts, let alone interest repayments.

I argue how Bt cotton in Kavarampur was enmeshed with uncertainty and climatic risk from droughts and rainfall variability in rainfed agriculture for smallholder farmers, in large part exacerbated by the political economy of market integration in capitalist agriculture. Persistent harvest failures of input intensive GM crops under rainfed conditions were worsened by compounding indebtedness from usurious interest rates from credit used to finance input investments. This created a dependent debtor and creditor relationship between farmers and input dealers, whereby persistent crop failures rendered farmers being unable to repay debts and only pay the interest, doubling down on Bt cotton again to cover their losses. I therefore consider how and why GM technological fixes to adapt to climate change are flawed due to their narrow definition of success and lack of nuance for the differentiated social, political, environmental conditions of technology adoption which can undermine the technology’s ability to aid climate adaptation (Nightingale et al., 2019).

The rest of this paper is organized as follows. First, I explore the literature on GM technological fix narratives in agriculture in relation to climate change and market integration. Second, I contextualize the literature on Bt cotton as a technological fix and its adoption in India, before introducing the Telangana case study and methods. Third, I empirically analyse Bt cotton adoption in semi-arid rural Telangana. The final section situates the empirical findings of the research within wider debates on technological fix narratives for the climate.

2. GM biotechnological fixes and climate change

Bringing together agribusiness, philanthropic organizations, and international financial institutions, crop breeding since the 1980s has centred around the ‘gene revolution’ (Parayil, 2003, p. 1). As Schnurr et al (2022) explain, genetic modifications made to plant DNA are intended by GM crop proponents to overcome incumbent problems of climate extremes, pests or disease in the global South through technological innovation, fixing rural poverty through economic growth and industrial agriculture.

Bt cotton is a key GM crop that was introduced by Monsanto in the USA in 1995. It is a genetically engineered cotton variety containing a gene from a soil bacterium that binds to the bollworm pest DNA and kills it after the cotton leaf or bud is ingested (Flachs, 2016). Approved in China in 1997 and India through a joint venture with Indian seed firm Mahyco in 2002, the crop spread to 15 countries by 2019, of which 13 were developing countries (Tokel et al., 2021). In just over two decades, Bt cotton has covered 70% of the global cotton area of ∼35 million hectares, with 50% of Bt cotton area concentrated largely in India (ISAAA, 2019). As proponents Qiao and Huang (2019) explain, Bt cotton can control targeted bollworm pests, thus requiring less pesticide application to the crop. In turn, this also reduces labour costs on the farm dedicated to spraying pesticides and reduced crop losses, increasing yields overall relative to non-Bt cotton despite higher seed costs (Herring & Rao, 2012; Kathage & Qaim, 2012). Supporters also predict that faster pest growth in response to a warming climate also makes Bt cotton better served to control pests due to its in built technical traits compared to non-Bt cotton (Huang & Hao, 2020).

Critiquing these deterministic outcomes, Nightingale et al. (2019) find that GM supporters present a scenario where climate change will adversely affect smallholder livelihoods unless a specific technology is adopted through markets. Biotechnology is a new iteration of a technological fix within the linear agricultural development paradigm that is designed to overcome these environmental externalities, packaged as the solution to a ‘natural’ poverty state of the subsistence farmer in dryland settings which will be worsened under climate change (discussed by Gunderson et al., 2020).

As reviewed earlier, proponents argue that Bt cotton is better able to control pest populations under climate change. Efficiency improvements within one area however can be offset by dependence on a different resource. For instance, Bt cotton adoption has been shown initially to reduce pesticide usage relative to non-Bt cotton. Over a long-term period of 15–20 years, however, there are emerging signs in China and India that with rising temperatures and rainfall variability, the crop is unable to deal with new secondary pests and bollworms returning that are now resistant to Bt cotton, increasing pesticide and fertilizer use to offset diminishing productivity returns (Kranthi & Stone, 2020; Najork et al., 2022; Zhang et al., 2018). The ephemeral nature of fixes therefore entwines farmers within a constant loop of increasing dependence on new technologies that must be adopted to solve the side effects of previous technologies (Schnurr et al., 2022).

In Burkina Faso, Luna and Dowd-Uribe (2020) also show that smallholder farmers in day-to-day cultivation faced numerous and simultaneous constraints in matching the optimal growing conditions for Bt cotton. These included labour availability, unreliable rainfall, scattered landholdings and differentiated access to pesticides, fertilizer and oxen, affecting harvest losses and indebtedness (Luna and Dowd-Uribe, 2020). The reliance on the monsoon for example in semi-arid South Asia is important due to 70% of the variable and low annual rainfall falling in the four-month growing season between June and September. From this rainfall, only 30%–40% is utilized by crops, with the rest lost to evapotranspiration and surface runoff (Kuchimanchi et al., 2019). Planting too early before the monsoon arrives or leaving the sowing and weeding of land too late into the growing season present farmers with an extremely small margin of error to successfully harvest a cash crop (Taylor & Bhasme, 2021). As Bharucha (2019) describes, the determinant of a good season can be down to missing rain in one part of the village by a few days.

Climatic factors such as increasing temperatures, extremes in rainfall events such as droughts or floods and timing of rainfall are all predicted to further reduce agricultural productivity through yield losses and crop failures (Singh et al., 2019). Climate models in South Asia forecast increased seasonal variations of monsoon onset dates and timing, and frequency of drought and flood extreme rainfall events through increasing temperatures under climate change (Kuchimanchi et al., 2019; Wang et al., 2021). This has the effect of climate change amplifying economic risk with a high input and capital-intensive crop as Bt cotton on top of already unpredictable set of growing constraints of rainfall, soils and temperatures in dryland areas.

To not follow the same environmentally deterministic trajectory of analysis as GM proponents, Nightingale et al. (2019) reassert the fact that embedded dualities of nature and society remake each other under climate change and agricultural development, contingent on contextual political economy conditions. The next section of the literature review explains the role of markets as the primary contractual mechanism through which farmers adopt GM technologies, mediated through debt and knowledge relations with market intermediaries such as input dealers. Following this, I will identify the gaps in the literature this paper will address.

2.1. Markets and debt relations in commodity agriculture

A central economic mechanism that has often enabled the adoption of GM technological fixes is market integration, with a key role for intermediaries. Market intermediaries in agriculture refer to differentiated group of classes in rural peripheries acting as key centres of localized power in their economic relations with farmers, acting as extension advisors, creditors and output markets (Aga, 2019; Harriss-White, 2018). The role of market intermediaries in agriculture have grown in the global South with the advent of economic liberalization as part of capitalist development since the 1980s (Oya, 2011). As Bernards (2021) explains, policies such as scaling back public extension supports, input subsidies, price supports and credit for crops brought private, decentralized market mechanisms to the fore to fill these roles.

The interlocked factor market is a particular market form to invest in commercial crops for smallholders. Input dealers often disburse credit (with interest) to farmers to finance crop investments and in exchange, farmers sell their harvest at a fixed price back to repay debts (Ramprasad, 2019). Credit is the lubricant for sustaining agricultural investments as smallholder farmers have limited collateral land and cash to afford inputs up front. For supporters of interlocked markets, farmers benefit from buying inputs without collateral or cash, a fixed price for harvests if market prices crash and reduced barriers to entry for smallholder farmers lacking bargaining power with traders in markets (Jaleta et al., 2009; Pal, 2015). Likewise, intermediaries benefit from the contractual arrangement through a predictable harvest income. Markets are therefore seen as neutral institutions for efficiently allocating resources through profit oriented mutual exchange and giving freedom of individual choice between buyers and sellers (discussed in Jan & Harriss-White, 2012).

Critical scholars on the other hand argue that apolitically modelled market outcomes miss the complexity of how markets actually exist and the differentiated power relations laden within them (Taylor & Bhasme, 2021; Vasavi, 2019). For instance, the underlying reasons for smallholder farmers to be integrated into markets and grow cash crops are itself complex.

Luna (2020) provides a useful overview of the economic compulsion for smallholder farmers to use agricultural technologies such as Bt cotton as a response to debt and livelihood survival. Termed a ‘treadmill’ or a ‘reproductive squeeze’ by agrarian scholars (Stone & Flachs, 2018; Watts, 1983), agricultural technologies become increasingly adopted by smallholder farmers for growing cash needs for the household for food, fuel, education, healthcare (Louis, 2015). Under constrained conditions where farmers have few other choices to earn cash for the household, cash crops continue to be grown despite cumulative problems from crop adoption itself, such as rising input costs, indebtedness and fluctuations in market prices relative to incomes. In this way, farmers are kept on a treadmill to survive to cope with additional losses and debt repayment from technology adoption as well as rising household expenses, as these outpace their household incomes (Taylor, 2013).

Furthermore, literature on market intermediaries and Bt cotton in India has focused on the hybrid properties of the crop. Hybrid seeds have a yield advantage to non-hybrids, but require higher fertilizer, labour, irrigation and pesticide usage as well as being designed to be planted at lower densities (Siddiqui, 2020). Hybrid cotton seeds only responded to heavy fertilization and intensification of the same piece of land, as well as being more prone to predation by pests than indigenous varieties (Stone & Flachs, 2018). With Bt cotton, seeds need to be purchased every year, providing a guaranteed a continuous market for private firms, as well as profit from interest on credit they use to sell to farmers in interlocked markets.

Profit incentives by input dealers also distort marketing and extension advice to farmers. Chaudhuri (2019) for example reveals that seed dealers in India made between 10% and 15% profit margin on each Bt cotton seed packet, incentivizing them to market and sell Bt cotton inputs on credit directly to farmers in thousands of towns and village peripheries across India (Bownas, 2016). Input dealers also over prescribed fertilizer and pesticide use to farmers, which over time became greater due to the diminishing returns of Bt cotton in light of secondary pest outbreaks or pest resistance, keeping farmers on a treadmill of increasing input costs and decreasing incomes from the crop (Stone & Flachs, 2018). In the advent of withdrawal of state extension supports, the input dealer acted as the sole source of cultivation knowledge, erasing farmer knowledge of crop cultivation (Flachs, 2019; Stone et al., 2014).

In this paper, I address a gap in the literature on Bt cotton in India through examining the conjugated processes of how the uncertainties of the monsoon in semi-arid areas where the crop has been adopted combine with climatic variability in rainfall patterns to exacerbate debt relations within markets which finance crop investments. I build on the ‘treadmill’ concept by Stone and Flachs (2018) and apply it to indebtedness to show how continuing to grow the so called climate fix Bt cotton in rainfed settings to repay existing debts in fact keeps farmers on the treadmill, beholden to usurious financial arrangements as debts rise. This, I argue, undermines the narrative of Bt cotton thriving under climate change, as the heavy investments required for the crop cultivation place smallholder farmers into situation of higher economic risk in need of consistent returns. This means farmers are actually less able to weather bad years.

The predominant pattern of growing Bt cotton under rainfed conditions, on top of climatic effects on rainfall patterns makes the chances of yield stability more difficult over time, as does recouping crop investments for market intermediaries, let alone enjoying a profit from cultivating Bt cotton. Failure to honour interest repayments or input investments by smallholder farmer due to a bad harvest from drought may leave them within engrained debt relations with intermediaries. They often face higher interest repayments due in exchange for their lack of collateral landholdings, whereas larger farmers can secure lower interest, formal sources of credit and benefit from more flexible repayment arrangements.

I critique GM technological fix narratives of success as existing within narrowly defined optimal conditions that ignore the power relations within markets and agroecological realities of adoption for smallholder farmers. The key question to ask therefore is not whether Bt cotton helps farmers adapt to climate change but to understand the nuances of: ‘who controls the technologies and how they are inserted in a social and economic system and regime of accumulation’ (Borras et al., 2021, p. 11).

In the next section, I introduce the methods and case study research in rural Telangana to understand the long-term adoption consequences of Bt cotton through the lenses of climate and debt.

3. Methods and case study

The paper utilizes a mixed methods case study based ethnographic approach consisting of 94 semi structured interviews and 151 household surveys (98% village sample size) over eight months in 2019 in the south Indian state of Telangana. Telangana has a semi-arid climate characterized by unpredictable monsoons and frequent droughts, and red laterite soils that have low fertility, storage, infiltration capacity and high runoff rates (Kuchimanchi et al., 2019).

Originally a region within the state of Andhra Pradesh, Telangana was granted independent statehood in 2014 (Benbabaali, 2016). The liberalization of seed and credit in the late 1980s (Vakulabharanam, 2005) and the explicit promotion of biotechnology by the Chief Minister Chandra Babu Naidu and the Indian government in the 2000s (Aayog, 2015; Bowles, 2003) brought a private sector driven commercialization of smallholder farming via Bt cotton. The cultivation area rose from averaging ∼200,000 hectares between 1955 and 1985 to 2.3 million hectares in 2021 (Directorate of Economics and Statistics Telangana, 2022). Ninety-five per cent of cotton cultivated in India by 2014 was Bt cotton across 10 major growing states, after its approval in 2002 (Kranthi & Stone, 2020).

The justification for choosing Telangana was that it is the second biggest Bt cotton producer in India out of 10 major Bt cotton growing states (Directorate of Economics and Statistics Telangana, 2022). Approximately 13% of Bt cotton grown in Telangana is irrigated compared to the all India average of 33%, a figure that has stayed similar for non-Bt and Bt cotton since 1985 (Directorate of Economics and Statistics India, 2020). This is a higher proportion than Maharashtra’s 3%, the third biggest cotton grower in India. Together, Telangana and Maharashtra produce 40% of India’s Bt cotton and cover 50% of its area, so the choice of studying Bt cotton in rainfed conditions in Telangana reflects how much of Bt cotton is grown in India (Directorate of Economics and Statistics India, 2020).

The main fieldwork was conducted with farmers in the village of Kavarampur,¹ Nalgonda district.

Kavarampur typified the semi-arid soils and climate in Telangana. The village received 80% of total annual rainfall in the monsoon (June to September) of approximately 460 mm (TSDPS Telangana, 2021).

The village contained 155 households and was populated exclusively by Other Backward Castes (OBC), Scheduled Tribes (ST) and Scheduled Caste (SC) groups, who make up almost 90% of Telangana’s population (Benbabaali, 2016). Historically, these castes had lower average landholding sizes, literacy rates and rural household incomes relative to higher castes (Kuchimanchi et al., 2019). 85% of the farming households in Kavarampur owned less than 2 ha, like Telangana, where the figure is 88% (DES Telangana, 2021). The dominant caste in the village were OBCs, with the village leader and the largest landholder in the village (5 ha) being from the caste. Bt cotton farming was a primary occupation in the village during the main monsoon growing season from June to October.

In Bt cotton cultivation, males tilled and cleared the soil of stones or debris using oxen or tractors, purchased inputs, sought extension advice from input shops, spread seeds before cultivation and sprayed pesticides. Women handpicked cotton from the plant buds at the end of the harvest season, ploughed fields using oxen and removed weeds in fields prior to and during the Bt cotton cultivation. Work in the dry season consisted of migrant agricultural labour within Telangana, construction labour in nearby towns Dittam and Hyderabad city, MNREGA work and a small minority of caste occupations such as OBC Goud and Yadav sub-castes who sold palm wine and herded sheep and goats for milk and for sale.

Figures 1 and 2 provide maps of Telangana and Kavarampur.

Figure 1. Telangana in India.

Figure 2. Kavarampur Map. Source: Wikipedia (2006, 2018).

Would like to replace Figure reference with the following: Wikipedia. (2018). Nalgonda in Telangana. https://commons.wikimedia.org/wiki/File:Nalgonda_in_Telangana_%28India%29.svg.

Table 1 below summarizes the key characteristics of Kavarampur.

Table 1. Kavarampur profile.

Category	Value
Population	710
Number of households	155
Average annual household income ($)	1,950
Cropping patterns	100% grow Bt cotton
Per capita income per day ($)	1.20
Average landholding size (ha)	1.2
Farming Households	105

Source: Author (2019).

To select Kavarampur, I utilized a snowball sampling approach by contacting rural NGOs and government agricultural extension offices in Hyderabad, the capital city of Telangana. This enabled me to recruit a local NGO research assistant in Nalgonda district. We scoped for villages over a two-week period based on areas the NGO worked in and villages which roughly corroborated demographics of castes, landholding size, Bt cotton cultivation and village population size in rural Telangana using Nalgonda district census figures. In Kavarampur, the village head acted as the gatekeeper for the research process, holding a meeting with village participants before I began the research to clarify the process and purpose of the research.

Qualitative semi-structured interviews enabled a nuanced understanding of farmer’s everyday experiences and the processes behind agrarian change from growing Bt cotton in the village. Quantitative household surveys of 98% of the households allowed me to use the interviews to speak to representative trends at the village level, so for example perceptions of rainfall, Bt cotton outcomes and debt were not anomalies. Most of the research took place within participants’ houses in the village, but I also conducted field visits over the course of the eight months to inform my questions on the Bt cotton growing season and rainfall patterns. In addition, I conducted supplementary interviews with input dealers in nearby towns to Kavarampur to understand the context of when and how Bt cotton was cultivated, the processes within interlocked factor markets and to gain their responses to farmer’s struggles with the crop.

Prior to conducting research, I recorded oral consent from farmers in Telugu, as adult literacy in the village was only 19%. The interviews were roughly split between men and women within households and proportionally according to landholding sizes and caste populations in the village. I asked questions on the history of Bt cotton in the village, motivations for growing the crop, farmer knowledge on Bt cotton cultivation, rainfall patterns, debts and risks from Bt cotton cultivation, long term perspectives on yields and incomes and farmer relationships with input dealers. I spoke to the farmers in their native language Telugu.

Male and female local research assistants recruited from within the village recorded the interview on a dictaphone while I asked the questions and wrote notes in a paper handbook. Each interview participant was given a pseudonym based on local male or female names in Telangana. Interviews were transliterated and transcribed into English using a transcription software. The interviews were coded and analysed thematically based on literature on monsoon variability, semi-arid climates, Bt cotton cultivation in rainfed conditions and debt and risk from Bt cotton within market relations with input dealers. These built on gaps identified in the literature on the promotion of Bt cotton as a cash crop for climate change but the crop primarily being grown in an already unpredictable semi-arid agroclimate in India where monsoon variability is embedded with high input costs and economic risk from the debt relations with input dealers.

Household surveys were in paper form and the questions were written in both Telugu and English, but the questions were asked in Telugu. The household surveys asked questions on farmer recall data on the baseline year 2018 on rainfall onset, quantity and variability for 1, 5 and 10 year recall period, to specify when they had good and bad harvest years for Bt cotton and how these trends changed over time. In addition, I asked questions on input costs, yields, indebtedness, cropping patterns. The answers were recorded in English. I manually uploaded the household survey data each day onto an Excel spreadsheet that I created which matched the question formatting of the household surveys. I used Excel to generate descriptive statistical tables to present household incomes from Bt cotton for the recall year 2018.

My positionality in this research is as a male, urban, middle class and upper caste Indian origin researcher. I was born in India in Hyderabad, the capital of Telangana and can speak the local language Telugu. Despite being a naturalized UK citizen, I frequently visit India to visit extended family and maintained Telugu proficiency, which engendered trust amongst the villagers, as well as informal and formal repeated interactions over eight months of the research. The local research assistants were also instrumental in aiding the research process due to familiarity with villagers and clarifying the questions as I asked them when required, e.g. using simpler vernacular terms to ask about Bt cotton yields, debts and monsoon variability and recall of historical Bt cotton cultivation in the village.

The next section empirically analyses Bt cotton technological fixes in Telangana.

4. Investigating Bt cotton technological fix narratives in Telangana

4.1. Rapid adoption of Bt cotton, persistent rainfed crop failures under climate change

Cotton had long been grown in the Telangana region under rainfed conditions using indigenous cotton seeds, but averaged between 2% and 5% of the total cultivable area between 1955 and 1985. Valukabharanam (2005) notes that cotton adoption took off in the late 1980s and 1990s in Telangana, spurred by factors such as a 1988 Indian government policy to partially liberalize seed imports, as well as withdrawal of public sector credit support in the 1990s. The emergence of private sector input dealers promoting hybrid cotton on informal credit further helped increase cotton adoption in the 1990s even before Bt cotton’s approval in 2002. By 2001, cotton adoption had already trebled to 14% of the total cultivable area.

Cotton in Kavarampur was grown by 10–20 households prior to Bt cotton, the first by Govinda in 1993, an elderly OBC farmer with 2 ha, the dominant caste in the village. He purchased hybrid cotton seeds after working as a security guard in Hyderabad for agricultural extension officers, learning about how the crop was grown (Govinda, farmer, 4/4/19).

Bt cotton entered the village in 2004/5. In surrounding town peripheries, input shops proliferated in the 1990s and early 2000s, owned primarily by upper caste Reddys and OBCs for new income opportunities (Venkat, input dealer, 29/4/19). They began to market and sell Bt cotton on informal credit to farmers, who could finance the crop despite little collateral and landholdings of under 1 ha. Farmers followed and observed initial successful harvests by lead farmers such as Govinda, which precipitated Bt cotton adoption via imitation across caste and classes (Maertens, 2017; Stone et al., 2014). By 2019, Bt cotton was adopted by 100% of the 105 farming households in the village. Within five years of adoption however, farmers in Kavarampur began noting the increasing prevalence of crop failures, which according to them stemmed from monsoon variability.

For Chandrasekhar, an elderly SC farmer, planting hybrid cotton was more successful in the late 1990s than Bt cotton today. He explained:

[With hybrid cotton], the market price compared to inputs was high and the rainfall in that time was good. [Now], Bt cotton needs perfect conditions for it to grow well. (Chandrasekhar, farmer, 22/4/19)

This account corroborated with village wide household surveys farmers across all social categories struggled to benefit from growing Bt cotton.

Marginal farmers, owning an average of 0.57 ha, were often the worst hit in per hectare terms due to their low collateral and inability to borrow at low interest rates. The majority of marginal farmers were from SC castes, historically the bottom of the caste system and facing amongst the highest rural poverty rates, lowest education rates and smallest landholdings (Guérin et al., 2013).

Eshwar, an elderly SC farmer with 1 ha, lamented his persistent losses from Bt cotton, observing that:

It hasn’t been a good time for a long time … Since the last 10 years it has not been profitable. For the last five years at least especially the situation has been bad … There is nothing to do now. (Eshwar, farmer, 17/6/19)

Lower caste, classes of farmers copied early adopters with Bt cotton adoption, but faced the highest economic risk as they had least collateral and landholdings to finance the crop, so paid higher interest repayment on input investments than higher castes. Lerche and Shah (2018) refer to this constellation of caste and class based differentiation as conjugated oppression, whereby the individual categorical deprivations become entwined in entrenched poverty outcomes for the lowest class SC and ST groups.

In her survey of cotton, Bt cotton production also amongst SC farmers in Karnataka, Ramamurthy (2011) finds the dual edged opportunities and crisis that cotton capitalism brought. On the one hand, for SC families (especially women and children) who previously worked as cotton labourers for upper castes, the high market prices of cotton relative to subsistence crops allowed SC families to grow cotton rainfed from small parcels of land to achieve ‘class leaping returns’. In the 2006 survey year, however, cotton pushed most families into indebtedness as they piled up fertilizer, seed and pesticide investments on credit with input firms, leaving them better off to not have cultivated cotton and earned daily wage labour.

Farmers felt that monsoons were worsening with increasing frequency, leaving their Bt cotton investments to become unrealized and turn into indebtedness from the crop. Additionally, long-term monsoon trends in Nalgonda district and Telangana state corroborate long-term changes in monsoon rainfall in Kavarampur, due to a lack of village level rainfall data. Using local perceptions of rainfall with aggregated climate data is important as ‘only farmers who perceive a problem will implement strategies to adapt or respond to it’ (Simelton et al., 2013, p. 124). Figures 3 and 4 below shows the monsoon trends for Nalgonda district and Telangana. In both cases, the long-term trend lines slope downwards between 1951 and 2019, supporting qualitative farmer insights of a declining monsoon over time in the village and its effects on Bt cotton outcomes.

Figure 3. Telangana Monsoon Rainfall 1951–2019. Source: DES Andhra Pradesh (2006); TSDPS Telangana (2021).

Figure 4. Nalgonda Monsoon Rainfall 1951–2019. Source: DES Andhra Pradesh (2006); TSDPS Telangana (2021).

A technological fix is inevitably unequally differentiated in terms of access to the benefits of the adaptation intervention (Taylor, 2013). Bt cotton was a short-term fix as it was unable to sustain yields in a climatically extreme setting. It entered dryland environments through the lure of cheap credit and high market prices from intermediaries. Glover (2010) notes how Bt cotton’s initial yield successes in India were premised on controlled, irrigated trials in field sites, from which the crop was marketed as a success compared to non-Bt cotton varieties. In reality, in the actual dryland environments where it was eventually adopted, these gains could not be realized. As Gutierrez et al (2020) show, India has adopted a particular long season, low-density hybrid Bt cotton type which lends itself to low yields and low planting densities, due to the prohibitive input costs. In low rainfall environments such as in Maharashtra and Telangana where Bt cotton is grown predominantly, seed, fertilizer and pesticide costs have risen 2–3× since its introduction to India in 2002 (Gutierrez et al., 2020). In the face variability in rainfall and risky and high interest debt financing to pay for inputs, it has made yield instability a major problem for smallholder farmers, who face severe economic distress when weather fails, undermining its utility as a climate fix.

In addition to the quantity of rainfall over time, the timing of rainfall was also crucial for farmers. Bt cotton is a very fickle crop to grow in rainfed conditions, requiring at least 500 mm of rainfall and application of pesticides and fertilizer at various stages of the four-month growing cycle between monsoon onset in June and September. In 2010-2011, per acre cultivation costs of Bt cotton were 10× that of non-Bt cotton, primarily due to high seed costs (Dhanagare, 2016). Each Bt cotton harvest involved large investments to be recouped as a minimum to make a profit and for the gamble to pay off. Too much rainfall early in the growing cycle or during flowering of cotton buds close to harvest can reduce crop yield (Sharma et al., 2018). In the fieldwork year of 2019 for example, it rained 460 mm in the monsoon as well as the monsoon onset not arriving as late as July, so farmers did not plant cotton at all. In an interview with Latika, an elderly OBC farmer with 1 ha, she revealed that in 2019 too the Bt cotton crop investments failed for the second consecutive year, stating:

We haven’t even started planting this year the rains are so late … Bt cotton has pests or it dries up … We had so much reliable and predictable rain back then versus now. (Latika, farmer, 18/6/19)

In this erratic rainfall setting, Bt cotton as a technology actually perpetuated losses by farmers intensifying in a semi-arid climate (Taylor, 2018). An econometric model by Guntukula and Goyari (2020) shows that 76% of Bt cotton yields are determined by climate variables, in particular rainfall and temperatures, so cotton yields decrease with lower rainfall and higher temperatures.

Under climate change, 40% of the global Bt cotton area may see shorter growing seasons and droughts (The Express Tribune, 2021). Relative to other crops such as cereals, lentils and oilseeds at 30%, Bt cotton yields suffer up to 60% yield loss from climate related factors (Sankaranarayanan et al., 2020). In India, rainfall variability and higher temperatures are linked to Bt cotton production decline of 15%–25% by 2050 (Sankaranarayanan et al., 2020). This makes Bt cotton cultivation an even riskier economic prospect in the future than present in Kavarampur. Bt cotton outcomes in Kavarampur over time have seen fleeting successes followed by persistent failures through the existing semi-arid and rainfed conditions in which it was adopted in, undermining the technological fix to adapt to climate change.

Matthan (2022) underscores the notion that smallholder agriculture under climate uncertainty is buffeted between market risk and climate risk, each negatively intersecting with one another. High investment costs for cotton seeds, pesticides and fertilizer are financed by informal credit market contracts, and the returns from the crop are mediated by adequate rainfall through the growing cycle. The greater the variability in climate extremes of drought or erratic rainfall, the worse off it makes farmers with regards to repaying their debts and keeping the household afloat economically. For the Bt cotton technical fix to function, it needs high investment costs financed through these prevalent high interest rate arrangements in interlocked debt relations. Yet in bad years, not only does a poor harvest engrain farmers deeper within debt relationships, but undermines the usefulness of Bt cotton as a crop in being unable to adapt to the fluctuations in climate.

The next section outlines how crop failures from rainfed Bt cotton not only undermine its usefulness in adapting to climate change, but are also enmeshed within the central role of debt relations necessary for smallholder farmers to finance the crop in the first place. Bt cotton adoption perpetuates a model of debt relations within industrial agriculture where seeds and inputs were controlled by private firms to benefit their profit maximization, not farmer prosperity (Gras & Cáceres, 2020). In combination with the climatically extreme environments the crop is grown in, the high capital intensive and credit dependent investments required for Bt cotton makes farmers less able to adapt to climate change due to indebtedness arising in bad years.

4.2. Indebtedness treadmills and dried up Bt cotton

Most farmers in Kavarampur were unable to afford Bt cotton investments, so purchased seeds, fertilizer, and pesticides on credit from input dealers in nearby towns such as Dittam. At the end of the harvest season, farmers sold their cotton harvest at fixed price to the input dealer plus 24%–36% annual interest on top of the repayment. Profits made net of these costs could be used for household consumption, reinvestment, or savings. In recent years however, Bt cotton crop harvests increasingly failed because of erratic monsoons.

Farmers were now compelled to reinvest in Bt cotton again out of few other cash earning options to repay their debt and interest obligations under droughts, low fertility soils or extreme rainfall, increasing their vulnerability to climate change through the risk of further debt from crop failures. The process of rural indebtedness in this way created a self-reinforcing indebtedness treadmill, where the farmer fell into further debt each time they had a harvest failure, but needed to rely on further investments on Bt cotton out of desperation, creating a dependency on commodification (Taylor, 2013). Farmers had loan and interest obligations to repay each year and so often doubled down on Bt cotton investments to earn cash in the short term for the household consumption needs and avoid defaulting on their debts or being forced to sell off assets such as land. A longing memory of successful Bt cotton harvests in years gone by became a mirage to chase for farmers to alleviate their problems. The sunk cost fallacy of Bt cotton investments has proven to be penurious in the long term, however, for farmers in Kavarampur, as it only worsened their indebtedness levels and kept them on the treadmill, due to the unviability of the growing conditions in Telangana for it to consistently provide positive incomes net of rising input costs (Gutierrez et al., 2020).

For farmers, yields were insufficient relative to the input expenses they invested, making them indebted from Bt cotton. If the farmer profits, then the input costs and interest are repaid, and the farmer takes a new loan out for the coming season. If the farmer faced a loss, input dealers offered to renew current loans even if the principle was untouched, if the interest repayments were on time. However, this only delays the debt repayment as interest accumulates into the next harvest year, concentrating the risk into one crop and increased indebtedness amongst all farmers (Kandikuppa, 2021). Progressively, farmers were even unable to repay the interest, which brought frequent reprisals from input dealers in the form of public shame and threats of violence. Jagannath, an elderly SC farmer was unable to pay his interest for the past two years due to Bt cotton failures, and received frequent visits from input shop employees, stating:

He will come and scold you he will humiliate you and remove your dignity that’s what happens. (Jagannath, farmer, 17/6/19)

Informal debt for Jagganath came as a necessity to be able to renew loans to keep the household afloat but was regarded as unsafe due to the repercussions one could face for failure to repay interest. This was in addition to the fact that the debts could not be waived like bank loans could (Kandikuppa, 2021). For SC families such as Jagganaths too, with low landholdings and little collateral, bank loans were difficult to access (Guérin et al., 2013). Farmers were willing to tread water and pay off the interest year on year even though the interest rates charged were much higher at up to 36% per year compared to a maximum of 24% for formal sources. Cumulative debts from Bt cotton amounted to an average of $2200 per household across caste and class in Kavarampur, with an additional $528–792 for annual interest repayments of 24%–36% on top of the debts. For the creditors, it was in their interest to sustain credit flows and maintain interest repayments rather than foreclose on farmer collateral for failing to repay (Gerber, 2013). The maintenance of a large customer base repaying interest was preferred to losing customers, as Jayant, a middle-aged OBC farmer who last made a profit from Bt cotton three years ago explained:

We pay the interest we don’t have the money to pay the principle. That is what the brokers want too, they don’t want you to pay the principle they want you to pay the interest. (Jayant, farmer, 28/6/19)

For farmers, input dealers’ flexibility in offering new loans without repayment allowed them to borrow to cover harvest failures. Farmers were willing to tread water and pay off the interest year on year even though the interest rates charged were much higher at up to 36% per year compared to 24% for formal sources. The more farmers proceed with agricultural intensification, droughts, low fertility soils or extreme rainfall may lead to failed harvests, so the farmer has nothing to pay back to the intermediary at the end of the season (Taylor, 2013). This keeps them indebted to intermediaries and can mutually exacerbate environmental degradation the more they intensify (Gerber, 2013). This compels farmers to deepen their intensification to repay their rising debts, creating a self-exploiting cycle of debt relations and environmental degradation. Yet, when questioned, intermediaries in nearby towns to Kavarampur deflected their role in adverse Bt cotton outcomes for farmers. An exchange with a local degree educated Reddy input dealer Akash in the nearby town of Dittam revealed:

The farmer [comes to me] four to five times. Once for seeds, twice for fertiliser and twice for pesticides … there is no such thing as a loss because of the cotton it is due to rain … it is according to the rain the last several years have not been good they have been minus years. (Akash, input dealer, 4/6/19)

For the input dealer, farmers’ losses from Bt cotton were due to the variability of the rain, and not anything wrong on their side. They in fact made money despite the yields that farmers achieved. All this time, input dealers keep renewing loans and business for another year, getting repeat investments in Bt cotton while the risks of doing so in a rainfed environment only created uncertainty and further losses from cultivating the crop. Loan renewals were a guaranteed repeat business in Bt cotton. Regardless of how farmers fared with Bt cotton, input dealers received a minimum of ∼$500 in annual interest repayments per household from over 100 households in the village just in Kavarampur. Indebtedness in Kavarampur reflects the rural indebtedness overall in Telangana. Telangana has among the highest rural indebtedness rates in India at 74%, compared to an all-India average of 34% (NSSO, 2021). This has been linked to a large extent to cash crop cultivation and intensification such as Bt cotton by smallholders who adopted it rapidly in recent decades (Kranthi & Stone, 2020).

The short-term credit dispersal for inputs enabled farmers to access the seeds and inputs to cultivate Bt cotton, but farmers pinned their debt repayment hopes on its high yields. The risks of doing so in a rainfed environment only created uncertainty and further losses from cultivating the crop for farmers. The culmination of existing agroecological conditions, climate variability and indebtedness to market intermediaries made Bt cotton a failed technological fix in the long run in Kavarampur. The plight caused by Bt cotton was summarized by Thirumala, a young OBC farmer with 1.6 ha land who took losses of $583 from Bt cotton in 2018. He was desperate to keep providing for his young family given the repeated failures from Bt cotton investments and debt repayments it caused, stating:

Farming doesn’t bring much return it falls on our heads … in this time whatever work we are putting in we are not realising the results. For me [Bt cotton] is enough to pay the input costs and my children’s costs – that’s it  …  what can we do? (Thirumala, farmer, 5/4/19)

GM fixes overlook the debt relations that perpetuate climate vulnerability and poverty in dryland settings (Taylor & Bhasme, 2021). They are packaged into an input and capital-intensive monocrop market commodity for the benefit of corporate capital. Farmers in Kavarampur, betrayed by Bt cotton after fifteen years of experimentation, could no longer depend on empty promises from GM technological fixes. Instead of increasing incomes through markets, Bt cotton in a rainfed setting forced farmers into losses induced from a technology that was supposed to thrive under climate change. Technological fixes through market delivery deepened indebtedness treadmills through interlocked factor markets via the already existing economic precarity of the rainfed setting Bt cotton was grown in, to worsen outcomes for farmers to adapt to climate change.

5. Conclusion

GM technologies in agriculture aimed at thriving under climate change and increasing productivity have unquestioningly grounded technological fixes as the prime focus for agricultural policy solutions. In this paper, I extended Nightingale et al.’s (2019) critique of technical fixes, by arguing that pre-existing environmental conditions and market relations a GM technology is adopted in undermines the technological fix’s capacity to adapt to climate change. GM narratives disregard the challenges of the suitability of GM technologies for diverse socio-economic and geographic conditions that farmers face under climate change.

In this paper, I have shown that Bt cotton outcomes in Telangana have not exhibited linear progress via its pest and climate resilient properties as corporate technological narratives predicted, but a chaotic, boom and bust cycle. Farmers across caste and class lines were trapped into growing an input intensive cash crop prone to the vagaries of a declining monsoon, affecting SCs and marginal farmers the most. The draw of high yields and profits from Bt cotton prompted farmers to delve knee deep into risky high interest debt from input dealers to fund their investments. In turn, rainfed Bt cotton’s failures only made the risk more severe and plunged farmers into heavy losses and debts, but farmers out of no alternatives doubled down on Bt cotton in a recurring treadmill to unsuccessfully recoup losses.

To exacerbate the problem of Bt cotton failure, profit incentives of intermediaries in interlocked markets reinforced economic distress. These were encapsulated in indebtedness treadmills with farmers by renewing loans annually and accumulating interest repayments from crop failures in an unreliable rainfed semi-arid setting. Losses for many farmers in the longer term pushed them to struggle to make ends meet in Kavarampur despite GM technological fix driven agricultural commercialization. Losses from Bt cotton in Kavarampur under a changing climate left farmers in a worse position than prior to planting the crop with regards to trenchant indebtedness problems afflicting rural livelihoods.

Constraining the remit of responses to climate change within a narrow technical frame depoliticizes and conceals alternatives to the industrial model, maintaining power with private firms and their representation of the solution to climate change (Clapp et al., 2018). In doing so, technological fix narratives ignore the role of climate change, framing it as an external threat to be dealt with rather than recognizing the commodification in farming as a causal factor worsening the ability to respond to climate change in the first place.

Acknowledgements

This article is dedicated to the people of Kavarampur village without whom this research would not be possible. Equally to Yadagiri for providing excellent research assistance. I’m grateful to my PhD supervisors Admos Chimhowu, Ben Neimark and mentor Phil Woodhouse for helping shape the initial PhD chapter. Thanks also to Marcus Taylor and Tom Lavers for substantial improvements to the PhD thesis after the viva. To Alex Loftus and Nithya Natarajan, I have been privileged to have your critical comments and support which greatly shaped this work. Finally, thank you to the three anonymous reviewers who were meticulous in their suggestions and improvements to the paper. All errors are mine.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This article draws on PhD research funded by the Economic and Social Research Council, Award Reference: 1774620.

Notes on contributors

Ambarish Karamchedu

Dr. Ambarish Karamchedu is a critical development studies scholar whose work focuses on India and explores the themes of risk, debt and technology adoption within agrarian and climatic change in rural areas. He is interested in the political economy of GMcrops, the hydropolitics of groundwater irrigation infrastructures and industrial livestock production.

Notes

1 Pseudonym.