On risk and regulation: Bt crops in India

Abstract

Genetic engineering in agriculture raises contentious politics unknown in other applications of molecular technology. Controversy originated and persists for inter-related reasons; these are not primarily, as frequently assumed, differences over scientific findings, but rather about the relationship of science to ‘risk.’ First, there are inevitably differences in how to interpret ‘risk’ in situations in which there are no established findings of specific hazard; ‘Knightian uncertainty’ defines this condition. Science claims no method of resolution in such cases of uncertainty. Second, science has no claim about risk preferences in a normative sense. In genetic engineering, Knightian uncertainty is pervasive; declaring uncertainty to constitute ‘risk’ enables a precautionary politics in which no conceivable evidence from science can confirm absence of risk. This is the logic of the precautionary state. The logic of the developmental state is quite different: uncertainty is treated as an inevitable component of change, and therefore a logic of acceptable uncertainty, parallel to acceptable risk of the sort deployed in cost-benefit analysis in other spheres of behavior, dominates policy. India's official position on agricultural biotechnology has been promotional, as expected from a developmental state, but regulation of Bt crops has rested in a section of the state operating more on precautionary than developmental logic. As a result, notwithstanding the developmental success of Bt cotton, Bt brinjal [eggplant, aubergine] encountered a moratorium on deployment despite approval by the regulatory scientific body designated to assess biosafety.

Abbreviations

GEAC

Genetic Engineering Approval Committee

GMO

genetically modified organism

Risk and Uncertainty

Genetic engineering in agriculture raises contentious politics unknown in other applications of molecular technology. Controversy originated and persists for inter-related reasons. The pivot is commonly held to be risk. In pharmaceuticals, medicine and industrial applications, recombinant DNA technology has been widely accepted as providing useful tools; in agriculture, products using these same tools have been codified as ‘GMOs,’ evoking almost universally an aura of unique risk, worthy of special regulation. Science is evoked by state institutions, social movements and developers of technology as the touchstone for assuring safety in the face of this risk construction: there must be ‘bio-safety’ procedures and assurances. Yet assurances produced by both state science and global consensus of working scientists often prove insufficient to support approval of genetically engineered crops. Opponents attack mainstream science as flawed or instrumental – produced or paid for by interested parties such as corporations – or insufficient for assuring the safety of ‘GMOs.’ Both sides in the global rift over biotechnology thus evoke Science, but science actually has very little to do with regulatory outcomes. Rather, politics hinges on how uncertainty is framed in public policy and by mass politics, and how it is inserted into regulatory logic. The ‘GMO’ is a construct that evokes not only unique biological risk, but also exceptional externalities that broaden the risk palette: risks of dependency of both nations and farmers on corporate control. The ‘risk’ relevant to regulation is both biological and socio-economic.^1-3

The socio-economic arguments premised on unacceptable risk seem refuted by farmers’ voting with their plows for biotechnology: whatever risk there may be has proved an acceptable risk. Farming is inherently risky, and farmers have decided from experience, in large numbers, that the risk-benefit calculus for Bt crops favors adoption. Evidence from India – as well as China – indicates that the framing of Bt crops as inaccessible to and dangerous for small farmers is empirically inaccurate. Most farmers of Bt cotton in Asia, for example, operate less than a few hectares, often with family labor, often using animal traction; there is nothing ‘industrial’ implied by the addition of a gene for insect resistance, nor risk in principle any different from that introduced by other new cultivars, contrary to global critiques. How a trait gets into a crop's germplasm has no necessary connection to how farming is done; even organic farmers in India may use Bt cottonseeds.⁴ Though associated with wealthy economies historically, genetically engineered crops grown in ‘developing countries’ in 2012 exceeded total acres grown in the so-called ‘developed countries’ for the first time.⁵ India was the 16^th country to approve a genetically engineered crop: Bt cotton in 2002. The results were as expected by state support of biotechnology: wide adoption, improved production, yields and profits, escape from a situation of import dependence to one of significant cotton exports, and reduced pesticide applications for cotton's major pest.⁶ Despite persistent media stories, reports of ‘studies’ and globally-distributed films alleging mass suicides driven by Bt cotton, farmers of all size classes and social standing have adopted the technology rapidly, almost universally, and to good effect.⁷ Seed technology has proved to be scale-neutral and useful. Most observers expected this success story to be followed rapidly by other Bt crops in India.

Controversy over India's second Bt crop – brinjal [aubergine, eggplant, Solanum melongena] – however proved to be intense and centered on risk. Bt eggplant carried the same transgene for insect resistance as the first Bt cotton hybrids (Cry1Ac) and faced the same regulatory process. Controversy was not over the socio-economic effects so dramatic in the politics of claiming Bt cotton suicides, but centered almost entirely on the adequacy of risk assessment for regulation.

Risk is fundamental to the global debate on genetically engineered crops, but the concept is used in surprisingly loose and lazy ways, on both sides of the ideational divide. Indeed, what makes risk politics powerful is that ‘risk’ is an elastic and elusive concept in common discourse, arousing anxiety that is inherently difficult to dissipate: our evolutionary history programs us to be cautious. In normal science, in contrast, risk has a precise, but deceptively simple, meaning: risk equals the probability of some hazard. Anyone booking a flight, taking prescription drugs or scheduling surgery recognizes potential hazards. There are real risks, though probabilities are often imprecisely known. We regularly take some risks because of expected benefits, or because the risk of doing nothing is higher. The question is always: compared to what? Ideally, regulation of any technology would reach some threshold of acceptable risk – balanced with benefits – for a whole society.

Conceptually simple, these comparisons are notoriously difficult in practice. Often neither hazard nor probability is known, or cannot be measured. This situation is not one of risk, but of uncertainty. This foundational distinction was established in the 1920s by economist Frank Knight.⁸ Under conditions of uncertainty, risk is of necessity a social construction. The common cell phone is a good example. there is some suggestive but insubstantial evidence of hazard, no proof of hazard and no estimate of probabilities; one might say there is hypothetical risk, but the technology has such obvious utility that any such hypothetical risk is discounted by nearly everyone. Uncertainty in the case of the mobile phone has been widely, but not universally, coded as acceptable risk or no risk at all. Science is not helpful in these categorizations until some procedures establish a hazard and, optimally, a probability distribution of said hazard. Science cannot assess uncertainty, nor determine appropriate risk preferences. At the individual level, these are matters of cognition and personal attitudes toward risk aversion. For a whole society, technology subject to collective risks and externalities – nuclear power, for example – deciding on an acceptable level of risk is of necessity a political decision. Some technologies rise to the level of arousing risk perceptions large enough to provoke legislation, surveillance and control (dioxin in drinking water, for example) and others do not (mobile telephones, for example).⁹ Agricultural biotechnology has provoked a global movement¹⁰ organized around precautionary logic premised entirely on hypothetical risk.

Though rDNA plants have been assigned to a category of special risk, meaningful regulation requires much more specificity. The common standard of acceptable risk would ask of science, as a precondition for costly surveillance and control: what is the probability of hazards in transgenic plants compared to cultivars bred by other means? Here the science is critical. Though there may well be new hazards, as pre-emptive regulatory systems assume, none has been demonstrated in mainstream science to date. The European Commission Directorate-General for Research assessed available regulatory science for environmental and food-safety risks in A Decade of EU-funded GMO Research (2001–2010): “The main conclusion to be drawn from the efforts of more than 130 research projects, covering a period of more than 25 y of research, and involving more than 500 independent research groups, is that biotechnology, and in particular GMOs, are not per se more risky than e.g. conventional plant breeding technologies (p. 16).”¹¹ Many studies support this basic finding, as do conclusions of national academies and other expert bodies.^12,13

Our species is increasingly aware of very real hazards of global warming precisely because the atmospheric science of climate change has produced an institutionalized global consensus through the Intergovernmental Panel on Climate Change (IPCC) – though it is one rejected by some nations, some social groups, and some scientists. Even so, there is a center of gravity around evidence of hazards, within a range of probabilities. There is no comparable institutionalized global consensus on genetic engineering, and no demonstration of hazard; as a result, national implementations of regulatory science embedded in fields of power and administrative structure diverge significantly. These differences in institutions of state science contribute to our understanding of the puzzling variation in acceptance and rejection of rDNA plants over nations and time.¹⁴ State science in India for regulation of Bt crops was located in the Genetic Engineering Approval Committee, which adopted a permissive or promotional policy toward agricultural biotechnology, though its authority to regulate came under an act for environmental protection; consequently, the GEAC reported to the Ministry of Environment and Forests. The structural location of that committee under the Ministry of Environment, during the reign of a specific Minister of Environment, led to a change in India's regulatory stance from promotional to precautionary in the transition from approval of Bt cotton in 2002 to the rejection of Bt eggplant (brinjal) in 2010. Politics in this period likewise led to a restructuring of state science for regulating biotechnology in India, a process still contested in the Parliament.¹⁵

India from Bt cotton to Bt Brinjal

India supported biotechnology for its potential benefits, like China and Brazil, as a project of the developmental state. Nevertheless, state backing for biotechnology did not result in public-sector Bt cotton – as it did in China – but rather in funding research and in evaluating and approving hybrids from the private sector. However, both India and China promoted Bt cotton for similar reasons. Bt cultivars utilized an insecticidal protein derived from a common soil bacterium, Bacillus thuringiensis, hence ‘Bt,’ to control Lepidopteron pests. Results demonstrated benefits consistent with field trials conducted under the auspices of the GEAC: higher harvested yields, better pest control, improved farmer incomes, reduced pesticide spraying on cotton.^16-18 Political mobilization around ‘the failure of Bt cotton,’ conjoined with a farmer suicide narrative, gained international attention, but proved irreconcilable with both farmer behavior and peer-reviewed science.¹⁹ Because the Cry1Ac gene had proved effective and safe in Bt cotton, application to a popular vegetable crop – eggplant, or aubergine, or – more commonly in the subcontinent, brinjal – seemed an obvious step.²⁰ Damage to the crop by Lepidopteron pests was extensive, uncontrolled by conventional pesticides even at great cost to the farmers and some risk to consumers from residues on marketed fruits.

Benefits to farmers and to agriculture did not dominate consideration of India's second transgenic crop, however. Construction of risk in the new Bt crop demonstrated a change in operative politics. Risk discourse around Bt cotton had largely been about agronomic failure – farmer suicides in particular²¹ – though there were peripheral claims of biological externalities such as poisoned livestock as well.²² Both narratives were equally without empirical grounding. Bt eggplant presented a more capacious opening for risk politics because food crops are inherently more susceptible to anxiety framings. The insecticidal protein produced by the cry transgene was widely discussed as a generic ‘toxin’ or ‘poison’ [zeher in Hindi], implying that humans would be subject to the same deadly effects as larvae. No rational person consumes poison.

Assessment of risks and benefits was allocated by statute to the Genetic Engineering Approval Committee [GEAC], the apex body for scientific vetting of genetically engineered crops. After 9 y of tests involving 7 government agencies and departments, the GEAC approved release of Bt brinjal by both the private and public sector developers, based on comparative assessment of options.²³ No hazards from the insecticidal protein were found through standard safety protocols; GEAC findings conformed to the European Union's general conclusions quoted above. Investigation of existing practices of eggplant cultivation in India did, however, uncover hazards to both farmers and consumers: heavy pesticide application – even some pesticides unapproved for food crops. The developmental case for Bt eggplant was then quite similar to that for Bt cotton. The Expert Committee of the GEAC concluded:

… the current practices of using extensive pesticides is not only harmful to the health and environment but also non-sustainable in future for control of FSB [fruit and shoot borer, Leucinodes orbonalis] in brinjal crop. In view of the above, there is an urgent need for developing alternative control strategies. Adoption of transgenic crops engineered primarily using the cry proteins to prevent damage caused by insect pests has given excellent results in cotton and maize worldwide resulting in significant economic benefits. A similar approach in brinjal is expected to provide substantial benefits to farmers.²³

These conclusions were based on field trials from 2004 through 2008; FSB-directed insecticides were reduced by 80 percent; total insecticides, by 42 percent. Tests for weediness, pollen flow, growth, toxicity, composition, nutrition and allergenicity found no substantial differences from other brinjal cultivars when compared to the Bt cultivars. The GEAC concluded that the risk profile of the new crops was lower than demonstrated hazards incurred by existing practices. Those hazards extended beyond damage to local agro-ecologies. The GEAC Expert Committee II documented specific risks to humans in current practice. Moreover risks of debt, financial failure or crop loss to farmers were addressed by the GEAC, but were conspicuously absent from the political debate over Bt brinjal. Using pooled data for the years 2004–2009 aggregated across all Indian sites, analysis indicated that Bt hybrids roughly doubled the yield of marketable fruits over national checks. This pattern held over both high- and low- yielding hybrids. The combination of lower pesticide costs and increased harvest resulted in an increase of INR 69,239 per hectare over non-Bt counterparts and INR 85,291 over national checks. This gain is conceptually the net economic benefit to the farmer: the value of higher yields added to savings on pesticides.²⁴

India's regulation of biotech crops includes provisions for establishing whether or not Bt technology would result in a net reduction of risks to farmers: the government does not want to approve cultivars that pose additional financial or agronomic risks to the rural population. Therefore, different varieties were developed for and tested in different regions of the country by public institutions, since agronomic variation is great. Tamil Nadu Agricultural University developed 4 open-pollinated varieties (OPVs) for Southern India; for the Southwest, the University of Agricultural Sciences, Dharwad, developed 6. The Indian Institute for Vegetable Research in Varanasi developed both hybrids and OPVs for the northern and eastern regions. Though Monsanto became the object of political attacks on Bt eggplants, more OPVs from the public sector than hybrids from Monsanto's private sector partner Mahyco were planned for release – in marked contrast to Bt cotton. For farmers this would mean a choice between 2 types of insect-resistant brinjals: lower-cost and savable seeds of varieties or higher-yielding, more expensive hybrid seeds. In this sense, concern for the riskiness of new seed technology was better served by plans for diffusion of Bt brinjal than Bt cotton.²⁵ Knowledge of this bio-property arrangement, however, was largely missing from public debate, at every level; even the Minister of Environment publically mentioned the risk of domination of India's food supply by Monsanto.

Findings of statutory state science were not decisive, however. Minister of Environment Jairam Ramesh concluded that the GEAC studies were inadequate; risks to food safety and the environment had, in his opinion, not been decisively disproven.²⁶ Of the 2 hypothetical risks cited by the Minister, neither could be demonstrated by a finding of some specific hazard, nor were these ‘risks’ explicitly compared to known hazards of existing practices. Instead, uncertainty was constructed as ‘risk,’ in Knight's formulation, which in the event trumped official scientific evidence.^8,27,28 Food safety was the most telling example. Only one study, not peer-reviewed and funded by an international campaigner against biotechnology (Greenpeace), was cited as evidence of hazard: organ damage and death.^24,26 However, Professor Séralini's claims about Bt proteins and organ failure had been rejected by the GMO Panel of the European Food Safety Authority; his subsequent article positing cancer risks from genetically engineered maize was retracted after publication by Food and Chemical Toxicology – a rare and embarrassing step for a journal.²⁹

Nevertheless, the Séralini paper evoked a precautionary reaction from the Minister. He declared the GEAC findings only ‘advisory’ and turned to the court of public opinion. The mechanism was public meetings in 7 cities across India. Not surprisingly, significant opposition surfaced in urban public meetings, and reinforced the Minister's worry that safety was unproved; the Bt eggplant was not released for cultivation. Defenders of the GEAC asked in a public campaign against this decision: why should meetings with 8000 people in urban India constitute national public opinion? The question is reasonable, but Indian politics in this sense reflected a global pattern of urban bias: farmers receive lagged and uncertain access to products of the genomics revolution taken for granted by urban populations, and urban populations have much to say about what crops farmers can grow.³⁰

Counter-mobilization of scientists was extensive, both internationally and in India, but politically ineffectual. The reason had nothing to do with Science, but with institutionalization of state science in India – that is, where regulatory authority to interpret scientific findings is constituted within the state. Defenders of GEAC science had no institutional leverage where it could count. Sections of the state supported the GEAC conclusions, but none had a determinant voice on the issue; the cabinet split.³¹ The Indian Council for Agricultural Research and the Department of Biotechnology, as well as the Minister of Agriculture and Minister of Science and Technology, and indeed the Prime Minister, were convinced, but statutory authority lay with the Minister of Environment. An analysis by 6 elite scientific academies produced a judgment that the Bt brinjal was safe, but this conclusion did not reverse the moratorium.³²

It's not about the science

The global rift over ‘GMOs’ often centers science. Opponents are accused of being anti-science or Luddite in their approach to technology; proponents are accused of being duped by ‘science’ produced for corporate interests backed by corporate money.³³ But the more consequential conflict is the conceptual divide between risk and uncertainty. State science in India (GEAC) found no credible evidence of hazard in consumption or growing the second Bt crop – that is, no incremental risk. The minister called the tests incomplete and inadequate: there was instead uncertainty. He called upon public meetings in 7 cities and continuous dialog to satisfy “all stakeholders” or establish “public trust and confidence.” Many scientists in India – and some Cabinet Ministers – viewed this standard as setting the bar impossibly high. There are no tests from normal science that could conceivably eliminate all uncertainty among all stakeholders. Sociology of knowledge alone suggests the reason: organized groups around the world remain unconvinced by the science supporting evolution, the safety of vaccines, viral origins of AIDS, or global warming.^34,35 Some positions against settled science are motivated by ideology, but there are rational bases for caution in the face of uncertainty. There is folk wisdom in the joint aphorisms of Donald Rumsfeld: ‘unknown unknowns’ establish pervasive uncertainty. This truism is reinforced by Rumsfeld's observation that ‘the absence of evidence does not constitute evidence of absence.’ Are tests adequate, sufficiently fine-grained, long enough in duration?

Unknown unknowns are pervasive, resulting in an ideational opportunity structure for opposition to evidentiary claims that is literally limitless. Risk-averse publics demand certainty; science can respond only with “given the current state of knowledge.” All science is thus inevitably vulnerable to politics by its very epistemological commitments: tentative results subject to continuous revision driven by better evidence or better tests or both. Science has no answer to any particular risk construction under precautionary logics when both hazard and probability distribution remain unknown. Ironically, it is the very absence of unique hazard in genetically engineered field crops¹¹ that renders them vulnerable to precautionary risk politics.

Precautionary logic characterizes some nations and some sections of states more than others. Differences reflect conflicting logics of developmental states and precautionary states. Developmental states undertake interventions in the economy that might produce positive results but are unlikely to be undertaken by private actors – railroads, research, ports, social overhead capital.³⁶ India, China and Brazil have all invested in agricultural biotechnology as developmental projects; other nations are taking a similar path.³⁷ Development entails change, and developmental states embrace some uncertainty, to be balanced with prudence. The logic behind risk-taking resembles that of individuals facing surgery or new prescription drugs: continuation of the status quo likewise entails uncertainties, and often known hazards, so absolute risk avoidance is in practice impossible. Developmental states take what they determine to be acceptable risks, believing that benefits are worth pursuing. Precautionary states, in contrast, privilege caution over opportunity; uncertainty is coded as unacceptable risk. The obvious, and paralyzing, dilemma is that it is impossible to know how much precaution is cautionary enough – especially in the absence of evidence of hazard. Yet, there are real hazards, and thus real costs, in excessive caution. For example, in contrast to hypothetical or anticipatory risk – ie, uncertainty – India's GEAC Expert Committee II documented very specific hazards in current cultivation practices to which any anticipated risk of change would have to be compared.²¹ Urban consumers, in my experience in India, know of neither the risks to themselves from pesticide residues on eggplants nor the risks to farmers and the environment in growing them with existing practices.

Political groups, nations, and departments are divided between these predispositions – the developmental and the precautionary – over time and over issues. The location of official science in the state then affects outcomes, not Science itself. One might think of food safety or agricultural productivity as matters more appropriately allocated to agencies other than the Ministry of Environment, but the national specificity of state science in India gave the environment minister veto power. Ministries of Environment tend to be preservationist, hence precautionary; their political constituencies are likewise predisposed to oppose development on grounds of preservation. Ecological systems are notoriously difficult to model, and sensitive to small and unanticipated perturbations with results that may be irreversible. The GEAC was established under legislation protecting the environment, reflecting logic of the Cartagena Protocol, itself a treaty under a framework convention protecting biodiversity. Predictably, a more precautionary construction of risk dominated the Environment Minister's approach to Bt eggplant. This outcome is true to mission. Ministries of Agriculture or Science and Technology have different missions, more in common with the logic of developmentalism; their mission involves change of the status quo, not its preservation. Benefits to farmers and farming would have been weighted more substantially by an agriculture ministry than an environment ministry. Ministers with these missions supported the GEAC decision in India.

This dialectic of risk and benefit encounters the Goldilocks Paradox of all regulation; the level of cautionary restriction should be not too much, not too little, but just right. Excessive regulation is suffocating and adverse to equity. Too little precaution might produce hazards that entail unacceptable risk. The strictest regulation enables multinational life science corporations with capacity and connections to win at the expense of small firms and public institutions.³⁸ Rejection of GEAC conclusions in the case of Bt brinjal was discouraging to public-sector scientists who might contribute to new crops for India: if criteria of risk are essentially arbitrary, and state science can be overruled by a single politician, investing a career in plant breeding is a risky choice, perhaps irrational. Uncertainty about regulatory criteria also depresses investment in both public and private sector research and development. If there is no definable hazard, and no rational way to set limits of acceptable risk, scarce public resources are diverted to surveillance and control rather than agricultural innovation. If effective and safe technologies are blocked, agriculture is needlessly crippled. In terms of assessing democratic development, regulatory regimes empower politicians and officials over farmers. Much of the ‘GMO’ debate evokes ethical principles of democratic decision-making and equality; it seems ethically difficult to justify depriving farmers of the same technical progress urban people take for granted. Finally, with climate change continually producing daunting challenges to agriculture, ruling out any tools for response is itself a risky position to take.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

Ronald Herring is Professor of Government at Cornell University. This piece is a summary of issues raised in a lecture at the Nehru Museum and Memorial Library, New Delhi, January 22, 2014. He acknowledges clarifications from Shri Jairam Ramesh at this lecture.