Assumptions underlying different models of sustainable agriculture must be tested for robustness – in this way the risks of path dependence can be reduced.
The unsustainability of agriculture and food systems is receiving unprecedented news coverage due to the increasingly well documented range of interlinked adverse impacts: the role in driving climate change, by accounting for about 37 per cent of anthropogenic greenhouse gas emissions; the contribution to deforestation and declining biodiversity; the input to air and water pollution; the degradation of soils; and the role in driving the emergence and/or spread of disease (e.g. COVID-19) or antimicrobial resistance. Furthermore, increasing concern is being paid to the resilience of food systems to climate change, and the role of climate change itself in undermining food security, supply-chain disruption and food price spikes. Given these headline topics, the question of how agriculture can be made ‘sustainable’ is increasingly at the forefront of debates taking place at international meetings (e.g. the first UN Food Systems Summit in September 2021, the UN Framework Convention on Climate Change’s COP26 in November 2021 and the forthcoming COP15 of the UN Convention on Biological Diversity in late 2022), as well as discussions in the media and on social media. Getting agriculture ‘right’ within the context of wider global food systems can contribute to improvements both in environmental outcomes for climate, soils, water and biodiversity, and in nutritional outcomes for public health.
While there is no agreed definition of sustainable agriculture, there is a vast body of literature on the many environmental impacts which result from farming practices. It is evident from the literature that agreeing on what makes agriculture sustainable is difficult, for the following reasons:
- Agriculture can have multiple impacts on the environment (through pollution, affecting soil, air or water quality; emissions of greenhouse gases; soil degradation; or the reduction or degradation of habitats, resulting in biodiversity loss). Reducing one metric (e.g. greenhouse gas emissions) may bring about a trade-off in terms of other metrics (e.g. biodiversity loss).
- Context dependence must be taken into account. The same practice, undertaken in different places, can have different outcomes on the same aspect of sustainability, due to local variations in environment or climate.
- Many environmental impacts vary with the scale of the intervention (or its frequency). For example, the intensive management of a single field in a large area of agroecological farmland may not affect local biodiversity to a measurable extent; but if the same agricultural practice was scaled up to landscape level, it would. Hence, it is difficult to characterize a practice as sustainable – or unsustainable – without reference to the scale of its implementation.
- In a market-driven system, the demand for agricultural outputs gives rise to linkages between production areas, even if they are in different countries. This means that if yield is sacrificed to reduce environmental impacts in one place, then in order to meet demand the market will incentivize production in other places, potentially with lower standards for environmental governance. Increases in production could be achieved through agricultural intensification, or by clearing land in order to convert it to farmland: and, in aggregate, this process reduces the sustainability of the overall food system.
Therefore, there may not be a simple approach to defining sustainability in agriculture without simultaneously also defining sustainability in food systems as a whole – and, more broadly, examining how land and other key resources are used by different sectors globally.
The following thought experiment serves as a useful illustration of the complexities involved. Let us imagine that research highlights a new farming system which reduces greenhouse gases, conserves biodiversity, stores carbon, maximizes animal welfare and preserves the cultural, recreational and amenity value of the landscape. While yields are only 70 per cent of those achieved through standard intensive farming practices, the public goods are improved to such an extent that they more than compensate, in terms of economic value, for the relatively lower yields at the level of individual farms. Whether this farming system is more sustainable is illustrated by considering the following two hypothetical cases:
- (1) Demand for food stays the same, i.e. consumers continue to eat the same types and amounts of food: in which case, if the ‘sustainable agriculture’ system were to be adopted globally, it would imply that the area of agricultural land would have to be expanded by at least 42 per cent (i.e. 1 divided by 0.7) to make up for the reduction in yield. Land clearance for this purpose would release greenhouse gas emissions, cause biodiversity losses, and degrade water quality and availability around the world. This would suggest that, despite being more sustainable locally, the proposed form of agriculture might lead to a less sustainable system as a whole.
- Alternatively, in hypothetical condition (2), demand for food changes: people are incentivized not to waste food, to adopt a healthy diet rich in plant-based foods, and to avoid excess consumption. With a fall in demand of 30 per cent, all farms could adopt the new, hypothetical, farming system, while delivering the raft of environmental benefits and without needing to expand the land ‘footprint’ of agriculture. So, the same farming system could be seen to be unsustainable in case (1) but sustainable in case (2).
As this scenario illustrates, the criteria for what makes agriculture sustainable or not are often confusing and contested. This confusion often arises due to a lack of clarity about whether sustainability is being discussed in the literature as a property of the agricultural system, or the farm, or the context, or the aggregate or relative impact (in terms of the food system’s environmental footprint, or the footprint per kg of agricultural product). The aim of this paper is to unpack some of the framing issues and examine the underlying assumptions being made in different discussions, thus bringing some clarity as to how particular arguments are being made – from different viewpoints – which underpin perspectives on how agriculture may become more sustainable.
The goal of the paper is to provide transparency as to why there are different – logically consistent and plausible, but nonetheless deeply contested – models for sustainable agriculture and food systems. This systemic appraisal is built on an examination of the robustness of the underlying assumptions, and is relevant to a wide range of actors within global food systems, including political decision-makers, investors and corporate stakeholders, as well as civil society actors. These assumptions may not be apparent: instead, they may be implicit, or unexamined, but often they are based rather less on strong evidence and more on what proponents may think is desirable, including from an ideological perspective. On one level, this paper is about definitions, but it is, to a greater extent, about the enabling assumptions that are made to underpin and legitimize different directions of travel in order to achieve the ‘sustainability transition’.
In short, the assumptions that are made to support a particular version of how agriculture affects the sustainability of food systems serve to channel policymakers and investment in certain directions. In turn, choosing a specific version of sustainability of agriculture has huge implications for both the amount and type of foods that are produced and for the impact of agriculture on climate change, pollution, biodiversity loss and human health. Given the urgency of tackling each of these issues, examining the robustness of the underlying assumptions that support a given direction of travel will reduce the risks of a ‘path dependence’ that aims to increase sustainability but does not achieve the necessary sustainability goals.
Unpacking the narratives that support different ‘versions’ of sustainable agriculture
To analyse the principal narratives on sustainable agriculture, this paper contrasts alternative ‘versions’, conceptually similar to ‘mindsets’ or paradigms, originally described by Donella Meadows in 1999 thus: ‘The mindset or paradigm out of which the system – its goals, power structure, rules, its culture – arises’. The ‘versions’ of sustainable agriculture within sustainable food systems presented below (Table 1) bear some similarities to the contrasting hypothetical cases described above. Version 1 focuses on agricultural intensification to raise yields per unit of area through production gains, while minimizing environmental impacts (primarily focusing on the efficiency of agricultural production), and ‘sparing’ land from agriculture. Version 2 focuses on agroecological (nature-friendly) approaches that produce lower yields but with lower environmental impacts, and thus contributes to ‘sharing’ land with nature/wildlife.
The key assumptions made in the arguments that link agricultural practice at the farm level and the impacts on food systems more generally are laid out in Table 1 (and discussed at length below). Both versions are logical and internally consistent subject to their assumptions. The core difference between the two versions is in the assumption about the future of demand (and thus the amount of land and mode of farming that are necessary to meet it). How demand may (or may not) grow depends on a range of factors, including the way the market works (or how it might work). These assumptions often have an ideological basis that is sometimes, but not always, explicitly acknowledged.
As an example of the debate, a common perspective is that intensive farming wearing the guise of ‘sustainable’ intensification is a more valid, more pragmatic, approach to sustainable food production than nature-friendly, agroecological farming. This is in spite of the fact that the vast majority of the academic literature comparing conventionally intensive and agroecological farms shows that the latter are associated with greater biodiversity, more stored carbon, more productive soils and, perhaps, less pollution. In other words, the empirical data support the claim that intensive farming is more environmentally impactful (less ‘sustainable’) at the field, farm and landscape levels, but the assumptions made about scale effects lead to the reverse conclusion. In essence, the debate is therefore not so much about the evidential basis of sustainability, but the assumptions made in its interpretation.
