Options for reforming agricultural subsidies from health, climate, and economic perspectives

The study investigates how agricultural subsidy reforms can align health, climate, and economic objectives. Agricultural subsidies strongly influence production choices, yet current food systems are neither healthy nor sustainable. The research question is which reform options—removal, repurposing toward nutrition-sensitive and low-emission foods, and/or global restructuring of payments—can reduce greenhouse gas emissions and improve population health without diminishing economic welfare. To address this, the authors apply an integrated economic–environmental–health modelling framework to quantify global, regional, and national impacts to 2030.

The paper situates its analysis within prior work on food systems, health, and environmental limits (e.g., Lancet Commission on healthy diets; Nature studies on environmental limits), and on agricultural subsidies and policy reforms (OECD monitoring, CAP reform discussions). It references evidence linking subsidies and nutrition/obesity outcomes, systematic reviews on input subsidies and food security, and proposals for greener CAP pathways. Methodologically, it builds on established CGE policy-assessment models (MAGNET) and comparative risk assessment frameworks (akin to Global Burden of Disease). It also acknowledges literature on environmental footprints of foods and farming system impacts, and policy discourses such as “public money for public goods.”

The authors use a coupled modelling framework integrating economic, health, and environmental assessments. Economic model: a computable general equilibrium (CGE) model with agricultural detail (MAGNET) to estimate impacts of changes in agricultural support on production, trade, consumption, and economic welfare. MAGNET explicitly represents land resources and agricultural policies, including different payment types (output, intermediate input, land-, capital-, labour-based) and degrees of coupling. Subsidy data: OECD Producer Support Estimate (PSE) budgetary transfers as implemented in GTAP; market price support is reflected in GTAP tariff data. Model resolution: 28 countries/regions and 34 sectors. Downscaling: Production and consumption results are downscaled to country level using FAO production data and FAO food availability adjusted for household food waste, tracing flows from primary production through processing and food service to final consumption. Health model: a global comparative risk assessment linking dietary and weight-related risk factor changes to cause-specific mortality (CHD, stroke, type-2 diabetes, cancers including colorectal, and respiratory disease). Risk factors include high red meat; low fruits, vegetables, nuts, legumes, fish; and underweight, overweight, obesity (via energy intake). Relative risks are taken from meta-analyses of prospective cohort studies. Feedbacks: Health outcomes (diet-related deaths, including among working-age population) feed back into the CGE model by adjusting population and labour force, affecting output and welfare. Environmental assessment: GHG emissions footprints paired to production/consumption changes, focusing on methane and nitrous oxide (CO2 allocated to other sectors per IPCC). Animal-source food footprints include direct (enteric/ manure) and indirect (feed) emissions. Projections to 2030 include anticipated technological and management improvements consistent with socioeconomic trajectories. Scenarios (target year 2030, baseline on middle-of-the-road population and GDP growth, labour growth, and projected biophysical yield developments): (1) RMV—complete removal of subsidy payments; (2) Repurposing—redirecting 50% (S50) or 100% (S100) of subsidy budgets to low-emitting, nutrition-sensitive commodities (vegetables, fruits, legumes, nuts) in a budget-neutral way; (3) Repurposing plus global redistribution—keeping the global subsidy total fixed but allocating by country according to GDP (GDP scenario) or population (POP scenario), thereby extending subsidies to previously non-subsidising countries, and using them for nutrition-sensitive, low-emitting production. Uncertainty: structured analysis across alternative socioeconomic pathways (varying population and economic growth); parameter uncertainty within pathways not included.

Removal of subsidies (RMV): • Production/emissions: GHG emissions moderately reduced in OECD (−1.8%), slightly reduced in non-OECD (−0.1%), slightly increased in non-subsidising countries (+0.5%). • Diet/health: Per-capita fruits and vegetables decreased across all regions (about −6 g/d on average; 1–9 g/d range), total energy intake decreased (~−11 kcal/d; 2–21 kcal/d range). Net increase in diet-related mortality: +75,000 deaths in 2030 (95% CI 71,000–80,000), driven by reduced fruit and vegetable intake, partially offset by reductions in overweight/obesity. • Economics: Mortality increase reduced welfare by ~USD 1 billion, but gains in allocative efficiency from reduced distortions amounted to USD 11 billion; terms-of-trade effects reduced gains for net importers (OECD ToT −USD 1.1 billion) and increased them for net exporters. Net economic impact positive in most regions. Repurposing to nutrition-sensitive, low-emitting foods (S50/S100): • Production/emissions: Horticultural production increased markedly in OECD (+19% under complete repurposing) and modestly in non-OECD (+3%); slight reductions in non-subsidising countries (−2.4%). GHG emissions moderately reduced in OECD (−1.7%), roughly unchanged in non-OECD (−0.2%), slightly increased in non-subsidising countries (+0.5%). • Diet/health: With 100% repurposing, fruits and vegetables consumption rose by +55 g/d (~10%) in OECD, +31 g/d (~5%) in non-OECD, and +2 g/d (0.3%) in non-subsidising countries via lower world prices. Avoided deaths: 444,000 fewer in 2030 (95% CI 429,000–460,000). • Economics: Increased labour supply from improved health generated benefits (USD 12 billion). However, greater sectoral targeting reduced allocative efficiency, especially in OECD (−USD 20 billion). Small allocative efficiency gains in non-OECD (+USD 0.3 billion) where subsidies offset taxes on horticulture. Net effect negative in OECD under full repurposing; half repurposing (S50) reduced allocative efficiency losses and mitigated welfare reductions but also roughly halved health benefits. Repurposing plus global redistribution (GDP/POP): • Production/emissions: More even global increases in fruit/vegetable production, with large increases in previously non-subsidising countries, especially in POP scenario (+4%). GHG reductions similar or larger than repurposing-only: −0.3% (GDP) to −0.4% (POP). • Diet/health: More equal distribution of fruit/vegetable consumption gains; in previously non-subsidising countries: +12 g/d (POP) and +5 g/d (GDP), vs +2 g/d under repurposing-only. Avoided deaths: ~370,000–379,000 in 2030 overall, with more equal per-capita mortality reductions across regions. • Economics: Global welfare increases of ~USD 1.8–5.5 billion. OECD regions experienced net gains as allocative efficiency losses diminished or reversed and labour gains accrued. Non-OECD impacts similar to repurposing-only. Previously non-subsidising countries faced allocative efficiency losses partly offset by labour gains; potential for full compensation via transfer payments as other regions’ net gains were about twice the net losses.

The integrated analysis shows that subsidy reforms can contribute to healthier and more sustainable food systems, but trade-offs depend on design. Removing subsidies improves allocative efficiency and reduces GHG emissions modestly, yet worsens population health via reduced fruit and vegetable intake, partially offsetting economic gains through labour-force losses. Repurposing subsidies to nutrition-sensitive, low-emitting foods improves diet quality and reduces diet-related mortality and emissions but can reduce allocative efficiency, particularly in high-subsidy OECD contexts. Partial repurposing can balance health and economic outcomes. Combining repurposing with global redistribution by GDP or population maintains substantial health gains, yields similar or larger GHG reductions, distributes benefits more equitably across regions, and achieves positive global welfare, though it creates losses in newly subsidising regions that may require compensatory transfers. Overall, incorporating health-related welfare into economic evaluations is critical to capture feedbacks between diet, labour supply, and efficiency when assessing agricultural policy.

Reforming agricultural subsidies toward nutrition-sensitive, low-emission production can deliver meaningful health and climate benefits while remaining economically feasible, especially when partially repurposed and complemented by a fairer global distribution of subsidy payments. The study’s integrated framework demonstrates that up to half of current subsidies could be repurposed to support fruits, vegetables, legumes, and nuts without compromising economic welfare, and that combining repurposing with redistribution by GDP or population yields comparable health gains, similar or greater emissions reductions, and net global welfare improvements. Policy approaches such as “public money for public goods” could enhance feasibility by aligning payments with public health and environmental objectives. Future research should design mechanisms that jointly incentivise healthier consumption, low-emission production, and improved farming practices across broader environmental dimensions, while addressing regional equity and political economy constraints.

The analysis couples subsidies to sector-level production characteristics but does not differentiate management practices within sectors or consider environmental impacts beyond GHGs (e.g., biodiversity, water use, pesticide use), which can vary for horticultural systems. Sub-national impacts, including farm-level income changes and distributional effects within countries, are not assessed. Many determinants of production beyond subsidies (e.g., geography, climate, resource availability, market integration, consumer demand) are not directly modelled as policy levers here. Politically, coupling subsidies to specific commodities may be ambitious in some contexts (e.g., OECD), and implementation would require alignment with evolving policy frameworks (e.g., EU Farm to Fork, UK agriculture bill). The uncertainty analysis spans socioeconomic pathways but does not include parameter uncertainty within pathways. Newly subsidising countries in redistribution scenarios may incur allocative efficiency losses that would necessitate compensatory transfers to share in global gains.