Studies and Reports of the Month: April 2026

This study, funded by the European Research Council, analyses FAOSTAT Food Balance Sheets for 188 countries across 51 years (1970 to 2021), using quantile regression to track how each country moves through the global distribution of dietary composition over time. The framework distinguishes income-driven nutrition transition from compositional change driven by societal norms or policy.

• Global nutrition transition is real and quantified. Every tenfold increase in per-capita GDP corresponds to a 13% rise in the dietary share of animal-sourced calories and a 15% decline in cereals. The pattern holds across 188 countries and 51 years.
• Six high-income countries diverge from the pattern. Canada, Finland, Norway, New Zealand, Switzerland, and the United Kingdom show both relative and absolute declines in animal-sourced food calories. Canada’s animal-sourced share fell 12.3% since 1970, with absolute animal calories down 15.4% (154.8 kcal per capita per day).
• Brazil and Portugal accelerate beyond the trend. Brazil and Portugal show animal-product calorie share rising 13.0 and 15.1% respectively (+175.6% and +144.6% in absolute terms), exceeding the global trajectory. Brazil’s transition is faster than its income trajectory alone would predict.
• No global protein transition is occurring. Legume calorie share keeps falling worldwide and sits far below recommended intake. The six divergent countries are no different: their animal-share decline runs mostly into vegetable oils, cereals, and poultry, while legumes barely move (the largest increase among the six is about 38 kcal per capita per day, against an EAT-Lancet reference near 426).
• Even diverging HICs remain above EAT-Lancet thresholds. The six divergent countries have reduced animal-sourced food intake but remain above the EAT-Lancet planetary health diet recommended levels. The finding documents directional change, not target attainment.

The data establish that the nutrition transition is largely income-driven, but not destiny. Six high-income countries demonstrate that compositional shifts away from animal-sourced foods can occur even at high income levels, supporting the argument that structural and norm-based interventions can reshape dietary trajectories beyond what income alone predicts. The Canadian trajectory is a useful empirical anchor for North American advocacy: a high-income, agriculturally productive country cut its animal-sourced calorie share by 12.3% over five decades. The honest qualifier belongs with it: roughly 69% of Canada’s plant-based calorie gain was vegetable oil, cereals were second, and legumes were essentially flat. This is evidence that compositional change is possible, not that a protein transition occurred.

Three implications for food systems transformation:

1. The absence of any global protein transition toward legumes confirms that demand-side momentum on its own is not redirecting consumption to nutritionally and environmentally preferable plant proteins; targeted infrastructure, procurement, and pricing policy is required.
2. Brazil and Portugal’s accelerated trajectories against the global trend identify priority geographies for engagement before the consumption pattern locks in further.
3. The finding that even leading HICs remain above EAT-Lancet thresholds disciplines messaging: directional progress has not equaled target attainment.

FAOSTAT supply data measure availability rather than consumption; food losses, waste, and within-country distribution inequities are not captured. The analysis is calorie-based and does not address protein quality, micronutrient density, or processing levels. In the six divergent countries, the analysis cannot identify which specific policies, cultural shifts, or market dynamics drove their trajectories – mechanism-level case studies are the natural next step. Within-country variation is masked by national averages, particularly relevant for large diverse countries where coastal and interior, urban and rural, or income-stratified trajectories may diverge sharply.

A comprehensive review, by an eight-author team across several countries, synthesizing 205 sources on the bidirectional relationship between food production and environmental change. Coverage spans climate, soils, nutrients, water quality and availability, biodiversity, croplands, livestock systems, and blue foods. The paper documents both how food production drives environmental degradation and how that degradation now constrains food production itself.

• Animal-sourced foods drive disproportionate environmental pressure per calorie. The plant-sourced calorie-pressure slope is roughly eight times steeper than the animal-sourced slope, meaning plant foods deliver 8x more calories per unit of environmental damage. Blue foods remain low in both calorie contribution and environmental impacts.
• Livestock and feed production account for around 60% of all food-system greenhouse gas emissions. Total food-system emissions reach an estimated 17 Gt CO2-equivalent per year, equivalent to 25-33% of total annual anthropogenic emissions. Business-as-usual food-system emissions alone could breach the 1.5 °C and even the 2.0 °C warming targets.
• Animal feed absorbs over 40% of agricultural water, 35% of cropland, and 44-47% of synthetic nitrogen, plus 100% of pasture. The upstream resource demand of animal-sourced food is structurally larger than the foods themselves.
• Methane emissions from enteric fermentation of ruminants alone (112 Tg per year) approach those of the entire fossil fuel sector (120 Tg per year). Combined with landfills, waste, and rice cultivation, food-system methane (212 Tg per year) substantially exceeds fossil fuel sector methane.
• Over 90% of blue food production is vulnerable to human-induced environmental change, affecting food security for 3.2 billion people. Coral reefs have declined by half since the mid-twentieth century, and ocean warming is projected to shift global fishing patterns markedly by 2050.

Supply-side technological gains alone cannot break the food-environment feedback loop. Mitigation pursued without adaptation produces rebounds, with reductions in food waste partly offset by increased consumption. Adaptation pursued without mitigation locks in resource-intensive consumption patterns. Integrated strategies that couple plant-rich diets such as the Planetary Health Diet with circular feed systems, agroecology, climate-smart agriculture, and redesigned food procurement could reduce land use by 71% and emissions by 29% per capita in Europe while ensuring nutritional adequacy. Translating these limits into national and sub-national targets will require explicit attention to power, ownership, and equity so that environmental gains do not come at the cost of a just transition.

The review treats demand-side change primarily through a behavioural feasibility lens rather than through structural reform of food environments such as procurement standards, fiscal instruments, and institutional defaults. Corporate concentration and political economy dynamics that lock in current production patterns receive limited attention. The brief discussion of alternative proteins – cited as 88% CO2-equivalent reduction and 82% land-use reduction versus conventional animal-sourced foods – does not engage the rebound and political economy literature that complicates that framing. Further research is needed on policy instrument design (taxes, subsidies, regulation, choice architecture) to dissuade consumption of high-impact foods without creating affordability shocks for lower-income consumers.

This review synthesizes evidence across seven intervention domains to map the levers available for transitioning global food systems toward healthier, more sustainable, and more equitable diets. Drawing on 154 references across behavioural science, economics, and public health, the paper connects the behaviors of consumers, producers, and the midstream actors that influence both.

• Hidden costs of food systems: Current food prices capture only about one-third of their true societal costs. The authors argue that the hidden environmental and health externalities total an estimated $20 trillion annually, more than double the value of food consumed globally.
• Livestock land use: Approximately 90% of US agricultural land is tied to livestock production, including both grazing areas and cropland devoted to producing feed crops such as corn, soy, and hay.
• Price-driven dietary change: A meta-analysis found that a 20% reduction in fruit and vegetable prices yields a 16.6% increase in purchases. Yet equalizing food access and prices between low- and high-income US households would close only about 10% of the nutritional gap, with the remaining 90% driven by demand-side factors.
• Food-as-medicine potential: Simulations suggest a 30% Medicare and Medicaid subsidy on fruit and vegetable purchases could prevent 1.93 million cardiovascular events and save $39.7 billion in healthcare costs.

The review reframes the strategic priority from awareness campaigns and individual behavior change to structural reforms: public procurement standards that incorporate true-cost accounting, regulation of food environments including emerging digital delivery platforms with an estimated 3 billion users globally, and R&D investment in making healthy and sustainable foods more affordable, accessible, and appealing. The paper identifies seven complementary intervention domains and argues that contextualized bundles of these interventions, not isolated measures, are needed for systemic change.

The review does not quantify the relative cost-effectiveness of different intervention types or their interactions when combined, which is the critical question for decision-makers allocating limited resources. Most evidence comes from high-income countries, particularly the United States, limiting direct applicability to low- and middle-income settings where dietary transitions are most urgent. The paper also flags GLP-1-based weight-loss drugs as an emerging disruptor whose implications for food system sustainability and equity warrant close attention but remain largely unstudied.

A Bayesian path analysis of 2,083 German adults integrated the Theory of Planned Behaviour (the dominant cognitive model of food behaviour) with egocentric social network methods. Participants reported their own attitudes, perceived behavioural control, intentions and behaviours regarding healthy eating and meat consumption, alongside the perceived attitudes and behaviours of up to 15 named members of their close social network. The study asked a sharper question than most behaviour research has tackled: when both cognition and observed network behaviour are measured, which one predicts what people actually eat?

• Observed peer behaviour outperformed every cognitive variable in predicting what people actually ate. What close ties were doing predicted both meat consumption and healthy eating over and above participants’ own attitudes, perceived behavioural control, and the social pressure they reported feeling. The independent effect on meat consumption was 0.43 and 0.22 on healthy eating, both after full adjustment for the cognitive model.
• Network effects ran on behaviour, not on intention. When intention was the outcome, network variables added essentially no explanatory power. When behaviour was the outcome, network behaviour was independently and substantially associated even after full cognitive adjustment.
• What runs through the social channel is perceived normality, not approval. Network behaviour was most strongly linked to descriptive norms (estimate 0.61 for healthy eating; estimate -0.45 for meat consumption), not to injunctive norms about what others endorse. People update on what they see happening around them.

These findings sit alongside a converging body of work on how norms shift. Cass Sunstein’s coverage of preference falsification shows that norms operate as costs on deviation, so populations systematically underestimate the readiness for change because public conformity masks private dissent. Damon Centola’s research explains why behaviour change in such conditions requires reinforcement through multiple trusted ties rather than broadcast exposure to information. This study supplies the food-specific empirical anchor: cognitive models systematically undercount how much eating behaviour is shaped by what is seen rather than what is said.

The implication for institutional food environments is that changing what is offered or made default is likely a more reliable lever than changing what is recommended. The authors recommend three intervention pathways grounded in their results, all of them coordination interventions rather than persuasion ones:

1. Partner-involved programmes rather than individuals alone, creating reinforcement among close ties.
2. Modelling opportunities such as workplace lunch groups, creating visibility of adoption.
3. Structural measures including price subsidies for healthy foods, changing the choice architecture so that going against a high-meat-consuming network is not also financially costly.

Because the study measured only the 15 closest ties, it captures how strong-tie reinforcement turns an already-considered change into action, not how a new norm is first introduced. That seeding step runs through weaker ties and cross-cluster brokers, so reinforcement and introduction are complementary jobs rather than interchangeable ones.

The study has the design limitations of its kind. Cross-sectional data cannot disentangle selection (people choosing similar others) from influence (being shaped by them). All measures are self-reported, including network members’ attitudes and behaviours, which were not validated by the network members themselves. Networks were truncated at 15 close ties (approximately the size of Dunbar’s ‘sympathy group’ of inner-ring relationships). This cap captures the strong-tie reinforcement channel cleanly but is structurally blind to the weak-tie ring of acquaintances, coworkers, and cross-cluster brokers, where other research suggests novel norms are first introduced. The sample is German – a country with high baseline meat consumption – so the magnitudes will not transfer cleanly to other dietary contexts. The longitudinal, real-world studies the authors call for would convert correlational evidence into tested causal mechanism.

This is the first comprehensive assessment of Japan’s food system against the EAT-Lancet Healthy Reference Diet (EHRD), integrating consumption, supply, self-sufficiency, and environmental footprint analysis. The authors adjusted the global EHRD for Japanese sex- and age-specific energy requirements and calculated three sufficiency rates using the National Health and Nutrition Survey (2011 to 2019) and Food Balance Sheets (1960 to 2020). They developed a Japan-specific environmental footprint model based on input-output life cycle assessment, covering CO2 emissions, water use, and biodiversity loss derived from both.

• Severe red meat oversupply. Beef, lamb, and pork consumption reached 569% of the recommended level in 2019, up from 463% in 2011. Domestic supply runs at 448% of the recommended level, sustained by a 45% domestic production base, indicating that overconsumption is sustained primarily by imports. The 2019 mix is approximately 60% pork, 20% beef, and 20% processed meats.
• Persistent plant-based deficits. Tree nut intake sits below 5% of the recommended level, fruit intake at 27%, and unsaturated oils at 29%. Legume self-sufficiency is 8% (24% if restricted to edible beans). Six decades of supply data show these gaps are structural and have not closed over time.
• Hidden feed dependency. Although domestic self-sufficiency for chicken (66%) and eggs (97%) appears high, Japan’s concentrate feed self-sufficiency was only 12% in 2019. The apparent self-sufficiency of these animal products masks heavy upstream import dependence on feed crops.
• Modeled environmental gains and trade-offs. Aligning Japanese adult diets with the EHRD would reduce CO2 emissions by 5.6%, water use by 2.3%, and water-related biodiversity loss by 10.3%, with red meat reduction the dominant driver. Reducing red meat creates synergistic environmental benefits at home and abroad, while increasing legumes, fruits, and nuts under current production structures shifts environmental pressures to producer countries.

The findings reframe Japan’s dietary misalignment as structurally produced rather than the result of consumer choice. Supply trends, trade dependence, and feed reliance together mean voluntary, awareness-based interventions cannot close the gap alone. The authors propose three coordinated policy directions: developing a Japan-specific dietary guideline that integrates environmental sustainability alongside health and culture; using subsidies and public procurement to lower the price of recommended foods, particularly fruits, legumes, and nuts; and reforming the animal protein supply chain, including feed sustainability, to ensure that protein transitions do not simply move environmental burdens abroad. The hidden feed dependency insight applies wherever poultry and egg sectors report high domestic self-sufficiency while relying on imported feed, extending the relevance of the analysis to other high-income contexts.

The environmental modeling assumes constant production technologies and trade structures, so the reported reductions are likely conservative for red meat (where livestock contraction would amplify gains) and may understate trade-offs for plant-based foods (where increased demand could induce land expansion in producer countries). The analysis covers only three of the nine planetary boundaries, leaving land-system change and biogeochemical flows unaddressed. Whole grains, palm oil, and tree nuts were excluded from the supply-side analysis due to data aggregation, and added sugars and oils intake estimates are likely underestimated because national surveys do not fully capture processed-food contributions. Future work using individual-level data and dynamic land-use modeling would address these gaps.

This systematic content analysis examined the most recent sustainability reports and websites (2021 to 2024) of 33 of the world’s 35 largest meat and dairy companies, applying the Nemes et al. (2022) greenwashing framework to every environmental claim. The framework codes claims for vague language, missing evidence, selective disclosure, irrelevance, false hope, false comparisons, and shifting blame.

• Greenwashing is near-universal. Of 1,233 environmental claims identified, 98% (1,213) contained at least one greenwashing indicator. Only 29% of claims were supported by any evidence, and only three of 1,233 cited scholarly literature. 38% (467) were unverifiable promises about the future.
• BlackRock ownership correlates with claim volume. Companies with higher BlackRock ownership stakes generated more environmental claims, consistent with sustainability-linked financial pressure manifesting as communications rather than emissions reductions. JBS raised $1 billion in sustainability-linked bonds during the study period.
• Pattern matches fossil fuel greenwashing. The mechanisms documented here are structurally identical to those previously demonstrated for the fossil fuel industry, where similar near-universal greenwashing patterns are now subject to active litigation, regulatory action, and investor risk repricing.

The findings document at scale what regulators, litigators, and investors increasingly suspect: voluntary corporate sustainability reporting in the meat and dairy sector is not delivering the emissions reductions it advertises. Several active processes can use this evidence directly. The New York Attorney General’s lawsuit against JBS for misleading climate claims, California’s SB 253 and SB 261 emissions disclosure laws, the EU’s Green Claims Directive negotiations, and pending misleading-advertising cases against Danish Crown, Tyson, Arla, and Fonterra are all sharpened by a peer-reviewed near-census showing that 98% of the sector’s claims contain greenwashing indicators.

The structural conclusion is that voluntary disclosure regimes do not produce reliable environmental information from this sector. Mandatory disclosure with verification, third-party assurance comparable to financial audit, and enforcement against false advertising are the policy instruments that would convert promises into reductions. The fossil fuel parallel is informative: the same communications-as-substitute-for-action pattern was identified there, and the same regulatory and litigation infrastructure now applies pressure that voluntary regimes did not.

The analysis covers public-facing communications and does not measure actual operational emissions, so it cannot independently quantify the gap between claims and performance. Two of the 35 largest companies were excluded due to insufficient publicly available material. Coding is necessarily judgment-based even with a structured framework, though intercoder reliability is reported and adequate. Future work should track whether mandatory disclosure regimes (California, EU) reduce greenwashing prevalence, and should pair claims analysis with operational emissions data to quantify the misrepresentation gap directly.

This UK dietary modelling study assessed the nutritional, environmental, and cost outcomes of nine substitution scenarios applied to a gender-stratified Food Foundation basic basket for lower-income adults. The substitution categories (processed meats, milk, yogurt) were each paired with three selection criteria: most popular, most nutritionally balanced, and least expensive. Novel plant- and fungi-based food selection was grounded in 52 weeks of Kantar Worldpanel real purchase data identifying 72 NPBFs across the GB market.

• All nine substitution scenarios reduced dietary environmental footprints. Greenhouse gas emissions fell by 1.47 to 6.41%, land use by 0.72 to 4.55%, and water use by 0.39 to 6.94%. Processed meat substitutions delivered the largest environmental savings.
• Processing level and nutritional quality are separable. 100% of the 25 most-purchased plant-based meats and 100% of plant-based drinks in the UK were classified as healthy under the UK Nutrient Profiling Model, despite over 88% falling under NOVA category 4 (ultra-processed). The processing axis and the nutritional axis behave differently for these products.
• Nutritional adequacy preserved when nutritionally balanced products were selected. All micronutrient weekly recommended dietary allowances continued to be met. The main exception was iodine: the most popular plant-based drink (almond-based) reduced men’s iodine intake to 52% of the RDA, indicating that fortification choice matters for substitution at scale.
• Cost effects were split between dairy and meat substitutions. Plant-based dairy alternatives mostly saved money, up to £1.01 per week for women’s yogurt substitution. Plant-based processed meats increased costs by £0.97 to £2.17 per week, representing up to 3.8% of the men’s weekly basic basket.

The findings address one of the most common objections to plant-forward dietary advice for lower-income households: that single targeted substitutions undermine nutritional adequacy or are unaffordable. In a UK low-income context, single substitutions of animal-sourced foods with commercially available novel plant- and fungi-based foods preserved nutritional adequacy across more than 15 micronutrients and reduced environmental footprints in every scenario tested.

The findings also separate the processing axis from the nutritional axis, with implications for blanket ultra-processed-food taxation proposals that could inadvertently penalize plant-based alternatives scoring healthy under nutrient profiling. The price differential for plant-based meats creates a direct policy ask for affordability interventions, including price parity measures, public procurement standards, and demand-side subsidies, to remove the cost barrier that currently limits substitution at population scale in lower-income households.

The cost analysis is based on a single major UK retailer in March 2024 without promotional discounts; price dynamics are evolving rapidly and vary by retail context. The substitution model captures only single-item swaps within an otherwise unchanged basic basket, not whole-diet shifts; broader environmental and nutritional effects of integrated dietary change cannot be inferred from these scenarios. The basic basket itself sits below UK Eatwell Guide thresholds, reflecting affordability constraints; the cited figure that only 0.01% of the UK population currently meets the Eatwell Guide indicates how far the realistic baseline sits from any aspirational target. Supply-chain and corporate-concentration dynamics, which condition whether population-scale substitution is structurally feasible, are not addressed.

This preprint reports a prospective health impact assessment using a life table approach to model how dietary shifts embedded in four contrasting French net-zero emission scenarios would affect adult mortality between 2025 and 2050. The four pathways are drawn from the ADEME Transition(s) 2050 framework, an established French climate-energy modelling exercise that grounds the dietary shifts in technically feasible decarbonization scenarios. Baseline diet is taken from the NutriNet-Santé cohort (n=28,245); dose-response relationships for 13 food groups are sourced exclusively from AMSTAR-2 high-rated meta-analyses; uncertainty is propagated via Monte Carlo simulation across 1,000 iterations and six sensitivity analyses.

• All four scenarios prevent 19-24% of all-cause adult mortality by 2050, around 152,000 to 191,000 deaths avoided yearly, with a floor near 154,000 under the most conservative pathway. The result holds across six sensitivity analyses, where 100,000 to 200,000 deaths are prevented under every parameter combination.
• The largest mortality reductions come from increased plant foods and decreased meat. Modelled increases in nuts (+486 to +918%), whole grains (+145 to +284%), and legumes (+68 to +457%) combine with decreases in red meat (-38 to -79%) and processed meat (-34 to -91%). Both directions of change contribute meaningfully.
• Life expectancy gains of 17.6 to 22.8 months by 2050. Had dietary transition begun in 2016 following the Paris Agreement, 6,000 to 10,000 additional deaths could have been prevented by 2026, quantifying the cumulative cost of delay.

The most policy-relevant finding is that every modelled decarbonization pathway delivers large health co-benefits. This neutralizes the common objection that dietary transition only works if a specific climate strategy is chosen. Diet emerges as a leading channel for health co-benefits across all climate mitigation sectors, with mortality reductions larger than those typically modelled from air pollution improvements or active-transport increases alone.

The quantified delay costs reframe inaction as an active harm rather than a neutral choice. France’s net-zero planning and ongoing dietary guideline revisions can directly use these results, and the broader implication for European policy is that integrated climate-health frameworks should treat the dietary transition as a primary public health intervention with co-benefit framing, not as a secondary outcome of climate action.

The study is currently a preprint and has not completed peer review; findings should be monitored for publication and any methodological adjustments. The NutriNet-Santé cohort is a convenience sample skewed toward health-conscious participants, which the authors mitigate by using the lowest organic-consumption quintile as the baseline, but residual selection bias remains. The model is France-specific and transferability to other European or non-European contexts will require country-specific baselines and dose-response calibration. Mortality outcomes are projected from observational dose-response relationships rather than experimental data, so all effects are subject to residual confounding even with the 25% conservative attenuation applied. Equity and affordability dimensions of the dietary transition are not modelled.

This study analyzes 61 years of panel data from 139 countries (1961 to 2021) to map how urbanization interacts with agricultural modernization to shape nitrogen fertilizer use, nitrogen yield, and nitrogen use efficiency in croplands. The researchers use dynamic panel models with country and year fixed effects, stratified by World Bank income group, to capture nonlinear and stage-dependent relationships. Data sources include FAOSTAT, the USDA International Agricultural Productivity database, FAO AQUASTAT, the World Bank, and the CHANS nitrogen budget model.

• Urbanization’s effect on nitrogen outcomes is nonlinear and development-dependent. In low-income countries, urbanization drives fertilizer use up while suppressing yield and efficiency. In high-income countries, the relationship reverses: at 90% urbanization, advanced modernization shifts nitrogen use efficiency from a 4% decline to a 12% gain.
• Global nitrogen fertilizer application has risen sevenfold since 1961. Global mean application rose from 9 kg N per hectare in 1961 to 67 kg N per hectare in 2021. Most regions followed a steady upward trajectory reflecting the post-WWII diffusion of synthetic fertilizers; Eastern Europe is the notable deviation.
• In high-income countries, further farm consolidation makes things worse. Farm size expansion acts as a negative moderator in high-income contexts, exacerbating declines in both nitrogen yield and use efficiency. This reverses the conventional assumption that bigger farms are always more efficient at advanced development stages.
• In low-income countries, larger farm size buffers early efficiency losses. Farm size is the dominant moderator at low income levels, shifting the turning point of the urbanization-yield curve from 52% down to 34% and providing resilience against the early-stage efficiency penalties otherwise associated with urbanization.

The development-stage contingency dismantles the assumption that a single nitrogen management strategy can be promoted globally. For high-income agricultural systems, the policy implication is that mechanization and irrigation investment, not further consolidation, are the levers that recover nitrogen efficiency under continued urbanization. For low- and middle-income countries, the pathway is structurally different and includes a productive role for farm consolidation up to specific thresholds.

The findings have direct relevance for nitrogen pollution mitigation, agricultural greenhouse gas emission reductions (particularly nitrous oxide), and the nitrogen pollution boundary in the planetary boundaries framework. They also indicate that international nitrogen policy targets cannot be set as a single global standard, and that climate-policy and agricultural-policy alignment must be calibrated to country development stage rather than imposed uniformly.

The paper does not disaggregate cropland by end use, which is its largest analytical limitation from a food systems standpoint. Roughly one-third of global cropland produces animal feed, and feed crops have fundamentally different nitrogen efficiency profiles per unit of human nutrition delivered. The headline nitrogen use efficiency figure aggregates across food and feed production, masking the question of how the urbanization-nitrogen relationship would shift under different cropland composition scenarios. The associations are not experimentally established as causal, and endogeneity is mitigated but not eliminated by the dynamic panel specification.