Low-carbon diets can reduce global ecological and health costs

The study addresses how dietary changes—specifically reductions in animal-sourced foods (ASF)—can mitigate not only greenhouse gas (GHG) emissions but also broader external costs to human health and ecosystems that are not reflected in food prices. Given that the global food system contributes about one-third of anthropogenic GHG emissions, identifying effective decarbonization measures is critical for meeting Paris Agreement targets. While strategies include waste reduction, improved agricultural practices and efficiency gains, dietary change offers considerable mitigation potential without requiring new technologies. Prior work has documented higher emissions and environmental pressures from ASF relative to plant-based foods (PBF) and associated public health benefits of shifting diets. However, the indirect socio-economic and ecological cost repercussions from food production externalities remain under-explored at the global level. This study aims to quantify the hidden production-related external costs embedded in national diets worldwide and to evaluate the potential reductions in these costs, as well as GHG emissions and health burdens, under nine low-carbon dietary change scenarios that progressively reduce ASF.

Previous research has quantified environmental externalities of diets using bottom-up and top-down approaches, but typically with limited scope: single countries, specific food types (for example, fruits and vegetables, ASF, pork) or single impact categories (for example, GHG emissions, nitrogen flows). Comprehensive assessments that jointly appraise global ecological and socio-economic collateral costs of dietary change across multiple environmental impact categories and regions have been lacking. Earlier studies have shown climate benefits from reducing ASF, contributions to planetary boundaries (land use, biogeochemical flows, biodiversity, water) and health improvements, but they often omit monetized production-related external costs and emphasize either consumption-related health effects or narrow environmental endpoints. This work fills that gap by integrating global food supply data, life cycle impact assessment across multiple impact categories, and monetization of endpoint damages to provide a holistic estimate of externalities and potential savings from dietary shifts.

Study design: A global, consumption-based, bottom-up assessment linking national food supplies to production-caused environmental impacts and endpoint damages, monetized to estimate hidden external costs, and combined with a comparative risk assessment to quantify consumption-related health benefits of dietary change. Data: National food supply (kg/cap/yr, kcal/cap/day) for 2018 from FAO Food Balance Sheets (FBS) covering 101 countries (~91% of global population). National final consumption expenditure (FCE) on food and non-alcoholic beverages (at-home) from USDA ERS (2018). Life cycle inventory (LCI) data for 708 food items from ecoinvent v3.5, Agri-footprint v4.0 and ESU World Food databases, including additional items for mealworms, processed plant-based meat and milk alternatives. LCA/LClA: Cradle-to-gate system boundaries (farm/fishery to primary processing node; excludes retail and use phases). Midpoint impacts and endpoint damages computed in SimaPro v9.1 using ReCiPe2016 (hierarchist, 100-year horizon). Midpoints include terrestrial acidification, freshwater/marine/terrestrial ecotoxicity, freshwater/marine eutrophication, land use, global warming, ozone formation, water use, human toxicity (cancer, non-cancer), particulate matter formation, ionizing radiation, ozone depletion. Endpoints: human health (DALYs) and ecosystem quality (time-integrated species loss over next century). Conversion of LCI flows to midpoints and midpoints to endpoints followed ReCiPe2016 characterization factors. Country-specific vs global-average intensities applied for domestic production; export-weighted global average intensities used for imports. Impacts of imported food allocated to importing country (consumption-based accounting). Monetization: Human health DALYs monetized via budget constraint method (Weidema), valuing 1 DALY at US$176,624 (2018; uncertainty US$148,288–200,236). Ecosystem damage monetized as willingness to pay for species protection with benefit transfer from 2000 to 2018 using US GDP per capita (PPP) growth and income elasticity 0.38, yielding US$17,891,594 per time-integrated species loss (uncertainty US$4,472,899–44,728,985). Dietary scenarios: Nine hypothetical global scenarios applied to each country relative to 2018 baseline (BASE): EAT (EAT-Lancet reference diet); No red meat (NRM); NRM with insects (NRM-I); Pescatarian (PESC); PESC with insects (PESC-I); Vegetarian whole-food (VEG); Vegetarian with processed plant-based meat (VEG-P); Vegan whole-food (VGN); Vegan with processed plant-based meat and milk substitutes (VGN-P). Substitution rules applied on a calorie basis: for NRM/PESC/VEG/VGN, lost ASF calories replaced by two-thirds whole plant proteins (beans, legumes, soybeans) and one-third fruits/vegetables; variants replace one-third with insects (mealworms) or processed plant-based alternatives. Total supplied calories per capita kept constant across scenarios (excluding EAT which follows EAT-Lancet targets). Comparative risk assessment: Estimated consumption-related health benefits (avoided DALYs) for four dietary risk factors (low fruit, low vegetable, low legume, high red meat intake) and four disease endpoints (coronary heart disease, stroke, cancer, type-II diabetes) using literature relative risks and population-attributable fractions, assuming uncorrelated risks. Uncertainty: Propagated via 1,000-run Monte Carlo on LCA impact intensities and bounds on monetization factors; reported 95% confidence ranges for global externalities and avoided DALYs. Key assumptions and scope: Cradle-to-gate only (lower bound on total impacts); no production feedbacks (intensities unchanged with scale shifts); many LCI entries from HICs; insect proxy limited to mealworms; expenditure data exclude out-of-home consumption.

- Global hidden external costs of food production embedded in diets in 2018 are US$14.0 trillion (US$5.9–32.8T), about 17% (7–39%) of world GDP, exceeding China’s GDP. Of this, US$8.3T (US$4.9–13.4T) is linked to human health burden and US$5.7T (US$1.1–19.4T) to ecosystem quality decline. - For every US$1 of consumer food expenditure, average hidden production-related external costs are US$1.94 (US$0.82–4.56): US$1.15 to human health and US$0.79 to ecosystems. Thus, the true cost of the average global diet is nearly triple the consumer outlay. - External costs as a share of GDP vary strongly by income: ~7% (3–16%) in high-income countries (HICs) versus ~102% (43–228%) in low-income countries (LICs). - Per capita monetized externalities are highest in North America and Oceania (~US$4,200 per year) and lowest in South Asia and Sub-Saharan Africa (~US$1,100). - Food group contributions: Meat (beef, pork, lamb, chicken) accounts for 51% of global production-related externalities; cereals are also major contributors, especially in lower-income regions; legumes/nuts/pulses contribute more to externalities in lower-income countries due to higher reliance. - Mitigation via diet change: Removing only red meat (NRM) reduces externalities by ~US$4.0T (US$1.8–9.4T) and GHG emissions by ~2.3 Gt CO2e (2.0–2.8). A vegan whole-food scenario (VGN) yields the largest savings: ~US$7.3T (US$3.2–17.0T) and ~4.5 Gt CO2e (3.9–5.8), equivalent to 9% (4–20%) of global GDP and ~110% (75–280%) of the required food-sector GHG reduction to meet the EAT-Lancet climate boundary. - Regional responsibility: Upper-middle-income and high-income countries (UMICs/HICs) deliver 81–100% of potential externality abatement across scenarios; eliminating ASF in LICs/LMICs yields modest savings, and adopting EAT in LICs/LMICs could increase GHG emissions. - Alternative substitutes: Insect protein and processed plant-based foods can reduce externalities but generally less than replacing ASF with whole legumes, fruits and vegetables; VEG and VGN externalities are 14% and 21% higher, respectively, when ASF are substituted with processed plant-based products rather than whole foods; little difference between NRM and NRM-I. - Health and ecosystem endpoints: Moving from BASE to VGN could avoid up to 25.6 million production-related DALYs (17.4–38.3 million). Meat elimination (PESC) captures most of this potential (≈19.9 million DALYs; 13.7–29.0), largely via reduced agricultural water use. Species-loss prevention: up to 155,000 species lost (110,000–218,000) avoided with VGN; most from meat removal (PESC ≈132,000; 95,300–181,000), driven by lower water consumption and land use. - Health benefits split: Approximately one-third of total avoided DALYs from dietary shifts are due to reduced production-related environmental damages, and two-thirds are due to direct consumption-related risk reductions; neglecting production effects underestimates health benefits of plant-forward diets.

The findings reveal that current diets embed substantial, monetizable external damages to human health and ecosystems, indicating that consumers and societies underpay for food when externalities are excluded. Monetization facilitates comparison across heterogeneous damage categories and can inform policies to internalize external costs, such as eco-labels, taxes, or incentives for lower-impact foods. Dietary transitions toward lower shares of ASF can concurrently mitigate climate change and reduce broader external costs, especially in HICs and UMICs where ASF consumption and per-capita externalities are highest. However, implementing an externalities accounting framework at scale requires consensus on LCA methods, better data coverage, and agreed valuation of DALYs and biodiversity impacts. Processed plant-based and insect-based alternatives can support transitions but may not achieve the same externality reductions or nutritional profiles as whole-plant foods; they may nonetheless reduce behavioral barriers by preserving culinary habits. For LICs and LMICs, priorities include improving access to healthy, diverse diets and advancing sustainable production practices for all foods; adopting EAT-like patterns may increase GHG emissions without concurrent production-side improvements. Addressing dominant drivers (water use, land use, particulate matter formation, etc.) through agricultural practice changes—sustainable intensification, agroecology, improved livestock feed and management, and decarbonizing on-farm energy—can further reduce externalities independent of consumer behavior. Overall, integrating production- and consumption-linked health perspectives broadens the understanding of diet-health-environment interdependencies and underscores the importance of targeting high-income regions for the largest near-term gains.

Dietary change toward lower proportions of animal-sourced foods, as part of broader food system transformation, can substantially reduce production-caused damages to human health and ecosystem quality, translating into trillions of dollars in potential external cost savings and significant GHG abatements. Quantifying hidden externalities embedded in national diets clarifies the societal benefits of low-carbon dietary shifts and supports policies to incentivize these transitions, particularly in HICs and UMICs where savings potential is greatest. The study contributes a global, monetized appraisal of production-related external costs and demonstrates that including production-linked health effects is essential to fully valuing the benefits of plant-forward diets. Future research should enhance LCI coverage across regions and production methods, include post-farm stages and feedbacks from scaling production, broaden dietary risk factors and disease endpoints, refine valuation methods for biodiversity and health, and assess policy instruments for internalizing external costs without compromising nutrition and equity, especially in lower-income settings.

- System boundaries exclude retail, packaging, transport beyond cradle-to-gate and cooking/use phases, likely underestimating total impacts and externalities and the differences between higher- and lower-income countries. - LCI data gaps: Many impact intensities rely on global averages and are skewed toward high-income regions; limited representation of production method variability for some food groups. - No feedbacks modeled: Impact intensities assumed constant despite potential system responses (e.g., expansion to less suitable land increasing inputs; economies of scale for insects reducing intensities). - Monetization uncertainties and ethical considerations: Valuing DALYs and biodiversity loss involves assumptions (budget constraint method, benefit transfer, income elasticity) with wide uncertainty ranges. - Dietary health modeling limited to four risk factors (low fruit, vegetable, legume intakes; high red meat) and four disease endpoints, assuming uncorrelated risks and average exposure at the population level. - Expenditure data reflect at-home consumption only; hidden cost per US$1 may be underestimated relative to total (including out-of-home) consumption. - Insect impacts based on mealworms only; generalization to other edible insects is uncertain. - EAT diet modeling may increase emissions in LICs/LMICs absent concurrent production-side improvements; results are scenario-based and not prescriptive dietary recommendations.