The COVID-19 pandemic, political instability and the climate crisis have renewed focus on the capacity and resilience of global food systems to deliver adequate food and nutrients to the growing global population. With 702–828 million people affected by hunger in 2021 (ref. 1), and more than 2 billion people suffering one or more micronutrient deficiencies, commentators have called for rigorous monitoring and evaluation of food system performance to guide policies and promote accountability2.

Reporting in Nature Food, Lividini and Masters present nutrient balance sheets (NBS) that account for the production and dietary supply of 36 nutrients ‘from farm to fork’3. Estimates reported globally for the years 1961–2018 draw on food balance sheet (FBS) data from the Food and Agriculture Organization Corporate Statistical Database (FAOSTAT). The NBS provide a unifying framework to assess and characterize food system dietary nutrient supplies, and to explore future scenarios and intervention options where deficiencies in dietary nutrient supplies are apparent. The framework enables various components of the food system — food production, trade, processing, cooking, loss and wastage and consumption — to be explored in terms of dietary nutrient supplies and deficiency risks to populations.

The approach works across agriculture–nutrition domains, integrating food systems perspectives. For example, the authors show how combinations of staple-crop biofortification, food fortification and micronutrient supplementation interventions can fill shortfalls in dietary nutrient supplies from the prevailing food system in various countries. The NBS of Lividini and Masters is the latest in a growing body of studies that report frameworks for estimating the availability of dietary nutrients for human consumption. The advance comes in the reporting of a wider range of nutrients over the full timescale of available FBS data for most countries globally, and the use of the supply and utilization account data that sit behind the FBS data, giving greater granularity and allowing users to trace back to production stage.

The FBS data that underpin this study are suited to comparisons cross-country and over time, due in part to their consistent structure. FBS data are integral to various nutrition and food system models and tools such as the HarvestPlus Biofortification Priority Index tool, which considers the potential of biofortified crops in different countries; the International Food Policy Research Institute IMPACT model4, which provides estimates of future nutrient supplies under different future scenarios of food system change; and dietary risk factors in the Global Burden of Disease5. In addition, the FAOSTAT Food and Diet Domain project will report the supply of dietary nutrients available for human consumption, based on FBS, household survey and individual-level food consumption data.

Although FBS data are powerful, they are geographically coarse — providing estimates of the food available for consumption at national level, for up to 96 distinct food items. FBS data do not capture subnational variation in diets, including the variation between regions, socioeconomic or sociocultural groups, and gender and demographic groups. As such, the framework is only suitable for certain applications, and there may be instances where integration or triangulation with other data sources may be useful. Indeed, the authors suggest that the NBS could be strengthened through integration with household consumption and expenditure survey (HCES) data, including that of the family of Living Standards Measurement Study surveys6. Alternatively, where individual-level dietary data are available, these could be used to inform subnational distributions of intake and variation between socioeconomic and demographic groups.

There is growing use of HCES data in widescale assessments of dietary nutrient supplies and in nutrition modelling tools7; and although the HCES data are not available in all countries, the socioeconomic and spatial resolution that they provide is undoubtedly valuable. However, relatively little attention has been paid to increasing the quality and spatial resolution of food composition data. There is substantial variation in staple-crop nutrient composition due to soil, climate, agronomic and other factors, and the spatial scales at which this variability occurs is likely to drive important variation in nutrient intakes, particularly in contexts where food systems are predominantly localized8,9. There is a need to establish routine surveillance of crop nutrient composition, particularly for staple crops due their dominance in the diets of low-income populations, as well as the development of nutrient accounting frameworks that can incorporate these data at subnational scales. This should be a priority area of work to support the rigorous monitoring and evaluation of food system performance to inform policies in support of resilient and sustainable global food systems.