A few months ago, a new analysis came out with a shocking number: about 1.6 billion people could not afford a healthy, sustainable diet. The diet in question was the EAT Lancet Commission’s "healthy reference diet," based largely on vegetables, fruits, whole grains, legumes, and nuts and meant to promote human health, and the analysis was eye-opening about the challenges facing our food system in terms of making healthy diets affordably available.
But then, a few weeks ago, the publication of the 2020 "State of Food Insecurity and Nutrition in the World" (SOFI) report further rocked the food and nutrition community with an even more startling number: an estimated 3 billion people could not afford a healthy diet.
Clearly, the number of people experiencing such poverty could not have actually doubled in such a short period of time, nor could food prices (especially as the analysis was completed before the COVID-19 pandemic). So what, we wondered, was the difference?
The answer lies in two critical questions at the heart of analysing the affordability of healthy diets: what is "affordable," and what is a "healthy diet"?
What is affordable?
The main difference between the two estimates lies in what they consider to be ‘affordable’ to a given individual or household. The analysis of the EAT Lancet diet declared the diet unaffordable if its cost (per person) was above household income (also per person) - that is, under the assumption that the household could spend its entire income on food. Such a threshold is useful in one way: one cannot argue that a diet above this threshold could possibly be affordable. However, it is also deeply problematic. No household faces zero non-food expenses, and it would be counterproductive to improving nutrition (not to mention other social goals) if households were to not invest in health, education, sanitation and hygiene, and information technologies - and counterproductive to improving wellbeing not to leave households with some free budget left to spend on cultural, social, and recreational pursuits. Indeed, the paper’s authors note these limitations (but do not alter their main finding as a result).
In the SOFI analysis, in contrast, the affordability threshold chosen is 63% of per capita household income, based on the proportion of expenditures on food by lower-income consumers in lower-income countries. This lower threshold expands the authors’ count for those unable to afford a healthy diet from 1.86 billion people (under assumptions similar to those used in the analysis of the EAT Lancet diet) to 3.02 billion people. Certainly, this estimate is more realistic. But it is still restrictive: spending nearly two-thirds of one’s income on food leaves little for any additional necessary or wellbeing-enhancing expenses. Under more realistic assumptions based on actual current spending rates on food (15-50% of income, by country income group), the appendix to the SOFI report estimates that 4.6 billion people cannot afford a healthy diet. Adjusting the threshold further to allow for those in lower-income countries to spend more on enhancing the quality of their lives, beyond their diets, would raise the count ever higher.
What is a healthy diet?
The second main difference between the two analyses is the definition of a "healthy diet." The EAT Lancet analysis examined a reference diet that is based on ranges for grams/calories obtained from various general food categories (e.g., whole grains, dark green vegetables, fish, legumes); the diet itself was chosen with only health and nutrition concerns in mind, but the EAT Lancet Commission then related it to a set of environmental targets, showing that it could be provided to the global population while keeping the food system within a set of planetary boundaries related to environmental sustainability. Notably, this diet has much smaller amounts of red meat and starchy vegetables (e.g., potatoes) than are currently eaten in most world regions, but much more nuts, whole grains, legumes, fruit, and vegetables. The affordability analysis regrouped the recommended food categories into "functional groups" and chose the cheapest food within each group that was available in a given country. The diet is thus a hypothetical one and based on one food per group.
Woman picking avocados in the market in Surabaya. © GAIN / Andrew Suryono
The SOFI diet, in contrast, is based on ten national food-based dietary guidelines (from different world regions), which are then applied in each country by selecting the two least-cost food items in each food group that are available locally. The average cost of these ten least-expensive food baskets is then taken as the estimate for an affordable diet. This diet thus provides adequate calories and nutrients alongside a more diverse intake of foods (i.e., more than one food per group) and in line with actual cultural practices, as documented in the dietary guidelines. It does not include any accounting for sustainability. In general, the SOFI "healthy diets" tend to slightly more expensive than an EAT Lancet-derived diet (USD 3.27-4.57 per person per day, versus USD 3.31- 3.61, when similar data and assumptions are used to calculate costs) - implying that if the EAT Lancet affordability analysis were repeated using SOFI assumptions, the number unable to afford a healthy diet would be significantly higher than 1.6 billion.
In both cases, the diet considered is more expensive than one that delivers mere adequacy of nutrient intake - which SOFI estimates that 1.5 billion people would still be unable to afford.
Conclusions
Ultimately, the headline conclusion of the two analyses is the same: whether 1.6 billion or 3 billion, too many people currently cannot afford a healthy diet. Under either estimation technique, this should serve as a call to action.
The EAT Lancet affordability analysis made an important contribution in casting light on just how difficult it is to afford a healthy diet; the SOFI analysis took this further, by considering actual culture-based dietary practices, building in a certain amount of diversity, and using more realistic assumptions for the amount of income that can be spent on food. But it can likely be even further improved.
When setting targets, it is important that these embody the values that we wish to guide social progress. In the case of diets, this goes beyond mere nutrient adequacy to include quality of life that is supported by being able to spend much of one’s budget on non-food expenses and the dignity and wellbeing that come with being able to select a diet in line with one’s social and cultural preferences. It also includes the co-benefits of choosing a healthy and sustainable diet, thereby contributing to addressing an urgent need to reduce the environmental impacts of food systems.
Thus, while both affordability analyses have made important, policy-relevant contributions to thinking on this topic, future analyses may be able to set even more ambitious assumptions and targets.