E286: How 'least cost diet' models fuel food security policy

In this episode of the Leading Voices in Food podcast, host Norbert Wilson is joined by food and nutrition policy economists Will Masters and Parke Wilde from Tufts University's Friedman School of Nutrition, Science and Policy. The discussion centers around the concept of the least cost diet, a tool used to determine the minimum cost required to maintain a nutritionally adequate diet. The conversation delves into the global computational methods and policies related to least cost diets, the challenges of making these diets culturally relevant, and the implications for food policy in both the US and internationally. You will also hear about the lived experiences of people affected by these diets and the need for more comprehensive research to better reflect reality.

Interview Summary

I know you both have been working in this space around least cost diets for a while. So, let's really start off by just asking a question about what brought you into this work as researchers. Why study least cost diets? Will, let's start with you.

I'm a very curious person and this was a puzzle. So, you know, people want health. They want healthy food. Of course, we spend a lot on healthcare and health services, but do seek health in our food. As a child growing up, you know, companies were marketing food as a source of health. And people who had more money would spend more for premium items that were seen as healthy. And in the 2010s for the first time, we had these quantified definitions of what a healthy diet was as we went from 'nutrients' to 'food groups,' from the original dietary guidelines pyramid to the MyPlate. And then internationally, the very first quantified definitions of healthful diets that would work anywhere in the world. And I was like, oh, wow. Is it actually expensive to eat a healthy diet? And how much does it cost? How does it differ by place location? How does it differ over time, seasons, and years? And I just thought it was a fascinating question.

Great, thank you for that. Parke?

There's a lot of policy importance on this, but part of the fun also of this particular topic is more than almost any that we work on, it's connected to things that we have to think about in our daily lives. So, as you're preparing and purchasing food for your family and you want it to be a healthy. And you want it to still be, you know, tasty enough to satisfy the kids. And it can't take too long because it has to fit into a busy life. So, this one does feel like it's got a personal connection.

Thank you both for that. One of the things I heard is there was an availability of data. There was an opportunity that seems like it didn't exist before. Can you speak a little bit about that? Especially Will because you mentioned that point.

Will: Yes. So, we have had food composition data identifying for typical items. A can of beans, or even a pizza. You know, what is the expected, on average quantity of each nutrient. But only recently have we had those on a very large scale for global items. Hundreds and hundreds of thousands of distinct items. And we had nutrient requirements, but only nutrient by nutrient, and the definition of a food group where you would want not only the nutrients, but also the phytochemicals, the attributes of food from its food matrix that make a vegetable different from just in a vitamin pill. And those came about in, as I mentioned, in the 2010s. And then there's the computational tools and the price observations that get captured. They've been written down on pads of paper, literally, and brought to a headquarters to compute inflation since the 1930s. But access to those in digitized form, only really in the 2000s and only really in the 2010s were we able to have program routines that would download millions and millions of price observations, match them to food composition data, match that food composition information to a healthy diet criterion, and then compute these least cost diets. Now we’ve computed millions and millions of these thanks to modern computing and all of that data.

Great, Will. And you've already started on this, so let's continue on this point. You were talking about some of the computational methods and data that were available globally. Can you give us a good sense of what does a lease cost diet look like from this global perspective because we're going to talk to Parke about whether it is in the US. But let's talk about it in the broad sense globally.

In my case the funding opportunity to pay for the graduate students and collaborators internationally came from the Gates Foundation and the UK International Development Agency, initially for a pilot study in Ghana and Tanzania. And then we were able to get more money to scale that up to Africa and South Asia, and then globally through a project called Food Prices for Nutrition. And what we found, first of all, is that to get agreement on what a healthy diet means, we needed to go to something like the least common denominator. The most basic, basic definition from the commonalities among national governments' dietary guidelines. So, in the US, that's MyPlate, or in the UK it's the Eat Well Guide. And each country's dietary guidelines look a little different, but they have these commonalities. So, we distilled that down to six food groups. There's fruits and vegetables, separately. And then there's animal source foods altogether. And in some countries they would separate out milk, like the United States does. And then all starchy staples together. And in some countries, you would separate out whole grains like the US does. And then all edible oils. And those six food groups, in the quantities needed to provide all the nutrients you would need, plus these attributes of food groups beyond just what's in a vitamin pill, turns out to cost about $4 a day. And if you adjust for inflation and differences in the cost of living, the price of housing and so forth around the world, it's very similar. And if you think about seasonal variation in a very remote area, it might rise by 50% in a really bad situation. And if you think about a very remote location where it's difficult to get food to, it might go up to $5.50, but it stays in that range between roughly speaking $2.50 and $5.00.

Meanwhile, incomes are varying from around $1.00 a day, and people who cannot possibly afford those more expensive food groups, to $200 a day in which these least expensive items are trivially small in cost compared to the issues that Parke mentioned. We can also talk about what we actually find as the items, and those vary a lot from place to place for some food groups and are very similar to each other in other food groups. So, for example, the least expensive item in an animal source food category is very often dairy in a rich country. But in a really dry, poor country it's dried fish because refrigeration and transport are very expensive. And then to see where there's commonalities in the vegetable category, boy. Onions, tomatoes, carrots are so inexpensive around the world. We've just gotten those supply chains to make the basic ingredients for a vegetable stew really low cost. But then there's all these other different vegetables that are usually more expensive. So, it's very interesting to look at which are the items that would deliver the healthfulness you need and how much they cost. It's surprisingly little from a rich country perspective, and yet still out of reach for so many in low-income countries.

Will, thank you for that. And I want to turn now to looking in the US case because I think there's some important commonalities. Parke, can you describe the least cost diet, how it's used here in the US, and its implications for policy?

Absolutely. And full disclosure to your audience, this is work on which we've benefited from Norbert's input and wisdom in a way that's been very valuable as a co-author and as an advisor for the quantitative part of what we were doing. For an article in the journal Food Policy, we use the same type of mathematical model that USDA uses when it sets the Thrifty Food Plan, the TFP. A hypothetical diet that's used as the benchmark for the maximum benefit in the Supplemental Nutrition Assistance Program, which is the nation's most important anti-hunger program. And what USDA does with this model diet is it tries to find a hypothetical bundle of foods and beverages that's not too different from what people ordinarily consume. The idea is it should be a familiar diet, it should be one that's reasonably tasty, that people clearly already accept enough. But it can't be exactly that diet. It has to be different enough at least to meet a cost target and to meet a whole long list of nutrition criteria. Including getting enough of the particular nutrients, things like enough calcium or enough protein, and also, matching food group goals reasonably well. Things like having enough fruits, enough vegetables, enough dairy. When, USDA does that, it finds that it's fairly difficult. It's fairly difficult to meet all those goals at once, at a cost and a cost goal all at the same time. And so, it ends up choosing this hypothetical diet that's almost maybe more different than would feel most comfortable from people's typical average consumption.

Thank you, Parke. I'm interested to understand the policy implications of this least cost diet. You suggested something about the Thrifty Food Plan and the maximum benefit levels. Can you tell us a little bit more about the policies that are relevant?

Yes, so the Thrifty Food Plan update that USDA does every five years has a much bigger policy importance now than it did a few years ago. I used to tell my students that you shouldn't overstate how much policy importance this update has. It might matter a little bit less than you would think. And the reason was because every time they update the Thrifty Food Plan, they use the cost target that is the inflation adjusted or the real cost of the previous edition. It's a little bit as if nobody wanted to open up the whole can of worms about what should the SNAP benefit be in the first place.

But everything changed with the update in 2021. In 2021, researchers at the US Department of Agriculture found that it was not possible at the old cost target to find a diet that met all of the nutrition criteria - at all. Even if you were willing to have a diet that was quite different from people's typical consumption. And so, they ended up increasing the cost of the Thrifty Food Plan in small increments until they found a solution to this mathematical model using data on real world prices and on the nutrition characteristics of these foods. And this led to a 21% increase in the permanent value of the maximum SNAP benefit. Many people didn't notice that increase all that much because the increase came into effect at just about the same time that a temporary boost during the COVID era to SNAP benefits was being taken away. So there had been a temporary boost to how much benefits people got as that was taken away at the end of the start of the COVID pandemic then this permanent increase came in and it kind of softened the blow from that change in benefits at that time. But it now ends up meaning that the SNAP benefit is substantially higher than it would've been without this 2021 increase. And there's a lot of policy attention on this in the current Congress and in the current administration. There's perhaps a skeptical eye on whether this increase was good policy. And so, there are proposals to essentially take away the ability to update the Thrifty Food Plan change the maximum SNAP benefit automatically, as it used to. As you know, Norbert, this is part of all sorts of things going on currently. Like we heard in the news, just last week, about plans to end collecting household food security measurement using a major national survey. And so there will be sort of possibly less information about how these programs are doing and whether a certain SNAP benefit is needed in order to protect people from food insecurity and hunger.

Parke, this is really important and I'm grateful that we're able to talk about this today in that SNAP benefit levels are still determined by this mathematical program that's supposed to represent a nutritionally adequate diet that also reflects food preferences. And I don't know how many people really understand or appreciate that. I can say I didn't understand or appreciate it until working more in this project. I think it's critical for our listeners to understand just how important this particular mathematical model is, and what it says about what a nutritionally adequate diet looks like in this country. I know the US is one of the countries that uses a model diet like this to help set policy. Will, I'd like to turn to you to see what ways other nations are using this sort of model diet. How have you seen policy receive information from these model diets?

It's been a remarkable thing where those initial computational papers that we were able to publish in first in 2018, '19, '20, and governments asking how could we use this in practice. Parke has laid out how it's used in the US with regard to the benefit level of SNAP. The US Thrifty Food Plan has many constraints in addition to the basic ones for the Healthy Diet Basket that I described. Because clearly that Healthy Diet Basket minimum is not something anyone in America would think is acceptable. Just to have milk and frozen vegetables and low-cost bread, that jar peanut butter and that's it. Like that would be clearly not okay. So, internationally what's happened is that first starting in 2020, and then using the current formula in 2022, the United Nations agencies together with the World Bank have done global monitoring of food and nutrition security using this method. So, the least cost items to meet the Healthy Diet Basket in each country provide this global estimate that about a third of the global population have income available for food after taking account of their non-food needs. That is insufficient to buy this healthy diet. What they're actually eating is just starchy staples, oil, some calories from low-cost sugar and that's it. And very small quantities of the fruits and vegetables. And animal source foods are the expensive ones. So, countries have the opportunity to begin calculating this themselves alongside their normal monitoring of inflation with a consumer price index. The first country to do that was Nigeria. And Nigeria began publishing this in January 2024. And it so happened that the country's national minimum wage for civil servants was up for debate at that time. And this was a newly published statistic that turned out to be enormously important for the civil society advocates and the labor unions who were trying to explain why a higher civil service minimum wage was needed. This is for the people who are serving tea or the drivers and the low wage people in these government service agencies. And able to measure how many household members could you feed a healthy diet with a day's worth of the monthly wage. So social protection in the sense of minimum wage and then used in other countries regarding something like our US SNAP program or something like our US WIC program. And trying to define how big should those benefit levels be. That's been the first use.

A second use that's emerging is targeting the supply chains for the low-cost vegetables and animal source foods and asking what from experience elsewhere could be an inexpensive animal source food. What could be the most inexpensive fruits. What could be the most inexpensive vegetables? And that is the type of work that we're doing now with governments with continued funding from the Gates Foundation and the UK International Development Agency.

Will, it’s fascinating to hear this example from Nigeria where all of the work that you all have been doing sort of shows up in this kind of debate. And it really speaks to the power of the research that we all are trying to do as we try to inform policy. Now, as we discussed the least cost diet, there was something that I heard from both of you. Are these diets that people really want? I'm interested to understand a little bit more about that because this is a really critical space.Will, what do we know about the lived experiences of those affected by least cost diet policy implementation. How are real people affected?

It's such an important and interesting question, just out of curiosity, but also for just our human understanding of what life is like for people. And then of course the policy actions that could improve. So, to be clear, we've only had these millions of least cost diets, these benchmark 'access to' at a market near you. These are open markets that might be happening twice a week or sometimes all seven days of the week in a small town, in an African country or a urban bodega type market or a supermarket across Asia, Africa. We've only begun to have these benchmarks against which to compare actual food choice, as I mentioned, since 2022. And then really only since 2024 have been able to investigate this question. We're only beginning to match up these benchmark diets to what people actually choose. But the pattern we're seeing is that in low and lower middle-income countries, people definitely spend their money to go towards that healthy diet basket goal. They don't spend all of their additional money on that. But if you improve affordability throughout the range of country incomes - from the lowest income countries in Africa, Mali, Senegal, Burkina Faso, to middle income countries in Africa, like Ghana, Indonesia, an upper middle-income country - people do spend their money to get more animal source foods, more fruits and vegetables, and to reduce the amount of the low cost starchy staples. They do increase the amount of discretionary, sugary meals. And a lot of what they're eating exits the healthy diet basket because there's too much added sodium, too much added sugar. And so, things that would've been healthy become unhealthy because of processing or in a restaurant setting. So, people do spend their money on that. But they are moving towards a healthy diet. That breaks down somewhere in the upper income and high-income countries where additional spending becomes very little correlated with the Healthy Diet Basket. What happens is people way overshoot the Healthy Diet Basket targets for animal source foods and for edible oils because I don't know if you've ever tried it, but one really delicious thing is fried meat. People love it. And even low middle income people overshoot on that. And that displaces the other elements of a healthy diet. And then there's a lot of upgrading, if you will, within the food group. So, people are spending additional money on nicer vegetables. Nicer fruits. Nicer animal source foods without increasing the total amount of them in addition to having overshot the healthy diet levels of many of those food groups. Which of course takes away from the food you would need from the fruits, the vegetables, and the pulses, nuts and seeds, that almost no one gets as much as is considered healthy, of that pulses, nuts and seeds category.

Thank you. And I want to shift this to the US example. So, Parke, can you tell us a bit more about the lived experience of those affected by least cost diet policy? How are real people affected?

One of the things I've enjoyed about this project that you and I got to work on, Norbert, in cooperation with other colleagues, is that it had both a quantitative and a qualitative part to it. Now, our colleague Sarah Folta led some of the qualitative interviews, sort of real interviews with people in food pantries in four states around the country. And this was published recently in the Journal of Health Education and Behavior. And we asked people about their goals and about what are the different difficulties or constraints that keep them from achieving those goals. And what came out of that was that people often talk about whether their budget constraints and whether their financial difficulties take away their autonomy to sort of be in charge of their own food choices. And this was something that Sarah emphasized as she sort of helped lead us through a process of digesting what was the key findings from these interviews with people. One of the things I liked about doing this study is that because the quantitative and the qualitative part, each had this characteristic of being about what do people want to achieve. This showed up mathematically in the constrained optimization model, but it also showed up in the conversations with people in the food pantry. And what are the constraints that keep people from achieving it. You know, the mathematical model, these are things like all the nutrition constraints and the cost constraints. And then in the real conversations, it's something that people raise in very plain language about what are all the difficulties they have. Either in satisfying their own nutrition aspirations or satisfying some of the requirements for one person or another in the family. Like if people have special diets that are needed or if they have to be gluten free or any number of things. Having the diets be culturally appropriate. And so, I feel like this is one of those classic things where different disciplines have wisdom to bring to bear on what's really very much a shared topic.

What I hear from both of you is that these diets, while they are computationally interesting and they reveal some critical realities of how people eat, they can't cover everything. People want to eat certain types of foods. Certain types of foods are more culturally relevant. And that's really clear talking to you, Will, about just sort of the range of foods that end up showing up in these least cost diets and how you were having to make some adjustments there. Parke, as you talked about the work with Sarah Folta thinking through autonomy and sort of a sense of self. This kind of leads us to a question that I want to open up to both of you. What's missing when we talk about these least cost diet modeling exercises and what are the policy implications of that? What are the gaps in our understanding of these model diets and what needs to happen to make them reflect reality better? Parke?

Well, you know, there's many things that people in our research community are working on. And it goes quite, quite far afield. But I'm just thinking of two related to our quantitative research using the Thrifty Food Plan type models. We've been working with Yiwen Zhao and Linlin Fan at Penn State University on how these models would work if you relaxed some of the constraints. If people's back in a financial sense weren't back up against the wall, but instead they had just a little more space. We were considering what if they had incentives that gave them a discount on fruits and vegetables, for example, through the SNAP program? Or what if they had a healthy bundle of foods provided through the emergency food system, through food banks or food pantries. What is the effect directly in terms of those foods? But also, what is the effect in terms of just relaxing their budget constraints. They get to have a little more of the foods that they find more preferred or that they had been going without. But then also, in terms of sort of your question about the more personal. You know, what is people's personal relationships with food? How does this play out on the ground? We're working with the graduate student Angelica Valdez Valderrama here at the Friedman School, thinking about what some of the cultural assumptions and of the food group constraints in some of these models are. If you sort of came from a different immigrant tradition or if you came from another community, what things would be different in, for example, decisions about what's called the Mediterranean diet or what's called the healthy US style dietary pattern. How much difference do this sort of breadth, cultural breadth of dietary patterns you could consider, how much difference does that make in terms of what's the outcome of this type of hypothetical diet?

Will: And I think, you know, from the global perspective, one really interesting thing is when we do combine data sets and look across these very different cultural settings, dry land, Sahelian Africa versus countries that are coastal versus sort of forest inland countries versus all across Asia, south Asia to East Asia, all across Latin America. We do see the role of these cultural factors. And we see them playing out in very systematic ways that people come to their cultural norms for very good reasons. And then pivot and switch away to new cultural norms. You know, American fast food, for example, switching from beef primarily to chicken primarily. That sort of thing becomes very visible in a matter of years. So, in terms of things that are frontiers for us, remember this is early days. Getting many more nutritionists, people in other fields, looking at first of all, it's just what is really needed for health. Getting those health requirements improved and understood better is a key priority. Our Healthy Diet Basket comes from the work of a nutritionist named Anna Herforth, who has gone around the world studying these dietary guidelines internationally. We're about to get the Eat Lancet dietary recommendations announced, and it'll be very interesting to see how those evolve.

Second thing is much better data on prices and computing these diets for more different settings at different times, different locations. Settings that are inner city United States versus very rural. And then this question of comparing to actual diets. And just trying to understand what people are seeking when they choose foods that are clearly not these benchmark least cost items. The purpose is to ask how far away and why and how are they far away? And particularly to understand to what degree are these attributes of the foods themselves: the convenience of the packaging, the preparation of the item, the taste, the flavor, the cultural significance of it. To what degree are we looking at the result of aspirations that are really shaped by marketing. Are really shaped by the fire hose of persuasion that companies are investing in every day. And very strategically and constantly iterating to the best possible spokesperson, the best possible ad campaign. Combining billboards and radio and television such that you're surrounded by this. And when you drive down the street and when you walk into the supermarket, there is no greater effort on the planet than the effort to sell us a particular brand of food. Food companies are basically marketing companies attached to a manufacturing facility, and they are spending much more than the entire combined budget of the NIH and CDC, et cetera, to persuade us to eat what we ultimately choose. And we really don't know to what degree it's the actual factors in the food itself versus the marketing campaigns and the way they've evolved. You know, if you had a choice between taking the food system and regulating it the way we regulate, say housing or vehicles. If we were to say your supermarket should be like an auto dealership, right? So, anything in the auto dealership is very heavily regulated. Everything from the paint to where the gear shift is to how the windows work. Everything is heavily regulated because the auto industry has worked with National Transportation Safety Board and every single crash investigation, et cetera, has led to the standards that we have now. We didn't get taxes on cars without airbags to make us choose cars with airbags. They're just required. And same is true for housing, right? You can't just build, you know, an extension deck behind your house any way you want. A city inspector will force you to tear it out if you haven't built it to code. So, you know, we could regulate the grocery store like we do that. It's not going to happen politically but compare that option to treating groceries the way we used to treat the legal services or pharmaceuticals. Which is you couldn't advertise them. You could sell them, and people would choose based on the actual merit of the lawyer or the pharmaceutical, right? Which would have the bigger impact. Right? If there was zero food advertising, you just walked into the grocery store and chose what you liked. Or you regulate the grocery store the same way we regulate automotive or building trades. Obviously, they both matter. There's, you know, this problem that you can't see, taste or smell the healthiness of food. You're always acting on belief and not a fact when you choose something that you're seeking health. We don't know to what extent choice is distorted away from a low-cost healthy diet by things people genuinely want and need. Such as taste, convenience, culture, and so forth. Versus things that they've been persuaded to want. And there's obviously some of both. All of these things matter. But I'm hopeful that through these least cost diets, we can identify that low-cost options are there. And you could feed your family a very healthy diet at the Thrifty Food Plan level in the United States, or even lower. It would take time, it would take attention, it would be hard. You can take some shortcuts to make that within your time budget, right? And the planning budget. And we can identify what those look like thanks to these model diets.

It's a very exciting area of work, but we still have a lot to do to define carefully what are the constraints. What are the real objectives here. And how to go about helping people, acquire these foods that we now know are there within a short commuting distance. You may need to take the bus, you may need carpool. But that's what people actually do to go grocery shopping. And when they get there, we can help people to choose items that would genuinely meet their needs at lower cost.

Bios

Will Masters is a Professor in the Friedman School of Nutrition, with a secondary appointment in Tufts University's Department of Economics. He is coauthor of the new textbook on Food Economics: Agriculture, Nutrition and Health (Palgrave Macmillan, 2024). Before coming to Tufts in 2010 he was a faculty member in Agricultural Economics at Purdue University (1991-2010), and also at the University of Zimbabwe (1989-90), Harvard's Kennedy School of Government (2000) and Columbia University (2003-04). He is former editor-in-chief of the journal Agricultural Economics (2006-2011), and an elected Fellow of the American Society for Nutrition (FASN) as well as a Fellow of the Agricultural and Applied Economics Association (AAEA). At Tufts his courses on economics of agriculture, food and nutrition were recognized with student-nominated, University-wide teaching awards in 2019 and 2022, and he leads over a million dollars annually in externally funded research including work on the Agriculture, Nutrition and Health Academy (https://www.anh-academy.org), as well as projects supporting government efforts to calculate the cost and affordability of healthy diets worldwide and work with private enterprises on data analytics for food markets in Africa.

Parke Wilde (PhD, Cornell) is a food economist and professor at the Friedman School of Nutrition Science and Policy at Tufts University. Previously, he worked for USDA’s Economic Research Service. At Tufts, Parke teaches graduate-level courses in statistics, U.S. food policy, and climate change. His research addresses the economics of U.S. food and nutrition policy, including federal nutrition assistance programs. He was Director of Design for the SNAP Healthy Incentives Pilot (HIP) evaluation. He has been a member of the National Academy of Medicine’s Food Forum and is on the scientific and technical advisory committee for Menus of Change, an initiative to advance the health and sustainability of the restaurant industry. He directs the USDA-funded Research Innovation and Development Grants in Economics (RIDGE) Partnership. He received the AAEA Distinguished Quality of Communication Award for his textbook, Food Policy in the United States: An Introduction (Routledge/Earthscan), whose third edition was released in April 2025.