Healthy food is cheaper than junk food, except by the only metric that matters

The USDA Economic Research Service published a report in 2012 called Are Healthy Foods Really More Expensive? It Depends on How You Measure the Price.

The aim of the research was to compare the prices of healthy foods with unhealthy foods using a variety of metrics.

For the purposes of the report, “healthy foods” were defined as foods that:

Contain an amount of a food in at least one of the major food groups (vegetables, fruits, grains, dairy, and protein foods) equal to at least half the portion size that the Dietary Guidelines for Americans 2010 uses for measuring the nutrients in that food, and contain only moderate amounts of saturated fats, added sugars, and sodium.

Thus, “healthy foods” encapsulates not only fresh fruits and vegetables but also milk, fruit juice, whole-grains and even products like white bread, which would fall into the “enriched grains” subcategory. The “vegetable” category includes starchy foods like potatoes. Everything that doesn’t come under “healthy foods” counts as unhealthy food, or, in the language of the report, “moderation foods”.

I am naturally suspicious of any study that categorizes white bread or potatoes as healthy foods, but what really grabbed my attention was the report’s editorial musings on the value of the different metrics used to measure price.

The authors computed price in three different ways: price-per-calorie, price-per-edible-gram, and price-per-average-portion.

Readers of my previous post comparing the fiber content of grains, fruits and vegetables might already have discerned that price-per-calorie is the metric I’m most interested in. In my opinion, caloric budget is the most basic and essential dietary determiner for nutritionally-minded people. (It’s certainly more relevant and immutable than the number of grams of food one eats in a day or the number of average portions one eats in a day.)

But in many places the report takes a strong stance against price-per-calorie as a measure of food affordability. They outline their objections in three main points, each of which I’ll deal with in turn.

1. When less healthy foods are defined as energy-dense (a higher number of calories per edible gram), the metric suffers from mathematical coupling or negative autocorrelation. This occurs because the same variable—calories—appears in the numerator of the energy density (calories per edible gram) calculation as well as in the denominator of the price metric (price per calorie). Thus, high-energy-dense foods tend to have a low price per calorie because the price is divided by a large number of calories, while low energy-dense foods tend to have a high price per calorie because the price is divided by a small number of calories.

The most important part of this argument is the opening conditional statement: “When less healthy foods are defined as energy-dense”. The authors argue that it is tautological to say healthy foods are expensive per calorie if all healthy foods are calorie-sparse. In other words, it is automatically true that foods without many calories will cost a lot per calorie (in the general case).

But does anyone actually define healthy and unhealthy foods based on energy density? Oily fish is relatively energy-dense, but is regarded by most as healthy food. Sugar-free energy drinks are extremely energy-sparse, but are regarded by most as an unhealthy food. Moreover, the report itself doesn’t use energy density as a measure of the healthiness of food at all.

2. Since the price per calorie does not account for the amount of food consumed, it is not a good proxy for out-of-pocket food costs. As Frazão et al. (2011) explain, a gallon of skim milk has about half as many calories as a gallon of whole milk. Thus, the price per calorie is nearly twice as much for skim milk as for whole milk. Yet consumers often pay the same out-of-pocket cost for a gallon of milk, regardless of whether they buy skim or whole milk. This is because the price per calorie metric ignores the total costs associated with the total number of calories consumed. Frazão (2009) confirms this weakness of the price-per-calorie metric using data from Townsend et al. (2009) to show that the higher cost per calorie of the healthier, less energy dense diets does not translate into higher total daily food costs after accounting for the higher number of total calories in the less healthy diets.

To paraphrase this argument, “Price per calorie isn’t very important because it ignores the fact that many people don’t tailor the amount of food they eat to the number of calories they need”. Or even more basely, “Healthy food is more expensive, but people who eat unhealthy food eat more of it, so don’t spend any less on food”. While the proposition that most people don’t pay much attention to their intake of calories is undoubtedly true, it is also a dire indictment of the state of nutritional knowledge in the general population. People who drink whole milk may not adjust the rest of their diet to account for the extra calories compared to skim, but they undoubtedly should.

Thus, the argument used in the report is not invalid, but it is irrelevant to people who have the time, education and other resources to optimize their own diet for health. Imagine an athlete who closely monitors their intake of calories and drinks a glass of whole milk every morning. If they switch to skim milk, the cost is relatively more expensive than for whole milk, as either i) they consume more of it to maintain their specific calorie intake, or ii) they consume the same amount, and make up the calorie difference using a different food which itself will have non-zero costs.

3. The price-per-calorie metric is inconsistent with low-calorie marketing claims. Lipsky et al. (2011) raise the point that if consumers used the price of food energy in their decision making, then no manufacturer would want to advertise that a product had fewer calories, since that would result in a higher price. As it is, many foods carry a claim of fewer calories.

This argument seems like rudimentary economic reasoning applied naively to a real-world scenario. It assumes all consumers are rational, price-sensitive etc.

It’s obvious that people who buy sugar-free soda or chips made with Olestra are not doing so out of a primary consideration of the cost of food energy but rather from a desire for a pleasurable sensory experience. However, that doesn’t change the fact that such foods are tremendously expensive from a nutritional perspective in they they cost a relatively large amount of money and provide very little that the body actually needs.

The more general point is that while consumers may not on the whole be motivated by the price-per-calorie of foods, they are still constrained by it, in as much as they need to reach a certain number of calories per day.

In sum, the arguments used to disparage price-per-calorie as a metric of food cost seem to focus on observation of current, often unhealthy food choice behaviors. I am far more interested in tailoring discussion and advice towards people engaged with their diets and committed to finding healthier ways to eat.

Perhaps the reason that the report spends so much time attempting to preemptively torpedo price-per-calorie as a measure of affordability is that the results of the study show healthy foods are more expensive. Foods in the “Vegetable” category are more expensive per calorie than foods in every other category, and foods in the “Fruit” category are more expensive per calorie than foods in every other category bar vegetables. When interpreting those results, bear in mind that the “Vegetable” category includes starchy foods like potato and sweet corn.

In other words, the two healthiest groups of food are the most expensive on a per-calorie basis. The cheapest group was — unsurprisingly — grains.

In closing, here’s an image from the report that shows 100 calories worth of different foods:

The image and caption seem to suggest it’s ridiculous to compare two dozen strawberries with a small handful of  M&Ms, but my takeaway is that it’s unfortunate that the strawberries would be so vastly more expensive than the candy or potato chips. In fact, this image plays into the kind of pervasive portion-based (rather than calorie-based) thinking that actually suppresses sensible eating choices. I can imagine some of my friends reaction to seeing a portion of strawberries like that: “You can’t possibly eat all that! How can you eat so much and not get fat?”

My submission is, if we could get rid of that kind of thinking about food, we’d all eat a little bit healthier.