There are good reasons to criticize the US food stamps program. The fact that some poor people might use food stamps to buy “junk food” is not one of them.
I make this point in response to a new article by Chris Edwards at the Cato Institute. In the article, Edwards cites a recent study by the US Department of Agriculture (USDA) which analyzed the purchasing habits of some food stamp recipients. Among other things, the study showed that 22.6% of the food stamp purchases were of junk food. Edwards then extrapolates this figure to the broader food stamps program and estimates that some $15 billion a year in taxpayer money is spent on junk food through the food stamps program.
On a technical level, there appears to be nothing wrong with Edwards’s estimate here. The categories he used in his definition of junk food appear reasonable, and he was also careful to note that the USDA study’s may not be representative of the whole program since it didn’t rely on a random sample.
The issue lies in the solution this observation suggests–namely, narrowing the types of foods that can be purchased with food stamps.
There are several drawbacks to this proposal. Let’s go through them.
Reducing Choice Reduces Value to Recipients
The first problem is that it would reduce the value of the benefits received. One of the most important insights from economics is that value is subjective–that is, the value I place on oranges and almond milk might be different than the value you place on each of those things. From this, it follows that individuals are in the best position to decide for themselves what they want and need at any given time.
If health isn’t the top priority for them, they might not make healthy decisions. But they will make decisions that they believe will make them content.
If we place restrictions on an individual’s choices, they will likely be worse off than before. We might have restricted the very thing they wanted most, which means they will then be left with a second-best alternative.
This is why many economists believe giving money confers a greater benefit to the recipient than giving in-kind benefits (for example, giving someone food or clothing). If a person has money, they can acquire the thing they want most. That might be clothing, but it might also be something else.
This analysis gets slightly more complicated with food stamps because it’s not regular money, but it’s also not particular foods. It’s somewhere in between. It’s effectively a gift card that can only be used for food, and only at stores that accept it. Food stamp beneficiaries cannot use food stamps to buy alcohol, pet food, or certain other items, but they have considerable flexibility otherwise. From an economics perspective, this makes food stamps less valuable than money (because they’re still confined to food) but more valuable than if the government was just delivering a box of pre-determined food each month.
If government came in and decided to further restrict the types of foods that can be purchased, this would reduce the economic value that food stamps provide to recipients. This would make the recipients worse off because they would have to substitute foods they like less for foods they preferred.
Another downside of more food stamp restrictions is that it requires an unsettling expansion in paternalism by the federal government. Yes, we can all probably agree that eating more fruits and vegetables is better for us than eating a steady diet of soda and Oreos. But do we really want the government setting policy to nudge or force people to eat healthy food?
In this case, of course, it would only apply to a subset of the population (low-income folks), and it would only apply to a subset of their food choices (those paid with food stamps). But it also gets the government more involved in making these determinations. If conservatives go along with a junk food ban on food stamps, it will make it harder for them to take a stand against other related legislation on nutrition, be it a nationwide soda tax or something else. Needless to say, this would not be a positive development.
Confusing the Issue
From a libertarian perspective, the essential problem with the food stamps program is that it requires the redistribution of wealth through taxation. The harm involved here is caused by taxation. How the government happens to spend or give away the money after it takes it might be more or less appalling, but it’s not the key issue.
In the case at hand, there is not a compelling, self-interested reason for a taxpayer to care if the $67B food stamp program was used for soda or fruit; they’re not getting the money back either way.
Edwards suggests that eliminating the ability to purchase junk food on food stamps might reduce demand for food stamps and thereby reduce taxpayer expenditures. While this argument is plausible, it seems unlikely. Implicitly, it assumes that some people are using food stamps to supplement a sweet tooth and would simply not use all of their benefits if junk food was eliminated. That is, poor people would be offered free food and turn it down because it didn’t fit their preferences.
Given that the food stamps program is not terribly lucrative to begin with and is only available to low-income people, this response would be strange. Substituting other foods to still fully utilize their benefit would be a more rational and likely behavior.
Ultimately, eliminating food stamps isn’t a top priority for libertarians, and that’s as it should be. At $67B per year, it amounts to just over 1/10 of what the US spends on the military and war. And unlike spending on war, the worst thing the food stamps program can do is waste money.
But if we’re going to criticize food stamps, our approach should be informed by economics and a healthy skepticism of government paternalism. We should continue to promote the virtues of individuals making choices for themselves, even if those individuals happen to be on welfare and the choices are about how to spend it. Reducing the value of food stamps to recipients without reducing the cost to taxpayers is not in anyone’s interest.