The Food Stamp Program and birth weights

With the 50th anniversary of The War on Poverty, there has been a lot of discussion reflection on how well it worked and what the U.S. should do about poverty going forward. So I wanted to draw attention to a 2011 article, Inside the War on Poverty: The Impact of Food Stamps on Birth Outcomes by Douglas Almond, Hilary W. Hoynes, and Diane Whitmore Schanzenbach.

Abstract—This paper evaluates the health impacts of a signature initiative of the War on Poverty: the introduction of the modern Food Stamp Program (FSP). Using variation in the month FSP began operating in each U.S. county, we find that pregnancies exposed to FSP three months prior to birth yielded deliveries with increased birth weight, with the largest gains at the lowest birth weights. We also find small but statistically insignificant improvements in neonatal mortality. We conclude that the sizable increase in income from FSP improved birth outcomes for both whites and African Americans, with larger impacts for African American mothers.

One of my goals in starting this blog is to translate social science knowledge into language people without specialized training can understand. This article is good practice.

Figuring out the impact of a program like the Food Stamp Program (recently renamed SNAP) is challenging because people opt-in to treatment. That is, they choose to sign up for the program.

If we were trying to determine the impact of a new cholesterol drug we could get a group of people with high cholesterol, randomly divide them into two smaller groups, and give one the new drug and the other a placebo. If the group getting the new drug has better results and this differences is large enough, we can conclude that the drug is effective. Many readers will already be familiar with this. It’s called a randomized controlled trial (RCT).

(There are examples of this approach with anti-poverty interventions, often when a government designs rollout in a way that allows incorporates this kind of research. That’s a whole other topic, but I just wanted to briefly note it.)

So, the issue is that we generally have no good group to which we can compare people getting food stamps. If we compare them to the overall population and find out that they are more likely to be hungry (whatever measure related to hunger we use) the obvious answer is: “Well, I’d hope so. That’s who the program is for!” Plus to qualify they have to be low income and thus more vulnerable to hunger.

So we could compare them to people with incomes low enough to get food stamps but not getting them. Nope (or at least not as a simple comparison). By virtue of opting into the program they are clearly different than people who could have signed up but didn’t. So there’s a lot of research that tries to deal with this problem, such of which can be found on the U.S. Department of Agriculture (USDA) website.

I really like this article’s approach. The program arrived at different counties at different points in time. If it really was leading to better nutrition among some of the poorest people in the U.S., we should expect that mothers who were pregnant when the program started would eat better and therefore be less likely to give birth to underweight babies than those a few months before them when the program hadn’t arrived yet. The authors show that this was indeed the case. Well done.