Nudging Nutrition: How the Framing of Relative Price Changes as Taxes or Subsidies Affects Grocery Purchases in a Seven-Month Field Study

The rapid rise in obesity in the U.S., and the medical care costs it entails, has led to calls for governments to incentivize healthy diets.  For example, the Institute of Medicine, International Obesity Task Force, and editorials in the New England Journal of Medicine and the American Journal of Public Health have called for the U.S. to tax energy-dense foods.  However, it is not well known what effects such taxes would have.  A large literature has used naturally-occurring variation in prices to estimate the price elasticities of demand for certain categories of food, but such studies are hindered by the fact that existing taxes on energy-dense foods are very modest.

One important unresolved question is whether consumers’ food purchases are sensitive to how one frames the change in relative prices; i.e. whether one taxes non-nutritious food or subsidizes nutritious food.  A substantial literature in behavioral economics has documented loss aversion - people’s behavior can be responsive to whether a price change is framed as a reward for beneficial behavior or a penalty for undesirable behavior. 

This paper uses data from a randomized controlled field experiment to estimate the impact of changes in the relative prices of nutritious and non-nutritious food, and to test whether consumers respond to the framing of a price change; i.e. are they more responsive when it is presented as a tax on non-nutritious food, a subsidy for nutritious food, or both? 

The field experiment involved 207 households at a supermarket chain in the Northeast U.S. from 2010-2011.  There were three treatment groups, each of which was treated with a 10% discount of nutritious foods relative to non-nutritious foods, but for one it was framed as a tax on non-nutritious foods, for the second it was framed as a subsidy of nutritious food, and for the third it was framed as both.  Whether a food was nutritious was indicated by a shelf-label nutrition guidance system already in place in the supermarket.  The control group received a discount on all grocery purchases to compensate them for participation, but their discount did not change the relative prices of nutritious and non-nutritious foods. 

We estimate difference-in-differences models that compare the change before and after the price change in the treatment groups relative to the control group.  There was one month of grocery shopping data collected prior to the treatment, and six months of data collected during the treatment.  We present estimates of the effect of both the relative price change and how that price change is framed.  The results contribute to the literature on policies to prevent and reduce obesity and the literature on behavioral public finance.