Optimal Interventions for Increasing Healthy Food Consumption Among Low Income Households

The federal government currently spends over $100 billion per year on policies aimed to increase fruit and vegetable (FV) consumption among low income households. These include price-, nutrition education-, and access-related interventions. Currently, the government allocates funds to each type of intervention in an ad- hoc fashion, often resulting in surprisingly disappointing outcomes. For example, access-related interventions have seen mixed results in many case studies, resulting in debate about the importance of supply-side interventions. This paper introduces a novel consumer behavioral model for grocery shopping dynamics, which is nested into a bi-level model for optimizing the governments investments. The government’s goal is to increase fruit and vegetable (FV), or more generally healthy food, consumption among low income households by utilizing strategic portfolios of interventions. Based on primitives estimated from data, the model agrees with known empirical evidence and suggests several new policy insights, for example that access- related interventions are only beneficial under specific conditions which depend heavily on the consumer’s value of nutrition. In a setting where the government cannot provide completely personalized interventions, it is found that high levels of FV consumption by can be achieved by subsetting consumers based on key characteristics and deploying smart subpopulation-level strategies.