Press "Enter" to skip to content

Variations on the Thrift Food Plan: Model diets that satisfy cost and nutrition constraints

Published: January 2025
Bibliographic reference: Yiwen Zhao, Linlin Fan, Norbert L.W. Wilson, Angelica Valdes Valderrama, Parke Wilde. Variations on the Thrift Food Plan: Model diets that satisfy cost and nutrition constraints. Food Policy 130 (2025) 102781

Download Paper

Abstract

The Supplemental Nutrition Assistance Program (SNAP) supports Americans with low incomes in acquiring adequate and healthful diets. The maximum SNAP benefit is based on the Thrifty Food Plan (TFP), the lowest cost of four U.S. Department of Agriculture (USDA) food plans. This paper uses optimization models and data replicating those used to reevaluate the TFP in August 2021. The optimization models solve for a food plan that is as similar as possible to the national average diet of healthy-eating Americans, while meeting nutrition requirements and cost constraints. This study’s objective was to investigate which model components are most important in driving the results and explore economic tradeoffs between food costs, nutrition quality, and consumer preferences in the U.S. food marketplace. The results showed that model food plans differed greatly from current consumption, with only 29 of 97 food categories being selected. The TFP algorithm was driven
primarily by the cost and food group constraints rather than the objective function. The constraints with the highest Lagrangian semi-elasticities were, in order: the cost constraint, a food energy constraint, a vitamin E constraint, and particular food group constraints such as dairy. The implications for recommended SNAP benefit amounts depend on which constraints are used and on how much difference between the model diet and current consumption is considered acceptable. Relaxing certain food group constraints, such as dairy constraints, for nutrition goals would permit a lower cost target, while seeking model food plans more similar to current consumption would require a higher cost target.

This work was supported by USDA NIFA #2021–67023-34479 and Hatch Appropriations under Project #PEN04964 and Accession #7006590. Our funding sources were not involved in the study design, collection, analysis, or interpretation of the data. The views expressed in this paper are of the authors and should not be attributed to USDA.

Research Findings