Author ORCID Identifier

Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation


Agriculture, Food and Environment


Agricultural Economics

First Advisor

Dr. Yuqing Zheng

Second Advisor

Dr. Steven Buck


Grocery sales taxes represent a stable tax revenue stream for state and municipal government, but there is rare empirical evidence suggesting grocery taxes may adversely affect health. In addition, the how governments set grocery sales taxes is still unclear. Therefore, based on a novel national dataset of annual county and state-level grocery taxes from 2009 through 2016, the following three essays in the dissertation investigate the health impacts of grocery sales taxes and the causes of grocery sales taxes in a framework of tax competition.

In the first essay, we document the spatial and temporal variation in grocery taxes and empirically examine the statistical relationship between county-level grocery taxes and obesity and diabetes. We link the tax data to three-year, county-level estimates based on data from the Centers for Disease Control and Prevention on rates of obesity and diabetes and provide a nation-wide spatial characterization of grocery taxes and these two health outcomes. Using a county-level fixed effects estimator, we estimate the effect of grocery taxes on obesity and diabetes rates, also controlling for a subset of potential confounders that vary over time. We find a one percentage point increase in grocery taxes is associated with 0.588 and 0.215 percentage point increases in the county-level obesity and diabetes rates. In conclusion, Counties with grocery taxes have increased prevalence of obesity and diabetes. We estimate the economic burden of increased obesity and diabetes rates resulting from grocery taxes to be $5.9 billion. Based on this estimate, the benefit-cost ratio of removing grocery taxes is 1.90 across the United States if we only consider the effects on obesity and diabetes rates.

In the second essay, we aim to examine whether grocery sales taxes make significant impacts on individual’s body weight outcome. We merged the county-level grocery tax data with the individual longitudinal data from Panel Study Income Dynamics (PSID) and explore a fixed effect model to estimate the causal impact of grocery sales taxes on family food expenditures and individual BMI (Body Mass Index). After that, we conduct the analysis of heterogeneous effects by income category and obesity level to identify the policy impacts on different individuals and families. We find that a ten point-percentage increase in grocery sales rate leads to a rise of BMI by 0.61 (which roughly translates to a body weight gain of 1.68kg). The results are more significant for the overweight population whose BMI is greater than 25 but smaller than 30. We do not find significant results towards different income population.

In the third essay, we study the state-county tax policy interaction patterns and explore the causes of grocery sales tax changes considering spatial externalities under a Stackelberg tax competition model with three propositions. Derivatized from the model, county grocery tax rates are affected by states’ grocery tax rates (vertical effects), neighboring counties’ grocery tax rates (horizontal effects) and neighboring states’ grocery tax rates (diagonal effects). By employing the twelve-year data of state and county grocery taxes, we also empirically examine the three propositions in a spatial autoregressive model. The empirical results are consistent with the three theoretical proportions. The average county grocery sales tax rate is less than the average state grocery sales tax rate, and we find the county grocery tax rate changes negatively with its domestic state grocery sales tax rate. Neighboring counties play a large role in determining the local county grocery tax rates. For example, a county will increase its grocery tax rates by 0.780-point percentages if its neighboring county increases one percentage point tax rate on average. Neighboring state tax rates can also positively affect a county’s grocery tax rate. A county is expected to increase its grocery tax rate by 0.110 percentage point when its neighboring states increase state grocery tax rate by one percentage point on average.

Digital Object Identifier (DOI)