Xun LiFollow

Date of Completion


Embargo Period



Advertising; Food Environment; Obesity; Energy Shock; Food Price; Random Coefficient Logit Model

Major Advisor

Rigoberto A. Lopez

Associate Advisor

Yizao Liu

Associate Advisor

Chad Cotti

Field of Study

Agricultural and Resource Economics


Doctor of Philosophy

Open Access

Open Access


This dissertation consists of three empirical studies of food marketing, which directly and indirectly affects consumers’ purchase behavior, reshapes their eating habit, and alters social welfare. Chapter one investigates the advertising spillover effects using the carbonated soft drinks market as a case study. In this chapter, spillover effects are modeled using the conventional linear and constant elasticity of substitution (CES) advertising production functions. Empirical results confirm strong and positive brand advertising spillover effects across brands belonging to the same company as well as negative spillover effects from advertising by competitors. Empirical results also indicate that the CES advertising production function outperforms the linear function, providing strong support for decreasing returns to scale in advertising and imperfect substitution between brand advertising and advertising of other brands in the same company. Finally, the CES function results in significantly higher estimates of the price elasticities of demand as well as lower estimated markups.

Chapter two applies a stochastic frontier approach to rigorously ascertain the effects of food environment components on obesity outcomes. Using individual consumer data and food environment data from New England counties, empirical results indicate that supercenters and limited service restaurants are positively associated with weight outcomes, while fruit and vegetable stores and full-service restaurants are negatively linked to weight gain. In metropolitan counties, however, food environment factors that affecting weight outcomes are full-service restaurants and limited-service restaurants. In non-metropolitan counties, food environment components affect weight outcomes significantly only in counties adjacent to a metropolitan area. In counties that are not adjacent to a metropolitan area or which are completely rural, the associations between food environment components and weight outcome are consistently weak.

Chapter three contributes to the debate on food and energy prices by examining the relationship between milk and diesel prices and price pass-through. Empirical results indicate that energy price (e.g., diesel price) significantly impact the prices of milk products. The pass-through rates are around 0.6227. More interestingly, most of private labels have the lowest energy (diesel) pass-through rates, implying that comparing to other products, private labels are more invulnerable to energy price shocks.