Essays on Dynamic Demand Estimation

Thumbnail Image



Journal Title

Journal ISSN

Volume Title

Repository Usage Stats



This dissertation consists of three chapters relating to dynamic demand models of storable goods and their application to taxes that are imposed on soft drinks. Broadly speaking, the first chapter builds the estimation strategy for dynamic demand models of storable goods that allows for unobservable heterogeneous preferences in household's tastes. The second chapter uses the estimation strategy developed in the first chapter to study the policy implications of taxes that are imposed on sugary soft drinks. The last chapter explores and provides an explanation for the level of pass-through for soda taxes.

To be more specific, the first chapter develops techniques for incorporating systematic brand preferences in dynamic demand models of storable goods. Dynamic demand models are important for correctly measuring price elasticities of products that can be stockpiled. However, most of the literature excludes systematic preferences over consumers' brand tastes. This chapter resolves this issue by incorporating random coefficient Logit models into a dynamic demand framework and hence allows for realistic demand substitution patterns. It builds on Hendel and Nevo's 2006 Econometrica paper, where the authors introduce a model of dynamic demand that flexibly incorporates observable heterogeneity and estimates it via a three-step procedure that separates brand and volume choices. While a powerful tool, this method is tricky to execute. Therefore, this chapter also discusses the difficulties that may face implementers.

The second chapter predicts the effects of taxes on sugar sweetened soft drinks (sugar taxes) on both total consumption and the welfare of different types of consumers. It specifies and estimates a structural dynamic demand model of storable goods with rational and forward-looking households. It flexibly incorporates persistent heterogeneous consumer preferences and develops a computationally attractive method for estimating its parameters. Sugar taxes have been proposed at both the national and state-level, and passed in three states, as a means of slowing or reversing the growth in obesity and diabetes. To accurately analyze the effects of these policies, this chapter takes two specific aspects of soft drinks into account: storability and differentiation. It compares the results from this model to two benchmark studies: a static model with consumer heterogeneity and a dynamic model without households' persistent heterogeneous tastes. It finds that failing to account for dynamics (i.e. storability) results in overestimated reduction in consumption and failing to account for persistent heterogeneous preferences (i.e. differentiation) results in overestimated reduction in consumption and underestimated welfare loss. The model and method developed here are readily applicable to many studies involving storable goods, such as firms' optimal pricing behavior and anti-trust policies analyses.

The third and last chapter focuses on the incidence of soda taxes by studying the pass-through level of these taxes. It lays out a framework for thinking about the determinants of the pass-through level. More specifically, it builds theoretical models that examine the pass-through under more complex supply structures with multiple manufactures and retailers. In addition to providing some intuition behind theoretical predictions of the models, this chapter also presents empirical results found in the data along with their implications.






Wang, Yucai Emily (2011). Essays on Dynamic Demand Estimation. Dissertation, Duke University. Retrieved from


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.