A Framework for Empirical Models of Dynamic Demand∗

This paper provides a framework for modeling dynamic demand. A recent literature estimates the dynamics of consumer demand in a variety of different contexts, including durable goods, storable goods, and subscription services. To make these estimation problems tractable, researchers typically restrict consumer expectations using an Inclusive Value Sufficiency (IVS) assumption. Some versions of IVS make assumptions on flow utilities and others on continuation values. A flow utility assumption is more palatable because flow utilities map closely to fundamentals. However, we show that a flow utility assumption requires a further assumption that there are sets of products with common continuation values. We argue that these two sets of assumptions result in different economic implications and hence are suitable in different contexts. We also show that most models employ an assumption on the separability of utility from holdings from utility for new purchases, and we discuss this restriction. Finally, we categorize the assumptions made by various empirical models within our unified framework. ∗We wish to thank Igal Hendel, Robin Lee, Aviv Nevo, Pasquale Schiraldi and Philipp Schmidt-Dengler and participants at the 2013 Demand Estimation and Modeling Conference for helpful discussions on these topics. NSF Grants SES-0551348 and SES-0922540 provided support.