Empirical Analysis of Dynamic Decision Contexts

This course will focus on empirical tools for analyzing dynamic decision contexts, wherein current actions of firms or consumers have effects on future payoffs, profits and/or competitive conduct. The course will build the relevant material generally, but our applications will be mostly focused on empirical marketing, operations and industrial organization problems. We will have an applied focus overall, emphasizing the practical aspects of implementation, especially programming. The overall aim of the class is to help students obtain the skills to implement these methods in their research. By the end of the class, students are expected to be able to formulate a dynamic decision problem, program it in a language such as Matlab or C, and to estimate the model from data. The course starts with discrete choice markovian decision problems, and continuous markovian decision problems, and focus on building the computational toolkit for the numerical analysis of these problems. We then move on to specific applications, and discuss multi-agent dynamic equilibrium models. Finally, we discuss recently proposed advanced methods for alleviating computational burden in dynamic models.