Marketing

This class is intended to provide PhD students with an advanced treatment of causal inference as well as a gentle introduction to frontier machine learning techniques that are useful in economic applications. From the machine learning toolkit, we will cover Lasso and Random Forests in detail, along with recent approaches to inference with HD nuisance parameters. Deep Learning will be covered only briefly. R packages for implementation of Lasso and Forests will be introduced. We will then consider the three canonical approaches to causal inference: the Rubin framework, the Pearl framework and the Heckman synthesis. Randomized control trials and associated methods of inference (Fisher exact p-values and randomization inference) will be covered briefly. We will then consider recent approaches based on synthetic control and matrix completion. We next turn to methods based on selective choice, including Roy models and control functions. Finally, we will consider some issues related to IV approaches, including problems of weak or many instruments (and a Lasso-based solution) and the local/selected nature of resulting estimands (LATE and MTE).

This course introduces students to canonical modeling approaches for analyzing decision making by both firms and consumers, focusing on static environments. Central topics include demand estimation, models of strategic interaction, networks and platforms and auctions. Applications include firm pricing decisions, new product introductions, strategic entry and vertical relationships. The course generally includes coding assignments and student presentations, in addition to the weekly lectures on methods and applications.

This course examines consumer and firm behaviors that involve inter-temporal trade-offs and as a result involve dynamic optimization on the part of both consumers and firms. It begins with an overview of dynamic programming methods, in both single and multi-agent settings, emphasizing methods that link estimation with computation. Single agent topics include models of capital replacement, dynamic demand, inventory models and salesforce management. Multi-agent topics include strategic innovation, learning by doing, demand smoothing, and product repositioning. A strong emphasis is placed on recent methods and frontier topics. The course generally includes coding assignments and several student presentations, in addition to weekly lectures.

3 credits

This workshop provides a forum for the presentation of research ideas and completed research by students. The course includes discussion of current job market papers and job market presentations, journal reviewing, and generating new research ideas. In addition, some topics are covered to illustrate current research areas of interest for the faculty. All marketing PhD students who are not on the job market are expected to participate
actively.

Prerequisite: Permission of the instructor

This course is designed for first, second, and third year students to provide exposure to the literature related to core research methods used in quantitative marketing research and to build student appreciation of what goes into conducting research in quantitative marketing. The content of the course varies by year with similar core topics, but rotating papers, and some rotating topics. The core topics include choice models, aggregate demand models, Bayesian models, consumer heterogeneity, and state dependence. Rotating topics have included structural model identification, experimentation, causal inference, search, learning, advertising effects, and conjoint analysis, and rotate each year. In addition, the course provides exposures to other perspectives on quantitative research. Course evaluation includes coding assignments and homework’s, and the final exam has both coding and conceptual parts to it. In addition, students are expected to submit a paper and present their own research.

Prerequisite: Permission of the instructor

This course covers advanced topics in quantitative marketing research. The topics rotate each time offered and are selected based on current topical areas in the marketing, economics, and related fields as well as student and faculty research interests. Part of the evaluation in the course is to submit a paper and present their own research. All marketing PhD students who are not on the job market are expected to participate actively.

Prerequisite: Permission of the instructor