Structural Estimation

Structural Estimation (MACS 40200), University of Chicago, M.A. Program in Computational Social Science: Winter 2017, Winter 2018

The purpose of this course is to give students experience estimating parameters of structural models. We define the respective differences, strengths, and weaknesses of structural modeling and estimation versus reduced form modeling and estimation. We will focus on structural estimation. Methods will include taking parameters from other studies (weak calibration), estimating parameters to match moments from the data (GMM, strong calibration), simulating the model to match moments from the data (SMM, indirect inference), maximum likelihood estimation of parameters, and questions of model uncertainty and robustness. The course focuses on both obtaining point estimates as well as getting an estimate of the variance-covariance matrix of the point estimates. All the syllabus, references, and assignments for the course are available in the GitHub repository for the Winter 2018 section of MACS 40200.

Some of the examples in the course will come from economics, but the material will be presented in a general way in order to allow students to apply the methods to estimating structural model parameters in any field. We focus on computing solutions to estimation problems. We also study results and uses from recent important structural estimation papers.

One of the most attractive characteristics of this course is that just over one third of the points in the course come from a written paper structural estimation research project and a presentation of that project. Students get hands-on experience executing a structural estimation to answer a research question. Most students extend a topic of a recent paper, and many students use this class to add structural estimation to their MA thesis. This course was popular in Winter 2018 among MAPSS students specifically because it included a writing component. The writing portion of this class adds a dimension of learning that is often missing from a rigorous methods course.