授课内容:计量经济学
授 课 人:加拿大多伦多大学计量经济学领域中青年学者宛圆渊教授
授课时间:2019年12月10日—14日
授课地点:舜耕办公楼816
主办单位:yl8cc永利
授课对象:经济学研究生
具体时间安排:
12月10日 周二上午,1-4节
12月11日 周三下午,5-8节
12月12日 周四下午,5-8节
12月13日 周五上午,1-4节
This econometrics seminar course will give a brief introduction to literature of partial identification. In particular, I will use examples from the program evaluation and the empirical industrial organization literature to show how the paradigm of partial identification is useful in relaxing modeling assumptions and providing more reliable inference. Other examples will be covered if time permits. After this seminar series, students are expected to understand the ideology of the partial identification and how it can be applied to empirical work.
Part I: Basics of partial identification
Introduction of the idea of partial identification
The development of partial identification literature in recent years
Advantages and challenges
Part II: Program evaluation under partial identification
Introduction of popular casual inference methods
The marginal treatment effect (MTE) framework
Validity of the identification assumptions
Partial identification of MTE under weaker assumptions
Part III: Estimating oligopoly entry models under partial identification
Two-by-two discrete choice games for modeling market entry competition
Parametric vs Semi-parametric model
Identification of the strategic interaction under unknown profit shock distribution
Part IV: Other examples of partial identification
Incomplete English-auction model
Partial identification in dynamic panel data model
Interval observed data