Econometrics and Applied Economics workshops: Adam Rosen
| Dates: | 12 March 2015 |
|---|
| Times: | 16:15 - 17:45 |
|---|
| What is it: | Seminar |
|---|
| Organiser: | School of Social Sciences |
|---|
| Speaker: | Adam Rosen |
|---|
|
“Generalized Instrumental Variable Models, with Application to the Analysis of Treatment Effects”
Abstract: In this talk I will present an overview of joint work with Andrew Chesher on Generalized Instrumental Variables (GIV) models, as studied in Chesher and Rosen (2014). In contrast to traditional IV models, GIV models do not require the mapping from unobservable variables to endogenous variables to admit a unique inverse. This broadens the scope for application of instrumental variable methods to models allowing general forms of unobserved heterogeneity. Unobserved variables may be multivariate, and may enter structural functions nonseparably, thereby removing strong practical limitations on the role of unobserved heterogeneity. Examples include models with discrete or mixed continuous/discrete outcomes and continuous unobservables, random coefficient models, and auction models. To demonstrate the application of our results, I will expand on material from Chesher and Rosen (2013), which employs an IV model for binary response. We characterize sharp bounds on average response functions and average treatment effects (ATE), and apply the model to data on the effect of family size on female employment using data from Angrist and Evans (1998).
For more information: http://www.socialsciences.manchester.ac.uk/subjects/economics/our-research/workshop-and-seminars/econometrics-and-applied-economics/
Speaker
Adam Rosen
Role: Reader
Organisation: UCL
Travel and Contact Information
Find event
University Place 6.206
University Place
Manchester
