- Xavier D'Haultfoeuille - INSEE-CREST
Course : 18 hours
Objective
This course is an introduction to recent development in semi and nonparametric models in econometrics. Economic models rarely impose parametric restrictions on error terms. Theoretical restrictions rather take the form of exclusion restrictions. We investigate the consequence of relaxing such parametric assumptions, both in terms of identification and estimation. The course presents methods which are to become classical in applied econometrics, such as quantile regressions or single index models, but also general modern ideas such as control functions in IV models or partial identification.
Outline
- Linear models
Quantile regression
- Limited dependent variable models
Binary models: maximum score estimator, single index models.
Censored and truncated models: LAD and trimming estimator.
Selection models : semiparametric two steps estimators
- Instrumental variables in semi or nonparametric models
A benchmark: the linear model.
The “inverse problem” approach.
The control function approach.
- Partially identified models
Examples of partially identified models : missing data, incomplete models.
Inference on parameter sets.
Références
Chernozhukov V., H. Hong and E. Tamer (2007), Estimation and Confidence Regions for Parameter Sets in Econometric Models, Econometrica, 75, 1243-1284.
Horowitz, J. L. (1998), Semiparametric Methods in Econometrics, Springer-Verlag.
Imbens, G. and Newey, W. K. (2009), Identification and Estimation of Triangular Simultaneous Equations Models Without Additivity”, Econometrica, forthcoming.
Koenker, R. (2005) Quantile Regression, Econometric Society Monograph Series, Cambridge University Press.
Manski, C. (2003), Partial Identification of Probability Distributions, Springer.
Wooldridge, J. W. (2002), Econometric Analysis of Cross Section and Panel Data.
|
