Identification and estimation of semiparametric two-step models
Let H<sub>0</sub>(X) be a function that can be nonparametrically estimated. Suppose E [Y&7CX]=F<sub>0</sub>[X⊤β<sub>0</sub>, H<sub>0</sub>(X)]. Many models fit this framework, including latent index models with an endogenous regressor and nonlinear models with sample selection. We show that the vector β<sub>0</sub> and unknown function F<sub>0</sub> are generally point identified without exclusion restrictions or instruments, in contrast to the usual assumption that identification without instruments requires fully specified functional forms. We propose an estimator with asymptotic properties allowing for data dependent bandwidths and random trimming. A Monte Carlo experiment and an empirical application to migration decisions are also included.
Year of publication: |
2016
|
---|---|
Authors: | Escanciano, Juan Carlos ; Jacho-Chávez, David ; Lewbel, Arthur |
Published in: |
Quantitative Economics. - New Haven, CT : The Econometric Society, ISSN 1759-7331. - Vol. 7.2016, 2, p. 561-589
|
Publisher: |
New Haven, CT : The Econometric Society |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Identification and estimation of semiparametric two-step models
Escanciano, Juan Carlos, (2016)
-
Identification and estimation of semiparametric two-step models
Escanciano, Juan Carlos, (2016)
-
Identification and estimation of semiparametric two step models
Escanciano, Juan Carlos, (2010)
- More ...