Testing parametric models in the presence of instrumental variables
In this note we develop tests for checking the hypothesis whether a regression function, which is identified via instrumental variables, belongs to some parametric family of functions. We show that this testing problem can essentially be reduced to testing whether a regression function in an ordinary nonparametric regression model vanishes. The asymptotic distribution theory of two test statistics is developed, and their power properties are investigated in a simulation study.
Year of publication: |
2008
|
---|---|
Authors: | Holzmann, Hajo |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 78.2008, 6, p. 629-636
|
Publisher: |
Elsevier |
Keywords: | Instrumental variables Inverse estimation Lack-of-fit tests Nonparametric regression Statistical inverse problem |
Saved in:
Saved in favorites
Similar items by person
-
Distribution Dynamics of Regional GDP per Employee in Unified Germany
Vollmer, Sebastian, (2010)
-
Vollmer, Sebastian, (2010)
-
Perspectives on the world income distribution: beyond twin peaks towards welfare conclusions
Holzmann, Hajo, (2007)
- More ...