Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

Simon Fraser University

$13,000.00

Dimension reduction in specification testing of econometric models

EF 2007-08

Résumé
Nonparametric methods permit estimation of econometric models without strong a priori parametric assumptions. This flexibility comes at a cost, known in the literature as the ``curse of dimensionality."' If the number of considered variables is large, then a huge amount of data is necessary for accurate estimation. Specifying a particular parametric model avoids the curse of dimensionality, but it is then crucial for empirical analysis to test whether this is adequate. However, specification tests contrasting a parametric estimator with a nonparametric one in turn suffer from the curse of dimensionality, and in particular lack power. Hence their practical usefulness is questionable in many applications. In this research program, I propose a novel set of tests of parametric and semiparametric models against a nonparametric alternative that avoid the curse of dimensionality. The key principle is to replace the nonparametric alternative by a ``class'' of a semiparametric alternatives, thus reducing the dimension of the alternative while preserving consistency in testing. I will expand the scope of this new dimension reduction approach to a set of relevant situations in econometric modeling. This will significantly broaden the applicability of specification testing in econometrics, and more generally in applied statistics. I will develop new procedures for: (i) testing a parametric regression function; (ii) testing general conditional moment restrictions; and (iii) testing a single-index regression model. I will allow for panel and time-series data, and for the presence of both continuous and categorical variables. I will also develop a data-driven method for choosing the smoothing parameter in the newly developed tests. Asymptotic analyses, simulation studies and econometric applications will be undertaken. Software modules will be made available to the scientific community for implementation of the different tests.
Programme
Discovery Grants Program - Individual
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
328474-2006
Name nom
Lavergne, Pascal
Department département
Economics
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2007
Competition year année de concours
2006
Award amount
13000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
14
Committee name en
Statistical Sciences
Committee name fr
Statistique
Area of application code
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
3011
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Multivariate analysis
Research subject fr
Analyse multivariée
Partie
2
Nb partie
3
Application title
Dimension reduction in specification testing of econometric models
Keyword
NA
Application summary
Nonparametric methods permit estimation of econometric models without strong a priori parametric assumptions. This flexibility comes at a cost, known in the literature as the ``curse of dimensionality."' If the number of considered variables is large, then a huge amount of data is necessary for accurate estimation. Specifying a particular parametric model avoids the curse of dimensionality, but it is then crucial for empirical analysis to test whether this is adequate. However, specification tests contrasting a parametric estimator with a nonparametric one in turn suffer from the curse of dimensionality, and in particular lack power. Hence their practical usefulness is questionable in many applications. In this research program, I propose a novel set of tests of parametric and semiparametric models against a nonparametric alternative that avoid the curse of dimensionality. The key principle is to replace the nonparametric alternative by a ``class'' of a semiparametric alternatives, thus reducing the dimension of the alternative while preserving consistency in testing. I will expand the scope of this new dimension reduction approach to a set of relevant situations in econometric modeling. This will significantly broaden the applicability of specification testing in econometrics, and more generally in applied statistics. I will develop new procedures for: (i) testing a parametric regression function; (ii) testing general conditional moment restrictions; and (iii) testing a single-index regression model. I will allow for panel and time-series data, and for the presence of both continuous and categorical variables. I will also develop a data-driven method for choosing the smoothing parameter in the newly developed tests. Asymptotic analyses, simulation studies and econometric applications will be undertaken. Software modules will be made available to the scientific community for implementation of the different tests.