KIT | KIT-Bibliothek | Impressum

Semiparametric estimation with generated covariates

Mammen, Enno; Rothe, Christoph; Schienle, Melanie

Abstract: We study a general class of semiparametric estimators when the infinite-dimensional nuisance parameters include a conditional expectation function that has been estimated nonparametrically using generated covariates. Such estimators are used frequently to e.g. estimate nonlinear models with endogenous covariates when identification is achieved using control variable techniques. We study the asymptotic properties of estimators in this class, which is a non-standard problem due to the presence of generated covariates. We give conditions under which estimatorsare root-n consistent and asymptotically normal, derive a general formula for the asymptotic variance, and show how to establish validity of the bootstrap.

Zugehörige Institution(en) am KIT Institut für Volkswirtschaftslehre (ECON)
Publikationstyp Forschungsbericht
Jahr 2016
Sprache Englisch
Identifikator DOI(KIT): 10.5445/IR/1000051816
ISSN: 2190-9806
URN: urn:nbn:de:swb:90-518162
KITopen ID: 1000051816
Verlag KIT, Karlsruhe
Umfang 43 S.
Serie Working paper series in economics ; 81
Schlagworte Semiparametric estimation, generated covariates, profiling, propensity score
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft KITopen Landing Page