Publication Date: September 2008
Revision Date: August 2011
This paper is concerned with tests and conﬁdence intervals for parameters that are not necessarily identiﬁed and are deﬁned by moment inequalities. In the literature, diﬀerent test statistics, critical value methods, and implementation methods (i.e., the asymptotic distribution versus the bootstrap) have been proposed. In this paper, we compare these methods. We provide a recommended test statistic, moment selection critical value method, and implementation method. We provide data-dependent procedures for choosing the key moment selection tuning parameter kappa and a size-correction factor eta.
Asymptotic size, Asymptotic power, Bootstrap, Conﬁdence set, Generalized moment selection, Moment inequalities, Partial identiﬁcation, Reﬁned moment selection, Test, Unidentiﬁed parameter
JEL Classification Codes: C12, C15
See CFP: 1372
Computer programs also available on the Econometric Society website