Semiparametric Efficiency in GMM Models of Nonclassical Measurement Errors, Missing Data and Treatment EffectsAuthor(s):
Publication Date: March 2008
We study semiparametric eﬀiciency bounds and eﬀicient estimation of parameters deﬁned through general nonlinear, possibly non-smooth and over-identiﬁed moment restrictions, where the sampling information consists of a primary sample and an auxiliary sample. The variables of interest in the moment conditions are not directly observable in the primary data set, but the primary data set contains proxy variables which are correlated with the variables of interest. The auxiliary data set contains information about the conditional distribution of the variables of interest given the proxy variables. Identiﬁcation is achieved by the assumption that this conditional distribution is the same in both the primary and auxiliary data sets. We provide semiparametric eﬀiciency bounds for both the “verify-out-of-sample” case, where the two samples are independent, and the “verify-in-sample” case, where the auxiliary sample is a subset of the primary sample; and the bounds are derived when the propensity score is unknown, or known, or belongs to a correctly speciﬁed parametric family. These eﬀiciency variance bounds indicate that the propensity score is ancillary for the “verify-in-sample” case, but is not ancillary for the “verify-out-of-sample” case. We show that sieve conditional expectation projection based GMM estimators achieve the semiparametric eﬀiciency bounds for all the above mentioned cases, and establish their asymptotic eﬀiciency under mild regularity conditions. Although inverse probability weighting based GMM estimators are also shown to be semiparametrically eﬀicient, they need stronger regularity conditions and clever combinations of nonparametric and parametric estimates of the propensity score to achieve the eﬀiciency bounds for various cases. Our results contribute to the literature on non-classical measurement error models, missing data and treatment eﬀects.
Auxiliary data, Measurement error, Missing data, Treatment eﬀect, Semiparametric eﬀiciency bound, GMM, Sieve estimation
JEL Classification Codes: C1, C3
See CFP: 1223