Semiparametric optimal estimation with nonignorable nonresponse data

When the response mechanism is believed to be not missing at random (NMAR), a valid analysis requires stronger assumptions on than standard statistical methods would otherwise require. Semiparametric estimators have been developed under parametric model mechanism. In this paper, new test proposed guarantee identifiability without using instrumental variable assumption. Furthermore, we develop optimal semiparametric estimation for parameters such as population mean. Specifically, propose two that do require any other Asymptotic properties of are discussed. An extensive simulation study presented compare with some existing methods. We present an application our method Korean labor and income panel survey data.