EcoSta 2018: Registration
View Submission - EcoSta2018
Title: Calibration estimation for semiparametric copula models under missing data Authors:  Kaiji Motegi - Kobe University (Japan) [presenting]
Zheng Zhang - Renmin University of China (China)
Shigeyuki Hamori - Kobe University (Japan)
Abstract: The estimation of semiparametric copula models under the presence of missing data is investigated. Our models comprise nonparametric marginal distributions and parametric copula functions. The two-step pseudo-likelihood method is infeasible when there exist missing data. Inspired in a recent work, we propose a class of calibration estimators for both marginal distributions and the parameters of interest without imposing additional models on the missing mechanism. We establish consistency and asymptotic normality for our estimators of copula parameters. We also present a natural procedure for consistently estimating the asymptotic variance of our estimators.