A0608
Title: Bayesian benchmarking small area estimation via entropic tilting
Authors: Yuki Kawakubo - Chiba University (Japan) [presenting]
Shonosuke Sugasawa - Keio University (Japan)
Genya Kobayashi - Meiji University (Japan)
Abstract: Benchmarking estimation and its risk evaluation is a practically important issue in small-area estimation. While hierarchical Bayesian methods have been widely adopted in small-area estimation, a unified Bayesian approach to benchmarking estimation has not been fully discussed. An entropic tilting method is employed to modify the posterior distribution of the small area parameters to meet the benchmarking constraint, which enables obtaining benchmarked point estimation as well as reasonable uncertainty quantification. Using conditionally independent structures of the posterior, general Monte Carlo methods are first introduced for obtaining a benchmarked posterior, and then it is shown that the benchmarked posterior can be obtained in an analytical form for some representative small area models. The usefulness of the proposed method is demonstrated through simulation and empirical studies.
