A0955
Title: High-dimensional conditionally Gaussian state space models with missing data
Authors: Joshua Chan - Purdue University (United States) [presenting]
Aubrey Poon - University of Strathclyde (United Kingdom)
Dan Zhu - Monash University (Australia)
Abstract: An efficient sampling approach is developed for handling complex missing data patterns and a large number of missing observations in conditionally Gaussian state space models. Two important examples are dynamic factor models with unbalanced datasets and large Bayesian VARs with variables in multiple frequencies. A key insight underlying the proposed approach is that the joint distribution of the missing data conditional on the observed data is Gaussian. Moreover, this conditional distribution's inverse covariance or precision matrix is sparse, and this special structure can be exploited to speed up computations substantially. The methodology is illustrated using two empirical applications. The first application combines quarterly, monthly and weekly data using a large Bayesian VAR to produce weekly GDP estimates. The second application extracts latent factors from unbalanced datasets involving over a hundred monthly variables via a dynamic factor model with stochastic volatility.
