Online Program Home
My Program

Abstract Details

Activity Number: 340 - SPEED: Bayesian Methods, Part 1
Type: Contributed
Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #306884
Title: High-Dimensional Posterior Consistency in Mixed Frequency Bayesian Vector Autoregressive Models
Author(s): Nilanjana Chakraborty* and George Michailidis and Kshitij Khare
Companies: University of Florida and University of Florida and University of Florida
Keywords: Vector Autoregressive Models; Mixed frequency Bayesian VAR; Posterior consistency; Spike and slab prior

In the recent past, considerable advances have been made in the research on Vector Autoregressive (VAR) models. It is widely used for simultaneous forecasting of multiple temporally observed variables and has got many applications in macroeconomics, finance and other areas. We often find economic data collected at various frequencies. Though some studies have been done in the recent years on mixed frequency VAR models, not much theory is yet developed on the posterior distribution of mixed-frequency Bayesian VAR models in high-dimensional setup. Here we consider a mixed-frequency VAR model speci?ed at the lowest sampling frequency; treating high frequency observations as different series occurring at the low frequency. To introduce sparsity in the model coefficients, we use spike and slab prior, which is a mixture of a degenerate distribution at zero (spike) and Gaussian distribution(slab), for the autoregressive coefficient matrix. Under standard regularity assumptions, we establish posterior consistency for this prior when the dimension of the VAR model grows with the sample size. The efficiency of the proposed estimation method is demonstrated on synthetic and real data sets.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program