Current browse context:
stat
Change to browse by:
References & Citations
Statistics > Computation
Title: Sparse Bayesian vector autoregressions in huge dimensions
(Submitted on 11 Apr 2017 (v1), last revised 4 Dec 2019 (this version, v3))
Abstract: We develop a Bayesian vector autoregressive (VAR) model with multivariate stochastic volatility that is capable of handling vast dimensional information sets. Three features are introduced to permit reliable estimation of the model. First, we assume that the reduced-form errors in the VAR feature a factor stochastic volatility structure, allowing for conditional equation-by-equation estimation. Second, we apply recently developed global-local shrinkage priors to the VAR coefficients to cure the curse of dimensionality. Third, we utilize recent innovations to efficiently sample from high-dimensional multivariate Gaussian distributions. This makes simulation-based fully Bayesian inference feasible when the dimensionality is large but the time series length is moderate. We demonstrate the merits of our approach in an extensive simulation study and apply the model to US macroeconomic data to evaluate its forecasting capabilities.
Submission history
From: Gregor Kastner [view email][v1] Tue, 11 Apr 2017 11:14:41 GMT (2174kb,D)
[v2] Wed, 27 Jun 2018 16:04:35 GMT (8029kb,D)
[v3] Wed, 4 Dec 2019 07:29:27 GMT (3233kb,D)
Link back to: arXiv, form interface, contact.