References & Citations
Economics > Econometrics
Title: Combining Shrinkage and Sparsity in Conjugate Vector Autoregressive Models
(Submitted on 20 Feb 2020 (v1), last revised 26 Aug 2020 (this version, v2))
Abstract: Conjugate priors allow for fast inference in large dimensional vector autoregressive (VAR) models but, at the same time, introduce the restriction that each equation features the same set of explanatory variables. This paper proposes a straightforward means of post-processing posterior estimates of a conjugate Bayesian VAR to effectively perform equation-specific covariate selection. Compared to existing techniques using shrinkage alone, our approach combines shrinkage and sparsity in both the VAR coefficients and the error variance-covariance matrices, greatly reducing estimation uncertainty in large dimensions while maintaining computational tractability. We illustrate our approach by means of two applications. The first application uses synthetic data to investigate the properties of the model across different data-generating processes, the second application analyzes the predictive gains from sparsification in a forecasting exercise for US data.
Submission history
From: Niko Hauzenberger [view email][v1] Thu, 20 Feb 2020 14:45:38 GMT (537kb,D)
[v2] Wed, 26 Aug 2020 13:38:21 GMT (1245kb,D)
Link back to: arXiv, form interface, contact.