We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

econ.EM

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Economics > Econometrics

Title: Gaussian Process Vector Autoregressions and Macroeconomic Uncertainty

Abstract: We develop a non-parametric multivariate time series model that remains agnostic on the precise relationship between a (possibly) large set of macroeconomic time series and their lagged values. The main building block of our model is a Gaussian Process prior on the functional relationship that determines the conditional mean of the model, hence the name of Gaussian Process Vector Autoregression (GP-VAR). We control for changes in the error variances by introducing a stochastic volatility specification. To facilitate computation in high dimensions and to introduce convenient statistical properties tailored to match stylized facts commonly observed in macro time series, we assume that the covariance of the Gaussian Process is scaled by the latent volatility factors. We illustrate the use of the GP-VAR by analyzing the effects of macroeconomic uncertainty, with a particular emphasis on time variation and asymmetries in the transmission mechanisms. Using US data, we find that uncertainty shocks have time-varying effects, they are less persistent during recessions but their larger size in these specific periods causes more than proportional effects on real growth and employment.
Comments: JEL: C11, C14, C32, E32; KEYWORDS: Bayesian non-parametrics, non-linear vector autoregressions, asymmetric uncertainty shocks
Subjects: Econometrics (econ.EM)
Cite as: arXiv:2112.01995 [econ.EM]
  (or arXiv:2112.01995v1 [econ.EM] for this version)

Submission history

From: Niko Hauzenberger [view email]
[v1] Fri, 3 Dec 2021 16:16:10 GMT (1834kb,D)
[v2] Mon, 17 Jan 2022 10:25:32 GMT (1861kb,D)

Link back to: arXiv, form interface, contact.