Publications Office of the EU
New methods for timely estimates
DisplayCustomHeader
MyPortal
Malti
BasketSummary
Web Content Display
Publication Detail Actions Portlet

Utilitajiet

Publication Detail Actions Portlet
 Klassifikazzjoni tal-utent
custom-survey-notification
Publication Detail Portlet

Publication detail

Home

New methods for timely estimates

Nowcasting euro area GDP growth using quantile regression : 2020 edition

Metadejta tal-pubblikazzjoni

This paper uses an application to explore the utility of quantile regression methods in producing (density) nowcasts. Our quantile regression modelling strategy is designed to reflect important nowcasting features, namely the use of mixed-frequency data, the ragged-edge and increasingly large numbers of indicators (big data). An unrestricted mixed data sampling strategy within a Bayesian quantile regression is used to accommodate a large mixed frequency dataset when nowcasting; we use a shrinkage prior to avoid parameter proliferation in what becomes a large dimensional quantile regression. In an application to euro area GDP growth, using over 400 mixed frequency indicators, we find that the quantile regression approach does not produce as accurate density nowcasts overall as the density combination approach of Mazzi et al. (2014) unless the indicators are orthogonalised and shrunk to a smaller number. The quantile regression approach overstates the uncertainties associated with GDP growth. This is reflected by high probability estimates of GDP at Risk (i.e. negative GDP growth) even during the period of strong GDP growth prior to the global financial crisis. However, when the nowcasts are formed early, at 30 or 15 days before the end of the quarter of interest, the quantile regression approach is better able to detect the ensuing recession than the density combination approach.
Ara aktar
Ara anqas
ISSN 2315-0807
ISBN 978-92-76-16865-2
DOI 10.2785/26603
Numru tal-Katalogu KS-TC-20-004-EN-N
PDF
ISSN 2315-0807 ISBN 978-92-76-16865-2 DOI 10.2785/26603 Numru tal-Katalogu KS-TC-20-004-EN-N
  • Released on EU publications website:
Publication Viewer

Viżwalizzatur tad-dokument

Pop up window annotations