Picture of UK Houses of Parliament

Leading national thinking on politics, government & public policy through Open Access research

Strathprints makes available scholarly Open Access content by researchers in the School of Government & Public Policy, based within the Faculty of Humanities & Social Sciences.

Research here is 1st in Scotland for research intensity and spans a wide range of domains. The Department of Politics demonstrates expertise in understanding parties, elections and public opinion, with additional emphases on political economy, institutions and international relations. This international angle is reflected in the European Policies Research Centre (EPRC) which conducts comparative research on public policy. Meanwhile, the Centre for Energy Policy provides independent expertise on energy, working across multidisciplinary groups to shape policy for a low carbon economy.

Explore the Open Access research of the School of Government & Public Policy. Or explore all of Strathclyde's Open Access research...

Forecasting with dimension switching VARs

Koop, Gary (2014) Forecasting with dimension switching VARs. International Journal of Forecasting, 30 (2). pp. 280-290. ISSN 0169-2070

[img]
Preview
Text (koop-IJF2014-Forecasting-with-dimension-switching-VARs)
koop_IJF2014_Forecasting_with_dimension_switching_VARs.pdf
Accepted Author Manuscript
License: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 logo

Download (157kB)| Preview

    Abstract

    This paper develops methods for VAR forecasting when the researcher is uncertain about which variables enter the VAR, and the dimension of the VAR may be changing over time. It considers the case where there are N variables which might potentially enter a VAR and the researcher is interested in forecasting N ∗ of them. Thus, the researcher is faced with 2 N − N ∗ potential VARs. If N is large, conventional Bayesian methods can be infeasible due to the computational burden of dealing with a huge model space. Allowing for the dimension of the VAR to change over time only increases this burden. In light of these considerations, this paper uses computationally practical approximations adapted from the dynamic model averaging literature in order to develop methods for dynamic dimension selection (DDS) in VARs. We then show the benefits of DDS in a macroeconomic forecasting application. In particular, DDS switches between different parsimonious VARs and forecasts appreciably better than various small and large dimensional VARs.