Article ID: | iaor20101932 |
Volume: | 29 |
Issue: | 1-2 |
Start Page Number: | 71 |
End Page Number: | 93 |
Publication Date: | Jan 2010 |
Journal: | Journal of Forecasting |
Authors: | Artis Michael, Okubo Toshihiro |
Keywords: | forecasting: applications |
The paper uses annual data on real GDP for 12 UK regions and 12 manufacturing sectors to derive regional and regional/sectoral business cycles using a Hodrick–Prescott filter. The cohesion of the cycles is examined via cross‐correlations and comparisons made with the regional cycles for Japan, the USA and the euro area. The UK emerges as especially cohesive and efforts to explain the overall cross‐correlations of regional GDP are not very successful owing to the low variance of the explicand; when attention is turned to the sectoral/regional cycles, with their greater variance, it is possible to demonstrate that economic variables such as distance, dissimilarity in structure and level of output play a significant role in explaining the variance in the cross‐correlations. A significant feature of the cross‐correlations in relation to those of EU countries is that while they continue to provide support for the ‘UK idiosyncrasy’ they no longer do so as strongly as they did in earlier data samples.