Business Cycle Synchronization in EU: A Time‐Varying Approach
| Date | 01 September 2014 |
| Author | George Filis,Stavros Degiannakis,David Duffy |
| Published date | 01 September 2014 |
| DOI | http://doi.org/10.1111/sjpe.12049 |
BUSINESS CYCLE SYNCHRONIZATION
IN EU: A TIME-VARYING APPROACH
Stavros Degiannakis*
,
**, David Duffy*** and George Filis****
ABSTRACT
This article investigates the time-varying correlation between the EU12-wide
business cycle and the initial EU12 member-countries based on Scalar-BEKK
and multivariate Riskmetrics model frameworks for the period 1980–2012. The
paper provides evidence that changes in the business cycle synchronization corre-
spond to major economic events that have taken place at a European level. In
the main, business cycle synchronization until 2007 had moved in a direction
positive for the operation of a single currency, suggesting that the common mon-
etary policy was less costly in terms of lost flexibility at the national level. How-
ever, as a result of the Great Recession of 2007 and the subsequent Eurozone
Crisis, a number of periphery countries, most notably Greece, have experienced
desynchronization of their business cycles with the EU12-wide cycle. Neverthe-
less, for most countries, any questions regarding the optimality and sustainability
of the common currency area in Europe should not be attributed to a lack of
cyclical synchronization.
II
NTRODUCTION
This article investigates the time-varying business cycles synchronization
between the initial EMU12 member-countries and the EMU12-wide business
cycle,
1
using quarterly data from 1980 until 2012. In addition, we investigate
this relation for Denmark, Sweden and the United Kingdom, the non-EMU
members, but originally EU15 members. The motivation for the selection of
this group of EU countries is that business cycle synchronization is an impor-
tant pre-requisite to forming a successful currency union as implied by the
insights of Optimal Currency Area Theory. In the United Kingdom, for
instance, one of Gordon Brown famous 5 tests for joining the Euro was the
assurance that the UK and the European-wide business cycles would be
*Athens University of Economics and Business
**Bank of Greece
***University of Ulster
****Bournemouth University
1
Luxemburg was omitted due to data unavailability. The results are not sensitive to the
exclusion of Luxemburg data due to its small size. The European Union-wide business cycle
is estimated in the same spirit with de Haan et al. (2007) and Artis et al. (2004). Stylized
facts for the European-wide business cycle are provided by Artis et al. (2004).
Scottish Journal of Political Economy, DOI: 10.1111/sjpe.12049, Vol. 61, No. 4, September 2014
©2014 Scottish Economic Society.
348
synchronized. In addition, the recent economic crisis signified the importance
of business cycle synchronization in the EU with regards to the application of
a suitable union-wide monetary policy response. This study explores a current
economic topic in light of recent economic developments.
Pioneers in the study of business cycles include, inter alia, Mitchell (1927),
Burns and Mitchell (1946) and Kuznets (1958). Since then, a significant
amount of literature has been produced on the study of business cycle syn-
chronization. Papageorgiou et al. (2010) and de Haan et al. (2008) provide an
extensive review of the literature.
Previous studies have used a wide range of techniques and data to study
the level of synchronization in European business cycles and other bilateral
business cycles synchronizations. The various techniques that have been
applied to this research question range from constant contemporaneous and
lagged correlations for entire periods or sub-periods
2
to Vector Autoregressive
(VAR) models
3
and from frequency-domain dynamic correlations
4
to rolling
windows correlations.
5
Although these methods provide a sufficient under-
standing of the business cycle synchronization in Europe, they share some lim-
itations. To start with, a static correlation figure is not able to capture any
fluctuations of the correlation level across time. In addition, the robustness of
the results obtained for rolling windows correlations is subject to the length of
the rolling window.
6
Furthermore, choosing sub-periods exogenously in an
effort to produce a quasi time-varying correlation could have several draw-
backs (see Sebastien, 2009, for additional explanation of these drawbacks).
These shortcomings are important and it is a development of this paper that
the proposed techniques do not suffer from such drawbacks.
This study directly addresses all the above issues by employing two robust
quantitative techniques, namely the Scalar-BEKK and multivariate Riskmet-
rics models, as were suggested by Baba et al. (1990) and J.P. Morgan (1996),
respectively. To our knowledge, these techniques have not previously been
applied to investigate the time-varying correlation between the individual
European member-countries and the European Union-wide business cycle
without a priori imposing regime switches. In addition, for robustness pur-
poses, we use two different filtering methods for the extraction of the cyclical
components, namely the Hodrick-Prescott filter (Hodrick and Prescott, 1997)
and the band-pass filter proposed by Baxter and King (1999).
7
These filters
were chosen in order for our study to be more easily comparable with
previous literature.
2
See, inter alia, Gogas and Kothroulas (2009), Ferreira-Lopes and Pina (2011), Furceri
and Karras (2008), Artis and Zhang (1999), Fatas (1997), Inklaar and de Haan (2001).
3
For further details on VAR models the reader is directed to Bergman and Jonung (2010).
4
For further details on frequency-domain correlations the reader is directed to Concaria
and Soares (2009), Azevedo (2002), Croux et al. (2001), as well as references in de Haan
et al. (2008).
5
See, for example, D€
opke (1999).
6
See Savva et al. (2010) for additional explanation.
7
We considered these two filtering methods as they are the most commonly used methods
and thus our results can be directly comparable to the existing literature.
BUSINESS CYCLE SYNCHRONIZATION IN EU 349
Scottish Journal of Political Economy
©2014 Scottish Economic Society
To continue reading
Request your trial