The euro crisis and Swedish GDP growth – a study of spillovers

Abstract

In this article, a Bayesian VAR model is used to study the effects of euro area shocks on GDP growth in the small open economy of Sweden. A novel feature is that the new policy uncertainty index of Baker et al. (2013) is introduced in the model. The model behaves well in terms of reasonable impulse response functions. The specific effects of the euro crisis is investigated through a historical decomposition which shows that shocks to euro area GDP growth have been a reasonably important factor for Swedish GDP growth, supporting it during 2010 and holding it back thereafter. Generally, shocks to policy uncertainty have held back Swedish GDP growth during the euro crises.

Keywords:

• small open economy
• Bayesian VAR
• policy uncertainty index
• historical decomposition

JEL Classification:

• C32
• F43

Notes

¹ See, for example, Pesaran et al. (2004), Galvão et al. (2007), Österholm and Zettelmeyer (2008), Bagliano and Morana (2010), Bayoumi and Swiston (2009) and Erten (2012).

² This mean-adjusted specification was developed by Villani (2009).

³ The European policy uncertainty index is based on data from France, Italy, Germany, Spain and the United Kingdom. The United Kingdom is obviously not part of the euro area but we argue that the index should serve well as a proxy for euro area policy uncertainty given the relative weights.The high-yield bond spread is sometimes interpreted as reflecting risk appetite; see, for example, Levy Yeyati and González Rozada (2005) and Österholm and Zettelmeyer (2008). It has also been shown to have predictive power for the real economy; see, for example, Mody and Taylor (2003).

⁴ The priors for are given in detail in Table A1 in the Appendix. Priors for all variables except the policy uncertainty index follow those in Österholm (2010). For the policy uncertainty index, the prior mean is chosen to be 100. This is based on the fact that Baker et al. (2013) normalize the index to have a mean of 100 prior to 2011.

⁵ See, for example, Doan (1992) and Villani (2009). The overall tightness is set to 0.2, the cross-variable tightness to 0.5 and we employ a lag decay parameter of 1.

⁶ The impulse response functions are based on a Cholesky decomposition of the covariance matrix with the ordering of the variables given by Equation 2.

⁷ This choice of the starting point for the euro crisis is obviously arbitrary but by 2010Q2, signs of problems were showing in terms of rising spreads on Greek government debt.

⁸ It can first be noted that the domestic control factors have had a very small effect on Swedish GDP growth during the time for the euro crises (as can be seen from the bottom two rows of Fig. A3).