Macroeconomic instability in the European monetary system?

Abstract

This article analyses the impact of the establishment of the European Monetary System (EMS) on a number of macroeconomic variables, such as exchange rates, money, interest rates and prices for member countries participating in the Exchange Rate Mechanism (ERM). Instability is examined in terms of multiple structural breaks in the variance of the series. Two procedures are followed for this purpose: the OLS-based tests to detect multiple structural breaks, as proposed by Bai and Perron (1998, 2003), and several procedures based on Information Criterion together with the so-called sequential procedure suggested by Bai and Perron (2003). Results indicate that there is some evidence of structural breaks in volatility across investigated variables, with the realignments in the ERM playing a significant role in reducing volatility in some countries and sub-periods. In this regard, the results tend to support the hypothesis that the EMS has contributed to reducing macroeconomic volatility in member countries.

Acknowledgements

The authors are very grateful to Oscar Bajo-Rubio, Mark Taylor and an anonymous referee for useful comments and suggestions on a previous draft of this article, substantially improving the content and quality of the article. Financial support from the Spanish Ministry of Science and Technology (SEJ2005-09094/ECON) and by the Centro de Estudios Andaluces (project ECO-9) is also gratefully acknowledged. The views expressed here are those of the authors and not necessarily those of the institutions with which they are affiliated.

Notes

¹ Initially, only Denmark and Greece participated in ERM-II. On 27 June 2004 the Estonian kroon, the Lithuanian lita and the Slovenian tolar joined ERM-II. On 2 May 2005 three other Member States joined ERM-II: Cyprus, Latvia and Malta. Slovakia followed suit on 25 November 2005. Finally, after having achieved a high degree of sustainable economic convergence with the other Member States, Greece adopted the euro in January 2001, Slovenia in 2007 and Chipre and Matta in 2008.

² We concentrate on the last two approaches given that the ICSS algorithm presents several weaknesses (see, for example, Sansó et al., 2004 and Valentinyi-Endrész, 2004).

³ We are particularly grateful to Bai and Perron for providing us with the GAUSS code for computations.

⁴ Similarly, Stock and Watson (2002) use the absolute value of the fitted residuals of a VAR model to analyse changes in variance. Alternatively, Valentinyi-Endrész, (2004) uses the squared errors from an AR(1)-GARCH(1,1) model to compute changes in variance.

⁵ For further analysis see Bai and Perron (1998, 2003).

⁶ This full period differs between countries depending on the data availability.

⁷ For this analysis we have chosen the variables with the most homogenous sample period.

⁸ In order to save space, we present in the tables only the numerical results of four of the statistics described in Section III. The other results are available from the authors upon request.

⁹ To select the number of breaks of the final model, we follow the practical recommendations by Bai and Perron (2003). Although in most of the cases the number of breaks selected by the procedures is very similar, where it is different we concentrate on the number indicated by the SP (SP is better at selecting the number of breaks than the other procedures). Moreover, the performance of the SP can be improved by considering the UDmax or WDmax tests to see if at least one break is present. The number of breaks can then be decided on the basis of an examination of the sup F_T (l + 1/l) statistics. See Bai and Perron (2003) for further analysis. All the other results are available from the authors upon request.

¹⁰ The Basle-Nyborg Agreement in September 1987 strengthened interventions in the foreign exchange market, implying the acceptance of the Deutschmark as the anchor of the system (Braga de Macedo et al., 2001).