Assessing economic convergence in the EU: is there a perspective for the ‘cohesion countries’?

Abstract

This article analyses the stochastic income convergence within the EU-15. The empirical analysis uses per capita GDP, in PPP and in constant prices of 2005 for the period 1950 to 2010. Apart from the traditional DF type tests we also account for possible structural changes. In this direction, we employ the Zivot-Andrews (1992) and the Lee-Strazicich (1999, 2003) testing procedures, for one and two breaks, endogenously determined. Furthermore, we apply the Carlino and Mills (1993) methodology proposed for the detection of β-convergence. The overall evidence supports the existence of two discrete clubs, the first by the ‘cohesion countries’ (Portugal, Ireland, Greece and Spain) and the second by the remaining members. In particular, there is a clear evidence of convergence within each club, whereas between clubs there is a luck of catching-up effects. Furthermore, investigation of correlation between relative per capita GDP of each country and several factors that are often identified as growth stimulants, namely Total Factor Productivity, FDI, investment and openness confirm, with the exception of Greece, a strong association between these factors and the convergence process. However, progress in the convergence has not been uniform across countries and over time, reflecting the specific interactions between domestic and international factors and their impact on the convergence process of individual countries.

Keywords:

• unit root tests
• structural breaks
• convergence
• EU

JEL Classification:

• C1
• C22
• O470
• O520
• P51

View correction statement:

Erratum

Acknowledgement

We are thankful to an anonymous referee for his valuable comments.

Notes

¹ Baumol (1986) tested for a negative relationship between average annual rates of growth and initial levels of income and concluded that a first convergence club appears to include the industrial countries, another, only moderately converging, consists of middle income countries, whereas the low income countries actually diverge over time. He also noted that among these groups exists very little evidence of convergence.

² See Durlauf and Quah (1999) for a survey of the convergence testing literature based on both cross-sectional and time series data.

³ Numerous studies have applied the Carlino and Mills’ (1993) methodology with mixed results, mainly because of the period under examination and the country considered as a benchmark. See, for instance, Tomljanovich and Vogelsang (2002), Nieswiadomy and Strazicich (2004), and Dawson and Sen (2007).

⁴ Ireland, Portugal, Spain, and Greece, form the group of ‘cohesion’ countries within the European Union. This definition was established with the enlargement of the then EEC to southern Europe, in 1981 and 1986, and was born out of the consideration that integration into the European Communities of the peripheral countries would imply measures to take into account differentials in development levels (Lains, 2006). In terms of per capita GDP, all four were below 75% of the EU average and still contain some of the poorest regions in the Union. In addition to the economic disparities, all four are on the periphery of the Union, while two among them, namely Ireland and Greece are geographically remote from all the rest.

⁵ Since 1 January 2004, Ireland is no longer included in the group of so-called ‘cohesion countries’ given the level of its Gross National Income (GNI) per head compared to the EU average.

⁶ The data for Greece covers the period 1951 to 2010 and for Germany the years 1970 to 2010.

⁷ Dawson and Sen (2007) note that if the only rejection of the unit root null hypothesis occurs in the tests without trend, as in the case of Austria and Belgium, the implication is a zero trend indicating that the country may have already reached its steady state.

⁸ A number of recent studies have been more sceptical about the robustness of this impact and suggest that the statistical significance of this correlation depends on the specification of the empirical model, the period under consideration and the proxy variables for these factors.

⁹ Rodriguez (2007) discusses recent empirical research regarding the link between openness and growth in cross-sectional data.

¹⁰ Several scholars have criticized the robustness of these findings especially on methodological and measurement grounds (see, example.g., Levine and Renelt, (1992; Rodríguez and Rodrik, 1999; Vamvakidis, 2002).

¹¹ Ireland, the smallest among the four countries, stands out as a clear exception. The country was the most export-oriented economy of the group and had arguably embraced outward-orientation more fully than had the other three countries (Barry, 2003). In contrast, Greece, a small economy too, remained throughout the period very closed. Similarly, although Portugal and Spain began to converge around 1986, the countries also remained closed after joining the EU.

¹² Borensztein et al. (1998) noted that certain characteristics in the host countries play an important role and conclude that FDI contributes to economic growth only when the economy has the capacity to absorb the advanced technology. Alfaro et al. (2004) also found that FDI positively impacts on economic growth in countries with sufficiently developed financial markets.

¹³ Concerning the cohesion countries, our results about the FDI-GDP connection are in line with the results found in Kaitila (2004). In Ireland’s case, the difference between the correlation coefficient of our own research and that of Kaitila (2004) is due to differences in the period under consideration. Indeed, for the period 1990 to 2000 the estimated coefficient is 0.847, which does not differ from that estimated in Kaitila (2004).

¹⁴ In Ireland, this is assisted by large inflows of FDI by United States’ Multinationals, which have used Ireland as a platform for the service of the European Single Market.

¹⁵ The explanatory variables most often proposed in empirical studies as determinants of TFP growth are the following: physical and human capital accumulation, government size and fiscal policy, monetary and price stability, openness to international trade, financial sector development, institutions and political stability, foreign aid, terms-of-trade, and other exogenous shocks.

¹⁶ For the case of Greece we can refer to Miller and Upadhyay (2002), who suggest that countries which fail to attract FDI, and therefore to adopt new technical advances will fall behind in productivity improvements and will not converge.

¹⁷ For the relationship between investment and economic growth see empirical studies by Kormendi and Meguire (1985), De Long and Summers (1992), Levine and Renelt (1992), Mankiw et al. (1992), Auerbach et al. (1994), Barro and Sala-i-Martin (1995), Khan and Kumar (1997), Sala-i-Martin (1997), and Bond et al. (2001).