Fighting Gravity: Institutional Changes and Regional Disparities in the EU

abstract

A thorough investigation of the relationship between the deepening and the widening processes of EU integration and the historic evolution of regional disparities is missing in the related vast existing inductive literature. This usually focuses on EU15 countries or Central and Eastern European Countries, generally involves relatively short and recent periods, and takes into consideration just one institutional change at a time, if any. This article aims at filling these gaps by providing a theoretical interpretative framework on the effects of each institutional change on inter- and intranational disparities, and by verifying the fitness of the theoretical expectations with a longitudinal trend analysis and with an econometric event analysis. Theory was overall right in claiming that widening and deepening of the EU would have exacerbated intraregional disparities. Counteracting such tendency could look like fighting gravity, since disparities continuously reappear in different forms and spatial levels, transmitting a sense of unescapable normality.

Key words:

• disparities
• institutional changes
• EU enlargement
• single market
• common currency
• Euro
• Theil index

Acknowledgments

The descriptive empirical part of this work is based on a research project carried out by the authors for the European Commission, Directorate General for Regional and Urban Policy [Contract No 2017CE16BAT125].

Notes

¹ The title itself of Ohlin’s (1933) book underlines the profound difference between international and interregional trade, suggesting that in the latter case, equalization could depend mainly on out-migration processes from lagging regions rather than on pure development.

² Article 130a of the Single European Act (1986) states that “[I]n order to promote its overall harmonious development, the Community shall develop and pursue its actions leading to the strengthening of its economic and social cohesion.” Single European Act, February 17, 1986, 1987 O.J. (L 169) 1, 25 I.L.M. 506. The Treaty of Rome (1957) itself was mentioning the goal of a harmonious development of economic activities (art. 2) “en réduisant l’écart entre les différentes régions et le retard des moins favorisés” (Preambule) [by reducing the gap between the different regions and the backwardness of the least advantaged (Preamble)]. EEC Treaty or Treaty of Rome: Treaty Establishing the European Economic Community, March 25, 1957, 298 U.N.T.S. 3, 4 Eur. Y.B. 412, https://eur-lex.europa.eu/legal-content/FR/TXT/PDF/?uri=CELEX:11957E/TXT&from=IT.

³ Given data availability, the empirical analysis starts in 1980. We are therefore forced to analyze the enlargements from that of 1981, leaving aside the 1973 one. Theoretical considerations presented in “The Common Market and Regional Disparities,” however, hold also for the six founding members and for the 1973 enlargement.

⁴ On the synchronization of business cycles, the reader can refer to Petrakos, Rodríguez-Pose, and Rovolis (2005).

⁵ The years leading up to official membership status constitute what can be defined as enlargement shock (Gomulka 2003; Charron 2013).

⁶ We refer here to the establishment of the Common Market (six members), and the subsequent enlargements to Denmark, UK, and Ireland in 1973 (EEC9), to Greece in 1981 (EEC10), to Spain and Portugal in 1986 (EEC12), and to Austria, Finland, and Sweden in 1995 (EU15).

⁷ The reduction of transaction costs came along with a reduction of uncertainty.

⁸ Other possible reasons for this low impact, assessed both ex ante and ex post, were the fact that, according to the comparative advantage theory, the advantages of international specialization do not reside in the differences between production costs of the single countries, but rather in their variation around the international (necessarily lower) average (Scitovsky 1958); and the fact that most of the new trade among European countries happened at the intraindustry level (or two-way trade), where the advantage mainly refers to variety of available goods rather than to their cost (Balassa 1975).

⁹ These would be, according to the title of chapters, “The Problems of Currency Union” (Ch. II.B) and “Some Likely Consequences of Integration” (Ch. III.4) (Scitovsky 1958). These interregional effects of integration are likely to be faster and wider than the international ones. They can also generate demographic abandonment and even desertification of some internal or peripheral areas, due to easier interregional mobility of labor and absence of reequilibrating automatic mechanisms in the economic sphere such as currency devaluation and downward flexibility of prices and wages—existing at the international level (Camagni 2002). Empirical evidence in the EU confirms rapid increases in the early phases of integration (Italy in 1955–65; Spain and Portugal in the 1990s, New Eastern accession countries in the 2000s), correlated with phases of acceleration of the integration process, like in the years 1985–95 (Camagni and Capello 2014).

¹⁰ In economic terms, regions do not compete on the basis of a Ricardian, comparative advantage principle, but rather on a Smithian, absolute advantage principle (Camagni 2002). The Ricardian principle, guaranteeing each country a role in the international division of labor, whatever its relative efficiency or competitiveness, is based on two alternative adjusting mechanisms—wage-price flexibility or currency devaluation—which are not available at the regional level.

¹¹ Maastricht Treaty, TEU or Union Treaty: Treaty on European Union, February 7, 1992, 1992 O.J. (C191) 1, 31 I.L.M. 253.

¹² For a debate on the costs of the common currency, see Wyplosz, Nickell, and Wolf (2006).

¹³ In fact, each enlargement shifted the center of gravity of the EU.

¹⁴ For this work, data were used at the NUTS2 level.

¹⁵ The weighted coefficient of variation and the mean logarithmic deviation were also considered. The first, in particular, is a very common indicator, especially in official EU Commission documents (European Commission 2017), but has the limit of not being decomposable. In Appendix A it is shown that the Theil index, the weighted coefficient of variation, and the logarithmic deviation provide consistent results. Consequently, the decomposable Theil index is preferred.

¹⁶ 1981: Greece. 1986: Portugal and Spain. 1995: Austria, Finland, and Sweden. 2004: Malta, Cyprus, Estonia, Latvia, Lithuania, Hungary, Czech Republic, Slovak Republic, Poland, and Slovenia. 2007: Bulgaria and Romania. 2013: Croatia. The first two enlargements could not be explicitly taken into account within the present work due to data available only since 1980.

¹⁷ This result is mainly a statistical effect, since, in general, new entrants are poorer with respect to incumbents’ average. An exception exists in this regard, represented by the 1995 enlargement, when richer countries—namely, Austria, Finland, and Sweden—entered the EU with a slightly positive effect on between-country disparities.

¹⁸ With the reunification of East and West Germany, disparities among European member countries decreased. This is particularly evident in the evolution of disparities in the group of the six founding members. With the reunification (continuous blue line in Figure 1(a)), disparities decreased in 1991 (the first year for which data on Eastern German regions are available) and remained constant for some years. Such positive evolution in disparities is due to Germany becoming poorer and thus moving toward the average of the other countries. The dashed-dotted blue line in Figure 1(a) represents, instead, how disparities would have evolved without the German reunification.

¹⁹ The light blue line representing the six founding members is much steeper in Figure 2(a) where East German regions are taken into account with respect to Figure 2(b). This is true for all other lines, representing different groups of countries, although not so evident given the lower weight of German regions in the aggregates.

²⁰ “[T]he united German government invested heavily in eastern infrastructures of all kinds. Physical, administrative, social, educational and scientific infrastructures were torn up at a speed that may in due course be regarded as one of the most remarkable achievements in rebuilding the basics of a modern market economy in just a few years’ time.” (Paqué 2003, 110).

²¹ Four years have been chosen as a reasonable time to grasp the short-term effects of entering the EU; it is short enough to be available for those member countries that entered very recently and long enough to capture some tendencies.

²² Campos, Coricelli, and Moretti (2019) even find negative effects from EU membership in the case of Greece.

²³ A debate exists in the literature on the actual presence of this inverted U curve, or rather the presence of a J curve (Petrakos, Kallioras, and Anagnostou 2011; Monastiriotis 2014; Artelaris and Petrakos 2016).

²⁴ In the previous period before the one considered in this work, namely, from the creation of the European Economic Community in 1957 to 1980, the Williamson curve proved right. Up to 1970 disparities increased, especially in weaker countries, like Italy and Belgium, but then decreased: large cities were left by industry, old industrial areas were hit by the crisis of their specialization sectors (iron and steel, shipbuilding, textiles) and new intermediate regions emerged (the Third Italy, that is the northeastern and central regions of the country, Flanders, Southern France, Southern Germany) (see Camagni and Cappellin 1985; Camagni 1995).

²⁵ As also highlighted by Petrakos, Rodríguez-Pose, and Rovolis (2005), the sharp fall in the Theil index in the case of Portugal depends on the fluctuations in the GDP in the region of Alentejo during the 1980s. As a consequence, these data should be viewed with caution.

²⁶ Excluded countries are Cyprus, Estonia, Latvia, Lithuania, Luxembourg, and Malta. In addition, Croatia, Ireland, and Slovenia were dropped, since by having only 2 NUTS2 regions according to the 2013 nomenclature, their Theil index may be unreliable.

²⁷ In fact, CEECs enter the database only at the beginning of the 1990s so that the overall sample is unbalanced in terms of number of observations over time.

²⁸ All the data were retrieved from Cambridge Econometrics.

²⁹ Smaller regions weight less and generally have more noisy data, due to smaller sample sizes, which introduce biases when analyzing changes in disparities.

³⁰ See also Ezcurra, Pascual, and Rapún (2007b).

Additional information

Funding

This work was supported by the European Commission, Directorate General for Regional Policy [Contract No 2017CE16BAT125].