Institutional change and the development of lagging regions in Europe

ABSTRACT

This paper assesses whether both the levels and the degree of change in government quality influence regional economic performance in the European Union and, in particular, in its lagging regions. The results of the econometric analysis, covering 249 NUTS-2 regions for the period 1999–2013, suggest that (1) government quality matters for regional growth; (2) relative improvements in quality of government are a powerful driver of development; (3) one-size-fits-all policies for lagging regions are not the solution; (4) government quality improvements are essential for low-growth regions; and (5) in low-income regions basic endowment shortages are still the main barrier to development. In particular, low-growth regions in Southern Europe stand to benefit the most from improvements in government quality, while in low-income regions of Central and Eastern Europe, investments in the traditional drivers of growth remain the main factors behind successful economic trajectories.

KEYWORDS:

• economic growth
• government quality
• institutional change
• regions
• European Union

JEL:

• R11
• R50

ACKNOWLEDGEMENTS

The authors are grateful to the special issue editors, the managing editor and the three reviewers for their insightful suggestions and comments made during the refereeing process. Özge Öner was particularly helpful in challenging some of the points made in the econometric analysis in past iterations. Earlier versions of the paper benefited from the feedback received by participants at presentations in Athens, Beijing, Boston, Brussels, Frankfurt, Groningen, Jena, L’Aquila, Madrid, Milan, Oxford, Porto, Rome, Santiago de Compostela, Split, Stavanger, Stockholm, Tokyo, Turin, Utrecht and Warsaw. The content of this paper does not reflect the official opinion of the European Union. Responsibility for the information and views expressed therein lies entirely with the authors.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Notes

1. Nomenclature of Territorial Unit for Statistics (NUTS), as defined by the European Commission.

2. Data availability limits the analysis to the use of mostly formal institutions – although corruption levels are also included – of government quality. Another interesting strand of research would be to consider the role played by informal institutions, such as social capital, trust or shared values (e.g., Beugelsdijk & Van Schaik, 2005; Akçomak & Ter Weel, 2009). This, however, goes beyond the scope of the paper and would require a separate analysis.

3. Potential drawbacks of PCA include relying on linear assumptions and on orthogonal transformations, as well as issues linked to the use of variance to determine the influence of a particular dimension.

4. The road accessibility data are based on road network data for 2001, 2006, 2011 and 2014, with the remaining years being extrapolated or interpolated. The raw data were gathered by Klaus Spiekermann and provided by Lewis Dijkstra at the European Commission. The distance-decay function is a fairly steep exponential function that becomes close to zero after four hours of travel.

5. Croatia, Cyprus, Denmark and Malta had to be excluded from the analysis due to missing data. Additionally, some individual regions were not included in the analysis for the same reasons. These comprise Ceuta and Melilla, all French overseas departments (Guadeloupe, Martinique, Guyane and Réunion), Açores and Madeira, and North Eastern Scotland and the Highlands and Islands. Recent changes in NUTS-2 boundaries led to the exclusion of the Finnish regions of Helsinki-Uusimaa, Etelä-Suomi, Pohjois-Suomi and Itä-Suomi.

6. The survey – one of the largest ever conducted at a regional (i.e., subnational) level – is based on around 200 participants per region and consisted of 34 quality of government- and demography-related questions, amongst others on education, healthcare and law enforcement. For a more detailed information on the survey as well as the construction of the indices, see Charron et al. (2011, 2014).

7. All independent variables were classified as endogenous in all regressions and the fourth and third lags were employed as (internal) instruments for the endogenous variables.