Is growth corrupted or bureaucratic? Panel evidence from the enlarged EU

ABSTRACT

This article aims at analysing the issue of conditional convergence in the new enlarged European Union (EU) over the period 1995–2012 by means of panel data techniques. We examined the issue of conditional convergence in the enlarged EU giving particular attention to the effects of corruption and bureaucracy on growth controlling for a widely used set of explanatory variables such as investment (domestic and foreign), human capital formation, inflation, general government final consumption and trade openness. Furthermore, we examine if growth responds differently to corruption and bureaucracy in the new EU members by means of two group-specific interaction variables to capture possible different responses to corruption and bureaucracy. The analysis reveals evidence of conditional convergence in the enlarged EU, with investment share, foreign direct investment, human capital, and country openness appearing as robust growth drivers. In contrast, inflation and government consumption rather hamper growth. Furthermore, the effects of corruption and bureaucracy on growth seem to differ across old and new EU members.

KEYWORDS:

• Convergence
• growth determinants
• corruption
• bureaucracy
• GMM

JEL CLASSIFICATION:

• B23
• C33
• O40
• O52
• O43
• O47

Acknowledgment

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

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Surveys of the empirical growth literature suggest a large number of explanatory variables grouped into ‘categories’ or distinct growth theories. See, for example, Durlauf and Quah (1999) and Durlauf, Johnson and Temple (2005).

² The EU-13 members are Bulgaria, Croatia, Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Romania, the Slovak Republic and Slovenia.

³ Corruption is not a new issue to discuss but has its roots 2000 years ago. The word corrupt when used as an adjective literally means ‘utterly broken’ and was first used by Aristotle.

⁴ 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.

⁵ Barro (1997) underlined the importance of accounting for the possibility of reverse causality. Especially in open economies, a positive coefficient may reflect a positive association between growth opportunities and investment since there is a choice between investing at home or abroad.

⁶ It is generally accepted that government expenditure affects growth in the short run whereas public debt has a long-run impact on growth. For instance, when government spending is extremely high inevitably leads to high public debt.

⁷ Several scholars have criticized the robustness of the empirical findings of the openness–growth link, especially on methodological and measurement grounds (see e.g. Levine and Renelt [1992], Rodríguez and Rodrik [1999] and Vamvakidis 2002).

⁸ Luxembourg was excluded as it is typically considered an outlier in the relevant empirical literature.

⁹ EU-14 members are Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Portugal, Spain, Sweden and the United Kingdom.

¹⁰ See Baumol (1986), De Long (1988), Grier and Tullock (1989), Barro (1991), Barro and Sala-i-Martin (1992) and Sala-i-Martin 1996.

¹¹ See Durlauf and Quah (1999) for a survey of the convergence testing literature based on both cross-sectional and time-series data.

¹² In a first stage, the pooled OLS method was used to obtain some preliminary results regarding the significance of the included determinants.

¹³ GMM methodology entered the growth literature with Caselli, Esquivel and Lefort (1996).

¹⁴ The results are available under request.

¹⁵ Based on the relevant theory, the speed of conditional convergence is calculated as $\left(1-{\mathrm{e}}^{-bT}\right)/T$, where b is the estimated coefficient of the lagged income variable.