Decentralization and economic growth in Europe: for whom the bell tolls

ABSTRACT

The effects of fiscal decentralization on economic growth are analyzed. A theoretical framework is developed that builds on the relationships between government size and growth and between decentralization and government size. The framework is tested empirically on a panel of 25 European countries observed between 1995 and 2015. The econometric results show that the relationship between expenditure decentralization and growth is bell shaped. The paper also focuses on expenditure composition. In this respect, the relationship between investment decentralization and growth is an inverted bell-shaped curve: there is a critical mass of decentralized investments beyond which it is possible to enhance growth.

KEYWORDS:

• decentralization
• capital expenditure
• economic growth

JEL:

• H7
• O4

DISCLOSURE STATEMENT

The views and opinions expressed in this paper are those of the authors and do not necessarily reflect the official position of Prometeia. No potential conflict of interest was reported by the authors.

Notes

1. This restriction allows a positive maximum in the interval $\left(G/Y\right)\in (0,1)$, since obviously government size cannot be greater that 1 or lower than 0 (see Appendix A in the supplemental data online).

2. Many empirical papers find that intergovernmental transfers increase local expenditure (Grossman, 1989; Oates, 1993, 2005; Ahmad & Searle, 2006).

3. In fact, this restriction on parameters implies a positive minimum in the interval $\left(G_{sn}/G\right)\in (0,1)$. In fact, the decentralization index cannot be greater that 1 or lower than 0 (see Appendix B in the supplemental data online for the algebra).

4. In this case, the function described by equation (2) is monotonically decreasing in the interval $\left(G_{sn}/G\right)\in (0,1)$ (see Appendix A in the supplemental data online). The third section shows the possible implications of these restrictions on parameters.

5. Some local policies could produce catching-up phenomena in growth rates (Brakman, Garretsen, & Van Marrewijk, 2002).

6. The countries considered are: Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland and the UK. Switzerland is included even if it does not belong to the European Union, while Croatia, Romania, Malta and Cyprus are excluded because there are too many gaps in the data.

7. Table B1 in Appendix B in the supplemental data online sets out the variables used in the analysis, with their sources. Data on per capita real GDP, denoted by y, and expressed in euros at constant 2010 prices, are from the International Monetary Fund (IMF).

8. This is the standard measure of fiscal decentralization widely used in previous studies. The most common database come from the IMF and the OECD. As acknowledged by many scholars (Ebel & Yilmaz, 2002; Stegarescu, 2005), given that decentralization is a complex and multifaceted concept that spans fiscal, political and administrative dimensions, this index of decentralization does not perfectly reflect such complexity. However, for cross-country comparisons, no reliable alternatives are possible.

9. We also tried alternative specifications including the RAI index which is a measure of the degree of devolution in regions. It reflects, in part, the fiscal autonomy of regions. The results indicate that when both decentralized expenditures and the RAI index are included in the model as covariates, the effect of devolution, for a given level of decentralized expenditures, is not statistically significant, mainly because these two variables are quite correlated (Pearson correlation coefficient = 0.4788).

10. This empirical strategy follows the approach of other previous studies (e.g., Facchini & Melki, 2011). An alternative model could be a growth equation in which the per capita growth rate enters directly as a dependent variable. The advantage of the model adopted in this paper is that it allows one to gain degrees of freedom.

11. Stationarity tests results are available from the authors upon request.

12. According to the results, the government size that maximizes the growth rate of GDP per capita is given by: $\text{exp( }-\left({\hat{\alpha }}_1/2{\hat{\alpha }}_2\right))=0.372$.

13. The results are available from the authors upon request.

14. The Barro–Armey curve still remains valid since equations (1) and (5) do not consider any decentralization ratios.

15. The share of subnational investment is in most countries around 60%.