European Union and nationally based cohesion policies in the Italian regions

ABSTRACT

This paper analyzes the effects of cohesion policies on the gross domestic product (GDP) per capita of the 20 Italian administrative regions for the period 1994–2013. The analysis includes both European Union and national funds. It estimates average partial effects through a control-function approach based on the funds’ allocation rules, and allows for the role of the regional environment on the impact of regional policies. A positive impact of European Union funds is found, as well as a less significant impact of (nationally financed) subsidies to firms. Quality of government has no relevance for European Union funds, but it enhances the impact of subsidies to firms.

KEYWORDS:

• European Structural Funds
• nationally based regional policies
• quality of government
• Mezzogiorno
• dynamic panel data

JEL:

• O11
• O43
• O47
• R58

ACKNOWLEDGEMENTS

The authors are thankful to the referees as well as the participants to the AISRE 2016 and SIE 2016 meetings for helpful comments made on a previous draft of this paper; to Giuseppe Taliano for helpful clarifications on some data issues; and to Alessia Brinciotti for skilful research assistance. The usual disclaimer applies.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

ORCID

Gianluigi Coppola https://orcid.org/0000-0002-7715-0950

Sergio Destefanis https://orcid.org/0000-0003-3568-3414

Notes

1. The treatment effect framework is beginning to catch on in the literature on the effectiveness of European regional policy (Becker et al., 2010, 2012; Pellegrini et al., 2013). These papers, which generally find a positive impact of cohesion policies on economic growth, are however based on a cross-sectional regression discontinuity design and do not explicitly deal with the policy assignment mechanisms.

2. A more complete specification would display (D.pop_it + g + d), where g is the rate of technological change; and d is the rate of depreciation. As g and d are not observable, one can substitute D.pop_it for (D.pop_it + g + d) in actual estimation, and use this variable as a simple control. On this, see Arnold, Bassanini, and Scarpetta (2007).

3. Interestingly, Arnold et al. (2007) show that equation (1) also encompasses the structure of the Lucas–Uzawa growth model. In order to discriminate between these competing models, one should have a measure for the rate of accumulation of human capital, which is lacking here. Yet, provided that the omission of this variable does not bias the estimates, equation (1) allows one to measure the impact of cohesion policies on GDP per capita in a set-up compatible with two very important growth models.

4. As we include year-fixed effects in equation (1) and the estimation period relates to three SFs’ programming periods, an additional PERIOD_1*SOUTH interactive term would lead to a dummy variable trap.

5. For reasons of data availability, we could not produce a series of capital account subsidies to firms separated by the rest of capital account expenditures.

6. These estimates are available from the authors upon request.

7. We also tested for the presence of contemporaneous policy effects, which turned out to be insignificant.

8. These values become 2.2% and 2.8% respectively when the Bun–Kiviet bias approximation formula is used for the lagged dependent variable coefficient; full estimates are available from the authors upon request.

Additional information

Funding

Gianluigi Coppola and Sergio Destefanis gratefully acknowledge financial support from the University of Salerno [grant number FARB 2016].