Is public capital productive in Europe?

Abstract

This paper addresses the issue of whether and by how much public investment or public capital can increase GDP. In comparison with the literature on the subject, we apply many different methodologies to answer these questions. A vector autoregressive (VAR) model (for France, Italy, Germany, the UK and the USA), a panel composed of 6 European countries (Austria, Belgium, France, Germany, Italy and the Netherlands) and a regional panel (French regions) are estimated. Public investment is shown to be a significant determinant of output; this is also true for public capital but to a lesser extent than public investment with a VAR methodology. The size of the estimated coefficient is also more realistic than those obtained in the literature. This empirical result confirms that the focus of some economists on safeguarding the level of public investment is not misplaced. The debate on the introduction of a ‘golden rule of public finance’ in the European Monetary Union is legitimate in this respect.

Keywords:

• public capital
• public investment
• VAR
• panel data
• European countries

JEL classifications:

• E62
• H54
• C23

Acknowledgements

We wish to thank an anonymous referee for helpful comments on a preliminary draft of this paper. We are also grateful to participants of Journées Internationales d’Economie Monétaire et Bancaire, University of Lille 3, June 2006, for their remarks.

Notes

1. Another novelty had to do with the sustainability of public finances which is now a specific objective of the SGP; the debt to GDP ratio below 60%, which was absent from the former regulations dealing with the SGP, is now a key component for the assessment of ‘viable public finances’. Public debt hence serves as a means for discriminating between ‘good’ and ‘bad’ countries and for blaming the latter; Member States with high debt ratios have to reduce their deficit further and more quickly than Member States with low debt ratios.

2. See also EC (2003) and notably, Table III.3 which shows that fiscal consolidation induced by high debt levels and the need to satisfy the Maastricht criteria co‐incided with relatively larger cuts in public investment.

3. This channel was at the core of Aschauer (1989b) paper. He showed that the slowdown in productivity growth in the US private sector during the 1970s and the 1980s was the consequence of a shortage of investment in public infrastructure.

4. These complementarities are at the heart of a recent contribution by Martinez‐Lopez (2006) applied to Spanish regions.

5. In comparison with Table 1 in Kamps (2005), we did not report studies for which statistical significance of estimated effects were missing.

6. The main impulse response functions (IRFs) are reported in Figures 1–6. Other IRFs are available from the authors upon request.

7. This break is ad hoc in statistical terms, but it has an economic meaning: the change in US monetary policy in the early 1980s had major consequences on European economies (see Fitoussi 1995).

8. Estimations on the same subsample for the USA and the UK blur the first results.

9. See Creel and Poilon (2006) for details. It is noteworthy that we did not expressed variables in first difference. The reason is that panel‐induced common response to non‐stationarity might not be appropriate if the variables are subject to measurement errors and some are stationary (Garcia‐Mila and McGuire 1992).

10. A sectoral decomposition of investments is not available over a long period for France. The sectors are: agriculture, industry (excluding construction), hotel and restaurants, wholesale and retail trade, transport, storage and communication, finance services, public administration, education, health.