Portuguese regional innovation systems efficiency in the European Union context

ABSTRACT

Current evidence on European regional innovation systems efficiency shows some conflicting results. Whereas some studies find support to a core-periphery distribution of efficiency, others find that lagging regions can be as well or even more efficient than rich regions in using their resources. This paper contributes to this debatable topic by providing additional evidence on the main determinants of the region's innovation efficiency and on efficiency differentials across EU regional innovation systems. Using data from 206 European regions and applying a stochastic production frontier methodology, our results corroborate the importance of interactions among regional agents on the region's efficiency score. More importantly, the distribution of efficiency scores across regional innovation systems does not entirely confirm the core-periphery divide among European regions. Instead, the mode of doing innovation appears to be a crucial explanatory factor of innovation efficiency at regional level. In the case of Portuguese regional innovation systems, they perform slightly below the average of their EU counterparts, except Lisbon's, and appear to be constrained by their mode of doing innovation.

KEYWORDS:

• Regional innovation systems
• production frontier
• technical efficiency
• European Union

JEL Codes:

• O11
• O18
• O32
• O47

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Two methods can be used to measure efficiency, a deterministic one – Data Envelopment Analysis (DEA, or a stochastic one – Stochastic Frontier Analysis (SFA). DEA has been found more competent in analysis of multi-output scenarios (e.g. Guan & Chen, 2010, p. 2012), with the additional advantage of not imposing an explicit functional form for the underlying technology and an explicit distributional assumption for the inefficiency term. Yet, it has the cost of not controlling for unobserved factors and statistical noise.

2 See EC (2014) for a description of the normalization procedure.

3 In the case of Portugal only five regions were included in the analysis (Norte, Centro, Lisbon, Alentejo and Algarve); Autonomous Regions of Madeira and the Açores were excluded due to lack of data.

Additional information

Funding

This work was carried out within the funding with COMPETE reference POCI-01-0145-FEDER-006683 (UID/ECO/03182/2013), with the FCT/MEC's (Fundação para a Ciência e Tecnologia) financial support through national funding and by the ERDF through the operational program on “Competitiveness and Internalization-COMPETE 2020 under the PT2020 Partnership Agreement.”