Regional innovation evolution and economic performance

ABSTRACT

Conceptual reflections and empirical evidence on the evolution of regional innovation processes are increasing in recent years. This paper contributes to this debate by focusing on the long-term implications that evolutionary changes in regional innovation patterns – intended as alternative combinations of territorial structural conditions and phases of the innovation process – can have on regional economic dynamics. By applying the regional innovation pattern framework in a dynamic perspective, it shows that structural changes in regional innovation patterns positively influence regional economic performance. From these results, reflections on regional innovation policy in the European Union context are drawn.

KEYWORDS:

• regional patterns of innovation
• evolutionary dynamics
• economic dynamics

JEL:

• O31
• R11

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

ORCID

Roberta Capello http://orcid.org/0000-0003-0438-6900

Camilla Lenzi http://orcid.org/0000-0002-0880-3516

Notes

1. NUTS = Nomenclature of Territorial Units for Statistics.

2. Regional innovation patterns represent a ranking in terms of innovation capacity and complexity, but not necessarily in terms of economic performance. In previous research we demonstrated that in fact each type of pattern has an efficiency in generating growth (Capello & Lenzi, 2013b).

3. The relevance of complementing and combining intra- and extra-regional relationships has been convincingly illustrated by Bathelt, Malmberg, and Maskell (2004) in the debate on local buzz and global pipelines.

4. Implicitly, then, the RIS approach suggests that both the knowledge creation and the knowledge exploitation subsystems should be fully developed and interacting in order sustain local innovation at most, and that the weaknesses, unbalances and/or underdevelopment of any of the two undermines local innovation capacity. For a similar discussion, see Capello (2015).

5. This perspective has also been applied in the frame of the RIS approach, with the goal to study the emergence and unfolding of new industrial paths of local development (Trippl et al., 2017) and not the evolution of regional innovation structures.

6. For a thorough theoretical discussion of the notion of path dependence and its application in a spatial perspective, see, among others, Henning et al. (2013). For a more in-depth discussion of the notion of path dependence, see, among others, Sydow, Schreyögg, and Koch (2009), Vergne and Durand (2010), and Garud, Kumaraswamy, and Karnøe (2010).

7. The discussion of the local conditions enabling the different pathways/strategies is out of the scope of this paper and is fully presented in Capello and Lenzi (2018a, 2018b).

8. The authors thank an anonymous reviewer for suggesting this interpretation.

9. Appendix A in the supplemental data online provides a summary of the variables used in the cluster analysis implemented to detect innovation patterns in European regions and the variables representing the key territorial features of the different groups of regions. For further details, see Capello and Lenzi (2013a).

10. In our previous research, we demonstrated that changes in innovation patterns are achieved as the result of long-term intentional behaviours and actions of local actors (in other words, strategies) that drive the region to cumulate the characteristics favouring that particular change. The accumulation of an endowment of particular characteristics, in fact, is not random but the outcome of years of specific behaviours and actions. In particular, the endowment of specific characteristics supporting the presence of a complex learning trajectory/paradigms in a region belonging to a less complex trajectory/paradigms indicates that that particular region has a higher probability to change its innovation trajectory/paradigms. This approach is indeed consistent with the view that a change occurs when niche, deviant behaviours become dominant and prevailing at the regional level (Simmie, 2012). The presence of a change is therefore the outcome of longer term and slow evolution processes and strategies. In this way, we cannot establish and observe the exact starting point in time, but only the final stages; this is, however, not of detriment to our reasoning. This also explains the time span used to compute the variable measuring the change in innovation patterns, even if structural long-term effects are the subject of analysis. Finally, two additional reasons help to explain the choice of the two periods considered. First, innovation data should be measured before the crisis period as to avoid possible co-founding effects of the crisis upon firms’ innovative behaviour. Second, there are constraints in the availability and comparability of innovation data over time. Innovation data used to identify regional innovation patterns are drawn from Community Innovation Surveys (CISs), which show limited comparability over time, especially in earlier editions before the crisis period. Before 2007, only the 2002–04 and 2004–06 surveys can ensure full comparability.

11. This is not a counterintuitive result since the changes analyzed are of a structural nature.

12. For a similar approach, see Capello and Lenzi (2016b).

13. Regions in the European science-based areas (20 observations) by definition cannot experience any upward change. In the empirical analysis, therefore, they have been excluded. A robustness check of the results of estimation of equation (2) has been carried out also by including these regions. Results are qualitatively unchanged and are available in Table A7 in Appendix A in the supplemental data online.

14. The period 2009–12 is more volatile presenting both recovery and decline with respect to the others (see Figure A1 in Appendix A in the supplemental data online). The other periods instead present a clearer trend and are therefore preferable as they enable a clearer interpretation.

15. Alternative methods have been developed and applied in the literature as to compute the initial capital stock (e.g., Bottazzi & Peri, 2007). We adopted the method proposed by Dettori et al. (2012) and Marrocu et al. (2013), which, differently from others, has been conceived in a spatial framework.

16. As shown in the following sections, our results are robust to alternative coding of the variable accounting for the change of innovation pattern.

17. Country group dummies are defined as follows: South, the reference case (Greece, Italy, Portugal and Spain); North (Denmark, Finland and Sweden); West (France, Ireland and the UK); Centre (Austria, Belgium, Germany, Luxemburg and the Netherlands); Baltic countries (Estonia, Latvia and Lithuania); Central and Eastern Europe (Czech Republic, Hungary, Poland, Slovakia and Slovenia); Romania and Bulgaria; Malta and Cyprus. The Eurozone dummy variable takes value 1 for regions in the following Eurozone countries: Austria, Belgium, Cyprus, Germany, Greece, Spain, Estonia, Finland, France, Ireland, Italy, Latvia, Luxemburg, Malta, the Netherlands, Portugal, Slovakia and Slovenia. Lithuania was excluded because adoption of euro occurred in 2015, after the period under consideration.

18. When including interactions, all the interaction terms, i.e., interaction effects and simple effects, should be included unless there is a good (theory-based) reason not to do so as to avoid the risk of a possible bias due to the omission of the individual terms of the interaction (Brambor, Clarck, & Golder, 2006). Interpretation should be done in relative terms with respect to the reference case (in this case the Imitative innovation area) and not in absolute terms.