Is Smart Specialisation Strategy coherent with regional innovative capabilities?

ABSTRACT

The implementation of the Smart Specialisation Strategy (S3) has required European regions to identify the technological domains in which they show superior innovative capabilities. This choice should promote specialization and facilitate diversification into new sectors. Given that regional specialization shows path dependence, successful diversification can be achieved in domains closely related to the existing knowledge base. The paper provides a first empirical assessment of the coherence between the technological domains chosen by Italian regions and those in which they show actual innovative capabilities, as measured by their patenting activity.

KEYWORDS:

• Smart Specialisation Strategy
• innovative capabilities
• regional development
• regional innovation policy
• open data

JEL:

• L52
• O38
• R11
• R58

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

ORCID

Diego D’Adda http://orcid.org/0000-0002-1683-1787

Donato Iacobucci http://orcid.org/0000-0001-8463-1106

Notes

1. We use indicators of technological specialization in order to measure the existing innovative capabilities within a region. In particular, for each technological domain, we use three different measures: (1) the relative specialization (using a Balassa index); (2) the absolute specialization (using a binary indicator equal to 1 when the number of patent is above a certain threshold); and (3) the increase in the specialization (the presence of a positive trend in relative specialization). See the data and methodology for further details.

2. Here, the term ‘coherence’ refers to the degree of ‘overlap’ between the technological domains as declared in S3 documents by regions and their actual knowledge base. Using S3 terminology, it relates to the concept of embeddedness: S3policies have to be embedded in the local reality, in local assets and strategic design capabilities (Camagni & Capello, 2013; McCann & Ortega-Argilés 2015). In other words, they have to show coherence with the regional existing capabilities. Note that in this paper the term ‘coherence’ does not refer to the construct used, for example, by Nesta and Saviotti (2005), Quatraro (2010) and Rocchetta and Mina (2018). In fact, we are not measuring the degree of internal coherence of the knowledge base with which a region is endowed. This latter concept refers to the coherence within a regional knowledge base, or, more specifically, the coherence between the knowledge domains in which a region is specialized. Using S3 terminology, this type of coherence points to the concept of relatedness (e.g. McCann & Ortega-Argilés, 2011).

3. For example, Foray (2015, p. 84) states that ‘the centre of gravity of the smart specialisation dynamic is the firms since they are best placed to conduct entrepreneurial discovery processes.  … The strategy is much more broadly a tool for economic development through research and innovation that must associate all the actors concerned in projects not necessarily centred on public research or universities’.

4. See www.wipo.int/ipccat/.

5. For a list of all the technological domains indicated by Italian regions and the corresponding IPC classes, see the Appendix in the supplemental data online. Abruzzo is not included among the regions analyzed because the S3 document indicates the general domains of specialization (namely, agrifood, life science, information and communication technology (ICT)/aerospace, fashion/design), but did not provide a detailed description of the technologies included in these domains. This made it impossible to determine the IPC classes associated with the specialization domains.

6. By way of example, say that a region is specialized (has a normalized RCA index > 0) in IPC classes A01 (agriculture; forestry; animal husbandry; hunting; trapping; fishing) and A21 (baking; equipment for making or processing doughs; doughs for baking) and it is not specialized in A22 (butchering; meat treatment; processing poultry or fish). Say that the same region has declared in its S3 strategy that the priority technological domains are A21 and A22 but not A01. In this case, the regional strategy is ‘coherent’ for what concerns the choice of A21 and not coherent for what concerns the choice of A22, since it is not specialized in this latter technological domain. The resulting overall coherence index based on the relative specialization measure will be 0.5 (one domain chosen in which the region is actually specialized over the two domains chosen by the region).

7. The latter result is not surprising if we consider that the number of technological domains chosen by regions is only one-third of those in which they showed a relative specialization.