Regional innovation system determinants in Croatia
Katarina BAČIĆ[1]
Zoran ARALICA[2]
Abstract
According to the Innovation Union Scoreboard from 2014 (EC, 2014b) metrics, Croatia is a moderate innovator along with several other new EU members. The latest IUS indicators for Croatia show the slowing of innovation activity due to prolonged recession circumstances and point to the risk of regressing into the modest innovator group. So far innovation process determinants in Croatia have been researched on the national level alone (Radas and Božić, 2009). Given the fact that indicators of innovation activity have regressed in the Age of Austerity, a detailed assessment of innovation determinants on the regional level provides an additional insight into interrelatedness of the innovation processes and performance. Using NUTS 2 level classification that was valid up to 2013, we found statistically significant differences in innovation performance among three Croatian regions by using analysis of variance (ANOVA) on a sample of 3.404 innovative firms in the period 2006-2008 and 3.390 innovative firms in the period 2008-2010. The innovation performance of Northwest Croatia is superior across various types of innovation. These findings can be explained with Northwest Croatia’s technological structure of industries that is more technology- and knowledge-intensive than in the other two regions, thus contributing to the more competitive international position of firms. Within-firm factors were also found important, attributing more to regions with larger firms and higher number of employees with university degree. The paper aims at providing relevant research findings for the configuration of national innovation system and for national innovation policy.
Keywords: regional innovation system, Community innovation survey data, innovative firms, innovation performance, analysis of variance (ANOVA)
1. Introduction
There has been a growing awareness about regional innovation performance in the last twenty years. The main reasons lie in the fact that the knowledge accumulation consisting of knowledge creation and knowledge diffusion, is also perceived as a regional phenomenon (Evangelista, et al. 2001). Recent European Union (EU) knowledge economy agenda (EU 2020) reflects this point of view in that it sees a new role for innovation. It is perceived as a facilitator of “inclusive” and ‘sustainable’ growth, contrary to the previous agenda where the emphasis was on “smart” growth alone (cf. Lazonick, Mazzucato, 2013: 1093). Following this context, institutions’ development and their (in)formal interactions with other institutions and/or companies on the regional level may be considered a prerequisite for efficient firm innovation performance. The systemic approach is therefore valid both at the national and at the regional level, as clearly reflected and refined in a definition of Regional Innovation System (RIS) by Doloreux (2004: in EC, 2014b)) where RIS “is a set of interacting private and public interests, formal institutions, and other organisations that function according to organisational and institutional arrangements and relationships conducive to the generation, use and dissemination of knowledge”.
Acknowledging the fact that the regional institutional setup and policy in Croatia are still not present per se, we follow the uncritical approach, that does not differentiate the between the type of RIS, but considers every regional innovation system as RIS regardless of whether some constituent elements are missing. To the best of our knowledge, there has been no research on what role RIS have played in innovation output in Croatia. Assessment of innovation determinants on the regional level may contribute to a better understanding of the impact of the spatial and the economic dimension such as within-firm differences and between firms’ determinants on innovation activity. Using NUTS 2 level classification that was valid up to 2013, we can differentiate among innovation output in North-West (NW) Croatia, Central and Eastern (CE) Croatia and Adriatic Croatia. Our hypotheses are focused on the interplay between regional innovation performance and innovation dimensions. The goal is to recognise which systemic dimensions can be associated with the regional innovation performance.
Our main hypothesis is based on the “structural” approach: innovation performance of regions with higher shares of knowledge- and technology-intensive industry structures will be relatively more dynamic in the phase of economic boom as, but also in the phase of economic downturn due to a competitive edge that allows an international perspective and less reliance on domestic economy. Hence, we expect differences in regional innovation output to become more visible during recession period due to the differences in the technological and knowledge level of industrial structures.
Centrality hypothesis: Given the lack of regional innovation infrastructure, institutional dimension of regions appears irrelevant for the innovation performance of firms. Yet, centrality of NW region due to the presence of the capital city that hosts national institutional infrastructure provides comparative advantages (pro-innovation environment due to more qualified workforce, more competitive environment for firms, availability of suppliers) should result in more dynamic innovation performance in NW Croatia.
To this end, we use analysis of variance (ANOVA) on a large sample of firm-level innovation data across regions. This allows us to capture both development and economic innovation determinants on the regional level.
The purpose of research in this paper is to set an integrative approach to RIS in Croatia, to examine the relations between the national innovation system (NIS) and RIS as well as to learn what have been the differentiating factors of regional innovation performance in Croatia. The paper aims at providing relevant research findings for the configuration of national innovation policy in a systematic perspective.
2. Conceptual Framework
Two avenues of researches have been emerging conceptually in the last twenty years. The first group of literature was focused on analysis of regional innovation determinants and their influence on innovation output. Theoretical foundation of this group group of literature, stretching from Schumpeter (1934) to the new growth theory (e.g. Romer 1990; Grossman and Helpman, 1994; Aghion and Howitt, 1998), argues that innovations lead to the increase in firms’ productivity. New growth theorists argue that knowledge produced by individuals firms in turn upgrades competiveness of the industry and enables growth of the national economy (Growth, Helpman, 1994). Following this logic, whereby firms outperform their counterparts in terms of innovations and influence the competitiveness of industries and the national economy could be applied to the subnational (regional) level. The second group of literature is focused on the system of innovation logic (e.g. Edquist, 2004, Cooke et al, 2004). Authors within this group analyse specific territorial models of innovation combining spatial agglomeration, intensive and informal knowledge flows and networking, with a focus on the optimal innovative practice (Carricazeaux, Gashet, 2006). The general indications, drawn from the recent theoretical and empirical literature in accordance with the latter approach, are that the process of technological accumulation takes place at local or regional level, even in the era of globalisation, and that technological spillovers tend to be highly concentrated at the geographical level.
At the same time knowledge creation and knowledge diffusion within specific region require strong interaction among the stakeholders (firms and institutions) which facilitate development of RIS. These propositions explain why regions have become fundamental units of analysis in the cost/benefit evaluation of the EU economic integration and in the studies which examine the process of economic convergence (or divergence) in Europe. There are two types of measurement of RIS (EC, 2014a: 76): the linear approach that builds an innovation system on a structure (such as Regional Innovation Scoreboard) and the dynamic approach that has as a goal to observe the dynamic capacities of innovation system.
The sectoral structure of a regional economy and its effect on innovation activity may be observed through these concepts. Though the structural approach, the structure of innovation activities, their complexity and R&D in a region will come as a result the manufacturing and service sector specialisation/diversification. A region with presence of a technology-intensive or knowledge-intensive structure (i.e. advanced regions) will have a better chance at improving its competitiveness. It seems that core regions and peripheral (non-core) regions present a suitable typology for analysis (McCann, Ortega-Argilés, 2013). In this context, the core regions should be characterised by a productive system based on medium-high and high-tech industrial sectors with strong reliance on innovation activities. On the other hand, less advanced regions are based on low-tech industries combined with low knowledge-intensive service sector. Following this approach, regional innovation progress and performance come as a result of sectoral structure changes.
The empirical studies which examined regional characteristics focused on the existence of RIS (Evangelista et al, 2001) and analysed specific innovation determinants within national borders. Moreover, a comparative analysis between Western European mega-region and Eastern European mega-region verified the existence of specific innovation patterns in performance among Eastern European countries that were based on low technology and medium-low technology (Heidenreich, 2009). In the case of Spain and UK, comparative analysis by Mate-Sanchez-Val and Harris (2013) emphasised the role of firm development stage in innovation performance. Firms in Spain lagged behind the UK counterparts in capturing benefits from R&D activities. Additionally, linkages with international markets have proven to be more important for the UK companies.
3. Regional innovation system – empirical findings
Croatia’s national innovation system (NIS) development has been extensively researched and could thus represent a starting point in considering the setup of RIS in Croatia. Strong and weak elements of the innovation system found at the national level are also likely to be found at the regional level. On the contrary, the regional level analysis can shed more light on the innovation output/performance dynamics seen at the national level. Innovation Union Scoreboard 2014 (EC, 2014b) revealed that Croatia’s innovative position relative to the other EU members has deteriorated. Although Croatia remained within the moderate innovator group, the country runs a risk of regressing into the modest innovator group since 2011, when its rate of innovation activity began to slow down. Poor innovative performance of the country can be associated with the 2008-2014 recession period, implying that the Croatian innovative capacity is strongly dependant on national economic trends. This certainly raises some questions as to whether there is a functional NIS in place, and if there is one, if it is appropriately set up? Institutional support needs of innovative firms in a transition economy may be different than in a mature market economy. Another issue to consider is that the legal and institutional environment in Croatia has been exposed to continual changes in the last 20 years, both in the light of the economic transition and well as in the light of the EU enlargement. Yet some key reforms that could have relieved firms’ fiscal and legal burden have not been carried out, including the public sector reform. These types of changes require a lot of adaptation practices in businesses, and innovation planning may become more difficult and costly in an environment that is continually legally and institutionally changing.
The sectoral structure within region could be relevant for the innovation capacity of Croatian regions. As it appears that RIS in Croatia have not emerged as a result of a systemic approach, the innovation capacity may strongly depend on the sectoral structure of the regional economy. Rondé and Hussler (2005) found that French innovation system was more regional than sectoral and innovation was more likely to occur within neighbouring industries in regions, as opposed to the innovations within the same industry in different regions. That is why we presume that, given the lack of institutional support, there will be reliance on internal innovation strengths (within firm) or supplier-to-buyer cooperation as a source of innovation while there will be lack of university-firms type of cooperation. Import-export activity and foreign direct investment (FDI) can also be seen as a source of innovation in the former context, as was corroborated in Hashi and Stojčić’s research (2013): knowledge spillovers generated through international trade play an important role in innovation activities.
That is why we include indicators of trade activity in this paper and expect that regions that are more open to trade to have better innovation output results. Innovation performance will also depend on the sectoral structure and development level of regions, as this is reflected in GDP per capita level and net wages. The Northwest (NW) Croatia’s development and trade indicators such as export and import per capita indicators are outstanding compared to the other two regions. The Adriatic Croatia is a runner-up region, and Central and Eastern Croatia’s trade indicators are at the lowest level both in 2006 and in 2010 (table 1). These indicators will enable us to understand the development of RIS elements and their relationship to innovation output across regions and question where the asserted prepositions can be applied to Croatian regions. A shift is the sectoral structure is most pronounced in NW Croatia – a 3.3% and 0.3% loss in industry and agriculture share to the rising market and public services share (3.6% and 0.8% increase respectively). The trends are comparable in the other two regions, but the magnitude of the sectoral shift is not as wide as in NW Croatia. Another distinction is the disproportionate increase in the public services share to market service share in the other two regions compared to NW Croatia. The dynamic of regional income and regional trade indicators in 2010 compared to 2006 is quite similar both in NW Croatia and in Adriatic Croatia – roughly above 16% rise in monthly earnings and in the regional GDP in EUR, above 15% increase in GDP per capita in EUR, above 8% increase in export per capita in EUR and the reverse trend in import per capita in EUR – decrease in the rough range between 11-13.5%. CE Croatia, according to its development indicators, significantly lags behind the development dynamic with its income per capita increasing by less than half the rate of the more advanced regions in the observed period. Export per capita increase, on the other hand, is comparable to the other two regions.
Table 1: Selected economic data within the statistical regions in 2006 and 2010
2006 / 2010 / 2010/2006NW
Croatia / CE
Croatia / Adriatic
Croatia / NW
Croatia / CE
Croatia / Adriatic
Croatia / NW
Croatia / CE
Croatia / Adriatic
Croatia
1 / 2
Shares of persons employed in legal entities across sectors (%) / (2-1)
Agriculture / 1.1 / 5.8 / 1.9 / 0.8 / 5.7 / 1.6 / -0.3 / -0.1 / -0.3
Industry / 33.6 / 36.9 / 28.6 / 30.3 / 36.1 / 28.2 / -3.3 / -0.8 / -0.4
Services / 65.3 / 57.3 / 69.5 / 68.9 / 58.2 / 70.2 / 3.6 / 0.9 / 0.7
Public / 21.9 / 28.3 / 24.2 / 22.7 / 30.0 / 25.5 / 0.8 / 1.7 / 1.3
Average monthly paid off net earnings, in EUR* / % change (2/1)
667 / 556 / 611 / 783 / 653 / 712 / 17,5 / 17,5 / 16,5
GDP, in million EUR
18.397 / 8.333 / 12.372 / 21.336 / 8.608 / 14.497 / 16,0 / 3,3 / 17,2
GDP per capita, in EUR
11.037 / 6.340 / 8.480 / 12.738 / 6.746 / 9.875 / 15,4 / 6,4 / 16,5
Export per capita, in EUR
2.651 / 1.172 / 1.513 / 2.878 / 1.250 / 1.638 / 8,6 / 6,6 / 8,2
Import per capita, in EUR
7.357 / 1.098 / 2.115 / 6.537 / 1.021 / 1.831 / -11,2 / -7,0 / -13,4
*Aggregated across counties using country shares in regional employment as weights.