1. Introduction
In the context of the 2030 Agenda for Sustainable Development (SDG) (https://sdgs.un.org/2030agenda), entrepreneurship plays a pivotal role in enhancing society's quality of life, including for disadvantaged groups, as it contributes to building resilient infrastructure, promoting inclusive and sustainable industrialization, and fostering innovation. Entrepreneurship is closely linked to SDGs 4 and 8, reviewed in 2019. SDG target 4.4 aims to significantly increase the number of youth and adults with relevant skills, including technical and vocational skills, for employment, decent jobs, and entrepreneurship. Simultaneously, SDG target 8.3 supports development-oriented policies that promote productive activities, decent job creation, entrepreneurship, creativity, and innovation while encouraging the formalization and growth of micro-, small-, and medium-sized enterprises (MSMEs) as critical agents for and beneficiaries of inclusive development through access to financial services.
Entrepreneurs with a strong commitment to sustainable development contribute to achieving almost all SDGs as they create businesses that support employment, alleviate poverty, and enable decent work and economic growth. Furthermore, entrepreneurship benefits efforts to reduce hunger, promote good health and wellbeing, achieve affordable and clean energy, and strengthen industries. Thus, entrepreneurship can be the driving force behind transforming our world and overcoming diverse global challenges (Apostolopoulos, 2018).
Recognizing the importance of entrepreneurs and entrepreneurship, the European Commission has launched the "Entrepreneurship 2020 Action Plan" which aims to unleash Europe's entrepreneurial potential, eliminate existing obstacles, and modernize the culture of entrepreneurship in Europe. These goals can be achievable through fostering an entrepreneurial mindset through entrepreneurship education in higher education institutions. Universities' teaching activities shape students' entrepreneurial orientations and competencies. Recent research suggests that university entrepreneurship programs may not increase the rate of entrepreneurship, but they do help students to better identify their entrepreneurial potential and improve the performance of their startups, leading to higher success rates, improved employability, and better management skills. It is crucial to consider regional differences and analyze regional innovation indicators to foster entrepreneurial spirit, create prosperity and wellbeing, and facilitate European system growth. This necessitates understanding each region's trajectory through tools like the Smart Specialization Strategy (S3) and Smart Specialization Strategies for Sustainable and Inclusive Growth (S4). To better comprehend regional situations, we propose an "ad hoc" index reflecting the entrepreneurial tendencies of specific regional territories.
The remainder of the paper is organized as follows: Following the introduction in Paragraph 1, Paragraph 2 elucidates the theoretical underpinnings of the IRE index. Paragraph 3 delineates the methodology for selecting the appropriate variables and illustrates the steps in creating the IRE index, whereas Paragraph 4 discusses the results. Finally, Paragraph 5 provides the concluding remarks.
2. Theoretical background
The Entrepreneurship 2020 Action Plan is a comprehensive strategy for catalyzing and disseminating the culture of entrepreneurship in Europe through transforming entrepreneurial universities and nurturing entrepreneurial spirit. The plan is structured around three key pillars: entrepreneurial education and training, fostering an environment in which entrepreneurs can flourish and grow, developing role models, and engaging specific groups with untapped entrepreneurial potential or those not reached by traditional business support methods.
The first pillar emphasizes the importance of expanding and enhancing entrepreneurial education and training, which is considered one of Europe's most significant investments. The acquisition and implementation of entrepreneurship skills have also been highlighted within the framework of EU co-operation in education (Jenner, 2012). Overall, investments in entrepreneurial education are efficient tools for raising public awareness of entrepreneurs and supporting underrepresented groups among entrepreneurs. Only if a large number of Europeans perceive an entrepreneurial career as a rewarding and attractive option will entrepreneurial activity thrive in Europe in the long term.
However, Benneworth and Osborne (2015) argued that entrepreneurship education has not yet reached its full potential, partly due to poor integration with other university knowledge activities. The authors suggest that future research on university entrepreneurship education should focus on how entrepreneurial activities align with universities' core knowledge, providing a more coherent understanding of universities' contributions to fostering entrepreneurial attitudes.
The global configuration of knowledge and technology is one of the most critical factors influencing the pursuit of internationalization and the rise of the global economy (Posselt et al., 2019). The rapid evolution of technology and market demands necessitates the development of an entrepreneurial spirit, digital literacy, and innovative learning and teaching methods. In this context, entrepreneurial universities must adapt by cultivating leadership, navigating complexity, adopting a lifelong learning approach, and transforming failure into success (Gibb et al., 2013).
These four essential features of entrepreneurs and entrepreneurial universities play a crucial role in promoting entrepreneurship in Europe. By focusing on these core values and principles, the Entrepreneurship 2020 Action Plan aims to revolutionize the landscape of entrepreneurship and support Europe's long-term economic growth and development.
The current literature has several well-established indexes and frameworks for assessing innovation and entrepreneurship at the national and regional levels. To provide a comprehensive overview and contextualization of the IRE Index, we have included a brief overview of several indices, in particular, the Global Innovation Index (GII), European Innovation Scoreboard (EIS), Regional Innovation Scoreboard (RIS), Global Entrepreneurship Index (GEI), OECD Science, Technology and Industry Scoreboard.
The GII provides detailed metrics that capture the state of innovation performance in countries around the world (Global Innovation Index, 2023). It includes indicators such as institutions, human capital and research, infrastructure, market sophistication, business sophistication, knowledge and technology outputs, and creative outputs. By analyzing 132 economies, the GII is seen as an "action tool" for innovation policy, as it enables policymakers to identify strengths, address weaknesses and implement effective strategies to improve their countries' innovation performance.
The EIS provides a comparative analysis of the innovation performance of EU Member States, other European countries, and regional neighbours (European Commission, 2023). It classifies indicators into four groups: framework conditions, including human resources, attractive research systems and digitalization. Second, investments cover finance and support, business investment, and information technologies. Third, innovation activities, including innovators, linkages, and intellectual assets. Lastly, there are impacts, such as employment effects, turnover effects, and environmental sustainability. Based on the scores, countries are classified into four performance groups: Innovation Leaders, Strong Innovators, Moderate Innovators, and Emerging Innovators.
While the EIS assesses innovation performance using a set of national-level indicators, the RIS (European Commission, 2021) provides a detailed regional perspective using indicators adapted to capture regional dynamics and specificities. The RIS uses a subset of the EIS indicators adapted to regional specificities. By providing insights at the regional level, the RIS allows the comparison of different regions within countries, highlighting regional disparities and identifying specific areas for policy intervention.
The GEI comprehensively measures a country's entrepreneurial ecosystem by assessing individual and institutional factors (Szerb et al., 2020). The data is collected at the individual level, such as the attitudes, abilities, and aspirations of the local population, reflecting their readiness and potential to become entrepreneurs. These metrics are then weighted against the broader social and economic infrastructure, which includes elements such as broadband access, transport links to external markets, the regulatory environment, and the availability of financial resources. This approach enables GEI to capture the population's inherent qualities and the external conditions that facilitate or hinder entrepreneurship activity.
OECD Science, Technology and Industry Scoreboard includes indicators such as R&D expenditure, innovation outputs such as patent applications and high-tech exports, and human capital measures such as tertiary education attainment levels and the number of researchers (OECD, 2023). It also tracks metrics on the digital economy, data on collaboration and networks, government policies and support mechanisms, and the wider economic impact of innovation, including productivity growth and job creation. This in-depth analysis helps policymakers, researchers, and business leaders make informed decisions to improve national and regional innovation ecosystems and drive sustainable economic development.
In the following subsections, we present an exploration of the key indicators that were scrutinized during the development of the IRE index. The aim is to inform you of their respective significance and their impact on a given region's innovation readiness.
2.1 The role of tertiary education on the population aged 25–34
Education is essential both at the individual level, as it is key to navigating the world, and at the national level, as it contributes to a country's economic and societal development. Recent research has emphasized the critical role of education as a supply-side factor in the context of entrepreneurship and innovation, particularly in tertiary education, also known as post-secondary education or advanced studies (Crecente-Romero et al., 2018; van Praag & van Stel, 2013).
Tertiary education is shaped by government policies and is seen as an effective instrument for advancing human and national development. According to Peña-Vinces & Audretsch (2021), its primary aim is to enhance employees' educational and professional levels across various specializations. This, in turn, increases the number of highly skilled workers, which boosts company performance and, consequently, the country's economy.
Furthermore, research conducted by Millán et al. (2014) revealed that education levels are positively related to entrepreneurial success, as highly educated individuals contribute to a company's prosperity by running their own businesses or providing their skills and expertise. Similarly, Jiménez et al. (2015) argued that the impact of tertiary education is twofold. On the one hand, it increases formal entrepreneurship due to higher self-confidence and lower perceived risk, while on the other hand, it reduces informal business activity. The negative relationship stems from an increased awareness of and sensitivity to the potential negative consequences of certain activities. In parallel, Barreneche García (2014) found that the more individuals are involved in tertiary education, the greater the positive entrepreneurial dynamism.
Taking together the evidence from the literature, the indicator of tertiary education is seen as a vital factor for fostering entrepreneurial activities. As a result, we considered it a general measure of the supply of advanced skills across various industries and sectors.
Figure 1. Percentage of the population aged 25–34 having completed tertiary education.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
Table 1. The selected indicators and data sources.
| Indicator | Numerator | Denominator | Data sources | Data availability |
|---|
Percentage population
aged 25–34 having
completed tertiary
education | Number of persons in age class
with some form of post-secondary
education | The reference
population is all age
classes between 25-
and 34-years inclusive | Eurostat, regional
statistics, Regional
innovation scoreboard
2021 | NUTS 2: 2012 – 2019 |
Percentage population
aged 25–64 participating
in lifelong learning | Number of persons in private
households aged between 25 and
64 years who have participated
in the four weeks preceding the
interview, in any education or
training, whether or not relevant
to the respondent's current or possible
future job | Total population aged
between 25 and 64
years | Eurostat, regional
statistics, Regional
innovation scoreboard
2021 | NUTS 2: 2012 – 2019 |
R&D expenditure in the
business sector | All R&D expenditures in the
government sector (GOVERD) and
the higher education sector (HERD) | Regional Gross
Domestic Product | Eurostat, regional
statistics, Regional
innovation scoreboard
2021 | NUTS 2: 2011 – 2018 |
Innovative SMEs
collaborating with others
as percentage of SMEs | Number of SMEs with innovation
co-operation activities. Firms with
co-operation activities are those
that have had any co-operation
agreements on innovation
activities with other enterprises or
institutions. | Total number of SMEs | Community Innovation
Survey: Eurostat and
National Statistical
Offices | NUTS 1 and 2 for
different countries
for CIS 2012, CIS
2014, CIS 2016, CIS
2018 |
PCT patent applications
per billion regional GDP | Number of patents applied for at
the European Patent Office (EPO), by
year of filing. | Gross Domestic
Product in Purchasing
Power Standard | Numerator: OECD,
REGPAT. Denominator:
Eurostat | NUTS 2: two-year
averages for 2012
- 2019 |
Sales of new-to-market
and new-to-firm
innovations in SMEs as
percentage of turnover | Sum of total turnover of new or
significantly improved products for
SMEs | Total turnover for SMEs | Community Innovation
Survey: Eurostat and
National Statistical
Offices | NUTS 1 and 2 for
different countries
for CIS 2012, CIS
2014, CIS 2016, CIS
2018 |
2.2 The role of lifelong learning in innovation activities
The European Commission defines lifelong learning as "all learning activity undertaken throughout life, with the aim of improving knowledge, skills, and competencies within a personal, civic, social, and/or employment-related perspective." In other words, it refers to all formal, non-formal, and informal learning. Furthermore, Jarvis (2007) expanded the concept of learning by specifying two types-- vocational and non-vocational- and emphasized that learning covers any opportunity to acquire new knowledge, skills, attitudes, values, emotions, beliefs, and senses through social institutions or any process.
The concept of lifelong learning can be viewed from three perspectives: individual lifelong learning, learning organizations, and learning societies (Tight, 1998). These dimensions underscore the value of learning for development at the individual, company, and national levels. For example, Coelli & Tabasso (2019) study found that engaging in education improves labour market performance and, specifically, employee outcomes such as working hours or wage rates. Moreover, lifelong learning benefits society by potentially increasing company output and tax revenues. In the context of learning organizations, McKelvey (1998) emphasized that the connection between learning and innovation enables companies to overcome challenges through their ability to learn and adapt.
For further analysis, the lifelong learning indicator is essential, as it strongly correlates with innovation activities (such as the development of artificial intelligence and nanotechnologies). Overall, lifelong learning, including formal and informal education, enhances knowledge, skills, and competencies.
Figure 2. Lifelong learning.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
2.3 The role of R&D expenditure in the entrepreneurial ecosystem
Research and Development (R&D) encompasses innovative and structured actions to obtain new knowledge and enhance existing knowledge (Eurostat). In recent years, extensive literature has focused on the relationship between R&D expenditures and firm performance. This interest stems from Schumpeter's theory, which posits that companies attempt to foster innovation in order to bolster their competitive advantages, potentially leading to higher profitability, productivity, and even market monopolies (Capasso et al., 2015). Today, interest in innovation continues to grow due to the rapid development of technologies and the availability of highly skilled labour. Another contributing factor is the opportunity to expand activities in foreign markets, causing R&D to become increasingly supply-driven as companies adopt a more global orientation (Siedschlag et al., 2013).
However, R&D investments can be challenging, as the positive impact of R&D on a company's innovative performance may increase up to a certain point, after which further R&D spending may lead to diminishing performance (Berchicci, 2013). Nonetheless, most scholars emphasize the beneficial role of R&D, particularly for a firm's productivity growth (Cin et al., 2017; Kancs & Siliverstovs, 2016; Wakelin, 2001), marketing performance (Sharma et al., 2016), employment rates (Di Cintio et al., 2017), and overall future performance (Ruiqi et al., 2017). Consequently, company growth tends to improve the economic situation within a country. For instance, research found that R&D expenditures and educational factors are most efficient in increasing GDP per capita. However, the analysis also revealed that the impact of R&D varies among regions; it is significant only in the most developed regions of Europe, while education is relevant in all cases (Sterlacchini, 2008).
R&D expenditures are considered one of the most influential drivers of economic growth at both the company and national levels. Additionally, they are crucial for transitioning to a knowledge-based economy, as they advance production technologies, leading to economic growth. For this reason, we included this measure in our analysis, as it may provide a more comprehensive view of the degree of innovativeness in European regions.
Figure 3. R&D expenditure in the business sector.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
2.4 The role of innovative SMEs and their collaboration activities
Due to their size, small and medium enterprises (SMEs) are often more vulnerable to challenges in the business environment. Nevertheless, most academic evidence suggests that innovations can effectively enhance SMEs' performance. For example, Hall et al. (2009) investigated the relationship between innovations and firm productivity, discovering that R&D intensity and investments in equipment are positively related to performance outcomes. Moreover, Laforet (2013) found that SMEs with an innovation orientation are more successful in the market, as they can quickly respond to market demand with better quality products or services.
In addition to improving domestic performance, innovations can help SMEs expand their international activities. By implementing organizational and product innovations, companies can bolster their marketing innovations. Integrating these innovations with technological advancements enables SMEs to increase exports (Bodlaj et al., 2020). Meanwhile, the analysis of Saridakis et al. (2019) observed that innovative SMEs are more likely to extend their activities overseas than their non-innovative counterparts. Furthermore, the impact of innovation on internationalization varies among SMEs depending on the nature of the innovation and the degree of novelty. Similarly, Rosenbusch et al. (2011) identified additional factors affecting innovation outcomes, such as the company's age and the cultural context.
Measuring innovative activities in SMEs advances research by revealing current market trends in terms of SMEs' adoption of open innovations within the European context. This indicator specifically focuses on SMEs implementing, promoting, and collaborating with educational organizations on innovative initiatives. More precisely, it measures the flow of knowledge between public research institutions and firms and between firms themselves.
Figure 4. Innovative SMEs collaborating with others as a percentage of SMEs.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
2.5 PCT patent applications
The context of innovation is closely related to patent applications, as they enable companies to distinguish their inventions overseas and protect intellectual property rights (Ervits & Zmuda, 2018). Patent applications are commonly used as indicators of innovation output within academic literature. For instance, Bronzini & Piselli (2016) assessed firm innovation through patenting activity, finding that implementing regional subsidy R&D programs led to an increase in patent applications, which positively impacted a company's innovation activities. Meanwhile, their analysis revealed company size as a distinguishing factor, with SMEs tending to exhibit higher intensity and likelihood of patenting than large firms. In contrast, Athreye et al. (2021) argued that larger companies can produce more patentable innovations due to existing cost barriers. Similarly, the study of Arundel & Kabla (1998) established a link between firm size and patent propensity rates. Furthermore, patents were identified as tools for protecting products and processes from being copied by competitors in the market.
Another line of research has focused on exploring the role of patent applications at the regional or national level. From a regional perspective, Lin et al. (2022) developed a two-stage model of the value-creation process within a regional innovation system, considering various patent statuses. Their findings revealed that invention patents play a crucial role in the overall performance of regional innovation development, as increasing patent applications lead to higher regional innovation scores. In another study, Whitacre et al. (2019) examined methods for measuring business innovation, creating an innovation index that allowed for the analysis of the relationship between innovation and firm- and regional-level outcomes. As a result, the research determined a positive impact of innovations on both firm and regional outcomes, with companies benefiting from increased employee wages and market share. Simultaneously, at the regional level, innovations influenced household income, the percentage of employees in the creative class, and poverty levels. In a broader view, Ervits & Zmuda (2018) conducted a cross-country analysis focusing on the impact of corruption and the business climate on patenting activity. Notably, countries with higher scores in institutional environments tend to exhibit greater incentives for patent applications. Consequently, companies, particularly SMEs, that lack ownership advantages and other resources may benefit from such environments.
Patent applications can be seen as an indicator of a firm's ability to develop new products, leading to increased competitive advantages. As such, it is a critical measure of a company's innovative activities. More specifically, we considered this indicator the number of patent applications per year within European regions.
Figure 5. PCT patent applications.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
2.6 The role of sales of new-to-market and new-to-firm innovations
Current EU innovation policies are primarily research-oriented, aiming to achieve an R&D investment rate of 3.1% of GDP. However, SMEs' innovative activities depend on internal factors, including both R&D and non-R&D factors and external factors, such as partnerships with companies and research centers. The recent study of Hervas-Oliver et al. (2023) argued that the success of innovation policies depends on a region's potential, as high investments in R&D may not necessarily lead to better performance for SMEs in less developed regions. Consequently, the degree of SME' innovation varies among European regions. The findings suggest that SMEs in more developed locations benefit mainly from all factors, while companies in less innovative regions heavily rely on external support.
Due to existing constraints, such as scarcity of resources and capabilities for developing R&D activities (Hausman, 2005), weak network embeddedness (Srinivasan et al., 2002), or a lack of highly skilled employees (Romano, 1990), SMEs are compelled to seek solutions to overcome these barriers. One possible approach is through collaboration. Golonka (2015) posits that a more market-focused co-operation strategy could enhance SMEs' innovativeness. Thus, to minimize regional disparities, more collaborative and location-sensitive policies are required to advance SMEs' innovation activities.
For further investigation, we introduced an indicator that measures the turnover of new or significantly improved products. This includes new products for the firm and new products for the market. This approach enables the capturing of the creation of cutting-edge technologies (new-to-market products) and the diffusion of these technologies (new-to-firm products).
Figure 6. Sales of new-to-market and new-to-firm innovations.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
3. Methods
To achieve the objectives of the research, the methodology for the development of the IRE index was executed through a multi-step procedure. After the above-mentioned review of extant literature to identify the critical variables that influence innovation, R&D, and entrepreneurship, with particular attention paid to the role of tertiary education to create the foundational basis for selecting the variables to be included in the IRE index, we focused on the following points:
1. Analysis of the 2021 "Regional Innovation Scoreboard". We examined the latest version of the Regional Innovation Scoreboard (RIS) (European Commission, 2021) to assess the performance of innovation systems across 240 regions from 22 EU Member States. The list of selected countries was expanded to include Norway, Serbia, Switzerland, and the United Kingdom. Cyprus, Estonia, Latvia, Luxembourg, and Malta were also considered, as in these countries, the Nomenclature of territorial units for statistics (NUTS) 1 and NUTS 2 levels are identical to the country's territory (European Commission, 2021).
2. Selection of an ad-hoc set of indicators from the "Regional Innovation Scoreboard" 2021. We selected at least one indicator most closely related to young and student entrepreneurship from the four main types of activities - Framework conditions, Investments, Innovation activities, and Impacts. The six indicators selected from RIS 2021 are reported in Table 1. It is worth noting that proxies for EIS indicators are included for all enterprises (European Commission, 2021).
The indicators highlighted in the previous paragraph were related to administrative areas generally comparable to the NUTS 2 administrative level. However, they have been adjusted and selected to be homogeneous in terms of the indicators calculated. NUTS 2 level data were mapped particularly for mainland Europe and islands, not considering, therefore, overseas dominions and including non-EU member states such as the UK, Norway, and Switzerland. The IRE indicators, as highlighted in the following sections, as well as GDP and population data, among others, were retrieved and homogenized in time – referred to 2021 – and in currency – data were expressed in Euro.
From EU GISCO databases, geographical data were derived and acted as the basis for the computation of the different elaborations. The data consisted of a selection and combination of NUTS 1 and NUTS 2 levels – and similar - administrative units, where ad-hoc indicators were attributed, analyzed and mapped. A total of 245 units were used in the analysis. Data regarded EU countries together with the UK and Switzerland.
3.1. Creation of the composite index Innovation Readiness Environment (IRE)
We applied the Mazziotta-Pareto composite index to summarize the data of the mathematical combination of the selected indicators to create the composite index "Innovation Readiness Environment" (IRE) that represents, with a single score, the overall performance measured by the six indicators (Mazziotta & Pareto, 2013; Mazziotta & Pareto, 2018). We conceived it as a synthesis of all indicators to assess the overall performance (as opposed to a single-indicator performance) of the European Regions.
The MPI can be employed to compare inequality levels across different regions, periods, or social groups, and it can also be used to evaluate the impact of specific policy interventions on inequality patterns.
The MPI building proceeds in the following two stages:
1) Normalization
Let X = xij be the matrix with n rows (countries or geographical areas) and m columns (indicators), and let Mxj and Sxj denote the mean and the standard deviation of the j-th indicator:
The standardized matrix Z = zij is defined as follows:
where the sign ± depends on the relation of the j-th indicators with the phenomenon to be measured (+ if the individual indicator represents a dimension considered positive and - if it represents a dimension considered negative).
2) Aggregation
Let cvi be the coefficient of variation for the i-th units:
where
Then, the generalized form of MPI is given by:
where the sign of the penalty (the product Szi cvi) depends on the kind of phenomenon to be measured and then on the direction of the individual indicators (De Muro et al., 2008).
If the indicator is increasing or positive, i.e. increasing values of the indicator correspond to positive variations of the phenomenon, then MPI– is used. Vice versa, if the indicator is decreasing or negative, i.e. increasing values of the indicator correspond to negative variation of the phenomenon, then MPI+ is used.
Given our phenomenon, the MPI is calculated with the negative sign.
Thus, in our IRE index, created using the Mazziotta-Pareto formula shown above, the indicators previously reported in Table 1 co-influence, and with reference to Figure 7, the results show only 11 European countries ranking very high in innovation performance, followed by 44 high-performing countries, 60 (medium-high), 66 (medium-low) and, finally, 64 (low).
Figure 7. Innovation Readiness Environment.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Silvia Battino.
3.2 An analytical view of IRE as a predictor of innovation
To understand the indicator's potential, it was decided to perform an analysis of spatial clustering of the same indicator, as well as an exam of a possible relationship between IRE and GDP. This set of analyses involved relying on different tools for performing such tasks as the linear regression and LISA methods.
Linear regression. A linear regression was performed, considering a potential relationship between the IRE and other elements that can be considered relevant, such as the GDP. The analysis was performed using Apache OpenOffice 4.1.14 CALC suite. A regression analysis was performed on the study area, and the spatial units were considered. The linear regression attempts to model the relationship between the two variables by fitting a linear equation to observed data. The function calculates the statistics for a line using the least squares method to calculate a straight line that best fits data and then returns an array that describes the line. This method calculates the best-fitting line for the observed data, minimizing the sum of the squares of the vertical deviations from each data point to the line.
The general equation for the line is:
Y = aX + b
Where X is the explanatory variable, and Y is the dependent variable. The slope of the line is b, and a is the intercept (the value of y when x = 0).
Autocorrelation and LISA. Since the data is considered spatial in its extent, Exploratory Spatial Data Analysis (ESDA) appears paramount in observing the related phenomena in its territorial component. Spatial clustering methods are useful for making sense of complex geographic patterns (Anselin, 1995). Events in space, in fact, are rarely randomly distributed but present, instead, a certain degree of local similarity among them. Regions in space, particularly, tend to have features locally similar, fading as distance increases or, as Tobler (1970) “all things are related, but nearby things are more related than distant ones”. Data can, in fact, mutually influence geographical shape and, spatial proximity and values attributed to the same units. That means observing a selected variable's behaviour in relation to its position in space and proximity. Such a characteristic is known as spatial autocorrelation. The most interesting property of spatial autocorrelation is the capability to analyze at the same time locational and attribute information (Goodchild, 1986; Lee & Wong, 2001) defined spatial autocorrelation as follows:
Where:
1. n is the number of objects;
2. i and j are two objects;
3. xi is the value of object i attribute;
4. cij is a degree of similarity of attributes i and j;
5. wij is a degree of similarity of location i and j;
if cij = (xi – ) (xj – ) Moran Index I (Moran, 1948; O’Sullivan & Unwin, 2010) can be defined as follows:
Moran I and Geary’s G, not considered here, represent a global indicator of spatial autocorrelation, which is considered in the overall study region. Local variations are better observed by means of local indicators of spatial association; in particular, the local moran index makes it possible to evaluate the similarity of each observation with nearby geographical objects for each position. This can be seen as the sum of all local indices and is proportional to the value of the Moran one:
Where:
N is the number of geographical units;
Xi is the variable describing the phenomenon under investigation in region i;
represents the sample average and (Xi – X) it is the variable’s average deviation;
is the Standard deviation;
wij is the weight matrix.
The area units considered need to be weighted by means of a contiguity matrix, reproducing the spatial relationship among the regions considered, using a binary set of existing / non-existing contiguity among the different areas, following a ‘Queen rule’ of connection (Anselin, 1995).
From the application of the above-mentioned method, it is possible to obtain five combinations:
(1) high values of the phenomenon and high levels of similarity with the nearby areas, known hot spots (High-High), observable on the upper right quadrant of the global Moran’s I graph;
(2) low values of the phenomenon and low levels of similarity with the nearby areas, called cold spots (Low-Low); observable on the lower left quadrant of the global Moran’s I graph
(3) high values of the phenomenon and low levels of similarity with the nearby areas are detected, referred to as potential outliers (High-Low);
(4) low values of the phenomenon and high levels of similarity with nearby areas are highlighted, referred to as potential outliers (Low-High);
(5) no significant autocorrelation values are detected (Not Significant).
Spatial autocorrelation can also be considered bivariate, typically as the correlation between one variable and the spatial lag of another variable. In the case considered, the implemented indicator, IRE, was related to measures of local GDP to understand the possible relationship among the different variables and their spatial extent. Recently, LISA’s local Moran’s I was used to analyze, among others, socio-economic phenomena on area unit data on various applications, from migration movements (Borruso, 2009; Murgante & Borruso, 2012) to the efficacy of cohesion policies (Balletto et al., 2020). The analysis was performed using GeoDA as Free and Open Source Software developed by Anselin (1995).
4. Results
IRE applied to the 245 European NUTS 1 and NUTS 2 units was considered to be related to per capita GDP values in € in 2021. A set of analyses were performed, as anticipated in the above-mentioned methods. Namely, the regression analysis and the Local Moran over the IRE index. As an early visual observation of the IRE maps realized that they portrayed a certain level of similarity of most innovative regions with the highest GDP’s ones, as those of the European core or the heart of the so-called ‘Blue Banana’ (Brunet, 1989), a regression analysis was initially carried out to observe if a certain relationship among the variables could be considered. IRE and GDP values were plotted, and a linear regression was performed, providing interesting results given the high level of relation that can be observed among the different variables (R2 of 0.525; Figure 8 a) in the direct relationship, and better fitting values considering Ln GDP (R2 of 0.6256; Figure 8 b).
Figure 8.
Regression Analysis a) IRE and GDP; b) IRE – Ln GDP. Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Graph by Giuseppe Borruso.
Such initial results made us consider the spatial component and behaviour of both IRE and its relationship with European GDP. As in the previous paragraph, Global and Local Moran’s I were performed on IRE (Figure 9 and Figure 10), Ln GDP (Figure 11 and Figure 12) and as a cross, bivariate Moran’s I on IRE and Ln GDP (Figure 13 and Figure 14).
Figure 9. Local Moran’s I Cluster Map on IRE.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Giuseppe Borruso.
Figure 10.
a) Local Moran’s I Significance map on IRE and b) Moran’s I. Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map and graph by Giuseppe Borruso.
Figure 11. Local Moran’s I Cluster Map on Ln GDP.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Giuseppe Borruso.
Figure 12.
a) Local Moran’s I Significance map on Ln GDP and b) Moran’s I. Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map and graph by Giuseppe Borruso.
Figure 13. Bivariate Local Moran’s I Cluster Map on IRE & Ln GDP.
Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map by Giuseppe Borruso.
Figure 14.
a) Bivariate Local Moran’s I Significance map on IRE & Ln GDP and b) Moran’s I. Source: elaboration from GISCO – Eurostat data; Data as in Table 1. Map and graph by Giuseppe Borruso.
Local Moran’s application presents spatial clusters of high-high autocorrelation of the IRE in the NUTS areas in proximity of the European core, covering from Eastern France, Southern Germany, part of Austria and Belgium, and most of England units. Scandinavian peninsula’s units represent a relevant cluster as well. Negative autocorrelation can be observed on a wide cluster of units belonging to Eastern European countries and on one in Southern Spain (Figure 9).
Cluster map appears significant, with a relevant value of Moran’s I (0,708; Figure 10 a and b).
Univariate local Moran’s I was also performed on Ln GDP, reporting a – quite expectable – autocorrelation and spatial clustering of areas in the European economic core, rooted into Northern Italian Regions and extended northwards through Southern and Western Germany, Eastern France, BeNeLux and Ireland. Scandinavian regions represent another cluster. An East-West division appears with a wide cluster of negative autocorrelation regarding the regions belonging to Eastern European countries (Figure 11).
Also, the cluster map appears significant in this case, with a relevant value of Moran’s I (0,788; Figure 12 a and b).
A final analysis was performed employing a bivariate local Moran’s I, comparing IRE and Ln GDP (Figure 13). As bivariate local Moran's I describes the statistical relationship between the first variable at a given location and the spatially lagged second variable at neighbouring locations, the ‘high – high’ values (red) represent the spatial autocorrelation of the area units (also observable in the upper right quadrant of the Moran’s I graph, Figure 14 b), characterized by high values of IRE, spatial contiguity, and high values of Ln GDP. An interesting picture can be observed, with clusters of NUTS areas in the Central and Southern parts of the European core – Southwestern Germany, Western France, Southern Austria – Belgian NUTS and most of England Units. Scandinavian NUTS represent a relevant cluster as well. Negative spatial autocorrelation covers a relevant part of Eastern European Countries’ NUTS.
The results from the bivariate local Moran’s I computation can be considered significant (Figure 14 a and b) with a relevant value of Global Moran’s I (0.672).
5. Conclusions
In the evolving discourse on the spatial dynamics of innovation clusters within European regions, it is crucial to highlight the critical role of the newly developed Index of Regional Entrepreneurship (IRE) in outlining the intricate patterns and associations with per capita GDP distributions.
If it is true that a European core tend to be quite settled and consolidated as in the Brunet (1989) Blue Banana metaphor, innovation tend to follow partly similar patterns but also other, more specific ones. Focusing on the British Isles and the Scandinavian peninsula, the IRE index serves as an invaluable tool for nuanced exploration of different regional policy frameworks, revealing institutional and structural dimensions beyond GDP narratives.
Weaving this analysis into the current literature requires an alignment with the objectives of Entrepreneurship 2020. The strategy is based on three pillars, all of which reflect on the IRE index: a robust framework for entrepreneurial education and training, promoting an environment ripe for entrepreneurial growth, and cultivating role models to unleash the latent entrepreneurial capacity within different social groups.
Strengthening entrepreneurial education and training is central to this strategy, representing a key investment to enhance Europe's position in the global entrepreneurial landscape. In this context, the IRE Index emerges as a cornerstone that allows for a deeper understanding of the region-specific nuances that influence entrepreneurial developments, as outlined in the writings of Jenner (2012) and in the guidelines that are encapsulated within the DG EAC - ET2020.
However, Benneworth and Osborne (2015) argued that the untapped potential in entrepreneurship education is largely due to inadequate integration within university knowledge frameworks. Thus, the IRE Index can act as an analytical tool, guiding research efforts to synergize entrepreneurial initiatives with university knowledge, fostering the ecosystem where entrepreneurial attitudes can be developed. Simultaneously, the rapidly evolving global knowledge and technology landscape highlighted by Posselt et al. (2019) requires a proactive stance in advancing entrepreneurial universities capable of navigating this complexity. IRE Index, in this regard, is emerging as a tool to guide universities in fostering leadership and innovation, thus, realizing Entrepreneurship 2020 Action Plan's vision that supports sustainable economic growth and development in a coherent and inclusive way.
Synthesizing these findings within the broader framework of the 2030 Agenda for Sustainable Development (SDG), the IRE Index is seen as a central tool for improving societal wellbeing, particularly for disadvantaged groups. The relationships between entrepreneurship and the SDG, in particular Goals 4.4 and 8.3, comes alive through the lens of the IRE Index, providing critical evidence on the role of entrepreneurship in fostering skills development and innovation, an, as a result, inclusive growth. Besides, promoting entrepreneurship, prosperity and wellbeing across Europe requires taking into account regional differences and a deeper understanding of regional innovation indicators. Using tools such as the Smart Specialisation Strategy (S3) and the Smart Specialisation Strategies for Sustainable and Inclusive Growth (S4), the IRE Index provides an in-depth understanding of regional entrepreneurial trends to guide future research and policy directions. As we explore the complex spatial dynamics of innovation clusters within European regions, integrating sustainability into governance mechanisms, as detailed by Radosevic and Soete (2023) and creating value networks, Wostner (2017), offers a critical perspective that complements our analysis. The governance challenges identified by the authors, including the need for new forms of governance that emphasize reflexive and experimental approaches, align with the findings derived from the IRE index, highlighting how entrepreneurship and innovation are not just correlated with economic outputs like GDP but also with sustainable regional development practices and the creation of value networks, both vertical and international. The integration of sustainability dimensions in smart specialization strategies provides a foundational approach to understanding and enhancing the regional entrepreneurial ecosystems mapped by the IRE index. This integration is essential for achieving the 2030 Agenda for Sustainable Development and specifically supports Goals 4.4 and 8.3 by promoting skills development and sustainable innovation.
While IRE provides insights into the dynamics of innovation clusters in European regions, several limitations and avenues for future research emerge. First, although the IRE index is comprehensive, it focuses on certain indicators, such as R&D spending, patent rate, firm size, and educational levels. As a result, it may not capture all relevant factors influencing regional innovation and entrepreneurship. Future research could expand the index by including additional dimensions, such as Gross Expenditure on Research and Development (GERD), to provide a more holistic understanding of innovation ecosystems. In addition, technological progress and its impact on entrepreneurial activity have been overlooked. By integrating technological aspects, such as measures of technology adoption, digitalization levels, or investment in new technologies, future research can provide a broader understanding of factors driving innovation and entrepreneurship in European regions.
Furthermore, IRE's reliance on existing policy framing, such as the Entrepreneurial 2020 Action Plan and Smart Specialisation Strategies, may not fully capture regional innovation ecosystems' dynamic and complex nature. A comparative analysis between the IRE index and alternative frameworks or methodologies could be applied for future research to overcome this limitation. In addition, the inclusion of longitudinal data and qualitative research methods such as case studies and interviews with key stakeholders could be used to gain insights into the contextual factors shaping regional dynamics.
