Research Design and Methods
Based on the deliberations above, the overarching research question for our quantitative analysis reads: Does political prioritization of ICT by EU regions have an impact on their ICT performance?
Considering data availability, we break down and operationalize political prioritization as 1) the existence of a dedicated regional ICT strategy and 2) the investment of ERDF funds in ICT. Thus, we arrive at two main concrete sub-questions:
Each research question is tested separately with comparisons of means for the binary independent variables (existence of an ICT strategy and the dedication of ERDF funding to ICT) in order to see whether our working hypothesis of a positive correlation between political prioritization and ICT performance holds in the different sub-dimensions of prioritization.
Data Collection, Variables and Descriptive Analysis
Case Selection
Starting out with the universe of all EU regions, we initially restricted our selection to EU-15 countries to establish basic comparability on the grounds of longer common EU history and roughly similar economic development in comparison to the newer EU member states. Further, we only selected countries with more than one NUTS2Footnote 2 regions to observe variance within a given country. As a last selection criterion, we used the languages read by our team of authors as to be able to interpret regional Operational Programmes (legally binding investment plans for ERDF and other EU funding programmes) and search for regional ICT strategies in the data collection phase.
Thus, we use data from the regions of the following European countries: Italy, Spain, UK, Ireland, Belgium, Portugal, Austria, Sweden and Germany. The dataset contains information about 97 European regions, which represent 35.9 % of the total NUTS2 regions of the European Union and 9 member states (or 60 %) of the EU-15 (Table 1).
Table 1 Number of regions included in the analysis per country
Independent Variables: Political Prioritization—ICT Strategies and ERDF Expenditure
As pointed out in the research design, we use two independent variables based on ICT strategies and ERDF expenditure (Fig. 1). The binary variable ‘Dedicated ICT Strategy’ indicates whether a region has developed a dedicated ICT strategy during the period 2007 to 2013. The values for this variable are determined by drawing on two different sources: the Regional Innovation Monitor database of the European Commission and a systematic Internet-based search of the top 20 results of a Google search for ‘ICT policy OR Information Communication Technology OR Innovation Policy AND Name of the Region’ in the respective regional/national language.Footnote 3 Consequently, we establish that about one third of the regions in our sample possess a dedicated regional ICT strategy (see Table 2).
Table 2 Variables included in the analysis
Information about investments of ERDF in particular policy fields is drawn from the Operational Programmes (OPs) presented by the regions and approved by the European Commission for the ERDF funding period 2007–2013.Footnote 4 In their OPs, the regions classify the expenditure of their ERDF funds according to 74 standardized categories, out of which six (categories 10 to 15) are related to ICT. Since most of these six ICT-related categories are quite ambiguous and therefore hard to link to concrete policy performances, we focus on the clear-cut Category 10 ‘Telephone infrastructure (including broadband networks)’.
Accordingly, the binary variable ERDF Investment indicates whether a region has allocated ERDF funds to category 10 or not. In our sample, roughly 40 % of the regions have invested ERDF funds for telephone and broadband infrastructure (see Table 2).
Dependent Variable: ICT Performance—Household Internet Access 2008/2012
As we narrowed down ICT-related ERDF expenditure on investments in telephone and broadband infrastructure, we operationalize ICT performance as the composite variable ‘Household Internet Access 2008/2012’ which draws on the two Eurostat indicators that can be directly linked to these investments: One indicator that measures the percentage of households where any member of the household has the possibility to access the Internet at home at the regional level and a second indicator for the percentage of households that are connectable to broadband Internet (based on Eurostat data for 2013).
To measure relative performance that can be compared across regions, we calculate the improvement of Household Internet and Broadband Access from 2008 to 2012, which lies within the time frame of the 2007–2013 cohesion policy framework:
$$ \mathrm{Household}\ \mathrm{Internet}\ \mathrm{Access}\kern0.35em 2008/2012=\frac{\left(\mathrm{Internet}\kern0.5em \mathrm{Access}\kern0.5em 2012+\mathrm{Broadband}\kern0.5em \mathrm{Access}\kern0.5em 2012\right)}{\left(\mathrm{Internet}\kern0.5em \mathrm{Access}\kern0.5em 2008+\mathrm{Broadband}\kern0.5em \mathrm{Access}\kern0.5em 2008\right)} $$
The resulting indicator Household Internet Access 2008/2012 has a standard deviation of 0.21. On average, Household Internet Access has increased by 40 % from 2008 to 2012 in our sample regions. The least performing region achieved an increase of 4 % (Norra Mellansverige, SE), while the maximum improvement in the sample is 127 % (Emilia-Romagna, IT). For further analyses, the regions have been grouped into three performance classes: ‘low performance’ refers to regions with less than 30 % improvement in Household Internet Access, ‘moderate performance’ to regions with 30 to 45 % improvement and ‘high performance’ to regions with more than 45 % improvement (see Table 2).
Main Hypothesis
The main assumption in this study is that policy prioritization matters for policy performance. If this holds true, we expect to find that regions which developed a dedicated regional ICT policy and allocated ERDF budget to broadband infrastructure display high performance with regard to the improvement of Household Internet Access (H1). On the other hand, we expect moderate performance from those regions where only a dedicated ICT strategy or relevant ERDF investment can be found (H2). Finally, we expect regions that neither developed a dedicated ICT policy nor invested ERDF to telephone and broadband infrastructure to show low performance (H3); see Table 3 for a matrix of our expected results.
Table 3 Expected distribution