Gender equality is a topic of increasing relevance worldwide, but its measurement is mainly limited to national levels, even in the UE. During the last 20 years, various experiences tried to exploit gender equality at a subnational (regional) level, but they were country-specific studies focused on single or a few domains (e.g.: education, health, labour market) not providing a complete overview of the territorial disparities on gender equality.
In this project, we propose a regionalization of the most complete gender equality indicator, the Gender Equality Index (GEI) of the European Institute on Gender Equality (EIGE). EIGE’s GEI measures gender gaps and different levels of achievement of the 27 Member States of EU + UK over time and across a range of relevant 6 core domains: Power, Knowledge, Work, Money, Time, Health for a total of 14 sub-domains and 31 different variables (called indicators). To avoid proposing another brand-new indicator, in this project we will develop a regional version of the index following, as close as possible, the EIGE’s GEI methodology building what we call Regional-GEI (in brief, R-GEI).
However, the regionalization of EIGE’s GEI is not a straightforward process and clashes with two serious issues. First, albeit most of the variables used in the GEI are based on European surveys which are representative also at the regional (NUTS 2) level, a large part is not. Secondly, some variables are meaningless if measured at the regional level (think, for example, to the number of women seating in the national Parliament). Consequently, for either of the two reasons, some variables (approx. 40%) must be replaced by alternatives chosen to be as close as possible to the original ones. That is a very complex process that requires a careful analysis of the data available at the regional level for any single country provided by the single national statistical institutes.
Therefore, the first main activity of this project will be the assessment of the actual availability of data at NUTS2 (regional) level in four major EU countries (Italy, France, Spain and Germany), covering a total of 98 regions for the construction of the R-GEI index, identifying a consistent system of indicators which can be used as a common basis for gender equality measurements at regional level in Europe and identifying adequate substitutions for those that are not statistically representative at regional level in the EU-27 wide surveys.
EIGE’s GEI is the result of the hierarchical composition of the 31 indicators in the 14 sub-domain indicators, in the 6 domain indicators and in the single indicator. Each step is the result of choices (e.g.: weighting schemas) that, if changed, may provide different scores resulting in different rankings of the European Regions. The second activity of this project will deal with the reliability and stability of the Regional ranking according to different composite indicators approaches. The methods that will be applied to compare Regional rankings will be: EIGE’s GEI procedure as a benchmark, Principal Components Analysis, Benefit-of-Doubt and methods based on partially of full non-compensative techniques (e.g.: Mazziotta-Pareto Index, Partially Ordered Sets theory).
In addition, it is often difficult for local authorities to identify benchmarks of reference, to compare their Region to the national and, even more hardly, European context. The third activity of this project will be the use of statistical machine learning methods to construct a socio-economic proximity matrix of the regions of these four countries, to define clusters of homogeneous regions and, for each region, the regions closest to it (nearest neighbours). The measurable output of the project will be:
- the building of a dataset of the indicators at regional level for the countries considered for at least three time points in the 2005-2018 time period;
- the definition of sub-domain, domain and global equality indicators scores for each region and their alternative rankings;
- the identification of the five closest regions for each of the 98 regions in the study countries.