Measuring gender equality: a comparison of different techniques to build synthetic indicators

Gender equality represents a central issue in the socio-economic background of our society and, consequently, its study is gaining an increasing attention in the international debate (consider, for instance, the recent constitution of the W20 group). During the last 20 years, the international literature proposed a number of indicators that aim at measuring gender equality (or gender inequality). Among the others, it is worth recalling the Gender Gap Index by World Economic Forum, the Gender Development Index by United Nations and the Gender Equality Index (GEI) proposed by the European Institute for Gender Equality (EIGE). Although sharing some common characteristics, these indicators differ from one another on crucial points, above all in the set of variables used to define the domains of gender equality, selection that is naturally driven by the data availability. In Italy various experiences on measuring gender equality at sub-national level have been proposed but the resulting indicators are generally focused on specific domains and there are no studies that analyze this phenomenon as a whole. The aim of this work is twofold: on one hand we propose a regional decomposition of the EIGE gender equality index (R-GEI); on the other we compare the synthetic indicator obtained following the EIGE methodology with a poset based synthetic indicator (POR-GEI). Albeit many variables used by EIGE can be derived at NUTS-2 (regional) level using Eurostat microdata (EU-SILC; EU Labour Force Survey), sample size is not always adequate to guarantee acceptable standard errors for regional estimates; moreover, some variables are not available (or do not make sense) at subnational level. Therefore, we decided to replace some of the original GEI variables with equivalent variables available or reasonable at regional level. The new R-GEI is obtained reproducing the EIGE methodology and it is compared to the POR-GEI index that exploits poset theory for aggregating indicators. Our findings show that R-GEI and POR-GEI produce quite similar results but the use of POR-GEI brings to some interesting advantages in the interpretation of the territorial differences in gender equality in Italy