Pedro Franco (p), Eduarda Marques da Costa

Andre Carrascal-Incera

Services of general interest (SGI) are a vector in battling regional disparities. Thus, these cannot be dissociated from the analysis of regional health inequalities, particularly social services of general interest (SSGI), especially in the context of an increasingly aging Europe. As such, in this study we analyze regional disparities in health in the light of economic, social, territorial, demographic and SSGI characteristics of the regions, trying to understand which factors have the most impact on these disparities. Thus, a Principal Components Analysis (PCA) was performed, followed by a Geographically Weighted Regression (GWR). In carrying out the PCA, socio-economic, territorial and SSGI indicators were used. Based on the extracted factors, a GWR was developed where the components played the role of the independent variable and health outcome indicators played the role of the dependent variable. The results show a good adaptation of the models to the observed reality. It should be noted that the factors involving the socioeconomic dimension and the provision of certain health services denote a higher regression coefficient. Concluding that SSGI are a factor to be taken into account in explaining regional inequalities in health, with particular incidence in some causes of mortality.

Miguel Á. Márquez (p), Marcelo Lufin Jesús Pérez

Pedro Franco

The fight against income inequality is on the agenda of many national governments. At the same time, political decentralization has been a significant tendency in the last decades. Paradoxically, to the best of our knowledge, no studies have examined which part of within-country inequalities can be related to every governmental level, making it difficult to assign responsibilities. In this paper, we try to fill this gap by connecting a vertical and horizontal (spatial vs a-spatial) decomposition of the overall regional inequality with the multilevel governance. The empirical analysis is focused on the standard three-level hierarchical structure (regions, provinces and municipalities) in Spain as the benchmark to decompose the overall Spanish municipal inequality. This decomposition provides information about the underlying relationship between the configuration of the Spanish subnational governments and the responsibilities to be assigned to each level of government in the fight against inequalities. An approximation to the assignment of responsibilities to the three levels of government would have important implications for both the diagnosis of the governmental performance and the design of effective policies aimed at fighting against inequalities among the municipalities in Spain. Our results suggest that policies to mitigate income inequalities among the Spanish municipalities should be implemented by contemplating multilevel governmental initiatives: policies coordinating neighbouring regions, initiatives linking to neighbouring municipalities and local policies.

Cristiana Fiorelli, Massimo Giannini (p), Barbara Martini

Miguel Á. Márquez

The concept of economic convergence refers to the process by which regions with similar economic characteristics and levels of development tend to converge over time. The European Union (EU) prioritizes economic convergence and social cohesion. Recent studies have shown that certain regions have surged ahead economically while others have been left behind, contributing to increasing regional divergence within countries. This paper aims to i) identify convergence clubs of NUTS 2 units in Europe by means of new machine learning and deep learning techniques; ii) study the dynamics of the clusters over time; iii) investigate the determinants of cluster dynamics and, therefore, explain what forces allow regions to remain in or leave the clubs. We use CatBoost, a new classification algorithm that allows for working with ordinal and categorical variables in a neural network. It is able to classify a large dataset and to perform distribution estimation of the data and is based on gradient boosted decision trees (GBDT). The algorithm is applied to identify how many "clubs" are in the European regions by considering economic, financial, social, political, and environmental variables, from the largest possible perspective. By analyzing the determinants of cluster dynamics, the study aims to explain what forces allow regions to remain in or leave the clubs.

Andre Carrascal-Incera (p), Geoffrey Hewings

Massimo Giannini

In recent years, there has been a growing awareness of the negative consequences of interregional inequalities, especially about left-behind places worldwide (Rodriguez-Pose, 2018). In particular, this affects some European countries such as the UK, France, Belgium and Denmark, which are more geographically unbalanced today than some decades ago, experiencing an accumulation of income dominated by metropolitan and central areas. These extreme interregional inequalities appear as a result of complex interrelationships between the effects of economic geography and modern production processes in a globalised context. In this paper, the EUREGIO database (Thissen et al., 2018) is used to compare the income distribution structure in 2000 and 2010. By means of an extended multiregional Input-Output model, I explore the structural roots of these inequalities in the European regions, offering results for interregional inequality within and between countries, and their impacts across the multiregional system. These models highlight the systemic contribution and structure of income interdependence, revealing important asymmetries that would remain hidden otherwise.