Understanding democratic transition using self-organizing maps : a special focus on Arab Spring Countries

Abstract EN

The aim of the paper is to identify, in the light of the Arab spring, the conditions favorable for democratic transition and to analyze the specific experience of Arab countries concerned by regime change.

We use a large set of indicators and the self-organizing map (SOM) approach as an alternative analytical tool.

Countries are then mapped considering data related to 33 indicators from 121 countries, four years (1984, 1991, 2002 and 2013), including economic, social, demographic, societal and political variables, as well as indicators related to institutional framework.

Mapping allows us to link countries configuration to their democratization levels.

Then, the SVM clustering methodology is used to cluster the prototypes produced by the SOM in order to discriminate the future successes and to determine the most influential variable on democratic success.

Countries neighborhoods are based on similarity in particular characteristics.

Our mapping highlights especially population age structure, globalization, health indicators, education and women participation in the society as the most important variables determining each country location which could be related to democracy level and its change over time.

In Addition, unemployment, corruption, democratic accountability, and law and order, are the most influential variables distinguishing failure from democratic transition success.

JEL Classification: P16, C45, C53