Abstract (eng)
Based on Robert Castel's work, the leading hypothesis of this master's thesis is: People, who feel insecure, either due to pecarious employment or private isolation, demand better social security systems than those, who do not feel insecure. Do people who feel insecure have a higher positive acceptance for welfare state measures? Using data of the European Social Survey 2016 (special module on welfare) and mixed-effect models, the impact of contextual and individual variables on the acceptance in 21 European countries is tested. Tests include questions about: Which impact does one's political self-evaluation on a left-right-placement-scale have on the assessments of welfare benefits? And to what extend can country-specific differences be explained e.g. by the economic performance? Do people living in countries with higher social inequality (as indicated by the GINI coefficient) tend to favour redistribution?
The master's thesis follows concepts developed by Robert Castel and Gøsta Esping-Andersen. Both describe the integration in employment structures and a good family support as socially important. Therefore, the political and social approach in each country might influence how the attitudes of citizens towards the welfare state are shaped in different European countries. Following Robert Castel's ideas on uncertainty and disaffiliation, the relationships between one's own perceived insecurity and one's attitudes toward social welfare benefits are estimated. Furthermore, the influence of the general socio-economical condition of countries (as indicated by GDP/capita, the GINI coefficient, amongst others) on the perception of welfare benefits is taken into account as well.
A multidimensionality can be seen in the results. At the individual level, personal job insecurity and one's political self-evaluation show significant effects. The distinct differences between countries can neither be explained with Esping-Andersen's typology nor with the contextual variables. The results rather indicate that contextual effects are best examined on smaller regional units (canton, region, etc.), because the Pearson correlation coefficient is highest in the mixed-effects-model with NUTS-regions.