Geoscience Reference
In-Depth Information
I also include all the available standard variables in the literature on preference
formation: age, gender, education, ideology (as captured by a question asking
respondents to place themselves in a left-right scale), and a variable asking
whether respondents reside in rural or urban areas. Unfortunately, the data
set does not include a direct measure of respondents' individual income. The
data set does include, however, a variable capturing respondents' occupations
which allows me to identify whether they are working (
α
in the model) or
receiving income support from the welfare state (1
in the model). Taking
advantage of this variable, I run the statistical analysis twice, once for the full
population and once isolating the preferences of those whose income depends
upon national welfare systems. A detailed presentation of the statistical results
appears in table format in Appendix B (Table B.1).
The analysis yields very interesting results.
15
In line with the argument of
this study, the models reveal a strong and statistically significant effect of
macrolevel income and labor-market conditions on individual preferences for
the centralization of welfare policy at the EU level. In other words, there is a
direct link between the Union's economic geography and citizens' preferences
for the centralization of interpersonal redistribution. The link operates through
three channels.
Consider first the two variables measuring differences in the structure of
labor markets among member states. The results reported in Table B.1 suggest
that as the incidence of unemployment increases so does the average support
for the harmonization of social welfare systems among the citizens of any given
country. By implication, as member states diverge in their labor market per-
formance, the disagreement about the desirability of centralizing interpersonal
redistribution increases. In addition, the empirical models also support the
claim that regional economic specialization limits the feasibility of a common
European social policy. Recall that, as a proxy for cross-regional differences
in economic specialization and the organization of labor markets, I included
a measure of national government's budgetary effort to support active labor
market policies. To be sure, the higher the effort in active labor market policies
in a particular country, the lower the support for the harmonization of welfare
systems in the EU. For instance, keeping all individual characteristics constant,
the average citizen of a country that is moderately rich and unequal and puts
very little effort in the activation of labor markets has a twenty percent chance
of showing “strong support” for such an institutional change. If his/her govern-
ment spent about 1% of GDP on active labor market policies, the probability
that same citizen offered the same answer would decline by ten percentage
points. These results are consistent with a recent literature that links varieties
of capitalism and welfare regimes with individual preferences for European
integration. While focused on a broader issue, the central intuition is very simi-
lar: individual attitudes toward integration are conditioned by the organization
of labor markets and welfare states across countries (Brinegar and Jolly
2005
;
−
α
15
The individual level control variables display the anticipated results.
