HOW DOES THE WELFARE STATE INFLUENCE INDIVIDUALS' SOCIAL CAPITAL?

ABSTRACT

In the current literature there is emerging consensus about a positive correlation between countries' welfare efforts and the social capital of their inhabitants: the larger the welfare state, the more social capital its inhabitants have. This paper pulls the discussion an important step further by asking what mechanisms can be responsible for this correlation. Based on theory and previous findings, we formulate a conceptual model of how the macro–micro relationship could come about. The hypotheses are tested using Eurobarometer data. While most of the mechanisms hypothesised are also found empirically, a substantial part of the direct macro-to-micro effect remains after controlling for the presumed most important intervening factors. Our conclusion is that the relationship is still not well understood and needs more academic attention in future research.

Keywords:

• crowding out hypothesis
• social capital
• access to informal help
• welfare state

Notes

¹Methodological individualism is an approach in social science that argues that a society-level phenomenon should be understood as the aggregation of decisions and behaviour by individuals.

²Conceived of in this way, social capital shares important characteristics with the concept of social support (see e.g., Vaux 1992; Veiel and Baumann 1992).

³See Table 1 for a list of included countries.

⁴Cronbach's alpha for the scale is 0.762 (country results vary from 0.537 for Bulgaria to 0.853 for Spain). We note that the dependent variable includes actually two dimensions: access to informal help in terms of time, and in terms of money. We investigated whether findings differed between analyses using a single variable which combines the two dimensions, and analyses with the dimensions as two separate dependent variables (not reported here). Overall, the results indicated that effects differed only minimally between the two approaches. Therefore we report the findings of the analysis using the combined variable.

⁵Pooled descriptive statistics of the exogenous control variables are reported below:

Table

⁶As none of the macro level control variables had a significant effect, they were dropped from the analysis.

⁷Due to space limitations we do not discuss the results of the socio-demographic control variables.