Does Poverty Increase the Risk of Social Isolation? Insights Based on Panel Data from Germany

ABSTRACT

Sociological literature suggests a causal effect of poverty on the risk of social isolation. While previous research on this link is only based on cross-sectional data, this article presents more valid evidence for this assumption by applying fixed-effects regression methods and panel data (German Socio-economic Panel). First, the study shows that the poverty effect on isolation occurs regardless of time-constant unobserved heterogeneity. Second, it examines whether poverty and isolation are correlated because of a mutual association with critical life events (health deterioration, dissolutions of couple relationships, and job loss) that may increase both the risk for poverty and the risk for isolation. Regarding men, the effect of isolation on poverty is not attributable to the cited biographical transitions. Regarding women, it is only in part—but not in whole—attributable to dissolutions of couple relationships.

KEYWORDS:

• Social exclusion
• loneliness
• social capital
• social relationships

Notes

1. Earlier waves disregarded couple relationships involving separate households.

2. The corresponding survey questions are: “Which of the following activities do you do in your free time? Please enter how often you practice each activity … get together with friends, relatives, or neighbors” and “… help out friends, relatives, or neighbors, when needed.” Respondents were asked to answer “each week,” “each month,” “less often,” or “never.”

3. International comparisons find that social isolation in Germany falls within the middle range of other European countries (Lelkes 2010, 2011).

4. Although it is unusual to fit a binary dependent variable into a linear regression model, the linearity assumption seems less problematic in this case because the FE regression model already takes into account that there is a subject specific overall probability of social isolation that differs largely between people. Also, the problems of heteroskedasticity and serial correlation can be dealt with by using robust standard errors.

5. The additional analyses also show that the effect of poverty on isolation is strongest when a poverty threshold of 40 percent is used. A higher poverty threshold produces comparatively weaker effects of poverty on isolation. Nevertheless, this study uses the 50 percent poverty threshold because of the small number of cases of income below the 40 percent poverty line.

6. Because the network discussion instrument is implemented only in two GSOEP waves, it is not possible to apply fixed-effects regression for this indicator.

7. Additional analyses show that both health and couple relationships are relevant factors in explaining the effect of part-time work and nonemployment on women’s risk of isolation. Here, the contribution of couple relationships turns out to be more important than that of health.

Additional information

Funding

This work was supported by the Deutsche Forschungsgemeinschaft [EC 433/3-1].

Notes on contributors

Jan Eckhard

Jan Eckhard is research fellow and lecturer at the Max-Weber-Institute at Heidelberg University. Latest publications: Migration and the Partner Market. European Journal of Population, Jg. 38: 58-86 (2017, with J. Stauder); Indicators of Social Isolation. Social Indicators Research (2017, online first).