Ethnic Diversity and Its Effects on Social Capital in Germany

Sarah Ludwig-Dehm, The Pennsylvania State University

Many countries in Europe have seen considerable increases in immigration and ethnic diversity during the last few decades. Europe has seen increased immigration during the years 2015 and 2016. Germany was one of the countries that led in many more Syrian refugees than other European countries. This let to sometimes strong resistance from the political opposition and parts of the population. Germans are concerned about the integration of immigrants into their society, and this has spurred new interest in the consequences of ethnic diversity.

Theories predict very different outcomes from ethnic diversity. Contact, conflict, and constrict theory are most commonly used and their predictions on the effect of diversity on social capital differ a lot. Numerous studies have been examining the association between ethnic diversity and social capital. Findings from these studies are as diverse at the topic itself, which leaves researchers with many unanswered questions about the association between the two concepts and possible mechanisms.

Therefore, in this study I will look at the association of ethnic diversity and social capital in Germany, with the goal to disentangle the relationship between the two concepts. The research will be guided by the following two questions: first, which dimensions of social capital are affected by ethnic diversity in Germany? And second, can these relationships be explained by individual or contextual characteristics?

To address my research questions, I use restricted data from the Allbus study of 2008 in Germany, which allows me to join the individual level data with contextual data from the German federal bureau of statistics to include ethnic diversity measures on the municipality level for each respondent.

Multi-level models are used for each one of the social capital dimensions. Preliminary results show no significant relationship between diversity and social capital when individual and contextual controls are included.


Many countries in Europe have seen considerable increases in immigration and ethnic diversity during the last few decades. Europe has seen increased immigration during the years 2015 and 2016. Germany was one of the countries that led in many more Syrian refugees than other European countries. This let to sometimes strong resistance from the political opposition and parts of the population. Germans are concerned about the integration of immigrants into their society, and this has spurred new interest in the consequences of ethnic diversity.

Putnam''s (2007) widely cited article fostered concerns about multiethnic societies, not only in Europe but also in the U.S. by suggesting that greater ethnic and racial diversity has detrimental effects on the social capital of societies. Social capital can be understood as a measure of social cohesion capturing civic participation and social networks. Social capital is associated with several positive outcomes, like mutual support and cooperation between individuals, better child development, but also higher effectiveness of institutions and governments (Putnam, 2000).

Theories predict very different outcomes from ethnic diversity. Contact, conflict, and constrict theory are most commonly used. Numerous studies have been examining the association between ethnic diversity and social capital to test the propositions of these three theories. Findings from these studies are as diverse at the topic itself, which leaves researchers with many unanswered questions about the association between the two concepts and possible mechanisms.

Therefore, in this study I look at the association of ethnic diversity and social capital in Germany, with the goal to disentangle the relationship between the two concepts. The research will be guided by the following two questions: first, which dimensions of social capital are affected by ethnic diversity in Germany? And second, can these relationships be explained by individual or contextual characteristics?

To address my research questions, I use restricted data from the Allbus 2008 study in Germany which allows me to join these with contextual data from the German federal bureau of statistics. The contextual data are on the municipality (Kreis) level, the smallest administrative subdivision in Germany. My independent variable is ethnic diversity measured with two different approaches: a percentage measure and the fragmentation index. The independent variables cover a wide range of social capital dimensions. They include generalized trust as well as trust in friends, relatives, acquaintances, and strangers, contact with friends, and membership in political, community, and leisure organizations.

Additional variables are included in the models to control for possible confounding effects and to test certain mechanisms that might be important in explaining the association between ethnic diversity and social capital. On the individual level, I include the respondents’ age, gender, family status, children, education, monthly household income, and labor force status. On the municipality level, I include the percentage of inhabitants who are unemployed and the number of in- and out-moves to measure disadvantage and stability of the area.

To answer my first research question, regression models are computed in which only the independent and the dependent variable are included. I calculate two models for each dependent variable to look at the different influences of measuring ethnic diversity by percentage and fragmentation. I use linear, logistic, and ordered logit models depending on the characteristics of the dependent variable. All of these models are multilevel models with two or three specified levels. In a second step, I include control variables on the individual and then also on the municipality level to assess my second research question. My preliminary results only include Germans and no non-Germans.

My preliminary results show the following: First, there is no difference between using a percentage or fragmentation measure of ethnic diversity in Germany. All models show very similar results for the two variables. Second, using a squared diversity term in addition to the linear term does not yield any significant results. However, interestingly, using a categorical diversity term showed sometimes surprising significant results. This could mean that diversity is not necessarily linearly related to social capital, but rather there might be a threshold effect. This will be examined in future analyses. Third, the results show that different dimensions of social capital are differently related to diversity. While most of the dimensions do not show any relationship with diversity, some are positively related (generalized trust, trust in strangers) and some negatively (trust in acquaintances, membership in leisure organizations). And fourth, the preliminary results show that when controlling for individual and contextual control variables, diversity has no significant influence on any of the dimensions of social capital.

Presented in Session 1235: Posters