Abstract
Do institutions or culture determine levels of social trust in society? If quality of institutions determines levels of social trust, migrants from countries with lower-quality institutions should enhance their level of social trust in countries with higher-quality institutions. If, on the other hand, the migrants' level of social trust is determined by their culture, it should not be affected by a different institutional setting. Furthermore, culturally diverse immigrant groups should have different levels of social trust in the same host country. Analysing migration from several non-western countries to Denmark, this paper demonstrates that institutions rather than culture matter for social trust.
Keywords:
Acknowledgements
The authors acknowledge financial support for their Social Capital Project (SoCap) from the Danish Social Science Research Foundation. They are especially grateful to Guido Tabellini and the participants in the conference ‘Quality of Governance’, 17–19 November 2005 in Gothenburg, Sweden, for useful comments and suggestions on a first version of this manuscript. Any remaining errors of fact or judgement are the sole responsibility of the authors.
Notes
[1] By the term ‘social trust’ we denote trust in people in general, in contrast to trust in specified identifiable groups of people or specific individuals. This form of trust is also known as ‘generalised trust’.
[2] The WVS report levels of social and institutional trust separately for Serbia and Montenegro, while we are unable to distinguish between Serbs and individuals from Montenegro in the SoCap04-survey. We hence combine the two samples in the WVS when comparing immigrants from Serbia and Montenegro in Denmark to their countrymen.
[3] ‘The Danish legal system does not treat immigrants and Danes differently’ and ‘Danes and immigrants will always be treated in the same way by the administration’.
[4] ‘The police will always treat immigrants worse than Danes’ and ‘The government looks primarily after the interests of Danes’.
[5] The questions were asked on a five-point Likert-scale with the response alternatives ranging from ‘agree completely’ to ‘disagree completely’.
[6] We did consider using instrumental variable estimation of the relationship between institutional trust and social trust, but could find no suitable instruments.
[7] Another explanation for the moderate correlation is measurement error in institutional trust (the independent variable) stemming from the fact that this variable (as is social trust) is imperfectly measured by survey instruments. That said, by measuring institutional trust by means of a composite scale, measurement error in this variable is reduced.
[8] We use the respondents' own answer to the question ‘Are there many (more than half), quite some, a few or no people from your own ethnic group living in the neighbourhood where you reside?’
[9] Note, however, that Dinesen and Sønderskov (Citation2013) find a negative effect of neighbourhood heterogeneity in the general Danish population using a much more fine-grained measure of heterogeneity than available here. Their results suggest that this effect may also hold for immigrants.