Loved and feared: citizens’ ambivalence towards free movement in the European Union

ABSTRACT

The right of citizens to live and work in any member state is a foundational pillar of the European Union. The views of EU citizens on free movement are characterized by a puzzle: the border-free Europe is seen as the most important achievement of European integration but also appears as a main driver of Euroscepticism. In this article, I argue that this is because of a tension between citizens’ own mobility rights and the mobility rights of citizens from other EU countries. This idea of ambivalence towards free movement is tested with observational data and a survey experiment across 28 EU countries. The results suggest that many citizens hold ambivalent views towards free movement due to a tension between the value of one’s own mobility and the fear of immigration. Their effective support depends on the relative salience of inward and outward mobility. This finding has important implications for the public support of international integration more generally.

KEYWORDS:

• European Union
• free movement
• immigration
• ambivalence
• reciprocity

Acknowledgements

The article has benefited substantially from generous comments by Dominik Schraff, Marc Helbling, Pascal Sciarini, Jeremias Stadlmair, and Jack Blumenau as well as from the seminar participants at the European University Institute and the University of Geneva. Special thanks also go to the think-tank foraus under its umbrella I designed and commissioned the survey experiment. The three anonymous reviewers helped to further improve the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes on contributor

Philipp Lutz is a post-doctoral researcher at the Department of Political Science and International Relations at the University of Geneva.

Notes

1 Ambivalence does not imply contradiction or irrationality since different considerations that result in opposing evaluations can be fully rational views independent from each other. It only implies that citizens hold opposing evaluations towards the same object.

2 For each country approx. N = 1000 respondents were randomly selected, with the exception of Cyprus, Luxembourg and Malta where the sample size is about N = 500. The Eurobarometer separates Northern Ireland and East Germany. I combine this data with Great Britain and West Germany to obtain only one sample for the United Kingdom and Germany. Non-citizens are excluded reducing the sample by 770 observations.

3 The two survey items are closely interrelated, thus justifying the index-building not only theoretically but also empirically. The correlation is 0.71 for free movement in general and 0.8 for free movement in one’s own country. The aggregate support levels are almost identical between the two survey items.

4 Ambivalence is measured based on the variables ‘general support’ and ‘support own country’ where values below 0.5 are classified as opposition and values above 0.5 are classified as support for free movement. Thereby, I measure ambivalence as a difference in nature and not a difference in degree.

5 The analysis relies on immigration attitudes and their relative salience because it is where citizens show relevant variation, whereas regarding utilitarian benefits almost all citizens favour more rights over fewer rights.

6 I use attitudes towards non-EU immigrants as a robustness check.

7 Due to the skewed distribution and to facilitate interpretation, I log-transform and normalise the variable to a range from zero to one.

8 The salience for the country represents citizens’ socio-tropic evaluations that are considered to be more important for individuals’ views on immigration than ego-tropic evaluations (Sides and Citrin 2007). Replacing it with personal salience yields a larger interaction term in the model without changing the substantive findings.

9 Dropping the left-right orientation for potential endogeneity concerns, does not alter the results in a meaningful way.

10 Since the Eurobarometer was fielded in early 2016, I use values from the year 2015.

11 All models using Eurobarometer data are based on population-weighted data using the post-stratification weights (w23) provided for all 28 EU countries.

12 The intra-class correlation (ICC) is 0.057 for the general support, 0.055 for the support in the own country, 0.056 for the support-dummy and 0.017 for ambivalence. The contextual variation is rather low and in particular, ambivalence does not vary substantially across countries.

13 They survey was part of the EuroPulse omnibus survey https://daliaresearch.com/europulse/.

14 Non-citizens are excluded and therefore the sample includes only citizens.

15 The answers 'don't know' are coded as missing (N=611).

16 The two items are highly correlated.

17 See Appendix for the detailed operationalization of these variables.

18 Correlation between immigration attitudes and individual immigration salience is r = 0.08.

Additional information

Funding

This research was supported by the National Center of Competence in Research nccr – on the move funded by the Swiss National Science Foundation.