Terrorism abroad and migration policies at home

ABSTRACT

Do security concerns lead to more restrictive immigration policies? In this article, we contend that transnational influences can shape legislative output on immigration at home. Terrorist attacks in a neighboring country affect the salience of security concerns in the focal state, the policy solutions for addressing them, and the political will to implement these changes. In proximity of countries targeted by terrorism, politicians have specific incentives to manipulate immigration regulations following pressure from public opinion, for political opportunism or in anticipation of their neighbors’ policy choices. Using data on 33 OECD countries, we find that proximity to targeted countries leads to the implementation of a more restrictive migration policy regime. The public’s common perception of a linkage between migration and terrorism thus has important policy consequences.

KEYWORDS:

• Immigration
• national migration policies
• OECD
• terrorism

Acknowledgments

We thank the editors, Jeremy Richardson and Berthold Rittberger, the anonymous reviewers, and participants of presentations at the EPSA conference in Belfast and at University of Trento for their constructive feedback. This research was partly funded by the British Academy (SRG19\190780, in partnership with the UK Department for Business, Energy, and Industrial Strategy).

Disclosure statement

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

Notes on contributors

Vincenzo Bove is professor of political science at the University of Warwick, UK.

Tobias Böhmelt is professor of government at the University of Essex, UK.

Enzo Nussio is senior researcher at the Center for Security Studies, ETH Zurich, Switzerland.

Notes

1 Our arguments about the role of geographic contiguity hinge on some sense of cultural similarity across borders. In fact, cultural commonalities are likely to be stronger the lower the geographic distance between societies (Guiso, Sapienza, & Zingales, 2009). At the same time, cultural proximity likely shapes how media frames security threats and affects the extent to which attacks abroad are covered and how the images are conveyed (Kwon et al., 2017; Nossek & Berkowitz, 2006).

2 Evidence about objective linkages between migration and terrorism is scarce. Recent studies have found at best mixed support for migration inducing terrorism or that more restrictive migration policy lowers terrorism risk (Bandyopadhyay & Sandler, 2014; Bove & Böhmelt, 2016; Choi, 2018; Dreher et al., 2017). Restricting migration policy can even be counterproductive if it creates feelings of discrimination and signals foreign-born populations that they are less welcome (Gillum, 2018; Lyons-Padilla, Gelfand, Mirahmadi, Farooq, & van Egmond, 2015). Yet, the few notable cases of migrants staging terrorist attacks, like the Berlin Christmas Market attack in 2016 (Nussio, 2018), rise to public consciousness via news media.

3 While restricting migration can reduce a country’s overall economic growth (Bove & Elia, 2017), voters may not be affected directly. Furthermore, there are several important distributional effects and immigration can lower the wage in specific segments of the host country’s labour force (Borjas, 2014). As such, restrictions to immigration do not necessarily translate into costs in voter support.

4 In Switzerland, a country without major terrorist attacks so far, nearly half of the population saw Islam as a threat to security according to a 2017 survey (Farman & Nussio, 2018).

5 We expect incumbents to exploit anti-immigrant sentiments regardless of extreme-right competition, as restrictive immigration policies are not only a prerogative of the right-wing parties (Alonso & da Fonseca, 2012).

6 This motivation should be most common for right-leaning governments. However, as reported in the appendix, we do not find right-leaning governments to react in a more restrictive manner.

7 We select OECD countries with a view towards increasing homogeneity among cases. Although perhaps at the expense of being able to generalize our findings to apply to all countries across the globe, our sample comprises states that are rather similar in several aspects at the macro level, e.g., economic development or regime type. Although there are exceptions over time and, e.g., Turkey performed rather poorly in terms of democratic accountability in recent years, only focusing on the states outlined in Figure A.1 ensures that our sample represents the population we wish to describe. Note that terrorism data are available from 1970, yet some of our control variables are unavailable for a number of countries before 1980. Results are qualitatively the same when we include data for 1970–1980.

8 The codebook is available at: https://www.imi-n.org/files/data/demig-policy-codebook.pdf.

9 Our arguments apply to general migration policies and we do not distinguish between types of policies, e.g., regulations and control mechanisms. The appendix presents a robustness check on the share of restrictive policies of all migration policies.

10 The policy-tool variable captures the instrument used to implement the policy measure above and consists of 28 codes, ranging from surveillance technology to work permits. The migrant-category variable identifies the migrant group targeted (e.g., low-skilled workers) whereas the geographical-origin variable includes the origin of the targeted migrant category (e.g., EU citizens.).

11 We do not distinguish between national and transnational attacks, but we disaggregate this variable in the appendix.

12 Stricter immigration policies may lead to more terrorist attacks abroad. If such reverse causality exists, the effect of terrorism abroad on migration policies at home will suffer from downward bias (Epifanio & Plümper, 2018). That said, and although we control for a large number of potential confounding factors, our estimates do not necessarily demonstrate a causal mechanism.

Additional information

Funding

We thank the editors, Jeremy Richardson and Berthold Rittberger, the anonymous reviewers, and participants of presentations at the EPSA conference in Belfast and at University of Trento for their constructive feedback. This research was partly funded by the British Academy (SRG19\190780, in partnership with the UK Department for Business, Energy, and Industrial Strategy).