Local Immigration and Support for Anti‐Immigration Parties: A Meta‐Analysis
No TL;DR found
Abstract
Does the share of immigrants in a community influence whether people vote for anti-immigration parties? We conduct a systematic review of the causal inference literature studying this question. We collect estimates from 20 studies and develop a new Bayesian meta-analysis framework to account for both between-study heterogeneity in effect sizes and the possibility of reporting bias. Although meta-analysis methods that do not adjust for reporting bias suggest a moderate effect of local immigration, our Bayesian model finds that the effect of local immigration on far-right voting is on average negligible once we account for reporting bias. However, the analysis also reveals a large heterogeneity in effects across contexts, suggesting that local immigration may be important for anti-immigration vote shares in certain settings. Verification Materials: The data and materials required to verify the computational reproducibility of the results, procedures and analyses in this article are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/TEPAK4. Immigration has become a prominent political issue in recent decades (Grande, Schwarzbözl, and Fatke 2019), with increasing public support for anti-immigration parties in many European countries (Arzheimer 2018). A rapidly growing research literature has addressed this political shift to the extent that even some researchers in the field view the attention as disproportionate (Arzheimer 2018; Mudde 2013). Despite this attention, there is a lack of consensus on whether immigration causally triggers anti-immigrant voting. Although Arzheimer’s literature review (2018, p. 160) argues that immigration tends to increase the vote share of the radical right, Golder (2016 p. 485) notes that only a minority of studies have employed research designs that allow for causal inference.1 In this article, we conduct a meta-analysis of the causal inference literature. In a systematic search of the literature, we screen papers against a set of criteria, the first of which requires that a study employs a research design developed explicitly to address bias due to selection and reverse causality. We identify 20 studies with a total of 147 qualified estimates. From the 147 estimates, we choose one main estimate per estimation technique per study, resulting in a sample of 31 estimates that we use in the main analysis. A simple unweighted average across the 31 estimates indicates that a 1 percentage point increase in immigrant share is associated with a 0.57 percentage point increase in the vote share of anti-immigration parties. This average masks a substantial heterogeneity, as individual Sara Cools is Research Professor, Institute for Social Research, P. O. Box 3233, Elisenberg, 0208 Oslo, Norway (sara.cools@ samfunnsforskning.no). Henning Finseraas is Associate Professor, Norwegian University of Science and Technology P. O. Box 8900, Torgarden, 0208 7491 Trondheim, Norway (henning.finseraas@ntnu.no). Ole Rogeberg is Research Professor, Frisch Centre for Economic Research Gaustadalleen 21, 0349 Oslo, Norway (ole.rogeberg@frisch.uio.no). We would like to thank Don Green, Eirik Strømland, and seminar participants at the Frisch Centre, Hertie School of Governance, NTNU, University of Oslo, OsloMet, the 2019 EPSA conference, and the 2019 Meeting of Norwegian Political Scientists for useful comments. Grant number 270687 (Norwegian Research Council) is acknowledged. See also Amengay and Stockemer (2018), who count the share (38%) of estimations that yield a positive and statistically significant coefficient for immigration on radical right voting. Kaufmann and Goodwin’s (2018) meta-analysis finds that increases in ethnic diversity tend to be associated with more negative views on immigration. American Journal of Political Science, Vol. 00, No. 0, XXXX 2021, Pp. 1–19 © 2021 The Authors. American Journal of Political Science published by Wiley Periodicals LLC on behalf of Midwest Political Science Association DOI: 10.1111/ajps.12613 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.