This paper explores the connection between individuals’ barrier preference toward immigration and refugee policy and how they are linked to voting behavior. We use three different datasets to tackle these two questions. We utilize multi-country survey data with abundant information regarding voters’ subjective attitudes toward immigrants to study the determinants of far-right voting in Europe. Notably, even after controlling for education, income level, or current employment status, which represents the human capital endow-ment of natives, industry-level comparative advantage and disadvantage, and political scale from left to right, we still find that a preference for immigration barriers has strong explanatory power for far-right voting behavior. We find that educational attainment and income level are important in explaining vot-ers’ anti-immigration attitudes, as the Heckscher-Ohlin model predicts. On average, voters with higher education and higher income are relatively more open to immigration and aware of their role in the economy.

In addition to the cross-country election studies that allow us to study the anti-immigration attitude, we also use election surveys conducted in the Neth-erlands and France to add more details on the role of individual immigration and refugee policy preference in the voting behavior. The election survey on the parliamentary election of the Netherlands provides a special identification approach that is not available in cross-country election surveys. The panel structure of the Dutch survey allows us to observe the voting history of re-spondents in 2012, 2017, 2019, and 2021 parliamentary elections. Using this information, this paper clarifies the characteristics of voters who newly join supporting right-wing populism. In other words, we study determinants of ‘switching’ voting behavior from non-populist parties to right-wing populist parties focusing on the refugee barrier preference. We explore whether atti-tudes towards refugees affect this switching behavior in the politics of the Netherlands. The election survey on French voters covers the most recent presidential election in France. We elaborate on how immigration policy pref-erences of French voters are related with the popularity of right-wing popu-list presidential candidates in France.

Considering that non-economic factors contribute to explaining the far-right voting behavior, utilizing other non-economic variables such as individ-uals’ sentiments towards various entities (including the government, legal sys-tem, and the European Parliament), and subjective evaluation of their own health and happiness can also add explanatory power, as numerous previous findings imply. For systematic consideration of these abundant variables in the ESS data, we use LASSO for variable selection. For this exercise, we con-sider two approaches in selecting the optimal penalty term, namely cross-validation (CV) and Bayesian information criterion (BIC). It is worth high-lighting that almost all of the immigration barrier preference variables are se-lected when using both CV and BIC for variable selection criteria.

Executive Summary

1. Introduction

2. Background

3. Data
3-1. European Social Survey
3-2. Dutch Parliamentary Election Panel Study
3-3. French Electoral Study

4. Econometric Specifications
4-1. European Social Survey
4-2. DPES and French Electoral Study

5. Estimation Results
5-1. Estimation Results Using ESS Data
5-2. Estimation Results Using DPES
5-3. Estimation Results Using French Electoral Study
5-4. Variable Selection Using LASSO

6. Conclusion

References