Perceived technological threat and vote choice: evidence from 15 European democracies

Abstract

The political consequences of workplace technological adoption have become a focus of recent party politics research. This article contributes to this literature by directly examining how the perceived threat of technological change relates to support for populist and non-populist left and right parties. It does so in two ways: first, by examining subjective rather than objective automation exposure, and second by distinguishing between personal and collective threat perceptions. Using vote choice data from 15 European countries, this article shows that subjective perception of personal automation exposure relates to increased support for left parties and decreased support for populist-right parties, while concern over collective risk relates to increased support for the populist right. These patterns suggest that fear of workplace technological change elicits both material and status concerns. The article concludes with counterfactual analyses demonstrating that both non-populist and populist left-wing parties could benefit by mobilizing voters who feel personally threatened by automation.

Keywords:

• Technological change
• automation and artificial intelligence (AI)
• populist and radical right voting

Acknowledgments

The authors would like to acknowledge helpful feedback from colleagues at Stanford University’s Center of Advanced Study in the Behavioral Sciences and Simon Fraser University’s Department of Political Science.

Disclosure statement

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

Notes

1 A separate body of literature has focused on subjective measures of automation exposure, but in the context of policy attitudes such as support for income protection and other redistribution (e.g., Busemeyer et al. 2023; Gallego et al. 2021; Jeffrey 2021; Kurer and Haüsermann 2022), rather than vote choice.

2 See Gingrich (2019) and Kurer (2020) who make a similar argument, but applied to individual material and status threat, rather than collective threat.

3 Although it does not consider vote choice, related work on policy preferences shows that sociotropic automation threat has little influence on voters’ prioritization of economic issues (Heinrich and Witko 2022).

4 Given the missingness in our data, we also show in Online Appendix A.12 that our core findings are robust to multiple imputation.

5 We note that findings examining the subset of countries that have a populist-right but no populist-left contender are most sensitive to the inclusion or exclusion of the left-right variable—particularly the finding that personal threat increases left voting. Results for the subset of countries with both populist-right and populist-left contenders are less sensitive to controlling for ideology. Why this is the case is a matter for future work, but we consider left-right ideology to be an important control for the reasons we outline above.

6 While we theorize populism and nativism to be pre-existing latent traits that shape how people interpret economic shocks, it is also plausible that such shocks can increase populist and nativist sentiment. Although we do not analyze this possibility here, we note that our analyses are generally not sensitive to the exclusion of both variables. The one notable exception is the relationship between personal automation threat and populist-left voting, which disappears when populism and nativism are not included as controls.

7 Although we consider the UK’s Sinn Féin to be a populist-left party, it only runs candidates in Northern Ireland, so most of the country is not presented with a populist-left candidate. For this reason, we include the UK in the group of countries with only a populist-right option.

8 Note that our regression models use a continuous measure of age, but we describe the correlates of automation threat using age intervals for expository purposes.

9 We also examine several alternative specifications. In Online Appendices A.7–A.9, we report multinomial models that examine 5-, 10-, and 25-year personal and collective automation threat separately as independent variables. In Online Appendix A.10 we report results from a multinomial model that separates out Green parties.

10 All other variables are held at their observed values.

11 As noted, because Ireland is the only country in our sample with a populist-left but not populist-right party, we set it aside for further analysis. Online Appendix A.13 shows how personal and collective automation threat relate to vote choice in Ireland. Although there appears to be no relationship between personal automation threat and non-populist or populist-left voting in Ireland, we do observe a positive relationship between collective automation threat and populist-left voting. In future work, we hope to further examine the relationship between automation and populist-left voting in a larger sample of countries with populist-left parties.

12 We consider Italy to be a country with populist-left and populist-right parties. However, only a small number of respondents voted for Italy’s populist-left party, Potere al Popolo! (n = 6). As a robustness check, we therefore also ran models where Italy was included in the subset of countries with only populist-right parties. Results are substantively similar, see Online Appendices A.14 and A.15.

Additional information

Funding

This project was funded with support from Canadian Social Sciences and Humanities Research Council [IDG 430-2000-498].

Notes on contributors

Sophie Borwein

Sophie Borwein is an Assistant Professor in the Department of Political Science at the University of British Columbia. She is interested in the politics of place, technological change, and political behaviour. [[email protected]]

Bart Bonikowski

Bart Bonikowski is an Associate Professor of Sociology and Politics at New York University. His research examines political culture in the United States and Europe with a particular focus on nationalism, populism, and the mainstreaming of radical right parties. [[email protected]]

Peter John Loewen

Peter John Loewen is the Director of the Munk School of Global Affairs and Public Policy and the Robert Vipond Distinguished Professor in Democracy at the University of Toronto. He is interested in voters, politicians, and technologies. [[email protected]]

Blake Lee-Whiting

Blake Lee-Whiting is a doctoral candidate in the Department of Political Science at the University of Toronto. His research pertains to the politics of technology, survey methods, and the disruptive effects of emerging technologies. [[email protected]]

Beatrice Magistro

Beatrice Magistro is a postdoctoral research associate in computational social science at the California Institute of Technology. She is interested in technology and political behaviour. [[email protected]]