Undemocratic visions of democracy? Evidence from a conjoint experiment in five European countries

In recent years, pundits and social scientists alike have observed a broad trend towards democratic backsliding across Europe. This includes the success of far-right populist parties, both in opposition and in government, the rise of authoritarian attitudes among the people, and the undermining of institutional checks and balances by governing elites. At the same time, however, we have also observed consistent popular support for democratic values. For example, many cross-national surveys indicate that, when asked general questions about democracy, people tend to be supportive. The extant literature on democratic trade-offs, especially in highly polarised countries such as the US, tries to explain this apparent paradox by arguing that political polarisation is associated with greater tolerance for undemocratic behaviour. Using novel data from a cross-country survey with an embedded conjoint experiment in five European countries, this paper contests the applicability of the polarisation argument in the European context. Specifically, we focus on the following five European countries, each as a representative of a specific European region: Estonia (Baltics), Spain (Southern Europe), Germany (Western Europe), Poland (Central Europe), and Sweden (Scandinavia). Our survey asks respondents a range of questions aimed at assessing their understanding of democracy as well as their positions on a range of democracy-related issues. Next, we present respondents in each country with two sets of candidate choice experiments (ten choices per set) to assess their willingness to support an ideologically proximate candidate who might hold an undemocratic position. Our findings suggest that polarisation is not necessarily associated with tolerance for undemocratic behaviour in the five European countries we focus on. Specifically, our analyses reveal that the policy issues for which voters are willing to sacrifice democracy are not always polarised. In addition, there is a stronger willingness to forgive undemocratic behaviour for social identity related interests than for socio-economic ones. Finally, we show also important differences across party identification. In sum, while our results are in line with the argument that voters are increasingly prepared to support ideas that are in principle democratic through undemocratic means, we find that in Europe these attitudes are not driven by political polarisation. Our paper contributes to the broad literature on democratic stability in advanced democracies by showing that the dynamics of democratic trade-offs are different in the context of European multi-party systems.