Are citizens willing to tolerate violations of democratic rules and norms by competent political leaders? I examine this question employing well-powered conjoint experiments in the United States, the United Kingdom, the Czech Republic, Mexico, and South Korea. The findings yield good as well as bad news for democracy. I find that undemocratic behavior by political candidates decreases voter support, whereas candidate competence increases support. Contrary to expectations, the effects of undemocratic behavior and competence do not interact. This means that competent candidates are sanctioned for violating democratic principles but also that support for undemocratic candidates increases with their competence. These findings can help explain the successes and failures of undemocratic political leaders around the world: although these leaders can gain support by appearing to be competent, competence does not make citizens tolerate undemocratic behavior entirely.

Democratically elected incumbents have engaged in subverting democracy without losing popular support in several countries in recent years. It remains a puzzle that this phenomenon has occurred in relatively mature democracies. In this research note, I argue that citizens become less likely to sanction undemocratic behaviour as their country gains democratic experience because they lose incentives to form opinions on the basis of threats to democracy. Using macro-level panel data on democratic experience, undemocratic incumbent behaviour and incumbent approval across 43 democracies from 1962 to 2018, I find that undemocratic incumbent behaviour decreases incumbent approval when democratic experience is low. As democratic experience increases, however, this effect fades out. These findings question whether we can count on citizens to be the backbone of democratic stability in countries with democratic experience.

The rise of authoritarian government in recent decades have sparked scholarly debate on what makes citizens tolerate and punish undemocratic behavior. Extant explanations hold that citizens may vote for undemocratic politicians if these politicians represent their political interests or hold a competence advantage over competitors. In this paper, I consider a different explanation. Specifically, I focus on whether citizens punish undemocratic behavior more when they perceive democracy in their country as vulnerable and provide experimental tests spanning over two data collections from 2020 and 2021 with a total of 12 country samples. I find that voters punish undemocratic behavior, that they punish it more severely than prior studies have suggested, and that undemocratic behavior drives citizens toward voting for alternative options rather than decreases turnout. Contrary to expectations, I also find that democratic vulnerability sometimes but only in a very restricted set of cases matter for sanctioning of undemocratic behavior.

What free speech restrictions would citizens like to impose? We provide answers to this question by combining global, descriptive evidence on support for free speech with conjoint experiments fielded in the United States, Germany, and Hungary. In support of our theoretical arguments, we find that willingness to restrict free speech is widespread. Citizens are more willing to restrict statements on the sensitive issues of economic stability, epidemics, and national security or offensive statements toward minorities and national symbols than they are to restrict standard government criticism. We also find that willingness to restrict outright antidemocratic statements is relatively modest in comparison with offensive statements and statements on sensitive issues, and that principled support for free speech is a poor predictor of actual willingness to restrict. Our findings imply that citizens’ (lack of) demand for fundamental democratic rights such as free speech may play a key role in democratic backsliding.

Do the electoral constraints imposed by two-party systems make citizens less likely to defect from undemocratic candidates? I employ two innovative candidate choice experiments fielded in England to answer this question. Specifically, I implement two designs manipulating the number and effective number of parties displayed between two and three, exploiting the characteristics of England's party system. Contrary to expectations, I find that Labour and Conservative identifiers do not defect more from undemocratic in-partisan candidates when they face three (effective) parties---Labour, the Conservatives, and the Liberal Democrats---rather than just the two major parties. Instead, defection from undemocratic in-partisans to the out-party drops and relocates to the Liberal Democrats even when the latter have no chance of winning. These findings highlight that having three rather than two parties does not generate more defection from undemocratic politicians---and that voters prefer defecting to the option ideologically nearest to the in-party even when chanceless.