The Anti-democratic U-Turn: The curvilinear relationship between trust in democratic political institutions and support for anti-democratic governments

Trust in political institutions is central to representative democracies, as numerous studies have shown that it significantly relates to participation, voter turnout and choice, and policy compliance (Devine, 2024). Recently, many democracies have experienced declining trust in partisan institutions (Valgarðsson et al., 2024), alongside growing support for anti-democratic alternatives (Lührmann & Lindberg, 2019). Authors such as Foa and Mounk (2017) suggest that low trust reinforces the appeal of anti-democratic alternatives, while others that low trust is merely an epiphenomenon of the emergence of a “critical citizen” who still supports democracy (Norris, 2022). This thesis investigates the relationship between trust in democratic political institutions and support for anti-democratic governments, challenging the conventional assumption that trust uniformly strengthens democratic stability. Adopting a multi-study approach, the research begins with an exploratory analysis (Study 1) based on a large quota sample of the Italian adult population (COCO dataset), which reveals a counterintuitive positive association between institutional trust and support for anti-democratic governments. Further analyses demonstrate that this relationship is non-linear, taking the form of a U-shaped curve: both low and high levels of institutional trust are associated with greater anti-democratic support. To assess the robustness of this finding, Study 2 employs both longitudinal and cross-national data. Using repeated measures from the Italian sample over six waves (April 2020–October 2022), the curvilinear relationship is shown to be stable over time. Additionally, analyses of the European Values Study (2017–2022) across 37 countries replicate the U-shaped pattern across diverse political contexts. This phenomenon is labelled the “Anti-Democratic U-Turn” (ADUT). Study 3 explores potential psychological moderators of this relationship, drawing on the Motivated Social Cognition framework (Jost et al., 2003). Using secondary data, it tests the moderating roles of authoritarian predispositions and epistemic needs through trust in science as a proxy for need for cognition (NFC). Results indicate that high authoritarian predispositions strengthen the positive association, while trust in science can reverse it at high levels of institutional trust. These findings suggest that institutional trust varies in form and implication depending on underlying psychological characteristics. Study 4 builds on this insight using a new Italian quota sample (N = 1,202) and improved measurement instruments. It examines a comprehensive set of moderators, including authoritarian predispositions, dogmatism, cognitive rigidity, need for cognitive closure (NFCC), and NFC. The results confirm that the positive slope at high levels of trust is contingent on specific individual differences: authoritarian predispositions and NFCC amplify it, dogmatism activates it, and cognitive rigidity and low NFC can reverse the typical negative association. These patterns support the distinction between critical and uncritical forms of institutional trust. Finally, Study 5 employs a 2×2 between-subjects vignette experiment to test causality. By manipulating both institutional trust and critical processing, the study shows that uncritical high trust increases support for anti-democratic governments relative to critical high trust, while this distinction is non-significant at low levels of trust. Overall, the thesis demonstrates that institutional trust can have ambivalent effects on democratic support. The identification of the ADUT effect challenges dominant theories and highlights the importance of distinguishing between critical and uncritical trust through their psychological foundations. The findings have implications for democratic resilience, suggesting that fostering reflective and critical forms of trust is essential to prevent democratic backsliding.