Do crises and external shocks undermine public support for authoritarian governments? This paper examines whether natural disasters trigger a public backlash against authoritarian incumbents, focusing on the February 2023 earthquakes in Turkey. With national elections held three months later and local elections a year after the earthquake, the Turkish context provides a unique opportunity to test how natural disasters affect public accountability in an authoritarian setting.

Combining georeferenced earthquake intensity data with electoral results across 45,000 Turkish localities, I use synthetic difference-in-differences to test pre-registered hypotheses about the earthquake’s effect on incumbent vote share and turnout. These analyses are complemented with survey data using an unexpected event (earthquake) during survey approach.

Findings show that voters hold authoritarian incumbents accountable for natural disasters, though the effects are complex and heterogeneous. In the 2023 national elections, the earthquake reduced AKP support in affected regions by about 2.5 percentage points. Turnout exhibited a parabolic response: moderately hit areas saw higher mobilization, while severely affected localities experienced demobilization due to displacement. By contrast, the 2024 local elections showed a weaker punishment effect.

This study contributes to research on public accountability in non-democracies and the electoral consequences of crises. It demonstrates that even in competitive authoritarian regimes, natural disasters can expose government failures and trigger backlash. The magnitude and duration of this effect depend on disaster intensity, government responses, and the political landscape. These findings highlight the need to account for effect heterogeneities in disaster studies and caution against generalizing across disaster severities.