Shifting Ground: The 2023 Earthquake and Voting in Turkey

• In Event: Autocratic Dynamism and Democratic Backsliding in MENA

Abstract

Do voters punish authoritarian governments for calamities of nature? Previous studies have shown how natural disasters can affect voting in democratic countries (Blankenship et al., 2021; Gasper & Reeves, 2011; Heersink et al., 2017; Morris & Miller, 2023; Ramos & Sanz, 2020). However, less is known about how natural disasters affect voting in authoritarian contexts, where political competition is limited and accountability mechanisms are weaker. Moreover, disasters in non-democracies affect more people than in democracies with good governance (Kahn, 2005; Persson & Povitkina, 2017). Therefore, it is important to understand how citizens react to natural disasters in these settings.

We address this question by analyzing the 2023 Turkish elections, which took place three months after a devastating earthquake. Turkey has been characterized by scholars as a competitive authoritarian regime, being dominated by president Recep Tayyip Erdogan and the ruling AKP party (Laebens & Öztürk, 2021; Levitsky & Way, 2020). Importantly, elections in Turkey remained a viable channel for political accountability (Svolik, 2023). Despite the president and his party securing their continued tenure, the 2023 elections were the most competitive within Erdogan's rule. We hypothesize that at least a part of the vote share decrease of the incumbent president and his party was due to the earthquake.

Using the electoral outcomes at the locality level and georeferenced data on the earthquake intensity from the United States Geological Survey, we employ causal panel data methods (difference-in-difference and matrix completion) to estimate the effect of the 2023 earthquake on voting for the incumbent authoritarian president Erdogan and the ruling AKP party. We test five literature-driven hypotheses that were pre-registered on EGAP (https://osf.io/2gb6n) before the election. Specifically, we test the effect of the earthquake on the vote for the incumbent and the turnout, while also exploring the moderating role of the experienced damage, government efforts and prior partisanship. For testing these hypotheses, we complement the electoral results and seismological data with the damage estimates from the official government registry of property units, relief effort records as well as socioeconomic statistical data from the Turkish Statistical Institute. To our knowledge, our study is the first one to use nationwide Turkish neighborhood and village-level data on electoral outcomes, as previous studies have resorted to district-level data (Livny, 2021).

Methodologically, we rely on both traditional difference-in-difference methods and a cutting-edge matrix completion technique (Athey et al., 2021), which has yet rarely been used in studies of authoritarianism. Thus, we address some potential identification concerns as matrix completion does not rely on the parallel trends assumption and usually provides more conservative and less biased estimates than diff-in-diff (Athey et al., 2021). Both methods indicate a negative effect of the earthquake on support for both President Erdogan and the incumbent AKP party, around one percentage point for each point of earthquake intensity that varied between 3 and 7.7. As the whole country was directly or indirectly affected by the earthquake, we regard our findings as a lower bound of a potential disaster effect relative to a situation where control units would have remained unaffected.

These findings shed new light on political accountability in non-democratic settings. We contribute to this field (Akbiyik & O’Donohue, 2023; Beazer & Reuter, 2019; Rosenfeld, 2018; Svolik, 2023) by exploring whether voters hold authoritarian governments accountable for their performance exposed by exogenous shocks. Finally, we add to the growing literature on the political consequences of natural disasters (Blankenship et al., 2021; Gasper & Reeves, 2011; Heersink et al., 2017; Morris & Miller, 2023; Ramos & Sanz, 2020) by examining how a major earthquake affects turnout and vote share for the incumbent president and his party in a competitive authoritarian regime.

Authors

• Konstantin Bogatyrev, Bocconi University

• Hande Tugrul, Bocconi University