From Shock to Disengagement: How Natural Disasters Reshape Politics in Low-Trust Democracies

Technical Summary: From Shock to Disengagement: How Natural Disasters Reshape Politics in Low-Trust Democracies

Problem Statement and Theoretical Context
This research addresses a critical gap in the literature regarding the political consequences of natural disasters within advanced democracies characterized by low levels of political trust. While existing scholarship has extensively documented how citizens react to crises—often citing "rally-around-the-flag" effects or electoral punishment of incumbents—these models typically presuppose a baseline of state-citizen trust. The authors argue that in low-trust environments, such as contemporary Spain, standard mechanisms of democratic accountability may fail. Specifically, when trust in representative institutions is already fragile, natural disasters act as "performance shocks" that do not merely shift support between incumbent and opposition parties. Instead, they may push political confidence past a "tipping point," leading to a disengagement from representative institutions, the strengthening of anti-political attitudes, and a shift toward non-political actors or radical alternatives.

The study posits that political trust is the most performance-dependent dimension of trust, distinct from the more resilient social or personalized trust. In low-trust contexts, the lack of a "buffer" means that even modest declines in perceived government performance during a crisis can cause citizens to exit the representative system entirely, rather than seeking alternation through established opposition parties.

Methodology and Data
The authors test their theoretical framework using the October 29, 2024, DANA floods in the Valencia region of Spain, a case selected for its severity (232 fatalities, ~$11 billion in damages) and the widespread public criticism of the regional government's crisis management. Spain serves as an ideal context due to its historically low levels of political trust, which have remained depressed since the 2008 financial crisis.

The empirical strategy employs a mixed-methods design combining observational analysis, a pre-registered conjoint experiment, and panel data:

• Sampling: The study utilizes a field survey conducted in two waves: immediately following the disaster (December 2024–January 2025) and one year later. The sample consists of 1,500 observations across three groups defined by crisis exposure intensity:
  • High Exposure (HE): Residents in municipalities most severely affected (designated "zona cero").
  • Medium Exposure (ME): Residents in Valencia province municipalities not severely affected.
  • Low Exposure (LE): Residents in municipalities outside Valencia province but within the region.
• Variables: The primary treatment variable is crisis exposure intensity. The study controls for gender, age, education, and ideology, with standard errors clustered at the municipality level.
• Robustness: To address potential confounding factors, the authors conducted Coarsened Exact Matching (CEM) and verified that the groups were balanced on individual sociodemographic characteristics and most electoral characteristics. However, they explicitly note a significant pre-existing imbalance: municipal PP vote share in the 2023 local elections was significantly lower in High Exposure municipalities (31.8%) than in Low Exposure ones (39.4%). By contrast, municipal VOX vote share and registered unemployment rates were balanced across groups. The authors also utilized an alternative operationalization of exposure based on geodesic distance to the flood footprint.

Key Results
The analysis yields four primary findings corresponding to the study's hypotheses:

1. Erosion of Political Trust (H1): Highly exposed residents exhibit significantly lower levels of trust in political actors (specifically regional and central governments) compared to low-exposed residents. This effect is most pronounced for the regional government (a drop of ~0.67 points on a 0-10 scale in the first wave) and persists, though attenuated, one year later. Notably, trust in non-political implementing institutions (e.g., the army, security forces) and societal actors (volunteers, experts) remains stable or unaffected, indicating a targeted erosion of confidence in representative institutions.
2. Strengthening of Anti-Political Attitudes (H2): Increased exposure correlates with stronger anti-political attitudes, including higher support for populist ideas and technocratic decision-making.
3. Preferences for Crisis Management (H3): A conjoint experiment reveals that citizens broadly prefer crisis committees composed of institutional and technical actors over political representatives. However, this preference is uniform across all exposure groups, suggesting that the desire for non-political management is a general sentiment rather than a specific reaction to the disaster's intensity. Crucially, citizens do not view grassroots societal actors (volunteers/NGOs) as substitutes for political management; societal committees were the least selected option overall.
4. Electoral Behavior and Disengagement (H4): The study finds that dissatisfaction does not translate into a standard shift toward mainstream opposition parties. Instead, the primary electoral consequence in the most affected areas is persistent electoral demobilization (abstention). There is tentative and intensifying evidence of a limited shift toward the radical-right party VOX in the most severely affected areas, but no evidence of a shift toward traditional opposition parties. The apparent punishment of the regional incumbent (PP) is largely attributed to pre-existing differences in PP support across exposure zones rather than a causal effect of the disaster itself.

Significance and Contributions
The paper makes four distinct contributions to the literature:

1. Contextual Focus: It directs attention to the neglected context of advanced democracies with low political trust, challenging the assumption that high state capacity automatically mitigates the political costs of disasters.
2. Theoretical Mechanism: It develops a "tipping-point" argument, suggesting that in low-trust contexts, disasters push citizens past a threshold of disengagement, leading to electoral demobilization and support for radical alternatives rather than conventional alternation.
3. Precondition vs. Outcome: It reframes political trust not merely as an outcome of disasters but as a critical precondition that shapes how citizens interpret performance shocks.
4. Attitudinal and Behavioral Integration: By combining attitudinal measures with electoral behavior, the study demonstrates that the political consequences of natural disasters extend beyond vote switching to include deeper disengagement and the erosion of faith in representative institutions.

The authors conclude that in low-trust democracies, natural disasters may exacerbate political fragmentation and disengagement, creating a vacuum that is not filled by mainstream opposition but rather by demobilization or support for anti-system forces.