Relief Didn’t Save Hoover: 1927 Flood Cut His Vote by Over 10 Points

📚 Why This Question Matters

Do natural disasters help or hurt incumbents at the ballot box? Prior work offers conflicting answers: Achen and Bartels (2002, 2016) argue that voters punish incumbents indiscriminately after disasters, while other studies find voters reward or punish officials according to the quality of relief efforts. The literature’s focus on contemporary cases may shape these divergent conclusions.

🗂️ The 1927 Flood and the Historical Opportunity

The 1927 Mississippi River flood in the American South created a large, catastrophic shock paired with unusually broad and fair distribution of federal disaster aid. Herbert Hoover—later the 1928 Republican presidential candidate—was personally responsible for overseeing the relief effort, providing a rare historical test of whether high-quality, well-distributed aid shields politicians from electoral punishment.

🔬 How Evidence Was Tested

• County-level comparisons of vote outcomes in the 1928 presidential election are used to measure electoral change in affected versus unaffected areas.
• Synthetic control methods are applied to construct counterfactuals and assess robustness of estimated vote-share changes.

📈 Key Findings

• Despite unprecedented levels of disaster aid that were broadly and fairly distributed and Hoover’s direct role in relief, affected counties saw Hoover’s vote share fall by more than 10 percentage points.
• Results are robust to synthetic control analyses, suggesting the estimated negative effect is not driven by simple preexisting trends.
• Even if voters can distinguish low- from high-quality responses, the aggregate effect of this catastrophic disaster was broadly negative for the politician associated with relief efforts.
• Findings lend some support to the notion of blind retrospection advanced by Achen and Bartels, while also highlighting the need to unpack the causal mechanisms by which damage and relief influence vote choice.

💡 What This Means

The historical case shows that large, well-managed relief does not guarantee electoral reward; aggregate electoral responses to disasters can be negative even when relief quality is high. This outcome challenges simple expectations that good performance on relief will shield incumbents and motivates further research into the pathways—information, attribution, timing, or emotional reactions—through which disasters reshape voter behavior.