This study investigates Brazilian voters' inconsistent behavior regarding mayor malfeasance, using a field experiment partnered with the State Accounts Court of Pernambuco during 2016 municipal elections. We show that while voters express strong negative reactions to official wrongdoing in hypothetical vignettes, this sentiment does not translate into actual voting decisions.

Methodology & Context

* Collaborated with Brazil's independent anti-corruption court (Tribunal de Contas do Estado de Pernambuco - TCE-PE) during the 2016 municipal elections campaign.

* Conducted a field experiment where mayors in targeted municipalities were informed of their malfeasance.

Key Findings

* Voters who received information about official wrongdoing did not change their self-reported voting intentions.

* The discrepancy highlights the gap between general anti-corruption norms and local electoral behavior.

Why It Matters

* This reveals that voters' knowledge of corruption does not automatically translate into action against it at polling stations.

* Election day decisions are influenced by other practical concerns, such as employment or health services, outweighing the salience of anti-corruption norms.