Key Question 📚 State repression effectiveness remains unclear despite its prevalence in nondemocracies. This research examines county-level voting patterns near military bases before Chile's 1973 coup using a novel spatial analysis approach.

Data & Methods 🔬 We analyze base locations pre-coup and correlate them with two key measures: government killings/disappearances during the Pinochet era, and electoral registration/voting "No" decisions in the crucial 1988 plebiscite. Our identification leverages geographic discontinuity to isolate repression's spatial impact from political motivations.

Key Findings ✅ Counties near military bases experienced higher government violence under Pinochet yet showed increased likelihood of:

• Registering voters
• Casting a "No" vote in the 1988 plebiscite, despite living closer to violent incidents

This geographic paradox suggests repression's effectiveness may be inversely related to its visible intensity. Two potential mechanisms emerge: residents near bases receive more direct information about repression costs, potentially altering their political calculus; alternatively, proximity might reduce salience or trust in the regime but requires testing with additional data.

Why It Matters 🌍 The 1988 results represent a pivotal moment in Chile's democratic transition. Intriguingly, this study finds no evidence that the spatial effects of repression influenced political preferences post-transition despite their documented electoral impact during Pinochet's rule.