Does a change in Chilean electoral law from compulsory to voluntary voting affect voter turnout?
This note introduces new methods for causal inference with ecological data. By combining aggregate election results and individual demographic information, it addresses the challenge of estimating effects on individual behavior.
Using Aggregate Data & Individual Demographics
Researchers combine large-scale voting patterns with detailed personal characteristics from separate datasets to infer individual-level causality.
Applying Sharp Bounds Analysis
The analysis uses recent advancements in partial identification literature to provide precise limits for causal effects, despite the 'selection on observables' assumption.
Case Study: 2012 Chilean Election Reform
Data from mayoral elections before and after the law change reveals that voter turnout estimates were previously overstated. The bounds analysis shows the reform actually decreased expected turnout.
This approach demonstrates how to identify causal effects in ecological inference problems, offering clearer insights for political scientists studying voting behavior.
