Overview
Voter turnout has been declining in democracies around the world, and low participation raises real questions about whether election outcomes actually represent the public. Some countries have tried to fix this by making voting mandatory. But does it work?
I estimated the causal effect of compulsory voting laws on voter turnout using a country-year panel covering 202 countries from 1945 to 2025. The design is difference-in-differences: I compared changes in turnout before and after countries adopted compulsory voting laws, relative to countries that never adopted them. I used two-way fixed effects (country and year) with standard errors clustered at the country level. An event study validates the parallel trends assumption.
Compulsory voting increases turnout by about 10 percentage points (p = 0.005). With an average turnout rate of 71%, that's roughly a 14% relative increase. The event study shows turnout is stable in the years before adoption and jumps immediately after, suggesting the effect is both real and lasting. Switcher countries and never-treated countries had nearly identical turnout before adoption (70.3% vs 70.1%), which strengthens the causal interpretation.
Data
Turnout is measured as the percentage of voting-age population that casts a ballot, from the International IDEA Voter Turnout Database. The treatment variable is a binary indicator for whether a universal compulsory voting law was in effect, manually coded from official government records. The merged dataset has 2,441 country-year observations. Since I'm only measuring whether the law exists (not how strictly it's enforced), this is an Intention-to-Treat estimate.
Robustness
- Adding country-specific linear time trends: effect is 6.75 p.p.
- Leave-one-out test (dropping one country at a time): effect ranges from 7.75 to 11.88 p.p.
- No single country drives the result.