Electronic Voting in Brazil

electronic voting effects on enfranchisement, health, and political outcomes

hand placing ballot in ballot box with brazil flag in background

Brazil launched electronic voting (EV) in 1998. In 1994 elections were done through paper ballots only and by 2002 the transition was completed to full EV. The introduction of the EV system aimed at reducing the time and costs of vote counting. Simultaneously, other features of the technology, such as the use of candidates’ photographs as visual aids, the use of error messages for voters about to cast residual votes, and guiding the voting process step by step, facilitated voting and reduced errors. Thomas Fujiwara’s analysis provides detailed insight: Voting Technology, Political Responsiveness, and Infant Health: Evidence from Brazil. He uses a regression discontinuity design embedded within the EV rollout in Brazil. EV devices were only distributed to municipalities with more than 40,500 registered voters while the rest used paper ballots.

In 1994, only 65% of ballots were valid. In 1998, paper ballot validity was 75% and electronic ballot validity was 90%. In 2002, electronic ballot validity was also 90%. Fujiwara concludes a de facto enfranchisement of 10% of voters, 3.4 percentage point (pp) increase in health expenditure, 7.7pp increase in the likelihood that a pregnant woman would attend at least one prenatal care visit, and a 0.5pp decrease in low-weight births by less educated women. He finds that the increased enfranchisement of these less educated voters increased the number of valid votes for left-wing candidates and led to higher rates of health expenditure by 3.4pp which then translates to the observed health outcomes.

The RD design Fujiwara utilizes provides a strong causal inference that the EV rollout caused the observed improvements in health outcomes. There is a significant, discontinuous increase in ballot validity of 15pp in 1998 comparing the paper ballot to electronic voting. The same trends are not seen in 1994 and 2002. Furthermore, the possibility of manipulation in voter population is low as the populations of municipalities show a continuous proportion of registered voters and turnout across all municipalities. Additionally the series of robustness checks including looking at municipality fixed effects like monthly income, income inequality, education variables, geographical location, etc. to show that populations on both sides of the cutoff are indeed comparable.

A potential critique of the study is that it focuses on a specific context and may not be generalizable to other countries. Brazil has unique political, social, and economic characteristics that may not be present in other countries or systems. Thus, it is important to be cautious when drawing conclusions about the effects of electronic voting technology in other contexts. Furthermore, while Fujiwara controls for several factors that could affect the validity of votes, he does not account for the interest of voters in voting itself. It is possible that voters who are more interested in the political process are also more likely to cast valid votes. This could be a potential confounding factor that may have contributed to the observed increase in ballot validity. Interest surveys could have potentially mitigated this issue. Finally, the long-term effects of the policy are not entirely clear. It is possible that the effects of electronic voting technology on health outcomes and political responsiveness were temporary and will not improve over time. Additionally, it is unclear whether the observed increase in health expenditure was a one-time increase or a sustained increase that will continue to benefit the population. These questions can be explored in future research of electronic voting.


Ishil Puri