Brazil: Decision Making Process on Electronic Voting
Brazil began shifting toward electronic voting in 1994. The impetus for the initial move to e-voting was largely led and managed by the TSE. The TSE has jurisdiction over all aspects of elections in Brazil and regulates the functioning of political parties. Over its history, the TSE has developed a reputation for trustworthiness, competence and autonomy in the management of the electoral process. In addition to its election management role, the TSE is also responsible for revising the electoral law every two years and submitting it to the legislature for approval, as required by Brazilian law. Because of its good reputation, the electoral law submitted by the TSE is rarely debated; and this gave the TSE significant leeway to purse electronic voting as a solution to challenges faced by the electoral process. While outside actors had some input, the move to electronic voting was largely an autonomous process carried out by the TSE, and consequently, actors within the judicial institution made most of the major decisions.
There were two primary reasons why the TSE adopted electronic voting machines (EVM). The first was to combat endemic fraud in the paper ballot tabulation process. The second was to address issues related to electoral accessibility and spoiled ballots in the paper voting system. Due to Brazil’s complex electoral rules, voters regularly have to choose from thousands of legislative candidates. This makes results tabulation a logistical challenge because the paper voting system involves hundreds of thousands of vote counters who were often government employees from State-owned banks or the postal service. Because of the scale of the task, vote counting could take weeks and this post-election period was a time of great uncertainty and tension.
Most importantly, the lengthy tabulation period increased opportunity for vote counters allied with candidates to manipulate the vote count because the lengthy vote count was difficult for partisan and other civil society actors to fully monitor. The most common type of fraud was manipulation of the tabulation sheets known as “maps” where vote counters who were allied with candidates would subtract votes from one candidate’s tally and add them to the favored candidate’s count.45 This type of electoral fraud became a national issue after the 1994 presidential and legislative elections when a scheme to manipulate the election results involving electoral judges was uncovered in Rio de Janeiro. The local branch of the TSE was forced to annul the results for the legislative elections and hold a new one, leading to questions about the pervasiveness of fraud in elections.
A secondary motivation for switching to electronic voting was due to accessibility problems in the paper-based system. This system was a hugely complicated, as it required voters to write in the name or identifying number of their preferred legislative candidate. Two factors, the large number of candidates in legislative races, as well as the level of illiteracy in the country (approximately 20 percent, according to the 1990 census) resulted in almost 40 percent of votes being blank or invalid in 1994 legislative elections. These factors were compounded by the fact that, in legislative elections, voters voted for multiple offices and would fill in several names or numbers to cast votes for all offices. TSE officials argued that the high number of blank votes cast could be attributed to illiterate voters, who did not want to take a long time writing in a name, revealing they could not write.
The disenfranchising effect of complex ballots also made fraud easier, as described by Federal Deputy Tourinho Dantas:
If an illiterate voter doesn’t know how to read or write, how can he vote? They humiliate themselves at the moment in which they vote. When he goes to the ballot booth and he doesn’t know what to do, he casts a blank vote. This vote, in the majority of places, is filled out by those perpetrating fraud. It is by this means that fraudulent votes are cast in so many places.46
The initial decision to switch to electronic voting was made by President of the TSE, Minister Sepúlveda Pertence, in 1994. He cited the Rio de Janeiro scandal as a factor:
After the experience we have lived through, not in the poorest regions, but rather in one of the most important cities in the country [Rio de Janeiro], we cannot retreat from the imperative of automation, or if that is not possible, the “mechanization” of the vote.
The impetus to change voting technologies came almost wholly from within the TSE, and was based in part on previous positive experiences with the use of technology in voter registration and results tabulation. When the decision was made between 1994 and 1995, there were no other major societal actors such as political parties, civil society organizations or other government bodies advocating for the abandonment of paper ballots.
In Brazil, a pilot was not carried out to test electronic voting. Instead, a gradual introduction of universal electronic voting was achieved over the course of three elections: in the 1996 elections, 30 percent of voters (33 million) directly voted through the electronic voting machines; in 1998, an additional 30 percent (35 million voters) voted through e-voting machines; and in the 2000 elections, the entire nation voted through electronic voting (100 million voters).
45 In Portuguese slang, this practice was known as “mapism” (“mapismo”).
46 Dantas, Tourinho. Díario do Congresso Nacional, October 27, 1994, p. 13,331
Brazil: Building the System for Electronic Voting and Counting