Electronic Voting Reduced Blank Votes But Altered Null-Vote Competition in Brazil

📌 Context and Question

Electronic voting machines were introduced nationwide in Brazil for the first time in the 1998 general elections. In that year, some cities used electronic machines while others continued with paper ballots, creating variation useful for assessing how voting technology changed invalid-ballot behavior. The focus is on two types of invalid ballots—blank votes and null votes—and how their frequencies changed across four offices: federal deputy, state deputy, president, and governor.

📊 Comparing 1994 and 1998 Election Results

• City-level results from the 1998 elections were compared with results from the previous general elections in 1994.
• The comparison leverages cities that adopted electronic voting machines in 1998 versus those that retained paper ballots to isolate changes associated with the new technology.
• Outcomes measured: percentage of blank votes and percentage of null votes for each of four posts (federal deputy, state deputy, president, governor).

🔎 Key Findings

• Electronic voting reduced the percentage of blank votes across all four offices: federal deputy, state deputy, president, and governor.
• Prior studies had shown significantly lower invalid votes for federal deputy in electronic-voting cities; this analysis extends that observation to multiple offices.
• Effects on null votes were mixed:
• Null votes reduced competition for the posts of federal deputy and state deputy.
• Null votes increased competition for the more visible executive posts of governor and president.

⚠️ Why It Matters

These results show that the introduction of electronic voting machines changed not only the incidence of blank ballots but also the role of null votes across different kinds of offices. Changes in invalid-ballot behavior can alter perceived competitiveness and voter expression in both legislative and executive races, with implications for how voting technology shapes electoral outcomes and voter signaling.