Voter Fraud, Mobilization and Administration of Mail Ballots in US Elections

• In Event: Evidence about Voter Suppression, Voter Fraud, and Administrative Discretion

Abstract

The 2018 US House Election in New Mexico’s 2nd Congressional District raised concerns about fraudulent mail ballots. After polls closed on Election Day, multiple news organizations called the race for the Republican candidate. These calls were retracted the next morning when a large number of uncounted mail ballots were reported. Final results showed a narrow victory for the Democratic candidate. The number of mail ballots requested in the district’s largest county, Dona Ana, smashed prior levels of mail ballot use in mid-term elections. A large number of provisional ballots were cast on Election Day by voters who election officials reported as requesting a mail ballot, but said they did not do so.



Local media and election officials reported that multiple Democratic aligned civic and political organizations conducted programs to encourage voters to use mail ballots in the 2018 election. Using the official voting records from the New Mexico Secretary of State and additional records from the Dona Ana County Clerk on mail ballot requests, provisional ballots, and the date of voting transactions, we investigate a series of questions to identify possible explanations for the surge in mail ballot use and incidence of provisional ballots.



The official voting data indicates the surge in mail ballot use is consistent with reports of Democratic aligned organizations encouraging mail ballot use. The use of mail ballots among registered Republicans and Republican leaning demographic groups is similar to prior elections. The large increase in mail ballot requests and mail ballots cast is entirely among registered Democrats and Democratic leaning demographic groups. Since past field experiments have found encouraging voting by mail to be an effective voter mobilization tactic, especially in lower salience elections, the surge in mail ballot use seems likely to be due to effective campaign tactics, and not voter fraud.



Digging further into the data, we explore the relationship between the ballot request date and casting a provisional in person ballot. If mail ballot recruitment is conducted far before Election Day, it may lead to high rates of provisional voting. The likely (but untested) mechanism is that voters unfamiliar with mail ballots “forget” about the request. In addition to insights about voting behavior, this patterns suggests campaigns should reconsider when to conduct mail ballot recruitment efforts.



We also explore whether there is a relationship between ballot mailing date and casting a provisional in person ballot? Election officials seeking to spread out their workload often mail ballots as far in advance of Election Day as legally permissible. However, this may be prior to voters looking for the ballot in their mail, and thus voters do not recognize the mailing when it arrives. In addition to insights about voting behavior, this pattern suggests that election official reconsider ballot delivery dates.

Authors

• Lonna Rae Atkeson, University of New Mexico

• Christopher B. Mann, Skidmore College