Administration of Mail-In Voting
Measuring Data Across a Nationwide Patchwork of Policy Decisions
The Election Performance Index (EPI) tracks a variety of metrics related to voting by mail—including the traditional domestic vote-by-mail process and the process for military service members and overseas citizens, (also known as UOCAVA voters). To be more specific, the EPI measures data pertaining to the administration of mail voting through four indicators focused on the number of mail ballots unreturned, mail ballots rejected, military and overseas ballots unreturned, and military and overseas ballots rejected in all 50 states and Washington D.C. These indicators have changed dramatically since 2008, when the EPI first started tracking election administration; in addition, the administrative tasks and policies they track have undergone political and administrative upheaval since the COVID-19 pandemic in 2020.
Before we get started, we should note that we generally follow the conventions of the Election Administration and Voting Survey (EAVS) when it comes to terminology. The EAVS no longer refers to domestic, non-UOCAVA mail ballots as “absentee” ballots, since a majority of states no longer require an individual to be physically absent in order to vote by mail. UOCAVA ballots are still categorized as absentee ballots under the EAVS. The EPI follows this categorization in its methodology, website, and other relevant publications. Additionally, although the EAVS collects more detailed information about return modes (mail, fax, email, etc.) for UOCAVA ballots, the EPI calculates the rate of rejected and unreturned UOCAVA ballots using the combined return modes, rather than splitting each out individually.
As we explore these indicators, it is also important to remember that election officials are tasked with enacting a slate of policy decisions determined at the local, county, state, and federal level. To further complicate matters, federal law—the Uniformed and Overseas Citizens Absentee Voting Act—also governs voting by military and overseas citizens in federal elections. While the EPI measures the outcomes associated with those policy decisions, it does not paint the full picture of voting by mail administration. This blog will focus on what the EPI is able to measure, though we also hope to provide additional context about the policy decisions that directly impact these indicators.
Voting by Mail
A Note About State-Level Vote by Mail (VBM) Policies
The Election Administration and Voting Survey (EAVS) serves as the data source of many EPI indicators, including those pertaining to voting by mail and UOCAVA ballots. According to the most recent EAVS report, in 2024 there were:
- 66,867,671 mail ballots transmitted,
- 47,957,093 mail ballots returned,
- 47,629,437 mail ballots counted, and
- 584,463 mail ballots reported as rejected
Compared to 2020, there was a significant decrease—26.3%—in the number of mail ballots transmitted to voters in 2024. However, the ballot return rates for voters who received a mail ballot were similar for both elections (72% in 2024 and 77.8% in 2020).
Most policies related to mail ballot administration are determined at the state and local level. While this is one of the key features of the decentralized network of election administration in the United States, it has created a patchwork of policies across the country, and even within states, governing how mail ballots are administered and returned. In other words, some states’ elections are run entirely with mail ballots, some states have specific jurisdictions that vote entirely by mail, and some states allow a mix of in-person and vote by mail options. The legal landscape of mail voting became even more complicated following the 2020 election, when many states expanded access to mail-in voting during the election to help voters cast their ballots safely. Some states made these changes permanent, others temporary; since 2020, many states have added additional security protocols to mail voting, including restrictions to mail ballot verification processes, required excuses to obtain a mail ballot, and changed deadlines by which mail ballots can be received.
States use a variety of policies to verify that mail ballots have been properly cast by a registered voter. These include states that require a voter’s signature, their signature and identification, a witness’s signature, or signatures from both the voter and a witness.
In addition to variations in verification requirements, states across the country have their own deadlines by which mail ballots must be received (for example, one state may require that a ballot is postmarked by Election Day, which may mean that valid ballots arrive and are counted after Election Day passes, while another requires that a ballot must be received by Election Day in order to be counted), or when administrators are allowed to begin counting verified mail ballots. While the EPI doesn’t measure state mail voting policy directly, it is important to keep these policy-level differences in mind when analyzing the index, as they provide context to the metrics which are reflected in the EPI.
Rejected Rate
The EPI’s “mail ballots rejected” indicator measures the degree to which mail and absentee ballots are rejected as a percentage of all votes cast. The nationwide average in 2024 was 0.34%, the same as 2020 and slightly above the mail ballot rejection rate of the 2016, 2012, and 2008 elections. There are state-level outsiders (like Mississippi, which does not report enough data to be included) in the mail ballot rejection rate, which are likely due to a variety of factors dictated by states’ specific mail ballot policies.
One factor at play here is the variation between states in how much of a state’s voters use mail ballots at all. For example, eight states and Washington D.C conduct exclusively mail ballot elections; their experiences and processes will inevitably look very different from states with much lower levels of mail voting.
Another factor in rejection rates is varying policy regarding mail ballot verification across states. In 2024 and 2020, issues with signature verification were the most common reason a mail ballot was rejected. As we note in the 2024 EPI Methodology, 40.7% of mail ballots were rejected because the signature on the return envelope did not match one on file, and another 10% were rejected because they did not have a signature at all. Another 17.8% of ballots were rejected because they were not received by the state’s deadline, which tends to be a larger challenge for UOCAVA voters.
Unreturned Rate
The EPI’s “mail ballots unreturned” indicator measures the percentage of mail ballots that are sent out, but not returned for counting, as a percentage of all mail ballots transmitted in each state. Like the rejected rate, the nationwide average for unreturned mail ballots in 2024 (11.15%) was in line with the rate in 2022 (11.54%), and slightly higher than the rates in 2016, 2012, and 2008. Also like the rejected rate, there are some outliers in the unreturned rate between states; Alabama, Connecticut, Mississippi, and Nevada did not report enough data to be counted in 2024. The eight states that conduct entirely VBM elections are also excluded from this indicator; because these states send out ballots to all registered voters, the unreturned rate would be overinflated, and disproportionately impact EPI scores as a result.
Ballots go unreturned for countless reasons: voters may experience unexpected moves, deaths, natural disasters, decide not to vote, or simply forget to return their ballot by the deadline. Like with all materials that travel through the mail, occasionally returned ballots may be lost; other ballots may go unreturned when permanent mail ballot lists are not updated frequently enough and become out-of-date.
UOCAVA Voting
What is UOCAVA, and Who Qualifies?
A small percentage of eligible voters are serving in the military or living overseas; these individuals qualify to cast a mail ballot under the Uniformed and Overseas Citizens Absentee Voting Act (UOCAVA). UOCAVA voters can face significant geographic barriers to casting a ballot; as a result, voters in this demographic are allowed to send their ballot by mail, fax, or even electronically, depending on the jurisdiction where an individual is registered to vote. In 2024, the number of ballots transmitted to overseas citizens was more than double the number of ballots transmitted to voters who were uniformed services members—continuing a trend which started as early as 2016.
UOCAVA was first passed by Congress in 1986, and was amended in 2009 by the Military and Overseas Voter Empowerment Act (MOVE Act). The MOVE Act required state election officials to transmit requested UOCAVA ballots to voters at least 45 days before a federal election. In addition to these federal guidelines, UOCAVA voting procedures are also governed at the state-level; while the federal government stipulates ballots must be sent out at least 45 days before an election, each state is able to decide whether its UOCAVA (and other mail ballots) must be postmarked or received by Election Day to be counted, and if the state accepts postmarked ballots after Election Day, how long that grace period is. Despite this extra complexity, the rate of requests for UOCAVA ballots has increased since 2008.
Tracking the rejected and unreturned rates for UOCAVA ballots can be a challenge for researchers using the EAVS, including our team. States do not adhere to a nationwide standard of naming conventions for tracking these ballots, and more research needs to be done (like the work currently being done by the Federal Voting Assistance Program) to determine exactly why UOCAVA ballots are rejected or unreturned. Even so, the UOCAVA unreturned rates in the 2020 and 2024 EPI’s were one of the primary indicators driving changes in state scores, although it is unclear if these changes are a return to pre-pandemic benchmarks or genuine fluctuations caused by policy.
Rejected Rate
The rejected rate for UOCAVA ballots measures the percentage of military and civilian overseas ballots that are returned, but rejected. The EPI measures rejections at the statewide level and divides reasons for rejections into four categories determined by the EAVS: missed deadline, problem with voter signature, lack of postmark, and other. By far, the most common reason for rejection was that a ballot was received after a state’s deadline. While the deadline for mailing out UOCAVA ballots is federally mandated, each state sets its own deadline for when ballots must be returned, as with mail ballots.
Digging into specifics, the EAVS data we use in the EPI shows that of the 30,401 returned UOCAVA ballots rejected, 14,370 were because they were received after a state’s deadline, or 49% of all UOCAVA ballot rejections in total. Compared to previous years, there was a marked decline in the number of UOCAVA ballots rejected; the national average in 2024 was 2.69%, up slightly from 1.62% in 2020, but still lower than previous elections measured by the EPI.
Unreturned Rate
The EPI also has an indicator for the UOCAVA unreturned rate; that is, the percentage of military and overseas civilian ballots that were requested but unreturned by voters in each state. This indicator displays similar trends to the unreturned rate of domestic mail ballots, as UOCAVA ballots are often affected by the same hurdles when it comes to both governing policies and voters themselves.
In 2024, the unreturned rate of UOCAVA ballots was 23.35%; slightly higher than the national average in 2020 and 2016. Additionally, four states did not report enough data to the EAVS to be counted in the EPI. As with standard mail ballots, more research is needed as to why and how requested UOCAVA ballots go unreturned.
Measuring Vote by Mail and UOCAVA in the Future
Voting by mail has become increasingly popular in the US, especially since 2020. It is a convenient way for voters to make their voices heard, even as election administrators have faced increased scrutiny on how they enact their states’ unique vote by mail policies. Each state enacts its own laws, deadlines, and policies for mail ballots, leading to variation at the local, county, state, and national level for election administrators and voters alike.
In the EPI, the rejected and unreturned rate for mail ballots and UOCAVA ballots are essential indicators—serving as point-in-time measures of the outcomes of nationwide VBM policies during a given year. While the EPI does not track state and local level changes in VBM policy, it does account for them when calculating and validating each indicator. When combined with research into state-level policies, the EPI can help paint a picture of the state of voting by mail and voting using a UOCAVA ballot in the United States.
