# Mail-in Voting Casting a ballot delivered and returned outside a polling place. The vault's [[Voter Suppression]] cluster covers restrictions on *who may vote*; this page covers a **method**, and the reason it needs its own page is that the method itself became the contested object — first of a disinformation campaign, then of federal regulation. > [!note] "Absentee" and "mail-in" are the same thing > The distinction Trump drew in 2020 — absentee good, mail-in bad — has no administrative basis. Verification is identical. The USPS's own 2026 rule states that "the terms 'mail-in ballot' and 'absentee ballot' are synonymous and interchangeable." Florida, where Trump votes by mail, is a no-excuse state. ## How much of the electorate | Year | Share of voters casting mail ballots | |---|---| | 2016 | 21% | | 2018 | 23.1% | | 2020 | **43%** | | 2022 | 31.8% | | 2024 | **29.0%** | Census Bureau, Current Population Survey Voting Supplement — voter self-report. The EAC's administrative count for 2024 is about **30%** of all ballots cast (roughly **48 million** mail ballots), which is reassuringly close. **Use "roughly three in ten," not "about a third."** One-third was accurate for 2022 and overstates 2024. The honest framing: about 29–30% today, up from about one in five before the pandemic, having peaked at 43% in 2020. Within no-excuse states, 22.1% of voters used mail in 2024, against 5.2% in states requiring an excuse. **The rules drive the usage**, which is what makes rule changes consequential. ## The legal landscape - **29 states** have no-excuse absentee voting; **8 states and D.C.** conduct all-mail elections; **13 states** still require an excuse. - All-mail states: Oregon (2000), Washington (2012), Colorado (2014), Utah (2019), Hawaii (2020), California, Nevada and Vermont (2022), D.C. (2023). Nebraska and North Dakota allow county opt-in. - **Oregon was first**, by citizen initiative in November 1998, implemented in 2000. California is usually credited with the first no-excuse absentee law in 1978, though NCSL's statute table records earlier enactments in Idaho (1970) and Washington (1974). - "All-mail" is really *ballots distributed by mail* — in 2024, **21% of mail ballots were returned to a drop box or election office** rather than mailed back. Origins are military: Union and Confederate soldiers voted from their units in the Civil War, casting roughly 230,000 ballots in 1864 — about 7.5% of the vote in states that allowed it. Congress passed soldier-voting laws in 1942 and 1944; the durable frameworks are UOCAVA (1986) and the MOVE Act (2009). ## The partisan reversal **This is the fact most often gotten wrong in both directions.** Nationally, mail voting was close to partisan-neutral before 2020. MIT's survey series found Democrats slightly likelier to use it between 2008 and 2016, but that difference "was primarily an artifact of which states had chosen to conduct their elections entirely by mail" — geography, not voter type. The causal literature agrees: universal vote-by-mail has **no measurable effect on partisan turnout or vote share**. At the *state* level, though, the "Republicans built this" story is true and well documented. Arizona's no-excuse mail voting came from a near-unanimous **Republican-sponsored bill in 1991**, with the Permanent Early Voting List following in 2007. In 2020, over 80% of the million-plus Arizona Republicans who voted were on that early-voting list. A former Republican Maricopa County Recorder: mail voting "was always more popular with Republicans… The party really did push it." Then it inverted, and then it partly converged: | Year | D voting by mail | R voting by mail | Gap | |---|---|---|---| | 2020 | 60% | 32% | 28 pts | | 2022 | 46% | 27% | 19 pts | | 2024 | 37% | 24% | 13 pts | **The gap closed from the Democratic side.** Republican usage went 32 → 27 → 24; the convergence is Democrats pulling back, not Republicans returning. Republican *support* for universal mail voting ran 14% (2016), 10% (2020), 15% (2022), 15% (2024) — it never recovered. MIT's finding is that the divide "developed first at the elite level, and then at the grassroots," which is the causal direction worth carrying: this was led, not demanded. ## The 2020 campaign against it Trump's attacks were specific and dated. April 4, 2020: "I think a lot of people cheat with mail-in voting" — said shortly after requesting his own mail ballot for the Florida primary. May 2020: the false claim that Michigan "illegally" sent ballots to 7.7 million people (they were *applications*). June 22: "MILLIONS OF MAIL-IN BALLOTS WILL BE PRINTED BY FOREIGN COUNTRIES." July 30: universal mail voting would make 2020 "the most INACCURATE & FRAUDULENT Election in history," with a suggestion to "Delay the Election." The 2020 cycle produced **more than 300 election lawsuits by late October and over 340 through Election Day** — most of them about mail voting rules, with restriction suits brought primarily by the Trump campaign and Republican operatives. Post-election, Trump and allies lost roughly **59 of 60** challenges, before judges appointed by presidents of both parties. **What the evidence actually shows.** The Heritage Foundation's own election fraud database — the most-cited conservative source — contained **1,412 proven instances as of January 2023**, spanning every kind of election fraud since 1982, against billions of ballots. Heritage describes it as "a sampling," not a census. The News21 database found **491 absentee-fraud cases from 2000 to 2012**. Brennan Center research puts fraud incident rates between **0.0003% and 0.0025%**. Oregon has counted over 100 million mail ballots since 2000 with about a dozen proven fraud cases. The American Statistical Association finds "no reason to believe that fraud is more common in VBM states." ## The real problem: rejection, not fraud Mail ballots do get thrown out, at rates that dwarf documented fraud. | Year | National rejection rate | |---|---| | 2016 | 1.0% | | 2018 | 1.4% | | 2020 | 0.8% | | 2022 | 1.5% | | 2024 | 1.2% | In 2024 that was **584,463 rejected ballots**. The leading cause was **non-matching signature — 40.7% of all rejections**. Rates rose in 39 states and D.C. between 2020 and 2024, and vary enormously by state: Arkansas 8.8%, Indiana 0.2%. **Rejection falls unevenly.** Florida studies covering ~9.6 million (2016) and ~8.2 million (2018) ballots found disproportionate rejection among younger voters, Hispanic voters, unaffiliated voters and military dependents. In 2016, voters aged 18–21 cast 9% of Florida mail ballots and accounted for **31% of rejections**. Later work found racial disparities that persist after controlling for individual voting experience — i.e. not fully explained by inexperience. As the Brennan Center puts it, these "are problems of access and administration, not fraud." About three-quarters of states let voters cure envelope errors. ## The counterexample that is real **North Carolina's 9th district, 2018** — and it cuts against the narrative it is usually recruited for. Republican consultant Leslie McCrae Dowless, working for candidate Mark Harris, ran a paid absentee-ballot operation in Bladen and Robeson counties: collecting unsealed ballots door to door, marking blanks, completing partials, forging signatures and witness attestations. It was caught by **public absentee data** — 61% of county mail ballots went to the Republican while only 16% of mail voters were registered Republicans. The state board refused to certify, the seat sat vacant into the 116th Congress, and on **February 21, 2019** the board unanimously ordered a new election — the first congressional race overturned in about forty years. Dowless was indicted; seven associates pleaded guilty; Harris withdrew, and Dan Bishop (R) won the September 2019 do-over. The board found Harris had no knowledge of the scheme. Dowless died in 2022 before trial. Brookings drew the distinction that matters: this was **election fraud, not voter fraud** — party-aligned operatives interfering with voters, not voters casting ballots they weren't entitled to. It was committed for a Republican, exposed by transparent election data, and remedied. Brookings also predicted in real time that it would be repurposed as justification for restricting mail voting. ## 2026: from persuasion to machinery The argument stopped being rhetorical. [[Executive Order 14399]] of March 31, 2026 directs the [[USPS]] to impose federal design and data requirements on ballot mail and to refuse non-compliant mailings; the implementing rule created the [[USPS Federal Ballot Mail Portal]], which state officials must clear before the Postal Service will carry their ballots. That program — and the litigation and whistleblower disclosure around it — is tracked on those two pages, not here. On **September 14, 2026** the Supreme Court refused to let the rule operate in this election; the rule is blocked, not voided. A second federal operation runs alongside it from a different agency: the [[Unlawful Voter Initiative]], DHS's August 2026 diversion of USCIS fraud investigators onto state voter rolls, also the subject of a whistleblower disclosure. The shift is the point. From 2020 to 2024 the campaign against mail voting ran through claims, lawsuits and state legislation, and mostly lost. In 2026 it runs through **a federal agency's acceptance criteria**, where the question is no longer whether a voter is eligible but whether the state filled in the form correctly. ## See Also - [[Executive Order 14399]] · [[USPS Federal Ballot Mail Portal]] — the 2026 federal machinery - [[Unlawful Voter Initiative]] — the DHS half of that machinery - [[USPS]] — the carrier turned gatekeeper - [[Voter Suppression]] — the front-end restrictions; this page is the method - [[The Big Lie]] · [[Stop the Steal]] · [[2020 Presidential Election]] — the campaign against it - [[Election Assistance Commission]] — the federal source for the administrative numbers - [[2026 Midterm Elections]] — the election now at stake - [[Help America Vote Act]] · [[National Voter Registration Act]] — the statutory backdrop - [[Provisional Ballot]] — the in-person analogue