Oregon vote by mail: the first all-mail state and what it measured

Oregon vote by mail is the system under which every registered Oregon voter receives a paper ballot by mail and returns it by mail or at a drop site, with no traditional polling places. Oregon voters adopted it for primary and general elections through Ballot Measure 60 in November 1998, and the state has run all of its elections this way since 1999.⁠^1,2 The record shows a durable, well-documented system with low measured fraud and moderate cost savings, and a turnout effect that peer-reviewed studies find to be much smaller than early claims.

What happened

Oregon vote by mail grew in steps over almost twenty years before it became universal. According to the Oregon Secretary of State’s own timeline, the Legislature approved a test of vote by mail for local elections in 1981 and made it permanent in 1987, after which a majority of counties used it for local and special elections.⁠² The Legislative Fiscal Office cites the 1981 authorizing law as Oregon Laws 1981, Chapter 805.⁠¹

Statewide use began with special elections. The Secretary of State records the first statewide special election by mail in June 1993, with 39% turnout, and a second in May 1995, with 44% turnout.⁠² In 1995 the Legislature approved expanding vote by mail to primary and general elections, and the governor vetoed the bill.⁠²

The best-known early test came with a federal vacancy. In December 1995 Oregon held a primary entirely by mail to nominate candidates for a vacant U.S. Senate seat (58% turnout), and in January 1996 it held the general election for that seat entirely by mail, electing Ron Wyden to replace Bob Packwood, with 66% turnout.⁠² The Legislative Fiscal Office describes the 1996 contest as the first election for federal office conducted entirely by mail in any state.⁠¹

The remaining step came through the initiative process rather than the Legislature. A 1997 bill to expand vote by mail passed the Oregon House and died in a Senate committee.⁠² In the May 1998 primary, still a polling-place election, the Secretary of State records that 41% of registered voters were permanent absentee voters, that absentee ballots made up nearly two thirds of ballots cast, and that turnout hit a record low of 35%.⁠² Supporters then qualified an initiative for the November ballot, and the Secretary of State notes that no paid signature gatherers were used.⁠² On November 3, 1998, voters approved Measure 60 by 757,204 to 334,021.⁠² The Legislative Fiscal Office records that the measure was approved November 3, 1998 and took effect December 3, 1998, and that Oregon has conducted all elections by mail since 1999.⁠¹ In November 2000 Oregon became, in the Fiscal Office’s words, the first state to award its presidential electoral votes from an election conducted entirely by mail.⁠¹

Later changes reshaped the system around the mailed ballot. On January 1, 2016, Oregon became the first state to register eligible citizens automatically when they interact with the Driver and Motor Vehicle Services Division, a law known as Oregon Motor Voter.⁠¹ Beginning January 1, 2020, the state paid return postage on every mailed ballot under SB 861 (2019).⁠¹ And under a 2021 law, HB 3291, which the Secretary of State says took effect for the 2022 elections, mailed ballots postmarked by 8 p.m. on Election Day and received within seven days are counted.⁠³

How it worked

The mechanics of Oregon vote by mail are set out in statute and administrative rule, and the Legislative Fiscal Office’s 2020 review describes them in plain terms.⁠¹

Who gets a ballot. Counties must mail ballot packets to all active registered voters between 14 and 20 days before each election.⁠¹ The packet holds the ballot, a secrecy envelope and a pre-addressed return identification envelope.⁠¹ The ballot itself carries no information identifying the voter.⁠¹ The Fiscal Office describes the secrecy envelope as optional.⁠¹

How a ballot is returned. The voter signs a declaration on the outside of the return envelope, which is required for the ballot to count, and returns it by mail or at any authorized drop site in the state.⁠¹ A ballot left at a drop site in the wrong county is marked as received on time and forwarded.⁠¹ Oregon has historically had a hard receipt deadline of 8 p.m. on Election Day;⁠¹ the postmark rule described above now adds a seven-day grace period for mailed ballots.⁠³ In June 2026 the Secretary of State said a U.S. Supreme Court ruling in Watson v. Republican National Committee, decided 5 to 4, preserved states’ authority to count ballots cast by Election Day and received within a post-Election Day grace period.⁠³

How identity is checked. The core identity check is a signature match. County staff compare the signature on the return envelope to the signature on the voter’s registration record, and only ballots with verified signatures are counted.⁠¹ The Fiscal Office reports that staff are trained in handwriting analysis.⁠¹ If an envelope is unsigned, the voter is mailed a statement to sign and return within 14 days; if the signature does not match, the voter is mailed a new registration card to supply an updated signature within 14 days, and voters who do not respond are made inactive.⁠¹

In-person and accessible options. Oregon is not purely remote. Counties must keep voting booths available, and every county elections office has a tablet and printer so voters with disabilities can mark an accessible ballot; counties also send assistance teams to voters’ locations.⁠¹

Counting and audit. Counties count in secure rooms with video surveillance and space for observers, read ballots with optical scanners that store an image of each ballot, test equipment before and after elections, and must audit results in every election, either through a risk-limiting audit before certification or a post-election audit after it.⁠¹ Ballots and signed envelopes are retained for set periods and, according to the Fiscal Office, are available for public inspection during that time.⁠¹ For the general idea of paper-based audits, see audits and verification.

Registration and the voter file. Oregon developed a centralized voter registration database, OCVR, in 2005 to comply with the Help America Vote Act, and that system supports signature verification and ballot preparation.⁠¹ Because every active voter is mailed a ballot automatically, the accuracy of the voter file directly determines who receives a ballot. That link matters for the Motor Voter error described under failures below.

Measured results

Oregon vote by mail has been measured on turnout, cost, fraud and partisanship, and the findings differ sharply in strength.

Turnout

The Secretary of State’s turnout history, generated September 12, 2025, reports general election turnout as a share of registered voters: 82.36% in 1992, 71.31% in 1996, and 59.02% in 1998 under polling places; then 79.80% in 2000, 86.48% in 2004, 85.67% in 2008, 80.33% in 2016, 78.50% in 2020, 66.90% in 2022 and 75.40% in 2024 (2,308,256 votes of 3,060,374 registered).⁠⁴ These are percentages of registered voters, not of eligible citizens, and the registered denominator changed when Motor Voter added people automatically in 2016.⁠¹

The Legislative Fiscal Office compared polling-place elections from 1992 to 1998 with mail elections from 2000 to 2018. Combined primary and general turnout averaged 55.5% before and 60.5% after; general elections 70.4% and 76.1%; primaries 39.6% and 43.9%.⁠¹ The Fiscal Office concluded that vote by mail “appears to have marginally increased voter turnout 2-5%, depending on the type of election”.⁠¹ It also warned that Motor Voter likely affects turnout percentages and that the state lacked data to separate the effects.⁠¹

The peer-reviewed literature is more cautious. The most cited early estimate, by Southwell and Burchett, was a 10 percentage point turnout increase. Gronke and Miller, in American Politics Research (2012), report that they were “unable to reproduce earlier findings, either via replication or extending the time series to include 2010 electoral data,” found evidence of a novelty effect, and found “a consistent impact of voting by mail on turnout only in special elections”.⁠⁵ A broader causal study by Thompson, Wu, Yoder and Hall in PNAS (2020) used the staggered county rollout of universal vote by mail in California, Utah and Washington from 1996 to 2018 and found an increase of roughly 2% in turnout as a share of the citizen voting-age population.⁠⁶ That study did not include Oregon, but it is the strongest design-based estimate of the same policy.

Who votes

Berinsky, Burns and Traugott, in Public Opinion Quarterly (2001), studied Oregon individually and concluded that vote by mail “increases turnout, but it does so without making the electorate more descriptively representative,” that it “advantages the resource-rich by keeping them in the electorate,” and that it “does little to change the behavior of the resource-poor”.⁠⁷ They found no influence on the partisan composition of the electorate.⁠⁷ Thompson and colleagues likewise found that universal vote by mail “does not appear to affect either party’s share of turnout” or either party’s vote share, at least in normal times.⁠⁶

Cost

The Legislative Fiscal Office reports that the May 1998 polling-place primary cost Oregon counties $3.4 million and the May 2000 mail primary $2.8 million, a saving of nearly $600,000 (17%), with cost per ballot falling from $5.10 to $3.03.⁠¹ Comparing the 1998 and 2018 general elections after inflation adjustment, it estimates a saving of about 18%.⁠¹ The same report cautions that counties do not report costs consistently and that “a fair comparison of the direct costs of vote by mail elections and polling place elections is not possible at present,” and that the figures exclude the state-paid return postage that began in 2020, estimated at $1.6 million to $3.1 million for the 2019 to 2021 biennium.⁠¹

Fraud

The Fiscal Office reports that Oregonians cast about 60.9 million ballots in statewide and special elections from 2000 to 2019, and that the Oregon Department of Justice estimates 38 criminal convictions relating to voter fraud over that period, a rate it calculates as 0.00006%.⁠¹ The most common forms, according to the Department, were duplicate voting across states and ballot forgery, such as one household member signing another’s ballot.⁠¹ A conviction count measures detected and prosecuted cases, not all attempts; see the competing views below.

Failures and criticisms

Oregon vote by mail has critics on several distinct grounds, and the strongest versions deserve a fair statement.

Coercion and the secret ballot. The 2005 Commission on Federal Election Reform, chaired by former President Jimmy Carter and former Secretary of State James Baker, wrote that “absentee ballots remain the largest source of potential voter fraud” and that citizens who vote at home “are more susceptible to pressure, overt and subtle, or to intimidation”.⁠⁸ The same commission said Oregon “appears to have avoided significant fraud” through safeguards including signature verification, but that “even in Oregon, better precautions are needed to ensure that the return of ballots is not intercepted”.⁠⁸ The underlying concern is structural: a polling booth guarantees that no one watches the voter mark the ballot, and a kitchen table does not. Oregon’s own fraud data identify one household member signing another’s ballot as a recurring offense.⁠¹ Conviction counts cannot measure pressure that is never reported. For a general treatment, see intimidation.

The turnout promise was oversold. Supporters argued that universal mail voting would raise turnout substantially.⁠¹ The replication by Gronke and Miller and the multistate causal study by Thompson and colleagues both point to small effects, concentrated in low-salience elections in Oregon’s case.⁠^5,6 The Carter-Baker commission put it bluntly in 2005: vote by mail “appears to increase turnout for local elections,” but there was “no evidence that it significantly expands participation in federal elections”.⁠⁸

It may widen, not narrow, participation gaps. If convenience helps people who already vote keep voting, as Berinsky, Burns and Traugott found,⁠⁷ a reform sold as inclusive can leave the electorate as unequal as before.

Signature matching rejects lawful voters, unevenly. Signature comparison is a human judgment. The Fiscal Office notes that DMV-captured signatures are sometimes unclear, which can produce mismatches when a voter’s signature has changed.⁠¹ Oregon has not published, in sources verified for this page, an analysis of rejection rates by age or race. Washington, which copied Oregon’s model, has: the Washington State Auditor’s February 2022 audit found a statistically significant difference in rejection rates between white voters and voters of other racial and ethnic groups (Native American and multiracial voters excepted), and higher rejection odds for younger and male voters, while also finding no evidence of bias in election workers’ decisions.⁠⁹

The voter file becomes the ballot list. In an all-mail system, a registration error is not a latent problem; it sends a ballot. In September 2024 Oregon’s DMV found 1,259 people who had been registered through Motor Voter without the required proof of citizenship, and on October 7, 2024 it reported 302 more; the DMV attributed the problem to a clerical error, and Governor Tina Kotek called for an independent audit.⁠¹⁰ The reporting attributes the fault to automatic registration, not to mailed ballots, but the case shows how tightly the two are linked.⁠¹⁰

Deadlines and finality. The seven-day postmark window added in 2022 means more ballots arrive after Election Day. Critics of grace periods argue that a hard deadline gives a cleaner, faster result; supporters argue that a voter should not lose a ballot to mail delays. The U.S. Supreme Court’s 2026 ruling, as described by the Secretary of State, settled that states may use such a window, not whether they should.⁠³

Supporters’ rebuttals. Defenders answer that polling places also carry risks: the Fiscal Office argues that vote by mail “eliminates the prospect of voter intimidation at physical polling places,” that it leaves a paper trail for audits, and that fraud is “no more widespread in vote by mail elections than it is polling place elections”.⁠¹ Each side is describing a different threat. Neither has data that settle the question of unreported household pressure.

What transferred and what did not

Oregon’s model spread, slowly and selectively. As of September 2020, the Legislative Fiscal Office counted five states conducting elections entirely by mail: Colorado, Hawaii, Oregon, Utah and Washington.⁠¹ The National Conference of State Legislatures, in a table updated August 7, 2026, lists states that conduct all elections by mail with their enactment years: Oregon 1998, Washington 2011, Utah 2012 (with counties permitted to choose), Colorado 2013, Hawaii 2019, California 2021, Nevada 2021, Vermont 2021 (general elections only) and the District of Columbia 2022.⁠¹¹

What transferred was the mailed ballot, signature verification, drop boxes and ballot tracking. What often did not transfer was Oregon’s long incremental runway: two decades of local and special elections by mail before universal adoption.⁠² Several later adopters moved faster, some during the 2020 pandemic. Other features, such as Oregon’s in-person booths at county offices, vary by state.⁠^1,11

The Carter-Baker commission also warned that what works in one state may not transfer to another, writing that vote by mail “is, however, likely to increase the risks of fraud and of contested elections in other states” where safeguards are weaker, and citing problems with ballot tracking in Washington’s King County in 2004.⁠⁸

Lessons for My Vote Forward

The following points are interpretation and design proposals, not findings about My Vote Forward, which has not been tested.

• Convenience is not representation. Oregon’s best evidence shows that making participation easier mostly retains existing participants.⁠⁷ A framework that wants input from people who do not usually take part cannot assume that a convenient channel will reach them. Report who took part against who could have; see measurement.
• Measure turnout claims against a counterfactual. The gap between the early 10-point estimate and the later replication,⁠⁵ is a warning about evaluating any participation tool by before-and-after comparisons alone; see evaluation.
• Remote participation changes the coercion problem. Oregon accepted some household risk in exchange for convenience and paper auditability. My Vote Forward separates public priorities and voluntary opinions from private voting precisely because remote expression and secret voting carry different risks; see private voting and intimidation.
• Identity checks have error rates, and errors fall unevenly. Signature matching is a reminder that any eligibility check rejects some legitimate participants. The Washington audit⁠⁹ shows why rejection rates should be published by group where privacy allows; see identity and registration.
• Upstream data errors flow downstream. The 2024 Motor Voter case⁠¹⁰ shows that the list of who can participate is itself a security boundary.
• Publish the method, the denominator and the audit. Oregon’s turnout tables give numerators and denominators⁠⁴ and its audit rules are public.⁠¹ That is the standard for any published release.

Unresolved questions

• How much undetected household pressure occurs in all-mail systems? Conviction counts cannot answer this, and no source verified for this page measures it directly.
• What are Oregon’s signature-rejection rates by age, race and language, and how do cure rates differ? A Washington-style audit for Oregon was not found in the sources consulted.
• How much of Oregon’s post-2016 turnout change reflects Motor Voter rather than mail voting? The Legislative Fiscal Office said in 2020 that the state lacked data to separate them.⁠¹
• What did the independent audit of Motor Voter called for in 2024 conclude? Its findings were not verified for this page.
• Does the seven-day postmark window change how quickly results are accepted as final, and with what effect on public trust? This has not been measured in the sources consulted.