States Ranked by Election Administration Quality

The instrument: what the EPI actually measures

The Elections Performance Index was built by the MIT Election Data and Science Lab, with origins in a Pew Charitable Trusts project, primarily by and for election practitioners. It scores every state on 19 defined indicators: registration and absentee ballot rejection rates, data completeness, wait times, disability access, website usability, post-election audit adoption, turnout of registered voters, and more (MIT Election Lab, EPI project page). It is the closest thing American election administration has to an official scorecard, and it is deliberately nonpartisan in construction.

The design principle is worth stating plainly. The EPI does not ask whether a state's policies are wise. It asks whether the machinery worked: did ballots get counted, did lines stay short, did the data get reported, did rejected ballots stay rare. A state can have policies a given voter dislikes and still administer them competently, and the index is built to catch competence, not ideology. That is why this ranking leans on it. It is the rare measure in American elections that both parties helped build and neither party controls.

The trend it documents is improvement. In 2008, the national average EPI score was about 63 percent. In the 2024 edition, released April 16, 2026, it was 80 percent (MIT Elections Performance Index blog, 2026). Whatever else is contested about American elections, the machinery has gotten measurably better under governors and legislatures of both parties. A 17-point rise in a national administrative average over sixteen years is not noise; it is a documented institutional upgrade, and it happened while the surrounding political rhetoric grew louder, not quieter.

Why turnout is not in the score

The single most common mistake in reading election rankings is treating high turnout as proof of good administration. This ranking rejects that equation on purpose. Turnout measures voter behavior, driven by competitiveness, demographics, weather, and habit; administration measures whether the machinery counted the votes cast quickly, accurately, and transparently. The two can move in opposite directions, and one state settles the argument by itself.

New Mexico ranks first on this administration composite and 47th in the nation in 2024 turnout at roughly 55.5 percent (UF Election Lab compilation, 2025). A state does not administer elections badly because fewer of its residents chose to vote; it administers them well when the ballots that were cast were counted, audited, and reported cleanly, which New Mexico's 87 percent 2024 EPI score documents (New Mexico Secretary of State, 2026). The reverse holds too: a high-turnout state with sloppy data reporting or high mail-ballot rejection would score poorly here regardless of how many people showed up. Texas, 46th in turnout, is not in the bottom tier of administration measures, and Wisconsin, among the highest-turnout states in 2024, earns its rank here on same-day registration and statutory audits, not on the turnout number itself (Wisconsin Elections Commission, 2024).

Keeping turnout out of the score is what makes the ranking legible across party lines. Turnout correlates with partisan and demographic factors that have nothing to do with administrative competence, and folding it in would smuggle those factors into a measure that is supposed to be about machinery. The EPI's own designers separate the two, and this report follows them.

2024: strong administration under maximum scrutiny

The 2024 election was administered under more legal and political pressure than any in modern memory, and the measured result was strength. The MIT Election Lab's summary of the new EPI edition is titled plainly: U.S. Election Administration Remained Strong in 2024 (MIT Elections blog, 2026). More than 158 million ballots were counted at roughly 65 percent turnout of the citizen voting-age population, documented across thousands of local jurisdictions (EAC EAVS, 2025). The scale is worth pausing on. That is one of the largest administrative undertakings any democracy performs, executed by tens of thousands of mostly local officials, and it cleared without a documented systemic tabulation failure.

The companion Survey of the Performance of American Elections found average wait times fell for both Election Day and early voters. Only 11 percent of Election Day voters waited more than 30 minutes in 2024, down from 14 percent in 2020, and the improvement held even as many voters returned to in-person voting after the pandemic-era mail surge (MIT SPAE, How We Voted in 2024, 2025). Polling place disruptions were infrequent, and voter confidence was high across the board, with the long-standing partisan confidence gap narrowing considerably.

The federal data agrees. The U.S. Election Assistance Commission's 2024 Election Administration and Voting Survey, released June 30, 2025, documented the same operational picture across thousands of local jurisdictions (EAC, 2025). Three independent instruments, an academic index, a national voter survey, and a federal administrative census, reached one conclusion. When three tools built by different institutions using different methods converge, the finding is no longer a claim. It is a measurement.

The top tier: a photo finish among diverse states

New Mexico holds the number one spot on this composite for a reason no other state can match: it is the only state to have finished first in a fully published EPI table, in 2022 at 88 percent, and then repeated in the top ten of the 2024 edition at 87 percent (New Mexico Secretary of State, 2024 and 2026). Its score climbed from 82 percent in 2020 to 87 percent in 2024, a rise of 18 ranking positions in four years. That is the largest documented improvement among the leaders, and it happened in a state that ranks 47th in turnout, which is the cleanest possible proof that administration and participation are different things.

Michigan and Colorado follow, and neither placement is close to arbitrary. Michigan ranked second nationally at 88 percent in the 2022 EPI while standing up nine days of early voting and expanded mail rules that voters approved in 2022, then ran a 74.7 percent turnout presidential election in 2024 that was canvassed and certified without a documented tabulation failure (Michigan Secretary of State, 2024; Verified Voting). Colorado scored 87 percent and owns the audit distinction the rest of the country is now copying, which the next section takes up in full. Vermont and Nebraska round out the top five at 87 percent each, and the pairing is deliberate: a small Democratic-leaning state and a Republican-governed one, separated by rounding, judged by the identical ruler.

The honest caveat lives here. In the 2024 edition, ranks two through seven all round to 87 percent (New Mexico Secretary of State, 2026). At that resolution the ordering of the top tier is a matter of tenths of a point and multi-edition consistency, not a claim that one state decisively beat another in a single year. This ranking's composite reflects performance across editions and across the audit and convenience supplements, and it says so rather than pretending the EPI's top cluster is more separated than it is.

Audits are the quiet revolution

The strongest trend in election administration over the past decade is the spread of post-election tabulation audits. Colorado ran the first statewide risk-limiting audit in 2017 (Colorado Secretary of State). Rhode Island enacted an RLA statute the same year. Virginia now requires them on a recurring statutory schedule, Georgia requires statewide audits and executed a full hand tally of the 2020 presidential race, and states including Washington, Michigan, and New Mexico run RLA programs or pilots (Verified Voting audit-law database). A risk-limiting audit checks a statistical sample of paper ballots against machine counts, sized so that a wrong outcome would be detected with high probability. It is the difference between asserting a count is correct and demonstrating it.

This is a both-parties story, stated plainly. Republican-governed Georgia and Utah, Democratic-governed Colorado and New Mexico, and divided states like Virginia and Michigan all built audit regimes in the same decade. The SPAE found support for most election security practices rose from 2022 to 2024, particularly among Republicans (MIT SPAE, 2025). Verification is the rare election policy with a bipartisan adoption curve, and the curve is steepening: the practice that was novel enough to make national news when Colorado tried it in 2017 is now routine infrastructure in a growing bloc of states.

The audit dimension is also where this ranking's subscores diverge most sharply from raw EPI order. Colorado, Washington, Virginia, Georgia, and Rhode Island carry the highest Audit Strength meters on the list precisely because their verification regimes are statutory and mature, not aspirational. Georgia earns a high audit meter despite ranking fourteenth overall, because the hand tally it executed after 2020 was the largest post-election audit in American history and its RLA requirement is real (Georgia Secretary of State; Verified Voting). Administration is measured by what a state does, not by how quiet its elections happen to be.

Convenience infrastructure and the vote-by-mail states

A second cluster on this list earns its place through access architecture. Oregon has run universal vote-by-mail since 2000 and in 2016 became the first state to enact automatic voter registration, a pairing later copied in whole or part by additional states (Oregon Secretary of State). Washington runs entirely by mail with signature verification and ballot tracking for every voter, and it still posted 67.4 percent turnout in 2024 with no marquee statewide contest on the ballot (UF Election Lab compilation, 2025). Colorado's universal mail system rejected only about 1.0 percent of returned ballots in 2022, a low rejection rate that reflects a mature curing process rather than a permissive one (Ballotpedia, 2023).

The same-day registration states form their own high-turnout tier. Wisconsin, Minnesota, New Hampshire, and Maine all let eligible voters register and vote on the same day, and all four posted 2024 turnout well above the national rate, from Wisconsin and Minnesota near 76 percent down to Maine at 73.0 percent (Wisconsin Elections Commission, 2024; Minnesota Secretary of State, 2025; UF Election Lab compilation, 2025). Wisconsin did it through more than 1,800 municipal clerks, a radically decentralized system that nonetheless canvassed a sub-one-point presidential margin without a documented incident (official returns, 2024; Verified Voting). Decentralization is often assumed to be a weakness. Wisconsin's 2024 record is evidence it can be a strength when the local administration is competent.

Utah is the entry that most directly refutes the idea that mail voting is a partisan project. It moved to statewide universal mail-ballot delivery by 2020 under unified Republican governance, pairing it with signature verification and county audit requirements (Utah Lieutenant Governor's Office; Verified Voting). Alongside Nebraska, the highest-ranked deep-red state in the 2022 EPI at 87 percent, Utah is the standing proof that convenience infrastructure and conservative governance coexist without contradiction (MIT Election Lab).

What the ranking does not claim

Three disclosures keep this honest. First, the top of the 2024 EPI is a photo finish: New Mexico reports it ranked seventh at 87 percent with ranks two through six also rounding to 87 (New Mexico Secretary of State, 2026). Ordinal rankings at the top are separated by tenths of a point, and this report's composite reflects multi-edition performance, not one year's rounding. Second, administration is not participation: New Mexico ranks near the top here and 47th in 2024 turnout, while Texas, 46th in turnout, is not in the bottom tier of administration measures. Third, the wait-time indicator was not available for the 2024 EPI edition, so wait-time evidence here comes from the SPAE instead, and the report says so rather than papering over the gap (MIT Elections Performance Index, 2026; MIT SPAE, 2025).

A fourth disclosure belongs with them. The subscores in this report are analytical translations of documented facts into a common 0-to-20 scale, not figures MIT or the EAC publishes in that form. The EPI Score meter is the state's real EPI percentage; the other four meters convert documented facts, same-day registration, mail rejection rates, SPAE wait-time performance, and statutory audit regimes, into comparable components, with bad metrics inverted so a fuller meter always means better. Every underlying fact traces to a named source in the list below, but the packaging into meters is this report's, and the composite is ordinal, not official.

Only actions matter here: statutes passed, audits executed, data reported, lines measured. States earn their place on this list by administrative record, and every claim above carries the name of the institution that measured it. Where the evidence is thin, this report says so. Where two defensible orderings exist, it says that too. The goal is not to declare a winner by fiat; it is to show the receipts and let the reader see exactly how the ordering was built.