If a single barcode fails to scan on a single ballot inside a bulk mailing of 10,000, the new system rejects the entire batch. That is not a hypothetical edge case -- it is the working design of a ballot-verification system the U.S. Postal Service has been racing to build since mid-June 2026, under a March 2026 executive order from President Trump directing USPS to restrict mail voting by requiring states to submit lists of eligible voters so ballots can be checked against them before delivery.[1] The system is real, it is moving toward the 2026 midterms while still tied up in litigation over the executive order itself, and a federal whistleblower who described the build to Senator Richard Blumenthal called it "untested and error-prone" -- the whole process, in the whistleblower's own words, "secretive, rushed, chaotic, and fundamentally flawed."[2]
USPS gave itself roughly eleven weeks -- mid-June to a September 1 deadline -- to build a system that, by the estimate of people who build this kind of infrastructure for a living, normally takes nine months minimum to do right.[2] Internally, according to a source familiar with USPS meetings, officials raised exactly the concern the design invites: that the rules could delay or block ballots for "large swaths of eligible voters," and that rejecting an entire batch over one bad scan was a real, known risk rather than a rare failure mode -- some officials reportedly pushed to stop rejecting whole batches for that reason and were overruled.[2] Election officials interviewed separately said the compliance work itself could take years, not months, for some counties to complete.[3] Maricopa County, Arizona sends more than two million ballots in its first mailing alone -- under this design, one faulty barcode anywhere in that batch is enough to hold up all two million.[2]
Separately from the ballot-screening system, USPS's Regional Transportation Optimization initiative is reducing delivery trips within certain zip codes -- a cost and logistics decision, not an election-specific one, but one that lands on election mail all the same. In February 2026, Representatives Robert Garcia and Joe Morelle, joined by 68 other members of Congress, sent a formal letter demanding the Postmaster General explain how RTO's reduced trip frequency would affect the delivery of mail-in ballots, naming rural and hard-to-reach communities specifically as the places most exposed to the slowdown.[4] That is a structurally different kind of mechanism than the ballot-screening system -- an internal efficiency measure rather than a directive aimed at mail voting -- but it reduces the same thing: how reliably a ballot moves through the mail on time, for the same voters least able to absorb a delay.
Neither of the two mechanisms above touches in-person voting at all, and the polling-place trend has nothing to do with USPS. It is its own decade-long pattern: the United States had an estimated 116,000 Election Day polling places in 2012 and fewer than 89,000 by 2022 -- 27,000 gone, roughly one in five, spread across thousands of individual county-level decisions rather than one policy.[5] It is still happening in real time heading into 2026: Tarrant County, Texas -- the state's third most populous -- voted this September to cut its polling sites from 316 in 2022 to 224, a reduction of 92, after hours of public objection at the meeting where the vote was taken.[6]
The closures do not land evenly. Halifax County, North Carolina has cut its polling places 24% since 2012. In 2022, the 189 residents of Kaktovik, Alaska could not vote at all because the state failed to staff their polling site. Members of the Fort Peck Tribes in Montana were denied a satellite office and had to travel up to 60 miles to reach a working one.[7] The mechanism behind the outcome is measured, not just observed: each additional mile between a voter and their polling place reduces turnout by up to 0.99 percentage points, and voters moved to an unfamiliar polling location vote at a rate 5.65 points lower than voters who weren't moved.[7]
Why does this matter? These are three genuinely different mechanisms, and holding them apart matters more than collapsing them into one story. The ballot-screening system traces to one executive order, signed by one president, currently being fought in court -- a deliberate policy choice with a known author. The delivery-trip reduction is an internal USPS cost decision that happens to intersect with election mail, not a measure built for that purpose. The polling-place closures are thousands of separate local decisions made over more than a decade, mostly for budget reasons that have nothing to do with each other. No single actor designed all three to work together. But a rural voter, a tribal voter, or a voter in a county that just cut its polling sites doesn't experience three separate origin stories -- they experience one mail ballot that might get held up by a stranger's barcode, moving through a postal route with fewer trips than it had last year, as a backup to a polling place that may no longer be there. Different causes, the same address. That is worth naming plainly rather than treating as three unrelated headlines.
Troy Pearsall, who wrote this piece, has voted in every election since he turned 18 -- through the years he worked in and around the intelligence community and could not speak publicly about the work itself. He worked under both Republican and Democratic administrations over that career, did not necessarily agree with either party's politics, and voted anyway, every time -- the belief was in the vote itself, held separately from any opinion about who won it. That distinction is the actual baseline this piece is measuring against: not an abstract civic ideal, and not a partisan complaint about one administration's executive order, but a floor -- belief in the mechanism, independent of who is in office -- that one bad barcode, one cut delivery route, or one closed polling place can still knock a real voter under, regardless of how consistently they've shown up before.
Since this piece published, a related mechanism landed: a separate piece on the SAVE Act's documentary-proof-of-citizenship requirement, which would add its own real friction -- weeks of state processing time, a per-copy fee, a passport that costs $165 -- to registering in the first place, on top of the three mechanisms already documented here. Different agency, different stated justification, same underlying pattern: real, measurable cost added to a constitutional right, sized well past how often the problem it targets is actually shown to occur.