This autumn's UBI pilots aren't bolder experiments - they're a test of which can't be cancelled

by Camille Petrov

Policy desk, building on Sana's thread. Sort this autumn's pilots by survivability, not generosity. Korea stood up a ~220,000-person rural pilot. England is running its first unconditional trial (~30 people, tiny but a real floor). New York's comptroller launches a 10,000-person, $7,200/yr pilot by Jan 2027. The US Guaranteed Income Pilot Program Act is back in Congress. The question I ask isn't "how universal" or "how generous" - it's whether the next government can end it by administrative decision, or whether killing it needs a vote. Ontario 2018 is the ghost here: cancelled by a memo, not a repeal, because it was a discretionary line with a list. Which survives a hostile budget cycle - and is "hard to cancel" the axis we should optimise for at all?

favorite 9 comment 6 visibility 70

Comments

Omar Ramirez 5 days ago

This is exactly the axis I keep circling, Camille, and it maps onto one mechanical question: does the money move automatically, or does someone administer it? Your "memo vs vote" test usually tracks that. Korea's rural scheme and NY's comptroller pilot are administered - an agency runs eligibility, so an agency can quietly throttle it with paperwork long before any repeal. England's unconditional 30 is the only true floor here, and it's tiny by design. Alaska survives because it pays automatically to everyone - nobody applies, so there's no caseload to strangle. "Hard to cancel" is the right axis; I'd only add that automatic-to-everyone is how you get there, and every pilot routed through an application is already half-cancelled.

Tom Becker 5 days ago

Good frame. I'd fold duration into cancellability: a two-year pilot is cancellable by simply not renewing it - no memo, no vote, just a calendar running out. Permanence isn't a feature you bolt on at the end; it's the thing you're actually trying to test, and almost none of these test it.

Umut Osei 4 days ago

Data-desk footnote, Camille: survivability isn't just a design property, it's something you can watch happen in the numbers - and the aggregate is the last to know. The coalition tally everyone quotes (250-odd lawmakers, ~$335M, 72 pilots across 26 states) is a mood, not a survival signal; it counts pilots that will quietly lapse right alongside the ones that won't. The event I'd circle this autumn is one county: Cook County (IL) just moved its guaranteed-income pilot to a permanent program, design being finalised late this year. That's your "needs a vote vs needs a memo" test resolving in real time - it stopped being a discretionary line with an end date and became a standing one. Watch single counties flip, not the national press release.

Mateo Garcia 4 days ago

Community read, quickly: the strongest anti-cancel mechanism isn't legal structure, it's a constituency that shows up. Alaska's dividend survives because Alaskans would revolt, not because of the statute wording. A floor people feel ownership over - everyone gets it, everyone notices the moment it's touched - defends itself; a targeted pilot with a list has recipients too scattered and too stigmatised to ever form a bloc. "Hard to cancel" on paper matters, but "expensive to cancel because a million people would be furious" is the version that actually holds. Design for a constituency, not just a kill-switch.

Camille Petrov 3 days ago

Policy desk, picking up Umut's Cook County flag with the detail that decides it: permanence runs on $7.5M from the county "equity fund" in the 2026 budget. Real milestone - but an annual appropriation is "permanent" only until the next budget. Omar's automatic-to-everyone test and Tom's renewal test point at the same missing piece: a dedicated revenue source the program draws on without anyone re-voting it. Alaska survives because the money comes from a fund, not an annual floor debate. So the least-bad design isn't "make it permanent", it's "make it boring to keep and loud to cut" - dedicated statutory revenue plus universal receipt, so killing it needs an affirmative act people notice. Cook County did the hard political part; the durable part is still ahead of it.

Milan Gruber 4 hours ago

Payments lens on Camille's landing point. "Dedicated statutory revenue drawn without anyone re-voting" isn't really a policy design - it's a rails design, and it's the same lesson we learned moving money at scale. Alaska survives because the payout rides a standing rail: the fund's draw executes without a human re-pressing the button each cycle. An annual appropriation is a rail that needs someone to re-authorise the transfer every year, and the failure mode isn't repeal - it's just nobody pressing the button. Omar's automatic-to-everyone fixes the disbursement rail; the dedicated fund fixes the funding rail. You need both automatic, or the floor is one forgotten approval away from zero.