Debates and responses

Editorial analysis

This largely stands and we are not going to dodge it. A full, universal, liveable UBI is not affordable under current tax systems — anyone claiming otherwise should open our fiscal sandbox and drag the sliders themselves.

Two things can be said. First, gross cost is not net cost: if UBI is clawed back from higher earners through progressive taxation and consolidates part of existing transfers (China spends roughly ¥4.2tn on social security and employment; US Social Security is about $1.6tn), the net fiscal impact is far below the gross figure. Second, "cannot afford the full version" is not "cannot afford any version". The real question is not whether to pay everyone $1,000, but at what level and what coverage the benefits start to exceed the costs. That question has an answer, and it is not zero.

Sources: Congressional Budget Office, Monthly Budget Review: Summary for Fiscal Year 2025 (2025) · 中华人民共和国国家统计局, 中华人民共和国2024年国民经济和社会发展统计公报 (2025)

Editorial analysis

This fully stands, and we think it is the most common error in the entire UBI discourse — one that needs correcting publicly. The compute/AI tax lever in our sandbox is deliberately capped low so you can watch it fail to close the gap.

If UBI is to be funded, it must come from mainstream tax bases: VAT, income tax, capital gains, carbon, land, or sovereign fund returns. Pinning it on a narrow levy against one young industry is both insufficient and politically fragile against lobbying. UBI supporters should stop saying it.

Sources: Tax Foundation, New Congressional Data Center Tax Proposals Threaten US AI Investment (2026) · ITIF, Taxing AI Compute Would Be a Mistake (2026)

Editorial analysis

This partly stands. Labour supply does fall; we do not dispute the measurement, and the OpenResearch data sits on our evidence wall.

But two counter-cases must be shown alongside it, and their designs are closer to real UBI. The Alaska Permanent Fund Dividend is genuinely universal and genuinely permanent, and Jones and Marinescu found no effect on aggregate employment — only a 1.8-point rise in part-time work — because the cash created local demand that offset individual labour-supply reductions. Iran's national transfer to 70 million people likewise showed no negative effect on hours or participation, with women's participation rising.

Here is the real crux: small, time-limited experiments measure "one person gets money", while a universal permanent scheme changes market equilibrium. The former understates demand effects and distorts behaviour because the money visibly ends in two or three years. This is not an excuse for UBI — it is a claim that the two bodies of evidence answer different questions.

One more layer: in Stockton full-time employment rose from 28% to 40% against a control moving 32% to 37%, because the money paid for childcare and transport. So "cash makes people work less" and "cash lets people go to work" are both true, in different populations.

Sources: NBER, The Employment Effects of a Guaranteed Income (2024) · American Economic Journal: Economic Policy, The Labor Market Impacts of Universal and Permanent Cash Transfers (2022) · Journal of Development Economics, Cash transfers and labor supply: Evidence from a large-scale program in Iran (2018)

Editorial analysis

Partly stands — and on housing specifically, we think it largely stands.

If UBI were financed by printing money or pure deficit, inflation would follow. But a tax-financed UBI is redistribution, not monetary expansion: it moves money from high-savers to high-spenders, raising aggregate demand without creating it from nothing. Iran is instructive — universal cash was paid precisely while energy subsidies were removed and prices rose sharply, yet the poor were more than compensated and poverty and inequality fell in subsequent years.

Housing is the genuine weak point. In cities with constrained land supply, a large share of any universal income supplement is absorbed by rent — a repeated finding in the housing-subsidy literature. The honest conclusion: without supply-side reform in land, housing and healthcare, a substantial portion of UBI is handed to asset owners. That is not a reason to skip UBI. It is a reason not to do UBI alone.

Sources: Economic Research Forum, From energy subsidies to universal basic income: lessons from Iran (2017)

Editorial analysis

The first half stands. The second does not.

We agree — and say so plainly on the pressure index page — that there is no evidence yet of economy-wide AI job loss. Claiming mass unemployment has arrived is wrong, and we do not claim it.

But "the aggregate is fine" hides what has already happened in the structure. Stanford Digital Economy Lab tracking shows employment of workers aged 22–25 in the most AI-exposed occupations now sits 19% below where it would be had it tracked their less-exposed peers, with no comparable gap for experienced workers. That gap has widened steadily since it was first documented in August 2025. The mechanism is not layoffs but reduced hiring of new entrants — which is exactly why it is invisible in aggregate data, and exactly why it is dangerous: the talent pipeline breaks years before the unemployment rate notices.

The historical analogy deserves to be taken seriously rather than traded as a slogan. Previous revolutions did create new jobs, but transitions ran for decades and the displaced generation bore the entire cost. The case for UBI does not require "this time is different". It requires only "last time the transition cost was real, and this time we could choose not to put it on the same people".

Sources: Stanford Digital Economy Lab, No Widespread Displacement, but the AI Employment Gap for Young Workers Has Widened to 19% (2026) · Stanford Digital Economy Lab, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (2025)

Editorial analysis

There is measured evidence against it, though the rebuttal is not complete.

Three arguments for universality. First, means-testing has costs, and its failures are asymmetric: missing a genuinely poor family is far worse than paying an extra rich one, and targeted welfare worldwide leaves large numbers of eligible people unenrolled. Second, universality removes stigma — a significant mechanism behind the mental-health gains in the Finnish trial. Third, and most important, targeted welfare creates the poverty trap: benefits withdraw as income rises, producing effective marginal tax rates above 60% that directly punish work. A UBI that does not phase out has no such problem — which is why "UBI reduces work" and "means-tested welfare reduces work" must not be conflated.

Where the rebuttal falls short: if UBI claws back the payment from high earners through progressive taxation, then in net terms it is already targeted, and "universal" becomes an administrative form rather than a distributive fact. Conceding that is more honest than pretending universality is magic in itself.

Sources: Finnish Ministry of Social Affairs and Health, Results of the basic income experiment (2020)

Editorial analysis

This largely stands, and we think it is more lethal than any economic objection. The Ontario case is real and it is on our evidence wall.

The one counter-example is Alaska: the Permanent Fund Dividend has paid out since 1982, across four decades and many changes of government, because of two design features — the money comes from an independent sovereign fund rather than the annual budget, and everyone receives it, so any politician cutting it faces the whole electorate rather than a small client group. Universality here is not a moral stance, it is political engineering.

Iran's programme has also run for over a decade. Durable cash transfers are therefore not impossible, but they must be designed not to require re-appropriation every year. We cannot claim this objection is solved. We can only say the opponents have identified precisely the part of the design that deserves the most work.

Sources: American Economic Journal: Economic Policy, The Labor Market Impacts of Universal and Permanent Cash Transfers (2022)

Editorial analysis

Partly stands, and it deserves the most serious treatment of any objection here — it is not an economic question, so economic data cannot dismiss it.

Measured evidence answers part of it. Finnish participants reported markedly less mental strain and higher life satisfaction than control. Mincome data from Dauphin showed hospitalisations down 8.5%, concentrated in mental health and accidents (though that finding was later challenged by a UBC reanalysis). Secure income did improve psychological wellbeing; the picture of recipients sliding into torpor did not appear.

But the objection points at something longer-run and social: if AI makes much work economically unnecessary, around what do people organise their lives? UBI cannot answer that, because it was never the tool for it. What UBI can do is separate "losing your job" from "losing your subsistence", so people in transition have time to find an answer rather than accepting any terms out of panic. That is a necessary condition, not a sufficient one. Anyone selling UBI as a solution to the crisis of meaning is overpromising.

Sources: Finnish Ministry of Social Affairs and Health, Results of the basic income experiment (2020) · Vancouver School of Economics, UBC, A Reanalysis of 'The Town with No Poverty' (2021)