Scott Santens has written a compelling case for universal basic income – subscribe to his Substack! – and I agree with the overarching goal of designing more effective and humane welfare systems. Unconditional cash transfers are central to the dignity and agency of poor households, and the operational case against means-testing — that it excludes the people it is meant to help, punishes work through claw backs, and wastes resources on gatekeeping — is well-grounded in evidence. But there is no frictionless version of Basic Income policy, and supporters of basic income ought to understand three complications before accepting the full package:
- the funding source determines whether a program helps or harms aggregate welfare;
- the labor supply evidence looks more favorable at the micro scale than the macro scale; and
- achieving genuine universality requires building administrative infrastructure, not dismantling it.
The Case for Universality
The operational case against means-testing is one of the best-evidenced arguments in the basic income literature.
- Means-tested benefits impose some of the highest marginal tax rates in the system on the lowest earners. As income rises, benefits phase out — the EITC alone withdraws at up to 21%, stacked on top of payroll taxes. When food assistance, housing, and healthcare clawbacks operate simultaneously, combined effective rates can exceed what any high-income earner faces. Hoynes and Rothstein (2019) identify this welfare trap as a primary structural argument for unconditional income support.
- Means-tested programmes routinely fail to reach the people they are designed to help. Around 20% of eligible US households do not claim the Earned Income Tax Credit (Tax Policy Center, 2024). Among eligible UK pensioners, Hernandez and Pudney (2006) find that 25–33% do not claim Income Support, a gap driven by real but measurable claim costs — stigma, complexity, and application burden. Wu and Meyer (2023) show that when Indiana automated its welfare system, SNAP enrolment fell 15% and TANF 24%, with the most vulnerable disproportionately lost at recertification.
- Replacing conditional programs with a universal basic income generates measurable welfare gains — not from the transfer itself but from removing embedded distortions. A 2024 Journal of Monetary Economics model finds that substituting the US income security system with an expenditure-neutral UBI raises output by 13%, labor force participation by 10 percentage points, and welfare by 2.8% in consumption-equivalence terms. The mechanism is elimination of asset-testing penalties and earnings-threshold disincentives, not the universality of the transfer.
- Santens is also right that a universal basic income and a phased-out negative income tax produce identical net transfers: recipients below the breakeven gain, those above pay back more in taxes than they receive. Chang, Han and Kim (2024) confirm both programmes generate comparable welfare outcomes at their respective optima.
The Case for Caution
The Funding Source
The macroeconomic literature is remarkably consistent on the point that the welfare outcome of a basic income scheme depends less on the transfer than on how it is financed. The transfer itself, in most models, is welfare positive. It is the financing mechanism that determines whether aggregate welfare rises or falls (Jaimovich et al 2022; Daruich & Fernández 2022).
Income tax financing is the most damaging route. Higher labor taxes reduce the after-tax return to working, drive marginal workers out of the labor force, and reduce capital accumulation — producing welfare losses that can more than offset the gains from the transfer (Jaimovich et al 2022; Prescott, 2004). One general equilibrium model found that a labor-tax-funded UBI reduced steady-state welfare by more than 20% (Daruich & Fernández 2022). Consumption tax financing performs substantially better and, in at least one model, generates positive welfare gains across all cohorts. Replacing existing means-tested programs — rather than supplementing them — is the most reliably positive scenario in the literature because it removes the distortions noted above without requiring equivalent new taxation. However, removing existing welfare programs is a highly fraught exercise that comes with its own massive political risks.
The Alaska Permanent Fund example, used as the paradigm case for political durability, is worth examining here. Alaska’s program is funded from oil revenues. Nobody pays higher taxes for it. A tax-financed national program would face a fundamentally different political economy, and the line between Alaska’s four-decade durability and a hypothetical federal UBI is less direct than the article implies. This also suggests a possible way forward in whole or in part, namely taxing Artificial Intelligence, however the devil is in the details.
Impacts on Labor Force Participation
The pilot and experimental evidence on labor supply is broadly encouraging. Systematic reviews of UBI-type experiments find no significant overall reduction in work effort. Pilots in Finland, Kenya, Stockton, and elsewhere show that recipients do not withdraw from the labor market in the ways critics predict.
The difficulty is that pilots operate in partial equilibrium. Participants receive a transfer, but the taxes required to fund a national programme are not imposed on anyone. The fiscal cost is invisible to the local economy. This is why pilots generally find positive results: they capture the benefit of the transfer while missing its primary cost. However, evidence from the largest US guaranteed income RCT to date — $1,000 per month for three years across 3,000 participants — found a 4.1 percentage point participation reduction with no detectable shift toward education, caregiving, or other productive activities.
When macroeconomic models impose the tax financing a national program would require, the picture changes. General equilibrium models find labor force participation reductions of 10 to 14 percentage points for larger programs, driven mainly by the tax channel rather than by recipient choices. That result is more sobering than the pilot literature as a whole, and it does not yet include the tax side of the ledger.
None of this means basic income is wrong. It means the labor supply evidence is context-dependent and scale-dependent in ways that advocacy summaries tend to smooth over.
Administration is transformed, not eliminated
All systems have boundaries and one should be cautious with claiming that universality obliviates the need to define and assess these boundaries. In this day and age, for example, how should non-citizens be treated? The claim in the universality argument is that a basic income eliminates bureaucracy – no applications, no eligibility checks, no caseworkers; the money simply arrives. This overstates the case. Achieving genuine universality — where the homeless person, the person without a bank account, and the rural elder without internet access all actually receive their payment — requires a universal population register, a reliable payment infrastructure, and an error-correction mechanism that can identify who has been missed without reintroducing surveillance. The EITC’s 20% non-participation rate — concentrated among the rural, self-employed, non-English speaking, and circumstantially irregular — illustrates that removing an income test does not automatically solve a delivery problem.
Building new administrative systems is likely to be costly and consume huge amounts of political capital. At the same time most countries have some systems already in place – there is no tabula rosa. In the U.S. context this may mean relying on and adapting existing systems, perhaps including the IRS and the SSA.
The Honest Case
Basic income is worth fighting for, and Santens makes a compelling case for why the current means-tested architecture fails the people it is designed to help. But the strongest version of that case is also the honest one: how these programs are funded matters and this is the first order policy choice, there may be non-trivial impacts on labor supply depending on the funding approach; and reaching everyone requires investing in administration, not assuming it away.
The very worthy goal of protecting the least among us in a way that lifts them up in dignity is a problem for which there is no single perfect solution, but we can say with certainty that a solution is within reach.
Be well.
There is no way to peace, peace is the way.
AJ Muste