The Math Nobody Wants to Do

The UBI debate has a step almost everyone skips, on every side of it. The actual math. Not "it would cost a lot" or "we could fund it with a VAT." The line-by-line arithmetic of what universal basic income costs versus what any proposed funding mechanism actually generates. The pilot data exists, the funding sources are documented, the biometric infrastructure is getting built. None of that changes what the arithmetic says.

So here it is: 267 million US adults times $1,000 a month equals $3.2 trillion a year. That's roughly 47% of total federal spending in FY2024, which came to about $6.75 trillion. Drop UBI to $500/month and you're still at $1.6 trillion. Push it to $2,000/month and you hit $6.4 trillion, which is nearly the entire federal budget.

The Best Proposal Is $1.9 Trillion Short

Andrew Yang's Freedom Dividend was the most detailed UBI funding proposal any major presidential candidate has put forward. His plan cobbled together a 10% value-added tax, consolidation of existing welfare spending, lifting the Social Security payroll tax cap, a carbon tax, and a financial transactions tax. The Tax Foundation scored the combined revenue at roughly $1.3 trillion per year.

That's $1.9 trillion short of the $3.2 trillion gross cost, every single year.

The VAT alone, which Yang pitched as the engine of the whole thing, illustrates the scale of the problem. The Tax Foundation estimated a broad-base 10% VAT would raise about $952 billion annually. The CRFB was more skeptical, putting the figure closer to $600 billion. For context, European VAT rates average around 20%. Yang proposed half that, and even the generous estimate covers less than a third of the cost. The CRFB went further and labeled Yang's overall revenue projections "Likely False."

Then there's the question of what deficit-financed UBI actually does to the economy. The Roosevelt Institute modeled both scenarios in 2017: a $1,000/month UBI paid for by deficit spending would expand GDP by $2.5 trillion over eight years. But fund it through taxes (the revenue-neutral version) and the growth drops to 2.62% over the same period. That's a massive gap, and it means the economic case for UBI changes completely depending on how you pay for it.

$7 Trillion in Projections. $44 Billion in Tax Revenue.

The implicit promise behind most UBI advocacy is that AI will generate so much economic value that funding a basic income becomes trivial. The projections look enormous. McKinsey estimated generative AI alone could add $2.6 to $4.4 trillion per year to global corporate profits. Goldman Sachs projected a 7% increase in global GDP, roughly $7 trillion over a decade. PwC put the total at $15.7 trillion in additional global GDP by 2030.

Run those through actual tax math. Goldman's $7 trillion is global, spread over ten years: roughly $700 billion annually worldwide. The US share is about 25%. Tax that at a 25% rate and you get approximately $44 billion in new annual federal revenue. That's 1.4% of the $3.2 trillion UBI cost. Even Accenture's more generous US-specific projection ($8.3 trillion by 2035) covers about 3.3% of the cost after the same tax math.

And so far? Goldman's own chief economist said in early 2026 that AI contributed "basically zero" to US GDP growth in 2025. The bank now projects AI won't meaningfully move US GDP until 2027, with a boost of 0.4 percentage points by 2034. No major institution has published modeling that connects AI economic gains to UBI funding at scale. The CBO called the revenue impact uncertain. The windfall that's supposed to pay for all of this hasn't arrived, and nobody projecting it has shown how it gets from corporate balance sheets to a monthly check.

Friedman's Warning, Now Coming from the Other Side

Milton Friedman originally pushed a negative income tax (a close cousin of UBI) as a way to replace the welfare state with something simpler and cheaper. He warned that the left would eventually "buy a Trojan Horse," adopting his market-friendly poverty fix without realizing it would gut comprehensive social programs.

Fifty years later, that's exactly what the left is worried about. The Trojan Horse critique is now the primary objection from welfare policy advocates: Robert Greenstein at the Center on Budget and Policy Priorities, major labor unions, disability rights organizations, and most poverty researchers who've actually run the replacement numbers. Their argument is specific: replace SNAP, Medicaid, housing vouchers, and disability benefits with a flat $1,000/month check, and people with high medical costs, disabilities, or kids in expensive housing markets end up worse off. A flat payment ignores the uneven distribution of need.

Yang tried to defuse this with an opt-in structure. Recipients of cash-equivalent benefits could choose UBI instead, but nobody would be forced to switch. Health care, housing assistance, and Social Security would still stack on top. Critics argue the opt-in is temporary. Political pressure to mandate the switch would build over time, and people who don't understand the trade-off (the math, again) would make choices that leave them with less.

Charles Murray's version from the right makes the Trojan Horse explicit. His 2006 book In Our Hands proposed $10,000 per year to every American adult, funded entirely by eliminating all federal welfare programs. Someone on combined disability, housing assistance, and SNAP pulling $2,200/month gets a flat $1,000 check. The flat payment treats everyone's needs as identical.

Every Alternative Has a Different Blind Spot

The UBI debate rarely acknowledges that the alternatives each solve a real problem while creating a different one.

A federal job guarantee makes the government the employer of last resort, but it does nothing for caregivers, people with disabilities, or anyone who can't clock in. Expanded EITC effectively reduces poverty among single mothers, but it's tied to earned income and misses everyone outside the labor market. Shorter work weeks distribute existing work but don't create an income floor. Every proposal is a patch for a different population, and no single program architecture fits 267 million adults at a price anyone will actually pay.

The Alaska Problem

Alaska's Permanent Fund is the one model that actually works at scale, and it's been running for 43 years. The fund hit roughly $83 billion by December 2024. It distributes annual dividends to every resident. Jones and Marinescu found no significant employment reduction over four decades. Part-time work increased by 1.8 percentage points. Somewhere between 15,000 and 25,000 Alaskans are kept out of poverty by the dividend each year.

The catch is obvious once you look at it. Alaska has around 730,000 residents and massive oil royalty revenue. Replicating this model for 267 million adults would require a sovereign wealth fund in the trillions, funded by resource revenues or financial transaction taxes that don't currently exist. The model that works is the one built on something the rest of the country doesn't have.

Who Needs the Math to Stay Vague

The most revealing thing about the UBI debate is how many different groups benefit from keeping the conversation vague.

Tech billionaires get cover. The same executives funding UBI research spent $1.1 billion in political spending during the 2024 cycle and through 2025, as Part 2 showed. Companies like Uber, Lyft, and DoorDash spent over $200 million on Prop 22 to keep gig workers classified as contractors, destroying the benefits-bearing employment that UBI is supposed to replace. Backing UBI lets them look concerned about displacement without touching the labor model that makes their businesses profitable. A job guarantee or worker reclassification law would cost them real money. A UBI check, paid for by taxpayers, would not.

Conservative policy advocates get a Trojan Horse. A UBI framed as "simple and efficient" opens the door to dismantling categorical programs that are harder to eliminate individually.

Progressive politicians get an aspiration without a price tag. The only active federal bill, Rep. Bonnie Watson Coleman's Guaranteed Income Pilot Program Act (H.R.5830), hasn't moved past committee referral. Watson Coleman announced in November 2025 that she won't seek re-election. The bill's primary sponsor is a lame-duck member in a Republican-controlled Congress.

The mechanism is the same across all three groups: keep the debate at the level of "should we have UBI?" and you never have to confront the $1.9 trillion gap.

Seven Questions Standing Between UBI and Reality

COVID stimulus proved the delivery infrastructure exists. The IRS sent $1,200, then $600, then $1,400 to hundreds of millions of people. A Pew Research survey found broad support. The logistics work. What doesn't work is everything else.

Four parts of this series surfaced questions that no UBI proposal, from any direction, has resolved. Here's the list.

The funding hole. Yang's was the most complete proposal and it's $1.9 trillion short every year. Nobody since has proposed anything more complete. The AI windfall that's supposed to close the gap covers 1.4% of the cost when you run the actual tax math. What fills the other 98.6%?

The tax base problem. Every UBI funding model assumes a tax base that holds steady. But if AI displaces workers or pushes the economy toward 32-hour weeks and contractor arrangements, earned income drops. Less earned income means less consumer spending, less payroll tax revenue, less of everything the funding models depend on. The same force creating the need for UBI simultaneously shrinks the tax base that would fund it. None of the major projections from Goldman, McKinsey, or PwC model this feedback loop.

The healthcare cliff. Over 160 million Americans get health insurance through their employer. UBI proposals either ignore this entirely or assume healthcare stacks on top. But if AI displacement breaks the employment model that delivers insurance, $1,000/month doesn't cover premiums. Murray's version eliminates healthcare along with everything else. Yang's preserves it, but what happens when the employer-sponsored system that covers half the country doesn't have enough employers? Without a public healthcare system, AI displacement isn't just an income problem. It's a coverage crisis that a monthly check can't touch.

The benefits infrastructure collapse. It's not just health insurance. Retirement contributions, disability insurance, unemployment insurance, workers' comp. All of it runs through the employer relationship. A flat $1,000 check doesn't replace any of those systems. It doesn't even acknowledge they exist.

The flat-check design problem. A flat payment treats a 22-year-old in rural Arkansas and a disabled single parent in San Francisco as having the same needs. Adjusting for cost-of-living, disability, or family size reintroduces the bureaucracy UBI is supposed to eliminate. Nobody has proposed a design that solves this without recreating the system it replaces.

The payment infrastructure question. Part 3 of this series documented the biometric gating and programmable money being built into digital payment rails. Who sets the access conditions? Who decides what counts as eligible spending? The infrastructure choices being made now will shape what UBI actually looks like if it ever arrives.

The political coalition that doesn't exist. The left wants UBI as a supplement. The right wants it as a replacement. Tech wants it as cover. The version that satisfies all three has never been proposed, because it can't be. The design decisions that matter most (replacement vs. supplement, amount, who's covered, what's cut) are the exact decisions that split every coalition that tries to build one.

None of this makes UBI a bad idea. The 43-year Alaska track record alone makes dismissing it lazy, and the pilot data across five continents is more consistent than almost any social policy ever tested. But the version of UBI that actually reaches 267 million adults, funded by mechanisms that generate $3.2 trillion, designed to protect people whose needs exceed a flat check, delivered on infrastructure that doesn't require biometric enrollment, built to survive the collapse of the employment model it's supposed to replace? That version hasn't been proposed. These are the questions standing between the idea and the implementation, and almost everyone in the debate has a financial reason to leave them unanswered.