Essay · Tax Policy

The Output Went Up.
The Tax Code Didn't Notice.

Right now, the U.S. tax code rewards a company for firing people. That's not a glitch. It's what a hundred-year-old system does when the economy it was built to measure stops looking anything like the one we live in.

Draft · ~1,900 words · 10 min read

Right now, the U.S. tax code rewards a company for firing people. Cut headcount and keep revenue flat, and your tax bill often goes down. That's not a bug anyone designed on purpose, it's what happens when a tax system built around wages meets a company that's figured out how to make more with fewer people. The United States has gone through a productivity renaissance, technology has improved the output of workers steadily for over a century, and the pace may be about to compound faster than at any point since the industrial revolution. With automation, artificial intelligence, and robotics, we might see the first trillion-dollar one-person company. If that happens, it wouldn't be a glitch in the tax system. It would be the tax system working exactly as designed.

Collecting next to nothing.

That's what's happening now. Here's how we got a tax system that wasn't built to notice it.

Why the gap opened

This is not our first rodeo with productivity gains. The United States is where it is because of continuous innovation, adaptation and technological advancement. The world of 1913, when income tax became the backbone of our tax base, looks nothing like the world of today. As the country industrialized, the old system of tariffs and excise taxes couldn't keep up. Its replacement has managed to last for over 100 years, but the cracks are widening, and we may be getting to the point where the dam finally breaks.

In the past, every fundamental shift in government policy has happened after a crisis: war, famine, civil unrest, recession, or some combination of them all. And we may be approaching that tipping point again soon. So let's not wait this time. It's not a right or left issue. We can and should de-risk as we walk into an uncertain future.

Productivity vs. Compensation, Indexed (1979 = 100)

1979 2026 0% Productivity +93% Pay +34%

Productivity grew 93% while typical hourly compensation grew just 34% from Q4 1979 through Q1 2026. Pay is average compensation of production/nonsupervisory workers (~80% of the private-sector workforce); productivity is net output per hour for the total economy. Source: Economic Policy Institute, Productivity–Pay Tracker, from underlying BLS Labor Productivity and Costs data.

Why income tax made sense in 1913

Before 1913, the federal government leaned mostly on tariffs and excise taxes; revenue collected at the border or on specific goods, which worked when the economy was smaller and moved slower but couldn't keep pace with an industrializing country. The 16th Amendment gave Congress the power to tax income directly. But the question is not what but why. Tariffs fell hardest on ordinary consumers, while the era's new industrial fortunes in oil, steel and rail were growing rapidly, largely without being taxed. State and local governments also leaned heavily on property taxes, but as wealth shifted from land into stocks, business income, and industrial capital, a land-based tax missed where the money increasingly was. We needed a tax solution where the wealth was accumulating. I'm proposing we're living through a similar time today.

Before computers or even the punch clock, hours of income was the right measuring stick to gauge economic output. Wages worked as a proxy for output because, in a factory based economy, the two moved together closely enough that taxing one was almost the same as taxing the other, and wages were already visible, since employers were already tracking and reporting on them.

In 1913, taxing wealth or output directly would have meant valuing land, inventory, equipment, and business assets across the entire country, with no reliable national records to work from. A paycheck, by contrast, was already a number employers were tracking, and it moved through banks. Income tax wasn't chosen because it was the most accurate measure of value creation, it was chosen because it was the only system the government could actually administer at a national scale.

Success, and tax-base erosion aren't sequential here. They're the same event.

Productivity's next phase

With every previous technological advance, human productivity exploded, and each time, new jobs eventually absorbed displaced workers. That pattern is the entire argument for not worrying about this one too. But it's also the pattern AI breaks, on purpose. Every past technology automated a task. AI is the first one explicitly built to automate the worker, the whole bundle of judgment, labor, and decision-making a paycheck used to represent. And it doesn't have to be proven absolutely correct or happen on any particular timeline, for the tax problem to already be real: the more AI succeeds at what it's built to do, the less income tax the very companies profiting from it will generate.

The foundation of our tax system and our main social safety net, Social Security, relies entirely on human wage labor. Every time we have a crisis, taxable income takes a hit at the exact time it's needed. Isn't it time to think about how we build taxes for a future not built around how many hours get clocked, but what gets bought and sold?

Two bases that don't erode the same way

I'm not proposing a complete replacement of income tax overnight. Instead, build a Value Add Tax infrastructure alongside income tax, and we can scale VAT up as needed when it's needed, not after the next crisis has already occurred.

A value add tax is the only way to measure and tax a billion dollar company with 10 workers making $250k a year against a billion dollar company with 10,000 workers each making $48k a year. VAT is also the system that scales with technological advancements in a future that may or may not be tied to human labor.

VAT's tax base grows with GDP, which cannot currently be said definitively for income tax. Wages simply don't keep pace with output. VAT also removes the tax incentive to run lean. A company doesn't lower its tax bill by employing fewer people. It's taxed on what it sells, whether that took 5 employees or 500.

While we have sales taxes at the local and state level, that's applied at the final retail sale and often doesn't address service industries which have dominated the growth side of our economy in recent decades. VAT at the federal level taxes value at every stage of production, making it both harder to dodge and a better match for our service and technology based economy.

VAT also shines against traditional corporate income tax, where profits are easy to shrink on paper or "re-invest" into the business. The transaction is what is taxed regardless of whether the company chooses to show a profit or not.

VAT also has a track record most other taxes simply don't. It's harder to evade, less sensitive to the swings of a recession, and less exposed to the games multinational companies play to shift profits offshore than income tax is. Which is a large part of why more than 100 countries already run one. Exporters get rebated at the border, imports get taxed at the border, so domestic producers aren't put at a disadvantage against foreign competitors the way they can be under an income-tax-only system. Again, most of the developed world already runs on this. The US is the outlier.

Why we can't have nice things

What will kill a transition to VAT before it can ever be considered, is one political side claiming this policy comes from the other. In our bipolar world, if a meteor was hurtling toward earth and one side sounded the alarm the other would call it fake news or find a reason to be pro-meteor. And to be clear, the concept of wages not tracking with productivity has surfaced all across the political spectrum. But in our algorithm based, us vs them society, if we're ever going to make a real change, it'll have to be bipartisan.

Libertarian right Center-right Technical center
Cato: no country has successfully replaced income tax with VAT, it always layers on top. Objects that VAT is hidden in prices, letting spending grow unnoticed. Niskanen & Brookings: conservatives have long favored consumption taxation since it doesn't punish saving/investment. Rep. Bill Thomas once floated a corporate VAT. CRS & Tax Policy Center treat it as a design question, not ideological. Avi-Yonah: the "replace, don't add" framing has been wrong for 35 years — the U.S. will need both.

Sources: Cato testimony, Niskanen/Brookings, Avi-Yonah, "Why the US Needs Both an Income Tax and a VAT"

While there is support across the spectrum, there is skepticism on the libertarian right. Cato argues that no country has successfully replaced an income tax with a VAT. Every real-world VAT ends up layered on top of the existing income tax instead. Which happens to be what I'm arguing for. But their real objection is that a VAT is hidden in the price of goods and this lets government spending grow without voters being aware. Not going to dismiss the point but we can track VAT and make it transparent if that is a big enough concern. But I have to say, we've never as consumers been able to know what a company is making in profit. If consumers are comfortable buying a luxury good with a 500% markup and are happy with the purchase, it makes me feel like while this is a fair transparency concern, it's also one that applies to plenty of taxes we already accept.

The one thing nearly everyone agrees on: a VAT without an offset hits lower-income households harder as a share of what they spend. Which is valid, but we can offset the increase with income tax adjustments where needed.

It's about reliability and future outlook

While the world has been fixated on the bite of inflation, it is just one half of the coin. The cost of healthcare, housing and food continue to rise, and the government's ability to provide services is getting squeezed on both ends. And that squeeze doesn't show up gradually, it shows up all at once. Recessions are exactly when we need government coffers to be full but it's the exact time when the government's biggest revenue source, wage-based income tax, shrinks the fastest. The bills don't arrive on a schedule that matches when the tax base can cover them.

The numbers

Federal revenue, 2023
16.5%
of GDP, just under the post-WWII average of 17%
Individual income tax
49%
of total federal revenue — the single largest source
SS trust fund depletion
2034
per the 2026 Trustees Report
Benefits still covered
83%
from incoming payroll tax after depletion

Sources: Congressional Research Service, "Overview of the Federal Tax System in 2024"; SSA, 2026 Trustees Report.

Individual income tax supplies about 49% of total federal revenue, the single largest source, and CBO projects 65% of all income reported on individual returns in 2024 came from wages and salaries alone.

Right now older Americans rely on Social Security in retirement. Social Security is funded almost entirely by a 12.4% payroll tax on that same wage base. The ratio of workers paying in to beneficiaries drawing out has fallen from more than 5-to-1 in 1960 to under 3-to-1 today, and payroll tax is capturing a shrinking share of total wages as high earners' pay outpaces the taxable cap.

The system is already on the ropes, but a shift in employment or wages will put the entire system down on the mat in a way that it might never get up again. This is not about workers rights or fair wages, but this tax system is for us, the American people. We pay taxes because we expect to receive a benefit from them. With the future of Social Security in question, and future threats to the labor force, we need to be thinking about what is best for us now and in the future.

Where VAT falls short

Of course VAT has its issues and detractors. The most consistent criticism is regressivity, with lower-income households spending a larger share of what they earn, so a flat consumption tax takes a proportionally bigger bite from them. This shows up in Tax Policy Center, IMF, and Penn Wharton modeling alike.

Many countries have tried to offset this with exemptions for specific goods like food or medicine, an approach that's often cumbersome and inefficient, since wealthier households buy these too. A direct rebate, a flat payment or refundable credit might be a simpler way of offsetting what lower-income households pay in VAT. Tax Policy Center modeling found a VAT paired with a targeted rebate could hold the burden on the bottom fifth of households to a fraction of a percent of income and Canada's GST credit already does a version of this.

Another weakness of this proposed change is the one-time price shock when the tax is implemented. But I feel like our recent foray into tariffs has shown that the market can adjust well under uncertainty. There may be some adjustment and pain but a smooth rollout can be done.

VAT is also more resistant to capital flight than might be obvious. It taxes consumption at the point of sale, in the country where the sale happens, so a US company selling overseas still owes VAT wherever the customer is. That revenue can't be booked in a low-tax jurisdiction the way corporate profit can. There is no equivalent of an offshore subsidiary for a purchase a customer makes at home. Of course there is real mobility risk still but that isn't the VAT side, it's the capital gains side. Wealth and investment income can relocate, but that's a topic for another discussion.

The real question

Going back to where this started: a tax code that rewards a company for having fewer people on payroll. That's not a design flaw waiting to be patched, it's what happens when a hundred-year-old measuring stick is used on an economy it was never built to measure. Wages made sense as a stand-in for output in 1913 because it was the only number the government could actually use. It's a much worse stand-in today, and it's going to keep getting worse, on a timeline nobody, left or right, actually controls.

None of this requires betting the whole argument on how fast AI moves, or picking a side in that fight. VAT doesn't ask a company how many people it employs. It asks what it sold. That is the shift. Layer a consumption tax alongside the one we already have, let it scale as the wage base shrinks. It's the difference between a tax system that notices what's actually happening in the economy and one that's counting hours on a punch clock that stopped existing decades ago.

The real question isn't whether the wage-based tax system holds up forever. The numbers already say it's under strain. The real question is whether we build the replacement calmly, ahead of the next crisis, or wait and build it the way we always have: after the dam breaks, in a panic, with whatever's left.

Tyler Herman is a UX/UI designer and product manager writing about tax policy, AI, and the shape of the economy to come. More about this project.