One morning in late January, Tesla released its annual report, and up where the shareholders look was the number the company wanted them to see. Almost $5.7 billion of income earned in the United States in 2025. Roughly double what it had earned here the year before. A very good year, told the way a very good year is supposed to be told, to the people whose job is to be impressed by it.
Further back, in the tax footnote, where almost nobody looks, was the other number. On that $5.7 billion of American income, the current federal income tax for the year was zero. A record domestic profit on one page; on another, in the same document, nothing owed against it.
A disclosure rule that took effect for 2025 forced a third number into the same document, and it is the one that makes the first two impossible to wave away. Tesla’s cash income-tax payments that year ran about $1.2 billion worldwide. More than a billion of it went to China and other foreign governments. Twenty-eight million went to the United States, and even that was for older years, not for the $5.7 billion. Foreign treasuries collected. On the current year’s American profit, the American treasury collected nothing.
None of this is a crime. The Institute on Taxation and Economic Policy, which pulled every figure straight out of the filing, was careful to say there is no sign anything here was illegal. It came off through a short list of ordinary, legal deductions and credits, most of them written or widened by Congress, recently, on purpose. Over three years the company reported $12.58 billion of U.S. income and paid $48 million in federal tax on it — four-tenths of one percent. The number on the sign says twenty-one.
This is the same argument as the last eighteen posts. The money still gets out untaxed; it has just stopped belonging to a person and started belonging to a company, and the company only needs two sets of books to do it, legally.
The Rate Is The Part You’re Allowed To Fight About
The loud fight is the rate. In 2017 a tax law cut the federal corporate rate from thirty-five percent to twenty-one, and the argument ever since has been whether it should go back up — to twenty-eight, to twenty-five, to something. That fight is real, it is on television, and it is mostly beside the point, because the statutory rate is not what large companies pay and never was, in either direction. A small domestic company with an ordinary accountant pays something close to twenty-one. The largest companies in the country pay a fraction of it. The rate overstates the burden on one and understates it on the other at the same time. Arguing about the number on the sign tells you almost nothing about what is collected inside the building.
What is collected inside the building turns on a fact that sounds like an accusation and is actually just the law. A large public company keeps two different profit numbers, and both are true. One is book income: the profit it reports to shareholders under accounting rules, the number it wants as large as it can honestly make it, because the stock trades on it. The other is taxable income: the profit it reports to the IRS under the tax code, the number it wants as small as it can legally make it, because the tax is owed on that one. Same company, same year, two figures, each correct under its own rulebook. The space between them is not fraud. The space between them is the design, and the rest of this post is a tour of how the space is built.
It is worth knowing how large the space has gotten. According to the Office of Management and Budget’s historical tables, the corporate income tax supplied roughly a third of all federal revenue in the early 1950s. It supplies about nine percent now. Measured against the size of the economy it has fallen from around five percent in the 1950s to about a percent and a half over the last decade, and the statutory rate in those high-collection years was above fifty percent. The number went down because it was walked down, deliberately, over a long time, by people who benefited at every step, and it fell fastest in the decades they insisted there was no money for anything else.
A few honest things before the mechanisms, because the argument doesn’t survive without them and this series doesn’t build a strawman to knock down. This is not the claim that companies shouldn’t be taxed, or that every line on a corporate return is a dodge. Some of the gap is defensible. A company is taxed on its profit and its shareholders are taxed again on the dividend or the gain, which is the same dollar taxed twice, and the lower second rate is partly an answer to that. Letting a business deduct the cost of a machine quickly is a real incentive to buy the machine, not a trick. And a company that genuinely lost money for years and finally makes it back should not be taxed on the good year as if the bad ones never happened. Those are real, and the post will come back to each of them.
And mostly this is a consensus, and the post won’t pretend otherwise in either direction. Transfer pricing, the deferral of foreign profits, the rules that let a company largely choose how its own subsidiaries are taxed, and the simple decades-long fact that nobody made companies pay tax on the profit they show shareholders — that machinery was built and left standing by both parties, across many administrations. There is no clean villain there, and inventing one would be its own dishonesty. But two pieces are not symmetric, and flattening them into “both sides” would be the opposite dishonesty. The 2017 law that cut the rate, and the 2025 law that made the cut and the largest write-offs permanent — the same one-party reconciliation bill the last post named, the one that cleared the House two hundred eighteen to two hundred fourteen with no votes from the other side — were one side’s, on near-party-line votes. And the slow starving of the part of the tax agency that audits large companies has come overwhelmingly from one direction. The Treasury’s own inspector general reported this spring that the enforcement money is down to a few billion after rescissions; the agency lost roughly a quarter of its staff in 2025, including about a third of the revenue agents who run the slow, complex audits of big firms; and the last time resources were cut like this, the audit rate on large corporations was simply cut in half. A rule that is never enforced on the people best able to dodge it is, for them, not a rule. The consensus is real. So is the asymmetry. Saying both, and letting neither cover for the other, is the job.
How The Big Number Becomes The Small One
Open the report again. The big number is near the front. The small number is in the back, in the tax footnote, in language built so you will not read it. Between them is a short list of named items, and each one is a legal bridge that carries profit out of the number the IRS sees while leaving it in the number the shareholders see. Nothing here is hidden. It is all printed, in the same document. The trick is not concealment. The trick is that the footnote is boring and enormous, and the fight you are offered instead is the one about the rate. Four bridges carry most of the weight.
The first is the gap itself, and the one tool ever built to close it. Because book income and taxable income run on different rulebooks, a company can show investors a record year and show the IRS almost nothing, legally, in that same year. In 2022 Congress finally aimed something at exactly this: a fifteen-percent minimum tax on the book income of corporations averaging over a billion dollars of it — a tax on the number they show shareholders, precisely because that number is the honest one. It was the first serious attempt in decades to tax the profit a giant company brags about. Then the 2025 law restored a set of deductions that shrink taxable income without touching book income, which widens the very gap the minimum tax was built to catch, and the Treasury was left with enough discretion to soften how hard the minimum tax bites. The Tax Policy Center’s analysis lays out the result plainly: the new breaks can lower the regular tax so far that the minimum tax is supposed to take over, but whether it actually does is now a question of administrative guidance, not law. The backstop exists. It is being loosened in the plumbing, where no one stands on a lawn to announce it.
The next line down is the one that did the actual work in the scene at the top. Accelerated depreciation lets a company deduct the cost of equipment far faster than the equipment wears out, and full “bonus” depreciation lets it deduct the whole cost in year one. The machine keeps producing income for a decade or two; the deduction is taken now, all at once, against this year’s profit. Stack enough current investment against current earnings and the taxable number goes to zero while the book number stays a record. Pair it with immediate expensing of research costs and the effect compounds. This is the bridge that carried roughly half a billion dollars off Tesla’s bill in a single year. And the 2025 reconciliation law made both of these permanent — the Joint Committee on Taxation’s own estimate puts the depreciation piece around three hundred sixty billion dollars and the research piece around a hundred forty over ten years — on the same one-party vote named above. There is a genuine investment-incentive argument for letting companies do this, and the post takes it seriously a few paragraphs from now. What matters here is narrower: a timing rule sold as encouragement to invest, made permanent, is also the cleanest legal road from a record profit to a zero bill in the same year.
The third bridge is the cleanest case in the whole footnote, and it is worth dwelling on precisely because it is small. When a company pays its executives in stock instead of cash, it records a non-cash expense on its books. Later, when the options are exercised, it deducts the actual spread between the grant price and the market price — real dollars off the tax bill, for compensation that cost the company no cash, and frequently a far larger deduction than the expense it ever booked. The deduction grows with the stock price, which means it is largest in exactly the years the company is winning most. In the scene at the top it was worth about a hundred seventy-two million dollars in one year, in a year the company’s chief executive was being handed a pay package the size of a small country’s budget. This is not a drafting accident that survived. It is one paragraph. It could be limited to the expense actually booked in a single sentence. It has been proposed and not done, repeatedly, which is the tell: a fix this small and this obvious, left undone this long, was left undone on purpose.
The fourth bridge is the oldest and the most bipartisan, and it is where the foreign-tax number in the opening came from. For decades the architecture has let a multinational book its profit where the tax is low and its costs where the rate is high — the patents “owned” by a subsidiary in a small low-tax country, the American sales paying that subsidiary for the rights, the profit landing somewhere it is barely touched. The names change; the Double Irish closed and the structures rearranged around it. The 2025 law reworked this layer again, renaming the main offshore-income regime and adjusting the rates and deductions inside it. The part that is new and asymmetric is what happened to the only real attempt to put a floor under the whole game. More than a hundred forty countries had agreed to a global minimum corporate tax. In June 2025 the United States negotiated its own multinationals out of it: a “side-by-side” deal, announced by the Treasury, that fully exempts U.S.-parented groups from the global minimum’s main enforcement rules, in exchange for the United States dropping a retaliatory tax it had threatened other countries with. The Bipartisan Policy Center’s tracking has the arrangement contested and unsettled into 2026. The plain version: a company can pay over a billion dollars in income tax to China and other governments and nothing to its own on the same year’s domestic profit, and the one international structure built to stop that was the structure the United States arranged for its own companies to skip.
What’s Real
A few honest qualifications, because the strongest version of the argument is the one that doesn’t pretend the counter-arguments aren’t there.
Double taxation is not a talking point invented for this. A corporation pays tax on its profit, and then the shareholder pays again on the dividend or the gain that comes out of that same profit. That is genuinely the same dollar taxed at two stages, and a coherent system has to account for it somewhere; the lower rate on the second stage is partly that accounting, not pure favoritism. The honest way to argue corporate tax is on the combined burden across both stages, not the headline rate alone, and a critique that hides that is arguing in bad faith.
The investment case for fast write-offs is real too. Letting a company deduct a machine immediately, rather than over twenty years, lowers the after-tax cost of buying the machine, and there is decent evidence that does pull some real investment forward. Research expensing has a stronger version of the same logic, because research throws off benefits that spill past the company doing it. These are not nothing, and a serious person can support the incentive while opposing what it has become.
Loss carryforwards are the honest one, and the last post’s draft conceded the principle for exactly the kind of company people love to be angry about. A firm that loses money for years building something and then finally turns a profit should not be taxed on the good year as though the lean years never happened; taxing only the up years and ignoring the down ones would overstate real lifetime profit and punish the act of trying. The argument against carryforwards is not that they should not exist. It is about scale and abuse — losses from a different era used to wipe out tax in a year of historic profit. It is worth being precise here: the company in the opening did not get to zero mainly through old losses. It got there through the depreciation, the credits, and the stock deductions. The carryforward is the bridge with the best defense, and it is not the one carrying most of this particular weight.
And the competitiveness argument deserves a fair statement. Most rich countries tax their companies mainly on profit earned at home and largely leave foreign profit to foreign governments, and the United States moved toward that system for a reason: tax worldwide income too hard and companies relocate the headquarters, not just the accounting. That pressure is real. None of these points is fake. The point of the post is not that they are. It is that they are the reasons given for a structure whose actual output is the distribution above, and that “this is hard” has been doing the work of “this cannot be done” for a long time, said by people who know the difference and benefit from the confusion.
What They’re Paying For
What the rigged corporate code actually produces, set against what the rate fight claims to be about, is four things stacked together. Each is worth real money to someone specific, and the someone is nameable in every case, which is the difference between this and a complaint about “corporations.”
It is a permanent, legal gap between the profit a company shows the people it wants to impress and the profit it shows the people it owes. The gap is widest at the top, because the bridges across it require lawyers, scale, and a tax department, and the corner manufacturer has none of those. The beneficiary is not abstract. It is the largest firms specifically, the ones with the resources to run both sets of books well, collecting a discount the small competitor down the road cannot reach.
And it is a timing rule, sold as a reason to invest, that once made permanent became a reliable way to report a record year and owe nothing on it in that same year. The beneficiary is the capital-heavy giant with enough current spending to bury its current profit, and the shareholders who get the record book number and the zero tax bill in the same report, with no contradiction the law recognizes.
And it is a deduction for compensation that left no cash, sized to the share price, largest exactly when the company wins biggest. The beneficiary is the executive paid in stock and the equity holders alongside, who convert a pay package into a tax shield, so that the better the stock does and the larger the package, the less the company owes for handing it over.
And it is a set of borders the profit crosses on paper while the work, the customers, and the public services stay here — defended now by a government that talked its own companies out of the one global floor built to stop it. The beneficiary is the multinational and the advisory industry that moves the paper, and the political actors who took the donations, kept the architecture standing, cut the agency that might have checked it, and arranged the exemption. That is the bill the argument about the rate has been keeping you from reading.
The Fixes Are Boring
The fixes below are structural, not slogans. The post is not going to put “close the loopholes” or “simplify the code” or a token bump in the statutory rate on the list, because a higher rate on the sign changes nothing for companies whose effective rate already ignores the sign, and “close the loopholes” with no loophole named is the sound the consensus makes while keeping them. These sound impossible because the people who profit have spent a long time making them sound impossible and making the rate the fight you have instead. Several are things the country itself recently did, then loosened. Roughly cheapest and most immediate to genuinely hard:
- Raise the buyback excise to a rate that changes the math, and stop letting the rules quietly shrink it. A one-percent tax on stock buybacks is small enough that companies treat it as a cost of doing exactly what they were going to do. The diluted version is already on the record: enacted at one percent in 2022, then narrowed by the Treasury’s final rules in November 2025 to exempt much of what it was meant to reach. Set a rate that bites and write the rules so they cannot be carved back out.
- Ring-fence multi-year funding for the unit that audits large companies and large partnerships, protected from the annual clawback. This is an appropriation, not a new agency. The 2022 funding was projected to return several dollars for every dollar spent; most of the enforcement money has since been rescinded and roughly a third of the complex-audit agents are gone. The watered-down version — fund it, then let each budget quietly take it back — has already been run, on purpose. Make the high-end enforcement money durable and unrescindable.
- Cap the stock-compensation deduction at the expense the company actually booked. No deducting more off the tax bill than the cost shown to shareholders for the same stock. It is one paragraph and could be written in a sentence. It has been proposed and dropped repeatedly, which is the evidence that the obstacle was never the drafting.
- Give the corporate minimum tax on book income real teeth, and take the discretion to soften it out of the rulemaking. The 2022 minimum tax was the right target — the profit shown to shareholders — and the 2025 law promptly widened the gap it was built to catch while leaving Treasury room to relax it. The diluted version is the one we have. Set the floor in statute, on the book number, with the loosening levers removed.
- End permanent full expensing, or pair it with the book-income floor so it cannot zero out the bill. Either return depreciation to something that tracks how the asset is actually used, or keep fast write-offs but make the minimum tax on book profit the backstop that catches the result. Every temporary version of full expensing was written to sunset and did; the 2025 law made the giveaway permanent on a party-line vote, which is the move this reverses.
- Tax the profit where the sales, the workers, and the customers are, not where the paperwork says the patents live. Apportioning a multinational’s tax to its real economic footprint — or a worldwide minimum with actual force — is the only thing that reaches profit-shifting at the level it operates. The diluted version is current events: the global minimum was negotiated down to a side-by-side arrangement that exempts U.S.-parented groups, the United States having opted its own companies out.
- Tax book income itself, at a real rate, with the administrability built in — and say plainly this is the hard one. A genuine tax on the profit companies report to shareholders is the only fix that closes the gap rather than patching a bridge across it, and it carries real problems: accounting standards can be lobbied, treaties constrain it, and the firms it reaches will fight it hardest. It is last because it is hardest, not because it is unserious. The honest position is that the difficulty is a reason to design it carefully, not a reason to keep treating the other six as radical so this one is never discussed.
Six of the seven are a statute or an appropriation — things a Congress or an administration could do without inventing a new institution, several of them by reversing a loosening enacted within the last few years. The seventh is the genuinely hard one, and it is honest to say so rather than pretend the list is all easy. None of them is the radical position. The radical position, measured against the decades when corporations paid the largest share and the country built the most, is the one we are living in: record profits reported to shareholders and nothing owed on them at home, defended by an argument about a rate almost none of the money ever pays.
Who Is This For
The same logic that lets the owner’s gain go untaxed lets the company’s profit go untaxed, walked in full in the last post; this one is just that machine seen at the corporate scale, where it doesn’t even need a death to finish the job. What you were shown is the rate. Where the money is is the gap. The next post turns the same question the other way around. The dodge is the money the company keeps that never gets taxed; the next layer is the money that flows the other direction — the bailouts, the subsidies, the public funds that arrive when the same companies need them, the free market that only runs one way. Part 20 is corporate socialism: privatized profit, socialized loss. Different layer. Same question.
Even that's Odd 
Leave a comment