In September, Treasury proposed new rules on which company costs count against two big corporate tax breaks. The short version: most multinationals come out ahead. The public can comment until November 10.

The quick version: Big US companies get two special tax deals, one for selling abroad and one on profits earned through foreign subsidiaries. How much those deals are worth depends on which costs get counted against them. Treasury's new rules say most head-office costs, interest and research spending no longer count, so for most companies the deals just got bigger.

When people argue about corporate tax, they usually argue about the rate. Is 21 per cent too high? Too low? But a rate is only half the story. The other half is the base: how much profit the rate actually gets applied to. And the base often depends on a boring-sounding question with big money attached: which costs does a company have to subtract from which slice of its income?

On September 10, 2026, the Treasury Department and the IRS released proposed rules, officially labelled REG-117273-25, that answer one of the biggest allocation questions left open by the One Big Beautiful Bill Act (OBBBA). That is the tax and spending law signed on July 4, 2025. The proposals were published in the Federal Register the next day.

The rules already apply to 2026 tax returns, and anyone can send Treasury comments, or ask for a public hearing, until November 10.

The two tax breaks, in plain English

Both breaks were created by the 2017 Tax Cuts and Jobs Act (TCJA) and reshaped by OBBBA.

The export break (FDII, now FDDEI). Say a US software company sells licences to customers in Germany, or a US manufacturer ships machines to Japan. Part of the profit from those foreign sales gets taxed at a lower rate than the normal 21 per cent. The idea is to give companies a reason to keep their valuable work, like research, patents and engineering, in the United States instead of moving it to a low-tax country. It used to be called Foreign-Derived Intangible Income, or FDII. OBBBA renamed it foreign-derived deduction eligible income, or FDDEI.

The minimum tax on foreign profits (GILTI, now NCTI). This one points the other way. When a US company earns profits through subsidiaries abroad, it has to pay at least some US tax on those profits each year, even if the money stays overseas. If the foreign country already taxed those profits at a high enough rate, the company gets a credit and owes little or no extra US tax. It used to be called Global Intangible Low-Taxed Income, or GILTI. Now it is called net CFC tested income, or NCTI. (A CFC, or controlled foreign corporation, is just a foreign subsidiary that a US company controls.)

Think of it as a carrot and a stick. The export break is the carrot for keeping profits at home. The minimum tax is the stick for parking them abroad.

What OBBBA did to the rates

Both breaks work through a deduction. The company subtracts a percentage of the relevant income before applying the 21 per cent rate, which lowers its effective rate.

From 2018 to 2025, the export break worked out to an effective rate of 13.125 per cent. The minimum tax on foreign profits was 10.5 per cent before credits. Under the 2017 law, both were set to go up in 2026, to 16.406 per cent and 13.125 per cent.

OBBBA cancelled most of those increases and made the new levels permanent. From 2026, the export rate is 14 per cent and the foreign-profits minimum tax is 12.6 per cent before credits. The law also let companies count 90 per cent of their foreign taxes toward the credit, up from 80 per cent. According to a summary by the law firm Cooley, the result is that a company whose foreign subsidiaries pay tax at 14 per cent or more should, in principle, owe no extra US tax on those profits.

OBBBA also scrapped a quirk that economists across the political spectrum disliked: a 10 per cent carve-out for factories and equipment that, in practice, rewarded building factories abroad and penalised building them at home.

What the two breaks charge, before and after OBBBA

Effective federal tax rate, before foreign tax credits and before expense allocation. Hover or tap a bar.

Tax break2018–2025Scheduled for 2026 (2017 law)OBBBA, 2026 onward
Export break (FDII β†’ FDDEI)13.125%16.406%14%
Foreign-profits minimum tax (GILTI β†’ NCTI)10.5%13.125%12.6%
Source: Cooley LLP summary of the One Big Beautiful Bill Act. The statutory corporate rate is 21%.

Why the costs matter as much as the rate

On the foreign-profits side, the problem is the credit for foreign taxes. A company can only claim credits up to a ceiling. That ceiling is roughly the US tax that would be due on the foreign profits. To work it out, the company has to assign some of its US costs to those foreign profits. Treasury rules that predate 2017 required a share of the parent company's interest payments, head-office overhead, legal bills and similar costs to be assigned that way, usually based on how much in assets or income sat in each category.

Every dollar of US cost assigned to foreign profits shrinks the foreign profits on paper. That lowers the credit ceiling. And credits that did not fit under the ceiling could not be saved for another year. They were just lost.

So companies had a long-running complaint. Their subsidiaries could pay foreign tax well above the advertised threshold, and they would still owe extra US tax, only because their US head-office costs had been pushed into the calculation. A September commentary by the law firm Eversheds Sutherland summed up the mood in its headline: under the new rules, "NCTI almost works the way GILTI was advertised."

On the export side, the problem was a mirror image. The export break applies to a company's eligible income minus the costs "properly allocable" to it. When interest and research costs were subtracted, the income that qualified for the low rate shrank. Exporters that borrow a lot or spend heavily on research, which describes much of American manufacturing, tech and pharma, got a smaller break than the headline rate suggested.

What Congress changed

OBBBA fixed both problems in the law itself, and left the fine print to Treasury.

For the export break, the law now says the eligible income is reduced only by "expenses and deductions (including taxes), other than interest expense and research or experimental expenditures, properly allocable" to it. In other words, interest and research costs no longer shrink the break.

For the foreign-profits credit, the law added a new rule, Section 904(b)(5). When a company works out its credit ceiling, it can now subtract only three kinds of cost from its foreign profits: the deduction for those profits itself, state and local taxes on them, and other costs that are "directly allocable" to them. Interest and research costs are explicitly excluded. Any cost that fails these tests gets moved over to the company's US income instead.

The law never defined "directly allocable". Filling in that blank was the main job of the September proposals.

What Treasury proposed

Summaries by Forvis Mazars, Grant Thornton, BDO and PwC describe five main points.

1. Broad definitions. "Interest" covers anything deductible as interest, and "research costs" covers research spending deducted or written off over time. For the first time, how a cost is labelled directly changes how much a company saves under both breaks.

2. Three buckets. For the foreign-profits credit, every cost goes into one of three buckets. Some always count against foreign profits: the deduction itself and related state taxes. Some never do: interest and research. Everything else counts only if it is directly allocable.

3. A tight test. Treasury reads "directly allocable" narrowly. According to Grant Thornton, a cost qualifies only if it would normally be assigned to foreign profits and is not the kind of cost that gets split up using assets or income as a rough guide. "The inquiry looks to the nature of the deduction rather than to how a particular item of deduction is apportioned," the firm wrote.

So what fails the test? Head-office costs of overseeing subsidiaries, legal and accounting fees, damages, settlements, and general overhead. What passes? Very little: certain currency losses on money brought home from foreign subsidiaries, and some carried-forward business losses. PwC summed it up by saying Treasury "would treat the vast majority of expenses" as not directly allocable.

4. A two-step process. First, a company assigns its costs the normal way, as if the new law did not exist. Then it moves any cost that landed on foreign profits but fails the test over to its US income. Research costs were never assigned to these foreign profits under the existing rules anyway, so in practice that part of the change makes little difference to the credit.

5. A big asterisk. The costs that get moved are treated as US costs "for all purposes" of the foreign tax credit rules. That phrase matters, and we come back to it below.

Companies can follow the proposals now, but only all of them, not just the parts they like.

Two simple examples

These examples are made up and simplified, to show which way the money moves.

The exporter. A US manufacturer earns $1 billion of income from foreign customers that qualifies for the export break. Under the old approach, it had to subtract $200 million of interest, $250 million of research costs and $150 million of other costs.

Old way of counting costs (same new rate)New way under OBBBA
Qualifying income$1,000m$1,000m
Interest subtracted$200mNot subtracted
Research subtracted$250mNot subtracted
Other costs subtracted$150m$150m
Income that gets the break$400m$850m
Deduction at 33.34%$133.4m$283.4m
Federal tax saved at 21%$28.0m$59.5m

The rate did not change. The break more than doubled, because the income it applies to grew.

The multinational. A US company has $100 of foreign profits (counting the foreign tax), and its subsidiaries paid $15 in foreign tax. That is a 15 per cent rate, above the 14 per cent threshold. After the 40 per cent deduction, $60 is taxable, so the US tax before credits is $12.60. The company can use 90 per cent of its foreign tax, or $13.50, as a credit.

Now suppose the old rules assign $20 of head-office costs to those foreign profits. The foreign profits drop to $40 on paper. The credit ceiling falls to 21 per cent of $40, or $8.40, and the company owes $4.20 of extra US tax, even though its subsidiaries paid more than the threshold.

Under the proposed rules, those head-office costs generally are not directly allocable. The foreign profits stay at $60, the ceiling is $12.60, the credit covers it, and the company owes nothing extra. The $20 of costs does not vanish. It now reduces the company's US income instead.

The catch: when moving costs backfires

That last point is where things get complicated. Because the moved costs count as US costs, they can push a company's US income into a loss.

The foreign tax credit rules have a fix for that, and it cuts both ways. A US loss gets set against foreign income in the same year, which can shrink the credit ceiling after all. In later years, the rules generally treat some US income as foreign to make up for it. So part of the benefit gets pushed into the future rather than lost.

Grant Thornton warned that, for affected companies, "the anticipated benefit of removing deductions from the Section 951A category may be significantly reduced."

Who is most at risk? Companies with modest US profits compared with their head-office costs. That includes groups with big foreign operations, lots of US debt, or a year of US losses. For a company with plenty of US profit, the change is close to a pure win.

The case for the rules

Supporters, including most of the advisory firms that have written about the proposals, say they simply carry out what Congress passed. Grant Thornton called the OBBBA changes "ultimately a favorable change" and said "for most taxpayers the result is likely favorable."

They make two main arguments.

First, fairness of design. A minimum tax on foreign profits is supposed to hit profits that are lightly taxed. If a company already pays enough foreign tax, extra US tax that shows up only because of accounting for head-office costs looks like a glitch, not a policy. Congress chose to fix that, and a narrow reading of "directly allocable" does it.

Second, jobs and research at home. Interest and research are mostly US costs. Subtracting them from export income meant the export break shrank the most for companies that borrow and do research in America. Removing them strengthens the reward for keeping research, engineering and manufacturing here, which was the whole point of the export break. Timing adds to this argument. The Federal Reserve raised its target rate to 3.75 to 4 per cent on September 16, its first increase in three years, so interest is a bigger cost for many companies than it was when the law was written.

The case against

Critics push back on two levels: the policy itself and how Treasury chose to carry it out.

The cost. In an October 2025 estimate, the Penn Wharton Budget Model projected that OBBBA's main international changes would reduce corporate tax revenue by $276 billion from 2026 to 2035, compared with the 2017 law's scheduled rates. About half of that, $143.4 billion, came from the export break alone. Penn Wharton said its estimate was larger than the Joint Committee on Taxation's $118 billion for the same provisions. The Tax Policy Center's Brosy put the cost of the international changes at about $170 billion over ten years. Neither group put a separate price tag on the cost-allocation rules, so how much the September proposals add is not known.

Who benefits. Progressive tax groups opposed the international package from the start. In a May 2025 analysis of the House version of the bill, Sarah Austin of the Institute on Taxation and Economic Policy (ITEP) said cancelling the scheduled rate increases would "squander $164 billion over the next 10 years," citing the Joint Committee. Those figures refer to an earlier draft, not the final law. But the objection still applies: rules that make the breaks bigger extend benefits that critics see as poorly targeted.

Profit shifting. Brosy argued that OBBBA does little to stop companies from moving profits to low-tax countries. One reason is that the minimum tax still lets companies blend income and taxes from high-tax and low-tax countries into one calculation, which he called "global pooling". A higher credit ceiling makes that blending more valuable.

Who decides. The law says "directly allocable". It does not say "only if not split up by assets or income". By ruling out whole categories of cost, Treasury has arguably picked the most company-friendly reading of a term Congress left undefined. Supporters reply that the phrase was added specifically to keep costs out, and a stricter reading would undo that.

What it means for the budget, the economy and everyone else

The budget. The big revenue cost was locked in when OBBBA passed. These rules matter at the edges: a narrow definition of "directly allocable" costs the government more than a broad one would. No official estimate isolates that, so the size is uncertain.

The economy. The rules lower the tax on income US companies earn by exporting from the United States, especially for companies that borrow and do research here. That is the intended effect. Whether it changes where companies actually invest, rather than just lowering their tax bills on decisions already made, is an open question. The Tax Policy Center estimated that the export break cuts the average effective tax rate on profitable export investments by about 2 percentage points.

Who gains. The direct winners are large US multinationals and big exporters. Most official and academic models say the bulk of a corporate tax cut ends up with shareholders, who are concentrated among higher-income households, with a smaller share going to workers.

The politics

The rules land in a busy autumn. Government funding runs only until December 11, under a stopgap bill signed on September 2. A year-end package is expected to carry a list of bipartisan tax extensions, and Republicans have also floated a second party-line bill in the lame-duck session. Neither has so far been tied to the export break or the foreign-profits minimum tax, but both are reminders that the rules can change.

Democrats have generally described OBBBA's international changes as a tax cut tilted toward big corporations. Republicans have presented them as a way to help US companies compete and to keep intellectual property and jobs at home. The allocation rules themselves have drawn little political attention. For now, the fight is among tax advisers and companies, and it will play out in the comment letters due on November 10.

What to watch

The comment letters, due November 10. Business groups are likely to push Treasury on three things: how the test applies to costs that serve both US and foreign operations, how moving costs interacts with the loss rules, and whether Treasury will update the loss rules it says it is still studying.

Labelling fights. Because interest and research now get special treatment under both breaks, expect more audit arguments over what counts as interest and what counts as research.

For companies, advisers agree on the practical steps. Sort your costs into the three buckets. Work out the benefit to your credit ceiling alongside the effect on your US income and losses. Then decide whether to follow the proposals for 2026, remembering that it is all or nothing.

This article is analysis for general information and is not tax, legal or investment advice.

Sources