Our corporate tax system is our best so far—don’t break it

International corporate taxation is a difficult puzzle. The 2017 GOP reform got a lot right.

Photo by qimono.

Part of the funding for the Biden Administration’s fiscal package can be described simply: raise corporate taxes. The most recent proposal from House Democrats would raise the headline corporate tax rate to 26.5 percent from a previous rate of 21 percent.

But the simple description also leaves a lot out: increasing globalization and digitization has made corporate income more and more difficult to apportion between jurisdictions. Global corporate income is a bit squishy or squirmy, especially when it comes from “intangible” sources like patents or brands or ideas. When a company operates subsidiaries in multiple countries, it can fairly easily use accounting tricks to shift costs and revenues from one jurisdiction to another. If the U.S. raises its corporate tax rate unilaterally, companies will try to shift more of their profits overseas, and the US might not get as much tax revenue as it is hoping for.

Things might be easier if the U.S. government could forget about national borders and just tax companies on their global income. That might work well if every global company was headquartered in the U.S. The problem is that many important companies are based abroad. If U.S. law taxes U.S. companies—but not foreign companies—on their global income, that would put them at a systematic disadvantage compared to French or Japanese or Korean companies. In the long run, it might encourage more companies to shift headquarters and other operations outside the U.S., harming the US economy.

So if you are going to design a corporate tax system in the 21st century, you have to design the international provisions very carefully and handle many different permutations of international taxation. When Republicans overhauled the tax code in 2017, they added several new provisions designed to discourage the most egregious accounting shenanigans and encourage companies to recognize more of their profits in the United States.

These provisions are all subject to debate or reform⁠—but I would not be too eager to change a system that works relatively well, given the extraordinary complexity of the problem. The current U.S. corporate tax system recently reached a record high in revenue, at least in nominal terms, despite the 14-percentage-point rate cut in 2017. The framework from the Republicans’ 2017 tax reform is the best one the U.S. has had so far.

As we will see, global tax negotiations are likely to create space for a higher tax rate on globally-mobile intellectual property. That should give Democrats an opening to raise some extra revenue without upsetting the careful balance struck by the 2017 law. But Democrats should be cautious about more aggressive changes that could easily backfire.

A basic corporate tax trick—and a provision that curbs it

Corporate income is, roughly, revenue minus expenses. Add up your revenues, deduct your expenses, and that gives you a first pass at your corporate income. But this gets trickier when a company operates in multiple jurisdictions, and has to pay taxes to each of them.

Say that there is a company with a presence in both the U.S. and Ireland, for example. You have to account for the transactions between the U.S. and Irish parts of the company. If the Irish part of the company sells components to the U.S. part of the company, that is a deductible expense against the U.S. corporate income tax rate, currently 21 percent, and revenue for the Irish part of the company, taxed at 12.5 percent (though a hike to 15 percent has been announced).

Here is the problem: every time the U.S. portion of the company pays an additional dollar to the Irish portion of the company for something, it subtracts 21 cents from its U.S. tax burden and only adds 12.5 cents to its Irish tax burden, saving 8.5 cents.

The company is therefore incentivized to overstate the Irish contribution and understate the U.S. contribution.

This strategy won’t always work: if you are selling a tangible product like Irish whiskey, you can’t plausibly attribute the value of the whiskey to the Irish subsidiary unless it’s really brewed there. But intellectual property, like software, is harder to pin down.

A software company’s U.S. component might earn a bunch of revenue in the U.S., but then say to the IRS something like this: “actually, the Irish part of our company did all of the hard work on this idea, and we need to pay royalties to the brilliant Irish subsidiary for all of the wonderful work they did. After accounting for the payment to our lovely counterparts, it turns out we really didn’t make very much money at all.” This is a problem for U.S. tax authorities.

They can use subjective judgments and laws to curb this behavior, but an automatic solution is a good backstop. In 2017, Republicans introduced BEAT, the Base Erosion and Anti-Abuse Tax. It kicks in if companies try to attribute too much of their profits to overseas subsidiaries.

It is a kind of alternative minimum tax. It names the most abusable cross-border deductions, and asks you to calculate your tax again, at a lower rate of 10 percent but without any of those deductions. You then owe the higher of the two numbers.

BEAT would never be a good tax base if it were our only tax base; it eliminates too many legitimate expenses in an attempt to go after illegitimate ones. If applied too broadly it could easily be considered akin to a tariff, increasing taxes on real imports. But BEAT only goes after the firms that are pushing their luck and relying too heavily on the most abusable cross-border tax deductions; in that context, it is acceptable to institutions like the World Trade Organization (WTO.)

The BEAT rate is already set to increase, even per existing law. (Republicans did this to improve the long-run revenue raised from their 2017 bill.) But Democrats may strengthen it further, or perhaps overhaul it with a related proposal called SHIELD (Stopping Harmful Inversions and Ending Low-Tax Developments).

How we tax U.S. firms’ ideas when they earn revenue abroad

The same gambit I described for BEAT—where income gets “shifted” to low-tax jurisdictions through abuse of cross-border transactions—could also be used when U.S. firms sell their goods abroad.

Consider a U.S. tech firm that has long headquartered itself in Silicon Valley, and clearly employs most of its best and highest-ranking employees there. Say that it expands internationally, and starts selling its software in Australia, which has a relatively high corporate tax rate of 30 percent. But instead of attributing its income from those sales either to Australia or to the U.S., it claims that much of the income was earned by patents held in a subsidiary in a small island country with little or no corporate tax rate, like Barbados.

This is obviously worthy of suspicion. It’s plausible that the value added to the company came from the Australian sales team. It’s plausible the value added came from the software developers at the Silicon Valley headquarters. But it is not plausible that the value added came from a tiny island where the company has minimal facilities.

This kind of income⁠—income in excess of the plausible returns on your physical capital invested⁠—was defined in the 2017 tax reform as Global Intangible Low-Taxed Income. Unless you pay at least a 13.125 percent rate on it abroad, you owe U.S. tax on that income to bring your rate up most of the way to that 13.125 percent threshold.

It might be tempting to increase this rate: clearly some U.S. companies under-attribute the value that their U.S. headquarters add to their global success, and that results in missing tax revenues. But at the same time, it’s important to note that companies without a U.S. headquarters don’t pay any extra U.S. tax on GILTI.

A global minimum

The regime therefore puts U.S. companies at a competitive disadvantage relative to foreign rivals, unless the foreign countries also add global minimum taxes. This has been one reason that very few countries besides the U.S. have ever had the confidence to tax global profits for their national champions.

Of course, the U.S. also isn’t the only jurisdiction that’s annoyed by this kind of behavior. Every large industrialized country would like to discourage these accounting gimmicks. But they face a collective action problem: if they move unilaterally, they will put their own national champions at a disadvantage.

So in recent years, the Organisation for Economic Co-operation and Development has been developing an international framework to set minimum standards for corporate taxation. The second pillar of this deal would require all countries to impose a minimum tax at a rate of at least 15 percent. The parameters of the income subject to the OECD minimum are not defined the same way as the U.S. defines GILTI, but the existing system could be modified to be closer in line with the international treaty.

“It is clear that the global minimum tax will be much narrower than GILTI,” the American Enterprise Institute’s Kyle Pomerleau tells me. It has more generous deductions in many respects, and it only applies to the largest corporations, while U.S. taxation of GILTI currently applies more broadly.

The OECD framework will not cure global tax competition; it will only alleviate the pressure for very low rates. Given the progress made on this agreement, it could make sense to raise the GILTI rate to 15 percent and shift to a treaty-compliant base. This would generate more revenue without harming competitiveness.

How we incentivize bringing intangibles home

GILTI is a stick the IRS can use to beat companies that try to shift profits to tax havens. The U.S. system also offers a carrot to companies that shift profits in the other direction. Consider again the Silicon Valley-based firm selling its software in Australia.

If the firm admits that most of the value added is attributable to U.S. ideas, it gets rewarded with a 13.125 percent rate on foreign-derived intangible income (FDII), similar to the rate the U.S. GILTI provisions encourage. In other words, because the FDII and GILTI rates are the same, one might as well admit that the developers in California are the ones responsible for generating the ideas.

What worries me here is that many Democrats don’t seem to see the connection between these two rates. They are eager to limit or repeal FDII, which would subject intangible income to higher taxes if it is attributed to the U.S. than if they are attributed abroad. This will backfire, precisely because that income is so squishy and mobile.

Economists at the Penn Wharton Budget Model are watching the differential between GILTI and FDII rates. When that Silicon Valley-based firm sells its software in Australia, it is incentivized to pretend its best ideas came from outside of the U.S. if it gets a global rate lower than the FDII rate. This is what happens under the House Democrats’ proposal: “We estimate that these reforms would increase the incentive to shift intangible investments and profits recorded by U.S. firms to other countries with a tax rate below 20.7 percent,” write Penn Wharton’s Alexander Arnon and Zheli He.

FDII rates should instead remain coupled with the GILTI or OECD minimum rate on intangible income, in order to make sure firms are incentivized to report intangible income to the U.S. where appropriate.