Last year’s giant tax bill established something unprecedented in the history of the income tax, which sounds dryly technical and thus unobjectionable. It’s called permanent “expensing,” and it allows all businesses to immediately deduct the cost of investments in almost all tangible property. But behind this dry financial jargon is a profoundly expensive and regressive change.

Expensing permanently privileges capital over labor when viewed in the context of our broader federal tax system. It effectively exempts normal business profits from tax if they stem from investments in capital. These business profits themselves are heavily concentrated among the very wealthiest. It inefficiently creates tax subsidies that can magically turn debt-financed investments that are unprofitable before taxes into investments that are profitable after tax.

Expensing costs on the order of $2 trillion over 10 years compared to something closer to economic depreciation, which was the law for most of the income tax’s history. This is almost 5 percent of federal income tax revenues. Such a massive investment would be far better spent on other priorities, whether expanding access to health insurance, reducing child care costs, or addressing the growing national debt, to name a few.

Surprisingly, many policymakers on both sides of the aisle support expensing. Supporters believe it will increase investment in the United States and thereby economic growth. But even if one’s goal is purely to increase business investment in tangible U.S. property, there are probably more cost-effective ways to do so.

Lawmakers should repeal this provision and return to the tax code’s prior, long-standing approach under which large businesses could deduct only the annual cost of their tangible investments through depreciation deductions that approximate assets’ decline in value.

Before Congress started tiptoeing towards permanent expensing about 25 years ago, here’s how it worked for large businesses. If a large toy company spent $100 million of its pre-tax profits on machinery that could produce wooden toys, it could not deduct that amount immediately on its tax return. Instead, each year, it could deduct a small portion of the cost that roughly matched the depreciation of those machines. Now, after the passage of the One Big, Beautiful Bill Act, it can deduct all $100 million, immediately.

Part of the reason why this huge tax cut has flown under the radar is that it is hard to understand its real-world effects.

The first important effect is that expensing an asset is like allowing businesses to pay no tax on all of the so-called normal profits that asset produces. Such income is heavily concentrated at the top. Capital income is two-thirds of all income of the top the 0.1 percent, but only 6 percent of all income of the middle class and low earners (the bottom 80 percent). While normal profits are a subset of capital income, they are also heavily concentrated at the top. Expensing is thus extraordinarily regressive.

Tax policy experts all agree on the following central principle, which is taught in any introductory tax class: Allowing taxpayers to immediately deduct the cost of an investment that produces income over time is, under certain assumptions, equivalent to exempting all of the income from that investment from tax.

While this is not intuitive, it holds for expensing as well.

To see why, imagine that corporation investing $100 million of its pre-tax profits in some machines to produce wooden toys. Suppose these machines will produce a 100 percent return over two years. If the corporation can deduct the cost of the machines, it can spend $100 million to earn $200 million over two years. If its tax rate is 25 percent, it will have $75 million of after-tax profits.

Likewise, if it has to pay tax on its $100 million of pre-tax profits, it will have $75 million to invest in the machines. These machines will produce $150 million of toys and—critically, if its profits are tax exempt—it too will have $75 million in after-tax profits.

The fact that the large corporation ends up with the same amount after-tax means that allowing the corporation to immediately expense the cost of machines is the same as exempting all their normal profits from those machines from tax.

During much of the income tax’s history, companies could deduct only the cost of assets—whether machines or buildings or computer software and hardware—through depreciation deductions over time, which were meant to very roughly proxy their annual decline. Only very small businesses were eligible for expensing. But Congress gradually moved toward full expensing in recent decades, steadily expanding what kinds of assets and businesses are eligible for immediate write-offs.

Partial expensing for larger businesses was first enacted in the second Bush Administration to address recessionary concerns after September 11, 2001, but it was temporary and limited to short-lived assets. Since then, Congress has periodically reinstated temporary and limited expensing for larger businesses, often as a temporary stimulus and typically under full or partial Republican control.

Last year’s bill changed all this. Congress made 100 percent expensing permanent for the largest businesses, and expanded its reach to even longer-lived assets. Now, the largest multinational companies can expense all assets expected to produce income for up to 20 years, and many assets expected to produce income for much longer, including some real estate.

One might ask, what’s the big deal? Businesses can also immediately deduct the cost of wages and salaries. It is true that, all else equal, expensing would equalize the tax treatment of capital and labor investments.

But all else is not equal. While wages and salaries are deductible for businesses, they are then taxed at the individual level when workers pay tax on their wages and salaries. Less than 30 percent of corporate income is actually taxed at the individual level. When business income is taxed at the individual level, it is typically taxed at much lower rates than wage income due to preferential rates for capital gains, dividends, and pass-through business profits. The income tax is filled with other tax incentives to invest in capital, whether the research and development tax credit or deductions for oil and gas exploration that exceed their costs. Moreover, around one-third of federal revenues are derived from the payroll tax, which applies to labor income but exempts capital income. Our tax system was thus already heavily biased towards investments in capital. Permanent expensing further tilts the playing field away from workers and towards investments in capital.

If this sounds bad, consider a second problem. Last summer’s bill also violated the closest thing there is to a bipartisan article of faith among tax policy wonks: Expensing should never be coupled with interest deductibility.

When businesses can deduct interest on debt used to finance investments eligible for expensing, they are taxed at negative rates on the income produced by the asset. Essentially, the government is writing them a check to make the investment. Policymakers and experts as diverse as Paul Ryan, Kevin Brady, Alan Auerbach, Jason Furman, Douglas Holtz-Eakin, and the George W. Bush Tax Reform Panel all agree with this should not be allowed. Even President Donald Trump’s first campaign acknowledged this principle, proposing full expensing only if a company gave up the ability to deduct interest payments.

These negative tax rates also are not intuitive, so here’s another example.

Let’s say a different large corporation doesn’t have access to the machines described above and is instead considering buying some mediocre machines. They produce wooden blocks that quickly disintegrate or are rough and give kids splinters. As a result, if it invests $100 million in these machines, they will only produce a 20 percent return over two years, or $120 million in sales. The corporation also doesn’t have cash to invest and can only borrow at a 12 percent annual interest rate.

This is a dumb investment before taxes: Borrowing at 12 percent to earn a roughly 10 percent annual return.

But suppose the corporation could deduct its interest payments? And, to simplify the example, suppose it can’t deduct the machines upfront, but all profits from the machines are tax-exempt? (As explained, this is equivalent to allowing expensing.)

In this case, the corporation could magically convert a dumb pre-tax investment into a smart after-tax investment, thanks to benefitting from a negative tax rate.

It would owe no tax on the $20 million in profits from the machine. But it could deduct $24 million in interest payments against its other income. If still taxed at a 25 percent rate, these deductions would be worth $6 million. So it would end up with $2 million in after-tax profits, even though it was losing $4 million before taxes. In effect, its tax rate on its -$4 million in profits would be negative 150 percent.

This hypothetical negative tax rate may sound extreme but it’s not far from reality. The nonpartisan Congressional Research Service estimates that the effective tax rate on debt-financed corporate investment in equipment is now negative 21 percent, and in intangible assets like AI models it is now negative 59 percent.

These negative tax rates are why tax experts on the left and right agree that allowing businesses to expense assets and also claim interest deductions is a horrible idea.

Unfortunately, a majority in Congress does not. At the same time last summer that Congress permanently allowed all businesses to deduct 100 percent of their investments in most tangible assets, it expanded their ability to deduct interest payments.

These tax subsidies for debt-financed investment are inefficient and unfair. They apply more to capital investments than profits generated by hiring workers, because it is easier to take out loans secured by tangible assets. They can result in negative tax rates not just on normal returns to capital investments, but also on unusually high returns. They only apply to debt-financed ventures, and thus encourage corporations to become more leveraged, generating risks for the economy more broadly.

Other countries have long recognized the senselessness of such subsidies. Many raise a significant share of their revenues from value-added taxes (VATs), which can be progressive if used to fund more robust safety nets. VATs allow businesses to expense investments. But they do not allow businesses to deduct interest payments at all.

Some, including people I deeply respect, think the best path forward is to disallow all business interest deductions and further expand expensing to all assets (such as buildings, land, and securities), rather than repealing expensing. This approach is called a “cash-flow tax.” While it would be theoretically better than the current situation, especially if business taxation then raised more revenues than current law, the arguments in favor of such an approach tend to ignore three key issues.

First, it is highly unlikely that Congress will have the political will to bar businesses from deducting all interest costs. Businesses have been able to deduct interest since the inception of the income tax more than 100 years ago. Not a single member of Congress has proposed eliminating interest deductions since permanent expensing became law. To be sure, repealing expensing would be a heavy lift politically. But not nearly the magnitude of disallowing all interest deductions.

Second, a cash-flow tax could undermine multilateral cooperation on limiting profit shifting. In 2021, more than 135 countries reached a historic agreement to implement a series of interlocking minimum taxes that would largely eliminate the incentive for large multinational enterprises to shift their profits on paper in low-tax jurisdictions. These minimum taxes now cover roughly 90 percent of multinational income. A cash-flow tax would, at a minimum, complicate implementation of this carefully constructed agreement because it doesn’t exist in any country.

But it could also be viewed fairly as a unilateral effort by the United States to get credit for taxing multinational profits at higher rates than we actually do, because of how the agreement operates. If this unraveled or destabilized the multilateral agreement, the net result would be more profit shifting to low-tax countries, thereby reducing tax rates on the largest multinationals when they should be increasing instead.

Third, analysts in favor of expensing tend to argue that it will increase capital investment in the United States, which will spur economic growth and can increase worker’s productivity and wages if that capital is a complement to their work.

Some capital investments do improve the productivity of workers, thereby justifying higher wages. But that is not always the case. Whether AI will make workers more productive or simply replace their jobs forcing workers into lower-paying jobs is hotly debated, for instance.

Moreover, even if one’s sole goal is to increase long-term U.S. capital investment, there are other reforms that are likely to have more bang-for-the-buck than expensing due to how large corporations actually make decisions about whether to increase capital investment and locate it in the United States.

Traditional economic models tend to assume that the CEOs, investors, creditors, and analysts of large businesses all have perfect information, and their incentives are all aligned on maximizing the present value of the firm’s long-term, after-tax profits. Unfortunately, none of these assumptions are correct, especially in the tax context.

Instead, more often than not, large firm managers incorporate taxes into their investment decisions through proxies for their tax rate that treat the value of expensing as zero. The most common are the firm’s statutory tax rate and “book” tax rate. A firm’s statutory tax rate doesn’t capture the value of all the deductions, exclusions, or tax credits that firms can and do claim, including expensing. The rules for calculating a firm’s “book” tax rate (its estimated tax rate as reported in its public financial statements) mandate that firms ignore the value of claiming deductions earlier in time, which is the entire benefit of expensing.

When CEOs make U.S. investment decisions based on these tax rates that disregard the (very large) value of expensing, that means expensing generates large windfalls for such firms, but is completely ineffective at increasing U.S. investment.

While this ineffectiveness may lessen concerns about expensing spawning a shift from labor-intensive to capital-intensive business models, it makes no sense to provide huge windfalls to large firms, let alone windfalls that rise with how capital-intensive the corporation is.

To the extent that CEOs focus on the “book” tax rate when making investment decisions, ironically it also means that a cash-flow tax could worsen tax incentives for U.S. investment. A cash-flow tax would allow expensing for all investments, including real estate and land, in exchange for disallowing interest deductions. The “book” tax rate would treat this even more robust expensing as having no value, while recognizing the disallowance of all interest deductions as a large tax increase.

To be sure, not all CEOs make decisions in the same way, nor do all disregard the value of expensing. But the best study on this subject to date found that 44 percent of large firm managers primarily focus on their statutory tax rate when making investment decisions, 41 percent on their book tax rate, and only 13 percent on their time-discounted marginal tax rate. Traditional economic models that tout the benefit of expensing for U.S. investment assume that 13 percent figure is 100 percent.

This means there are likely far more cost-effective ways to stimulate U.S. investment than expensing. Relying on the same study, I estimate that, under reasonable assumptions, one could create the same functional tax incentives for large businesses to invest in the United States by adopting other tax reforms with 40 percent of the cost on a present value basis, thereby saving almost $1.3 trillion in revenues. The reason is that these other reforms, whether a lower corporate tax rate or an investment tax credit, reduce corporations’ book and/or statutory tax rates, which are more pertinent in their investment decisions, while expensing does not.

It’s clear that $1.3 trillion is a lot of revenue to waste by inefficiently trying to stimulate U.S. investment. Moreover, many non-tax policies may be more cost-effective at increasing long-term U.S. investment or growth than any of these options, such investing in education or infrastructure.

Curious readers may ask why CEOs and other executives of giant firms disregard the value of expensing so often? Don’t they have the resources and motivation to consider the tax consequences of investments in detail? There are at least four reasons why.

First, investors and analysts have no way to calculate the value of expensing for a company, and CEOs care what these people think.

The tax returns of all businesses, including the largest multinationals, are not public. Publicly traded companies are required to issue financial reports. But financial regulations have long accorded them wide deference in what they report about taxes. As a result, even among publicly traded firms, the information corporations report about their taxes is sparse and inconsistent across firms. There is no way analysts can use it to calculate the value of expensing for the firm.

Publicly traded firms are required to publicly report their “book” tax rate. But, as noted, this tax measure ironically assigns zero value to expensing. This is because the financial accounting rules (both within the United States and globally) have long disregarded the time value of money, which is the whole reason why accelerating deductions and deferring taxes are so valuable.

Thus, even among publicly traded companies, analysts and investors have no way of calculating the present value of the firm’s tax liability on a marginal investment. This means they have no way of knowing what the firm’s time-discounted marginal tax rate is and incorporating it into their views of the wisdom of the corporations’ investment choices.

Second, CEOs and managers at large companies care what investors and analysts think, because their views determine the firm’s share price. The firm’s share price, in turn, affects these executives’ compensation, whether through stock grants, options, deferred compensation, or effects on their reputation.

Relatedly, CEOs care deeply about their firm’s reported earnings per share. Executive compensation and performance are typically benchmarked to this metric, as are analysts’ investment recommendations. But earnings per share only incorporates taxes through the “book” tax rate, which ignores the value of expensing entirely.

Third, a firm’s after-tax book income can often have real economic consequences for the firm. For example, contracts with lenders will often specify that the interest rate the corporation must pay will increase if it does not meet certain book earnings targets. Thus, even CEOs whose interests are aligned with maximizing the firm’s long-term value may rationally focus on maximizing after-tax book earnings, despite this metric disregarding expensing’s value.

Fourth and lastly, taxes are complicated, even for large corporations. In order to act in line with the assumptions of traditional economic theory, firm managers would need to continuously model the constantly changing deductions, credits, and rule changes in each national and sub-national jurisdiction in which they operate. Numerous studies have found that in the face of such complexity, many executives turn to simple proxies, like statutory tax rate differences, when making investment decisions.

This raises a second weakness of traditional economic models finding that expensing will increase long-term U.S. investment: They disregard the international context in which large corporations typically make investment decisions. Specifically, these models incorrectly assume choices about whether to invest marginally more in the United States are limited to investments generating “normal” returns, which are the returns that expensing exempts.

Economists distinguish between normal business returns and so-called rents, which are unusually large returns. Theoretically, a company will keep investing more as long as its investments will generate a higher return than its cost of capital, so its marginal investments will be those producing normal returns and not rents. According to long-established theories, a pure cash-flow tax would exempt normal business returns from tax, but would tax rents at a higher rate, holding revenue constant. Cash-flow tax proponents therefore argue that such a reform would increase U.S. investment because the tax on marginal investments, which all produce only normal returns, is zero.

But if one’s objective is to increase U.S. investment, it doesn’t matter whether a multinational increases its total capital investment globally, but whether it increases its investment in the United States. The location of investment is the margin that matters, not the amount.

From this perspective, many marginal U.S. investments do generate rents, contrary to the assumptions of cash-flow tax proponents. A multinational may decide to locate an investment generating rents in the United States or another country for many reasons, but one may be taxes.

Compared to an income tax with economic depreciation, a pure cash-flow tax may or may not increase the amount of such highly-profitable investments made in the United States. It will depend on the responsiveness of locational investment decisions to tax, the extent to which rents are larger in the cross-border context, and other factors. But, holding revenues constant, a pure cash-flow tax will increase the statutory and book tax rate on such highly-profitable investments, further undercutting the argument that a cash-flow tax will increase U.S. investment.

In the real world, though, these theoretical dynamics are a bit beside the point. Congress has enacted expensing at a tremendous cost without paying for it. It has simultaneously expanded interest deductibility, moving business taxation even further from a cash-flow tax. The practical effect is a huge tax cut for both normal returns and rents, and one that varies widely by the type of investment and whether it is debt-financed.

To be clear, any $2 trillion tax cut for businesses will increase U.S. investment, and there is evidence that expensing does. Some find expensing is more cost effective at increasing U.S. investment than a statutory rate cut in the first few years; others that an investment tax credit or statutory rate cut are equally or more cost effective. The evidence is stronger that expensing boosts investment among small firms and in the short-term, rather than increasing investment on net in the long-term. But the question policymakers and researchers should focus on is not whether a $2 trillion tax cut increases U.S. investment at all.

The questions they should ask are: Does expensing increase U.S. investment as cost-effectively as other policy alternatives, including non-tax options? Is it fair or efficient to provide large businesses with net subsidies for debt-financed tangible investments? And should our tax system as a whole put a thumb on the scale in favor of capital investments instead of hiring workers, irrespective of the magnitude of the effects?

Our old system of economic depreciation and full interest deductions arguably was equally or more effective at increasing U.S. capital investment, holding revenue constant. It did not create as large subsidies for debt-financing. Meanwhile, raising taxes on large businesses by repealing expensing in order to pay for investments in public infrastructure and human capital could generate comparable or larger growth effects without placing an even heavier thumb on the scale for capital over labor.

Through this lens, expensing is not worth the expense. There are cheaper and better ways to accelerate economic growth and ensure that it is widely shared.