Two conflicting ideas underpin a lot of the conversation surrounding businesses and the federal income tax. First, that our current tax code is too complicated, and second, that we should use the tax code to manage economic policy. Cost recovery of capital investment is a good case study. It’s really complicated and it has huge economic effects.

What people mean by saying we should make the tax code less complex is usually that we should have a broader base and lower rates. Cut less stuff out of what is actually taxed, and then less will be needed from that tax base. These cutouts are deductions—called loopholes when they are being criticized. They are conceptually simple to address, and grindingly difficult to undo politically. Business deductions, which are permitted to be taken below the line under §162 of the Internal Revenue Code, permit the regular cost of doing business to be deducted—salaries, rents, and the like. Business expenses are separate from capital expenses, the assets you purchase in order to go on with your business. These are often called “tangible” property, and like a lot of the tax code, the distinction has a qualitative nature that seems simple at first and eventually leads to a long, specific list. Unlike business expenses, which can be deducted as a whole in the taxable year in which they can occur, assets are deducted in chunks over time. This is called depreciation deduction, or cost recovery, and it’s unbearably complicated.

I’ll keep it short, and no one has to do any math.

Once you purchase a piece of equipment or other tangible property, like a horse, or a tractor, you deduct a certain amount of its value per year for a span of years, called its class life. Class lives are determined by the type of item, and it’s a back-of-the-envelope evaluation of how long that item is going to be useful to you. A truck is good for five years, and so is a computer; an Alaskan natural gas pipeline is good for seven. Business property gets 39 years, and residential property slightly less. Class lives are found in §168 of the Code, along with the other key parts of the math equation: when you start subtracting in the year, and how you do the division. I’m going to mostly skip that bit because there’s not much payoff, but for longer-term property classes, you deduct the same number every year, and for shorter term, you do more math and take more money out earlier (called the double declining balance method). At the end of the class life, you’ve deducted the full cost of the asset.

Okay. This is complicated, but why does it matter?

Because a dollar’s value moves in time. If I give you 20 dollars today, that is literally worth more than 20 dollars in five years. Or, to map this on to the depreciation apparatus, if I give you four dollars a year for the next five years, that is still worth less. This is called the discount rate, and it means it’s impossible to fully recover the cost of capital assets. As the cost is deducted each year, what that fixed dollar amount is actually worth goes down, because the economic value of that set number is less in the future than if you had it now. In a paper for the Tax Foundation, Scott Greenberg argues that overall, U.S. corporations only get back 87.14% of what they spend. Business investment is literally not worth it.

And business investment—long term buy-in to durable assets—has long been considered an engine of economic growth. The stuff that would help us economically has disfavored tax treatment. Now, that one will obviously follow fixing the other is a simplistic claim, and I’m not making it. It’s really hard to predict what business incentives will actually lead to increased growth, rising wages, and so forth. For example, though some studies show the cost of the corporate income tax is fully borne by workers, not all of them do, and the disagreement about whether lowering the tax would have benefits that do get passed on to workers falls largely along partisan lines. If it was simple, we wouldn’t all be fighting about it so much.

But there is some evidence that depreciation has a relationship to capital investment. Bonus depreciation existed from 2001 to 2004 and was reinstated in 2008. It allows businesses to deduct a full 50% of certain investments up front, and provided a case study for a paper by Erick Zwick and James Mahon. They argue bonus depreciation did raise capital investment, by 10.4% from 2001-2004 and 16.9% from 2008-2010. This is why the Republican tax plan contains a provision for full expensing.

What would full expensing do?

By permitting full deductions of capital assets for businesses in the year in which they are purchased, the argument goes, businesses get full cost recovery, which will in turn result in more investment, which will in turn boost the economy. The Republican plan only permits full expensing for investments that aren’t “structures,” continuing the bifurcation of federal income tax as applied to real estate and federal income tax as to everyone else. First, their plan would be significantly simpler than what we have now. Depreciation deductions are hard to wrangle, especially with the class lives that extend into decades, and the administrative burden is significant. Second, there would be a big price tag—up to $2.2 trillion dollars over the next decade, though a lot of that is the transition cost of shifting to a new framework. And as the Atlantic noted in August, it splits the business coalition, because full expensing benefits different businesses than those who want a much lower corporate rate.

In the end, tax policy is just policy. The distortions that full expensing would introduce into the economy would theoretically result in economic growth due to the incentives to make long term investments. Is that growth enough to make up for the trillions in lost revenue? That question is a political one masquerading as an economic one, at least in part because measuring these kinds of effects is hard.