The Republican tax bill is going to conference, and updates are happening so quickly they are time stamped to the minute. So, let’s talk about the individual Alternative Minimum Tax, a decades-old smoking gun demonstrating the failures of the federal income tax.

(There is also a corporate AMT, which is not the subject of this post, because A. it’s complicated and B. no one knows what is going to happen to it.)

What is the AMT?

Congress first enacted the Alternative Minimum Tax in 1969. Secretary of the Treasury Joseph Barr testified before the Joint Economic Committee that 155 people making over $200,000 a year had no federal income tax liability on their 1967 tax returns. People were very upset about this! That’s not surprising—adjusting for inflation, that’s over $1.3 million dollars in annual salary. 

In response to this PR nightmare (Congress famously received more complaints that year about those 155 than about the Vietnam War) the Tax Reform Act of 1969 created the AMT. It is basically a shadow framework for people who were beating the system. Normal tax calculation is as follows: you find your adjusted income, calculate your tax rate, take out your deductions, and pay what you owe. The Code preferences certain habits and behaviors by either letting you take that money out of your income (thus reducing your rate AND what you owe), or deducting it from your liability (just reducing what you owe). There is a laundry list of available deductions, and an entire cottage industry that exists just to help people calculate them.

The AMT math clicks in right after you’ve subtracted your deductions from your income. It takes that number (called your Taxable Income, because it’s actually what is taxed) and undoes a stack of deductions. It’s reconstructing the normal tax liability calculation as if most of those preferences didn’t exist, such as miscellaneous itemized deductions, ISOs, net operating losses, state and local income tax, and so forth. Then there’s a big exemption—$84,500 for married filing jointly—which phases out at a certain threshold—$160,900 for the same. It’s a ramp-up, followed by a cliff. There are two brackets: 26% and 28%. All of this combines to take aim at a certain kind of rich person: they make a lot of money, and they take a LOT of deductions. Over time, the AMT applied to more and more people, because it wasn’t indexed to inflation until 2013. That’s been corrected, but it’s still a lot of taxpayers: just under 4 million households paid it in 2016.

What’s the issue?

The AMT is strictly a rich person’s problem, which may explain why it’s regularly criticized but never actually repealed. In the most recent go-round, the Senate bill retains it, though as they’ve raised the exemptions and phase-outs it’s unclear who would actually pay it. The AMT is kind of a blueprint for a certain type of tax reform, though since it only covers high-income earners with a high ability to pay, it’s more thematic than realistic. But it does a lot of what some reformers want: broad base, lower rates, simpler math. The reason it bears criticism is because it’s a backup plan.  

First, it’s complicated: individuals who are targeted have to do two full calculations of their income tax liability—one the regular way, and one under the AMT. The AMT’s variant tax preferences require a lot of calculating. Even though only 3.9 million households paid it in 2013, nearly 10 million had to do the math. Compliance costs are a huge burden, which is why those who agitate for tax reform bang the drum for simplicity. Though, as we are seeing play out in real time, that’s easier said than done.

Second, it’s evidence of a system failure. Taxation compliance costs are necessarily baked into our system, because the various deductions are policy, not revenue-driven. Even taking that as a fact—even taking it as a normatively positive fact—the AMT shows the internal fractures. The existence of the AMT is philosophically at odds with the justifications for all that complexity. That reasoning is: yes, it’s complicated, but it’s worth it. These policy preferences—this carrot and stick of using money to push certain kinds of human behavior—are key, so all of the math is in service of something. The AMT admits that’s a total fallacy, because the end result of these policy preferences can have consequences that are bad. Instead of addressing those consequences, or, perhaps, confronting what a mess it is to be doing policy through the tax code anyway, the AMT is just a secret trapdoor to get money out of people that the original tax code was designed to return, because returning money is a mechanism for controlling human behavior.

There’s a line in a children’s book—bear with me on this—where snails are getting into the post office because glue on the stamps is cabbage flavored. So what does the postmaster do? Put toads in the letter boxes to keep down the snails. If you carry that logic through, pretty soon you’re going to have a full ecological system inside your post office (in the world in this book, the post office is still a reliable way to send and receive mail). That’s the AMT.

If you take as a predicate that income tax liability should mirror people’s ability to pay, then the absence of the AMT would lead to a bad outcome. I don’t think that justifies the existence of the AMT from a policy perspective, though. It doesn’t cure the illness at the heart—that our forest of deductions is confusing, counterintuitive, and ineffective.  The purpose of a tax code such as ours, which is in the business of picking winners and losers, if it is not conceptually bankrupt, must then be legitimated through the outcomes of those winners and losers. That is to say, if we’re going to have this dumb system, it should work.

The AMT stands for the claim that some codified legal tax preferences are loopholes and, worse, are understood in the tax code as such. That stacking these different preferences, while legal, intelligent tax planning, is also a cheat. This kind of thinking undergirds much of the public debate surrounding taxes, and while it’s very clickable, it’s disingenuous. It’s not fair to create a legal framework and then blame people for following it. The problem wasn’t with those 155 people; the problem was with the tax code.