Fix the Current AMT Before Creating a New One

January 31, 2012

In response to President Obama’s State of the Union address, in which he called for a “Buffett Rule” setting a minimum tax of 30% for households making over $1 million, a group of Democratic senators have introduced legislation doing just that. The motivation for this legislation is Warren Buffett’s claim that he pays a higher effective tax rate than his secretary – a claim that, as I’ve pointed out, is dubious, at least with regard to the specific numbers he gives. I’ve put together a chart showing that his claim does not reflect the distribution of tax burdens generally:

Tax Rates by AGI Group

(Note: this chart has been modified since it was first published.)

I’ve included all payroll taxes (both employer and employee side), but the shape of this chart is still what you’d expect in a country with a progressive tax system: high-income people generally pay higher rates. Admittedly, effective rates trail off slightly at the very high end of the income spectrum, reflecting the fact that most of this income is taxed as capital gains rather than as ordinary wage income. However, even when payroll taxes (which are regressive) are considered, the (very small) $10 million+ group still pays a higher effective rate than all groups making under $100K, and is about even with the $100K-to-$200K group. I’ll also point out that the Buffett Rule, as proposed, doesn’t target the extremely small number of taxpayers in this elite group, but instead casts a wider net with its $1 million threshold.

The point is that, anecdotes aside, the tax code is progressive, especially if payroll taxes are excluded (there are reasonable economic arguments for doing so, given that in many ways they are more like fees for specific social insurance programs than general government revenue raisers.) Those on the left, such as the President, would like to make the tax code more progressive, and those on the right would like to make it flatter. Without delving into a discussion of which direction we ought to choose, I’ll simply point out that, if you want to make the code more progressive, the Buffett Rule is probably the worst possible way to do so. It’s a rehash of a failed policy.

In 1969, the tax code had various tax preferences which primarily benefited wealthy households, to the point that a small number of high-income households paid no income tax at all. In response, Congress enacted a “Minimum Tax,” requiring all households to pay an add-on tax equal to 10% of these tax preferences in excess of a particular threshold. Then, as now, the motivation for the change was a perception that very wealthy households were unfairly benefitting from various forms of preferential tax treatment. In 1984, this idea grew into the modern Alternative Minimum Tax, which is (sound familiar?) an alternate tax system that sets a minimum amount each household much pay, regardless of various preferences and deductions.

Additional preferences and deductions have since been added to the AMT, and it has grown in complexity. It has essentially become an entirely parallel tax code, more or less equal in complexity to the regular tax code, except with alternate definitions of income, deductions, tax brackets, and rates. Many taxpayers need to do their taxes twice and pay whichever is higher. It’s the kind of absurdly complex system that leads the IRS to write tortured instructions like this:

Enter the amount from line 6 of the Qualified Dividends and Capital Gains Tax Worksheet in the instructions for Form 1040, line 44, or the amount from line 13 of the Schedule D Tax Worksheet in the instructions for Schedule D (Form 1040), whichever applies (as refigured for the AMT, if necessary) (see instructions)

-Instructions for IRS Form 6251, Line 37 (2011)

The Buffett Rule is the same thing: an alternate, parallel minimum tax (in this case, a flat 30% tax with very few deductions). Why should we expect that it won’t morph into the same kind of beast that the AMT has become? We already have a minimum tax whose original purpose was to fix the exact same problem that the Buffett Rule is trying to address; if anything, we should try and fix the current AMT rather than adding a second one.

There’s another problem with the Buffett Rule: current descriptions of it give no indication that it phases in at all. Rather, once a taxpayer’s income reaches $1M, the 30% minimum requirement suddenly activates. This is in stark contrast to the marginal structure of our income tax code, where higher rates only apply to income above the bracket threshold. The problem here is obvious: someone with an income near, but under $1 million (whose effective tax rate is almost certainly below 30%, according to the table above), has a huge incentive to keep his income below the threshold. If he exceeds it, his after-tax income suddenly falls significantly. (Citizens for Tax Justice, in a preliminary score of the proposal, assumes that it gradually phases in between $1 and $2 million, thus avoiding this problem to some degree, but as far as I know, this is just speculation on their part; it’s unclear if the actual proposal will have a phase-in when it’s introduced tomorrow.) (Update 2/1/12: According to Sen. Whitehouse’s official press release, the proposal does indeed include a phase-in between $1 and $2 million.)

Even if it does phase in, we’re left with a peculiarly arbitrary set of marginal tax rates, where the effective marginal rate for most people with incomes between $1 and $2 million is higher than it is above $2 million, because of the “bubble” rate effect caused by the phase-in. Nobody designing a tax code from scratch would produce anything remotely like this.

Finally, there’s no indication that the $1 million threshold is indexed for inflation. This is a problem with the current AMT; none of its parameters are indexed, so Congress passes temporary “patches” each year to avoid an effective tax increase. This is often a politically contentious (and occasionally farcical, at least with regard to the budget math) process, with the side effect that CBO’s “Current Law” budget outlook is rendered more or less meaningless. There’s the risk that the same problem with the Buffett Rule might cause similar absurdity to become another permanent fixture of the annual budget process, as well as the possibility of effective automatic tax increases every year if it’s left alone.

Reasonable people can disagree over the merits of a progressive tax system, but there are good ways and bad ways to tweak progressivity. Rather than adding yet another minimum tax, we should focus on either fixing the current AMT or, better yet, getting rid of it entirely and fixing the problematic deductions and preferences that motivated its creation in the first place. Let’s fix the tax code we already have, rather than patching over it with a new layer of complexity.


Related Articles