Our New Marginal Tax Rates Calculator
April 19, 2011
Yesterday we released our newest interactive tool, a marginal rates calculator and graph. It’s a flexible tool and is a good way to demonstrate the idiosyncrasies in the structure of our income tax code.
A good, reasonably simple example is the alternative minimum tax (or AMT.) It’s a provision that is confusing to many people because of the way it’s calculated. It operates as an entirely parallel tax system, with its own definition of taxable income, tax brackets and rates, and set of allowable deductions, and taxpayers owe whichever is higher.
The way in which AMT appears on tax forms obscures this nature. All taxpayers go through the 1040 calculating their tax the normal way. After you arrive at this number, you then go and calculate your tax again using the AMT rules, and the number you arrive at the end of this process is called your “Tentative Minimum Tax.” If your TMT is greater than your regular tax, you enter the difference on your 1040 form on the AMT line. This makes it appear like AMT is a simple add-on tax, but it’s not the case – total TMT is independent of your tax under the ordinary system, but AMT (the difference) obviously depends closely on the ordinary tax. Some strategies for reducing AMT liability do so simply by increasing regular tax – meaning that the difference between regular tax and TMT is smaller. However, the total tax owed is the same.
Our marginal rates graph displays tax liabilities under both AMT and the regular system. (The AMT calculation takes place prior to tax credits like the Child Tax Credit or the EITC, which reduce tax liability after this step, so the AMT line shows, essentially, tentative minimum tax, minus any tax credits.) If this line is above the regular tax line, the space in between represents AMT liability. This graph shows a family with AMT liability over a wide range of incomes.
The purple line is tax owed under the AMT system, and the dark blue line is the regular system. The AMT is engineered to have this “hump” shape, where it extends above the flatter, regular tax line in certain income range. This is accomplished by having a large income exemption which phase out over a wide range. Taxpayers in this range not only pay the statutory 26% or 28% AMT marginal rates on income, but also see their AMT income exemption reduced as well, leading to effective marginal rates of 32.5% and 35%, until the exemption is reduced all the way to zero, where the effective rate drops to match the statutory rate of 28%. This is slower than regular tax in this income range, so eventually regular tax retakes the lead, reducing AMT liability to zero, and returning the marginal rate to 35%.
Of course, from an economic perspective, the end result is all that matters; the details of the tax calculation are less important. With that in mind, let’s turn off the regular and AMT lines and look just at the actual tax owed and the effective marginal rate.
I doubt there’s an economist in the world that could come up with a clear economic reason for a tax structure like this. It’s arbitrary to the point of absurdity, and underscores the need for fundamental tax reform. We could start by getting rid of tax deductions entirely – that would eliminate the need for the AMT in the first place.
Another good use of the calculator is to shed light on “cliffs” in the tax code – abrupt changes where one’s tax liability jumps significantly. Two obvious examples are the $3,150 limit on investment income for EITC recipients, and the discrete phase-out of the tuition and fees deduction. Such things lead to negative incentives – in the first case, it’s possible for one’s total after-tax income to be higher by reducing investment income; in the second, there are two particular ranges where after tax income can be increased by reducing one’s gross income below a certain threshold.
Note: We’re having some technical difficulties with the calculator concerning the last two links. I’m aware of the problem and am working on a fix.
Note 2: The technical problems have been solved.