Note on Methodology

Start at Table 1, the first column, "Top 1%".

Take "Number of Returns" (first row) and multiply it by $466,950 , which I got from here, as the boundary for the top marginal rate. Fortunately, it's very close to the cutoff for the top 1% (second row from bottom). I used the rate for "married filing jointly." I take the product of those two numbers to be that income of the top 1% which is taxed at less than the top rate.

Subtract that number from "Adjusted Gross Income." I take the result to be that income of the 1% which is subject to tax at the highest rate. I'm using 35% as the highest rate; I use 2016 data because it was the newest I could find.

I now have $1.345 trillion. I next included an adjustment for the negative incentive effect of the tax. Romer and Romer estimate the elasticity of supply at at 0.2. The percentage change in the after-tax wage times the elasticity gives the percentage change in the supply. I got an estimated change of -5%, so I reduced the revenue accordingly.

Then I subtracted 23% as an allowance for deductions from taxable income. I got 23% here. It's the approximate average deduction in both the second and the third highest income groups, taken at the midpoints. I didn't use the highest group because there's no midpoint.

We multiply that by .50-.35=.15 to get the additional revenue from raising rate from 35% to 50%. I ended up with $147 billion for this piece.

Now we add to that the additional revenue from the 58% and 65% rates, which I'll take to apply to the .1% and the .01%, as I indicated in the post. Table 3 gives $966 billion for the income of 0.1%. I don't have an estimate of income for the "tippy top" group, but looking at Table I on p. 575 under "Pretax national income/Income share" it is apparent that there is a strong tendency for the top 10% of any group to get about 45% of the income for that group. So take 45% of $966 billion to get $434.7 billion for the top .01%.

Now we proceed as before. The calculated incentive effects this time are 9% for the increase from 35% to 58%, and 12% for the increase to 65%, and the deductions are as before. Since we already counted these people when we raised the rate to 50%, we multiply only by .58-.50 and, similarly, .65-.58. My totals are $54 billion and $20.5 billion, for a grand total of $222 billion.

Note that the increase to a 50% rate accounts for two-thirds of total revenue, and the increase to 58% another quarter, while the final increase to 65% accounts for less than a tenth.


I've done this wrong about seven times up to now, so if you see what looks to you like a major error, please email me at hfrant@post.harvard.edu.