OK, here’s a bracketless, marginless form.
agi = your adjusted gross income
lym = last year’s [del]mean[/del] median agi (I’m switching to median & I still think it’ll come out low)
1 - (7/8)[sup](log[sub]8[/sub][8*agi / lym])[/sup]
This is a lot harder to compute.
Using the smooth function:
1a. agi = _ 125,000: tax = _ $15,625.
2a. agi = _ 250,000: tax = ~ $43,125 (which is not exactly right, but close)
3a. agi = _ 500,000: tax = ~$100,000 (actually a little more, but close)
4a. agi = 1,000,000: tax = _$233,000
5a. agi = _ _15,625: tax = _$0
6a. agi = _ _62,500: tax = _ ~$5,960 (approx)
The smooth function gives rebates on low incomes as well, though pretty small ones.
7a. agi = $7,812: rebate = $368 (4.71%)
8a. agi = _ _ $1: rebate = _ $1.81 (which makes this a pretty sad negative income tax for the disabled)
Using a bracketed function as follows:
_ _ first $15,625 tax free
_ next $109,375 @ 12.5%
_ next $875,000 @ 23.4%
next $7,000,000 @ 33.0%
next $56,000,000 @ 41.4%
etc.
1b. agi = _ 125,000: tax = $12,547.
2b. agi = _ 250,000: tax = $41,797
3b. agi = _ 500,000: tax = $100,297
4b. agi = 1,000,000: tax = $217,297
5b. agi = _ _15,625: tax = _$0
6b. agi = _ _62,500: tax = _ $5,859
I haven’t decided how to handle incomes below $15,625.
Running those numbers, I think I may go to fifths instead of eighths, get a steeper curve. If I keep eighths, I think I’d have to make the baseline below the median, since I’m not trying to make this a massive revenue cut.