There have been some major changes to the U.S. individual income tax code over the past two decades — notably the tax cuts of the early 2000s, their partial rollback in 2012 and the 2017 tax reform/cuts. The net result of them all is maybe not what you’d expect: federal income tax rates fell for every income group since 2001, with the very highest earners seeing the smallest decline (in percentage terms at least) and those with incomes in the 50th to 60th percentiles down the most.
The Internal Revenue Service releases these numbers late every year with an almost two-year delay, so 2019’s are the most recent available. The income percentile groups are those chosen by the agency, with the twist that the IRS simply reports by top 50%, top 40% etc. so I had to do some subtracting and dividing to get the discrete slices shown above.
The very fine slices at the top of the income distribution are a relatively new addition — before 2015 the top 0.1% was as narrow as it got, although the IRS subsequently updated the numbers back to 2001 — and they allow one to see an interesting characteristic of the federal income tax. On the whole it’s quite progressive, in the sense that those with higher incomes face higher rates. But this progressivity reverses within the top 0.1% (taxpayers with adjusted gross incomes of $2.5 million or more in 2019), with the very highest earners paying out a markedly smaller share of their income in taxes than those in the bottom nine-tenths of the top 1%.
This is just federal income taxes we’re talking about. A 2007 study by economists Thomas Piketty and Emanuel Saez found that every income group below the 90th percentile paid out more in Social Security and Medicare taxes than in federal income taxes, resulting in an overall federal tax system that was much less progressive. In much of the country, regressive state and local taxes cancel out most or all of the progressivity that’s left, and in recent years Saez and Gabriel Zucman have taken to including health insurance premiums as a tax “because it’s mandatory and reduces wages,” which leaves taxpayers in the 50th to 90th percentiles facing a much higher relative burden than those above and below them in the income distribution.
But that’s a story for another day! For taxpayers in the top 10%, the federal income tax is what matters most, and changes in income tax rates within the top 10% over the past two decades make for an interesting tale. Rates haven’t ended very far from where they were at the beginning (this is the same data as in previous chart, with everybody in the bottom 90% combined into one group):
But wow some big stuff happened in between, especially within the top 1%. Here’s the shift from 2001 to 2007, by which time almost all provisions of the Economic Growth and Tax Relief Reconciliation Act of 2001 and Jobs and Growth Tax Relief Reconciliation Act of 2003 (together usually known as the Bush tax cuts) had taken full effect.
The 2001 tax law cut income tax rates for all income groups. The 2003 tax law accelerated those cuts, and also slashed taxes on capital gains and dividends. Capital gains in particular are concentrated at the top of the income distribution, with two-thirds of the nation’s long-term capital gains going to the top 1% in 2019 and those in the top 0.1% getting the majority of their income from them. These long-term gains have long been taxed at lower rates than ordinary income, which is the main cause of the income tax’s regressive turn at the very top of the income distribution. The 2003 law made the turn a lot more pronounced, with those in the top 0.01% facing a lower tax rate after it than those from the 97th to 98th percentile.
When there are few capital gains to be realized — as in the wake of the global financial crisis in 2009 — this effect is much reduced. Tax rates fell for everybody else that year because incomes were lower, but the top 1% saw their rates rise because they were more dependent than usual on ordinary income.
This effect was temporary, and by 2012 the tax-rate curve was almost identical to that of 2007. But the 2001 and 2003 tax cuts were mostly temporary too, a byproduct of the weird (and, frankly, bad) way in which laws have to be crafted to qualify as “reconciliation” legislation that doesn’t need 60 votes to make it through the U.S. Senate. As the Urban-Brookings Tax Policy Center sums up:
The content of reconciliation laws is limited in the Senate by the Byrd rule, which generally disallows items that do not affect outlays or revenue. The Byrd rule also prohibits initiatives that would increase the deficit beyond the fiscal years covered by the budget resolution.
With most of the tax cuts from a decade earlier set to expire, Congress passed the American Taxpayer Relief Act of 2012, which is technically the biggest tax cut in U.S. history but didn’t feel like because it simply made the Bush tax cuts permanent for all but the highest earners. The result was a big tax increase for the 1% and no change for everybody else.
The Tax Cuts and Jobs Act of 2017, which took effect in 2018, did not reverse much of this increase for the 1%. Instead, it gave its biggest breaks to the taxpayers just below them, in the 97th and 98th percentiles, with 2019 adjusted gross incomes of $291,384 to $546,434. (I’ve included the 2012 tax curve here for context, which makes the chart a little harder to read but I think rewards the extra effort.)
Back in the 2000s, Shawn Tully of Fortune magazine dubbed taxpayers in roughly this income range HENRYs, for “high earners, not rich yet,” and argued that they were being unfairly targeted by the alternative minimum tax. The AMT was created in 1969 to ensure that very high earners weren’t able to entirely escape taxation by means of deductions and credits. Because it wasn’t indexed to inflation, it started in the 2000s to squeeze what you might call the upper upper middle class. The 2012 tax law established a somewhat higher AMT exemption and indexed it to inflation. The 2017 law exempted all but the very highest earners from the tax, with the number of AMT payers dropping from 5.1 million in 2017 to 170,132 in 2019. As I’ve written before, this amounted to a windfall for the HENRYs, but before it their tax rates were actually higher than in 2001, so it wasn’t an entirely inappropriate one.
It also wasn’t entirely uncomplicated. To comply with reconciliation rules, most of the individual income tax provisions in the 2017 law were temporary. The AMT break is due to expire in 2025, which if allowed to happen will affect an estimated 7.6 million taxpayers that year, according to the Tax Policy Center. It also didn’t deliver its benefits evenly across the country. Thanks to another provision of the 2017 law that limited deductions for state and local taxes, HENRYs in low-tax states benefited much more from the AMT changes than those in high-tax states — although on average even the latter saw their federal income taxes reduced.
The permanent changes in the 2017 tax law were on the business side, with the biggest being a reduction of the top corporate income tax rate from 35% to 21%. Opinions differ on how a cut like that affects personal incomes. The mainstream view is that corporate shareholders bear most of the burden of the corporate income tax, and as high earners own most corporate shares, they pocket most of a corporate tax cut. But some economists argue that corporate taxes weigh most heavily on the incomes of rank-and-file workers because they’re less mobile than investors and thus less able to escape high rates. Income changes since the 2017 law actually give a little support to the latter view, with the share of overall adjusted gross income going to the top 1% falling in both 2018 and 2019.
Compared with the 2001 and 2003 tax cuts, then, the 2017 edition wasn’t all that skewed toward the very rich. Occasional Democratic claims that it was assume that (1) the corporate tax burden falls mostly on the highest earners and (2) all the temporary provisions in the 2017 law are allowed to expire, which given past experience seems unlikely to happen.
Yet the regressive turn at the top of the income distribution, while smaller than a decade ago, remains. A bit of it, to be sure, is an inflation-caused illusion: A $1,000 capital gain from the December 2021 sale of an asset purchased a decade earlier, for example, represents not $1,000 in real income but $809 in income and $191 in inflation, if you go by the consumer price index. Also, there are economic arguments for keeping capital gains taxes low (and some against) that I’m not going to get into here. On the other hand, the very wealthy have ways to reduce their taxable income that aren’t available to the rest of us, meaning that the above charts probably overstate the share of their true income going to taxes.
With that in mind, here’s the percentage of adjusted gross income going to each of the income groups from the bar charts at the beginning of this column, in 2001 and 2019.
Every group in the top 10% has seen its income share rise, while every group below it has seen its share fall. But the top 0.01%, the group that faces lower average tax rates than those just below it in the income distribution, still accounted for just 4.4% of all adjusted gross income in 2019, and the top 1% accounted for 20.1%. Again, that may understate those groups’ true income shares somewhat, especially that of the 0.01%, but it’s an indication that despite their gains since 2001, the very highest incomes still represent a pretty small share of overall incomes.
Given the information presented above, it’s understandable that some Democratic lawmakers want to find new ways to tax billionaires. Doing so might well make the tax system fairer. But it should also be clear that the bipartisan consensus that seems to have emerged over the past two decades that no one outside the top 1.5% or so should have their taxes increased (the Biden administration’s stated cutoff is an income of $400,000) limits how much money can be raised from the income tax.
The across-the-board reduction in effective tax rates since 2001 has brought with it a commensurate reduction in individual income tax revenue, from 9.4% of gross domestic product in fiscal year 2001 to 8.1% in 2019 (it fell even further, to 7.7%, in 2020 but some other things were going on in 2020). Fiddling around with the taxes of the 1% isn’t meaningless in this context — individual income tax revenue went from 7.1% of GDP in fiscal 2012 to 8% in fiscal 2014, after the tax increases on the highest earners had taken full effect — but it won’t pay all the bills.
To contact the author of this story:
Justin Fox at [email protected]
To contact the editor responsible for this story:
Susan Warren at [email protected]
- Yes, tax cuts may sometimes boost GDP,(1) I’m aware of no credible economic model in which the reductions in taxes since 2001 would have boostedGDP enough to pay for themselves and(2) actual GDP performance since 2001 has been dismal, with annual real growth of just 1.9%.
© 2022 Bloomberg L.P. All rights reserved. Used with permission.