Restraining the power of the rich with a 10 percent surtax on incomes over $2 million*

Excessive wealth and power commanded by a small group of multi-millionaires and billionaires pose an existential threat to America’s economic vitality, democracy and civil society.

It’s well-known by now that the richest 1 percent of American households have essentially doubled the share of national income they claim since the late 1970s. Less well-known is that inequality has even risen sharply within the top 1 percent, with the top 10 percent of that overall group—or the top 0.1 percent—accounting for half of all income within the top 1 percent.1 In 2016, the latest year of available data, households with adjusted gross income (AGI) of over $2 million made up just over 0.1 percent of tax filers, but accounted for 100 times as much (10 percent) of total AGI.

The political clout of this topmost sliver of households is likely even more outsized then their share of overall income. This group’s incomes overwhelmingly stem from owning financial assets, not working in labor markets.2 This means that they benefit from the preferential tax treatment given to income from wealth relative to income from work. The Trump tax cut at the end of 2017 was tailor-made for very rich, as its largest cuts accrue to business owners—both corporate and non-corporate business.3

Providing countervailing power to that wielded by households with multi-million dollar incomes will require ambitious policy changes across a range of issues. Steeply progressive taxes have recently been proposed by a number of policymakers and economists as key ingredients in the overall policy portfolio meant to restrain the power of the super-rich. One idea that has not yet gotten the attention it deserves in this discussion is a surtax on very high incomes. A surtax has a number of advantages as a tool for checking the power of the rich. First, it’s laser-targeted, and can be phased in only at the $2 million threshold. Second, it does not provide preferential treatment for wealth-based incomes relative to work-based incomes—it applies to every dollar of any kind over the income threshold. Third, this neutrality of the surtax across types of incomes means that in the long-run it would be hard to avoid or evade.

Proposal: A 10 percent surtax on incomes over $2 million

We propose a 10 percent surtax on all income over $2 million. For simplicity, the income threshold should be defined by a taxpayer’s adjusted gross income (AGI). In the last year of IRS data available (2016), $2 million came in just below the threshold for the top 0.1 percent of $2.3 million.4 The surtax would apply to all income above AGI, including dividends and realized capital gains.

How much revenue might a 10 percent surtax on income over $2 million raise?

Our preliminary estimate is that such a surtax would raise roughly $75 billion if implemented in 2020 and would raise $850 billion in its first decade. The methodology for this estimate is fairly straightforward. First, we estimate how much total AGI is recorded by households earning above $2 million.5 In 2020, we estimate that as roughly $1.08 trillion. Second, we estimate how much of the income of affected households would be exempt from the surtax by falling under the threshold, which is simply the number of tax filers in the top 0.1 percent multiplied by the tax’s threshold of $2 million. With roughly 157,000 filers estimated to have incomes over $2 million in 2020, this implies that about $320 billion exempt from the tax, making the overall base a bit over $750 billion. Third, we multiply this base by 10 percent to get $76 billion raised in its first year. In later years, we let the number of tax filers and their AGI rise at rates that have characterized the past twenty years and make similar calculations for each of those years.6 Over the next decade this implies revenue of roughly $850 billion.

That $75 billion in new revenue is a lot of money, even in the context of the federal budget. Those funds could, for example, essentially cover the cost of providing universal high-quality pre-kindergarten for all 3 and 4 year olds in the United States, plus providing substantial aid for paying for high-quality childcare for all 0-2 year olds. This sort of ambitious investment in America’s children would provide huge benefits to American society and economic efficiency.7

*A previous version of this post explored a surtax that applied to the top 0.1 percent of incomes. In 2016 (the last year of data available), that threshold was $2.3 million. The current version applies the surtax to incomes over $2 million, and does not annually index this threshold for inflation. Accordingly, the description of the proposal and some of the revenue estimates have changed.


1. Data on top 1 and top 0.1 percent income shares can be found in the online data on distributional national accounts maintained by Gabriel Zucman, based on work done by Zucman and co-authors Thomas Piketty and Emmanuel Saez.

2. The Piketty, Saez and Zucman (PSZ) data referenced in the footnote above indicates that non-labor income accounts for more than two-thirds of the income of the top 0.1 percent.

3. The single most-expensive component of the Trump tax cut of 2017 was the cut in corporate tax rates, which can account by itself for the entire net cost of the tax cut. Among its other provisions was a deduction for “pass-through” incomes – a form of income quite concentrated in the upper reaches of the income distribution. The PSZ data, for example, estimates that the top 0.1 percent claim almost 19 percent of all income generated by non-corporate businesses that is not paid to employees.

4. This threshold can be found here, a summary table from the Internal Revenue Service (IRS) Statistics of Income (SOI) data.

5. For estimated top 0.1 percent thresholds in 2019, see this table from the Tax Policy Center.

6. We assume the $2 million threshold is not adjusted for inflation.

7. For an overview of the broad benefits stemming from this sort of ambitious investment in early childcare and education, see Bivens, Garcia, Gould, Weiss, and Wilson (2016).