Societal Slack and Progressive Taxation: A Second Look (Part II-Income Tax)

When revisiting my old post "Societal Slack and Progressive Taxation" I decided to enlarge upon the matter, specifically considering how corporations have paid steadily less in tax since the end of World War II (from 4-5 percent of GDP, down to 1-2 percent in recent years).

In this post I turn to the distribution of the personal income tax burden.

According to recent data from the Tax Foundation, the top 0.1 percent of taxpayers took home 9.52 percent of all Adjusted Gross Income back in 2016. Some 141,000 people receiving some $2.1 million a year or more, they collectively got about $970 billion, and paid some $260 billion of that in tax.

As is often noted, the rate in the top tax bracket is far lower than it used to be. In 2016 it was 39.6 percent. By contrast, between the end of World War II and the early 1960s it was 91 percent, levied on, if one adjusts for inflation, roughly that level of income.

What would the '50s-era tax rate on the earnings of the top 0.1 percent above that $2.1 million level (such as was the case in 1950) raise? One can take the $970 billion figure and deduct from it the aforementioned income (the first $2.1 million made by that 141,000 persons). They would wind up with about $675 billion. Ninety-one percent of that sum would amount to over $610 billion--over twice that strata's entire tax contribution in 2016.

And of course, one could extend this pattern down through the rest of the 1 percent (1.4 million taxpayers with roughly $500,000 in income). Simply taxing everything between a half million dollars and $2.1 million at the 73 percent rate prevailing in 1950, one would wind up with another $460 billion--over $1 trillion in income raised on earnings above the half million dollar a year level, in contrast with the $500 billion actually paid by the top 1 percent. As this implies the practice does run into diminishing returns fairly quickly, but extending the approach down through the rest of the top 10 percent (with anything above $200,000 taxed at 53 percent, anything above $140,000 at 48 percent), would raise the tax collected on marginal income above $140,000 a year alone into the $1.7 trillion range, roughly $700 billion more than what the top 10 percent actually paid in 2016.

Moreover, there are respects in which these figures can be considered conservative. One is that the above, rough calculation did not take into account the intricacy of the bracket structure, which would have captured more wealth from each strata of taxpayer. (Income between $500,000 and $2.1 million would not all have been taxed at the 73 percent rate but, as income rose from the low to the high end of this range, been subject to seven progressively higher rates, with 1950 seeing anything above $1.65 million or so taxed at 90 percent.) Another is the way in which inflation and "bracket creep" translated to more people paying at higher rates (so that by 1963 anything above that $1.65 million, roughly speaking, was now subject to the maximum rate, and similar creep evident down the line). Still another is the fact that higher rates extended down somewhat below that most affluent tenth. (In 1963 anything above $14,000--$117,000 in today's terms--was taxed at 41 percent, higher than the maximum rate today.) The result is that a 1950s-like tax rate structure could raise still more than suggested above. Perhaps another trillion dollars a year would not be out of the question.

However, there is also another side to the issue, namely that nominal tax rates are one thing, and effective tax rates another. The exemptions built into the system, the possibilities for legal avoidance, the recourse to illegal evasion, doubtless translated to a lower effective rate, though by just how much is a highly contentious matter. Writers abiding by the libertarian "conventional wisdom" of the economics field are quick to insist, and vehement in insisting, that the '50s did not see the rich actually paying the extravagant rates the nominal figures imply.

The idea that effective tax levels are not really much lower than they were before strikes me as implausible in light of the sharp cuts in rates for which conservatives fought so long and hard; by the reality of the growing share of income (more on which below) enjoyed by the most affluent, which is very plausibly a function of their being more lightly taxed; and changing conduct on the part of corporations whose stockholders, inhibited from taking out as much in dividends as they theoretically could by high personal income tax rates, shifted to treating corporations like ATMs after the tax cuts (payout rates 50 percent in the '60s, dipping in the '70s and exploding to a whopping 94 percent by the turn of the century).

Moreover, there is a point that those dismissive of the significance of the higher rates overlook--namely that in the '50s the rich were less rich, absolutely and relatively, and therefore relatively less of their income, and everyone's income, subject to higher marginal rates. In 1950 scarcely 20,000 of 55 million taxpayers, 0.036 percent of the population, reported making $100,000 or more, qualifying for those top three brackets. Where the larger picture is concerned the share of income enjoyed by the most affluent 1 percent ran at 8-9 percent of the total during the decade, the top 10 percent about 30-35 percent of the total figure.

By contrast, hundreds of thousands now report making twice the equivalent of $100,000 in 1950s-era dollars (about $1 million), while that top 1 percent now commands 19 percent of the nation's income, and the top 10 percent, 45-50 percent of the total. Had the richest been so rich then, and the income distribution as skewed then as it is now, then the '50s-era system would have meant much more income subject to the highest rates. All others things being equal, the effective rates would have reflected this--and it can only be expected that, were such a rate structure in place today, we would also see that.

Still, as the large sums discussed here suggest, the additional revenue that could be generated by even an effective rate considerably below the nominal rates, capturing only part of the income discussed above, would be far from trivial. (Were new higher, nominal, rates to raise the effective rate on just the richest 0.1 percent by half from the current 27 percent to 40 percent as a whole, which may not be unprecedented--all the caveats about changing methods for calculating AGI over time notwithstanding--this would still translate to an extra $130 billion a year, $1.3 trillion in ten years.)

Additionally, there are grounds for thinking that the effective rates of the 1950s do not necessarily represent some theoretical maximum. Per capita income is three to four times what it was in the 1950s, and the reality is that much of our economic trouble has come not from there being to little capital available, but too few profitable opportunities for its investment. (The ultimate cause of slow growth and asset bubbles, it is why financial institutions were so willing to buy crummy mortgage-based securities back in the early 2000s, with the horrid implications with which we are all still coping.) There is also the reality that the data on household income, as Thomas Piketty argued, tend to underestimate rather than overestimate how much the wealthiest have, indicating hidden potentials, the proportions of which are hinted at in the multi-trillion dollar "black hole" in the world's balance sheets.

Consequently, the bottom line is that if it had to be done the obstacle to raising substantially more revenue by way of the income tax (even just raising it on the narrow stratum of the most affluent taxpayers) would not be the means to pay, but the political obstacles to changing tax rates.

1. IRS data indicates that in 1980 0.2 percent of taxpayers were paying 40 percent or more. Below that level another 7.5 percent paid between 30-40 percent.