In the Raleigh News and Observer last fall, David McAdams, associate professor of economics at Duke University’s Fuqua School of Business, claimed—contrary to even Keynesian economics—that President Obama’s proposed tax on millionaires would create jobs. The so-called Buffett rule, named after billionaire investor Warren Buffett, is supposed to ensure that “millionaires and billionaires” pay no smaller a percentage of their income in taxes than a middle-income family.
In a perverse twist on supply-side analysis of how marginal tax rates affect economic growth, McAdams argues that the Buffett rule, while forcing very-high-income individuals to pay more in taxes, would lower their marginal tax rate, stimulate investment, and hence create jobs. In reaching this odd conclusion, McAdams shows a level of understanding of tax analysis that is shockingly pedestrian, if not sophomoric, particularly for a professor at one of the most prestigious business schools in the world. (McAdams is not a professor in Duke’s highly respected economics department.)
McAdams illustrates his argument with the following example:
Consider . . . a millionaire whose income consists of $1 million in capital gains and $100,000 from a chain of hot dog stands. This millionaire pays taxes of 15 percent on capital gains and 35 percent on net income from the hot dog stands, for a total of $185,000 ($150,000 plus $35,000). This amounts to an average tax rate of 16.8 percent, less than the 25 percent marginal rate paid by many middle earners.
Under the Buffett rule he would have to pay 25 percent of all income in taxes, for a total of $275,000. However, he would be keeping 75 cents of every additional dollar generated by the hot dog stands, compared with 65 cents of every dollar without the Buffett rule. So this millionaire would have more incentive to expand his hot dog business, hire more workers, and so on. For the millionaires it would apply to, the Buffett rule would effectively cut marginal tax rates 10 percentage points even as it raises the overall tax burden.
Let’s assume for the moment that his example is analytically correct—which it is not—and the millionaire’s marginal tax rate for the year in question is lowered from 35 percent to 25 percent. This result is completely an artifact of the hot dog vendor’s particular situation during that tax year, and he could not have known about it in advance. In other words he doesn’t know that his marginal tax rate will be 25 percent going into the tax year. He discovers it only after the fact, when he does his taxes or, at best, speculates that it might be the case as the year draws to a close.
For the marginal rate of 25 percent to function as an economic incentive, a person would have to know that his investments are going to generate a million dollars in capital gains before the market actually does so. In fact he would need to know that this would continue to be the case for a considerable length of time. But obviously this knowledge is not available to anyone.
This millionaire-for-a-year hot dog stand owner would never make expansion decisions based on the expectation of a continued marginal tax rate of 25 percent. In fact the only reasonable thing for him to do would be to base his future investment on the (presumed) statutory marginal rate of 35 percent. McAdams seems not to understand the difference, in terms of incentives, between an ex ante marginal tax rate, which is all that matters, and an ex post marginal tax rate, which matters not at all. The fact that he conflates these two does not bode well for students at Duke’s Fuqua School.
Marginal vs. Average Rates
What is just as troublesome is that he confuses marginal and average tax rates. Before giving his example McAdams states:
[B]ut suppose “middle-class families” pay 25 percent of their income in taxes, which is the marginal tax rate for many middle-income earners. If a “millionaire” already pays more than 25 percent of his income in taxes, the change won’t affect him. But a millionaire whose income derives from both capital gains, taxed at just 15 percent, and business income, taxed at a marginal rate of 35 percent, may well be paying less than 25 percent of his total income in taxes—and therefore would pay more under the Buffett rule.
While McAdams is correct in saying that the marginal tax rate for middle-class families is about 25 percent, it does not mean those families “pay 25 percent of their income in taxes.” With a progressive income tax, the marginal rate is the percentage paid on the last increment of taxable income, not the rate paid on total income. A middle-class family pays 10 percent on the first $12,149 of taxable income and 15 percent on income from $12,150 to $46,250. The 25 percent bracket applies only to taxable income between $46,250 and $119,400, after which the marginal rate goes to 28 percent.
Thus a family facing a 25 percent marginal rate would not pay anything like 25 percent of its total income in taxes, as McAdams asserts. On an income of $75,000, the average rate would be 18 percent. As I said, this is a sophomoric mistake.
The same would be true for the hot dog vendor earning $100,000. McAdams confusingly claims the vendor would face a 35 percent marginal tax rate—wrong again—and pay $35,000 in taxes on the $100,000. Once again McAdams confuses marginal and average tax rates. Furthermore, not only is McAdams wrong about the amount paid by the vendor, he also gets the marginal rate wrong. In fact he’s off by two whole tax brackets and possibly three. A single person earning $100,000 in taxable income faces a marginal tax rate of 28 percent, not 35 percent, and if that vendor is a family man, making it an apples-to-apples comparison with the “middle-class family,” his marginal rate would be 25 percent. Because there is not enough information provided, it is impossible to go back to McAdams’s example and plug in the actual rates to see how far off his numbers really are.
Suffice it to say that when people write about taxation and job-creation incentives, they really should make sure they know what they are talking about.