Greg Mankiw links a former student courageously trying to protect the rights of Millionares not to pay more taxes
here.
Here's the thing -- there are no reputable studies on the elasticity of the effect of tax changes on total taxes collected, and there's no logical reason that that number should necessarily,
a priori, be less than one.
Here's why: in 2007, there were 495,000 tax returns filed for millionaires. That means a significant fraction were corporate CEOs, CFOs, finance people, and professional athletes/best-selling textbook authors/TV celebs.
Peyton Manning makes about $30 million a year -- let's explore his potential behavioral responses to changes in taxes. Let's raise Peyton's taxes by 10%. Under the logic of Alan Liard, Greg Mankiw's student, and under the logic that all economists know to be the truth, people respond to incentives. Peyton Manning is a person, so he responds to this tax hike by working 6% less, and decides now he's going to sit for the Colts playoff games since he makes less money per game, and he enjoys watching Tom Brady play in the playoffs more than being there himself. Doesn't really sound likely, does it?
Of course, Peyton Manning is going to play 16 NFL games and the playoffs even if you raise his taxes considerably. The same is true of a wide variety of other professions -- corporate execs usually have two choices, they can choose to work or not work -- there are no part-time CFO jobs, and it's probably tough to be a "part-time" hedge-fund manager as well... So, let's say Greg the textbook publisher or Chuck the hedge-fund manager decides, due to higher taxes, that they are just going to retire. In that case, the government loses 100% of the taxes Chuck or Greg would have paid! The multiplier is -10!!!
Except, according to logic which is totally obvious to a pre-schooler, if Greg the textbook author doesn't sell textbooks, then Thorstein the textbook publisher will. If Peyton the quarterback doesn't play in the playoffs or appear in Gatorade commercials, then Tom the quarterback will. If the CEO of Anthem, who routinely makes $40 million, quits due to high taxes, Anthem will pay the next CEO extravagantly. If Chuck the hedgefund manager doesn't manage Peyton's money, then Emilio the hedgefund manager will manage Tom's money...
So, in this case, microeconomists might derive an elasticity of -1 (a 1 percentage increase in the tax rate reduces tax collections by 1% if 10% of the rich decide not to work at all), while the true Macro elasticity would be closer to 1.
Note that this logic does not apply to everyone, but it is likely to apply to the vast majority of those who make more than 1 million dollars per year, and Greg Mankiw's students are completely oblivious to this kind of perfectly obvious logic...
Of course, another thing higher taxes might change is how much someone might try to cheat in order not to pay taxes -- well, if this is the case, then what we really need is to beef up the IRS.
The third thing it also might change is the pay that CEOs get. For example, if the top tax bracket were 99%, for all income over $5 million, tax opponents always describe this case as having the effect that people will just stop working after they get to $5 million. That's just not an option for the likes of Peyton Manning, however. What Peyton Manning (or any corporate CEO in the same position) would likely do is settle for just $5 million a year, since all income after that goes to the state. Then, once you get a Republican in office, and they cut that 99% top bracket from $5 million-plus to $25 million plus, Peyton Manning will renegotiate his salary up to $25 million, even though he's still playing 16 NFL games a year...
So, in this case, tax rates down implies total taxes collected from Peyton Manning up but total "production" from Peyton unchanged, but worsening inequality for the economy and less taxes paid for someone else. So, we'd need to look at whether the consumption spending of people like Tiger Woods, CEOs with their corporate jets, and investment bankers is better for long-run economic growth than the spending of the poor, who spend a large chunk of their disposable income on things such as education and health care...
This is essentially what has happened in the US, and that is why you shouldn't believe it when economists tell you we shouldn't tax millionaires.
But then, you might object, *nobody* would have an incentive to become CEO if their salaries are capped around $1 million (if I was designing a tax system, I would probably not go over 50% on taxes before $3-4 million, but I would make Peyton Manning's marginal rate closer to 70%...)
Well, I once took a poll of the students in my TA section. I asked them who they would rather date -- someone w/ a bachelor's degree who works 40 hours/week, makes $40,000 per year and pays no taxes, or someone w/ an MBA who makes $80,000, works 80 hours/week, and pays $30,000 in taxes.
Most of the girls said they'd go for the person w/ the MBA, even though, per hour, the 40 hours/week job would be more, b/c the MBA has a prestigious graduate degree, is ambitious, and makes a whopping $10,000 more overall. (I had them assume that the above were peak salaries...) I also asked them if they would take a promotion from the $40,000 job to the $80,000 job, even though it's much more work, and, after tax, a pay cut per hour, and again, most wanted to go for the higher-priced job.
The study I'd really like to see is if, since the early 1980s, as wages on Wall Street have gotten out of control, what has happened to the average age of retirement at firms such as Goldman Sachs. "Standard theory" of course does not even predict whether the income or substitution effect dominates, but I suspect that there are quite a few I-bankers on Wall Street who do not exactly love their jobs or working 100 hours a week, and will just retire once they've put away $10 million in the bank. Tax them more and that just delays retirement. And to the extent higher taxes preclude the possibility of putting away $10 million for your average I-banker and talent shifts out of finance (or NFL quarterbacking), I fail to see how this is such a bad thing for society...
Put another way, suppose the CEO of Anthem made $4 million rather than $40 million due to high tax rates on multi-million dollar incomes (which we do not currently have). While in the micro context, this would greatly reduce the government's revenue, in this case, Anthem would have higher profits, or could afford to pay it's other workers more, could cut prices to gain more market share, or might even be able to deny fewer claims. In other words, it wouldn't make the whole pie smaller, it would just make it more equally distributed. And, to the extent that the Anthem CEO no longer flies around in a private Jet, owns 6 houses, and drinks $1000 bottles of wine, redistributive taxes would quite likely make the pie bigger...
UPDATE: An email and commenter pointed out that the question I asked the students in my class was reinforcing cultural norms and stereotypes and came off as sexist. They are correct and I apologize. I had not properly edited my own post however -- when I framed the question in class, I did it in a much, much more pc way -- asking everyone how much each individual would want to make/work, and also asked the question about the spouse in a completely gender-neutral way, and also asked a variety of other questions, with all questions asked to both guys and girls (I edited my post to reflect this). The guys tended to prefer spouses who worked and made less. The girls also tended to want to make and work less -- confirming what microeconomists have already found about how the female labor supply is more sensitive to marginal tax rates than it is for men. And yes, these are due to sexist cultural norms.