A slightly off-center perspective on monetary problems.
Saez and Piketty are really starting to annoy me now. They’ve been going around suggesting that the optimal top rate on federal income taxes is about 70%. Yet research recently published by Saez and Diamond makes the following points:
1. It’s very hard to close loopholes, so the optimal tax rate on capital should be calculated by taking those loopholes as a given.
2. If the income tax continues to have its current loopholes, the optimal top federal tax rate is 48%. BTW, under current law the top MTR is not 35%; it is considerably higher. However the tax code is so complex I can’t tell you the exact figure. I recall ObamaCare added another 3.8%, and there are some other quirks in the tax law that make the effective top MTR slightly higher. Certainly a top rate of 48% would be an increase, but less than many assume. (I could live with a 48% top MTR (or even more) if applied to consumption, not income.)
There is overwhelming theoretical and empirical evidence that higher MTRs sharply reduce work effort. Let’s review the evidence:
1. Theory. Because tax revenues are recycled back into the economy through various programs (mostly social spending in Europe) the income-compensated labor elasticity number is appropriate, which is unambiguously positive.
2. Evidence. Europe had tax rates close to US levels back in 1970, and a similar work effort. Then Europe decided to test the Prescott model by sharply raising taxes above US levels. A beautiful natural experiment, because we know that cultural differences would not explain any big differences in work effort. After all, the work effort was similar when the tax rates were similar.
3. More evidence. Then low and behold European work effort (in taxable market jobs) began falling sharply, so that before the current recession it was around 25% below US levels. This is what the Prescott theory would predict.
So you have both theory, cross-sectional, and time series evidence all strongly pointing to big government depressing work effort in Europe. And all competing explanations are rather ad hoc.
And how does Piketty respond to all this evidence:
High taxes have depressed the labor supply in European economies, according to economist Edward C. Prescott, a Nobel laureate, who argues that Americans generally work more hours because U.S. tax rates are lower. The result: U.S. inflation-adjusted GDP per capita in 2010 was about 40 percent higher than the average for the euro area, according to data from the Organization for Economic Cooperation and Development in Paris.
"This is nonsense," said the Picketty, 40, a visiting professor at MIT in 2000-2001. "When people in Europe have five weeks of vacation, this is not in response to high tax rates."
Oh really? I kept reading to learn why “this is nonsense.” And there was nothing . . . no explanation was offered. In the past I’ve noticed that some people have trouble with this idea because they think of work effort as being in some sense socially determined, not responding to incentives at the individual level. But these social policy decisions are be very much influenced by the average preferences of the public. So yes, a 35 hour work week and 5 weeks vacation may be government-determined, but ask yourself why government responded to pressure for these changes in France, but not in America. Obviously growth in government may not be the only factor, but it’s nonsense to claim that this explanation is “nonsense.”
HT: Dilip
Tags:
70 Responses to “More nonsense on taxes”
Not so fast. This is shooting from the hip.
Glaeser, Alesina have an extremely comprehensive theoretical and empirical study. http://www.economics.harvard.edu/pub/hier/2005/HIER2068.pdf
Tax rate changes don’t explain the half of it because Prescott’s analysis is critically dependent on (surprise, surprise) patently unrealistic assumption of labour elasticity.
Glaeser/Alesina contend that unions in Europe have near-solved a massive social coordination problem. They do not put undue weight on his explanation, just its superiority to the alternatives of tax (Prescott) and culture.
(I could live with a 48% top MTR (or even more) if applied to consumption, not income.)
yes, a tax cut!
btw, my main issue with a lot of these studies is that they fail to treat the family as a unit: if one person works, but the husband/wife does not (but wants to) it does not necessarily make financial sense for him/her to go back to work due to taxes (among other things, like childcare). I know some families who are essentially paying for one person to work (negative revenue) but are doing it for non-monetary reasons like career development.
in other words, its household labor supply thats relevant.
I agree that Piketty is far too dismissive of the possibility that tax rates are responsible for a nontrivial chunk of the differences in optimization along the the labor/leisure margin. But to be fair, Alesina, Glaeser, and Sacerdote have a paper arguing that tax rates are probably not the main culprit, or at least not provably so: http://www.economics.harvard.edu/pub/hier/2005/HIER2068.pdf
Prescott’s paper was never very convincing. It argued its case in, frankly, a very Prescott-like way: assume a utility specification that implies a Frisch elasticity overwhelmingly at odds with the evidence from empirical labor economics, then point to a few numbers and argue (sans econometrics) that the numbers seem to fit. (By the way, in my view Chetty et al. firmly rejected the usual Prescott/Rogerson/Hansen argument about macro elasticities in the most recent NBER Macro Annual: http://obs.rc.fas.harvard.edu/chetty/ext_margin.pdf )
I’m actually open to the possibility that long-term labor supply elasticities are significantly higher than short-term ones, because institutions need time to adjust, etc; there are a lot of effects that short-term natural experiments just can’t pick up. But the (inevitably fuzzy) empirical evidence on this story doesn’t amount to much.
One thing I don’t understand about Piketty, Saez, and Diamond’s argument for the optimal top marginal tax rate is that once you move away slightly from their model, the implication that we should push the rich to the top of the Laffer curve disappears. For instance, if you assume that the rich and poor are not simply providing different amounts of a homogenous labor input, but instead providing fundamentally different kinds of services with an imperfect degree of substitutability, then in general equilibrium the poor will bear part of the incidence of a tax on the rich. Near the peak of the Laffer curve, the benefit from additional taxation is second-order while the loss to the poor from this general equilibrium effect is first-order. At the very least, this indicates that the optimal top marginal tax should be somewhat lower.
(And the idea that the rich and poor tend to be providing imperfectly substitutable labor inputs is not some sort of wild economic heterodoxy; it is, for instance, fundamental to the whole skills-and-technology literature.)
“When people in Europe have five weeks of vacation, this is not in response to high tax rates." I’m french, and I can tell you if people don’t work more, it’s mainly not because of taxes, frankly it’s really stupid to think that, but because most people think they have enough money, even if they don’t earn a lot. In my company we can make extra work sometimes paid twice more on sunday for example, and half of people are not interested, it’s always the same that agree to work more. So far people who are ready to work more are young men, many man above 40 are not interested, they have enough money, and women prefers to spend time with they family.
So at the end of the day, the difference is cultural: “these social policy decisions are be very much influenced by the average preferences of the public.” The French wanted to work less and have more leisure, so they voted for policies that enabled these preferences.
Neal: culture is endogenous.
We have also seen a flipping of leisure in the US so that wealthy citizens used to have more leisure, but now they have less. At the same time, tax rates on high incomes were reduced and the combination of higher low income support programs and higher payroll taxes has created much higher net marginal taxes on lower incomes. Has there been any research on this? The leisure reversal seems like a huge cultural transition, and the possible tax implications seem strong.
dwb, Yes, That’s a good point.
Ritwik, My response to that paper would be to cite Ljungvist’s comment on pages 65 to 77 of this link:
http://www.nber.org/chapters/c0073.pdf
Labor elasticity is an extremely complex subject, and I’ve always thought the very long run elasticities (which are hard to estimate) are much higher than short run elasticities. I think the cross sectional data is all we really have.
Also note that even if this article by Alesina, et al, is correct, it doesn’t help progressives like Krugman, who are very critical of right wing attacks on unions. Let’s say that the lower work effort (including a natural rate of unemployment that is much higher than Krugman would like) in Europe is jointly produced by unions and high tax rates. Then it would be completely accurate to say that the huge work gap, and real GDP gap, between Europe and the US is produced by progressive politics, broadly defined. Don’t hold your breath for American progressives to make that argument.
BTW, If I am “shooting from the hip” what is Piketty doing with an unsupported “nonsense” charge? At least I give reasons.
Matt, Good comment. I agree that long run elasticities are the key, which is why I only trust cross-sectional studies. BTW, the Asian data is also broadly consistent with the Prescott hypothesis, with Japan having more hours worked than Europe, and lower taxes, and Singapore having much higher hours worked than even America, and much lower taxes.
My other complaint is that this research (on top rate MTRs) assumes that higher taxes on Buffett would cause Buffett to reduce his consumption. That seems exceedingly unlikely in my view–but it gets us into the income vs consumption argument.
I agree with your final point, and indeed most of what you have to say.
Neal, See Alex, I meant cultural that is tax invariant.
kebko, Yup, I’ve made that argument several times. Think especially in terms of labor force participation by poor versus affluent women–there’s been a total reversal, roughly at the time the implicit MTRs reversed.
Professor Sumner, I don’t understand how higher taxes would lead people to seek a shorter work week and more vacation time. Just the opposite seems true.
JLS, You do realize that your argument actually supports Prescott, and not the alternative view that it is government mandated vacations, don’t you?
Bill, Income compensated supply curves are the key. Most people think in terms of income uncompensated supply curves.
FactCheck.org just did a post on the ObamaCare 3.8% tax. http://factcheck.org/2012/04/realtors-the-3-8-sales-tax-and-247-medicare-premiums/ Seems like a tax the rich provision that won’t raise a lot of revenue relative to the legislation. Nonetheless, to your point of complexity in the code, I agree.
What I don’t get here is why a higher level of work effort is considered de facto desirable. Shouldn’t the measure be average standard of living? And from there, an equivalent standard of living at a lower overall level of effort would be the more desirable outcome. Trying to force more people into an already crowded labor market doesn’t benefit anyone unless you think that people would rather live in a household where everyone has to work full time to make ends meet vs one where only one person is required to work and even then gets to retain a significant amount of their time to invest in their own personal interests.
I’d like to see an MTR study that adequately splits out social benefits from high MTRs. i.e. Are Europeans working less because of high MTRs, or because they have a generous government pension and don’t have to save as much for retirement? My wife and I both work more because we have to save more, or else we’ll be living our retirement in a single-wide in Tornado Alley. (I’m not advocating high MTRs; I am with Dr. Sumner in pining after a progressive consumption tax).
Read Full Article »