Sunday, December 28, 2014

Scott Sumner on taxes



Scott Sumner has a reply to a BV piece of mine in which I suggested that the labor disincentive effect of income taxation is relatively modest right now in the U.S.

Scott:
The required number of hours worked is itself endogenous.  In Europe, required yearly hours are less than in the US, due to higher taxes. 
This assertion might be true. If Scott has evidence to back up this claim, he does not present it in this post.

Scott:
Second, when you tax people, they are poorer, and they need to work more to maintain their standard of living. 
This is a common misconception that I see all the time.  There is no first order effect of taxes on national income, as the tax money gets recycled into the economy.  Now it’s true that the government might waste the tax money, leaving a country poorer, but in that case it would be more accurate to say that government waste causes people to work harder.  In modern economies most extra spending at the margin goes back in transfer programs.  Since national income doesn’t fall from the direct impact of taxes, there is only a substitution effect on labor supply, not an income effect.
Well, this would be a good point if taxes left income unchanged - in other words, if tax revenue was redistributed lump-sum to the people from whom it was extracted, as is the case in many economic models. However, if taxes represent income redistribution - as they often do in the U.S. - this is not the case, and we do need to think about income effects. In particular, our tax system is highly progressive - rich people pay most of the taxes, poor people get most of the services. Thus, if we want to think about whether taxes make the rich work less, we do need to think about income effects, contra Scott.

Scott:
You don’t want to use time series data, as there is a long-term downward trend in hours worked due to increasing affluence. 
Of course, but this does give you some general idea of how much we might increase labor supply (or fail to increase it) by cutting marginal tax rates. The long-term trend really seems to dominate. Of course, this by itself is not dispositive, but that's why I mentioned it.

Scott:
Lower income people face a much higher implicit MTR than in the 1950s and 1960s.
This is a good point, and I probably should have mentioned something about this in my article (I was focused on the idea of taxing the rich). Casey Mulligan has argued that implicit taxes (from benefit phase-outs) are high. Paul Krugman has also made this argument. I'd like to see more research on it, but it's a very important point, and one that is totally left out of the discussion in most U.S./Europe comparisons.

For a rundown of the relevant research on taxation and labor supply, see these fun slides. (hat tip to Claudia Sahm)

42 comments:

  1. Middle-Class and Upper-Middle-Class households definitely faced higher MTR rates in the 1950s and 1960s. The huge "91% in the 1950s" top rate gets undue attention, when what really hit most people hard were the many, many brackets below that.

    I disagree with his common "French worked the same as Americans in the 1960s due to similar taxes". French hours worked per person were actually higher in the 1960s than in the US, and both countries had major declines in hours worked through the 1970s even in spite of stable or declining marginal tax rates. After 1982, US hours worked mostly stabilized while France's kept declining, but it's rather off to claim that MTR rates suddenly affect hours worked a ton when they didn't seem to in the 1970s.

    If I had to guess, the higher immigration rate in the US probably had a big effect in maintaining US hours worked versus France. Immigrants to the US were disproportionately young and working, meaning that the US was more resistance to the general aging of the population that was happening in all of the rich countries.

    ReplyDelete
  2. Regarding your second point, this is something that I've always felt is lacking in discussions of the labor disincentive effect of taxation:, Laffer Curves, etc. it matters ENORMOUSLY who is the effective recipient of those taxes. If the gov't took all money/resources collected in taxes and destroyed them somehow, maybe buying gold and throwing it in the ocean, or sent it all to Zimbabwe as a form of tribute, people would still work very long hours even with tax rates just shy of 100%. After all, people in pre-industrial societies worked very hard despite making very little money by today's standards.

    On the other hand if it were redistributed without condition to lower income percentiles, say, via a basic guaranteed income, and then you'll see an effect much sooner. You can mitigate this by attaching conditions to those seeking funds, and so on.

    (This is all assuming perfect enforcement of course. I mean this just as a theoretical exercise.)

    ReplyDelete
    Replies
    1. Yuck, those typos make me look like one of those crazy rambling people on the internet who types out too many run-on sentences. Forgive me.

      Delete
    2. James9:07 PM

      In theory a 100% tax rate does not necessarily result in zero revenue, as assumed by the Laffer curve. Say for example you pay a 100% tax rate but receive it all back in the form of services, you're still getting something for your work.

      Delete
    3. Names9:36 PM

      I should have said 'receive it all back in the form of goods and services'.

      Delete
  3. Nick L8:56 PM

    I've never understood why economists treat the number of hours worked as endogenous in models that ascertain the effect of a tax on hours worked. The vast majority of people sell their labor and become employees, and it is their employers who decide how many hours they work, not the employees themselves. I can't go into my boss's office and say "due to my tax rate I'm going to start working 40 hours per week instead of 50." If I did that the boss would say "thanks, here's your pink slip."

    Obviously there are exceptions to this, but for the average full-time employee, at least in the U.S. and I'm willing to bet also in Europe, your boss decides your work schedule...it would seem prudent for economic discussions to take this into consideration.

    ReplyDelete
  4. James8:56 PM

    Noah,

    "our tax system is highly progressive - rich people pay most of the taxes, poor people get most of the services."

    Even with a flat tax, rich people would pay much more tax than poor people, because believe it or not rich people have much more money than poor people. That is not an indicator of how progressive the tax regime is.

    Also, who do you think really benefits more from the state and the various things the government does? People who have next to nothing, or really rich people?

    ReplyDelete
    Replies
    1. Really rich people. It's how the get to keep all their stuff.

      Delete
    2. Michael3:42 AM

      If we measured tax burden in terms of raw dollars paid, you are correct. I prefer to measure tax burden in terms of the percentage of income paid in taxes, and I'm talking about taxation on all levels of government. Remember when Romney looked down his nose at the 47% who had no Federal income tax liability? With all levels of taxation figured in, and most taxes are regressive, many of those 47% people, still had a higher overall tax burden than Romney, when expressed as a percentage of income. Here in CT, we had Tom Foley running for governor, owner of a $10M Greenwich mansion, a jet, a yacht, and a collection of valuable cars. Foley had no income tax liability for '11, '12, and '13, and probably won't be the end of that. I'd bet everything I have that Romney would refer to Foley as being a patriotic Yankee Doodle Dandy. There's welfare for the poor, and there's welfare for the rich.

      Delete
  5. 1. All in all, counting in state and local taxes, the tax codes are roughly flat, neither progressive or regressive with many regional and individual variations.

    2. Why do we accept the premise that labor should be taxed? Tax pollution, apply Pigou taxes, tax gasoline, tax luxury goods and services but please do not tax employers and employees.

    ReplyDelete
    Replies
    1. James9:34 PM

      what about capital income?

      Delete
    2. Michael3:56 AM

      Good observation. Every state, plus D.C., all have overall regressive state & local tax systems, with the exception of Oregon. My CT state/local tax burden is almost 20%, mainly due to my property tax. Whether you have income of 50K or 5M, the overall tax burden is in the low 30's for both. The main difference is, the 50K earner is struggling to maintain a decent standard of living, and to save for the future, while the 5M person has a beautiful standard of living, accumulating more wealth, and complaining about the amount of taxes he pays.

      Delete
  6. Bill Ellis11:43 PM

    Just a reminder. While it's true our taxes are very progressive it's not the big deal righties make it out to be because our distribution of revenue is not progressive.
    Our tax system reduces inequity far less than other western democracies.

    http://www.washingtonpost.com/blogs/wonkblog/wp/2013/04/05/americas-taxes-are-the-most-progressive-in-the-world-its-government-is-among-the-least/

    ReplyDelete
    Replies
    1. Those figures are misleading.

      http://ctj.org/ctjreports/2014/04/who_pays_taxes_in_america_in_2014.php#.VKCzWFJ2Cg

      Delete
    2. Bill,

      You and others, including clearly Sumner, get your story from looking only at the federal income tax. That is p;rogressive. But much of the rest of the tax system is regressive, most notably fica, which is the largest tax paid by the vast majority of the population, as well as state sales taxes, the largest source of revenue for states. Others have provided links showing that when all of those are taken into account, the US tax system is not clearly progressive or regressive. It is mildly progressive up into the upper 90th percentile, but goes regressive as one moves into the top 2 percent and higher.

      It really is about time that we stop letting anybody get away with repeating this total canard that the US tax system is progressive, particularly that it is strongly progressive. This is a widespread phoney meme.

      Barkley Rosser

      Delete
    3. Thanks, Barkley, you said exactly what I wanted to say.

      Other than that I agree with all else Noah wrote, in his original article and his reply to Sumner. The higher MTR faced by lower income people is a very important issue, which really really ought to be tackled by somebody who's not just trying to score points for his party.

      Delete
  7. Wait, if the lump sum distribution goes to other people than the taxed, then there is also an income effect for them. In other words, the taxed might increase their labor, while the recipients of transfers would lower it. In aggregate, the impact would be only due to the differential of income effect for the two groups. In other words, Scott is right, income effect is only of second degree, and substitution effect should dominate on an aggregate scale.

    ReplyDelete
    Replies
    1. I can see what Noah is doing. He is complaining about Scot by changing the question. Scot talks about the national aggregates, while Noah, all of a sudden, switches to taxing just the rich, even though his original article talked about aggregates.

      Classy.

      Delete
  8. NoahSense1:19 AM

    Noah, you're an interesting fellow. It seems to me that your attempt to make a name for yourself is wholly as an econo-troll. You seem to be the first econ blogger to be (quasi) accepted into the popular econ blogosphere discourse via trolling alone.........good job?

    ReplyDelete
    Replies
    1. jackcieo5:09 AM

      There's truth in them there words. Might even do Noah good to hear them.

      Delete
    2. Anonymous10:06 PM

      U mad bro that he gets paid for it and you do it on a voluntary basis like the rest of us? Dont answer that, your mom needs help salting the drive way. Just shake out the tears in your neck beard.

      Delete
    3. Noahsense10:37 PM

      Am I mad Noah gets paid for trolling in the blogosphere? He gets paid for this blog? I wasn't aware. Thanks for the (unverified and likely erroneous) information, I guess.

      Also, it's quite sexist of you to assume a woman (my mom) isn't able to salt the driveway without my assistance. This is the problem with economics, the constant oppression of women..........

      Delete
  9. As to the evidence for the impact of taxes, one just needs to look at the dramatic change in hours worked in EU between early 70's and late 70's, versus much milder drop in English speaking countries. Simple question: what happened? What drove the dramatic change?

    ReplyDelete
    Replies
    1. Anonymous9:40 AM

      What a clown.

      Delete
  10. His second point is weird. We spend an awful lot of time in grad micro discussing when it is safe to ignore wealth effects in the aggregate. You need either: 1)quantities are small relative to total wealth, 2) utility functions of a very special form, or 3) everyone is literally identical. None of these apply when talking about income taxation.

    ReplyDelete
  11. Anonymous9:29 AM

    In Europe working hours are mostly governed by law - you have basic week, I believe between 32 and 40 hours, depending on the country, and overtime (that has to be paid more than regular hours). Most countries have limit on overtime, between 48 and 60 hours a week total (normal + overtime). If you worked more than that, your employer is operating illegally. You can work more by not clocking hours (but then you are working for free), by working in your own company, where you can fudge your hours, or, in some cases, by working more than one job (I don't know rules for multiple jobs). Also, in countries with lax (that is, ineffective) legal systems, employers can force employees to work more, but they are still paid only allowable hours.
    These laws come from strong unions in Europe (and I believe they were especially strong in 1970's and 1980's). And they mean that Europeans are generally satisfied with free time versus more work. They are sign that Europeans have standard of living they can keep with 40-odd hour week, not that they want to work less because of taxes.
    As for second argument, and Krzys' first comment, the argument would say that construction engineer's willingness to take one more project (because he now wants more money) is offset by guy earning minimal wage deciding that working at fast food stall and waiting tables at the bar is enough, he can stop working part-time at Wal-mart. I don't see that.

    ReplyDelete
    Replies
    1. Lots of those laws date back to the EU directive from 1993. In other words, well after the big drop.

      As to the income effect, yes that's exactly what you would expect. If there is income effect, it must work either way on all levels. Yes, there might be a differential in response, but in aggregate it leads to only second order effect. In other words, substitution effect is expected to dominate. In aggregate, not for an individual. This is in fact, the standard assumption in optimal taxation theory. It's a bit surprising that Noah does not seem to realize that.

      Delete
    2. By the way, those limits are very far from binding. At 48 hrs/week, and say 45 working weeks (excluding 4 weeks of vacation and 3 of holidays), it's still almost 2200 hrs/year upper limit, while the average working hours are something like 1500/year for the eurozone. Hence, all those legal restrictions are irrelevant. Again, the question stands: what happened?

      Delete
    3. Anonymous9:29 AM

      All the current laws date back to the EU directive because of the way directives are treated: if some country has for thirty years the law that completely follows new directive, it still has to amend it to include "based on Directive...". Directive from 1993 wasn't created in vacuum, but based on the laws from EU member states. Those labor laws were developing over 2-3 decades.

      Legal restrictions are not irrelevant, because average in US includes also those who work 80 - 100 hrs/week. Since they are blocked at 40 - 60 hrs/week in EU, total average has to fall.
      Also, overtime is not always available. It depends more on the employer than on the employee. For those without the overtime, you can use say 37 hrs/week average, giving 1665 hr/yr average working hours, or very close to total average.
      Many EU countries have special provisions for mothers of small children (well, fathers too, but mothers take it a lot more) where they can work part-time (20-something hours). There is mandatory maternity leave, lasting between 6 weeks and 6 months, and optional, lasting up to 3 years, depending on country. Also, some if not all countries do not have "sick days", but workers go on sick leave when doctor confirms they are sick, and stay there as long as doctor says, even if it's whole year.

      Delete
    4. were there any of the maximum working times laws on the books in the 70's or 60's?

      All the small issues you mention are precisely that: small.

      Availablity of overtime is endogenous: affected by wage rates (and hence taxes).

      Yes, some people work 80-100 in the US, but they are a small minority and do not materially impact the averages.

      Maternity leaves do not impact any averages, since they are out of the labor force, and sick pay/days are available throughout EU.

      All the little issues do not explain the central fact : how come with a legal limit of 48 hrs x52 weeks : ca 2500 hrs/year, the Germans, for example, work barely over 1300 hrs/ year.

      The question remains: what is the source of the divergence?

      Delete
    5. Anonymous5:39 AM

      First, legally, each employee is guaranteed national holidays (I believe average is around 14 days) and vacation time (20 work days is mostly legal minimum, in some professions even 30) with additional days off if employee has to work on weekends and national holidays. So, absolute legal maximum is 45 working weeks in a year, and often a couple less.
      Most of government employees have no overtime, and lots of companies also allow no overtime. My best educated guess is that at least 50% of employees in Europe have no option to work overtime, and most of others have available less than maximum allowed overtime hours.
      As far as I know, women (and lately men) on compulsory maternity leave are counted in workforce, they are in their companies' books, they receive regular paycheck (but from government, not employer), medical, social security, pension fund, everything continues as normal. I believe that only extended, up to 3 years leave is counted as sabbatical, as if they are out of the workforce.
      If I understand US labor law correctly, employees usually have certain number of sick days per year. If they are sick more, tough luck. In Europe, sick leave is available on doctor's orders, for as long ad doctor prescribes it.
      Some countries, including Germany, have extensive programs for part-time work, mostly aimed at mothers of young children, but in some cases other people can be included. They work 20-odd hours a week, and retain full rights as full-time employees (of course, they get paid for 20, not 40 hours, but medical, pension rights,... they keep).

      Delete
  12. Anonymous11:26 AM

    The purported incentive effect would come from marginal income(gross income less taxes), not marginal taxes. Given the high income dispersion, we still have very high marginal income per hour.

    ReplyDelete
  13. This comment has been removed by the author.

    ReplyDelete
  14. ... Second, when you tax people, they are poorer ...

    Surely he could not be that stupid. Roads, power, clean water, education, medical care, even state and national parks and Social Security and Medicare - somehow all of these seem to add to more than just quality of life, they add to life itself.

    spelling edit. grrr

    ReplyDelete
  15. Anonymous10:47 AM

    that picture is absolutely perfect for this blog post.

    ReplyDelete
  16. I'm surprised you don't reach the obvious conclusion -- taxing rich people leads to means-tested redistribution which create larger MTRs for the poor. So, taxing the rich mainly makes the poor work less. This is a major reason why there's such broad support for the EITC and other attempts to address the high MTRs the poor face.

    But Greg's original point is valid -- people with high incomes and a large amount of discretion about how much to work feel a large effect from marginal rates, and the studies he cites support this. The effect on the poor is surely much larger, but that's small comfort when you need a brain surgeon and he's gone Galt for the year.

    Also, note that the US trails only S Korea in hours worked per person. This is almost certainly driven by the greater freedom to contract for labor. We can argue whether the European model of enforced leisure leads to better overall outcomes, but it certainly leads to lower overall incomes.

    ReplyDelete
    Replies
    1. Anonymous11:34 PM

      the person above read Ayn Rand in their teens and due to the flaws in evolutionary biology it stuck with them through their adult life.

      This is maybe a good reason to not believe in God: He creates vessels capable of being filled with shit.

      Delete
  17. "Lower income people face a much higher implicit MTR than in the 1950s and 1960s."

    I also pointed out to Sumner on his blog that I don't see how this helps his case. Probably Krugman would agree and say let's go back to the 50s tax rates.

    ReplyDelete
  18. Sumner's point in mentioning the 1950-60s re: MTRs is the smaller safety net then.

    It is a serious issue. Unfortunately there's very little thought given to it by people who actually might want to help. The Rs just use it as an argument to dismantle the safety net and the Ds won't touch the issue with a 10-foot pole.

    What annoys be about Sumner is, being a fairly classical liberal guy, I resent being scolded about classical liberal fundamentals by some nutter running a blog about NGDP targeting. NGDP is one of the most leftist economic projects out there. The only way to make it work would be to automatically and proportionally ramp up public spending whenever private spending slowed. And if the public spending is wasteful and private investment slows as a result, then you've got to ramp up public spending some more to keep to that NGDP growth trend. And at the same time he wants to play the free market fundamentalist? Give us a break, you weird old nut job.

    ReplyDelete
    Replies
    1. W. Peden4:31 PM

      "The only way to make it work would be to automatically and proportionally ramp up public spending whenever private spending slowed."

      Nope.

      Do you also think that public spending had to be cut everytime private spending increased? If so, why do you argue the NGDP targeting is a left-wing (by which you actually mean pro-state) policy? If not, why not?

      Imagine if all that was required to cut the size of federal spending was for people to spend more!

      In fact, of course, the outcome of a slowdown in private spending under NGDP targeting would be, all else being equal, an increase in private spending.

      Delete
  19. To the extent that pre-tax income for the wealthy is a positional good in itself, taxation does not affect it.

    To the extent that after-tax income buys positional goods, taxation rescales the competition without (to first order) changing outcomes.

    To the extent that the wealthy imagine taxation reducing their individual incomes without recognizing the lack of positional effects — that is, to the extend that they just imagine themselves having less money, as if for another, personal reason — they will oppose taxation more strongly than rational self-interest would suggest.

    Is this confusion ever discussed?

    ReplyDelete
  20. Anonymous7:04 PM

    "... this by itself is not dispositive"

    Mandatory reading before it gets worse: https://www.mtholyoke.edu/acad/intrel/orwell46.htm

    Twice a day, until over-average disease-related bodily uncoldness disbecomes nonnegligible.

    ReplyDelete