Sunday, November 25, 2012

Taxes and cliffs

(Update: John Batchelor show radio interview on this blog post)

The whole tax debate is supremely frustrating to anyone who survived econ 1.

The ill effects of taxation -- the "distortions" -- depend on the total, marginal rate including transfers. If I earn an extra dollar, how much more stuff do I get, or how much more of someone else's services can I receive? That calculation has to include all taxes, federal, payroll, state, local, sales, excise, etc. and phaseouts.

And, if you receive a benefit from the government that phases out with income, so every dollar of income above (say) $30,000 reduces your benefit by 50 cents, then you face a 50 percent marginal tax rate even if you pay no "taxes" at all. Taxes and benefits -- both in level and on the margin -- need to be considered together.
I've been looking for good calculations of marginal rates.  The CBO has just issued a nice report titled "Effective Marginal Tax Rates for Low- and Moderate-Income Workers" that begins (begins!) to shed some light on the right question.  Here's one important graph, titled "Marginal Tax Rates for a Hypothetical Single Parent with One Child, by Earnings, in 2012";

The CBO's headline (first page) says these low-income workers
face a marginal tax rate of 30 percent, on average, under the provisions of law in effect in 2012. ... Over the next two years, CBO estimates, various provisions of current law will cause marginal tax rates among this population to rise, on average, to 32 percent in 2013 and to 35 percent in 2014.
30% and rising to 35% is already news. (A lot of the rise reflects means-tested insurance subsidies under the ACA). But digging a bit deeper I see a more chilling story in the CBO report
...CBO also finds that under provisions of law in effect between 2012 and 2014, marginal tax rates vary greatly across earnings ranges and among individuals within the same earnings range.
Consider again the graph on the top. The marginal tax rate is not an even 30%. There are slices of income where the marginal rate approaches 100%. And, the graph is really misleading because it doesn't graph the "cliffs."  The figure caption says
The dotted lines indicate income limits for Medicaid and CHIP where taxpayers face “cliffs.” Similar spikes in marginal tax rates when the taxpayer loses eligibility for TANF and SNAP are not illustrated.
The CBO's artists apprently did not want to graph the vertical spikes in an honest solid line. (The CBO's "average" is, as far as I can tell, an average across taxpayers. No taxpayer reported income at exactly the cliff income, not one dollar more or less, otherwise the "average" would have been infinite. The CBO is not taking an "average" across income, which would include the cliffs.)

Another figure gets at the situation better, I think, though it takes more sophistication to digest

Now the "cliffs" show up. Overall, disposable income is very flat from $0 to $30,000 of income, and there are swaths where a discrete jump in income produces no increase in overall income.

Cliffs are particularly pernicious incentives. Even if people overlook marginal incentives for a while, "if you take this job you'll lose your health insurance" really focuses the mind.
And, this is only the start. Single parents with one child who are going to work need childcare, transportation, clothes, and so on. The calculation leaves out sales taxes and a range of additional means-tested programs and social services.  It  only represents marginal tax rates by people who actually do work and file taxes. The jump from out of the labor force, illegal work, or disability to employment is higher. (Box 3 p. 17 cites a 36% marginal tax rate for the jump to employment and 47 percent from part time to full time.)

Even within the same income group, there is a  tremendous variation in marginal tax rates.

 The CBO' introductory graph makes somewhat the same point. The huge spread in effective tax rates is as interesting as the average value

These estimates are also  understatements, as they only scratch the surface. The actual marginal tax + loss of benefit rates people face is very complex. A quick poll of faculty at the Booth lunch table showed the usual range of opinion but no serious calculations. Maybe it's deliberately complex so people will not make the obvious responses to big marginal tax rates!

And, in the name of simplicity, the CBO left out dozens if not hundreds of additional means-tested programs. As the CBO says (p. v) "Including additional programs would generally increase estimates of marginal tax rates." Yes, it would.

Why is there so much variation in tax rate across people of the same income? (p.v) The CBO here is looking at actual taxpayers, and
Survey data show that the majority of lower-income families do not receive means-tested transfers, either because they do not meet additional, nonfinancial eligibility requirements or because they are eligible but do not apply for benefits. Of those who receive transfers, the majority participate in only one program.  
This is both heartening and chilling. Written into law is a much larger welfare state than we actually have. Americans don't fully play the game. Yet. Witness the recent ad campaign to get more people to use food stamps. Once more people take more full advantage of the programs available to them, first the budget explodes. And second, they start to feel larger and larger marginal tax rates against growing out of the programs.

All of this  is official confirmation of the point Casey Mulligan has been making in his new book: Our system imposes huge disincentives for low-income people to move up. 
Greg Mankiw (H/T this is where I learned about the report) suggests
What struck me is how close these marginal tax rates are to the marginal tax rates at the top of the income distribution.  This means that we could repeal all these taxes and transfer programs, replace them with a flat tax along with a universal lump-sum grant, and achieve approximately the same overall degree of progressivity. 
The spread in tax rates means Greg is much more right than he knows. A $20,000 "universal lump sum grant"  and 30% flat tax rate would indeed be a better system -- not because it would approximate the current system more simply, as Greg implies, but  because it would dramatically lower marginal tax rates for so many low-income families.

The idea is worth pursuing. A "lump sum grant" means $20,000 voucher for food, housing, health insurance, and eliminating all the programs and their administering bureacracies. I didn't know Greg was such a radical!  However,  $20,000 x 100 million households = $2 trillion, on top of the military and all other federal spending. It's not clear a flat 30% even with no deductions at all is going to pay for it.

Perhaps we keep the $20,000 as provided benefits, as they now are, just lousy enough that rich people abandon them voluntarily as they bail out of public schools. That limits the budget impact a bit, though it will be pretty hard to explain to a $50,000 wage earner now paying next to nothing in Federal income taxes that they'll be writing a check for $15,000 next year, but don't feel bad because now they get to use food stamps and be on medicare.

We're going decidedly and inevitably in the other direction, which is the CBO's point in its gentle and understated  admonition that the "average" marginal rate is going to rise to 35%.

The response to our budget woes is more means-testing: Leaving deductions in place but capping them, adding to the phaseouts in the tax system,  more means-testing for Social Security, Medicare, and Medicaid, all of the low-income subsidies for the ACA which phase out with income of hours on the job.

It all sounds great if you don't understand margins.  Why should the government help rich people? But every time we do cap something or means-test it, we introduce another marginal tax. And some of the biggest marginal taxes hit the poor.

This is a deeply important point so let me reiterate it. If you means-test any benefit, you introduce a steep marginal tax rate at means-testing point. If you don't means-test a benefit, you blow out the budget. It's a hard nut, that you can't get around.

This is not a little problem. We worry about the distribution of income in the US, and how low-income people seem stuck. Well, faced with these barriers of course they're stuck. And the barriers are going to get worse. 

What to do?  To some extent this is why economics is called the dismal science. Draw the line any way you want, subject to the budget constraint that all redistributed money has to come from somewhere. If you make it high at the left end it has to have a low slope. Compassion breeds "dependency," a pejorative word for the simple fact that poor people are smart and respond to incentives.

But we can do a lot better!  At least we can measure and talk about total marginal tax rates including phaseouts and benefits -- and the CBO study is only a beginning -- rather than the silly Warren Buffet vs. his Secretary stories about average personal Federal income taxes in isolation.

And, we can avoid the big variation and the cliffs. We can avoid some people facing 100% or more margins and others facing no margin. We can  bring everyone closer to the "average" 30% rate.  The costs of a high rate are larger than the benefits of a low rate (and varying rates cause people to clump up on the high rates.)

Finally, perhaps more time limit as well as income limit will work as a sensible compromise. Unemployment benefits are limited in time, which is what has kept the US from developing the permanent underclass on the dole of some European countries.

Half-joke:  the Republican response to the Democrat's desire to raise the high bracket of  Federal income taxes to 39.5%, and raise the taxes on dividends and capital gains should be: Fine. You can have the Warren buffet lower limit. In return, we get the Greg Mankiw upper limit: (named after Greg's 90% marginal tax rate) If any taxpayer can show that his total marginal tax rate, including payroll, Federal, phaseout, state, local, excise, share of corporate, sales, property, and removal of benefits exceeds 75%, then his Federal income tax rate shall be reduced to that level. Well, maybe we should take this seriously on the low end of the income distribution.

Personal story: This is how I became an economist. Taking econ 1 as my humanities distribution requirement at MIT (pause for laugh), the professor showed the budget constraint for people on welfare, which at the time reduced benefits one for one with income, and kicked people out of public housing. For years I had felt at sea in the moral and cultural arguments about welfare dependency. In a flash, I saw it, there but for the grace of good fortune go I.

Next topic. In week 2 of econ 1 you learn that the distributional effects of taxation also are not read off the headline rates of the Federal income tax, but also depend on all taxation, all spending, and the burden of taxation through higher prices and wages, not who actually pays the taxes. That political argument is even sillier.


An excellent comment arrived by email:

Dear John... Regarding your post on marginal tax rates, the best paper I’ve seen on the subject is by Larry Kotlikoff and David Rapson, “Does it Pay, at the Margin, to Work and Save? Measuring Effective Marginal Taxes on Americans’ Labor and Saving.” This includes state programs (in Massachusetts, if I recall correctly), which add even more phaseouts. (Link to the NBER version). The chart on page 45 of the file is particularly striking. [reproduced below]

[Kotlikoff's abstract is great: 
The paper offers four main takeaways. First, thanks to the incredible complexity of the U.S. fiscal system, it's impossible for anyone to understand her incentive to work, save, or contribute to retirement accounts absent highly advanced computer technology and software. Second, the U.S. fiscal system provides most households with very strong reasons to limit their labor supply and saving. Third, the system offers very high-income young and middle aged households as well as most older households tremendous opportunities to arbitrage the tax system by contributing to retirement accounts. Fourth, the patterns by age and income of marginal net tax rates on earnings, marginal net tax rates on saving, and tax-arbitrage opportunities can be summarized with one word -- bizarre.]
 This anecdote from Jeff Liebman also illustrates the issue in a way that Kotlikoff’s charts might not:
Despite the EITC and child credit, the poverty trap is still very much a reality in the U.S. A woman called me out of the blue last week and told me her self-sufficiency counselor had suggested she get in touch with me. She had moved from a $25,000 a year job to a $35,000 a year job, and suddenly she couldn’t make ends meet any more. I told her I didn’t know what I could do for her, but agreed to meet with her. She showed me all her pay stubs etc. She really did come out behind by several hundred dollars a month. She lost free health insurance and instead had to pay $230 a month for her employer-provided health insurance. Her rent associated with her section 8 voucher went up by 30% of the income gain (which is the rule). She lost the ($280 a month) subsidized child care voucher she had for after-school care for her child. She lost around $1600 a year of the EITC. She paid payroll tax on the additional income. Finally, the new job was in Boston, and she lived in a suburb. So now she has $300 a month of additional gas and parking charges. She asked me if she should go back to earning $25,000.....
[Thanks! I also am not a specialist in this literature and am glad for pointers to good work.] 

Update 2: Another graph, thanks to MG.

Source, a great presentation by Gary D. Alexander Secretary of Public Welfare Commonwealth of Pennsylvania at the AEI


  1. I was hoping to see some grumpy grumbling about the, ehem, MTR-hike of the Republican proposal of making the largest tax bracket have their average tax rate equal to their MTR. Now, there's a nice, juicy cliff. Casey Mulligan could chime in too.

    You could get a kick out of Kleven and Waseem analyzing a similar system of "notches" in Pakistan in the QJE:

    This also has some lessons about in salient kinks and notches, not being that bad, esp. with rigidities around. (Also, some recent works of Chetty on the EITC.) But all of us should be working on making the welfare state smarter, for sure.

    Speaking of misperceived kinks, some ridicule is due to fall on the shrewd value creators, the backbone of America:

    By the way, re Greg's post, there is some recent work from Andreas Peichl at IZA (I found no public copy) that most OECD countries' tax and benefit schedules are remarkably close to linear in the end.

  2. "The ill effects of taxation -- the "distortions" -- depend on the total, marginal rate including transfers. If I earn an extra dollar, how much more stuff do I get, or how much more of someone else's services can I receive?"

    No, the ill effects of taxation - the "distortions" are the changes in relative prices between two or more goods and services. Taxation in and of itself serves to regulate the relative value of the government's currency relative to all other goods and services.

    Question: Why does a government tax?
    Answer: To create a demand for its currency.

    Question: Why does a government borrow?
    Answer: To create a supply of its currency.

    Since a government creates both a supply and a demand for its currency, we can say that the government regulates the value of its currency relative to all other goods and services.

    1. That is an interesting theory, Does that mean that if a nation taxed 90% of all income, and borrowed a corresponding amount, then it's economy would be fine because, after all, it is in equilibrium?

    2. "That is an interesting theory, Does that mean that if a nation taxed 90% of all income, and borrowed a corresponding amount, then it's economy would be fine because, after all, it is in equilibrium?"

      If the federal government taxed 90% of all income, used those taxes to pay down debt, and then borrowed and spent a corresponding amount, then an equilibrium has been reached (no new net debt). I can't say whether the economy would be fine or not fine. A lot would depend on how the newly borrowed money was spent.

      If the federal government taxed 90% of all income, borrowed a corresponding amount, spent half of the tax income and borrowed money subsidizing the production of widgets and half the money on rebates that encourage the buying of widgets, then an equilibrium price would be reached for widgets that would be lower relative to all other goods than if the federal government had not intervened.

    3. Kyle8,

      "Does that mean that if a nation taxed 90% of all income, and borrowed a corresponding amount, then it's economy would be fine because, after all, it is in equilibrium?"

      From the Keynesian point of view, it would be very difficult for the economy to grow. The Keynsian multiplier = 1/(1-c+t). As taxes get large the multiplier shrinks.

  3. Now consider that Romney was paying 15% on "carried interest" and complaining that was too high ("carried interest" should be taxed as regular income). If the point you are making were better understood, Romney would have been laughed at.

    I agree that effective marginal rates are a problem. Ignoring your attempt to put excise and property taxes in the mix, there would be much to be said for having an express limit on effective marginal taxes in the 50% range as a unifying principle for the design of the tax system and social programs. If the rich claim that 35% is a serious disincentive for them, they can hardly argue that 70% is not a disincentive for the poor.

    1. Actually Romney never complained about his taxes. He wanted to increase them by capping deductions. He complained about over-all rates of taxation and complexities and distortions of the tax code.

      The problem is that everyone was putting words in his mouth.

    2. Romney never proposed ending the carried interest tax break or hiking the tax rate on capital income. Since those were the primary determinants of his effective tax rate, none of the policies he proposed would have raised his tax burden by any significant amount.

    3. As an investment professional, the carried interest rule is a joke. The assets that the manger has in the fund can be taxed as a capital gain, but the 2/20 fee is clearly income. Mangement fees are fees. Fee income is earned income.

    4. "As an investment professional, the carried interest rule is a joke. The assets that the manger has in the fund can be taxed as a capital gain, but the 2/20 fee is clearly income. Mangement fees are fees. Fee income is earned income."

      As an investment professional, this should be food for thought:

      If a private equity partnership invests $100 in a company and that company increases in value to $200, there is a capital gain of $100 (taxed as long-term capital gain if held for more than one year. The partnership agreement determines how that capital gain will be allocated to partners. In your example, that would mean $80 allocated to the LP's and $20 to the GP. The LP then pays, out of his share, $2 to the GP as a "management fee". The $2 is deductible against LP's ordinary income, if any (within or without the LP).

      The mainstream media coverage on this issue normally suggests that "carried interest" involves "creating capital gain income from thin air" or "converting ordinary income into capital gain income", as an alchemist might be able to do. In fact, the amount of capital gain to be divvied up is fixed and "carried interest" is a voluntary contractual arrangement whereby LP's agree to cede part of their favorably taxed capital gain to the manager in return for a lower management fee (and such ordinary management fee would be a deductible expense to the LP). There is a clear matching principle involved here.

      Given the above, it should come as no surprise that the majority of the equity supplied in private equity carried interest deals comes from tax exempt entities, such as university endowment funds and other "not for profit" entities that reside in ivory towers immune from public scorn. They don't care about this exchange, because they don't pay tax anyway, and have no use for deductions. Therefore, the normal "matching" does not work as it should. Doing away with the tax exempt status for these investors would be as much a solution to this arbitrage as would taxing "carried interest" at ordinary income tax rates.

    5. Well said, Vivian.

      The only thing I would add is that if the government tries to eliminate "carried interest", then it will eliminate it for start-ups as well because "carried interest" is also known as "sweat equity" outside of VC and PE firms. If the government wishes to keep "sweat equity" untouched, it will have to narrowly define the firms which are not eligible for the tax treatment. I have little doubt that the government will be as bad at defining that as it is at defining "proprietary trading".

    6. "The only thing I would add is that if the government tries to eliminate "carried interest", then it will eliminate it for start-ups as well because "carried interest" is also known as "sweat equity" outside of VC and PE firms. If the government wishes to keep "sweat equity" untouched, it will have to narrowly define the firms which are not eligible for the tax treatment. I have little doubt that the government will be as bad at defining that as it is at defining "proprietary trading".

      That's a good point, Methinks (double entendre there!)

      Failing to recognize what the *real* causes of what a tax distortion is means that the solution is likely to be misdirected. This is a classic case. There is no reason at all why venture start up firms should not be able to contract at arm's length *with parties that are subject to the same tax rules*. This used to be called freedom of contract. Sure, the start-up's owners get favorable tax treatment, but that is at the expense of new investors. There is a clear voluntary trade-off here---one party gets tax benefits and the other doesn't. Unless there is improper "tax arbitrage", the Treasury doesn't stand to suffer.

      I've always thought that the real answer to "carried interest" was not to treat it as ordinary income per se, but to ensure that the trade can't be done when tax exempt partners trade part of their tax exempt benefits with taxable partners (that is exactly what happens with "carried interest"). It takes two to Tango.

      There is, I think, a ready solution to the problem: amend the code such that the trade-off within a partnership can't be used with a "tax exempt entity or person". The "earnings stripping" rules of section 163(j) contain a handy definition of "tax exempt" that would be useful for this purpose.

      Blaming tax exempts for at least half the problem doesn't get a lot of media interest, unfortunately. Nobody has the knowledge or the guts to admit that well, maybe the Princeton Endowment Fund is part of the problem.

  4. But let's parse this a bit more in detail. Take your anecdotal story about a woman going from 25K to 35K a year. Doesn't she have an enormous incentive to get out of section 8 housing? it's usually not in the best neighborhoods, is it? Shouldn't we take that into account when we talk about incentives?

    The real issue here is lack of affordable child care for single parents who want to work.

    1. Also, she could claim up to 5,000 in child care subsidies from flexible spending account

    2. Put the other way, our government gives people like this a big subsidy to stay in places with poor socioeconomic externalities.

    3. The causality is in reverse. For various reasons, we have places with poor socioeconomic externalities, and our government somewhat mitigates the plight of those people who live in them by giving them subsidies

    4. I don't see the causality problem. The argument over the marginal rates is not about whether or not we are trying to help them, it is about whether or not we are also punishing them for potentially trying to help themselves. The moral hazard is that the way we ensure a lowest standard level may also be an exile to that level.

    5. Marginal rates matter to the extent that they create disincentives to work. But people who are low income in the U.S. tend to live in violent neighborhoods with poor schools. They have enormous incentives to get out of those neighborhoods. There is no way that any marginal rate eliminates those incentives. The real issue is they are unable to.

    6. $10,000 is $10,000; if you would be $10,000 poorer accepting more work, that is a huge disincentive: the kind that makes you unable to provide food for your elderly parent or young child.

      On top of that there are a lot more poor people in places where I think you've forgotten to look because it is so easy to do so.

    7. anonymous you are changing cause and effect based on assertion. "they" (low income individuals) are unable to get out of violent neighborhoods in large part because of incentives to stay there thanks to modern welfare state. and why are the neighborhoods "violent"? lets take a holistic look. what about the "war on drugs" that turns poor blacks into criminals and incentivizes them to carry and use guns to enforce extra judicial "justice."

    8. In order to fix this problem with low-income individuals, you first have to see how the people have to live that are on the very edge of qualifying or if they are just out of reach of government assistance.

      My wife and I barely make above minimum wage, and after all we have to pay for if we do not budget correctly we barely have food to eat. Since we make above what the government deems "necessary" for government assistance, we are denied any help.

      This wouldn't be so bad if there weren't people who do not properly report their income are taking advantage of the same system denying us to get plenty of food stamps and welfare...

      Maybe taking some time to really screen and follow up on what people are reporting as income and how they're really living would help the money go to families who ARE actually in need.

      Great blog, John, I'm glad there are people like yourself out there helping people take control of their taxes and understand more about the financial world around them. Keep up the great work!

  5. Great post. I do have one complaint: I truly wish that economists talking about this issue would come up with a different term than the highly misleading "marginal tax rate". A phase out of benefits often has nothing to do with taxes and is not the result of any provision of Title 26 of the United States Code. You would be doing lay readers, who often find this most confusing, and even your fellow economists a big favor by coming up with something more accurate and appropriate. As noted here, one can have a high "marginal tax rate" even though no taxes are paid. I suggest that economists have a contest to come up with a better term that does not compete with and confuse the original meaning that has an entirely different connotation.

    1. This is a good point. Some economists use the word "wedge." As in, "the wedge between what I produce and the value of what I can buy is 70%." That's good because losing a benefit and paying a tax obviously count equally, and because it's clear we need to talk about all taxes and distortions that lie between working an hour and eating a pizza. But if we use that word, will anyone understand it?

    2. "We need to talk about all taxes and distortions that lie between working an hour and eating a pizza."

      We need to talk about all taxes and the distortions that arise when the production and consumption of pizzas are treated differently than the production and consumption of computers.

      The tax "wedge" exists only in the sense that it changes the relative value of one good or service versus another. Tax policy that affects the value of all goods and services relative to the government's currency is simply monetary policy by another name.

    3. "But if we use that word "wedge", will anyone understand it?"

      I doubt it, but why does it sometimes seem that economists have so little imagination? One could come up with any number of better terms than "marginal tax rate" or "wedge" for what is being described here. Here's my contribution to the contest: "Marginal Disincentive Rate", or MDR for short. Non-experts should readily understand it because It's even much more descriptive than "marginal tax rate", and much, much less confusing.

    4. I like this. I'd call it "Marginal Penalty Rate", but I'm just trying to help here. Thanks for the idea. I too believe that it should make both writing and reading about this topic easier.

    5. The number is 1 - Change in Disposible Income/ change in earned income.

      marginal income suck?

    6. I disagree. ”Tax” is precisely the right word for when the government takes something away from me as I earn more money.

      What true confusion exists (vs. ideological or partisan hackery) concerns the meaning of marginal vs. effective rates. Jeff Liebman's anecdote perfectly illustrates how this otherwise simple mathematical relationship is beyond many people.

    7. "I disagree. ”Tax” is precisely the right word for when the government takes something away from me as I earn more money."

      No, it is precisely the wrong word. When the government "taxes" in the traditional and precise sense, it takes something away from me that was mine in the first place. When the government phases out a benefit that it might otherwise provide (funded by *taxes* or by borrowing (as per Friedman, a form of *future taxation*), it is declining to spend something on me. Never was the difference between "taxing" and "spending" more cleverly disguised than combining both into a "marginal rate of tax". I see where this clever bit of obfuscation might be ideologically attractive to those who take the term "entitlement" literally, however.

    8. In the U.S. today, there is no difference between “that [which] was mine in the first place” and an entitlement.

      It may not be right, but it is nonetheless legal, practical, and ideological reality.

  6. Yes and all this does not even include all the other taxes the working poor pay. Remember the federal income tax is only 30% of all income taxes paid in this country


    A minimum wage full time worker aged 20
    A widow living on Social Security aged 75

    Both have an annual income of about $15,000

    They are among the 50% that the top few percent are forever claiming "pay no taxes." But the actual true statement is that they "pay no federal income taxes."

    However, both pay over $5000 in other taxes, like Social Security, sales, gas and real estate taxes - over 30% of their meager incomes, and no .. they do not qualify for the Earned Income Tax Credit or Food Stamps. They have to live on under $10,000 after taxes. On the other hand Warren Buffett pays only 11% of his 8 billion dollar investment gain .. yes, including his share of indirect corporate taxes .. leaving him $7,200,000,000 in investment gains in a recent year.

    The other taxes people like the minimum wage worker and Social Security benficiary pay are conveniently ignored by the top few percent and their pets in Washington when they claim millionaires can not pay a penny more in taxes as long as so many "pay nothing."

    See the details and the solution:

    1. No, they are not ignored. The minimum wage worker will pay more than the Social security recipient because he will pay FICA. Many free market economists have long pointed out the regressive nature of this tax and have advocated for a change in those entitlement programs.

      Warren Buffet pays taxes on dividend income, and again pays capital gains taxes when he sells stock. But it is misleading because this income has already been taxed. (and those companies he invests in pay sales taxes as well, and their employees pay taxes)

      Furthermore, if he purchases anything with his money then he also pays the sales taxes (and perhaps luxury taxes).

      People with a bleeding heart mentality are quick to point out the things that you point out. But often fail to realize that there is a trade off between economic growth and taxation. Taxation at all levels is harmful, but at the upper margins a small increase can have a large effect.

      Which helps poor people the most? More government bureaucracy and handouts, or a dynamic economy with rising wages?

    2. great post kyle8, thank you. the answer can't be to find new rationalizations and ways to tax the so called rich. the answer must be to keep a lid on the growth of government. every successful civilization in the history of world committed suicide by raising taxes too high (either directly or indirectly via debasement of currency). this is the provocative thesis (my words not his) of tax historian charles adams in his amazing analysis of tax history, "Taxes: For Good and Evil." a must read.

  7. Reminds me a lot of this post from AEI over the summer, incidentally:

  8. This is fantastic. The second graph tells the tale. Now I see what you were trying to explain to me in the other post about taxes and margins. I see this all the time with disabled people on SSDI, and retired people on S.S. do these calculations continually.

  9. Thanks for posting about the CBO report, but your first chart does not include the healthcare bill's marginal tax rates. Dig deeper and look at the full report:

    Page 34 shows the same single parent with one child but adds in the health care bills affects of marginal tax rates. With the health care bill the marginal tax rate is around 70% from $10,000 to around $37,000 for the same family!!!

  10. Conor et al,

    This is another great study this time from the state of PA

    I would also like to see this more comprehensive cliff analysis including other forms of redistribution (some smaller state programs that are hard to keep track off, but also significant non-state redistribution initiatives, like need based scholarships at pricey colleges).

  11. If all these benefits are phased out, rather than simply subjected to a hard cutoff, wouldn't these cliffs go away? As long as the net income graph is increasing monotonicly things don't seem too disastrous (the slope may be too low to overcome the marginal utility of leisure, but that happens anyway)

    1. In a word, yes. The CBO states: "[T]he phasing out of that assistance adversely affects incentives to work. As a result, lawmakers face difficult choices when targeting benefits to low- and moderate-income families. One choice relates to phaseout rates—specifically, whether benefits should be phased out gradually over a broad income range or more quickly over a narrower income range. The former approach would affect more families but allow for lower marginal rates of benefit reduction."

      The former approach, I believe, is most in line with what you are suggesting, Anonymous. It is the approach I prefer most. I am curious as to whether Professor Cochrane and others prefer this approach -- phasing out gradually over a broad income range to allow for lower marginal rates -- or prefer the other approaches listed in the report:

      The latter approach [phasing out more quickly over a narrower income range] would increase marginal tax rates more significantly for a smaller number of families in a narrower income range. Another choice concerns the maximum amount of the assistance, which also influences marginal tax rates. Over the same income range, phasing out a larger benefit leads to a higher marginal tax rate than does phasing out a smaller benefit."

      If someone's primary concern is alleged disincentives, my guess is that his or her top priority would be smaller benefits.

  12. Why would people abandon public schools? Corrected for parent income, the racial achievement gap, and the presence or absence of two parents, public schools perform as well or better than their private counterparts.

    What people perceive as school quality is overwhelmingly the result of the predisposition students have to succeed or fail based on demographic features. The correlations between these demographic features and the educational outputs of the students are robust, persistent, and vastly explanatory of any observed differences between different kinds of schooling. All of this is merely to say that the last four decades of developmental psychology have accurately observed the power and permanence of environmental impacts on young brains.

    1. Hmm, I wonder where President Obama and Chicago Mayor Rahm Emanuel send their kids to school? I presume you don't have kids and don't live in Chicago.

    2. In what district, and on what planet?

      Government schools, even in districts with the demographics you mention, are mediocre at best. That's my personal experience with my own kids, and it reflects the experience of every parent who's been disgusted with the poor quality of their local government school.

    3. John: for the record, Chicago public schools are pretty good. I have plenty of friends who send their kids to Chicago public schools. The problem with Chicago public schools is lots of kids who come from bad environments. But if your kids are in "gifted" programs or on accelerated track, the quality of education is excellent. Northside Prep? Walter Payton? those are some of the best schools in the nation, if not in the world and they are public.

      So it's unfair and biased to bash Chicago public schools.

      Don't confuse causality here. Do you expect teachers to perform heroics on kids who come from broken homes and broken neighborhoods?

    4. Some Chicago public schools are good. I graduated from one. Most are awful, and consign their charges to a lifetime of poverty. Charter schools and catholic schools do much better with the same students. The statistics and academic literature are just overwhelming. If CPS were so great, it would not need to worry about competition from charters and vouchers.

    5. Do you have any links to any literature that I could look at which controls for the student quality in order to judge the school quality?


      From the study:

      On average, across varying communities and policy environments, charter middle and high schools produce achievement gains that are about the same as those in traditional public schools

    7. also from the study:

      "Students who attended charter high schools were more likely to graduate and go on to college."


      "They also suggest a need to look beyond test scores to fully assess charter schools' performance."

      I think the last point is very important and is not so satisfactorilly explained by demographics alone

      thanks for the link!

  13. One other consideration is that if you attempt to offer everyone a broadly similar effective marginal tax rate, you are ignoring the fact that people have different levels of responsiveness to financial incentives. For instance, if you have young kids and a partner who works full time, you have a genuine choice about whether to work or not, and financial incentives will be a big part of your decision. But if you are a sole breadwinner, or adore your job, you're more likely to work regardless. Emmanuel Saez has a nice article on this (Using Elasticities to derive Optimal Income Tax Rates).
    So when we think about presenting an effective marginal tax rate schedule to an individual, the most efficient way to do it is to try to reflect those differences. Someone who is only ever likely to work part time should have lower EMTRs than someone who is likely to go for full time work. I first encountered this idea in an article by George Akerlof (The Economics of "Tagging" as Applied to the Optimal Income Tax, Welfare Programs, and Manpower Planning).

  14. I would love to see any evidence that the truly poor have a good understanding of their effective marginal tax+benefit rates and use that understanding to make strategic employment decisions.

    I would also love to see any evidence that they would use a lump sump primarily to care for their children.

  15. Great post. But I think your discussion of the promise of, and problems with, a "guaranteed income plus flat tax" regime is missing a couple of points. First of all, some-- perhaps many-- current welfare programs provide insurance against cost variance, not (or not just) funding of ongoing ordinary consumption. This isn't easily replaced by a flat income grant, though it can be replaced similarly with a little ingenuity. Second, if you are worried about the high MTR figure that might be required for an income tax to fully fund a guaranteed income, one obvious strategy is to fund it partly out of other sorts of taxes that provide less disincentive to work: consumption taxes, say, or Georgist/Pigouvian externality taxes.

  16. I think you left out the most interesting part of the quote in the extended anecdote. I'll reproduce from the original:

    "She tried taking an additional weekend job, but the combination of losing 30 percent in increased rent and paying for someone to take care of her child meant it didn’t help much either."

    This is not a story of substitution effects--it's the income effect. This unfortunate woman faced a marginal tax rate greater than 100%--she could increase her hours worked and see her take-home pay fall, and her ability to secure goods and services decline. If it's all about substitution, she should go back into more "leisure". But it has little to do with substitution: her initial response to the fall in income was to work even more! She only tried to "work less" when she failed to earn more by getting a second job.

    If a marginal tax rate >100% doesn't stop a person from trying to work more, then marginal tax rates can probably be said to be irrelevant.

    1. what if they try fail and then revert? will that convince you?

  17. This is a great article. Articles like this tend to focus on the impact of effectively high marginal income tax rates on low-income people. But Republicans and Democrats have agreed in the past (and continue to do so) to stealthily increase marginal taxes on higher-income households, which they do by phasing out deductions. Examples include ROTH IRA contribution limits or means testing Medicare Part B and Part D(currently), so-called "Cadillac" health plans (forthcoming), and limiting the mortgage-interest deduction (under discussion).

    I suspect that the analysis of the effects of these polices are more complicated than the ones for lower-income households. I would very much appreciate Prof. Cochrane pointing us in the right direction on this matter.

  18. My problem with the Kotlikoff - Rapson paper linked above (which is quite interesting, to be sure) is that it uses ESPlanner, a propriety software. Not sure how one could replicate results using the same data, if you do not have access to that simulation software.

  19. Ed Prescott and Lee Ohanian have an Op-Ed today in the WSJ in which they state that the *average* "marginal effective tax rate" in the United States is 40 percent. That is, on average, for every $1 earned, 40 cents goes to various taxes, including payroll taxes, income taxes, sales taxes, excise taxes, etc.

    It should not go unnoticed here that Prescott and Ohanian actually use the term "marginal effective tax rate" correctly---they do not confuse it with other disincentives, such as loss of federal, state or local benefits.

  20. Could one not setup government benefits to taper off with earned wages so that the net income of any individual is (almost) strictly increasing with increasing wages? So, e.g., rather than losing your health insurance you lose a percentage of it when you go above a certain threshold and this occurs in a number of phases. Such a model, if could be implemented effectively, would get rid off the cliffs.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.