Taxes on capital

I forgot to link to my most recent essay for The Economist.

Greg Mankiw has related piece, which makes my essay look overly simplistic.

PS.  I’ll be speaking at Boston College on Wednesday, March 14th in  Fulton Hall, Room 511, at 7:30pm.  I’m told that it’s open to the public.

Update: There’s a new blog by someone with market monetarist tendencies, and he’s a high school student (named Evan Soltas.)  If he’s already doing analysis this good in high school, I guarantee he’ll be far more successful than me by age 27.



48 Responses to “Taxes on capital”

  1. Gravatar of Martin Martin
    11. March 2012 at 08:08


    Two points:
    1. Have you heard of New Dynamic Public Finance? If I recall correctly Kocherlakota, one of the contributors, makes the argument that asset taxes should depend on labor income and that asset taxes from a social insurance perspective are non-zero.

    2. “If the current distribution of wealth is deemed unjust, then there should be a one-time wealth tax at the point of transition to the new tax regime, to be paid over a period of several years.”

    You’re going to have so much trouble with this. Just re-read it without the if-clause. People will ;).

  2. Gravatar of Negation of Ideology Negation of Ideology
    11. March 2012 at 08:11

    I understand the point you’re making in the column, but shouldn’t you use two different tax rates in your comparison?

    In your example, the total tax paid by the brothers is $80,000 in the payroll tax, while the total tax is $104,000 under the income tax.

    A fairer comparison would be the revenue neutral comparison – a 40% payroll vs. a 30.7% income tax for your example. I don’t know what it would be over the entire economy.

    Now maybe you’re assuming incentive effects that would make it closer, but there’s a limit to that because the second brother still has more after tax income so he has some incentive to invest.

  3. Gravatar of Beckett Beckett
    11. March 2012 at 10:21

  4. Gravatar of Greg Ransom Greg Ransom
    11. March 2012 at 10:54

    How good is Evan at math?

    If he doesn’t have a taste for math & some talent for math, he would need to chose a different career.

    Understanding economic science as an explanatory science has nothing to do with what is filtered for in the top Econ grad departments.

    And everyone knows it.

  5. Gravatar of J.V. Dubois J.V. Dubois
    11. March 2012 at 14:10

    I do not understand. Last time I checked, capital itself is not part of the tax. So if one year one brother put $60,000 into some project, and next year he would sell it for &66,000 he would pay tax only from the difference, not from the whole income again. He would only pay 40% of $6,000.

    The only valid objection I have to this is inflation. You would have to adjust your initial capital investment for inflation so that it can be correctly substracted from tax base when you finally monetise your capital gain income.

    But look, I do not object in general. In an ideal world, it would be best to have just value added tax. It is the embodiment of “broad base, low rate” tax rule as avery investment is consumed at some point. If you would have 40% VAT in your example it would immediatelly change the game as both brothers would have to pay 40% of their income in tax no matter what its source was.

  6. Gravatar of Alan Cole Alan Cole
    11. March 2012 at 14:37

    J.V. Dubois –

    A 40% VAT in Scott’s example would have the exact same result as a 40% payroll tax.

  7. Gravatar of TravisA TravisA
    11. March 2012 at 15:57

    Scott, I completely agree with your argument, but drawing the line between wage and capital income is often difficult.

    Should Buffett be taxed at the regular income rate on his Berkshire stock holdings? What if he originally invested the money in Berkshire from his savings?

    What if you don’t have a regular job, but buy and sell stocks with your savings? Does the answer change if you work 2 hours a week at McD’s?

    Also, how would you tax Mankiw’s different scenarios?

  8. Gravatar of J.V. Dubois J.V. Dubois
    11. March 2012 at 16:10

    Alan: No it would not have the same results. With 40% payroll tax only and no capital gains tax the thrifty brother would get to pay only $40,000 on taxes (his initial wage) while he would get to consume $120,000. That is 33% tax rate, right?

    And as for the Scott article, his analysis is obviously not true. The thrifty brother paid total tax of $64,000 from his income $160,000 {100,000 wage and 60,000 capital gains). That is exactly 40% not 52% as Scott tries to imply.

    Where is Scott’s 52% coming from? It is coming from his assumption that the thrifty brother would get to consume $200,000 if he would not have to pay any taxes {and double his wealth in 20 years). This is similar trick as with those math jokes when you mix subtraction and addition to create a paradox of missing number. Where is the trick here? Scott does not count with time value of taxation. Let’s imagine that government also has ability to double the value of taxes in 20 years and then deliver it to you. Then suddenly, we could say that thrifty brother get to consume $200,000 ($96,000 of his own income, $80,000 from doubled tax money he paid 20 years ago and $24,000 he paid today).

  9. Gravatar of dwb dwb
    11. March 2012 at 16:39

    off topic: Krugman has a post on North Dakota. totally wrong. Supply side policies DO matter (witness the inability to develop shale gas in western MD or NY despite low populations).

    Also there is LOTS of oil in Los Angeles: it was founded on oil. population is irrelevant. google oil rigs in Los Angeles!!!!

  10. Gravatar of ssumner ssumner
    11. March 2012 at 17:00

    Martin, If we are going the social insurance route, I’d rather do so with a luxury tax, not a income tax.

    2. I welcome “trouble.”

    Negation, I was just trying to show the unfairness of income taxes. It would show up either way.

    Greg, He’ll get a fine education at the school of Krugman/Bernanke/Svensson.

    J. V. See Alan’s comment, he’s right.

    TravisA, I agree that there are tough calls. But most people like me should just pay a progressive payroll tax. There are tough calls with any tax system, including our current system.

    If people are running a firm, or work to fix up places and sell them, then I’d assume the income from those activities is wage income. For the vast majority of people (like me) it’s an easy call–we are passive investors and hence should pay a zero tax rate on interest, capital gains and dividends.

    The goal of tax reform should be to put H&R Block out of business (except for residual business handling the self-employed.) I believe that in Sweden they simply present taxpayers with a bill. I like that approach. The government should tell me how much I owe them, not ask me to solve fiendishly complicated math problems.

    JV, You said;

    “Alan: No it would not have the same results. With 40% payroll tax only and no capital gains tax the thrifty brother would get to pay only $40,000 on taxes (his initial wage) while he would get to consume $120,000. That is 33% tax rate, right?”

    No, you can’t add current and future consumption as if they are valued equally. That’s like adding blueberries and watermelon, and calling the total “number of fruit.”

    The present value of that future consumption is only $60,000. So if $40,000 in taxes is paid in taxes right now, then the tax rate is 40%.

    What I’m saying here is not at all controversial among economists, both liberal and conservative public finance economists agree with me.

  11. Gravatar of ssumner ssumner
    11. March 2012 at 17:03

    dwb, Good point. Here’s another example. Lots of progressives like to argue that Texas’s success has nothing to do with a good economic model, even though neighboring states that also have lots of oil, but much less efficient tax structures, do far worse.

  12. Gravatar of AFG AFG
    11. March 2012 at 17:14


    The flaw in your argument is your narrow view of “Human Capital”, subsidizing education only favors one type of human capital over the others.

    1. Work experience – A carpenter builds his skills and gets more efficient the more he works. Wage taxes discourage work, which is probably the most important means of investing in human capital.

    2. Healthy habits and Medical Expenses – Eating well, exercising and treating disease all make a worker more effective and thus are investing in human capital. The payroll tax penalizes the return to these things.

    3. Immigration – Wage taxes discourage people from “importing” their human capital from other countries. If their human capital is more valuable here (Mexican farm workers making 10x the wages for the same job they could do there), then this is inefficient on a worldwide scale.

    4. Locked In Effect – If I find out my human capital is actually better suited for being a doctor than a carpenter, then a wage tax lowers the returns to taking a job where I have a higher marginal productivity. This is equivalent to the argument that says capital gains taxes discourage capital from travelling to its most efficient use.

    You like to say that “inflation” is a meaningless concept, everything should be viewed as NGDP. Well “labor” is a meaningless concept, everything should be viewed as different types of capital. Taxing wages over capital gains only makes sense insofar there is a distinction between “using” and “investing” in capital, but my examples above show how many of the things we consider labor actually increase the human capital base.

    The only way to have an efficient and progressive tax is to tax things that are immutably related to earnings (Mankiw once suggested, satirically, that we tax height), but you name me any tax, other than a lump sum tax, and I guarantee that I can show how it affects investment in some type of capital. You will probably argue that wage taxes are net better, because at least some of one’s wages reflects natural abilities. I will say two things:

    1. Same goes for the things you call “capital gains”. Your belief in EMH allows you to treat income for savings as merely delayed consumption (Present Value!), but even if we accept EMH investing is a skill based on natural abilities. In order to buy an index fund and hold it forever one must be smart enough to know they can’t beat the market and have enough self-control (something we know is partly or mostly genetic) not to sell when the market crashes. This is why we pay other people to invest for us. This is why there is no difference between labor and saving. The process of saving is labor whether you decide to do it yourself or pay someone else.

    2. You are far too much of a biological determinist if you think the majority of a payroll tax falls on so-called labor (something not improvable) and not on human capital (and thus investment).

    3. We may subsidize education. But we due tons of things to subsidize traditional “saving” as well.

    See here for more:

    I am going to have a much longer post of the subject soon. Hopefully it will help demonstrate how acknowledging the existence of human capital greatly complicates your views on this issue.

  13. Gravatar of AFG AFG
    11. March 2012 at 18:22

    I forgot to add that the logical conclusion of your EMH viewpoint is not at all what you think it is. Remember, the argument that savings is not income, but merely deferred consumption, rests on the idea that everyone’s returns will adjust to the same risk-weighted outcome in the long run.

    There is an equity problem with this argument. People with more existing wealth can accept higher risk and thus generate higher returns. Risk-adjusted returns may be equivalent, but at the end of the day the rich can use leverage to get more money. Shouldn’t we tax the higher interest rate that the wealthy can earn on their savings (by taking risks) and use it to fund a true safety net that allows the poor to take risks, invest in their human capital, and become entrepreneurs?

    But there is also an efficiency reason to do this. If EMH is true, then any capital gains over the market rate are pure luck, therefore a tax cannot disincentivize them. I propose that instead of using the purchase price as the base for calculating a capital gain, we use the purchase price adjusted for two times NGDP growth over that time period. If you buy any asset at $100, NGDP grows by 10%, and you sell it for $200, then you owe tax on $80 (200-(2*.1*100)). We can then tax that $80 at a 70% or higher rate.

    This will still allow the “I can win the lottery one day” effect for people to start businesses, but it will allow those that are lucky to fund the government. Zuckerberg would owe a lot of tax once he sells some stock. What do you think?

  14. Gravatar of dtoh dtoh
    11. March 2012 at 18:57

    A few things.

    1. The same argument can be made about foregoing current consumption and investing in human capital (education, training, etc.). Why is there double taxation on the extra wage income they earn later.

    2. I think you will have a very hard time separating income on capital from wage income, whether it’s at a hedge fund or a restaurant.

    3. The biggest argument against taxes on capital is that with asymmetric returns , it will quickly lead to negative average expected after tax returns. Quick example:

    10 Entrepreneurs
    20% Tax Rate
    Each entrepreneur invests $100 ($1000 total investment)

    After some period of time 9 entrepreneurs lose all their investment, but one realizes a gain of $1200. Total returns for the 10 entrepreneurs is ($300) and of course the average pre-tax return is 30%. Since the tax rate is 20% the 9 entrepreneurs who lost their investments pay nothing but the one successful entrepreneur pays $240. So as a group, on an after tax basis the entrepreneurs are left with only a $60 gain on an investment of $1000 or on average a 6% after tax return.

    If the tax is raised to 30%, then the investors who lost money still pay nothing, but the successful investor now pays $360. As result after investing $1000, the entrepreneurs are left with $940 which on average works out to a 6% after tax LOSS.

    With more asymmetric returns you will get negative expected after tax returns even at a 20% or 10% or 5% tax rate.

    Bottom line, even a low tax rate on capital is a huge disincentive to investment.

    4. Why even have a wage tax. It just complicates things and there are high administrative costs.

    5. The easiest thing would just be a simple progressive consumption tax, which could easily be implemented by having a fixed rate (e.g. 20%) and then giving everyone a card that will exempt them from tax on their first $20,000 of purchases.

  15. Gravatar of marcus nunes marcus nunes
    11. March 2012 at 18:59

    Another High School MM started bloguing today:

  16. Gravatar of Jim Glass Jim Glass
    11. March 2012 at 19:53

    I agree with everything in your Economist piece, but oh, what to do about it? There is a modest proposal — not new or original to me, nor theoretically pure, but potentially a practical improvement.

    Tax policy is extremely political, so any reform proposal to increase tax efficiency has to be *easily* politically sellable to both sides. Meaning to the typical and median voters who, as per public choice, are profoundly ignorant of the economics of taxation. (Obscure analysis of incidence and such isn’t going to convince *any* partisan skeptic.)

    Our last serious success was Tax Reform Act of 1986: The right got the top tax rate slashed to 28%, the left got mass extermination of many tax breaks for the rich (not least, favored treatment for capital gains). That was win-win enough to be enacted with full bi-partisan sponsorship. Governments the world around copied it, from Sweden on (even if we’ve been subverting it ever since). ISTM a similar political deal should be possible for capital investment, *if* the political forces were interested in it (which they aren’t). To wit:

    That capital shouldn’t be taxed is the reason most economists believe in consumption tax, but adopting one will never fly with the public: “the rich” should be taxed more, consumption taxes are regressive, etc.

    At the same time the US preferential tax rate for capital gain is severely defective in multiple ways — it is not a tax on capital (even at a favorable rate) but a tax on capital *transactions*: selling and re-allocating. It applies whether the capital subject to it is fully reinvested or consumed, hardly a logical “income” tax at all. As a transaction tax it impairs efficient capital allocation on its face, and incentivizes massive gaming of transactions to avoid the tax (say: “the tax shelter industry”). The rationale that it compensates for inflation is bogus, as there is zero relation between tax rate and inflation during the holding period. The average voter who considers it an income tax thinks that as a special rate it is either too high or too low, so it is politically unstable. In total, pretty bad.

    All this could be logically resolved simply enough. Many tax analysts have noted over the years (including in past comments here, I’d bet) that the current income tax can effectively be converted to a progressive consumption tax by creating a “super IRA” type account that can hold investments on a tax-deferred basis in *any* amount. Account funds cannot be consumed directly or indirectly (spent or borrowed against). Funds in the account are not taxed as long as they are invested or reinvested through it. Any amounts withdrawn are fully taxable as ordinary income. All favorable tax rates for capital income are repealed, there are only ordinary rates.

    A point of great political practicality is that such accounts already exist. IRAs today hold trillions of dollars (literally) of assets, have basically these rules, have operated stably for 30+ years, and everybody is familiar with them. They aren’t speculative, they work, the tax regs for them have already been worked out, etc.

    I’ve been able to meet the “but this just helps the rich!” argument among my egalitarian friends here in Moscow-on-the-Hudson with an illustration:

    Paris Hilton and her cousin Scrooge McDuck Hilton both live off of capital gains reaped from the bequest of old Barron, receiving the exact same amount of income subject to 15% tax. Paris, we know, spends hers on yacht parties hosting Faberge egg-throwing contests to see who can make the biggest splat. Her cousin is so little-known because he neurotically lives like a monk in a basement, reinvesting every penny. Paris is consuming the nation’s wealth. Her cousin is building the nation’s wealth (to be better able to pay for Medicare et. al. in the future.) Should they pay the same tax rate?

    Paris is being subsidized to consume the nation’s wealth. Her cousin is being penalized for building it up. Can that be right? Suppose they paid the same amount of tax total, but with Paris paying a 30% rate and her cousin 0%. Wouldn’t that be both better and more fair all around? Yeah, most of my liberal friends do get that. To those who don’t I point out Keynes’s observation that the great advance in human welfare came when the super-rich stopped spending on themselves via palaces and harems and started living much more modestly to invest in factories, railroads and banks. As long as people are investing instead of consuming, they are making *others* better off. That shouldn’t be taxed. When they consume they benefit themselves and should be taxed — at ordinary rates.

    This resolves the “Buffett tax” issue. To the extent Warren spends on himself, from Cokes to corporate jet, he’d pay a much higher progressive tax rate than his secretary — no 15% rate for him on money he spends. To the extent he keeps funds invested to benefit society he pays zero, as he should, and helps finance her future Medicare. Add corporate taxes and partnerships. If money is made available to consume, ordinary tax. If invested, no tax. If Stephen Schwarzman has a birthday he can throw a $5 million birthday party for himself paying tax on the $5 mil at ordinary rates or buy himself a cake from Carvel and keep the $5 mil tax free — but he can’t pay for the party using $5 million taxed at 15% through carried interest.

    In my experience, average people understand this picture clearly enough, no obscure econ analysis needed. That makes a TRA 86-type reform seem plausible. The left gets a *higher* tax rate on the rich on income that is *consumed* — no more 15% rate for capital liquidated to fund consumption, and no more tax shelters gaming the preferential-rate transaction tax. The right gets tax-free investment and wealth-building (as long as funds actually are invested, not gamed into consumption). And the mechanism to do it is already familiar to everybody and used on a large financial scale by millions.

    This is not a new idea, it’s been around a long time. Why hasn’t anybody pushed it?

    I think because political forces are much more interested in pushing impossible-to-phony proposals that motivate their own partisan base than constructive proposals that solve problems that don’t motivate their base. (Public choice logic.) But then after all my years doing tax law work, how could I not be a cynic?

  17. Gravatar of dtoh dtoh
    11. March 2012 at 20:03

    @Jim Glass
    You don’t need a super IRA to create a progressive consumption tax. All you need to do is eliminate all income tax, impose a a flat consumption tax on everything and then give everyone a card that exempts them from tax on the first $20k in purchases. If you want a graduated progressive tax, the card could give you a 100% exemption on the first $20k of purchases, 50% exemption on the next $20k of purchases, etc,. The whole system could be bid out to Master Card or Visa and administered at almost no cost.

  18. Gravatar of Martin Martin
    11. March 2012 at 23:54


    “Martin, If we are going the social insurance route, I’d rather do so with a luxury tax, not a income tax.”

    Wouldn’t a luxury tax very inefficient, due to the elasticity of demand of luxury goods?

    Anyhow, if you’re interested, I found some slides by and with Kocherlakota’s point. I used to favor zero asset taxes before reading about it, now I am not so sure.

  19. Gravatar of Brito Brito
    11. March 2012 at 23:55

    Man I’m jealous of that Evan Soltas kid, it’s very rare for us brits to go straight to a top US university straight out of secondary school, it’s hard enough getting into our own elite schools which are already among the best in the world. Makes me feel so inferior :(.

  20. Gravatar of Brito Brito
    12. March 2012 at 06:00

    Oops, disregard the above, I saw Exeter in his bio and him talking about the UK and just assumed he was from Exeter in the UK, I forget sometimes that the US steals all our place names. >:)

  21. Gravatar of JL JL
    12. March 2012 at 06:33


    So in your scheme someone could save up $1 million and receive $50K annual, untaxed capital income, while simultaneously benefitting from the progressive payroll tax due to his (near-)zero wage income: payroll taxt refund, food stamps, Medicaid and/or subsidized housing.

    We could means test the redistributive schemes, but then we’re back where we started: (a) people must file their capital holdings/income with the IRS, (b) the thrifty brother is punished for his thriftiness and (c) the original income (the $1M) is effectively being double taxed.

  22. Gravatar of Morgan Warstler Morgan Warstler
    12. March 2012 at 07:14

    Mankiw’s piece proves your wrong!

    Notice is ALL his examples – he’s assuming the deferred consumption is nothing like your assert. As I keep saying this, your own personal investment strategies are not important.

    $1M made in 1 year on $100K invested HAS NEVER been taxed.

    And Mankiw dices and splices but should basically end up here:


    Which means, in his examples:

    The guy who actually did the work, and deferred payment THAT GUY should pay less than the guy who just put in cash.

    This isn’t about the incentive to save. Any idiot can save.

    This is about the incentive to work based on risk. That’s a fine and noble skill. A waitress is a better worker than a McDonalds order taker – because she lives in part of tips.

    This doesn’t make her a entrepreneur, but it does explain the theory I’m putting forward.

    Waitress > McDonalds employee

    Sweat equity > Cash investment


    Scott YOUR PROBLEM, is that you refuse to respond to my trump card – you want to make this about what the saver gave up, and I’m arguing thats nice, but my guy who risks labor is more important.

    You want to argue capital vs. income.

    I want to argue about what is being risked.

    And I WIN, because you admit that all investments are equal (EMH), and admit that all risked labor is not (skill).

  23. Gravatar of ssumner ssumner
    12. March 2012 at 12:48

    AFG, You said;

    “There is an equity problem with this argument. People with more existing wealth can accept higher risk and thus generate higher returns.”

    No, that is only a problem with flat taxes, but I favor progressive wage taxes. And I also mentioned a wealth tax if the inital distribution is unfair.

    You raise some good points, but we can’t let the perfect be the enemy of the good. Moving toward a simple, progressive consumption tax would make things much more efficient and much fairer. At that point we can perhaps further tweak the system in ways you suggest.

    Regarding your last proposal, it’s worth thinking about, but I’d rather tax luxury goods like mansions and yachts. And what about people who lose money, does the IRS send them a check? See dtoh’s post.

    dtoh, I think I addressed those issues in the article. We subsidize education already, so I don’t think the human capital distortion would be all that bad. And there is already a big problem separating wage and capital income. But surely we can do better.

    I’m not opposed to your last idea, but some would object that it’s not progressive enough. Perhaps both taxes should be used in combination.

    Marcus, He seems pretty negative on MM.

    Jim Glass, You may be right about the politics, but I still prefer blowing up the income tax entirely. I think filling out tax forms is a huge waste of time (certainly for me it is.)

    Martin. It’s not inefficient if the goal is to have the rich consume less.

    Kocherlakota may be right, but his system would be very difficult to implement in practice.

    Brito, Me too.

    JL, I want the redistribution to be done through the very same payroll tax. A wage subsidy per hour worked.

    Morgan, I don’t agree, you are making emotional arguments, not rational arguments. There are no good guys and bad guys, just people whose talent, work ethic, luck and saving propensity differ.

  24. Gravatar of Major_Freedom Major_Freedom
    12. March 2012 at 13:05

    Morgan, what a bunch of nonsense. That’s no “trump card.” You don’t determine like some dictator that “sweat equity” is “more important” than “cash investment.” The value is a result of inter-subjective valuations between market participants in exchange, not your ex cathedra pronouncements.

    You write:

    Scott YOUR PROBLEM, is that you refuse to respond to my trump card – you want to make this about what the saver gave up, and I’m arguing thats nice, but my guy who risks labor is more important.

    This is the dumbest thing anyone interested in economics can possibly believe. What people give up is paramount in marginal value determinations. Opportunity costs are how we find out marginal utility of a further good, or action. Not only is this the proper framework in economic theory, it is also the proper framework in understanding the ethics of taxation. As much as I disagree with Sumner, he is absolutely correct that taxing capital is double taxation, and is economically destructive because it literally punishes people for giving up present consumption.

    Just step back and think about that for a second. You have $100, and you have the choice to consume, invest, or add to your cash. Capital taxation is punishing someone more for giving up cash balances, giving up present consumption, and choosing investment instead. Think about that. The state is taking more money away from people the more they abstain from consumption and cash hoarding, and the more they invest. If you don’t have a hole in your head, you would know that investment is the lifeblood of economic growth. Consumption decreases the size of the economy all else equal, and investment increases the size of the economy all else equal.

    To tax capital, which is really punishing people for delaying their current gratification in exchange for greater productivity in the future, is like the state telling people it’s wrong to grow the economy, but right to consume more in the present and make the economy smaller than it otherwise would have been. This is absolutely ridiculous, ethically and economically.

    Maybe you’ve been hoodwinked by Marxoid propaganda into emotionally hating “unearned income”, as if the only valid earning of money is from pushing buttons or delivering food, but in a free economy, ANYONE can invest cash and earn “unearned income”, and not only that, but the elimination of tax on capital will INCREASE the amount of aggregate investment in the economy in accordance with people’s true valuations, and that can only mean an economy more in line with people’s voluntary preferences, not the arbitrary decrees of the state imposing an unfair and economically destructive tax system.

  25. Gravatar of dtoh dtoh
    12. March 2012 at 13:58

    You said;
    “I’m not opposed to your last idea, but some would object that it’s not progressive enough. Perhaps both taxes should be used in combination.”

    Easy enough to solve. Just have a higher tax rate, and larger deductions. If you want graduated rates. Have the card provide a 100% exemption up to the first deduction level say $20k, then 75% exemption up to $30k, etc.

    The administrative advantage of totally eliminating income taxes and just having a progressive consumption tax would be huge. Plus you would eliminate the problem of human capital and of determining what is wage income and what is captial income.

  26. Gravatar of Jim Glass Jim Glass
    12. March 2012 at 16:31

    Jim Glass, You may be right about the politics, but I still prefer blowing up the income tax entirely.

    The problem is the politics matter. If we could just blow up systems regardless of politics then it would have been easy to fix Social Security long ago, rent control would have long been repealed in NYC, etc., etc., etc., just because these are demonstrably good things to do.

    Coase pointed out that transaction costs (including information cost, costs of misinformation, politics based thereon) are all that keep resources from flying instantly to their best use and creating Utopia today. But we’re not in Utopia, the transaction costs are there, and we have to plow our way through them as best we can to create an better world incrementally.

    Stylized examples of a world without tax on capital via eliminating income tax and the like are useful and valuable for educating the populace and reducing the cost of misinformation — though only very marginally. But there is no chance whatsoever they will be enacted politically.

    So if one actually wants to *really see* tax free capital investment one needs a proposal that adapts easily into the current tax system. The “super IRA” idea does this by keeping the entire rest of the tax system intact, while effectively converting it into the consumption tax economists want by stealth. That people keep their home mortgage deductions and all is not a fault but a virtue — because they are keeping them, period. So does one want tax-free treatment for capital or not? To 80% of voters the super-IRA is “everything we have now *plus* tax free savings if we want”, that’s sellable to them. To the 20% of partisans it is a deal, double the tax on the rich who consume capital, eliminate it on those who invest it. To get through the political system, *something* like this is the only possibility.

    It’s an old story, do we want the perfect to be the enemy of the good?

    Do you want to create a perfect theory of demand management that no central bank would ever apply, or a practical program that central banks may someday really use for the greater good? From what I’ve seen, the latter. Taxes are my business, and that’s how I feel about tax policy.

    I think filling out tax forms is a huge waste of time (certainly for me it is.)

    “Huge”? With Turbotax or some such an individual tax return shouldn’t take more than an hour or three a year. Unless one is self-employed owning a real business, when one’s accountant prepares it for one. Time preparing individual returns is the least problem of the system. And people really *want* all those exemptions and deductions and education subsidies and all the rest they get by spending that little amount of time.

  27. Gravatar of ssumner ssumner
    12. March 2012 at 16:45

    dtoh, You make a good argument. When we switch to electronic money that might be the way to go. Perhaps I’ll do a post sometime.

    Jim Glass, I did Turbotax last year, and it took many many hours, I have no idea how many–perhaps 20 hours.

    I certainly agree with making the perfect not be the enemy of the good, and would certain support super IRAs if they came up for a vote.

  28. Gravatar of AFG AFG
    12. March 2012 at 17:48

    I. My primary points are:(1)payroll taxes are not equivalent to consumption taxes once you acknowledge human capital, so pick a new type to advocate for

    (2) a political argument against the rhetorical choices your Economist column makes. This is how the debate always goes: Conservatives makes the theoretically sound claim that we shouldn’t tax capital, therefore repeal the capital gains and estate tax. Liberal says human capital! Conservative says exempt that. You can’t! Okay, but that’s why I really favor a consumption tax.

    But at the end of the day, there is far more political support for a capital gains tax cut so all we end up with distortionary subsidies for a certain type of capital. This is not, as you say, letting the perfect be the enemy of the good. If we have any tax on any type of income, the most efficient thing we can do is tax them equivalently to keep relative prices for different type of investment the same. Well technically, it should be proportional to the inverse of the elasticity, but unless you can show me conclusively that there is a greater response to physical than human capital investment. Those empirical studies probably exist, but they will be inherently flawed due to the impossibility of separating labor from human capital investment. A conservative government should be skeptical of politicians’ ability to central plan through weighing relative elasticities.

    In fact, I believe there is an apriori case that we are taxing human capital too much and physical capital not enough: (a) If the poor gain a greater percentage of their wealth from human capital and the rich gain a greater percentage of their wealth from physical capital, then we would expect net-subsidized physical capital to result in increased wealth inequality (and that is what we see). (b) A greater percentage of our national income is being generated the financial sector (and other things that earn from physical capital), but RGDP growth has decelerated (even before the 2008 crisis), this suggests that something is driving an allocation of capital from the human to financial sector and its resulting in lower efficiency.

    Yes, there are a thousand other factors that could explain these trends. But when theory suggests a particular distortion and we see a similar distortion show up in the data, a prudent, small-c conservative like yourself should be clamoring for your government to stop picking winners and RAISE the capital gains rate, until we can get a consumption tax.

    More than that, you, who has made a career about how the artificial semantic differences between inflation, interest rates, and NGDP can structurally constrain monetary policy, should see that having the government write non-existence difference between labor and capital into the tax code is only going to encourage avoidance through redefinition and systemically discriminate against the far more difficult to conceptual “human capital.”

  29. Gravatar of AFG AFG
    12. March 2012 at 17:49

    PS I apologize for the length, I am working on this issue for a much larger project anyway and I’d like to you your and other people’s thoughts.

    II. My secondary points are:
    (1) The Equity problem and risk-related returns – Fair enough on you suggesting wealth taxes. I want to emphasize though that it’s not that they “might” be necessary, but that we can expect progressive consumption taxes to lead to increasingly regressive outcome in terms of WEALTH not wage inequality in the long term. EMH might be true from the perspective of any given individual, but that doesn’t mean that the distributional consequences will be neutral in the long term.

    (a) The obvious one – large bequests that are not subject to any tax in your perfect world. Mankiw and other conservatives like to point out that the correlation between wealth transfers and wage inequality is surprisingly small. This is true, but misleading. The reason why it is small is mostly because those who inherit a massive amounts never had a reason to invest in their human capital (and thus don’t work or earn low wages “running” their fathers charitable foundation) and can simply pay someone to generate consistently higher returns than the market by taking advantage of their risk tolerance.

    But this is cheating on my part. For the rest of these examples, I will let you assume that everyone starts with the same initial distributions of wealth. In order to operationalize this, let’s impose a secret 100% Estate tax with no exemptions. It’s secret so it doesn’t affect incentivizes. After the person dies, the government goes around and wipes the memory (with those Men-in-Black flash things) of anyone who ever knew the dead estate-leaver so no one ever discovers the secret tax.

    (b) Assumption 1: There is only one type of Human Capital – Willpower.

    Assumption 2: People are born with different levels of Willpower (although it can be increased or decreased), and higher Willpower correlates with a higher wage rate (

    Assumption 3: Willpower correlates with higher savings rate (I don’t think it is controversial to suggest that impulsive people consume more)

    Assumption 4: Willpower increases Willpower. When people practice small acts of self-control, they increase their capacity for self-control (

    The second assumption is only here to delineate how the other effects correlate with income (you’re right, a progressive consumption tax conteracts that effect.)

    The conclusion to this should be obvious: One of the most important things in determining wages is willpower, but that same characteristic also largely determines willingness to defer consumption. In period 1, high willpower people will save more both on an absolute and relative basis. His act of self-control in period 1, will increase his total willpower in period 2, where his wealth will also be much higher than his low willpower brethren.

    Thus, in period 2, the high willpower person will increase his savings rate and command a higher wage, his wealth will allow him to take bigger risks and generate larger returns, and his act of self-control will further increase the gap between his and the poor’s willpower.

    There are diminishing marginal returns to these things. Eventually he won’t be able to raise his willpower any higher, so his wages, savings, and level of risk tolerance will remain constant. Although his risk tolerance won’t grow much in response to greater wealth, his massive existing stock of wealth will allow him to generate returns that consistently beat the market. His wealth will grow faster than he can consume it, so even though he is thrifty in the sense that he saves a lot, his absolute level of consumption starts to sky rocket. However, since his Willpower is now barely growing, his wages will become stagnant and most of his wealth accumulation will derive from saving. In an extreme case he might even stop earning “labor” income.

    (c) You should view this as the inverse of your parable about the twin brothers. Savings taxes may increase horizontal consumption inequality, but they will decrease vertical consumption inequality.

    Let’s consider two different consumption taxes and their effect on vertical consumption inequality.
    (Taxes/Cons.); (Taxes/Wages); (Taxes/Wealth). === The 3 ways of measuring consumption inequality.

    First, your progressive payroll tax. At a minimum, this system will become regressive, using consumption and wealth as the denominator. He taxes are constant, but his consumption and wealth increase infinitely so the effective rate goes to zero. There’s also reason to believe that it will become regressive in terms of the second, since people like Romney stop working when they wages become so insignificant compared to their investments.

    Second, a progressive tax on sales. In this case, the tax is potentially regressive in terms of wealth, if wealth increases faster than consumption. It may become regressive in terms of wages for the same reason as above. It definitely RETAINS in progressivity in terms of consumption.

    Third, a progressive tax on sales + extra taxes on luxury goods. Surely, this is even more progressive in terms of consumption. And more likely to be progressive in terms of wealth, since it raises the rate at which taxes grow. It may become regressive in terms of wages for the same reason as above.

    (d) Conclusion: Wages and labor are meaningless concepts. And Wage inequality is a horrible standard for measuring vertical equity, because it favors physical capital. There are only two important categories for tax policy: Consumption and Wealth Accumulation.

    Adding in the notion of human capital means that a sales tax is preferable to a payroll tax. It is both more efficient and preserves vertical consumption equality, whereas the payroll tax is regressive by EVERY measure.

    We should tax luxury goods, but in a practical sense ALL consumption taxes are likely to be regressive in terms of wealth inequality at the extreme end of the distribution. The above analysis shows how wealth inequality begets wealth inequality, so we likely need some kind of wealth tax to diminish it.

  30. Gravatar of dtoh dtoh
    12. March 2012 at 18:09

    There is no need for electronic cash in order to implement a progressive consumption tax. You just the run the card through a terminal that tells the sales clerk how much consumption tax to charge. If you don’t have the card with you, you pay the top consumption tax rate.

    Jim Glass,
    I actually think a straight forward progressive consumption tax is an easier sell. Any kind of tax free savings account is going to be attacked as just another tax break for the rich.

    On the other hand a tax marketed as a curb on excessive and conspicuous consumption could have much wider political and populist appeal.

    The only problem with getting rid of income (capital, wage and corporate) taxes is that it would be vehemently opposed by accountants, lawyers, and lobbyists as 90% of them would be put out of business.

  31. Gravatar of JL JL
    13. March 2012 at 02:03


    OK, so millionaires will drink whisky for minimum wage, 40 hours a week.
    It will cost a lot to expose these schemes, but the schemes will just get more complex: employing people for hobbies like sports (“I’m a 50 year old professional athlete”) or sending in bogus reports from home and conference calling with similarly ’employed’ people (“Our company sells digital widgets to anonymous customers on ebay. We employ 100 writers for minimum wage and just barely make a profit”).

    And we’ll still need welfare for the (longterm) unemployed, which the millionaires will be eligble for because you want to abolish means testing, because that isn’t fair for the thrifty brother.

    It gives a whole new meaning to the term “welfare queen”.

  32. Gravatar of JL JL
    13. March 2012 at 02:05


    OK, so millionaires will drink whisky for minimum wage, 40 hours a week.
    It will cost a lot to expose these schemes, but the schemes will just get more complex: employing people for hobbies like sports (“I’m a 50 year old professional athlete”) or sending in bogus reports from home and conference calling with similarly ’employed’ people (“Our company sells expensive reports. We employ 100 analysts for minimum wage and just barely make a profit”).

    And we’ll still need welfare for the (longterm) unemployed, which the millionaires will be eligble for because you want to abolish means testing, because that isn’t fair for the thrifty brother.

  33. Gravatar of ssumner ssumner
    13. March 2012 at 06:18

    AFG, You said;

    “My primary points are:(1)payroll taxes are not equivalent to consumption taxes once you acknowledge human capital, so pick a new type to advocate for”

    I think they are close enough, and you can always make them more similar by subsidizing education, or making it tax deductable. We already subsidize education, so that makes the systems almost identical.

    You said;

    “In fact, I believe there is an apriori case that we are taxing human capital too much and physical capital not enough: (a) If the poor gain a greater percentage of their wealth from human capital and the rich gain a greater percentage of their wealth from physical capital,”

    This seems way off base. Most poor people go to public schools where 100% of their education cost is subsidized. If anything, just the reverse is true, we are much tougher on physical capital than human capital.

    Let’s say I’m wrong, and you are right. I would still oppose a income tax, I’d favor the sort of consumption tax proposed by dtoh.

    Regarding your longer post, I don’t have time to address all the issues, as I get 100s of comments. I’ll just say that I don’t view wealth inequality as a problem, rather the problem is consumption inequality.

    dtoh, I see your point.

    JL, It seems to me you are arguing against any sort of welfare, as fraud occurs with the current system as well. Lots of people cheat on EITC. Lots collect welfare and unemployment insurance and still work in the underground economy. If you are really concerned about millionaires doing minimum wage jobs, I suppose we could have a wealth test for wage subsidies. I don’t see that as a big issue, so the distortionary effects of that would be small.

    For long term unemployed I’d get rid of minimum wage laws, and also have a back up government job for low pay cleaning up parks, etc. Make the government job so undesirable that no one wants to take it. Make it so they prefer the private sector low paid job.

  34. Gravatar of Negation of Ideology Negation of Ideology
    13. March 2012 at 07:21

    “I think they are close enough, and you can always make them more similar by subsidizing education, or making it tax deductable. We already subsidize education, so that makes the systems almost identical.”

    I think this brings us to a key point. It’s not the distortion of the tax system we should be concerned with, it’s the total distortion of the tax and spending system together.

    This relates to the earlier post about positive and normative economics. Capital gains taxes are a distortion against saving. Spending programs for retirement, emergencies (job loss, health care) and the like also reduce an incentive for saving. That falls into the positive category. As a society, we have made the judgement that those who profit the most from the system of private property rights enforced by the government should pay more of the costs of that government. We have also decided that a safety net is appropriate for those who are too old or too sick to work, or who lose their jobs. That falls into the normative category.

    So the way to deal with the distortion brought on by capital gains taxes is to balance it out with something else, as you suggest with education. We should keep the capital gain tax, with our eyes open, and balance the distortion out with large universal 401(k) accounts and health savings accounts. Singapore has a system like this.

    This benefits the rich in the long run, because it reduces the desire of future voters to redistribute wealth. Pure free market economics is incompatible with a democracy of voters who have no wealth.

  35. Gravatar of JL JL
    13. March 2012 at 08:09


    I am arguing against welfare for millionaires, e.g. by means testing.

    But means testing is also double taxation and it requires just as much paperwork as the capital income tax. It punishes the thrifty brother in exactly the same ways as the capital income tax.

    If you allow this – and you just did – then you have no actual argument against the capital income tax; you just think that the current rate is too high.

  36. Gravatar of JL JL
    13. March 2012 at 08:16

    p.s. May I suggest changing the title of the Economist article to:

    “The proper tax rate on capital income is $36,445”

    That’s the inflation adjusted Total Annual Value of the Welfare Package according to CATO for MA, source:

  37. Gravatar of TravisA TravisA
    13. March 2012 at 15:54

    @dtoh, I love your idea. But how would you handle the problem of poorer people renting out their ‘card’ to richer people? Even if you have an ID system with a picture and/or fingerprint, there is the the problem that stores have no incentive to enforce the mechanism (actually the opposite, since prices would be lower). Furthermore, you also have the Internet, where ID checking is not currently possible.

    Even if you tried to tie the ‘no sales tax card’ to a bank or credit account where payments are made from, you still have problems because the people who rent the card could make payments into that account.

    If you try to enforce the policy through penalties, how would you police such a system? Furthermore, how can you prevent people calling the policy ‘gifts’?

    Unless there’s a way around my objections, I think a super-IRA would be best. But I hope you have some solutions.

  38. Gravatar of dtoh dtoh
    13. March 2012 at 16:11


    Good point with a simple solution. Allow anyone to sell the card back to the government for the unused exemption amount discounted at the 1 year Tbill rate.

    As long as the the amount at which the full consumption tax rate kicks in is not set too high, this works perfectly. It also provides a subsidy to low income individuals whose consumption is below the threshold. A good thing in my IMHO, which could eliminate the a lot of welfare programs as well.

    Also remember, the stores are liable for the full consumption tax on their total sales. They only get a break for those exemptions properly executed through the card system.

  39. Gravatar of Rien Huizer Rien Huizer
    13. March 2012 at 22:56


    Provocative post/essay. No doubt you’ve seen Diamond and Saez (less utopian, more scholarly) article in the Journal of Economic Perspectives on various taxation issues, including capital (and yes with a discussion of the New Dynamic Public Finance approach). Their conclusion recommendation is against zero taxation of capital/capital,income, for a variety of theoretical and practical reasons.

    Of course if one could start with a clean sheet, tax design as a whole could be different from the present situation and then taxation of earned savings might be efficient, or fair in a Rawlsian way.

    But clean sheet reforms do not exist outside very extreme conditions (revolutions, devastation, etc). Tax lives in the land of the status quo bias.

  40. Gravatar of ssumner ssumner
    14. March 2012 at 05:02

    Negation, it seems to me your arguments point in exactly the opposite direction. You are right that programs like Social Security tend to reduce saving (as does Medicaid and Medicare and college financial aid.) The best second best policy would then be a subsidy to saving, the capital gains tax should be replaced with a capital gains subsidy.

    Jl, You are mixing up two unrelated issues, what’s best in theory and what’s best in practice. In theory I’d prefer no taxes on capital. In practice we might have to make a few tiny compromises. I haven’t thought through your proposal, but it sounded relatively minor. If it gets us 99% of the way toward eliminating capital taxes, I’m fine with it.

    Rien, Yes, I did a blog post that was very critical of that paper. They did exactly what you claim I did, assuming a clean slate in their call for very high MTRs. Even worse, they adjusted the clean slate assumption whenever it was necessary to support their ideological prejudices.

    So their paper wasn’t even internally consistent, at least you can’t say that about my essay.

    The point was to start with the optimal system. I realize that in the real world compromises must be made, but we could get pretty close to the ideal without any significant problems. None of the objections that people raised apply to me or 90% of other taxpayers. I should just pay a payroll tax. If there’s any reason why this is wrong, then I’d love to hear it. So far no one has offered objections that apply to more than a tiny percentage of the US population. I concede that the very rich might require some complexities that make a simple payroll tax infeasible, but even so we could free up more than 90% of the US population from the misery of the income tax system.

    BTW, see my reply to Negation.

  41. Gravatar of Negation of Ideology Negation of Ideology
    14. March 2012 at 05:37

    “The best second best policy would then be a subsidy to saving, the capital gains tax should be replaced with a capital gains subsidy.”

    That’s kind of what I was getting at with the Singapore style required investment accounts, which are a form of subsidy for saving. Calculate (or estimate) how much the disincentive for investment comes to (cap gains tax, social security, etc.) and offset it by requiring automatic savings accounts.

    “I concede that the very rich might require some complexities that make a simple payroll tax infeasible, but even so we could free up more than 90% of the US population from the misery of the income tax system.”

    Exactly. Getting rid of cap gains tax for the middle class might work, but for the rich I don’t see it as politically feasible. So offset the lost investment from the rich with more investment from the middle class and working poor.

  42. Gravatar of ssumner ssumner
    14. March 2012 at 10:32

    Negation, That would be progress.

  43. Gravatar of JL JL
    15. March 2012 at 02:38


    Your concession has totally changed (my perception of) your message.

    If you want to abolish capital income tax for the middle class, but are willing to tax the superrich (e.g. a tax on wealth above $1M) then I could respect your position.

    Without a continuous transfer from the superrich to the majority you will slowly accumulate a large class divide: a powerful, wealthy elite and a disenfranchised majority.

    From your ivory tower it may seem like a good idea to enact a one-time wealth transfer, but what usually happens is either a minimal transfer to pacify the masses, a violent suppression of riots and uprisings, or a violent revolution where the transfer succeeds but the elites end up killed.

  44. Gravatar of J.V. Dubois J.V. Dubois
    15. March 2012 at 08:16

    First Scott, thanks for the article. It helped me think about some issues. I would just make the following conclusions:

    1. It is very hard to make distinction between what is the capital and what is the labor part of the income. I think that this is the main reason why even capital gains are taxed now.

    2. That being said, the best possible taxation is what dtoh suggests and that it is the consumption tax. I believe it to be so because:

    – It does not discriminate the source of the income making various inefficiencies, and it stops people from trying to hide wage income as capital gains

    – It increase incentives for everybody to invest. With payroll tax, it is applied at the beginning by the first generation of savers. For children of wealthy parents their endowment is just given and the tax system does not create any incentives for the them to either spend less (due to absence of consumption tax) or work more (due to presence of payroll tax).

    3. As per progressive taxation, I believe that this is very important not for theoretical efficiency under given assumptions. I believe that it is very important for long-term stability of social contract and institutions. I believe that there needs to be a cap of how much power can a single person wield using their wealth. To much concentration of wealth – eve if gained justly for remarkable performance – is in my eyes dangerous for long-term stability of free society.

    However this is a separate issue. It makes perfect sense to have negative tax (either VAT rebate cards, subsidised education, food stamps) on one extreme of the scale and than luxury taxes on the other end of scale.

    But I am not that convinced that this is enough. What about someone very powerful that has a lot of money but lives “relatively” modestly? What about insane heirs of mighty commercial empires of their parents? I am not sure if capital gains tax diminishes the risk of such clusters of economic power in the grasps of individuals, but if it would I could see such reason alone as perfectly sufficient for enforcing such tax, “minor” economic efficiencies (compared to risk of lasting damage to social institutions) notwithstanding.

  45. Gravatar of TravisA TravisA
    15. March 2012 at 09:59


    That’s a clever attempt at a solution. As you say, it only works if the amount when the full tax kicks in is not too high. Plus, you have to consider the amount of the tax. It will probably have to be around 20% if it is to replace all existing taxes. Furthermore, I think a sharply increasing consumption tax would be beneficial — perhaps rates of around 40-50% for expenditures over $1 million. That would be difficult with a card, but not difficult with a super-IRA.

  46. Gravatar of dtoh dtoh
    15. March 2012 at 13:07

    Wow. I don’t think 40% or 50% would ever fly politically. To make the system simple and politically viable, what you need is a reasonable base rate that applies to all transactions. The card just gets you an exemption for the spending up to some threshold amount.

    To give a sense of the numbers, assuming 10 trillion of consumer spending (less an exemption on the first $10k of spending for every man, woman and child) leaves you with 7 trillion of taxable consumer spending. A tax rate of 20% would get you 1.4 trillion of revenue, which should be enough to replace all current income and corporate taxes, but not payroll taxes (FICA, etc).

  47. Gravatar of TravisA TravisA
    15. March 2012 at 15:26


    That’s a nice back of the envelope calculation. In order to replace all other taxes, we’d need a tax rate of around 37%. If we kept FICA, we’d need to still consider the definition of wages.

    A 20% tax rate for those earning $10,000-$20,000 per year is pretty regressive. So the rate is going to have to be graduated.

    A marginal rate of 40%-50% seems pretty reasonable to me. It’s not too far from the current top tax bracket, furthermore the rate can be avoided by not consuming at high amounts. I think it’s less disruptive to individuals to decide not to consume something than it is to decide to take exchange compensation for leisure, especially since a lot of wealth accumulation is done to assure that the family is taken care of.

    I like your idea and perhaps with tweaks it could be implemented.

  48. Gravatar of dtoh dtoh
    15. March 2012 at 17:05

    You may have better info on tax revenue than I do, but I believe tax revenue from personal and corporate income tax is around $1.2 trillion, but maybe my number for consumer spending is off.

    I’m just using the numbers as examples, but a family of four would pay no taxes up to $40k. You could tweak it so adults get a $15k exemption and children get a $5k exemption. IMHO the simpler the better, but the exact numbers are a political discussion not an economic one.

    The other question you raise is whether to put a special tax on high consumption. I don’t think this would fly in either party. Even the Democratic Party, is very dependent on big contributors who all pay lip service to progressive tax rates because with the loopholes they never pay the nominal rates. If on the other hand, you tell them that instead of getting taxed 15% on their clipped coupons, they are instead going to have pay 50% for their trips to Telluride, they’ll completely freak out.

    Also, I think you do a lot of damage to some specific industries with a “luxury” tax.

Leave a Reply