27 Oct 2011

When the Facts Change, I Silently Curse–What Do You Do, Sir?

Economics 24 Comments

Dangit, I realized there is yet another complication to the income vs. consumption tax saga. I don’t know about you guys, but this debate is like the first season of 24. In the beginning I thought it was incredibly interesting, I was hooked, I couldn’t get enough. But after 20 hours of it, there were so many plot twists and incredible assumptions that it bordered on absurd.

So yesterday two things happened that made me realize I have been an omitting something important from my repeated assurances that there is a legitimate sense in which the income tax dings you worse than the consumption tax. First, I gave a fairly rigorous “proof” that an individual really is worse off under an income tax, but I had to assume an individual’s lifetime consumption equals his lifetime income. (It doesn’t mean that’s true every period, just over the lifetime. E.g. you save a lot during your working years, and then you draw down yours savings during retirement, so that upon death you have no assets.)

So at the time, I thought I was bending over backwards to be fair to those who were arguing, “Hey I ultimately save in order to consume, so why aren’t the two taxes equivalent?”

But then I read Alex Tabarrok’s paper (from when he was a grad student) talking about the standard mainstream case for the superiority of an income tax versus an excise tax. If you’ve never seen a graphical presentation of these types of arguments, go look at his paper–it’s short. He doesn’t use indifference curves, he just draws a budget set and shows through a revealed preference argument that you are better off if the government takes (say) $1000 out of your income through a percentage tax, rather than $1000 out of your spending on a particular product through a percentage tax.

So at this point warning bells are going off: If Tabarrok is right that this type of “standard” analysis shows an income tax is better than an excise tax on a particular product, what changes to make an excise tax (as it were) on all products better than an income tax?

Well, there are two different things going on here. In Tabarrok’s demonstration, you assume the person’s pre-tax income is a fixed number. And since the person can shift out of the taxed good (when an excise tax is imposed), the government has to jack up the percentage, taking into account your response, in order to extract the same amount of revenue from you.

In contrast, when I was warning of the income tax’s distortionary effects, I focused on the fact that your income itself (over time) isn’t a fixed number, but depends on your saving and investment decisions which in turn are muted by an income tax. If you assume that a person will always consume 100% of his income (which I did in the “proof” linked above), then the only moving part is the person’s income, which will be affected by an income tax but not by a consumption tax (unless it makes you take more leisure).

In conclusion, my sophisticated, PhD answer to which tax hurts an individual more is, “It depends on your assumptions.” I think you can basically say that to any economic question, and you’ll be able to hobknob with the best of us.

Last point, in all seriousness: I want to stress that this is counterintuitive. I think a lot of you are going to say, “OK, I knew there was something fishy here. A consumption tax is just as bad as an income tax, because after all you save in order to consume in the future.” And I’m still saying no, that isn’t what’s driving my addendum above. In fact it’s the opposite (assuming I am now thinking about everything the right way). If you are only saving in order to consume in the future, then my original point holds up; see the “proof” linked above.

But the screwy thing occurs if you don’t consume all of your lifetime income. If you bequeath wealth to your kids, or donate to charities along the way, then your lifetime income is higher than your lifetime consumption. And now we’re back in the realm where a consumption tax can mess up your allocation decisions, this time between consumption and donations to others. So a consumption tax (if calibrated to extract the same amount of tax revenue) will cause you to consume less and donate more, relative to your original tradeoff. Just like an income tax will cause you to save less and consume more in the present. Without giving specific numbers and preferences, I don’t think we can say which effect is worse.

24 Responses to “When the Facts Change, I Silently Curse–What Do You Do, Sir?”

  1. Rob says:

    The link to the Tabarrok paper is not working.

  2. Brian says:

    Bob –

    First – a huge AMEN about the 24 series. You captured my feelings precisely.

    On the tax question, if I’m following this accurately – the point about a consumption tax is that it distorts the natural (free-market) time preference saving/consumption decisions. However, doesn’t this, in essence, assume no distortion in monetary policy, which obviously is not the case. In other words, assume that we have monetary policy which is too loose (since we almost never have the opposite, even if it’s equally bad theoretically). Wouldn’t a consumption tax in that situation be preferable to an income tax in order to counteract the skewing effects of monetary policy?

  3. Tho Bishop says:

    Thanks for fixing the comments! It is a relief knowing that it wasn’t something I was (wasn’t) doing.

    I don’t pretend to be able to offer anything valid on the pure economic evaluation of income v. consumption, but I have to ask whether there is some larger aspects of the debate that some may be overlooking – the actual collection of the tax.

    A consumption tax necessarily increases the tax base. An income tax only applies to those who get their income via “legal” means as well as applying only to the citizens of thet taxed region. A consumption tax, however, applies to everyone who consumes – black market participants (including illegal immigrants), lemonaid stand-quality business ventures and tourists. If we assume that the revenue is equal, the tax responsibility per taxed individual (on average) would necessarily be lower for the tax system that collects from the most people.

    Now I don’t know if this is relevant to the economic analysis but I was curious about that point.

  4. Steven E Landsburg says:

    The key point is that you’re better off when all consumption is taxed at the same rate. An excise tax on one good but not others violates that condition. An income tax (which includes a tax on interest) also violates that condition, because it effectively taxes future consumption at a higher rate than current consumption.

    So a consumption tax on all goods is better than an excise tax on one good, and also better than an income tax, and both for the same reason.

    To conclude that an income tax is better than an excise tax on one good would, I think, require additional assumptions. To conclude that a *payroll* tax is better than an excise tax on one good, though, follows as above, because a payroll tax is equivalent to a consumption tax.

    I haven’t looked at Alex’s paper (yet) but I wonder whether his “income” tax was really a payroll tax.

    • Bob Murphy says:

      But Steve are you assuming people consume 100% of their income eventually? How is a payroll tax equivalent to a consumption tax, if (say) I donate 50% of each paycheck to the Center for American Progress? Are you having my utility change because the CAP gets hit with the consumption tax when they buy computers etc. for Matt Yglesias to continue his fine work?

  5. Evan says:

    I’m still having trouble with this analogy. I can clearly see how an excise tax on a narrow range of goods is more distortionary than a broader tax. However, isn’t a tax on ALL income equally “neutral” to a tax on ALL consumption? I see in your proof, you show that a consumption tax increases the incentive to save, so that the guy earns more interest/dividend income over his lifetime. However, if that incentive applied to everyone in the economy, wouldn’t some of the earnings on increased savings tend to be cancelled out by a reduced general rate of return on investment due to the increased supply of saved capital causing a lower marginal demand for it? Also, wouldn’t the guy have less to save in each period because his consumption was more expensive (due to the consumption tax)?

    What exactly makes the income tax more restrictive of action? Does it have something to do with consumption being easier to defer to a later period than earning income is?

    • Bob Murphy says:

      Evan, I like the way Landsburg put it above. A consumption tax penalizes your consumption, obviously, but it doesn’t place a higher burden on one sector vs. another. But an income tax places a higher tax rate on future consumption versus present consumption. So if you get why a broad consumption tax is better than a tax on food, then you should see why a consumption tax is better than an income tax.

      (However, I now think the above argument is bulletproof only if we assume you consume all of your income. If you have a choice between spending your income on consumption versus spending it on philanthropy–and the latter doesn’t get hit with the consumption tax–then it’s back to being like an excise tax. Instead of taxing liquor or food, it’s taxing “consumption.”)

      • Rob says:

        I looked at Tabarrak’s paper. While he uses the example of a single good I think the methodology would still work if good x was assumed to be a bundle of all goods that consumers could buy rather than holding onto money and thus establishes the general benefit of an income tax from a consumers perspective.

        The paper assumes a single period and I’m wondering if one looked at later periods where money not spent in the first period were consumed then Steve Landsburg’s comments about income tax taxing future consumption would still be relevant .

        • Bob Murphy says:

          Rob wrote:

          The paper assumes a single period and I’m wondering if one looked at later periods where money not spent in the first period were consumed then Steve Landsburg’s comments about income tax taxing future consumption would still be relevant .

          Exactly. I mean re-read what you just wrote, Rob. You’re saying that it’s clear if there is no opportunity to save for the future, that an income tax (which dings saving) is OK. You can only see the effect Landsburg and I are talking about, if the individual has the option of deferring consumption today and earning more income (through investment) to consume in the future.

          However, what I realized yesterday–and what prompted this latest post on the topic–is that I was making a similarly silly move in my argument about a consumption tax. I was saying it doesn’t distort the saving/consumption tradeoff, but only because I was assuming people always consume eventually. If a person permanently consumes less than his income (by donating to others, including heirs), then my argument doesn’t work, just like Tabarrok’s argument doesn’t work if you are allowed to alter your income by your behavior.

          • Rob says:

            While I agree with you logic I am still struggling to reconcile it with the Tabarrok paper.

            His model takes savings into account (anywhere on his chart that is not on the horizontal axis includes some savings). This non-consumption could be for any purpose (future consumption, bequeaths to future generation or just to burn) and would in the case of the income tax option factor in possible taxation of future consumption. And this model clearly shows that income tax is still the best option for the consumer.

            I’m am not sure to reconcile this with your analysis.

      • Evan says:

        OK, I think I see what’s being said now. An income tax distorts the economy by singling out the acts of saving of productive investment for punishment.

        So when you say that the consumption tax (which presumably also distorts the economy by singling out acts of consumption for punishment), is preferable to the income tax (assuming revenue-neutrality) only because distorting the economy towards more savings would result in greater wealth? Or are you actually saying that it’s better from the point of view of individual preferences?

        • Bob Murphy says:

          Evan wrote:

          Or are you actually saying that it’s better from the point of view of individual preferences?

          Individual preferences. I have never been talking about some abstract devotion to “faster economic growth” or something like that. I have always been talking about individuals getting more utility as they subjectively perceive it.

          But, now I’m not so sure about the argument, for the reasons I gave in this post.

  6. Rob says:

    Here is my interpretation of what is meant:

    If you pay 10% of your income in some form of tax – then under income tax you always pay 10% every period, while under a consumption tax if you defer some spending in the current period to a later period then the present value of the future tax you pay will be lower assuming a non-zero time preference.

    If this is a correct view then everyone would benefit from a consumption tax except for individuals with high time preference who would always spend their money every period and would hence pay the same tax in either circumstances.

    • Joseph Fetz says:

      While it is true that a larger portion of people’s incomes will be saved, you are still less productive. It’s like saying, “yeah, I’ll take a higher savings rate in exchange for a lower overall utility.”

  7. Chris says:

    Suppose someone (guy A) invest $1000 at 10% p.a. for 10 years without an income tax:
    Then he takes out his renevues and goes to the shopping mall where he has to surrender 20% first before he can enter:
    He will have a net profit $1,275 to spend and has paid $319 in taxes.

    Now let`s look at the same person in an income tax regime(guy B) He invests $1000 at a 10% p.a. rate for 10 years but now he has to pay 20% income tax every year which efficiently reduces the rate to 8% p.a.:
    He has made a profit of $1,159 on which he does not have to pay any consumption tax anymore. He has considerably less net renevue than the person in the consumption tax regime. I am to lazy to calculate how much renevue the treasury makes in this example but it`s easy to understand that it will be far less than $318. At the same rate, you end up with less tax renue and less net profit for the investor in an income tax regime as compared to a consumption tax regime.
    If you take into account that guy B had to work harder to get the $1000 in the first place the disadvantages of an income tax become even more apparent.

    2nd point:
    Income tax is creating inequality between the regular work force and other means of production. With a consumption tax, if you employer wants to replace you with a machine, the machine has to be more productive than you are. If you have 20% income tax the machine only needs to have 81% of your productivity, thereby creating an incentive to replace productive labor with less productive machinery. Same for illegal workers, outsourcing to other countries and so forth. 

  8. Peter Marwood says:

    Being a revenue neutral arguement, why do I care? I’m I being asked to choose between death by firing squad vs. death by hanging? Much like the Laffer curve, why do I care what benefits the State?

    • Bob Murphy says:

      Peter, so you are indifferent between a 0% and a 100% income tax, enforced by the electric chair? Both give $0 to the government.

      • MamMoTh says:

        Actually the second option costs the government the electricity.

        But what a show!

  9. Karl Smith says:


    Your basic proof in the other section misses an important element. That is, the timing of government revenues. Indeed, under an assumption of equal return on savings you are likely to get the opposite result. The income tax requires a lower tax rate to earn the same revenue because it taxes some income twice.

    A numerical demonstration.

    Adam lives two periods and works in the first period to earn $100. He saves half of his income for the second period and earns an interest rate of 10%. Lets see how much tax Adam pays under a 20% income tax versus a 20% consumption tax.


    Year One
    Earns $100
    Consumes $50
    Saves $50
    Tax Paid $10

    Year Two
    From Savings $55
    Consumes $55
    Taxes Paid $11

    Total Taxes Paid $21

    Income Tax

    Year One
    Earns $100
    Saves $40
    Consumes $40
    Taxes Paid $20

    Year Two
    From Savings $44
    Consumes $44
    Taxes Paid $ 0.80 on Interest Earnings

    Total Taxes Paid $20.80

    This is where your analysis ends.

    However, in the consumption case the government took in $10 in period one, while in the income case the government took in $20 in period one.

    If the government spends the same amount over the periods then either it borrows less or saves more under the income scenario. Assuming in both cases the government spends $10 in the first periods and earns a 10% interest rate then in the second period the government comes in with $11.

    Thus total government revenue under the income scenario is $21.80 vs $21 in the consumption scenario.

    • Bob Murphy says:

      Karl, oops yeah to say “total revenue” we should mean “PDV of total expected revenues calculated at some point.” So my demonstration that the revenue-neutral income tax rate is higher than a consumption tax, for a person who ultimately consumes all lifetime income, is wrong.