17 Nov 2018

Economists Overlook Problems With a Carbon Tax

Climate Change, Shameless Self-Promotion 49 Comments

My latest at IER. One of the analysts there said he thought this was the single best explanation I’ve given, for what’s wrong with the common “tax ‘bads’ not goods” argument. Anyway, I am using Tyler Cowen, Scott Sumner, and John Cochrane as foils in this piece. An excerpt:

Try a different example: Imagine if the IRS, rather than applying the federal income tax to all workers, instead only applied the code to workers who were between the ages of 30 and 50. And furthermore, the government bumped up the marginal tax rates, in order to squeeze the same amount of total revenue out of workers in their 30s and 40s, as they currently derive from the entire workforce.

Again, besides basic issues of fairness, such an approach would be grossly inefficient, because it would give a huge incentive for workers to rearrange their activities, sucking as much income “forward” into their late 20s and pushing it “backward” to when they were 51 or older. With marginal tax rates of (say) 85 percent for people in their mid-years, lots of skilled workers would put forth less effort at age 48 (say) because of the punitive tax. Society would thus lose out on their productive work, and the IRS would have nothing to show for it, either, since you can’t tax income that isn’t generated in the first place.

Now that I’ve warmed the reader up, consider a third example: Suppose instead of raising $100 billion annually through an income tax, the government instead tries to raise that same amount of revenue each year by taxing the carbon content of economic activities. This too would have an immediate impact on behavior, causing people to rearrange their affairs in order to minimize their exposure to the punitive new tax. Because the “tax base” of carbon-intensive activities is smaller than the base of “activities generating income,” in this respect a carbon tax is obviously more economically disruptive than a conventional income tax.

If you’re tempted to say, “But Murphy, the whole point is that we want to discourage carbon emissions but encourage working,” then you need to read the article. I get that.

49 Responses to “Economists Overlook Problems With a Carbon Tax”

  1. Tel says:

    Carbon taxes are very broad indeed, because so many things depend on fossil fuels. Everything that gets transported (including raw materials and finished product). All agriculture uses fossil fuels for tractors, pumps, trucks, refrigeration. Building that requires diggers, bulldozers, cranes, tunneling machines. Then there’s everything that requires electricity, and the electricity used in processing, data centers, communications, air conditioning, elevators, building security, water supply, traffic lights, police, fire, hospitals, military.

    This is not comparable to tobacco tax in Florida… carbon tax is huge, and that’s why they want it so bad.

  2. Harold says:

    “It’s pretty standard in this literature for economists to recommend raising revenue in ways that disrupt the economy as little as possible, seeking to minimize what economists call “deadweight loss.”…One of the standard principles in optimal tax design is to have as broad a base as possible…”

    This is not actually true, or a poll tax would be universal. That has least deadweight loss as there is very little behaviour one can change to avoid it and the base is as broad as possible. This came up in a discussion of Thatcher here recently.

    We can conclude that in practice the minimisation of deadweight loss and widening the base is not the only or even overriding factor.

    ““Even if global warming is not a problem, there are obviously much worse taxes than a carbon tax,”
    Sumner is talking about the politics of getting a carbon tax through. His criterion is not the most economically efficient tax but the most hated by the GOP. he says “One solution would be to offset the carbon tax with an equally large reduction in even less popular taxes. ” Not less economically efficient taxes. You are therefore shooting at the wrong target.

    As Tel points out above, carbon tax may not be as broad a base as an income tax, but it is still pretty broad.

    “They tell their readers that such a tax swap would boost economic growth, and it would be good for the environment as a bonus.”

    If this is true then I agree it is not correct. Boosting growth by introducing a new tax and reducing other taxes is not likely to work else we would be doing it anyway. That is not really the point as I don’t think anyone suggests introducing a carbon tax without also claiming there are damages from carbon emissions. The message I take is that the environmental benefits can be obtained at a much lower cost than would otherwise be possible. making a carbon tax with offsets an even better deal.

    • Bob Murphy says:

      Harold wrote: “We can conclude that in practice the minimisation of deadweight loss and widening the base is not the only or even overriding factor.”

      Harold, I agree that my statements could have been clearer, but I didn’t mean to suggest that the ONLY principle in tax design was to minimize deadweight loss. After all, I said: “However, this seemingly sensible approach of “tax ‘bads,’ not goods” overlooks other standard results in the public finance literature.”

      Another standard principle is “fairness” and another is “ability to pay.”

    • Bob Murphy says:

      Harold wrote: “That is not really the point as I don’t think anyone suggests introducing a carbon tax without also claiming there are damages from carbon emissions. The message I take is that the environmental benefits can be obtained at a much lower cost than would otherwise be possible.”

      Harold I literally quoted Sumner on this. And yes there are lots of conservative/libertarian pitches in this genre, which explicitly say “Even if you think Al Gore is full of it…”

      • Harold says:

        “I literally quoted Sumner on this.”

        Your quote is “Even if global warming is not a problem, there are obviously much worse taxes than a carbon tax.”

        I am sure you would agree with this. There are always worse taxes.

        • Bob Murphy says:

          Harold sometimes it’s exhausting to argue with you. In context Sumner is clearly including standard income taxes in his statement. And in any event, you moved the goalposts. You originally said that nobody was saying, “Pretend for a moment this isn’t a problem,” and that instead all people were doing is seeking to minimize the negative impact of carbon taxes.

      • Harold says:

        To elucidate, my take is that Sumner believes there is a social cost to carbon emissions and therefore a carbon tax would be a good thing. I don’t believe he would advocate for a carbon tax in the absence of a belief in environmental harm.

        His article is about how to get this implemented politically.
        “One solution would be to offset the carbon tax with an equally large reduction in even less popular taxes.”

        the key here is that he says “even less popular taxes” rather than “even less efficient taxes”

        Therefore countering his argument by saying a carbon tax is less efficient that an income tax is missing the point.

        • Bob Murphy says:

          Harold from one post above: “the key here is that he says “even less popular taxes” rather than “even less efficient taxes””

          Harold from 2 posts above: “Your quote [of Sumner] is “Even if global warming is not a problem, there are obviously much worse taxes than a carbon tax.”

          I am sure you would agree with this. There are always worse taxes.”

          Harold, I’m not messing with you, I’m being dead serious: Can you see how it’s annoying to argue with you? It’s not your politics it’s that you do a 180 literally from one comment to the next.

          • Harold says:

            I am sorry you find it exhausting. I do try to make my points clearly but often I am less than successful, and I also make errors. However, I don’t think I have made a 180 here.

            I said
            “That is not really the point as I don’t think anyone suggests introducing a carbon tax without also claiming there are damages from carbon emissions.”

            Your response is to quote Sumner saying there are worse taxes than a carbon tax.

            This is not demonstrating that anyone wants a carbon tax in the absence of environmental damage.

            Maybe I am wrong, but I have not seen anyone calling for a carbon tax without also claiming there is a cost to carbon emissions. Please show me if I am wrong.

            Point 2
            Bob said “You originally said that nobody was saying, “Pretend for a moment this isn’t a problem,”

            No I did not say that. I was careful in my language.

            I have seen people argue that even if there were no damage from carbon it would be a good idea anyway. Due to point 1, I think this is obvious nonsense. If it was a good idea we would expect some people to reject the idea of environmental damage from carbon but still think a carbon tax was a good idea. maybe there are such people but I have not seen them. I think people are simply over-selling the benefits of C tax.

            However, your chart shows that this could actually be true if we replaced capital taxes. I strongly suspect that advocates for reducing capital taxes would not choose a carbon tax as a replacement, for the reasons you make clear in your article.

            Point 3.
            I don’t think Sumner is doing that here because he makes it clear that he is talking about politics of getting a carbon tax adopted. It is not at all clear he is saying that a carbon tax would be more efficient than income tax.

          • Harold says:

            Also see Josiah’s comment on Laffer, who does apparently claim economic benefits without believing in environmental benefits.

    • George Klingerhoffer says:

      “This is not actually true, or a poll tax would be universal…”

      Or maybe politicians don’t always do optimal tax design?

      • Bob Murphy says:

        Well in fairness I think Harold meant, “Or a poll tax would be universally recommended” which it isn’t, except by a few economists.

      • Tel says:

        You would expect that by definition it should be impossible to have a system that is simultaneously both “optimal” and widely hated by the people who live within it. Thus, if the politicians see the pitchforks and firebrands start coming out en masse (as happened to Thatcher and her poll tax) they quickly decide that the meaning of “optimal” has become something different.

        What I’m saying is, a “free market” is fundamentally about people doing the things that they want to do. Money is a mechanism that can help us achieve this purpose, but it’s an intermediary step at best. Although you can’t make everyone happy in this regard (some people will desire to do things that put heavy imposition on other people) you should at least aim to make quite a lot of people happy, most of the time. The politicians do have a well developed ear to detect public reaction to their policies… this being their primary survival skill.

        And yes, I understand that democracy quite often swings on a small number of individuals, you can get crooked politicians, badly done vote counts, confused voters who don’t understand what they are voting for and a bunch of other problems. Even after considering those issues, at least the system does feed public opinion back into the loop, which is something economists seem surprisingly reluctant to do.

        • Silas Barta says:

          Tel, I think I made your point here in different words: efficient taxes — because they’re unavoidable — are likewise anti-liberty.

    • Silas Barta says:

      >This is not actually true, or a poll tax would be universal.

      Well, even efficient taxes can become inefficient taxes, once the rate is high enough to encourage absurd avoidance behavior.

      I like to joke that, at a high enough head tax, people “give away” their heads to avoid it, with obviously bad consequences. (The grown-up version of the joke is funnier but I can’t say it here.)

  3. Transformer says:

    I don’t see how you can be confident in your conclusion that ‘a carbon tax would cause net economic harm, even if implemented as part of a broad “swap.”’.

    I agree that the politics of implementations would almost certainly render it sub-optimal. And I agree that judged purely as a revenue raising exercise the economic argument you make are sound.

    But if there really is a social cost of carbon (as the science strong indicates) then a global carbon tax (even sub-optimally implemented) does have the potential to reduce economic distortions arising from the externalizes of pollution not being adequately priced-in and so avoid rather than cause net economic harm.

    • Transformer says:

      Perhaps I’m missing something in the IER articles but

      – A tax on carbon that makes consumers of carbon pay its social costs will be socially beneficial

      – If the revenue raised is used to reduce other taxes then the overall dead-weight loss from these other taxes will be reduced,

      How can the net effects of these two beneficial things not be positive ?

      • Transformer says:

        And I think the answer may be that the Resources for the Future (RFF) study that Bob quotes only shows costs not benefits of a carbon tax.

        So in other words they claim that the negative effect of a carbon tax of GDP would be eliminated if the revenue raised was used to fund a reduction in capital tax. But that is without taking into the effect the other benefits to the environment and (I assume) to GDP from avoiding some of the economic costs of global warming that is the whole rationale of the carbon tax.

        • Bob Murphy says:

          Transformer you should click through and read my stand-alone article on the tax interaction effect. It’s like this:

          ==> Suppose there were no taxes, and social cost of carbon were $50/ton. Then the textbook Pigovian tax would be a carbon tax of $50/ton, with receipts returned lump-sum to citizens.

          ==> But then you realize we’re not starting with no taxes, we’re starting with distortionary taxes on labor income (for example). So what if we used carbon tax receipts not to give lump-sum rebate, but to reduce the marginal tax rate on labor income in a way that ends up being revenue neutral, then you might think the optimal carbon tax should be *higher* than $50/ton, because we’re (a) getting environmental benefits and (b) reducing deadweight loss from income tax.

          ==> But no, most modeling thinks that’s wrong, that actually you’d enact a carbon tax of less than $50/ton, because even though the rates end up lower, the income tax is *more* distortionary in the presence of a carbon tax.

          • Transformer says:

            Wow, that is a very interesting article and the Tax Interaction Effect is something that should definitely be more widely discussed in the carbon tax debate.

            As someone who is both pro-market and who has serious environmental concerns I see now that the correct approach is a carbon tax combined with policies to move the world closer to scenario 2 from your article !

      • Tel says:

        – A tax on carbon that makes consumers of carbon pay its social costs will be socially beneficial

        Only if you believe that phrases such as “social cost” and “socially beneficial” are meaningful and measurable. I’m guessing you have been discussing the matter with society; but speaking for myself I’ve never figured out how to do that. All reasonable suggestions will be considered!

        • Harold says:

          One method is to express everything in terms of money.

          • Transformer says:

            A reasonable definition of something being “socially beneficial” would be if the winners from that thing could (n theory) be able to fully compensate the losers and still be better off themselves.

            I discussed this with society and they agreed this applied to a carbon tax.

            • Tel says:

              If you are going to measure money then I would accept transactions that actually take place, on the basis that when someone voluntarily pays for something that means they probably wanted it … but theoretical transactions that perhaps could take place but never do is pure make believe, not a measurement.

              • Harold says:

                Things can be meaningful without being measurable.

              • Tel says:

                Abstract things like algebra don’t require measurement because they exist in their own abstract universe.

                Those things are meaningful in as much as they provide tools, but they don’t tell you anything about the world around you. Algebra in the abstract would continue to exist in every possible world, because it makes no reference to the world around it. This is one of the arguments as to why God cannot defy the very concept of logic, even though God might be materially powerful, and this would apply to any God-like being you can imagine.

                Getting back to the topic at hand, if you want a valuation that is not abstract, you need to take a measurement of something.

  4. Silas Barta says:

    I think your point highlights why the downside is so unintuitive, even for me. That is, it’s hard for most people to imagine there being a “too high” carbon tax (at least for some reasonable values). Even if you taxed it above some economist-calculated value, people reason that, “heck, it will have positive extrenalities in terms of super-energized incentives to improve energy efficiency or research into greener tech”.

    And, tbh, I personally wouldn’t mind if the income tax were abolished in favor of a carbon tax; it’s much easier to cut back my fossil fuel usage than my income in general.

    (And yeah, I know, that’s kind of the problem, isn’t it? A tax that’s good in terms of “easy to avoid” will be bad in terms of economic efficiency.)

    • Bob Murphy says:

      Silas can you find on your blog (years ago) when you took me to the woodshed and did a more intuitive proof of the TIE? The thing about the deadweight loss being proportional to the square of the income tax rate etc?

  5. Transformer says:

    When researching this yesterday I cam across some good old posts


    and this one from Silas that sounds like it may be the one you were asking about


  6. Transformer says:

    This one from Silas sounds like it may be the one you were asking about. It is by far the best explanation of TIE I found while researching yesterday.


    • Silas Barta says:

      Thank you, Transformer!

      Yeah, I really struggled with that when Bob first posted about TIE and did my best to come up with an intuitive explanation of

      But the post also concludes that there’s something of a paradox, in that the TIE argument implies you shouldn’t charge for scarce goods if there exist any other taxes, because “hey, it introduces an inefficiency”.

      (And thanks for not making a big deal about my bad Latin translation of “whose desire, his expense”…)

    • Silas Barta says:

      Actually, those other points I mention were in a follow-up post that discusses why a carbon tax shift would be a good idea nonetheless (it’s where I do my bad Latin):


    • Silas Barta says:

      Another interesting follow-up: a while back, Bob pointed me to this paper which bit the bullet on the paradox.

      That is, because of tax interaction effects, it may indeed be inefficient for a government to fight back against overfishing by imposing tradeable quotas!

      • Transformer says:

        An even simpler intuition on TIE may be just to see it in terms of how tax affects utility.

        If you are taxed 10% of your income you give up the least valuable 10% of your spending.

        If you are then taxed a further 10% the slice of utility you give up is greater than for the first 10%.

        A carbon tax reduces income in the same way as other taxes and the amount of utility it forces you to give up will be greater the more tax you are already paying.

        (I think this is pretty much what is happening with the supply and demand charts that are used to demonstrate dead-weight loss – as additional taxes cause the equilibrium price to increase each additional unit of tax causes proportionately more consumer surplus to be lost)

        • Transformer says:

          After a cup of coffee I realize this is probably a simplification too far in how deadweight losses work.

          • Bob Murphy says:

            Actually I think you are on the right track with it, Transformer.

            • Transformer says:

              Oh, thanks. Maybe I should give up coffee.

            • Josiah says:


              I don’t think your explanation works for this reason: We are considering a scenario where the new tax is paired with cuts in existing taxes. If the new tax increases deadweight loss from other taxes, then the tax cut ought to correspondingly lessen the deadweight loss from other taxes.

              • Transformer says:

                As long as the revenue from the carbon tax fails to eliminate the need for other taxes then there will still be a dead-weight loss from these other taxes.

                Any remaining dead-weight loss will then reduce the optimal level of the carbon tax since the utility loss it causes will be magnified by the existing taxes.

                In other words if there were no other taxes the losses to utility from a carbon tax would (if it were optimal) be exactly offset by the gains from the reduced carbon usage.

                But if there are existing taxes then the optimal carbon tax from an otherwise tax-free world applied to this other-tax scenario will cause utility losses greater than the gains from less carbon use so the optimal level of carbon tax will be lower.

                Can’t help but suspect that my explanation is less than crystal clear – but I’ve gone 10 hours with no coffee now.

              • Transformer says:

                Probably missed out the word “marginal” a few times.

              • Transformer says:

                And I think it quite possible that a carbon tax used to reduce other taxes will in aggregate reduce deadweight loss, but this will not necessarily prevent the interactions from other any remaining taxes reducing the optimal carbon tax level from what it would be with no other taxes.

              • Josiah says:


                Suppose you have a country with two taxes: a X tax and a Y tax. Someone proposes introducing a new Z tax and using the proceeds to eliminate the Y tax.

                It’s true that adding a Z tax will effectively increase the X tax as well, due to the tax interaction effect. However, the tax interaction effect also implies that getting rid of the Y tax would effectively cut the X tax. So you’d need to invoke something distinctive about the Z tax or Y tax to see whether the swap is a good idea.

              • Transformer says:

                Yes, I believe that what you say is correct.

                I don’t think the existence of TIE will determine if a carbon tax is a good idea or not, or even what effect a carbon tax of a certain level will have on total deadweight tax loss because as you say it is both reducing the deadweight loss of the existing taxes (by reducing them) and magnifying the deadweight loses of any taxes that it fails to eliminate.

                I think all that can be said is that the optimal level of a a carbon tax in an economy is probably going to be inversely correlated with the level of deadweight losses (post carbon tax) in that economy .

              • Transformer says:

                So to apply this to Josiah’s example:

                It is possible that the elimination of Tax Y will reduce the deadweight loss from Tax X tax more than the introduction of Tax Z will increase it. Net deadweight loss decreases as a result of Tax Z.

                Nevertheless the optimal level of tax Z (if its a tax on an externality) will likely be lower with the presence of Tax X that without it.

  7. Josiah says:


    As you know, your old boss Art Laffer favors a swap of a carbon tax for income tax cuts and believes this will boost the economy, even though he is, by his own admission, agnostic about global warming. Suppose you showed him the model results from RFF showing that a carbon tax/labor tax swap would have a slight net negative economic impact. What do you think he would say?

    • Harold says:

      Well, maybe I was wrong and Laffer is the guy. However, one article says ” At the same time, the U.S. allows something we want less of — carbon dioxide pollution — to be emitted without penalty.”

      Why does the USA want less of it? He says he is neutral on climate change, so he presumably does not thin k the USA wanting less CO2 makes any sense.

      However he does seem to want to swap these taxes for the economic rather than environmental effects.

  8. Tel says:

    Not directly related to tax, but this is a great way to convey meaning with a graph.


    • Harold says:

      It is persuasive whilst being misleading and very simplistic.

      This is what happens of you get your information from blogs which have an agenda. If you go to the science you find a different story.


      “Satellite altimetry has shown that global mean sea level has been rising at a rate of ∼3 ± 0.4 mm/y since 1993. Using the altimeter record coupled with careful consideration of interannual and decadal variability as well as potential instrument errors, we show that this rate is accelerating at 0.084 ± 0.025 mm/y2”

      These scientists have done a proper analysis using satellite data rather than naively fitting a straight line through data from one land based station.

      So if your intention is to mislead, I agree your chart is very effective.

Leave a Reply