22 Oct 2013

Ross McKitrick Explains Why There Probably Is NOT a “Double Dividend” With Carbon Tax

Climate Change 35 Comments

I rarely get down and beg, but if you are a serious student of the economics of climate change debate, I implore you to watch Ross McKitrick’s presentation from IER’s carbon tax conference over the summer. I’m embedding the video and then adding my commentary, but for more context go read my post at IER.

The takeaway message here is that McKitrick–who is truly a world expert on the subject–shows that the peer-reviewed literature comes down on the conclusion that even a 100% dollar-for-dollar swap of income tax cuts for a new carbon tax would probably hurt conventional economic growth. So the people running around saying a carbon tax is a “win-win” because now we’re going to “tax bads, not goods” are Consensus Deniers.

  • 1:00 – 2:00 McKitrick explains the “Pigovian” (named after the economist Pigou) analysis in which a tax on a “negative externality” is supposed to help, by reducing the behavior, such as emissions, toward the “socially optimal” level. But nobody really looked at how the tax revenue was used.
  • Around the 3:00 mark, McKitrick explains that in the 1990s, economists began studying the effects of a completely revenue-neutral tax on a negative externality. This is where the “double dividend” idea comes into play, since—in theory—it should be good to (a) levy a tax on the behavior causing the negative externality and (b) raise revenue that will allow taxes on socially useful things (like labor and saving) to be reduced. Some economists became excited that “revenue recycling” (where the revenue from an emissions tax, say, is devoted to income tax or capital gains tax cuts) provided a painless way for the government to help both the economy and the environment.
  • At 4:05 McKitrick raises the issue of the “tax interaction effect,” which caused researchers to pause in their tracks. The new tax—even if it’s levied on a negative externality—will still interact with the pre-existing tax code, exacerbating the original inefficiency in the code (or raising the “excess burden” of those original taxes). Thus the benefits of “revenue recycling” must be contrasted with the harms of the tax interaction effect. Is there a double dividend? Only if the former outweighs the latter.
  • At 4:30 McKitrick says the “overly strong” claim was clearly false: We can’t say that any emissions tax that is coupled with 100 percent revenue recycling is always better than the status quo, even though that’s what some economists had originally thought. Thus, the people today who say matter-of-factly that the government should “tax bads not goods” are decades out of date, with respect to the actual published literature on the topic.
  • At 4:45 McKitrick discusses a weaker version of the double dividend claim, which says that once we consider the possibility of revenue recycling, the government ought to set the emissions tax higher than the level that reflecting the negative externality. So for example, if the “social cost of carbon” is $35 per ton, then this weaker version of the double dividend hypothesis says that the government should set the actual carbon tax at higher than that, perhaps $40 per ton, since the carbon tax receipts not only help mitigate global warming, but also allow for pro-growth income tax cuts. Yet, McKitrick says that researchers in the field quickly concluded that even this weaker claim was probably false.
  • Starting around 6:45 McKitrick explains that the first economists to develop models with enough detail to show the possibility of a revenue recycling effect, discovered that in fact the government should set a carbon tax much lower than the “social cost of carbon.” In fact, with plausible estimates of the actual “social cost of carbon,” the optimal carbon tax could be $0 per ton. This shows the tremendous importance of the tax interaction effect.
  • Starting at 8:00 McKitrick tries to explain why the pre-existing tax code matters so much in these analyses. Part of the explanation is that the emissions tax has a smaller “base” than broader taxes such as income taxes. There is a general rule in the tax literature that a tax should be applied on as large a base as possible; this is one reason that a “revenue neutral carbon tax” can harm the economy.
  • Around the 12:00 mark, McKitrick showed that actually Sandmo in 1975 had provided the general framework for thinking about these issues. These remarks will probably only help professional economists make sense of the seemingly counterintuitive result that the government should set an “optimal” carbon tax less than the “social cost of carbon.”

35 Responses to “Ross McKitrick Explains Why There Probably Is NOT a “Double Dividend” With Carbon Tax”

  1. Darien says:

    “Pigovian” is the second-best economic adjective, trailing only “Walrasian,” a term used to describe Chinese walruses.

  2. joe says:

    Expert? McKitrick is a signatory to the Cornwall Alliance’s Evangelical Declaration on Global Warming, which states that “Earth and its ecosystems – created by God’s intelligent design and infinite power and sustained by His faithful providence – are robust, resilient, self-regulating, and self-correcting”.

    In other words AGW is nothing to worry about – God will save us! But if you are a Muslim, Hindu, Buddhist or atheist you are doomed?

    • Major_Freedom says:

      “All that counts is whether a doctrine is sound or unsound. This is to be established by discursive reasoning. It does not in the least detract from the soundness and correctness of a theory if the psychological forces that prompted its author are disclosed. The motives that guided the thinker are immaterial to appreciating his achievement. Biographers are busy today explaining the work of the genius as a product of his complexes and libidinous impulses and a sublimation of his sexual desires. Their studies may be valuable contributions to psychology or rather to thymology (see below p. 265), but they do not affect in any way the evaluation of the biographee’s exploits. The most sophisticated psychoanalytical examination of Pascal’s life tells us nothing about the scientific soundness or unsoundness of his mathematical and philosophical doctrines.

      If the failures and errors of a doctrine are unmasked by discursive reasoning, historians and biographers may try to explain them by tracing them back to their author’s bias. But if no tenable objections can be raised against a theory, it is immaterial what kind of motives inspired its author. Granted that he was biased. But then we must realize that his alleged bias produced theorems which successfully withstood all objections.

      Reference to a thinker’s bias is no substitute for a refutation of his doctrines by tenable arguments. Those who charge the economists with bias merely show that they are at a loss to refute their teachings by critical analysis.” -Mises, Theory and History, Ch. 2, Sec. 1.

      • Major_Freedom says:

        That’s called a smackdown, for any of you who missed it.

        • Bob Murphy says:

          Reminds me of the time I foolishly referred to Einstein as a world expert in physics, and someone corrected me with his quote, “God does not play dice.”

          • Matt Tanous says:

            Bob, don’t you know that religious believers cannot be experts! Because all scientific experts are automatically atheists, because that’s the only rational position to hold. I know, because I read that from some guy named Dawkins once, and he was really snide about it too…

          • Major_Freedom says:

            Hahaha

    • Matt Tanous says:

      In other words, what Mises was saying in that quote was simply this:

      The statement of fact here is “Earth and its ecosystems are robust, resilient, self-regulating, and self-correcting”. Whether they were created and sustained by God is a secondary statement, and the robustness and resilience of the ecosystem is not dependent on whether you accept that as true or false.

      • Major_Freedom says:

        That’s right.

        Whether or not there was a God that created the universe, is at the end of the day immaterial. It is immaterial because humans have reason, and can only ever understand the universe with reason. The source of reason can be assumed as “natural” in the sense of non-supernatural causes, or it can be assumed as a gift from a God, with no fundamental affect on the efficacy of it.

        Either way, whatever argument someone presents, must pass the test of discursive reasoning, in order for one to legitimately claim that the argument is soundor unsound.

  3. Grant McDermott says:

    As i) “a serious student of the economics of climate change debate” and ii) someone that actually learnt about the tax interaction effect from Agnar Sandmo himself, I have some comments…

    First up, Sandmo is Norwegian, not Swedish. More importantly, he also most certainly supports a putting a price on carbon… as does virtually everyone who understands the tax interaction effect (TIE) that I am aware of. You may claim that this is merely an appeal to authority, but I just want to point out that cognisance of the TIE is hardly viewed as decisive blow to climate policy by leading scholars.

    Second, I understand that this is a quick presentation intended for a particular audience. However, I find it strange that McKitrick insists on referring to the oldest and simplest CGE studies (e.g. Parry, 1995)… rather than taking a look at what more recent and complex papers have to say. For example, here, here and here. Taking the last paper as an example, since it is by a friend but also happens to be one of the most complete analyses that I have seen, incorporating the TIE into an Integrated Assessment Model means that “the optimal tax in 2010 is only about 3.5 percent lower than in first-best”. In other words, basically identical to the Pigouvian tax level in a setting without any TIE.

    Now, some previous studies have suggested that income distortions like the TIE mean a carbon price should be around 25% lower than the Pigouvian level. (E.g. the second paper I link to, which is by Barrage.) However, even that is clearly difficult to reconcile with McKitrick’s sweeping claim that the TIE renders a carbon price totally irrelevant.

    In an ironic departure from the way these things normally work, Bob’s nuanced headline actually happens to be more accurate than the overstated content of the post.

    • Bob Murphy says:

      Grant McDermont wrote:

      but I just want to point out that cognisance of the TIE is hardly viewed as decisive blow to climate policy by leading scholars.

      OK, and I just want to point out that there are a lot of people running around saying, “So long as they used the revenues from a carbon tax to reduce income taxes or payroll taxes, it would not only help the environment but also boost output.” They say it matter-of-factly, like the the only doubt is whether we can trust those rascally politicians and make sure they do indeed use the carbon tax receipts to cut other taxes.

      And I am saying they are ignoring the peer-reviewed literature. I didn’t say anything else on this point.

      I realize you didn’t claim that I did say more, but my point is that this is a HUGE issue that has been utterly neglected in the policy debate. I personally introduced it into the US debate about a year ago. Not that I invented the idea, but that I brought it into the debate because even people on “my side” had never heard of it.

      In fact, I learned of the 1996 Bovenberg and Goulder piece from an MIT (I think) paper where the authors claimed matter-of-factly that a carbon tax was a ‘win-win-win” and cited B&G’s 1996 paper to show why. (!!!)

      So I hope you can forgive me for telling everyone to watch McKitrick’s video. Nobody knows this stuff in the US policy debate; in fact the opposite conventional view has taken hold because of pro-carbon tax people saying stuff that isn’t true, and even citing B&G 1996 and telling their readers they showed the framework in which a carbon tax swap boosts conventional growth.

      • Silas Barta says:

        “So long as they used the revenues from a carbon tax to reduce income taxes or payroll taxes, it would not only help the environment but also boost output.” They say it matter-of-factly, like the the only doubt is whether we can trust those rascally politicians and make sure they do indeed use the carbon tax receipts to cut other taxes.

        This may be terminological issue, but I think the actual claim is that the tax shift would improve economic growth *net of the social costs of carbon*, That is, perhaps CO2-harm-ignoring economic growth would be lower, but with all costs accounted for it would be higher.

        • Matt Tanous says:

          “That is, perhaps CO2-harm-ignoring economic growth would be lower, but with all costs accounted for it would be higher.”

          Well, it’s obvious they aren’t claiming that, for instance, coal power plant production would be greater. They are clearly arguing that overall economic growth would be, though.

        • Bob Murphy says:

          Silas wrote:

          This may be terminological issue, but I think the actual claim is that the tax shift would improve economic growth *net of the social costs of carbon*, That is, perhaps CO2-harm-ignoring economic growth would be lower, but with all costs accounted for it would be higher.

          No Silas that’s absolutely not what they’re claiming. Skim this article, for example. I think this is the one that actually cited Bovenberg and Goulder for their claims, which is like me citing Rothbard on why Hayek’s knowledge problem was more important than Mises’ calculation problem.

    • Bob Murphy says:

      Grant, I’m not making any promises, but suppose I want to do a follow-up post on IER in response to your claims. Do you want me to keep you anonymous, or mention your name, affiliation, etc.?

  4. Harold says:

    i think you are using the word “consensus” in an unusual way. One persons opinion about what the literature says is not a consensus. A consensus is agreement among a group – in this context it would be a group of expert economists. Fortunately we do have access to the opinions of such a group – the IGM economic experts panel. They recently answered the question “…a federal carbon tax at this rate [ $20/tonne] would involve fewer harmful net distortions to the US economy than a tax increase that generated the same revenue by raising marginal tax rates on labor income across the board.” Nobody disagreed, and only 3% were uncertain. Now that is a consensus. The question is not quite the same, and it is quite possible that this consensus is compatible with McKitrick’ view – it would seem unlikely that every one of these economists would disagree with the literature. Can you explain why it is possible for the tax to less distortionary, but still for a dollar for dollar swap to be harmful?

    http://www.igmchicago.org/igm-economic-experts-panel/poll-results?SurveyID=SV_8oABK2TolkGluV7

    Ultimately, it is not the consensus that matters, it what is actually is in the research. The consesnsus is just a proxy, but usually a consensus of experts is quite a good proxy for the published information.

  5. Ross McKitrick says:

    Grant, thank you for the 3 references, none of which I have seen but which I will look at more carefully. From my first look, the RFF one reaches the familiar conclusion on page 8: when they compare their set-up to the Bovenberga and deMooij framework they show that they also derive a tax rule in which the optimal tax is below the Pigovian level by the amount of the TIE. What they do next is simulate the results of an arbitrarily-chosen $50/ton Ctax and they look at optimal rebates and leakage rates. The 2nd paper (Barrage) reports a TIE effect that reduces the optimal tax rate by 25-35% below the pigovian level, which is in line with the rough numbers in my presentation. What drives his claim of a large optimal Ctax of course is his model with large positive damages. The 3rd paper is a work in progress and arrives at a very small TIE, so i’ll have to see how that comes about, but the author notes that his result is in contrast to previous work on the subject.
    Large Ctaxes require not only large damage function parameters, but an assumption about how they develop in the future. My suggestion about handling the dynamics is to use state-contingent updating and a futures market: see http://www.sciencedirect.com/science/article/pii/S0140988310001106 and http://www.thegwpf.org/content/uploads/2013/07/McKitrick-Carbon-Tax-10.pdf.

    Harold – I suspect your panel of experts would also agree that a $20/tonne carbon tax would have little effect on US emissions, which is the point I was making at the end. Also, I doubt the panel were assuming that it was $20/tonne plus a long list of emission regulations, they were probably assuming it was a carbon tax instead of all the regulations.

  6. Harold says:

    “they were probably assuming it was a carbon tax instead of all the regulations” Elsewhere they pretty much all agree that a carbon tax would be better than regulations on emissions. I can only presume that they are dealing with the question as asked, which does not include regulations, so yes, I would agree that they were thinking of a carbon tax instead of the regulations. However, I think that is besides the point here. I am not an expert by any means, and I would not wish to argue with the results of research as disseminated in the literature, but it seems that saying a $ for $ swap of income tax to a new carbon tax would hurt growth is not a consensus, unless we can square hurt growth with less distortion. I notice that the question asked of teh panel was whether a new carbon tax would result in fewer harmful distortions than a new labor tax. This is different from asking would a new carbon tax combined with an equivalent cut in labor tax would result in less distortion than the status quo. Is this the significant difference?

  7. Blackadder says:

    Bob,

    Suppose we repealed the gas tax and made up the revenue by increasing marginal income tax rates. Is it your (and McKitrick’s) view that this would boost economic growth?

    • Bob Murphy says:

      Blackadder,

      Well I am always hesitant to endorse anything that boosts any kind of tax. So on the understanding that I’m answering your question as an objective question about predictions under certain assumptions, then:

      ==> I wouldn’t be surprised if the answer is “yes,” but notice that a gas tax is not equivalent to a carbon tax. A gas tax has often been instituted to take the place of road tolls etc. So, the burden of proof is higher on your scenario than on what McKitrick is talking about.

      • Blackadder says:

        Bob,

        I would think that the tax interaction effect would be worse for a gas tax than for a carbon tax, since the base is smaller. You are right that it would matter how you replaced the lost income (e.g. tolls vs. taxing the rich). My hypothetical was that you replace the lost revenue by raising marginal income tax rates.

        • Bob Murphy says:

          Blackadder well to the extent that a gas tax limits driving on congested roads, one could argue that it is a 19th-best solution to that problem (caused by governments setting tolls too low, or owning roads in the first place). So all I meant was, it’s easier to see how road congestion could hamper conventional economic growth in the near term, versus climate change that won’t cause net damages (according to Tol for example) for decades to come.

  8. Silas Barta says:

    Could you provide some of the intuition behind these “tax interaction effects” that negate the benefit of a shift it taxes from “goods to bads”? Your post is really thorough and yet it still comes across as “Yeah, you’d think that would be a net benefit but [magical stuff happens] and it turns out it’s not.”

    I suppose if I put aside 30+ minutes to listen I might “get it” but surely there’s some key insight that helps get a feel for why the intuitive benefits of the tax shift don’t manifest?

    Is it mainly because the benefits from reduced carbon emissions (per the consensus models) are so low to begin with?

    I mean, both taxes carry deadweight loss and evasion costs, but the situations are asymmetric since much of (what would be) the DWL on a carbon tax is offset by the social benefit of carbon reduction and thus is not a pure loss. Plus, the taxed “goods” — labor and investment — throw off benefits in the form of consumer/producer surpluses, which would increase as the taxes thereon decline.

    So yeah, I’d really like to get a feel for the mechanism behind this “tax interaction” that would somehow negate all that.

    • Bob Murphy says:

      Silas did you ever see this article? That should get you going, and then if you want to up the ante check out the 2000 (?) Goulder piece that I cite in the footnotes.

  9. Blackadder says:

    I’m kind of in the same boat as Silas here. I read your IER article when it came out and found it intriguing, but I’ve never been able to quite grok the argument.

    Take another case. We give out speeding tickets not because we think it’s a particularly efficient way of collecting revenue, but because we want to discourage speeding. If we didn’t care about speeding, no one would propose that we randomly stop drivers to collect an equivalent amount of revenue instead.

    Likewise, ignoring the externality issue, taxing carbon would be less efficient than, say, taxing income, and so the only reason we would even think of doing it is because we think there is an externality involved. Further, the loss from using less efficient carbon taxes has to count against the benefit from reducing the externalities associated with CO2 (how much is canceled out seems to be a matter of some controversy).

    That makes sense to me, but what does any of that have to do with “tax interaction”?

    • Bob Murphy says:

      Guys, I’m confused. I am providing you with people claiming that Americans should support a carbon tax because it will help the environment *and* boost the conventional economy. There are people pitching that argument specifically to conservatives, saying it is a great way to get Al Gore to support cutting income taxes, whether or not he’s right about global warming.

      You don’t think it’s worth me devoting a blog post featuring an expert saying that this is likely wrong, according to the literature? I shouldn’t object to bogus arguments for a carbon tax?

      And also, the numbers in the original B&G 1996 paper aren’t little. They are humongous. I’ll do another post on this at IER, because I’m surprised you think this is so trivial.

      • Blackadder says:

        Bob,

        I think you misunderstand me. I’m not claiming that your post is a waste of time, or attacking straw men, etc. In fact, I agree with you that a carbon tax would be a bad idea. It is genuinely the case that I don’t feel like I have a good grasp of the argument and I’m trying to understand it better.

        • Bob Murphy says:

          OK sorry if I’m defensive. I think the trolls should be taxed to prevent this in the future.

  10. Joseph Fetz says:

    “Aboot”

    Yep, he’s Canadian.

  11. Harold says:

    I am with Silas Barta above when he says it is difficult to grasp what the TIE effect is – in the article you link to you say “The precise reasons for the extra deadweight loss depend on the particular model, and ultimately it is hard to point to a “cause”" This does not make it easy. I think Silas was hoping you could sum it up, but perhaps it is just too complicated.

    So the take home message is that the carbon tax is not as much of a win-win as people have claimed. The first win is “correcting” the market for carbon to take account of the externality of carbon emmissions. The second win is stimulating growth through reducing other distorionaly taxes. It may be a win -1/2win, or possibly just a win, with no second win flowing from the switch in tax from wages to carbon. At the extreme end of the models, it could be an overall lose. Does this summarise it, or have I got the wrong end ot the stick again?

    • Bob Murphy says:

      Harold,

      OK you’re right, nobody (including me) has done a great job of summarizing the exact cause; partly the reason I haven’t done it, is that people offer different intuitions in the literature, and so it’s hard for me to summarize it since it’s not clear if there are multiple reasons, or if some people aren’t hitting the nail on the head when they give a shot at the intuition. It pops out of a general equilibrium framework; i.e. when you look at the economy with just the income tax you get a certain deadweight loss for a given amount of revenue, then you throw in the carbon tax and cut the income tax to hold revenue at the same level, you find the new equilibrium, and lo and behold there’s more deadweight loss than in the original scenario. But it’s hard to point to “what’s driving the result” except to say, “Adding the carbon tax drove the result.” See what I mean?

      I’m going to do a follow up post on this stuff.

      • Silas Barta says:

        Okay Bob, I’ve read the link you gave me above, and I think I have an explanation that conveys the intuition.

        Insight 1: the harm of a tax is more-than-proportional to its magnitute. (This is the assumption that the writing on this seems to assume and which I wish was made explicit here and in your article.) I believe Mankiw gives the rule of thumb that the DWL of a tax increases with the square of the revenue it taxes, or the tax rate, I forget which. Thus why you want to raise a given amount of revenue from as “broad a base” as possible.

        Insight 2 (most important): Because of the above, each increase in tax above the Pigouvian level is more harmful than the same increase from zero.

        Insight 3: Taxes on anything chase back to their original land/labor/capital factors. So a carbon tax amounts to a tax on land, labor, and capital, divided up per their relative supply curve slopes (or something like that).

        Given the above, the intuition becomes a lot clearer: a tax on carbon is like an income tax (with different levels for different kinds of income). Levied at the Pigovian rate, it merely cancels out the carbon harms. But if you have an additional (direct) income tax, you get a disproportionate harm for each (potentially) taxed dollar above the Pigovian level (compare to taxing from the first dollar) — *that* is the tax interaction effect.

        Furthermore, since the “green tax trade” tries to raise the same revenue on a smaller base (i.e. only those income sources touching carbon), the tax rates have to be much higher than they would be if they were on all income. This then causes major welfare-harming changes in behavior, far out of proportion to the assumed harms from carbon.

        (Incidentally, this is another one of those cases where a simpler explanation, in retrospect, saves a ton of time. Do you remember our scuffle over liability insurance, and how I finally understood it as soon as someone (you, maybe) pointed out that diminishing marginal utility = increasing marginal disutility?

        This is like that.)

        • Bob Murphy says:

          Silas, I have to think about your approach a bit, but in the meantime: Is your “Insight 1″ necessary? It’s true, but I don’t think you need it.

          • Silas Barta says:

            I’m pretty sure it’s necessary. Without it, it doesn’t matter where you place a tax; $X of revenue raised would do (less than) k*$X harm, whether it’s raised from a tax on shoes or a tax on all income.

  12. Harold says:

    I will await it with interest.

Leave a Reply