Carbon Tax or Cap & Trade: Does It Matter Which One?

Lynne Kiesling

The subject has roiled for the past few months: if the balance of the evidence has shifted toward the value of our taking more actions to reduce our greenhouse gas emissions, should the U.S. implement a carbon tax or a carbon market? Stated that bluntly, and based on centuries of experience with taxes and markets, the person who favors individual liberty and the use of market processes can be expected to favor a carbon market over a carbon tax, even if such a market involves government determination of the number of emission permits allowed in such a market.

Yet some prominent writers with classical liberal, pro-market inclinations have argued in favor of taxes over markets in this case, most recently Ron Bailey in Reason, Tyler Cowen at Marginal Revolution and George Mason University, Greg Mankiw at Harvard, and John Tierney at the New York Times. The imminently sensible John Whitehead sees little difference between the two policies.

I disagree, because I think all of these thoughtful and intelligent observers are making a fatal assumption and ignoring some crucial arguments that, for me, swing the balance toward carbon markets, notwithstanding the difficult design issues and costs. These arguments are the same reasons why I am not, and will never be, a member of The Pigou Club. They transcend, but encompass, the simple statement that institutions matter because they shape incentives and affect outcomes in a very complex and dynamic sense.

Starting with theory, as John Whitehead concisely and capably summarizes, if you make all of the standard neoclassical assumptions there’s no difference between the two policies. Or, as John puts it, “Both emissions taxes and cap-and-trade can achieve optimal emissions at the least cost to society.” They are most similar if the initial permit allocation process is an auction.

Ron Bailey’s argument for the tax, and Tyler’s support of his argument, cite some of the most important theoretical and practical reasons why these two policies would differ when implemented in reality. Ron channels Adam Smith a bit when he says

A carbon tax also offers less opportunity for corruption because it does not create artificial scarcities and monopolies.

There’s a public choice reason why the tax might have fewer implementation costs: the design of a carbon market, the establishment of the cap, the determination of the initial allocation of permits, all become prone to lobbying and political manipulation. One piece of evidence in support of this argument is the recent experience of the EU Carbon Trading Scheme, in which EU leaders caved in to political pressures when allocating the permits among countries and setting out rules for their distribution. Again citing Ron:

Business leaders see the policy handwriting on the wall and are rushing to help shape the emerging greenhouse gas (GHG) emissions regulatory scheme to their own best advantage. A government-created market in emissions permits would be particularly responsive to this kind of gamesmanship. In January, the U.S. Climate Action Partnership, consisting of ten big companies with a total market capitalization of $750 billion, including DuPont, Alcoa, General Electric, and BP America, issued a “blueprint for a mandatory economy-wide, market-driven approach to climate protection.”

Also in January, the Electric Power Supply Association, the lobby group that represents competitive power suppliers that account for 40 percent of the generating capacity of the U.S., acknowledged that “regulatory and legislative processes are moving forward seriously and with speed.” In February, the power-industry lobby group, the Edison Electric Institute, came out in favor of “federal action or legislation to reduce greenhouse gas emissions that…involves all sectors of the economy, and all sources of GHG.”

This is certainly true, and a valid and important concern. However, I don’t understand why setting a tax rate and determining its legal incidence (as opposed to its economic incidence, which reflects how producers pass on the tax in prices based on price elasticity of demand) is substantially less prone to political manipulation than establishing a carbon permit market. Tax supporters who ground their argument in the practical public choice foundation should explain why that’s the case, or else they are committing the Nirvana fallacy.

Another concern, again illustrated by the EU scheme, is the transparency of price signals and their transmission of true opportunity costs to the parties subject to carbon policy. A tax is simpler, clearer, and at least in terms of legal incidence more transparent and less potentially volatile. That creates a less risky environment for parties subject to carbon policy.

I have three critiques of that argument. First, Pigouvian taxes are indirect instruments because they are levied on an output or activity, not on the actual carbon emission. If you are going to tax my output based on an estimate of the amount of carbon its production process creates, why am I going to have any incentive to innovate my production process to make it less carbon-intensive? I’ll still get taxed as if I’m producing the higher amount of carbon. That’s a problematic disincentive to innovation unless we can monitor, measure, and tax the carbon directly. I don’t think the technology’s there yet.

Second, in the quest to reduce the regulatory risk we face a tradeoff here between simplicity of instrument and accuracy of price signal. Maybe the tax is simpler, and maybe cheaper to implement, but how well does it reflect the true opportunity cost facing those subject to carbon policy?

Finally, but most importantly, how do we know the right level at which to set the tax? This is where we get to the Coasian/Hayekian crux of the problems with Pigouvian taxes that the other commenters did not account for in their analyses. The most problematic aspect of Pigouvian taxes is that they rely on the assumption that the policymaker has sufficient knowledge to be able to set the optimal tax. The knowledge of the optimal level of emissions and the optimal tax rate is not, in Hayek’s phrase, given to any one mind. That is the knowledge problem. That means that the policymaker has to have information about production processes, production costs, the epidemiological and other health and consumption effects of the product, and the economic value of those epidemiological and other consumption effects. That is a heroic assumption. No one person or group of people, however well-informed and well-intentioned, can ever determine the optimal emissions or tax through a centralized process. A political process will stumble upon the optimal tax rate by luck; through analysis we can only approximate that rate. If we are wrong and have to revise it, then we violate the above-stated benefit of certainty and transparency. The best mechanism we have for discovering the optimal emissions is through a decentralized process that can aggregate this widespread, diffuse, personal knowledge; this decentralized process is a market.

Note that this diffuse knowledge criticism of Pigouvian taxes can also apply to cap-and-trade: no centralized process of determining the cap on emissions is going to result in the optimal level, except by luck. However, the distortionary effects of an incorrect cap feeding in to a decentralized market process may be small relative to the dynamic efficiency gains that are possible from the innovation incentives that are embedded in decentralized market processes. But here I would note the Bailey caveat from above; political manipulation of the market institutions can undermine many of those dynamic efficiency gains, so minimizing the politicization of the market design is important (and the EU, as usual, provides a cautionary tale in excessive politicization).

This critique, derived from Hayek, also provides a fundamental foundation for the Coasian approach to emissions trading, which involves reducing the transaction costs that prevent private parties from being able to bargain to a mutually beneficial solution. This is precisely the approach underlying the voluntary carbon market (and carbon futures market) at the Chicago Climate Exchange.

The most robust long-term policy approach is one that reduces the (sizeable) transaction costs preventing private parties from achieving mutually beneficial outcomes. Even in something as diffuse as carbon, which is admittedly more challenging than the sulfur dioxide market was in the U.S. in the 1990s, we should not lose sight of the importance of this fundamental objective. Another member of “The Coase Club” who has made a wonderful and thorough argument on these points is Steve Postrel.

Thus while I am sympathetic to the political economy arguments for the lower implementation costs of a carbon tax, the knowledge problem and the right set of institutions for inducing dynamic efficiency through innovation make me favor pursuing a carbon market, designing it well, and testing it carefully.

5 thoughts on “Carbon Tax or Cap & Trade: Does It Matter Which One?”

  1. Let’s try another approach to the question – an approach from first principles.

    Q)What is the objective of the process?
    A)Reducing carbon emissions.

    Q)By what percentage, or by what quantity, or to what quantity are carbon emissions to be reduced; and, over what period of time?
    A)Policy decision, hopefully based on sound scientific analysis of the impact of carbon emissions; ie, government decision.

    Once the policy decision has been made, either:
    1)a cap is set based on the policy decision; or,
    2)a tax rate is established which may or may not result in decisions which reduce the emissions by the desired percentage or amount. The cap always matches the policy decision (unless done the EU way). The tax rate is a SWAG under any circumstances.

    The other issue here is the “rat hole” opportunity. In the case of the cap, those who can reduce emissions cheaply and/or easily will do so; those who can’t will buy allowances from those who can. In a well functioning market, the purchase price of the allowances equals or exceeds the sellers cost of making them available. In the case of the tax, the funds flow to the government, which has a virtual monopoly on all of the rat holes down which the funds can disappear. The US federal government’s track record with regard to rat hole management is questionable at best. Therefore there is no guarantee (and little hope) that the funds collected by the tax will be used to develop or implement technologies for the reduction of carbon emissions.

    Therefore, I vote “cap & trade”.

  2. If you are going to tax my output based on an estimate of the amount of carbon its production process creates, why am I going to have any incentive to innovate my production process to make it less carbon-intensive? I’ll still get taxed as if I’m producing the higher amount of carbon. That’s a problematic disincentive to innovation unless we can monitor, measure, and tax the carbon directly. I don’t think the technology’s there yet.

    Sorry, why is the monitoring problem not equally an issue for cap-and-trade? You have to monitor emissions to figure out how many permits someone needs to buy…

    Note that this diffuse knowledge criticism of Pigouvian taxes can also apply to cap-and-trade: no centralized process of determining the cap on emissions is going to result in the optimal level, except by luck. However, the distortionary effects of an incorrect cap feeding in to a decentralized market process may be small relative to the dynamic efficiency gains that are possible from the innovation incentives that are embedded in decentralized market processes.

    Again from above: I don’t believe you’ve established there are more innovation incentives under cap-and-trade. Your argument seems to be that monitoring is easier under cap-and-trade, but I don’t see why.

    …expected to favor a carbon market over a carbon tax, even if such a market involves government determination of the number of emission permits allowed in such a market.

    Interesting. Are you aware of alternate ideas to use a decentralized process to determine the optimal quantity of permits?

  3. Hmm. Start from the other end.
    In a Pigou tax we are not trying to reduce emissions. We are trying to make sure that the externalities are incorporated into the price system.
    Thus if we know the cost of the externalities that’s what the tax should be. OK, we still meet the Knowledge Problem but it’s a more tractable version of it.
    Say we actually believe the Stern Review’s $ 85 a tonne CO2. That’s what the tax should be (say a $1 on a gallon of gas). Whether it reduces emissions or not is actually an irrelevance. Having included all costs (yes, even of Bangladesh both drowning and boiling at same time) into the price mechanism the result will, in aggregate, be socially optimal.
    Thus a tax, not cap and trade.

  4. How does one invoke Hayek in one breath and the speak of “designing” a market in the next while keeping a straight face?

  5. I was disappointed to see questionable reasoning like this in the article:

    Pigouvian taxes are indirect instruments because they are levied on an output or activity, not on the actual carbon emission. If you are going to tax my output based on an estimate of the amount of carbon its production process creates, why am I going to have any incentive to innovate my production process to make it less carbon-intensive?

    Why would such flawed, indirect methods be required at all?  Fossil carbon comes from just a few operations (extraction or import of coal, oil or gas, cutting of old-growth forest or draining/burning of peat under some definitions).  All that needs to be done is to apply the tax when the stuff comes out of the mine or well; all downstream processes see the cost in the price of inputs, and can minimize it accordingly.

    Paying for carbon sequestration is a tougher issue.  It would require measuring and assaying the material to be sequestered.  But again, this isn’t something that affects everyone.  If we’re going to pay farmers for plowing charcoal into the soil (sequestering carbon in the form of “terra preta”), someone is going to have to verify how much was tilled in and its carbon content.  Ditto the pumping of CO2 down injection wells.  But this only affects the very edges of the system, where carbon goes in or out (carbon losses to the atmosphere are taken care of automatically, like an unclaimed bottle deposit).  Tax rates are also fairly easy to verify and equalize across a treaty zone.

    “Cap and trade”, in contrast, has myriad avenues for shenanigans.  This is probably why the politicos like it so much.  The rest of us should dislike it for the same reasons.

    Unlike per-nation or per-industry carbon caps, a levelized carbon tax does not give an advantage to particular actors on the basis of political dealing.  Production will tend to shift to where it emits the least carbon, and the carbon emitted in transport will appear on the shipping bill.  The amount of the tax is indeed a political matter, but so is the allocation of carbon allowances under caps.  It will be far easier to agree to one number than individual allowances, as the collapse of the European allowance-trading market shows.

Comments are closed.