In my last article, I explained that two years ago the Manhattan Institute’s Senior Fellow Oren Cass wrote a masterful critique of the typical arguments for a U.S. carbon tax. His essay, “The Carbon Tax Shell Game,” is so good that I decided to spend two posts here at IER amplifying some of his strongest points. As I’ve been illustrating over the years with my own work (e.g., here and here), the case for a carbon tax falls apart once you start picking at it.

In the previous post, I focused on Cass’s claim that the carbon tax in U.S. politics is a “shell game,” because its proponents promise contradictory things to different groups. In this post, I’ll focus on Cass’s sophisticated critique of superficial justifications for a U.S. carbon tax based on the concept of a “negative externality.”

Assumptions About the “Discount Rate” Drive the Analysis

Cass reviews the standard economic argument for a carbon tax: Because greenhouse gas emissions could cause significant damage in the form of climate change, they constitute “negative externalities.” In the tradition of economist A.C. Pigou, a textbook response to negative externalities is a so-called Pigovian tax, in order to make actors in the marketplace—both businesses and consumers—take into account the full social cost of their activities.

Although this is very straightforward in theory, when we try to implement a carbon tax in practice things get much more dubious. (I have written exhaustively on these issues: see here and here for example.) There are many stumbling blocks in the path of acknowledging the possibility of manmade climate change damage, and setting an “optimal” carbon tax.

In this post, I’ll just review one of the issues, which we at IER have covered extensively: namely, the discount rate used to convert future damages into present dollars.

Because so much of the damage (which is simulated in computer models) from greenhouse gas emissions doesn’t occur until decades or even centuries in the future—the Obama administration’s team ran their computer simulations through the year 2300—it matters a very great deal how we translate, say, $100 billion worth of damage from climate change in the year 2210, into 2017 dollars. After all, if we set a carbon tax at $40 per ton today, that will cause people today to rearrange their activities to reduce emissions. On standard Pigovian terms, this $40 per ton is optimal only if the marginal ton of CO2 that would otherwise be emitted, would cause a total of $40 worth of damage—measured in 2017 dollars—over the next few centuries. In order to calculate that, we have to know how to compare 2017 dollars with (say) 2210 dollars.

(For an analogy, if we wanted to set a carbon tax in euros, even though our damage estimates were all expressed in dollars, then we would obviously need to know the exchange rate between euros and dollars. Likewise, since we are setting a carbon tax in today’s dollars, while our damage estimates in the computer simulations occur in various years through 2300, we need an “exchange rate” or “discount rate” to make them the same units.)

In case the reader wonders why I’ve spent so many paragraphs hammering home this point, it’s because the discount rate we plug into the formula drives the entire result. In other words, we don’t need to have a big argument about how sensitive the earth’s temperature is to a doubling of CO2 concentrations, and we don’t need to have a big argument over how much damage to human welfare a given amount of warming will cause. Even if we stipulate all of those conditions for the sake of argument, nonetheless I can derive an “optimal” carbon tax either close to $0, or on the other hand above $100 per ton, if you just let me adjust the dial of the discount rate.

Here’s how Oren Cass expressed the situation:

Consider the federal government’s official effort to develop [a social cost of carbon] assessment. The marginal cost of one ton of CO2, even after averaging out the range of hypothetical climate behaviors, varies from $0 to $129. (Marginal-cost estimates for the full range of modeling between the fifth and 95th percentiles vary from -$12 to $515.)

The choice of discount rate, meaning the relative importance of future costs versus current costs, overwhelms all other model attributes. Outputs generated with a 3% discount rate (the heaviest weighting typically used in regulatory analyses) are generally at least three times higher than comparable outputs for a 5% discount rate (a moderate weighting); one could more closely approximate the costs implied by a 5% rate with a tax of $0 than with a tax derived using the 3% rate. An assumption about how society values costs 100 years from now swings the result by more than an assumption about whether climate change exists at all. [Bold added.]

This is one of the main reasons we at IER have argued that the “social cost of carbon” is an inappropriate tool for regulatory analysis. The public and even policymakers are led to believe that there are scientists in white lab coats using fancy equipment in order to go out and measure “the social cost of carbon,” the way they might measure the charge on an electron or estimate the temperature of the sun’s surface.

Yet in contrast, an economic or even philosophical decision about how to treat the welfare of our great-great-grandchildren, relative to our own welfare, is what really drives the so-called “social cost of carbon,” and hence the size of an “optimal carbon tax” if we base it on such as estimate.

A Subtle Flaw With the “Double Dividend” Logic

Cass makes many sophisticated observations about the typical case for a carbon tax, and I would encourage even professional economists to read his essay. He makes many points that economists too often miss in their glib endorsement of “taxing bads, not goods” in the climate change policy debate.

For now, let me end by highlighting just one of Cass’s points. This is something I have not seen anyone else bring up. In context, Cass is addressing the claim that if we refund the revenues from a carbon tax in order to reduce other distortionary taxes, that we can (a) reduce emissions and hence reap an environmental benefit but also (b) boost the conventional economy, because those pre-existing taxes are so destructive of economic growth.

But there’s something missing from this typical analysis. It’s obvious once you think about it, but I confess I hadn’t even thought of this wrinkle until Cass brought it up—and I’ve spent years thinking about the problems with carbon taxes! Here he explains his insight:

Nor does describing a carbon tax as “revenue neutral” do anything to improve its appeal. Promising to use the revenue for tax cuts or a rebate does not guarantee its best use or a net positive economic impact, nor does it make the policy somehow free. To the contrary, a revenue-neutral tax is guaranteed to be costly precisely because it holds government revenue constant while also increasing costs to private actors by driving them toward higher-cost energy technologies. The effect is most obvious in a world where the tax has driven emissions to zero, and government revenue comes from all of its pre-tax sources, except consumers also find themselves motivated by the tax’s existence to pay the full cost of electric vehicles and solar panels. In this respect, the tax operates much like the minimum wage; it imposes large and plainly government-created costs in the form of “off-budget” spending for which the government is never held accountable. [Bold added.]

To repeat, the part I’ve put in bold is an amazing insight from Cass, which I haven’t seen discussed elsewhere in this entire debate.

To reiterate, Cass is here pushing back against the simplistic claims made by some—especially those pitching a carbon tax to conservatives and libertarians—that so long as the revenues are used to cut taxes on capital and labor, then the actual pain of a carbon tax will be minimized or even flipped negative. “After all,” so the reasoning goes, “we want to discourage people from emitting carbon dioxide, and we want to encourage them to work and save, so how could it possibly be a bad thing to levy a tax on emissions in order to fund cuts in the payroll tax and corporate income tax?”

Yet the block quotation from Cass above shows that this glib reasoning is leaving something out. Suppose that the U.S. government implemented an outrageously high carbon tax—something like $2,000 per ton—and enforced it vigorously. The price of conventional gasoline would skyrocket some $16 per gallon, while electricity prices would soar because coal- and natural gas-fired power plants would suddenly have outrageous taxes to pay.

In this regime, the amount of U.S. carbon dioxide emissions would fall drastically, so that even with the very high rate of carbon tax, it’s possible that within a decade very little revenue would be coming in. (Remember, total carbon tax receipts per year are annual tons of emissions multiplied by carbon tax per ton.) In this scenario, the tax rates on labor and capital would have to be basically what they were before the new, draconian carbon tax, because of the assumption of “revenue neutrality.” In other words, if the draconian carbon tax isn’t bringing in much revenue since the carbon base gets driven to basically zero, then there isn’t much in the kitty to offset the pre-existing taxes.

So what can we say about the state of the conventional economy? It clearly isn’t benefiting from any “pro-growth” kick emanating from “Pigovian tax reform.” No, in this extreme scenario, the pre-existing distortionary taxes haven’t been cut at all.

However, the situation isn’t simply a wash. Even though it’s not bringing in new revenue, the massive $2,000 per ton carbon tax is definitely forcing Americans to alter their behavior. Nobody would be driving gasoline-powered vehicles, and all coal- and natural gas-fired power plants would be shuttered. Americans’ standard of living would have collapsed, as transportation and energy had become outrageously expensive.

Now to be sure, there would be a drop in estimated future damages from human-caused climate change, since (by assumption) U.S. emissions would quickly fall to near zero. But the point is, the typical carbon tax advocates have been telling Americans that they could get a “win-win” when carbon tax receipts are devoted to “pro-growth tax reform.” Advocates have been telling Americans that they would get a healthier environment and faster economic growth.

Yet as this exaggerated example illustrates, the advocates need to do more homework. The numbers matter. The higher the carbon tax rate, the more it drives down the size of the carbon tax base. And even as the total amount of carbon tax receipts declines, the presence of an ever-higher carbon tax rate is altering behavior in ways that stifle conventional economic growth.

It’s not enough just to observe, “Greenhouse gas emissions are a negative externality while we want to encourage work and saving.” The numbers matter. Even if “taxing bads, not goods” leads to a “win-win” upfront, it’s possible that the numbers move and cause the flipside to occur in a decade or two.

Let me make sure the reader understands the point of my exaggerated example. I picked a ridiculously high carbon tax of $2,000 per ton to make Cass’s point crystal clear. But even at a more moderate level, we still have to take into account the subtle mechanism at work: The more successful a carbon tax is at inducing people to alter their behavior, then the less revenue it raises. The impact of a carbon tax on the conventional economy is not necessarily directly proportional to the amount of revenue it raises, and so in general we can’t rely on catchy slogans like “tax bads, not goods.” To assess the economic cost of complying with a carbon tax—which could be compared to the ostensible benefits of avoided future climate change damages—we need to look at specifics. A very low carbon tax rate won’t raise much revenue, and so won’t allow for much tax reduction elsewhere, but on the other hand a very high carbon tax rate might not raise much revenue either.

Oren Cass wasn’t picking one particular fiscal projection over another, but he was underscoring the huge leaps of faith in the typical case for a U.S. carbon tax when it comes to discussions of “revenue neutrality.” To repeat, the numbers matter. Anyone trying to sell the idea of a carbon tax just using basic principles and textbook intuition has not really considered the issue in depth.


Oren Cass’s essay on the politics and economics of a carbon tax was first released two years ago, but it is still very relevant for today’s policy debate. Cass picks apart the various (and contradictory) arguments for a carbon tax so beautifully that this is one of the single best pieces in the genre.

Print Friendly, PDF & Email