Cap-and-Trade vs. Emission Tax – Differences
August 15th, 2007 <-- by Joseph Aldy -->Building on my previous posts in this series providing an overview of cap-and-trade and emission taxes and a discussion of their similarities, this post illustrates some of the differences between cap-and-trade and an emission tax: the trade-off between cost certainty and emissions certainty, incentives for R&D, and revenue generation.
Cost certainty versus emissions certainty. In an uncertain world, it is impossible to design a policy that simultaneously guarantees an emissions outcome at a certain cost. An emission tax provides cost certainty – the incremental increase in energy prices is transparent and fixed under a tax – while cap-and-trade provides emissions certainty by capping aggregate emissions. Under a tax, emissions in aggregate can vary depending on the realized costs of abatement (that cannot be predicted ex ante with certainty), economic growth, relative changes in energy prices unrelated to a carbon policy, etc. These factors can likewise drive the variation in costs under a cap-and-trade program.
Should we prefer cost or emissions certainty? It depends on whether “making a mistake” in setting a cap or a tax substantially changes the costs or the benefits more. Suppose that in 2007 everyone expects a $25/ton CO2 tax to cause emissions to fall to 2005 levels by 2015, and likewise that a 2005 cap in 2015 would have a $25/ton CO2 allowance price. Now suppose that in 2015, abatement costs are unexpectedly high. Under the $25/ton tax, emissions would be higher than expected but the incremental cost to energy would reflect the $25/ton tax. Under the 2005 emissions cap, allowance and energy prices would be higher than expected but emissions would reflect the cap. Under this unexpected cost shock, if the cost to society of paying higher allowance prices under the cap-and-trade program exceeds the cost to society of foregoing some emissions abatement under the $25/ton tax, then the tax would be preferred. If instead we expect climate damages to be very sensitive to the level of 2015 emissions, then a cap-and-trade program would be preferred. Most economic research, because of the stock nature of climate change, tends to find that a tax is preferred to a cap.
A variety of cap-and-trade markets have experienced substantial price volatility in recent years. EU Emissions Trading Scheme CO2 allowance prices varied between about 7 euros and 30 euros per ton CO2 in its first 16 months. In April 2006, allowance prices fell by nearly 20 euros per ton in one week, and then down to less than a euro per ton today. The U.S. sulfur dioxide tradable allowance market has been nearly as volatile as the oil market, and much more volatile than the stock market. The southern California RECLAIM cap-and-trade program witnessed a more than ten-fold increase in allowance prices in 2000 during the California electricity crisis.
There are important caveats to these claims of cost and emissions certainty under a tax and cap-and-trade, respectively. A tax does not provide certainty about the aggregate cost to the economy of the climate policy; it only provides certainty about the incremental cost of consuming carbon-based energy. Cap-and-trade may not deliver certain environmental benefits. Complying with an emissions cap may cause emissions to increase outside of the cap – so-called emissions leakage – either to uncovered sectors of the U.S. economy (more on coverage in my next post) or to countries without carbon constraints. This leakage would offset the benefits of limiting emissions under cap-and-trade. Moreover, targets are not always met because of non-participation (the United States under the Kyoto Protocol) or non-compliance (the southern California RECLAIM program).
Incentives for R&D. Price volatility can create adverse incentives for firms to invest in R&D. Firms can form more precise expectations about energy prices under a carbon tax than a cap-and-trade program. An extensive economic literature on investment shows how uncertainty in returns can discourage investment. While some have argued that the possibility of higher than expected allowance prices creates the incentive for more R&D, the possibility of lower than expected allowance prices (and thus lower returns to R&D on abatement technology) will offset this effect. Suppose that the policies described above of a $25/ton CO2 tax (yielding 2005 emissions in expectation) and a 2005 cap (yielding a $25/ton allowance price in expectation) are considered again. The $25/ton carbon tax with zero probability of higher or lower costs will induce more R&D than the $25/ton allowance price with a non-zero probability of higher or lower allowance prices.
Revenue generation. As noted in my last post, an emission tax and a 100% auction cap-and-trade program are similar because they can both raise a lot of revenues that can be used to reduce distorting taxes on income, labor, and capital. Most Congressional proposals for cap-and-trade, however, include much less than 100% auction. Giving away emission allowances for free reflects the creation of an asset with a value equal to tens to hundreds of billions of dollars annually and handing it over to the largest emitters of greenhouse gases. It is not the polluter pays principle, but a pay the polluter principle. Large CO2-emitting corporations in Europe saw their stock prices rise under the EU ETS because of the substantial value of the assets – the emission allowances – they received. Only a modest fraction of allowances need to be given away to energy-intensive firms – on the order of 10-15% – to leave them financially equivalent under a cap-and-trade program than under no regulation.
My next post in this series will address cap-and-trade and tax design issues.
August 16th, 2007 at 3:06 pm
Great series. One significant difference you failed to touch upon is that a cap-and-trade system is at its core closed economically, once the initial credits are allocated, whereas a carbon tax doesn’t indicate where the revenues go.
When judging cap-and-trade systems and carbon taxes on their merits, it’s critical to ask where the revenues go to be able to judge their effectiveness. A cap-and-trade system that hands out allocations to particular sectors subsidizes their pollution and minimizes the incentive to reduce emissions and puts more efficient players at an economic disadvantage.
Similarly a carbon tax’s emissions reducing impact could be wholly undermined if revenues go directly into subsidies for the targeted industries.
August 18th, 2007 at 12:07 pm
Some environmental groups prefer cap-and-trade over emission fees because fixing the quantity of emissions ensures a specific environmental outcome. I think it is important to realize, however, that a cap-and-trade approach sets both the upper bound on emissions and also a floor for them. An emission fee sets no limit on emissions but also no limit on how much we’ll reduce emissions. The fee approach would bring more environmental benefit (bring the quantity of emissions below the cap-and-trade floor) if the fee is higher than the market determined permit price that would result from a cap-and-trade system.
August 18th, 2007 at 7:00 pm
hi nice post, i enjoyed it
August 28th, 2007 at 8:11 am
Thank you for these posts. They are helping me understand the nuances of the cap and trade, cap and auction and emission tax policy options for pollution control. I made a distinction between cap and trade and cap and auction because I still have trouble understanding the preference for cap and auction (when all the allowances are auctioned rather than allocated based on historical emissions).
Here is the part I don’t understand. For pollutants like SO2, NOX and Hg there are commercially available, proven control technologies to reduce emissions from existing sources. In addition to retrofit control equipment, both SO2 and mercury can be controlled by using different fuel sources that have lower amounts of sulfur and mercury. NOX can also be controlled by modifying the way the fuel is burned. It is also possible to reduce SO2, NOX and mercury by changing the type of fuel used. If the market is efficient then the cost of allowances will equal the abatement costs necessary to reduce emissions to the level of the cap. It will take investment money to implement those controls. So if those allowances are auctioned, won’t that just add to the cost of the program because in the full auction the affected firms have to buy the allowances to participate in the market?
The key point for CO2 is that it is different from those pollutants. There is no commercially available retrofit control technology for CO2, differences in the amount of carbon in any particular fuel type are negligible, and aside from trying to burn a fuel more efficiently the way fuel is burned does not affect CO2 emissions. Therefore, the only way that CO2 can be reduced from an existing source is to run less, run more efficiently (good only for small reductions) or convert to a different type of fuel. In my mind that makes an auction for CO2 a viable alternative.
However I still have a problem with the 100% auction. Your posts made the point that it is important that all sources face the same marginal cost of abatement. The emissions of CO2 per MW produced at a coal-fired plant are approximately double the emissions at a gas-fired electric generating plant. If all the allowances are auctioned, then doesn’t that mean that firms with coal-fired sources will face higher costs and as a result the total cost of the program will be higher?
September 4th, 2007 at 5:20 pm
This comment raises two important points. First, the energy costs consumers pay will not differ between cap-and-trade with 100% auction and cap-and-trade with 100% gratis allocation. At first, this may appear to be counterintuitive. Why should consumers pay any more for energy when big energy producers receive their allowances for free? Consider an analogy. Suppose that you participated in a charity raffle, and in doing so you win a new car. But you don’t need a new car – you just bought one. Since you don’t need the car, you decide to sell it. Do you offer to sell it in the used car market for a price of $0 because you did not pay anything for it? No, you would sell it at the market price for that make and model. Suppose that just before you sell your raffle car, an asteroid falls on your bought car and you find out that your insurance does not cover interstellar object damage. Now you use the raffle car, but there is an opportunity cost to using the raffle car – the income you would have earned if you decided to sell it instead.
This is the same for a utility that receives emission allowances for free. If the utility does not need all of the allowances, then it can sell them. If it doesn’t sell all of them, then there are still foregone economic opportunities (the possibility of selling allowances) that should be accounted for in the price of electricity. The price of electricity will reflect the cost of the last unit of labor, and fuel, and capital, and emissions abatement (which should equal the market price of permits in a well-functioning market) if utilities are setting prices competitively. (It is potentially much more complicated in states that have not deregulated their electricity markets and state public utility commissions still determine rates.)
The total costs to the economy and to consumers will be the same under a 100% auction and a 100% gratis allocation. The DISTRIBUTION of the costs, will differ, however, between these two allocations. A 100% gratis allocation will more than compensate the stockholders of regulated firms – this is what happened for some EU utilities that saw their stock prices increase as EU ETS allowance prices rose. Wealthy stockholders would benefit from such a policy. A 100% auction, especially if revenues are used to lower payroll taxes, would be much less regressive and could be designed to be distributionally-neutral.
Second, minimizing the costs of achieving a given emissions goal requires that all sources pay the same amount for emitting a ton of carbon dioxide. The coal-fired power plant should pay more to generate a megawatt-hour of electricity because it emits more carbon dioxide than a natural gas plant. Since it emits more, it has to hold onto (or buy) more emission allowances, or pay more in emission taxes. If a natural gas plant and a coal plant had to expend the same amount per megawatt-hour to cover their CO2 emissions, then utilities would never face the incentive to switch from coal to lower or zero-carbon emitting power sources. The key thing is to make sure that all sources pay the same amount per unit of pollution (e.g., per ton of CO2), not per unit of economic production.
September 5th, 2007 at 8:52 am
Thank you for the response. I am less confused but still don’t understand everything.
You explained the second point so I can understand it but I am still confused about the first point. My confusion is rooted in the difference between SO2, NOX and Hg controls and CO2 controls. Are the total costs to the economy for cap and trade vs. cap and auction the same for all pollutants? In a cap and trade program emissions have to be reduced and for SO2, NOX and Hg those reductions incur costs. As you say, the price of electricity will reflect the cost of those emissions abatement programs (which should equal the market price of permits in a well-functioning market) if utilities are setting prices competitively. If they have to buy the allowances in an auction and incur the abatement program costs doesn’t that add the auction cost to the abatement cost so the total cost to the ratepayer is higher?
I also have trouble with your analogy. I think there is a flaw that makes this statement incorrect: “If it doesn’t sell all of them, then there are still foregone economic opportunities (the possibility of selling allowances) that should be accounted for in the price of electricity.” Allowances represent a compliance obligation and the regulations impose penalties greater than the economic opportunities of selling them. So what foregone economic opportunities are there on allowances needed for compliance?