July-September 2000 Issue


Costs of the Kyoto Protocol: A New Assessment

n Energy Laboratory analysis suggests that enforcing the Kyoto Protocol may not be as expensive as most people think--as long as countries follow the intended guidelines. Most analyses consider the cost of reducing only carbon dioxide (CO2) emissions from fossil fuel use. But the protocol also recognizes reductions in other greenhouse gases (GHGs). According to the MIT analysis, taking advantage of opportunities to reduce those gases as well as CO2 can cut the cost of compliance in half. The protocol puts all gases on a common basis by using "global warming potentials" (GWPs) to weight the value of any emission reduction according to the gas's warming ability and lifetime in the atmosphere. In theory, then, meeting the Kyoto targets by cutting only CO2 or by cutting a combination of gases weighted using their GWPs should give the same climate impacts. To test the validity of the GWPs, the researchers used MIT's integrated atmospheric chemistry, climate, and ecosystem model to simulate the effects of the CO2-only and the multi-gas strategies for meeting the Kyoto requirements. The predicted global warming was similar for the two strategies. However, substantial differences appeared when the two strategies were used in response to a hypothetical policy involving deeper emissions cuts and participation by developing countries--assumptions that increase the role played by non-CO2 emissions. The GWPs are so flawed that such a stringent policy may or may not give the intended climate results, depending on which types of emissions are cut. The MIT team is now rethinking how to value emissions reductions, including considerations such as the timing of the avoided damage and the importance of protecting future generations from long-lived GHGs put into the atmosphere now.

In 1997, more than 80 countries signed the Kyoto Protocol on Climate Change. In it, the developed countries including Russia and the European economies in transition agreed to cut their emissions of GHGs to 5% below 1990 levels by the 2008-2012 period. Few countries have yet ratified the agreement, and it may not go into effect as anticipated. The basic targets and timetables laid out in the agreement still provide a useful set of assumptions for understanding the costs of climate policy. One reason for the reluctance to ratify Kyoto is that projections of the high cost of implementing it abound. For the most part, cost analyses have focused exclusively on reductions of CO2 emissions from fossil fuel use--a familiar area due to the intensive energy modeling and analysis of the 1970s. But the protocol actually addresses emissions not only of CO2 but also of other less-familiar GHGs--methane, nitrous oxide, perfluorocarbons, hydrofluoro-carbons, and sulfur hexafluoride--which come from such diverse sources as landfills, aluminum production, livestock, and electrical switchgear. Also recognized are measures to increase CO2 "sinks," which include forestation and other land-use changes that will increase the amount of CO2 absorbed from the atmosphere. Researchers are generally aware that their analyses should include the other GHGs and forest sinks, but most have believed that the impact on Kyoto's estimated cost would be small.

To check that assumption, Dr. John M. Reilly, Professor Ronald G. Prinn, and their colleagues at MIT and at the Marine Biological Laboratory at Woods Hole have made the first comprehensive assessment of the consequences of the multi-gas approach to implementing the Kyoto Protocol. They determined the cost of meeting the protocol's requirements in two ways: first by reducing only CO2 and then by reducing a mix of all six GHGs in the least costly way. (All analyses of multi-gas reductions include forest sinks as one component.) They then examined the impacts of those strategies on future global climate.

To perform their cost assessment, the researchers used their Emissions Prediction and Policy Assessment (EPPA) model, which analyzes economic activities and associated emissions in terms of sectors, technologies, and regions. They present their results in "marginal abatement curves" that show the cost of an additional one-ton reduction in emissions at various levels of abatement.

Their findings for the United States appear in the figure below. The higher curve shows the cost in 1995 dollars of eliminating a ton of carbon emissions if only CO2 emissions from fossil fuel use are considered. (One ton of carbon corresponds to 3.67 tons of CO2.) The lower curve shows the same information for emissions of all the Kyoto GHGs, each adjusted according to its contribution to atmospheric warming relative to CO2. As expected, in both curves the cost of eliminating a ton of carbon emissions increases as more reduction occurs. (One performs the least expensive reductions first.) Including more GHGs expands the available set of abatement opportunities, so cost at a given level of abatement is lower for the multi-gas strategy.

These marginal abatement curves show the cost to the United States of achieving a one-ton reduction in emissions of carbon as the total reduction in emissions changes. The top curve shows the cost of reductions in CO2 emissions. The bottom curve shows the cost of achieving a carbon equivalent reduction in emissions by cutting CO2 plus five other GHGs and enhancing carbon sinks such as forests--the approach specified in the Kyoto Protocol. If the US cuts its CO2 emissions by the percentage specified in the protocol, the required reduction would be 645 Mtce (marked "RR1") and the marginal cost $258/tce (P1). If instead it cuts its emissions of all six GHGs by that percentage, the required reduction would be greater--724 Mtce (RR2)--but the marginal cost would be less--only $176/tce (P2). But if the US meets the intended multi-gas target of 724 Mtce by reducing only CO2 emissions, the marginal cost would be $360/tce (P3). The implications for total annual cost (equal to the areas under the curves) are evident in the table below.

Examining costs at selected levels of abatement provides some interesting insights. The protocol calls for the United States to reduce its GHG emissions to 7% below its 1990 level by 2008-2012. If the US had to reduce CO2 by that percentage by 2010 (the selected ending date), the required reduction would be 645 megaton carbon equivalent (Mtce), marked on the figure as "RR1." (See also the summary table below.) At that level, the marginal cost of CO2 reduction (the cost of preventing the last ton of carbon from escaping) would be $258/tce (P1). The total annual abatement cost (measured by the area below the curve) would be $61 billion.

This table shows Energy Laboratory estimates of emissions reduction targets, marginal costs, and total annual costs for the United States in 2010 of fulfilling its requirements under the Kyoto Protocol. Three possible strategies are considered: setting a reduction target at 7% below 1990 levels based on carbon dioxide (CO2) emissions only; setting the 7% reduction target using Kyoto's "multi-gas" approach (a mix of greenhouse gases plus forest sinks) and achieving the target by reducing CO2 emissions only; and setting and achieving the target based on the multi-gas approach. (The estimates are derived from the curves in the figure shown above.) The researchers conclude that ignoring opportunities to reduce non-CO2 gases and to enhance forest sinks would double the total annual cost of implementing the Kyoto Protocol.

In contrast, if the United States had to reduce its 1990 emissions of all six Kyoto gases by 7%, the required reduction would be 724 Mtce (RR2). However, because of the added abatement opportunities, the marginal cost would be only $176/tce (P2) and the total annual cost only $43 billion. Thus, including a variety of GHGs provides more environmental benefit for less investment.

Looked at another way, the figure carries a clear warning. If the United States tried to achieve the Kyoto-specified abatement level of 724 Mtce by cutting only CO2 from fossil fuel use, the marginal cost of eliminating those CO2 emissions would be $360/tce (P3), and the total annual abatement cost would be $86 billion. Thus, the nation would pay twice as much to meet its Kyoto targets if it ignored opportunities to control emissions of all the named GHGs.

Using their EPPA model, the researchers also looked at emissions-reduction costs for the "Annex B" countries (those subject to emissions limits, namely, the members of the Organization for Economic Cooperation and Development as of 1990, plus Eastern Europe and most of the former Soviet Union). The calculated costs vary from country to country, but in every case they are minimized by pursuing multi-gas reductions. For Annex B as a whole, the cost of meeting the protocol's targets is about $174 billion if only CO2 abatement is undertaken and about half that much if emissions of all gases are cut and forest sinks are used. Thus, low-cost opportunities to cut emissions other than CO2 are more abundant--and more critical--than most experts have believed.

What about the climatic impacts of the emissions-control strategies? Reductions in emissions of the various gases cannot be compared on a ton-for-ton basis because the gases have varying abilities to capture heat and varying lifetimes in the atmosphere. The protocol therefore includes a set of indexes--the GWPs--that reflect each gas's potential for global warming over a set period of time relative to CO2. Adjustments based on those GWPs are intended to put reductions of all gases on a common basis. If the GWPs laid out in the protocol do their job well, the impact on climate will be identical whether the Kyoto target is met by reducing CO2 or a mix of gases. But if, for example, the GWPs presented in the protocol overvalue cuts in a given gas, countries may meet their reduction targets without eliciting the expected climatic effect.

To test the accuracy of the GWPs, the researchers used their "Integrated Global System Model" (IGSM) to estimate the consequences for the atmosphere, climate, and ecosystems of using different control strategies. Developed over many years by an interdisciplinary research team, the IGSM includes components that simulate economic growth and associated emissions (the EPPA model); complex feedbacks among ecosystems, atmosphere, and oceans; chemical interactions among gases in the atmosphere; the impact of climate change itself on various processes; the roles of carbon monoxide and nitrogen oxides; and the cooling effect of aerosols.

The first analysis focused on the Kyoto commitments for all of the Annex B countries, extended unchanged to the end of 2100. The researchers calculated the climate and ecosystem effects, first assuming only CO2 emissions are controlled and then assuming multi-gas controls. Both cases concluded that temperatures would increase by about 2.4°C between 1990 and 2100. The predictions are consistent, so the GWPs appear to be effective in calculating trade-offs among emissions in the various gases.

However, climate change experts generally agree that more dramatic steps are needed to stabilize atmospheric concentrations of GHGs in the future. Therefore, the MIT team devised a hypothetical policy requiring developed countries to achieve much larger cuts and developing countries to start a comparable program of cuts beginning in 2025. Meeting that more-stringent policy with a CO2-only strategy produced an estimated increase in temperature between 1990 and 2100 of 1.1°C. Assuming a multi-gas strategy yielded a temperature increase of only 0.5°C. Thus, the projected climatic impacts of the two strategies differ substantially when the cuts are deeper and the developing countries are involved--changes that increase the importance of the role played by gases other than CO2. Again, if the GWPs in the protocol are accurate, the projected temperature increases would be the same. Instead, they are different, and the difference is large enough to have important policy consequences: if policymakers take serious steps toward stabilizing the climate, the GWPs in the Kyoto Protocol will not provide a sound basis on which to set targets, credit emissions reductions, and value the emissions permits that, under the protocol, could be traded internationally.

How can the GWPs be improved? To clarify the issues and trade-offs involved, the MIT researchers used the IGSM to calculate GWPs under a variety of assumptions. Not surprisingly, assuming different time horizons gives very different outcomes. The GWPs in the protocol are based on the accumulated climatic impacts of a unit of an emitted GHG over the next 100 years. That time period is fine for some gases. Methane, for example, lasts only about a decade. But some gases last for many thousands of years, and their impacts beyond 100 years are lost. Allowing for interactions among gases in the atmosphere also changes the GWPs. IGSM analyses show that the climatic impact of a given GHG is directly influenced by concentrations of other gases. Indeed, large errors result from ignoring the interactive and climatic effects of non-Kyoto gases such as carbon monoxide, nitrogen oxides, and sulfur dioxide. And because the concentrations of other important gases are not constant over time, the GWP for a given GHG is likely to change over time.

Other factors must also be considered in deciding how to value cuts in various emissions. Controlling short-lived gases will yield climatic benefits sooner than controlling long-lived ones will. Cuts in short-lived GHGs are most valuable if we believe that the dangers from climate change are imminent. But if the danger of global warming is not immediate, cutting long-lived GHGs is a higher priority. Eliminating gases that last for thousands of years from today's emissions provides a benefit essentially forever, whereas today's emissions of short-lived gases like methane will be gone from the atmosphere in a few decades. Many people also place a high value on protecting future generations from harm, regardless of the source; and this consideration favors controlling the very long-lived gases. Finally, reductions in some emissions may bring "ancillary" benefits. In particular, the MIT analyses show that reducing emissions of certain GHGs also reduces emissions of critical urban air pollutants. (Cutting down fossil fuel use eliminates all sorts of emissions.) The researchers thus stress the need to consider global, regional, and local environmental concerns simultaneously when calculating the cost and effectiveness of any type of emissions-control strategy.

The researchers' analyses reveal serious shortcomings in current climate change policy--both in its design and in its likely implementation. Making improvements will require substantial advances in both scientific and economic understanding. Critical needs are better modeling capabilities; credible inventories of emissions of all GHGs, especially in the developing world; and a sound basis on which to extend analyses beyond 2010. In the near term, the MIT researchers recommend that the scientific community develop the best possible GWPs for the short-lived GHGs and that mechanisms be established that encourage countries to reduce or trade emissions of those gases to meet their Kyoto targets. However, they question whether gases that last for thousands of years should be part of such a trading system because of the difficulty of accurately comparing them to short-lived gases. An alternative policy would focus on simply reducing emissions of those long-lived gases to the lowest possible level.

John M. Reilly is associate director for research at the MIT Joint Program on the Science and Policy of Global Change. Ronald G. Prinn is the TEPCO Professor of Atmospheric Chemistry, head of the Department of Earth, Atmospheric, and Planetary Sciences, and co-director of the Joint Program. This research was funded by the MIT Joint Program on the Science and Policy of Global Change. Further information can be found in references.