Chapter 3. Emissions Pricing to Stabilize Global Climate*

Ruud Mooij, Michael Keen, and Ian Parry
Published Date:
September 2012
  • ShareShare
Show Summary Details
Valentina Bosetti Carlo CarraroSergey Paltsev and John Reilly 

Key Messages for Policymakers

  • Without significant emissions mitigation actions, projected “likely” global atmospheric temperature increases by the end of the century are approximately 2.5° C to 6.5° C above preindustrial levels.
  • Although there is much uncertainty, a global carbon tax starting at roughly US$20 in 2020 and rising at 3 to 5 percent per year should be in line with stabilizing atmospheric greenhouse gas (GHG) concentrations at 650 parts per million (ppm) or keeping mean projected warming to about 3.6° C. A starting tax of roughly twice this level would be recommended if the goal is to keep atmospheric GHG concentrations to 550 ppm or mean projected warming below 3° C.
  • However, keeping mean projected warming to 2° C (or stabilizing atmospheric GHG concentrations at current levels of about 450 ppm CO2 equivalent), the goal identified in the Copenhagen Accord (COP 15) and reiterated in the Cancun Agreements (COP 16) is highly ambitious and may be infeasible. Achieving this target would require the future development and wide-scale deployment of (still unproven) technologies that, on net, remove GHGs from the atmosphere. The Copenhagen pledges for 2020 still keep the 2° C target within reach—should these technologies be successfully developed—but highly aggressive actions would be needed immediately after that.
  • Even the 550 ppm target would become technically out of reach if action by all countries is delayed beyond about 2030. And required near-term emissions prices (in developed economies) consistent with this target escalate rapidly with delayed action to control emissions in developing economies. Postponing mitigation actions, especially in emerging countries where large portions of energy capital are being installed for the first time, can be very costly. Extra costs associated with the delayed actions escalate rapidly with the stringency of the target, and some more stringent targets become infeasible if action is postponed.
  • To reduce the cost while achieving an equitable sharing of them, decisions about where emissions reductions are taken and how they are paid for should be separated. Emission mitigation should take place where it is most efficient. Equity considerations can be addressed through agreed upon mechanisms that result in transfers from those better able to pay to those with less ability to bear these costs. Negotiating such a transfer scheme is likely one of the most difficult aspects of reaching an agreement.
  • Innovation, both on energy efficiency and alternative energy sources, is needed. Carbon pricing (e.g., carbon taxes or a price established through a cap-and-trade system) would provide a signal to trigger both innovation and adoption of technologies needed for a low carbon economy.

In this chapter, we discuss projected greenhouse gas (GHG) emissions pricing paths that are potentially consistent with alternative targets for ultimately stabilizing the global climate system at the lowest economic cost and under alternative scenarios for country participation in pricing regimes. The pricing projections come from models that link simplified representations of the global climate system to models of the global economy, with varying degrees of detail on regional energy systems. There is considerable uncertainty surrounding future emissions prices, given that different models make very different assumptions about future emissions growth (in the absence of policy), the cost and availability of emissions-reducing technologies, and so on. Nonetheless, projections from the models still provide policymakers with some broad sense of the appropriate scale of (near-term and more distant) emissions prices that are consistent with alternative climate stabilization scenarios and how much these policies cost.

In the next section we discuss where we might be headed in the absence of mitigation policy, in terms of future GHG emissions trends, and what these imply for the growth of atmospheric GHG concentrations and, ultimately, for the amount of likely warming over this century. We also discuss the benefits of different stabilization targets for atmospheric GHG accumulations in terms of potentially avoiding warming. The chapter then addresses projected emissions pricing, as well as the costs of mitigation policies, to meet stabilization targets in the ideal (but unlikely) event of early and full global cooperation and with efficient pricing across all emissions sources and over time. This is followed by a discussion of the implications of delayed emissions reductions by all countries compared with just developing economies. We briefly evaluate recent emissions reduction pledges by country governments in light of the climate stabilization goals. The following section discusses the distributional burden of mitigation costs across countries and the potential complications for negotiation of long-term climate policy. In the final section, we offer some thoughts on pragmatic policy steps in the near term.

Emissions and Warming Trends


There have been many efforts to project future emissions trends, and the range of projections over the twenty-first century has been wide. GDP and population expansion are major drivers of future emissions growth, although the role of the latter will gradually fade with the projected stabilization of the world population in the second half of the century. Some factors tend to dampen future emissions growth, such as potentially rising fossil fuel prices and improvements in energy efficiency (e.g., cars that can be driven longer distances per unit of fuel or buildings that require less energy to heat them). What differs most across forecasting models—and causes the uncertainty affecting emissions projections—are assumptions concerning future GDP growth; the availability of fossil resources; the pace and direction of technical change, in turn affecting the cost of low-carbon technologies and the energy intensity of the economy; and flexibility of fuel and technology substitution within the energy-economic system. Whether and when governments of high-emitting countries undertake meaningful GHG mitigation measures is an additional uncertainty on top of the various economic forces.

In the absence of (significant) mitigation action, energy-related carbon dioxide emissions (the primary GHG) are projected to increase substantially during the twenty-first century. Figure 3.1 shows the range of projections in a recent model comparison exercise organized by Stanford University’s Energy Modeling Forum (EMF-22), which engaged 10 of the world’s leading integrated assessment models.1 On average, fossil fuel CO2 emissions will grow from about 30 Gt CO2 in 2000 to almost 100 Gt CO2 by 2100.

Figure 3.1.Energy-Related CO2 Emissions Projections over the Twenty-First Century

Source: Authors’ calculations drawing from the EMF-22 dataset.

Note: The figure indicates a range of the median projections from each model used in the EMF-22 study.

The contribution of different regions to global CO2 emissions is more stable across models. The Organization for Economic Cooperation and Development (OECD) countries will contribute 15 to 25 percent to total emissions in 2100 (compared with just under half of global emissions at present). Although the United States continues as one of the main emitters, its projected global emissions share will decrease from 25 percent to 10 percent over the century. Brazil, Russia, India, and China (BRIC) will contribute about 45 to 50 percent of total fossil CO2 emissions by 2050, with at least 25 percent of the total emissions attributed to China alone from 2020 onwards. India accounts for a further 15 percent of global emissions by the mid-century. The rest of the developing world is projected to have an increasing role, moving from 17 to 25 percent of total emissions at present to 25 to 40 percent by 2100.

Anthropogenic CO2 emissions are mostly energy related, with a (small) contribution from industrial processes (mostly cement production) and a (more substantial) contribution from land-use change, although energy-related emissions are projected to grow faster than these other sources of CO2. Destruction of tropical forests and peat lands contributed 25 percent of global CO2 emissions in 2000, mostly from a subset of tropical countries including Brazil, Indonesia, and some countries in central and western Africa.

While CO2 is the major contributor to global warming, other GHGs also play a significant role. In particular, these include five other gases covered by the Kyoto Protocol: methane (CH4), nitrous oxide (N2O), and a group of so-called F-gases (HFCs, PFCs, and SF6).2 Currently, these non-CO2 gases contribute about 25 percent of total annual GHG emissions in CO2-e (i.e., CO2 warming equivalents over their atmospheric life span), although again, these emissions are projected to grow slower than CO2 emissions over the twenty-first century (IPCC, 2007).3

Implications for Future Atmospheric Concentrations and Temperatures

Once absorbed by the atmosphere, some GHGs are largely irreversible—CO2 emissions, for example, reside in the atmosphere for about 100 years.4 Without a significant emissions control policy, atmospheric GHG concentrations are projected to grow rapidly. The EMF-22 scenarios project atmospheric concentrations of 800 to 1,500 parts per million (ppm) CO2-e by 2100 (counting concentrations of the gases identified for control in the Kyoto Protocol). For comparison, concentrations in 2010 were about 440 ppm.5

To date, temperatures are estimated to have risen by approximately 0.75° C relative to preindustrial levels, with most of the warming attributed to atmospheric GHG accumulations as opposed to other factors like urban heat absorption, volcanic activity, and changes in solar radiation (IPCC, 2007). However, the full impact of historical concentrations has yet to be felt due to inertia in the climate system (gradual heat diffusion processes in the oceans slow the adjustment of temperatures to higher GHG concentrations).

According to IPCC (2007), in the absence of a GHG mitigation policy, projected “likely” temperature increases by the end of the century are in the range of 2.4° C to 6.4° C above preindustrial levels (“likely” refers to a 66 percent chance or greater). A recent study at the Massachusetts Institute of Technology (MIT) with updated climate and socioeconomic parameters projected even more warming—a 90 percent chance that temperatures will rise by 3.8° C to 7.0° C by 2100 with a mean projection of 5.2° C (Sokolov and others, 2009).

Yet another recent and especially comprehensive study by Prinn and others (2011) put together findings from intergovernmental panels (represented by the IPCC); national governments (including selected scenarios from the U.S. government Climate Change Science Program [US CCSP]); and industry (represented by Royal Dutch Shell). Prinn and others (2011) estimate global temperature increases of 4.5° C to 7.0° C from current levels by 2100 in the absence of climate policy. There are many risks associated with higher levels of temperature increase, some of which (particularly the risk of abrupt climate change) are poorly understood (see Chapter 4 and IPCC, 2007).

Avoiding Warming under Different Climate Stabilization Targets

Stabilization of GHG concentrations at levels often discussed in international negotiations would require very substantial emissions cuts. As indicated in Figure 3.2, some of the more stringent targets are already exceeded or will be exceeded in the not-so-distant future. In particular, the 450 CO2-e target for the Kyoto Protocol gases (consistent with keeping mean projected warming above preindustrial levels to approximately 2° C) is about to be passed.

Figure 3.2.Relationship between Different CO2-e Concentration Targets and Projected Concentrations in the Absence of Mitigation

Source: EMF-22 (Clarke and others, 2009).

Note: Figures in parentheses indicate mean projected warming (above preindustrial levels) if concentrations are stabilized at particular levels assuming a value of climate sensitivity equal to three.

Although the most stringent concentration targets might be beyond reach, even limited actions to reduce GHGs will appreciably reduce the probability of more extreme temperature increases. For example, according to results reported in Table 3.1, stabilizing GHG concentrations at 660 ppm rather than 790 ppm reduces the risk that warming in 2100 will exceed 4.75° C, going from 25 percent to less than 1 percent.6

Table 3.1.Cumulative Probability of Global Average Surface Warming from Preindustrial Levels to 2100
ΔT > 2°CΔT > 2.75°CΔT > 4.75°CΔT > 6.75°C
No Policy at 1400100%100%85%25%
Stabilize at 900100%100%25%0.25%
Stabilize at 790100%97%7%<0.25%
Stabilize at 66097%80%0.25%<0.25%
Stabilize at 55080%25%<0.25%<0.25%
Source: Adapted from Webster and others (2009).

Note: Results are based on 400 simulations of the MIT’s Integrated Global System Model under different assumptions about future emissions growth and parameters in the model. As the increase in global temperature from preindustrial levels to 2000 was about 0.75° C, probabilities for temperature increases relative to 2000 can be obtained by subtracting 0.75° C from the targets in the top row.

Source: Adapted from Webster and others (2009).

Note: Results are based on 400 simulations of the MIT’s Integrated Global System Model under different assumptions about future emissions growth and parameters in the model. As the increase in global temperature from preindustrial levels to 2000 was about 0.75° C, probabilities for temperature increases relative to 2000 can be obtained by subtracting 0.75° C from the targets in the top row.

But what scale of (near-term and more distant) emissions prices are needed to meet alternative stabilization targets and how much might these pricing policies cost? The answers depend, among other factors, on which countries participate in pricing regimes and the efficiency of the policies used to achieve emissions reductions. We turn to these issues in the next two sections, beginning first with the ideal global policy response with early and full participation in pricing regimes and then with more realistic scenarios with delayed action among all or a subset of countries.

Climate Stabilization with Global Participation of Countries

Here we consider a policy scenario with efficient (i.e., cost-minimizing) pricing of emissions across regions, different gases, and time, and full credibility of future policies in triggering long-term investments. The ideal case is useful to understand the basic dynamics of the system and to have a benchmark for evaluating how far more realistic policy scenarios diverge from the ideal policy.

Achieving global economic efficiency (i.e., reaching a climate stabilization target at the lowest global economic cost) involves pricing emissions at the same rate across different countries. This can be achieved by imposing the same GHG price across the countries through a system of carbon taxes or by allowing a full trade in emissions permits among all countries and all sectors of the economy.

One caveat is that the focus here is on the total global costs of policies and not the cost that might be borne by individual countries. As discussed further below, there are numerous possibilities for sharing the burden of mitigation costs across different countries. Another caveat is that possibilities for reduced deforestation in tropical countries and reforestation of temperate regions are not captured in the model results discussed below, even though they could contribute significantly to mitigation efforts.7

Emissions Prices and Emissions Reductions

Projected emissions prices (in CO2-e for all GHGs and in 2005 U.S. dollars) and CO2 reductions for different concentration targets in the EMF-22 exercise are summarized in Table 3.2, where the ranges include cases that do and do not permit transitory overshooting of the long-term stabilization target.

Table 3.2.Emissions Prices and Reductions under Climate Stabilization Scenarios: Full Global Participation
Atmospheric stabilization target, ppm CO2-eEmissions price in 2020 (2005 US$ per tonne of CO2-e)aPercent change in global CO2 emissions in 2020 relative to 2000Percent change in global CO2 emissions in 2050 relative to 2000
450a, b15-263−67 to 31−13 to –92
550b4-52−4 to 50−67 to 52
6503-2030 to 57−16 to 108
Source: Authors’ elaboration of the EMF-22 dataset.

Only a limited number of the models are able to solve for this case (even with overshooting) as it requires the development and wide-scale deployment of negative emission technologies.

Includes cases both with and without transitory overshooting of the long-term stabilization target. In the 650-ppm case, there is no overshooting.

Source: Authors’ elaboration of the EMF-22 dataset.

Only a limited number of the models are able to solve for this case (even with overshooting) as it requires the development and wide-scale deployment of negative emission technologies.

Includes cases both with and without transitory overshooting of the long-term stabilization target. In the 650-ppm case, there is no overshooting.

The global emissions price in 2020 that would be in line with a 650-ppm CO2-e target ranges between $3 and $20 per metric tonne of CO2-e across the different models. Emissions prices in 2020 are $4 to $52 per tonne under the 550 ppm CO2-e target (or $10 to $52 per tonne if overshooting is not permitted). For the 450 ppm CO2-e target, only two models find a solution when no overshooting is allowed: with overshooting, half of the models are able to find a solution, although the 2020 emissions price is generally quite high at $15 to $263 per tonne.

The reason models are less capable of finding a feasible set of actions for more stringent targets is that concentrations are already very close to 450 ppm CO2-e (Figure 3.2). Stabilizing at 450 ppm CO2-e would require an immediate and almost complete decarbonization of the entire global economy, which is most likely technically infeasible.8 Similarly, going back to the target after overshooting implies deployment on a massive scale of “negative emissions technologies” to remove CO2 from the atmosphere, particularly biomass power generation coupled with carbon capture and storage (BECS). Not all models envision the future deployment of such technologies, which are highly speculative at present.9

For cost-effectiveness over time, the emissions price should rise at (approximately) the discount rate (or rate of interest) to equate the (present value) of incremental abatement costs at different points in time (in emissions trading systems the allowance price would increase over time at this rate if there is perfect substitutability of trading in emissions permits and other financial instruments). Different modeling groups assume different (real) discount rates, usually in the range of 3 to 5 percent, so the carbon price would also increase over time at this rate.

Looking at emission reductions (expressed as percentage changes with respect to 2000 emissions), which need to be in line with the different targets (see the second and third column in Table 3.2) for the near- and medium-term, there is not much difference in appropriate emission reductions for 550 and 650 ppm—but very large emission reductions are required, even in the short term, for the 450 ppm CO2-e scenario.

Policy Costs

Ideally, mitigation costs would be measured by economic welfare losses (see Chapter 1 and Paltsev and others, 2009), although GDP losses are more commonly reported in climate policy models.

EMF-22 reports the net present value of GDP costs (discounted at 5 percent) in the range of $0 to $24 trillion (in 2005 U.S. dollars) for 650 ppm CO2-e stabilization, in the range of $4 to $65 trillion (in 2005 U.S. dollars) for 550 ppm CO2-e stabilization, and $12 to $125 trillion for 450 ppm CO2-e stabilization, considering the full range of scenarios.

Figure 3.3 reports costs for each participating model for the three different stabilization levels and considering different levels of participation (full participation and delayed participation) and paths with and without overshooting of the target. Concentrating here on the full participation cases, costs, expressed as a percent of the present value of world GDP, are approximately 0.1 to 1.5 percent, 0.3 to 2.8 percent, and 2.7 percent for the three concentration targets, respectively.

Figure 3.3.Policy Costs for the EMF-22 Dataset by Model Run

Source: Tavoni and Tol (2010).

Note: Green colors indicate models with biomass generation coupled with carbon capture and storage (BECS) and blue models without BECS. FP = full, immediate participation of developing economies, DP = delayed participation of developing economies, STAB = target not to exceed, and OS = target can be overshot.

US CCSP (Clarke and others, 2007) also reported the cost of climate policy as a percentage reduction in the global GDP, but rather than net present values, they reported the loss in different periods of time. The most stringent stabilization level in this study is roughly equal to 550 ppm CO2-e (450 ppm when only CO2 contributions are considered). The loss of the world GDP in comparison to a scenario with no climate policy is in the range of 1 to 4 percent in 2040 and 1 to 16 percent in 2100.

Emissions pricing will induce emissions reductions in the sectors where these reductions are cheapest. Models have different views about the timing of emissions reduction, but most of the projections agree that the power generation sector will be the first area where less-carbon-emitting (e.g., natural gas) or almost-zero-carbon-emitting technologies (e.g., nuclear, hydro, renewables) are introduced because of various economic substitutes that already exist in this sector.10 Less-emitting technologies in transportation (e.g., gasoline/electric hybrid vehicles, more fuel-efficient conventional vehicles) and energy-saving technologies in buildings and industry are also promising, but they currently look more expensive. Substantial reductions in GHG emissions in agriculture and cement production are also costly, but to achieve climate stabilization, emissions from all sectors of the economy need to be reduced drastically.

For more stringent climate stabilization targets, these reductions obviously need to begin in the near future. Deferring the bulk of mitigation action to later periods can make sense if we are optimistic about the availability, cost, and speed of deployment of low-emission technologies. A further degree of freedom is represented by negative emissions technologies. However, relying on a technological future that might not evolve as expected comes at a risk of missing the target completely.

Delayed Action and Incomplete Participation

Here we discuss how delaying mitigation action and incomplete participation among countries in pricing agreements affects the feasibility of climate stabilization targets and the emissions prices and costs associated with those targets.

For a given stabilization target, delayed global action implies that once global GHG emissions have peaked, they must then be reduced at an even faster rate, which could require an abrupt and very costly replacement of capital. In fact, if the world continues according to business as usual until 2030, according to most models, stabilization at 550 ppm CO2-e will no longer be possible (at least leaving aside highly optimistic scenarios for the development and deployment of negative emission technologies). This target might still be feasible if ambitious mitigation policies at the global scale are postponed until 2020, but this delay could substantially scale up global mitigation cost.

Rather than complete global inaction, more likely we will face asymmetry of actions across regions of the world. Significant mitigation actions are planned to take place in some developed economies within the next decade (e.g., the EU has pledged, by 2020, to reduce GHGs by 20 percent below 1990 levels). However, it is unlikely that emerging economies will make substantial emissions reductions in the coming decade.

Even asymmetric participation may rule out some of the more stringent targets, while scaling-up the global costs of those stabilization scenarios that remain feasible. Inaction in developing economies clashes with the fact that the bulk of emissions in the next decades will be coming from non-OECD countries.

If CO2 emissions are not regulated in some major emitting countries, several inefficiencies arise. At a given point in time, low-cost mitigation opportunities in countries without a mitigation policy will go unexploited, while other countries must bear a greater burden of mitigation costs. A dynamic inefficiency also arises, as unregulated countries are those where most new investments will take place. Investing instead in fossil technologies, fast-growing countries may lock in these long-lived technologies (e.g., a new coal plant may be in use for 50 years), and later conversion to low-carbon technologies becomes more costly, or simply impossible, if early scrapping is deemed unfeasible. Finally, nonparticipating countries might react to lower fossil fuel prices (arising from decreased fuel demand in participating countries) and increase their emissions, thus partially offsetting the environmental benefit of early movers, though studies suggest this carbon leakage effect is not too large.11 One solution, frequently discussed by economists, is the use of incentive systems (as for example an evolution of the Clean Development Mechanism) to induce reductions in developing economies while limiting leakage (e.g., Bosetti and Frankel, 2009).

Figure 3.3 again reports results from EMF-22, which looked extensively into climate agreements with incomplete country participation. We now concentrate on the delayed participation cases for each stabilization level.

A key result is that even in the limited number of models that suggest that the 450 ppm CO2-e stabilization scenario is feasible with early and full global cooperation over emissions mitigation (i.e., models with BECS technologies), the target becomes infeasible if only the OECD immediately undertakes mitigation action while BRICs and the rest of the world remain on their business-as-usual path until 2030 and 2050, respectively.

When participation of developing economies is delayed, half of the models cannot find a feasible set of investment actions that allow attaining the 550 ppm CO2-e scenario. Still, with overshooting, the CO2 emissions price faced by OECD countries in 2020 increases, on average, by a factor of three. On the other hand, delayed participation by developing economies does not make much difference to costs in the 650 CO2-e stabilization scenario. In this case, there is much greater scope for additional mitigation by all countries in the second half of the century to offset the foregone reductions early in developing economies, while still keeping within the concentration target.

A further point from Figure 3.3 is the wide range of disagreement across models, depending on assumptions about flexibility of substitution across technologies and, once more, on the assumptions concerning the availability of BECS (green versus blue markers in Figure 3.3 distinguish models with and without BECS technologies).

The set of technologies that will be available and the speed at which they will be deployed can crucially affect not only the costs of any climate policy, but also the time we can wait while still staying on track with a climate stabilization target. The stricter the climate objective or the later the mitigation effort starts, the greater the need to develop technologies (such as BECS and CCS) that have potential implications that we have not yet fully understood. This obviously requires a careful and realistic estimation of the costs and potentials of these technologies, the research, development, and demonstration requirements to make them available with a reasonable level of certainty, and the potential barriers and possible adverse side effects (e.g., CO2 leakage from storage sites) that might be linked to their deployment on a large scale.

How do projections we have discussed so far compare with the current state of climate negotiations? At the 2011 climate change meetings in Durban, South Africa (COP-17), for the first time, it was formally agreed that developing economies should be part of any future international emissions control regime (which is a step forward), although any control regime will not come into force before 2020 at the earliest. Prior to the meeting in Durban, countries agreed on submitting their emissions reduction “pledges” during the 2009 COP-15 in Copenhagen, Denmark, and the 2010 COP 16 in Cancun, Mexico, where most developed economies submitted their emissions reductions targets relative to emissions in 1990, 2000, or 2005.12 Brazil, Indonesia, Mexico, the Republic of Korea, and South Africa proposed reductions relative to their business-as-usual emissions,13 and China and India submitted carbon intensity reduction targets (i.e., CO2 emissions per unit of GDP). Some of the pledges have conditions attached, such as the provision of finance and technology or ambitious actions from other countries, while some pledges were provided as ranges. This leads to some flexibility in their implementation and a range of potential outcomes.

Therefore, implications of these pledges for 2020 global emissions will depend on what pledges are implemented and what rules will be applied. Many scientific groups have estimated global emissions in 2020 based on the Copenhagen Accord pledges. The 2010 Emission Gap Report (UNEP, 2010) collected these estimates and showed that emissions in 2020 could be as low as 49 Gt CO2-e (a range of 47 to 51 Gt CO2-e) when countries implement their conditional pledges in their more stringent form, or as high as 53 Gt CO2-e (a range of 52 to 57 Gt CO2-e) if pledges are implemented in their more lenient form.

Emission pathways consistent with a “likely” chance of meeting the 2° C limit generally peak before 2020, have emission levels in 2020 around 44 Gt CO2-e (a range of 39 to 44 Gt CO2-e), have extremely steep emission reductions afterward, and/or reach negative emissions in the longer term. Hence, the ranges implied by Copenhagen pledges do not necessarily rule out the 2° C target, as the two ranges are not severely distant from one another. However, as previously discussed, the larger the overshoot will be, the faster the decarbonization in the second half of the century will be needed, with all the implications that we have discussed in this chapter.

Who Bears the Costs of Abatement?

Distinguishing between who incurs mitigation costs and who actually implements mitigation activities is important. For example, mitigation might happen in developing economies but be financed by developed economies through an emissions offset program. Internationally allocating a given amount (typically determined by the stabilization target) of allowable emissions affects costs and who pays. This distributional issue would be extremely relevant both in the case of taxes and in that of permits. A number of possibilities for distributing the shares of emissions reduction among participating countries have been analyzed. Reductions might be based on equal percent reduction, GDP per capita, population, emissions intensity, historical responsibility, or many other alternative ways. As any of the schemes benefits (or imposes the cost on) countries unevenly in different aspects of socioeconomic indicators, there is no unique formula that would satisfy all participating countries.

There are two interacting equity concerns that would have to be dealt with in seeking the global emissions goal. First, incentives and compensation for developing economy participation will be required consistent with the principle of common but differentiated responsibilities. Second, since mitigation costs and compensation payments by developed economies will be substantial, they also will need to find an acceptable burden-sharing arrangement among themselves. Simple emissions reduction rules are incapable of dealing with the highly varying circumstances of different countries.

Successful climate negotiations will need to be grounded in a full understanding of the substantial amounts at stake. For example, for a 50 percent global GHG reduction by 2050 relative to 2000 (with full global participation in emissions pricing), Jacoby and others (2009) estimated that if developing economies (including China and India) are fully compensated for the costs of mitigation in the period up to 2050, then the average cost to developed economies is about 2 percent of the GDP in 2020 (relative to reference level), rising to 10 percent in 2050.14 The implied financial transfers are huge—over US$400 billion per year in 2020 and rising to about US$3 trillion in 2050—with the United States’ share amounting to US$200 billion in 2020 and over a trillion dollars in 2050.

It is surely extreme to assume that developing economies will demand complete compensation. Also, the amount of compensation is smaller if it only covers direct mitigation costs and not other losses, as might come through terms-of-trade effects.15 Nonetheless, international financial transfers under more aggressive climate stabilization targets would remain of unprecedented scale and seem highly unrealistic (at least in the near term), given large budget deficits at present.16 This further underscores the huge challenges to reaching a global agreement on rapid climate stabilization, challenges that only grow over time, when developing economies are expecting substantial compensation.

Yet another problem is that, besides being substantial, mitigation costs can also vary widely across countries. For example, mitigation costs are higher in energy-exporting countries, while energy importers have some terms-of-trade gains due to lower fossil fuel prices that allow them to reduce the cost of participating in emissions control regimes. Mitigation costs will also be greater in countries more dependent on carbon-intensive fuels and that employ inefficient mitigation instruments. Potential climate change damages also vary widely across countries but in a very different way. All these distributional impacts need careful study, as they complicate negotiation of long-term climate stabilization policy.


Advocates of rapid climate stabilization might be dismayed by some of the harsh technical, economic, and practical realities discussed in this chapter. Keeping mean projected warming above preindustrial levels to 2.0° C or stabilizing atmospheric GHGs at 450 ppm (about the current level) would require rapid widespread international adoption of emissions control policies and the development, and global deployment, of negative emission technologies later in the century to reverse atmospheric accumulations after a period of overshooting the long-term concentration target. Even the 550 ppm target (mean projected warming of 2.9° C) is extremely challenging, not least because required emissions prices escalate rapidly with further significant delay in controlling global GHGs, and the annual transfers to provide some compensation for developing economies are large and contentious to design. On the other hand, near-term emissions prices that are consistent with the 650 ppm target are more moderate, and delayed action on emissions reductions is less serious for this case, although obviously this target entails greater risks of dangerous warming.

The huge uncertainties—surrounding both the extent of climate change associated with a given atmospheric concentration target and our ability to develop technologies that would enable a rapid stabilization of the climate if the earth warms up rapidly—point to the importance of putting a policy architecture in place in the near term and delaying decisions about how rapidly emissions should be scaled back in the distant future until some of the uncertainties have been resolved. Aiming for a CO2 price somewhere in the ballpark of US$20 per tonne for 2020 applied across major emitting (developed and developing) countries seems reasonable and is roughly consistent with emissions prices suggested by the benefit-cost approach discussed in Chapter 4. Compensation issues for developing economies should also be more manageable at this level of pricing.

References and Suggested Readings

For details on the EMF-22 modeling exercise, see the following:

    Clarke, L., J.Edmonds, V.Krey, R.Richels, S.Rose, and M.Tavoni, 2009, “International Climate Policy Architectures: Overview of the EMF 22 International Scenarios,” Energy Economics, Vol. 31 (Supplement 2), pp. S64–S81.

    • Search Google Scholar
    • Export Citation

For an overview of future emissions reduction pledges by different countries, see the following:

For a discussion concerning the potential role of bio-energy and carbon capture and storage (CCS) technologies on the costs of stringent policy, see the following:

    Tavoni, M., and R. S. J.Tol, 2010, “Counting Only the Hits? The Risk of Underestimating the Costs of Stringent Climate Policy,” Climatic Change, Vol. 100, pp. 769–778.

    • Search Google Scholar
    • Export Citation

For a discussion about potential technological and economic obstacles for air capture technologies, see the following:

    Ranjan, M., 2010, “Feasibility of Air Capture” (master’s thesis; Cambridge, Massachusetts: Engineering Systems Division, Massachusetts Institute of Technology), available at

    • Search Google Scholar
    • Export Citation

Other publications:

    Babiker, M., J.Reilly, and L.Viguier, 2004, “Is International Emissions Trading Always Beneficial?” Energy Journal, Vol. 25, No. 2, pp. 33–56.

    • Search Google Scholar
    • Export Citation

    Bosetti, V., and J.Frankel, 2009, “Global Climate Policy Architecture and Political Feasibility: Specific Formulas and Emission Targets to Attain 460 ppm CO2 Concentrations,” NBER Working Paper No. 15516, (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    Bosetti, V., R.Lubowski, A.Golub, and A.Markandya, 2011, “Linking Reduced Deforestation and a Global Carbon Market: Implications for Clean Energy Technology and Policy Flexibility,” Environment and Development Economics, Vol. 16, No. 4, pp. 479–505.

    • Search Google Scholar
    • Export Citation

    Clarke, L., J.Edmonds, H.Jacoby, H.Pitcher, J.Reilly, and R.Richels, 2007, “Scenarios of the Greenhouse Gas Emission and Atmospheric Concentrations,” Subreport 2.1A of Synthesis and Assessment Product 2.1 by the U.S. Climate Change Science Program and the Subcommittee on Global Change Research (Washington: Department of Energy, Office of Biological and Environmental Research).

    • Search Google Scholar
    • Export Citation

    Edenhofer, O., C.Carraro, J. C.Hourcade, K.Neuhoff, G.Luderer, C.Flachsland, M.Jakob, A.Popp, J.Steckel, J.Strohschein, N.Bauer, S.Brunner, M.Leimbach, H.Lotze-Campen, V.Bossetti, Cian, M.Tavoni, O.Sassi, H.Waisman, R.Crassous-Doerfler, S.Monjon, S.Dröge, H.van Essen, P.del Río, and A.Türk, 2009, “RECIPE—The Economics of Decarbonization,” synthesis report. Available at:

    • Search Google Scholar
    • Export Citation

    Huang, J., R.Wang, R.Prinn, and D.Cunnold, D., 2009, “A Semi-Empirical Representation of the Temporal Variation of Total Greenhouse Gas Levels Expressed as Equivalent Levels of Carbon Dioxide,” Report 174 (Cambridge, Massachusetts: Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change), available at

    • Search Google Scholar
    • Export Citation

    IPCC, 2007, Climate Change 2007: Mitigation of Climate Change. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 2007. (New York: Cambridge University Press).

    • Search Google Scholar
    • Export Citation

    Jacoby, H., M.Babiker, S.Paltsev, and J.Reilly, 2009, “Sharing the Burden of GHG Reductions,” in Post-Kyoto International Climate Policy: Summary for Policymakers, ed. byJ.Aldy and R.Stavins (Cambridge, UK: Cambridge University Press), pp. 753–785.

    • Search Google Scholar
    • Export Citation

    Jacoby, H., F.O’Sullivan, and S.Paltsev, 2012, “The Influence of Shale Gas on U.S. Energy and Environmental Policy,” Economics of Energy and Environmental Policy, Vol. 1, No. 1, pp. 37–51.

    • Search Google Scholar
    • Export Citation

    Morris, J., J.Reilly, and S.Paltsev, 2010, “Combining a Renewable Portfolio Standard with a Cap-and-Trade Policy: A General Equilibrium Analysis,” Report 187 (Cambridge, Massachusetts: Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change), available at:

    • Search Google Scholar
    • Export Citation

    Paltsev, S., J.Reilly, H.Jacoby, and K.Tay, 2007, “How (and Why) Do Climate Policy Costs Differ Among Countries?” in Human-Induced Climate Change: An Interdisciplinary Assessment, ed. byM.Schlesinger, Haroon S.Kheshgi, JoelSmith, Francisco la Chesnaye, John M.Reilly, TomWilson, and CharlesKolstad (Cambridge, UK: Cambridge University Press), pp. 282–293.

    • Search Google Scholar
    • Export Citation

    Paltsev, S., J.Reilly, H.Jacoby, and J.Morris, 2009, “The Cost of Climate Policy in the United States,” Report 173 (Cambridge, Massachusetts: Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change), available at

    • Search Google Scholar
    • Export Citation

    PrinnR., S.Paltsev, A.Sokolov, M.Sarofim, J.Reilly, and H.Jacoby, 2011, “Scenarios with MIT Integrated Global System Model: Significant Global Warming Regardless of Different Approaches,” Climatic Change, Vol. 104, No. 3-4, pp. 515–537.

    • Search Google Scholar
    • Export Citation

    Sokolov, A., P.Stone, C.Forest, R.Prinn, M.Sarofim, M.Webster, S.Paltsev, A.Schlosser, D.Kicklighter, S.Dutkiewicz, J.Reilly, C.Wang, B.Felzer, J.Melillo, and H.Jacoby, 2009, “Probabilistic Forecast for 21st Century Climate Based on Uncertainties in Emissions (Without Policy) and Climate Parameters,” Journal of Climate, Vol. 22, No. 19, pp. 5175–5204.

    • Search Google Scholar
    • Export Citation

    United Nations Environment Programme, 2010, “The Emissions Gap Report: Are the Copenhagen Accord Pledges Sufficient to Limit Global Warming to 2° C or 1.5° C? A Preliminary Assessment” (Nairobi: United Nations Environment Programme).

    • Search Google Scholar
    • Export Citation

    Webster, M., A.Sokolov, J.Reilly, C.Forest, S.Paltsev, A.Schlosser, C.Wang, D.Kicklighter, M.Sarofim, J.Melillo, R.Prinn, and H.Jacoby, 2009, “Analysis of Climate Policy Targets Under Uncertainty” (Cambridge, Massachusetts: Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change), available at

    • Search Google Scholar
    • Export Citation

This chapter is based on a policy note prepared for the IMF Workshop on Fiscal Policy and Climate Mitigation on September 16, 2011, in Washington, DC. We are grateful to Michael Keen, Ian Parry, and all the participants of the workshop for comments and suggestions. The usual disclaimer applies.


See Clarke and others (2009); four of the integrated assessment models participated with two alternative versions for a total of 14 models.


The major sources of F-gases are air conditioning, semiconductor production, electrical switchgear, and the production of aluminum and magnesium.


Other substances will also affect future climate. These include chlorofluorocarbons (CFCs), whose emissions were largely phased out under provisions of the 1987 Montreal Protocol, but remain in the atmosphere as a powerful contribution to warming, and other short-lived warming substances like ozone and particulates. These substances add about another 30 ppm to atmospheric CO2-e. On the other hand, some substances, particularly sulfates, have a cooling effect through deflecting incoming sunlight.


Methane lifetime is about 12 years, nitrous oxide about 115 years, while F-gases lifetimes are thousands of years.


It is important to distinguish between the concentrations of all GHGs and a subset of the Kyoto gases. In 2010, the CO2 concentration was about 385 ppm and the Kyoto gases concentration was about 440 ppm CO2-e, while for all GHGs, concentration was about 465 ppm CO2-e. For more discussion on this issue, see Huang and others (2009).


The estimates in Table 3.1 should not be taken too literally, as they depend on assumptions about the probability distribution for warming at different long-term concentration levels, which are uncertain. The point is just that the risk of more extreme warming outcomes can be diminished sharply by stabilizing at lower GHG concentration levels.


There may be significant and relatively low cost opportunities for reducing emissions through protecting and enhancing global forest carbon stocks. Reducing emissions from deforestation and forest degradation (REDD) could lower the total costs of climate stabilization policies by around 10 to 25 percent or, alternatively, enable additional reductions of about 20 ppm CO2-e with no added costs compared to an energy sector–only policy (see Chapter 5 and Bosetti and others, 2011). However, implementation issues would need to be overcome (see Chapter 5): most of the rainforest countries have not yet developed the implementation capacity for monitoring and enforcing country-level projects, which might diminish the role of REDD in the next decade.


A small amount of GHGs can be emitted to offset the annual decay of GHGs in the atmosphere.


Another negative emissions possibility is filters for direct removal of CO2 out of the atmosphere, but these technologies (which are extremely costly and energy intensive at present) were not incorporated in the EMF-22 models.


Jacoby, O’Sullivan, and Paltsev (2012) provide an assessment of the role of natural gas in a potential U.S. climate policy considering recent shale gas development.


Most studies report carbon leakage from the Kyoto Protocol targets being in the range of 5 to 15 percent. See IPCC (2007), section at:


Typical targets for developed regions like Canada, the European Union, Japan, and the United States are in the range of 20 percent GHG reduction relative to 2000 levels.


Targets expressed with respect to baseline emissions are particularly tricky as they can be interpreted in very different ways depending on the baseline projection adopted.


The required carbon prices in this exercise are rising from about US$75/tCO2 in 2020 to about US$400/tCO2 in 2050.


In this case, the annual financial transfers to developing economies are lower by US$77 billion in 2020 and by US$108 billion in 2050 (Jacoby and others, 2009).


Even one of the Copenhagen Accord goals of raising US$100 billion per year by 2020 for climate finance from “a wide variety of sources” seems extremely challenging at this point (see Chapter 7).

    Other Resources Citing This Publication