Chapter

4. Carbon Taxes to Achieve Emissions Targets: Insights from EMF 24

Editor(s):
Ian Parry
Published Date:
March 2015
Share
  • ShareShare
Show Summary Details
Author(s)
Allen A. FawcettLeon C. Clarke and John P. Weyant 

Key Messages for Policymakers

  • According to the most recent Energy Modeling Forum (EMF 24), U.S. CO2 emissions are projected to be 2 percent below 2005 levels in 2020 (mean value across all reference case scenarios), though recent trends in energy markets that are not included in vintage of the models used in this exercise (e.g., the shale gas expansion) would tend to lower those projections.

  • On average, the estimated 2020 CO2 price consistent with the Administration’s pledge to reduce emissions by 17 percent below 2005 levels in 2020 is $35 per ton.* However, estimated prices vary dramatically with different modeling assumptions, especially in regard to the future costs and availability of emissions-saving technologies and complementary policies – e.g., recently enacted fuel economy regulations when combined with carbon capture and storage (CCS) requirements for new coal-fired power plants and a renewable portfolio standard (RPS) lower the average price to about $20 per ton in 2020.

  • Across all model runs, emissions reductions come from a broad range of sources including shifting away from fossil sources of primary energy through reduced energy consumption and replacement with non-emitting energy sources. In the latter regard, the largest reductions come from a shift away from coal, with smaller reductions coming from shifts away from oil, and in some cases from natural gas.

  • Technology options such as nuclear, carbon capture and storage, renewables, and energy efficiency are, when taken together, extremely important for cost-effective emissions mitigation, though no single technology is crucial. Instead, different combinations of technologies can lead to cost-effective decarbonization of the energy system.

  • This underscores the need for a credible and sustained carbon price, complemented with measures to further encourage a broad portfolio of technology developments. Complementary measures to contain uncertainty over future emissions (e.g., carbon budgets specifying allowable cumulative emissions over an extended budget period) may also be desirable.

*Prices here are in year 2005 – to express in year 2012 multiply by 17.5 percent.

1. Introduction

In the one hundred eleventh Congress, considerable attention was given to cap-and-trade legislation. The American Clean Energy Security Act of 2010 (introduced by Congressmen Waxman and Markey) passed the House, though the Senate version of the bill (Kerry–Lieberman’s American Power Act) was never put to the vote. With the prospects for cap-and-trade legislation stalled (and not helped by depressed prices in the European Trading System), and the need to strengthen the U.S. fiscal position, increasing attention is being paid to a domestic carbon tax.1

Although both of these policies would place a price on greenhouse gas (GHG) emissions, they behave differently under uncertainty (e.g., about future fuel prices and technology costs). A cap-and-trade policy (at least in its pure form without price stability provisions) specifies an emissions quantity target, leaving the carbon price uncertain, while a carbon tax specifies a price, leaving the quantity of GHGs uncertain. Economists generally recommend that emissions prices are stable (though rising) over time, to help equalize incremental abatement costs in different periods, and create a stable environment for clean technology investments (e.g., Pizer 2005). However, international negotiations generally focus on specific emissions reductions pledges, and the quantity of GHGs realized under a policy may be of significant domestic concern, so it is important to understand what emissions prices might be consistent with alternative emission reduction targets.

However, as the shale gas boom reminds us, the future, even the near future, evolution of the energy system is very difficult to predict. What is needed is not one model, with one future scenario, but rather a broad range of models and scenarios, to give some sense of what carbon price trajectories might be consistent with specific U.S. emissions reductions goals, and what the resulting transformation of the U.S. energy system might look like. This chapter therefore uses results from the Energy Modeling Forum (EMF) 24 (Weyant et al. 2014),2 which encompasses a broad range of future scenarios (e.g., for economic growth, fuel resources and prices, and the costs and availability of energy technologies) and expert opinion.

The discussion is organized as follows. Section 2 gives an overview of the motivation for and the design of the EMF 24 workgroup. Section 3 examines the results from a range of scenarios that all achieve emissions reductions of 17 percent below 2005 levels by 2020 and 50 percent below 2005 levels by 2050, under a variety of different technology and policy assumptions, focusing on emissions, carbon price, and primary energy impacts. Section 4 explores the results for alternative emissions targets ranging from constant 2005 emissions to 80 percent below 2005 levels by 2050. Finally, section 5 concludes.

2. Overview of the study design

The Stanford Energy Modeling Forum – which has engaged in 28 studies since it was established in 1977 – brings together leading energy and economic model builders and model users representing academic, corporate, and government perspectives. The primary goals of the Forum are to utilize the collective expertise of the participants to improve understanding of an important energy and environmental problem; identify policy-relevant insights and analyses that are robust across a wide range of models; explore and explain the strengths and limitations of alternative modeling approaches; and help identify high-priority areas of future research.

The EMF 24 exercise began with three motivating questions. What would the U.S. energy system transition look like under pricing policies needed to meet U.S. emissions goals that are roughly consistent with an ambitious international climate goal, such as containing mean projected global warming above pre-industrial levels to 2°C or 3°C? What are the potential implications of transportation and electric sector regulatory approaches to emissions reductions in meeting this goal? And, how might the technological improvements and technological availability influence the answers to both of the above questions?

To address these questions, each of the modeling teams involved in EMF 24 (insofar as possible) used a matrix of 42 scenarios, covering 8 different technology, and 15 different policy, assumptions, though only a subset of these runs is reported here. Table 4.1 summarizes these scenarios and those considered here (indicated by the shaded cells).

Table 4.1EMF 24 scenario matrix
Note: Emissions targets represent percentage reduction in 2050 relative to 2005 levels.

The technology assumptions consist of optimistic and pessimistic assumptions for end-use technology (e.g., energy efficiency), carbon capture and storage (CCS), nuclear energy, wind and solar energy, and bioenergy – the full list of technology assumptions used in EMF 24 is given in Table 4.A.1 in the Appendix.3 The policy assumptions include baseline or reference scenarios with no policy; emissions target scenarios of varying stringency, that all linearly reduce emissions over time to different percentages below 2005 levels; electricity and transportation regulatory scenarios combined with a 50 percent emissions target policy (by 2050); and various regulatory policies in isolation. The policy assumptions used in EMF 24 are further described in Table 4.A.2 in the Appendix.

This chapter focuses on two sets of scenarios. First, the 50 percent emissions target scenarios across all technology assumptions plus the 50 percent emissions target scenarios combined with regulatory policies. These scenarios present a common emissions target, linearly reducing GHGs from baseline 2013 levels to 50 percent below 2005 levels by 2050, across a wide range of technology assumptions and complementary policies, and allow an examination of the range of carbon prices and technology futures compatible with this target. Second, the chapter explores the range of carbon prices and energy systems that result from some of the alternative emissions targets.

Table 4.2 lists the modeling teams involved in EMF 24. Not all of the participating models submitted the full set of runs, and some models are limited to the electricity sector. Here we summarize results from eight of the models (ADAGE, CIMS, EC-IAM, FARM, GCAM, NewERA, US-REGEN, and USREP), without explicitly identifying results by model.4

Table 4.2Modeling teams
ModelFull nameInstitution
ADAGEAnalysis of the Global Economy ModelResearch Triangle Institute
CIMSCanadian Integrated Modeling SystemSimon Fraser University
EC-IAMEnvironment Canada Integrated Assessment ModelEnvironment Canada
FARMFuture Agricultural Resources ModelU.S. Department of Agriculture
GCAMGlobal Change Assessment ModelPacific Northwest National Laboratory/Joint Global Change Research Institute
NEMSNational Energy Modeling SystemEnergy Information Administration
NewERANewERANERA Economic Consulting
ReEDSRegional Energy Deployment SystemNational Renewable Energy Laboratory
US-REGENU.S. Regional Economy, GHG, and Energy ModelElectric Power Research Institute
USREPU.S. Regional Energy Policy ModelMIT Joint Program on the Science and Policy of Global Change

Note that the scenario design and model baselines for EMF 24 were locked down in early 2012, so the baselines do not reflect policies that were later adopted (e.g., the light-duty vehicle and corporate average fuel economy standards that were published in October 2012). Additionally, developments in energy markets such as the shale gas boom have altered baseline emissions projections since the EMF 24 scenarios were developed (e.g., the Energy Information Administration’s Annual Energy Outlook (AEO) for 2013 projects 2020 CO2 emissions to be 6 percent lower than the then current AEO 2011 projections).

3. Exploration of the 50 percent reduction target

The 50 percent emissions target scenarios in EMF 24 impose a GHG emissions cap from 2013 through 2050 that linearly reduces emissions from reference levels in 2013 to 50 percent below 2005 levels in 2050. Banking is allowed, but borrowing is not. And scenarios do not allow for international or domestic emissions offsets. The results indicate the emissions, or emissions reductions, associated with different emissions price trajectories (assuming all reductions are made in covered sectors).

With the exception of CO2 emissions from land use and land use change, the cap covers all ‘Kyoto’ gases in all sectors of the economy that the particular model represents.5 For comparison, the American Clean Energy Security Act of 2010 covered a smaller percentage of emissions and allows extensive offsets, so even though that bill required an 83 percent reduction in covered emissions below 2005 levels by 2050, actual U.S. emissions under the bill are more similar to the 50 percent emissions target scenario considered here.6

3.1 Emissions

Figure 4.1 shows covered CO2 emissions with baseline scenarios on the left and 50 percent emissions target scenarios on the right. The left-hand panel displays 58 separate model runs, 8 baseline scenarios, and 8 participating models (three of the scenarios were not run by two of the models). The right-hand panel displays 74 separate model runs, 10 policy models and 8 participating models (again with three scenarios not run by two of the models). The dark lines represent mean, and plus and minus one standard deviation, covered CO2 emissions levels in each year.

Figure 4.1Covered CO2 emissions – baseline and 50 percent emissions target scenarios

Taking the left panel first, in 2020 the baseline covered CO2 emissions range from approximately 5,400 to 6,800 MtCO2, with a mean value of 6,006 MtCO2. This mean is 2 percent below 2005 U.S. CO2 emissions levels.7 The range in 2050 is from approximately 5,200 to 8,600 MtCO2, with a mean value of 7,085 MtCO2, 16 percent above 2005 levels.

Generally the models show that emissions are growing more slowly than the economy – although population and per capita income are growing over this period, the resulting upward pressure on emissions is partly offset by declining energy intensity of GDP (due to improvements in energy efficiency) and declining carbon intensity of energy (due in part to a progressive shift away from carbon-intensive coal to other fuels). The spread in baseline emissions largely reflects different scenarios for the latter two factors (all of the models use similar population and GDP growth rates). In the context of a carbon tax, this uncertainty in baseline emissions leads to uncertainty about the level of carbon tax required to achieve a particular target, or alternatively, uncertainty about the emissions levels that would be achieved by a specific carbon tax.

Turning to the right panel, across all of the 50 percent emissions target scenarios, covered CO2 emissions range from approximately 4,100 to 5,600 MtCO2 in 2020, with a mean value of 5,083 MtCO2, 17 percent below 2005 U.S. CO2 emissions levels. The 2050 range is from approximately 2,500 to 4,000 MtCO2, with a mean value that is 49 percent below 2005 U.S. CO2 levels at 3,138 MtCO2.

There are two reasons for the spread in emissions across model runs in these policy scenarios – even though the same emissions targets are applied across all model runs. First, the policy allows for permit banking (though not borrowing), allowing covered emissions in a particular year to deviate from the annual emissions target (so long as cumulative emissions are equal to the cumulative emissions allowed under the cap). Second, the policy is specified as a cap on all Kyoto gases represented by a particular model; however, not all participating models incorporate non-CO2 emissions. Because abatement opportunities differ between CO2 and non-CO2 emissions, in models capturing both emissions sources, proportionate reductions in CO2 emissions may differ from proportionate reductions in GHGs required under the cap.

3.2 Carbon prices

For the 74 model runs covering all of the 50 percent emissions target scenarios we now examine the resulting carbon prices to see what carbon tax levels could be consistent with this emissions target.

Figure 4.2 displays the carbon price paths for all of the 50 percent emissions target runs across all models. The dark lines represent mean, and plus and minus one standard deviation.

Figure 4.2Carbon prices – all 50 percent emissions target scenarios

The average carbon price in 2020 is $35/tCO2 with standard deviation of $30/tCO2 (prices in year 2005). And for 2050, the average price is $163/tCO2 with standard deviation of $96/tCO2.

This figure therefore underscores the large uncertainties involved in choosing a carbon tax to meet specific emissions goals. Correspondingly, once a carbon tax is specified, there will be large uncertainties in the resulting emissions levels. One policy implication is that, just as many mechanisms have been considered for addressing price uncertainty under a quantity-based policy, it may be important to consider mechanisms for dealing with quantity uncertainty under a price-based policy if meeting specific emissions targets is of concern. One possibility would be to combine a carbon tax with a ‘carbon budget’ specifying the cumulative carbon emissions allowed to date along a path to meet an emissions goal. If at any point actual emissions exceed the carbon budget to that date, then a clock would start and if cumulative emissions in 5 or 10 years are not below the carbon budget allowed to that date, then a mechanism would kick in to adjust the tax or tax growth rate upward. This could provide a degree of certainty about meeting the emissions target, while avoiding any unanticipated price shocks.

Table 4.3 lists the mean and standard deviation across models for the carbon price in 2020 and 2050 across the eight models for six specific scenarios with different assumptions about technologies and other policies (see Table 4.A.1 in the Appendix for detailed descriptions of the technology assumptions in each scenario).

Table 4.3Carbon prices – mean and standard deviation by scenario
Policy assumptionTechnology assumption20202050
Mean ($2005/tCO2)Std. dev. ($2005/tCO2)Mean ($2005/tCO2)Std. dev. ($2005/tCO2)
AllAll$35$30$163$96
50% TargetOptimistic$30$26$133$66
50% TargetPessimistic$60$48$283$114
50% TargetCCS/Nuke$33$25$145$85
50% TargetRenewable$36$27$173$77
50% Target + Comp.CCS/Nuke$22$26$113$92
50% Target + Comp.Renewable$25$25$143$90
Note: The top row gives the mean and standard deviation of the carbon price across all scenarios that model the 50 percent emissions reductions target, including all scenarios that vary technology assumptions, and scenarios that include complementary policies along with the 50 percent target. Breakouts of specific policy and technology assumptions are given for all combinations that were completed by the full set of models included in this chapter. The ‘CCS/Nuke’ technology assumption represents the scenarios with optimistic assumptions for CCS, nuclear, and end-use technology, and pessimistic assumption for other technologies; the ‘Renewable’ technology assumption represents the scenarios with optimistic assumptions for wind, solar, biomass, and end-use technology, and pessimistic assumptions for other technologies.

As expected, generally lower carbon prices are required to reach the emissions targets under the optimistic technology assumption, and higher carbon prices are required under the pessimistic technology assumption. The standard deviations show that even within a specific scenario, considerable variation in the carbon price across the models remains. Carbon prices are fairly similar under the two mixed technology assumptions, the optimistic CCS/nuclear assumptions (which also limit renewable technologies) and the optimistic renewable assumptions (which do not allow CCS or new nuclear). This is an important result, the comparison between the optimistic technology and pessimistic technology cases clearly demonstrate the important role technology plays in enabling cost-effective emissions reductions, and the mixed technology assumptions show that there is no single silver bullet technology or technology portfolio, but instead there are multiple different technology pathways that can lead to cost-effective emissions abatement.

Under both the optimistic CCS/nuclear and the optimistic renewable technology assumptions in Table 4.3, the carbon prices fall dramatically when the emissions target policy is combined with complementary policies including a CAFE standard, RPS, and a CCS requirement for new coal-fired power plants. Under the optimistic CCS/nuclear technology assumptions the addition of the regulatory policies lowers the mean carbon price by 33 percent in 2020 and 22 percent in 2050. Under the optimistic renewable assumptions, the reductions are 31 percent and 17 percent in 2020 and 2050, respectively.

Although the carbon prices fall with the addition of the complementary policies, the overall economic cost of achieving these emissions reductions increases.8 In the presence of an emissions target policy, the complementary policies have little impact on emissions levels, but instead simply favor a particular set of technologies. From the perspective of a carbon tax policy, the presence of complementary policies can be thought of in two ways: they can lower the carbon tax necessary to achieve a desired emissions target, or they can increase the emissions reductions achieved under a particular carbon tax.

3.3 Primary energy

Next we turn to examining how the U.S. energy system is transformed in all of the 50 percent emissions target scenarios in order to demonstrate the diverse ways in which the emissions target can be achieved. Figure 4.3 compares the primary energy in 2020 and 2050 for the 58 baseline model runs. In this figure the model runs are sorted by quantity of fossil primary energy without carbon capture and storage (CCS) in 2050. The variation across model runs is due both the technology assumptions in each scenario and general differences between the eight models and their core driving assumptions such as the evolution of energy intensity. Taken together these runs represent a range of plausible business-as-usual futures for the U.S. energy system. In 2020, fossil fuels represent between 84 and 99 exajoules (EJ) of primary energy, and by 2050 the range expands to 78 to 130 EJ. The biggest differences are seen in the projections for the coal primary energy, which ranges from 19 to 33 EJ in 2020 and from 13 to 74 EJ in 2050. The amount of oil primary energy has a similar spread of results, between 30 and 46 EJ in 2020 and between 5 and 57 EJ in 2050.9 Natural gas primary energy varies between 23 and 33 EJ in 2020 and between 25 and 66 EJ in 2050.

Figure 4.3Primary energy – all baseline scenarios – total fossil in 2050 sort

For the non-fossil fuels, the technology assumptions in different baseline scenarios explicitly drive some of the variation in primary energy. For example, the pessimistic nuclear scenario requires the phase-out of nuclear power with no new construction and no lifetime extensions beyond 60 years. This assumption drives the lower end of the range of nuclear energy primary energy, which falls between 2 and 4 EJ in 2020, and between 0 and 12 EJ in 2050.10 Finally, the primary energy from all renewables ranges from 1 to 7 EJ in 2020, and from 1 to 13 EJ in 2050.

Turning to primary energy in the policy scenarios, Figure 4.4 shows primary energy across all 74 separate model runs for 50 percent emissions target scenarios. The expansion and contraction of different primary energy carriers are indicative of how the required emissions reductions are achieved. The first difference to notice is that the top white light gray bars represent the energy efficiency and demand reduction, which is between 3 and 21 EJ in 2020 and between 7 and 70 EJ in 2050. Primary energy from fossil without CCS falls across all runs to between 71 and 85 EJ in 2020 and between 45 and 75 EJ in 2050. The most dramatic and consistent shift across all runs is the reduction in primary energy from coal without CCS, which falls to between 2 and 25 EJ in 2020 and 0 to 15 EJ in 2050. Primary energy from oil and natural gas fall less dramatically, with 29 to 43 EJ of oil primary energy in 2020 and 7 to 50 EJ in 2050, and 21 to 32 EJ of natural gas primary energy in 2020 and 15 to 47 EJ in 2050.

Figure 4.4Primary energy – all 50 percent emissions target scenarios – total fossil in 2050 sort

In the policy scenarios we break out primary energy from sources with and without carbon capture and storage. CCS technology is not allowed by assumption in some scenarios, so the lower bound of primary energy with CCS is always zero. In cases where it is allowed, the maximum penetration is 5 EJ of primary energy with CCS in 2020, and 46 EJ in 2050. Nuclear energy provides between 1 and 4 EJ in 2020, and between 0 and 17 EJ in 2050, with the lower end being driven by scenarios that assume nuclear phase-out. Biomass renewables provide between 0 and 7 EJ of primary energy in 2020, and between 0 and 35 EJ in 2050. Finally, non-biomass renewables account for between 1 and 5 EJ of primary energy in 2020, and between 1 and 19 EJ of primary energy in 2050.

In order to tie the primary energy data back to the carbon price trajectories from Figure 4.2, Figure 4.5 displays primary energy across all 50 percent emissions target scenarios sorted by the carbon price in the corresponding year. One pattern that emerges is that many of the runs with lower carbon prices tend to have lower reference case primary energy in 2050, and the reverse is true for the higher carbon price runs, as can be seen by the total height of the bars including the energy efficiency and demand reduction category. Total reference case primary energy varies more across models than across scenarios within any particular model. Looking at the wide range of carbon prices and technology futures in the EMF 24 database that are consistent with the 50 percent emissions target emissions goals, we see that even with a broad set of sensitivities on technology assumptions and complementary policies, the range of possible outcomes projected by any single model is far narrower than the range presented in a multi-model comparison.

Figure 4.5Primary energy – all 50 percent emissions target scenarios – carbon price sort

4. Exploration of alternative emissions targets

In this section we move away from the 50 percent emissions target scenarios examined in the previous section, and look at the carbon prices and energy systems that are consistent with a range of alternative emissions goals. In addition to the 50 percent emissions target scenarios that cover all technology assumptions in EMF 24, the scenario matrix includes scenarios with policies that linearly reduce GHG emissions to 0, 10, 20, 30, 40, 60, 70, and 80 percent below 2005 levels by 2050. Aside from the emissions target levels all other aspects of these policies are identical to the 50 percent emissions target scenario (see the Appendix Table 4.A.2 for full descriptions of the policy assumptions). Unlike the 50 percent emissions target scenarios, these alternative targets are only analyzed under two sets of technology assumptions: first, a CCS and nuclear world, with optimistic assumptions on CCS, nuclear, and energy efficiency, and pessimistic assumptions for biomass, wind and solar; second, a renewables world, with optimistic assumptions for energy efficiency, biomass, wind and solar, and pessimistic assumptions for CCS and nuclear. In this section we examine two of these alternative emissions target scenarios, the 0 percent emissions target scenario and the 80 percent emissions target scenario.

Figure 4.6 depicts results for the 0 percent emissions target scenarios, which hold emissions constant at 2005 levels through 2050. Seven models submitted results for these scenarios. For three of the models, this emissions target is non-binding, so the carbon price is zero. Across all of the runs, the mean carbon price is $4/tCO2 in 2020 and $15/tCO2 in 2050 (in year 2005).

Figure 4.6Emissions, carbon price, and primary energy – 0 percent emissions target scenarios

The U.S. energy system remains unchanged in the three models where this policy is non-binding, while for the remaining models we find a variety of responses. For primary energy from coal without CCS the largest change from the reference case is a 6 EJ reduction in primary energy in 2020 and a 15 EJ reduction in 2050, and across all the models the mean reduction is 2 EJ in 2020 and 5 EJ in 2050. The largest reduction in oil without CCS primary energy is 1 EJ in 2020 and 4 EJ in 2050, and the mean change is 0 EJ in 2020 and a 1 EJ reduction in 2050. Natural gas without CCS primary energy shows a mean reduction of 0 EJ in 2020 and 1 EJ in 2050, and the largest reduction is 1 EJ in 2020 and 3 EJ in 2050. Only one model projects CCS to penetrate, finding 1 EJ of CCS in 2020 and 21 EJ of CCS in 2050.

The non-fossil sources of primary energy show much smaller changes in the 0 percent emissions target scenario. In 2020 and in 2050 the mean change is zero for nuclear, biomass, and non-biomass renewables. The largest changes are an addition of 1 EJ of nuclear and 1 EJ of biomass primary energy in both 2020 and 2050. Finally there is a mean 2 EJ of energy efficiency and demand reduction in 2020 and 5 EJ in 2050, with a maximum of 7 EJ in 2020 and 20 EJ in 2050.

The 80 percent emissions target scenarios, which reduce emissions linearly to 80 percent below 2005 levels by 2050, are shown in Figure 4.7. Seven models submitted results for these scenarios. The mean carbon price across all of the 80 percent emissions target runs is $65/tCO2 in 2020 and $439/tCO2 in 2050, with a standard deviation of $40/tCO2 in 2020 and $228/tCO2 in 2050. Note that the mean is driven up by one model that exhibits a large price spike in 2050 to near $900/tCO2, the remaining models find prices between $135/tCO2 and $593/tCO2. In 2045 the carbon prices are more closely grouped with a mean of $296/tCO2, and a $108/tCO2 standard deviation.

Figure 4.7Emissions, carbon price, and primary energy – 80 percent emissions target scenarios

The U.S. energy system is completely transformed in the 80 percent emissions target scenarios. In 2020 there is a mean of 12 EJ of primary energy from coal without CCS remaining, with a minimum of 2 EJ and a maximum of 22 EJ, and by 2050 the mean is only 2 EJ, with a minimum of 0 EJ and a maximum of 4 EJ. Oil sees less of a change than coal in this scenario, with between 29 and 41 EJ remaining in 2020, with a mean of 36 EJ, and between 5 and 44 EJ in 2050, with a mean of 22 EJ. Note that the low end of oil primary energy in 2050 is from a model that also sees very little oil remaining in the baseline. In 2020 primary energy from natural gas without CCS falls between 19 and 29 EJ, with a mean of 25 EJ, and in 2050 the range is from 4 to 16 EJ, with a mean of 11 EJ. The pattern of large reductions in primary energy from coal and smaller reductions in oil and gas primary energy occurs because the power sector is able to largely decarbonize by shifting to low or zero carbon technologies, whereas the transportation sector has fewer low-cost options in these models for switching away from oil.

In the 80 percent emissions target scenarios where it is available, CCS plays a large role. Three models project CCS penetrating in 2020 with between 1 and 5 EJ of primary energy from fossil with CCS. In 2050 all models have primary energy from fossil with CCS, with a minimum of 2 EJ and a maximum of 16 EJ. One model also includes an additional 8 EJ of biomass with CCS, for a total of 24 EJ of primary energy with CCS.

Primary energy from non-fossil sources grows in the 80 percent emissions target scenarios. Nuclear primary energy only increases in the scenarios where it is not assumed to be phased out, with little change in 2020, but an increase of between 2 and 9 EJ in 2050. Biomass without CCS increases between 0 and 3 EJ in 2020 and between 0 and 12 EJ in 2050. Non-biomass renewables increase between 0 and 5 EJ in 2020 and between 0 and 15 EJ in 2050. Finally, the biggest impact on primary energy in the 80 percent emissions target scenario is the reduction in the overall level of primary energy due to energy efficiency and demand reduction, which accounts for between 7 and 24 EJ in 2020 with a mean of 15 EJ, and between 9 and 78 EJ in 2050 with a mean of 46 EJ.

5. Conclusion

An important distinction between carbon taxes compared to pure cap-and-trade policies is that carbon taxes provide a degree of cost certainty at the expense of certainty about emissions outcomes, whereas cap-and-trade policies provide the reverse, emissions certainty and cost uncertainty. If a carbon tax is implemented with a goal of achieving a specific emissions goal (e.g., an international agreement that specifies required emissions reductions, or simply a domestically stated emissions goal), the runs contained in the EMF 24 database show that a very wide range of carbon taxes is potentially consistent with any given emissions goal. Factors influencing this range include assumptions about the business-as-usual emissions trajectory, assumptions about how the cost and performance of technologies evolve, and general differences across models. If the emissions goal must be met, and the carbon tax produces less abatement than expected, then the policy must be adjusted to meet the goal. Cap-and-trade policies have generally included features designed to address price uncertainty (e.g., safety valve, price collar, offsets), and carbon tax policies may need to consider equivalent mechanisms to address uncertainty in cumulative emissions reductions, while maintaining predictability in prices.

The EMF 24 database also sheds some light on the interaction between a policy that prices carbon and other complementary policies. The models find that policies such as a CAFE standard or RPS, when combined with a quantity-based emissions target, do not change the amount of emissions reductions, but instead change the way in which those reductions are achieved, which generally lowers allowance prices, but increases overall costs. When these complementary policies are combined with a carbon tax, there are two ways to consider the interaction. The complementary policies generate emissions reductions, and thus reduce the carbon tax that would be needed to reach any specific emissions goal, but may raise the overall cost of reaching that goal. Alternatively, the complementary policies increase the total amount of abatement achieved under any particular carbon tax.

The variety of technology assumptions and range of models in EMF 24 demonstrate the diversity of futures for the U.S. energy system that can be consistent with emissions reduction goals. The differences in energy futures are at least partially accounted for by how the energy system is assumed to evolve in a business-as-usual scenario, the specifics of the policy assumptions, the evolution of the cost and performance of different technologies, and a host of other factors that characterize the broad differences between models.

Notes

The views and opinions of this author herein do not necessarily state or reflect those of the United States Government or the Environmental Protection Agency.

For some recent discussions in the literature see, for example, Ramseur et al. (2012), McKibbin et al. (2012), Parry and Williams (2011), and Rausch and Reilly (2012).

This chapter uses the penultimate data submissions to the EMF 24 exercise. The final submissions may be subject to some modest revisions as we finalize the study that are not expected to impact the core results presented in this study. The final EMF 24 data is now publicly available, following publication of various papers in a special issue of the Energy Journal (Weyant et al. 2014).

For example, pessimistic CCS assumptions allow no implementation of the technology and pessimistic nuclear assumptions allow no new construction of nuclear power plants. Conversely, optimistic assumptions for nuclear and CCS specify that the technologies are available, but the cost and performance characteristics are left to the modeler’s choice.

The complete database of EMF 24 is publicly available, following publication of the EMF 24 special issue of the Energy Journal (Weyant et al. 2014).

This refers to the six main gases included in the Kyoto Protocol: carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), and sulphur hexafluoride (SF6).

Fawcett et al. (2009) discusses this further in the context of the similarly designed EMF 22 scenarios.

As noted above, the baseline scenarios for EMF 24 were locked down in early 2012. More recent projections would tend to forecast lower baseline emissions. This is in part due to changes in energy markets such as the natural gas boom, but also in part due to policies that have been adopted such as the light-duty vehicle corporate average fuel economy standard, which in this study is roughly represented in a policy scenario.

The policy overview paper for EMF 24 (Fawcett et al. 2014) includes an extensive discussion of the costs of alternative scenarios, and alternative possibilities for measuring economic cost.

Note that the extreme high amount of coal primary energy and corresponding extreme low amount of oil primary energy are results from a single model that tends to project a move in the transportation sector from petroleum fuels to coal-to-liquids. The model differences that drive these results are further explored in Clarke et al. (2014).

Primary energy values for non-fossil technologies are calculated using the direct equivalent method.

References

    Clarke, L. C., Fawcett, A. A., McFarland, J., and Weyant, J. P. (2014). Overview of the EMF 24 Technology Scenarios. Energy Journal35.

    Fawcett, A. A., Calvin, K. V., de la Chesnaye, F. C., Reilly, J. M., and Weyant, J. P. (2009). Overview of EMF 22 U.S. Transition Scenarios. Energy Economics31: S198S211.

    Fawcett, A. A., Clarke, L. C. and Weyant, J. P. (eds.) (Forthcoming). Energy Modeling Forum 24. Energy Journal Special Issue.

    Fawcett, A. A., Clarke, L. C., Rausch, S., and Weyant, J. P. (2014). Overview of the EMF 24 Policy Scenarios. Energy Journal35.

    McKibbin, W., Morris, A., and Wilcoxen, P. (2012). The Potential Role of a Carbon Tax in U.S. Fiscal Reform. Brookings Climate and Energy Economics Discussion Paper.

    Parry, I., and Williams, R. (2011). Moving US Climate Policy Forward: Are Carbon Taxes the Only Good Alternative?Resources for the Future Discussion Paper 11–02.

    Pizer, W. (2005). Climate Policy Design under Uncertainty. Resources for the Future Discussion Paper0544.

    Ramseur, J., Legget, J., and Sherlock, M. (2012). Carbon Tax: Deficit Reduction and Other Considerations. Congressional Research Service Report R42731.

    Rausch, S., and Reilly, J. (2012). Carbon Tax Revenue and the Budget Deficit: A Win-Win-Win Solution?MIT Joint Program on the Science and Policy of Global Change Report No. 228.

Appendix
Table 4.A.1EMF 24 technology assumptions
PolicyDescription
50% Emissions TargetThis represents the assumption of a national policy that allows for cumulative greenhouse gas emissions from 2012 through 2050 associated with a linear reduction from 2012 levels to 50 percent below 2005 levels in 2050. The cumulative emissions are based on the period starting from, and including, 2013 and through 2050. With the exception of CO2 emissions from land use and land use change, the cap covers all Kyoto gases in all sectors of the economy that the particular model represents. This includes non-CO2 land use and land use change emissions and emissions of GHGs not covered under many U.S. climate bills. To be explicit, CO2 emissions from land use and land use change are not included in the cap. For models that do not operate on annual time steps, the first year with a positive price on carbon should be the first model time step after 2012 (e.g., 2015 in a model with 5-year time steps), but the cumulative emissions are still to be based on an assessment of the emissions associated with a linear path starting from, and including, 2013 and through 2050. Banking of allowances is allowed, but borrowing of allowances is not permitted. Note that the emissions target scenarios with alternative targets use identical assumptions to allow for cumulative greenhouse gas emissions associated with a linear reduction from 2012 levels to X percent below 2005 levels in 2050, where X is the percentage reduction target associated with the scenario.
Renewable Portfolio Standard (RPS)The RPS applies only to the electricity sector. In this case, renewable energy includes all hydroelectric power and bioenergy. The RPS is defined as 20 percent by 2020, 30 percent by 2030, 40 percent by 2040, and 50 percent by 2050. Banking and borrowing are not allowed. If modelers are unable to meet these requirements within their model, they should create a scenario that includes a less aggressive RPS, but one that can be met by the model.
New CoalThis policy requires that all new coal power plants capture and store 90 percent or more of their CO2 emissions.
Transportation Regulatory PolicyThe transportation policy is a CAFE standard for light-duty vehicles (LDV) that specifies a linear increase in fuel economy of new vehicles, starting in 2012, to 3 times 2005 levels in 2050. If modelers do not have the ability to represent a CAFE policy, they can alternatively represent the policy as a cap that covers all LDV in the transportation sector, as defined in the particular model. This alternative policy is defined as a linear reduction in LDV emissions from 2012 levels to 55 percent below 2010 levels in 2050. Banking and borrowing are not allowed. It is understood that with rebound effects and differences in reference scenario, this LDV emissions cap policy structure will not be identical to the CAFE policy; however, we expect them to be similar (the 55 percent reduction in LDV emissions under the cap is consistent with the emissions reductions achieved in a test run of GCAM), and there are benefits to explicit analysis of CAFE standards. Note that biofuels, electricity, and hydrogen are assumed to be zero-emissions fuels for calculating the emissions cap.
Emissions target + RegulationsCombines all policies listed above.
Table 4.A.2EMF 24 policy assumptions
PessimisticOptimistic
End UseThe pessimistic technology scenario should represent evolutionary assumptions about the availability, cost, and performance of technologies that would reduce energy consumption at the end use or enhance opportunities for fuel switching. The precise assumptions are left to modeler’s choice.The optimistic technology scenario should represent plausibly optimistic assumptions about the availability, cost, and performance of technologies that would reduce energy consumption at the end use or enhance opportunities for fuel switching. For consistency between scenarios, modelers are encouraged to develop assumptions that would lead to roughly a 20 percent decrease in final energy in 2050 relative to the low-tech, no-policy case. The precise assumptions are left to modeler’s choice.
CCSNo implementation of carbon capture and storage technology.CCS is available. The cost and performance characteristics are left to modeler’s choice.
NuclearPhase out of nuclear energy after 2010. Phase out is defined as no construction of new nuclear power plants beyond those already under construction or planned (excluding proposed plants). This reflects the concept of the ‘off’ case being triggered by public skepticism about nuclear technology. In addition, modelers are encouraged to assume no lifetime extensions beyond 60 years as representing an environment that generally discourages the development and deployment of nuclear energy.New nuclear energy is fully available. The cost and performance characteristics are left to modeler’s choice.
Wind & SolarPessimistic techno-economic assumptions for solar and wind energy are left to modeler’s choice. However, modelers should choose their assumptions carefully, with the goal of developing a scenario that represents evolutionary technology development.Optimistic techno-economic assumptions for solar and wind energy technologies are left to modeler’s choice. However, modelers should choose their assumptions carefully, with the goal of developing a scenario that represents plausibly optimistic technology development.
BioenergyThe pessimistic scenario should represent a scenario of bioenergy supply on the lower end of what is deemed sustainable. The precise assumptions are left to modeler’s choice. However, modelers should choose their assumptions carefully, with the goal of developing a scenario that represents evolutionary technology development.The optimistic scenario should represent a scenario of bioenergy supply on the higher end of what is deemed sustainable. The precise assumptions are left to modeler’s choice. However, modelers should choose their assumptions carefully, with the goal of developing a scenario that represents plausibly optimistic technology development.

    Other Resources Citing This Publication