2. Impacts on Electricity Generation and Key Fuel Markets

Reference Case Trends

Over the next 20 years the demand for electricity is projected to grow by 1.8 percent per year, as compared with the 1990s, during which electricity consumption grew by 2.3 percent annually. Growth in electricity use is expected to slow as new, more efficient appliances enter the market and industrial production continues to shift away from energy-intensive industries. With 3.0-percent annual growth projected for the economy as a whole, the overall electric intensity of the U.S. economy—measured as the ratio of electricity use to gross domestic product—is projected to decline by 22 percent between 2000 and 2020.

To meet the growth in demand for electricity, 357 gigawatts of new generating capacity is projected to be needed (Figure 4). The vast majority of new plants are expected to be natural gas fired, with lesser amounts of new coal-fired and renewable capacity. New natural gas-fired combustion turbines and combined cycle plants are the most economical options for most uses. They generally have lower capital costs than other options and they are becoming increasingly efficient.

With the addition of many new natural-gas-fired plants, the share of electricity generated from natural gas is projected to grow from 16 percent in 1999 to 34 percent in 2020 (Figure 5). Generation from coal-fired plants is also projected to grow as a small number of new plants are added and as existing plants are used more intensively, but the share of generation coming from coal is projected to decline slightly. On the other hand, generation from oil and from nuclear power is expected to decline as some older plants are retired in the later years of the forecast. Generation from renewable plants is projected to increase, but not enough to maintain its current share.

Although fossil fuel use is expected to grow over the next 20 years, SO2 and NOx emissions are not projected to be higher in 2020 than they are today (Figures 6 and 7). As a result of the emission reduction programs established in the Clean Air Act Amendments of 1990 (CAAA90) SO2 and NOx emissions are expected to be lower in 2020 than they were in 1999. For example, the CAAA90 cap on power sector SO2 emissions is set at 8.95 million tons for the years 2010 and beyond, and that cap is expected to become binding in the later years of the reference case projections as power companies exhaust their supplies of banked allowances.¹⁰ For NOx, 19 States in the Midwest and Eastern regions and the District of Columbia are projected to see significant reductions in emissions beginning in 2004, when a summertime emissions cap takes effect. The summer season cap begins in 2004 and is maintained throughout the rest of the projections. Total U.S. NOx emissions are projected to increase slightly after 2004, but not enough to offset the earlier reduction.

Hg emissions from electric power plants are projected to remain fairly steady between 2000 and 2020—hovering around 45 tons from 2005—despite the expected increase in coal use (Figure 8). Some existing coal plants are projected to add scrubbers to reduce SO2 emissions and selective catalytic reduction (SCR) equipment to reduce NOx emissions, and all new coal plants are projected to have scrubbers, SCRs, and fabric filters. While these technologies are designed primarily to reduce SO2, NOx, and particulate emissions, they also help to reduce Hg emissions. The addition of this equipment is expected to nearly offset the increase in Hg emissions that would be expected with increasing coal use.

Unlike NOx, SO2, and Hg emissions, CO2 emissions (expressed in metric tons carbon equivalent throughout this report) are projected to rise steadily over the next 20 years as the power sector becomes increasingly dependent on fossil fuels (Figure 9). Between 1999 and 2020, the share of electricity generation from fossil fuels is expected to increase from 69 percent to 80 percent, and CO2 emissions from electric power plants are expected to increase by 217 million metric tons—39 percent—over the next 20 years. The actions projected to be taken to reduce SO2 and NOx emissions in the reference case in response to the CAAA90 are not expected to reduce power sector CO2 emissions, because they will not lead to significant fuel switching.

Despite the growing demand for electricity, prices are expected to decline by 9 percent in real terms over the next 20 years (Figure 10). The phase-in of competition in many regions of the country is one factor in the expected decline, in addition to falling coal prices and the declining cost and increasing performance of new natural gas technologies.

Reducing Electricity Sector NOx, SO2, and Hg Emissions

A number of options are available to reduce power sector emissions of NOx, SO2, and Hg. They include emission control options such as adding combustion controls and SCR and selective noncatalytic reduction (SNCR) equipment designed primarily to reduce NOx emissions, flue gas desulfurization equipment (scrubbers) to reduce SO2, and activated carbon injection (ACI) equipment to reduce Hg.¹¹ Other options for reducing NOx, SO2, and Hg emissions include fuel switching (either by changing fuels at existing plants or by retiring plants and replacing them with plants that use different fuels) and reducing consumer demand.

In the cases examined in this report all three of the options above are expected to play a role; but by a large margin, the key strategy projected to be used is the installation of emissions control equipment to reduce the three emissions. As shown in Table 3, the amount of equipment projected to be added increases as the emission caps on the three pollutants are tightened. For example, scrubbers are projected to be added to 90 gigawatts of capacity by 2020 in the 50-Percent Reduction case and to 151 gigawatts in the 75-Percent Reduction case. The values in Table 3 indicate that emissions control equipment is expected to be added to many of the existing U.S. coal-fired electric power plants, which currently total just over 300 gigawatts of capacity. The percentage of existing coal-fired capacity expected to have SO2 scrubbers is larger than suggested by the values shown in Table 3, because 90 gigawatts of that capacity already is equipped with scrubbers.

The projections are similar for NOx emission controls: SCRs are expected to be added to 98 gigawatts of capacity in the 50-Percent Reduction case and to 218 gigawatts in the 75-Percent Reduction case. The investment in SCR technology increases continuously across the cases as the required percentage reduction increases. The same is true for expected additions of SNCRs between the 50-Percent and 65-Percent Reduction cases; but when the required reduction is raised to 75 percent, power suppliers are projected to shift increasingly to SCR technology because it can achieve greater NOx reduction.

Relative to the reference case, less capacity is expected to be retrofitted with NOx control technology in the 50-Percent Reduction case, because the 19-State summer season NOx cap and trade program that is scheduled to begin in 2004 in the reference case is replaced by the national cap and trade program in each of the analysis cases. In the 50-Percent Reduction case the annual NOx cap, 3.1 million tons (roughly equivalent to an average annual emission rate of 0.25 pounds per million Btu of fossil fuel consumed in 2010 and 0.18 pounds per million Btu in 2020), can be met with less control equipment than is required to meet the seasonal cap in the reference case. The 19-State summer season NOx emissions cap represented in the reference is based on a target average emission rate of 0.15 pounds per million Btu of fossil fuel consumed. The NOx emission caps in the 65-Percent and 75-Percent Reduction cases—2.2 million tons and 1.5 million tons, respectively—lead to average annual NOx emission rates below 0.15 pounds per million Btu by 2020.

In many other aspects—including fuel use, generation by fuel, and capacity additions by type—the results in the three analysis cases are similar to those in the reference case. As the emission caps are tightened across the cases there is a slight shift from coal-fired generation to natural-gas-fired generation (Figures 11 and 12). For example, in the 75-Percent Reduction case, which is projected to have the largest shift, natural-gas-fired generation in 2020 is expected to be 10 percent above and coal-fired generation 10 percent below the reference case levels. The shifts in the two other cases are smaller.

As the emission caps are tightened across the cases, the projected allowance prices for NOx, SO2, and Hg are expected to increase, particularly as the caps are lowered to the limits in the 75-Percent Reduction case (Table 4). In this case, the emissions controls must be added to units for which the marginal costs per unit of reduction are higher. This is particularly true for allowance prices for NOx and Hg. For example, the annual NOx allowance price¹² in 2020 in the 65-Percent Reduction case is projected to be $1,457 per ton, but in the 75-Percent Reduction case it is 94 percent higher, at $2,825 per ton. Similarly, the Hg allowance price in 2020 in the 65-Percent Reduction case is projected to be $41,190 per pound, but in the 75-Percent Reduction case it is more than twice as high, at $85,225 per pound.¹³ The requirements to reduce Hg have a significant impact on the SO2 allowance price, especially as the Hg emission caps are initially phased in. The SO2 allowance price in the 75-Percent Reduction case in 2010 is lower than in the 65-Percent Reduction case, because efforts to meet the tighter Hg emissions limit in the 75-Percent case also reduce SO2 emissions.

The increasing cost of allowances across the cases is driven by several factors. For example, for a particular plant, the plant size, sulfur content of the coal used, and plant capacity factor are important in determining the cost of reducing SO2. Smaller plants are in general more costly (per unit of capacity) to retrofit with scrubbers than are larger plants. It is also more expensive on a per ton removal basis to control SO2 at a plant using relatively low-sulfur coal. Similarly, it is more expensive on a per ton removal basis to add a scrubber to a plant that is not used intensively. For example, for a large plant with scrubber capital costs of $200 per kilowatt, using a 2-percent sulfur coal and operating at a 75-percent capacity factor, the cost of removing SO2 is expected to be approximately $250 per ton. If the plant used a 1-percent sulfur coal the cost estimate would double, and if it operated at a 37.5-percent capacity factor the cost estimate would double again. For a smaller plant, with scrubber capital costs that could be $400 per kilowatt or more, the corresponding SO2 removal costs would be even higher. As a result, when controls must be added to smaller plants that are already using relatively low-sulfur coals and operating less intensively, the per ton costs of removal can be quite high.¹⁴

Investments in emissions control technology, combined with higher expenditures for natural gas, are projected to lead to higher supplier resource costs in the three emission reduction cases (Table 5). Supplier resource costs include electricity producers’ expenditures on fuel, nonfuel operations and maintenance costs, and investments in new plants and emissions control equipment. In the 75-Percent Reduction case, suppliers are projected to incur $89 billion (constant 1999 dollars) more in resource costs between 2001 and 2020 than in the reference case (Figure 13). On an average annual basis, the increases in resource costs in the three cases average $1.4 billion, $3.3 billion, and $4.4 billion, in the 50-Percent, 65-Percent, and 75-Percent Reduction cases, respectively.¹⁵

Changes in electricity prices are expected to parallel the changes in supplier resource costs in the three analysis cases (Figure 14). In percentage terms, electricity prices in 2010 are expected to range between 0 and 2 percent higher than in the reference case; and in 2020, as the emission caps tighten, they are expected to range between 2 and 6 percent higher. On an average annual basis, the projected increases in electricity revenues (prices times sales) relative to the reference case in 2020 are $4 billion, $9 billion, and $14 billion in the 50-Percent, 65-Percent, and 75-Percent Reduction cases, respectively.

Offsetting CO2 Emissions Growth After 2008

Because of the slight shift from coal-fired to natural-gas-fired generation, reducing power sector NOx, SO2, and Hg emissions is projected to have some impact on CO2 emissions (Figure 15). In 2010, CO2 emissions in the analysis cases are projected to be between 14 million and 33 million metric tons below the level expected in the reference case. (The projections for CO2 emissions are lower in the more stringent cases, because the expected shifts from coal to natural gas are larger.) In 2020, the range is slightly wider, between 12 million and 36 million metric tons. Even with these reductions, however, power sector CO2 emissions in 2020 are projected to be between 262 million and 286 million metric tons (between 55 and 60 percent) above the 1990 level.

The potential exists for an increase in the use of coal and in its associated emissions in sectors of the economy (i.e., residential, commercial, and industrial) not covered by emission cap programs. However, because coal plays such a small role in those sectors and the projected decreases in coal prices are generally expected to be less than a few percent, the potential for emission “leakage” appears slight.¹⁶ The increase in natural gas prices that is projected to occur because of increased use in the electricity sector appears to be more significant, leading to lower overall fuel consumption and lower emissions in the non-electricity sectors.

If a cap is imposed on power sector CO2 emissions at the projected 2008 reference case level of 672 million metric tons (197 million metric tons or 41 percent above the 1990 level), power suppliers will have to either take action to reduce their emissions or purchase offsets for between 65 million and 89 million metric tons by 2020 (Figure 16). (Again, fewer offsets are required in the more stringent cases, because the expected shifts from coal to natural gas as a result of the other emission caps are larger.) Note that no offsets are required until projected CO2 emissions in each of the three analysis cases exceed the assumed CO2 cap (the 2008 level expected in the reference case), which is projected to occur in 2010 in the 50-Percent Reduction case, in 2012 in the 65-Percent Reduction case, and in 2013 in the 75-Percent Reduction case.

To determine the prices that U.S. power suppliers might be willing to pay for offsets, the three analysis cases were rerun with CO2 emissions capped at the 2008 reference case level (Figure 17). The projected allowance prices in 2020 range between $33 and $54 per metric ton. As compared with earlier studies of the expected costs to the U.S. power sector of meeting the Kyoto Protocol requirements, these allowances prices are quite low; however, the CO2 emissions cap assumed in this analysis (41 percent above the 1990 level) is very different from the U.S. target specified in the Kyoto agreement (7 percent below the 1990 level). The key CO2 compliance strategy in these cases is expected to be a further shift from coal to natural-gas-fired generation. For example, in the 75-Percent Reduction case with no CO2 cap, coal-fired generation in 2020 is projected to be 10 percent below the reference case level, and natural-gas-fired generation is projected to be 10 percent above the reference case level. In the 75-Percent Reduction case with a CO2 cap set to the 2008 reference case level, the impacts are approximately doubled.

Because of the reduced reliance on coal projected in the cases with CO2 caps, the investments in NOx, SO2, and Hg control equipment are projected to be lower. For example, scrubbers are projected to be added to nearly 152 gigawatts of capacity in the 75-Percent Reduction case without a CO2 cap, as compared with only 115 gigawatts when the CO2 cap is incorporated. In the early years of the projections, the expected investments in control equipment to reduce emissions of NOx, SO2, and Hg in the cases with and without CO2 caps are similar; but they are much lower in the later years, when CO2 emission caps are imposed. The projected allowance prices for NOx and Hg are also lower when the CO2 emissions cap is included.

The projected CO2 allowance prices in the cases with CO2 caps represent the marginal cost of compliance within the U.S. power sector. They also represent the maximum price that power suppliers would be willing to pay for offsets. They would incur these costs only if they could not purchase offsets at a lower price. The price that U.S. power suppliers might have to pay to offset increases in CO2 emissions above the 2008 reference case level is difficult to determine, because it would depend on what the rest of the world did in response to any greenhouse gas emissions reduction agreement. It would also depend on how offset programs were defined, implemented, and verified.

The National Energy Modeling System does not represent energy or non-energy markets outside the United States, and EIA has not made an independent assessment of how world offset markets might evolve. Figure 18 shows world energy sector CO2 abatement supply curves (the upward sloping curves) produced by the Pacific Northwest Laboratory’s Second Generation Model (SGM), matched against the projected requirement for reductions if all countries complied with the Kyoto Protocol (the vertical lines).¹⁷

The supply curves in Figure 18 represent the projected CO2 emission reductions (abatement) from reference case projections that would occur in the energy sectors of all countries in response to rising prices for carbon allowances. Because worldwide trading is assumed, all countries—including those without greenhouse gas reduction targets in the Kyoto Protocol—are included in the supply curves, assuming full compliance with the Kyoto Protocol. Countries with greenhouse gas reduction targets can trade with other countries by using the Protocol’s clean development mechanism or joint implementation provisions. For example, if an Annex I country made investments that led to lower greenhouse gas emissions in China, the reductions would be counted toward the investing country’s reduction target. The estimates include offsets created in each country’s energy sector but exclude offsets that might be available from non-energy activities, such as changes in agricultural practices and reforestation activities. The reductions would be expected to come from numerous sources, including changes in fuel use, improvements in production efficiency (more efficient power plants), and reductions in consumer energy use.

The demand curves represent the estimated reduction in carbon emissions required by Annex I countries to reach full compliance with the Kyoto Protocol. The United States is included in both the supply and demand curves in Figure 18, assuming full U.S. compliance with the Protocol. The intersections of the lines of the same color represent the prices at which the market for energy sector offsets would clear in 2010, 2015, and 2020. Both the supply of offsets and the demand for them are projected to grow over time as a result of expected economic growth and changing technologies.

Because this study does not assume U.S. participation in the Kyoto Protocol, adjustments were made to remove the U.S. contribution from the demand curves in Figure 18. Without U.S. participation in the Protocol, the demand for offsets would be much lower than depicted in Figure 18. For example, if the rest of the world complied with the Protocol while the United States did not, the world trading price for energy sector carbon allowances would be fairly low—rising from just a few dollars in 2010 to roughly $5 in 2015 and $8 in 2020 (Figure 19). The supply curves in Figure 19 are the same as those in Figure 18, but the demand curves have been shifted to the left (lowered), because the U.S. carbon reduction requirement has been removed. The price would rise slightly as U.S. power suppliers entered the market to purchase the 65 to 89 million metric tons of offsets they would need; however, assuming a price of roughly $10 per metric ton in 2020, the total cost of offsets for U.S. power suppliers would be between $654 million and $888 million in the three analysis cases.

The net result of the these estimates is that if power suppliers are required to purchase offsets for any CO2 emissions above the level projected to be emitted in 2008 in the reference case, their costs in 2020 could rise by as little as $654 million or by as much as $888 million. The range in cost estimates results from the differences in offsets required in the three cases (between 65 million and 89 million metric tons carbon equivalent in 2020). The prices and costs could be lower if offsets from other greenhouse gases or carbon sinks were available. The analysis above is predicated on the assumption that the regional abatement costs curves provided by the SGM model are rreasonable estimates. Because we have no way to verify their reasonableness, that assumption increases the uncertainty of the cost estimates.