In the EIA analysis released in October 1998, it was assumed that the carbon targets would be achieved on average from 2008 to 2012, the first commitment period in the Kyoto Protocol. Carbon reductions would be phased in beginning in 2005, with approximately one-fourth of the total reduction achieved in 2005, one-half in 2006, three-fourths in 2007, and the full level achieved in the commitment period. The phase-in period would allow for a more gradual transition of energy markets, and it is consistent with the requirement in the Protocol that countries must make demonstrable progress toward their commitments by 2005. It was also assumed that some energy sectors would begin to respond before 2005 in anticipation of the coming changes in energy prices. Any carbon reductions achieved before 2008 would not, however, count toward compliance.

At the request of the U.S. House of Representatives Committee on Science, EIA has analyzed the impacts of an earlier start date for carbon emissions reductions. In this analysis, the carbon emissions targets are assumed to be phased in beginning in 2000, instead of 2005, reaching the emissions target during the commitment period 2008 through 2012. Again, carbon reductions achieved before 2008 do not count toward compliance, and the early start date does not imply any change in the timing or level of the U.S. commitments under the Kyoto Protocol. Assuming U.S. compliance with the Kyoto Protocol, the analysis investigates whether binding limits on carbon emissions, imposed through a carbon price, would be less costly in terms of overall economic impacts if the limits began in 2000 rather than in 2005. Only three of the six cases from the 1998 study—1990+24%, 1990+9%, and 1990-7%—are included in this analysis, representing the most stringent (1990-7%), least stringent (1990+24%), and intermediate (1990+9%) cases.

It is possible that an earlier start and more gradual phase-in of the carbon emissions targets could lower the carbon price required to achieving the specified targets in the commitment period. Starting carbon reductions earlier could induce consumers who are replacing energy-using equipment over the next decade to purchase more efficient, less carbon-intensive equipment than they might otherwise buy, making it less costly to meet carbon emissions targets in the commitment period. Lower carbon prices would have less negative impact on the economy in the commitment period, and a smoother adjustment of the economy to higher energy prices would be possible.

In order to achieve the early start date, carbon prices are imposed starting in 2000, raising energy prices and incurring macroeconomic costs earlier than projected in EIA’s 1998 study. Although the carbon price leads to higher energy prices before 2005 in the early start cases, it is assumed that consumers purchasing energy-using equipment will continue to rank energy prices along with other preferences in the same manner as in the cases with the 2005 start date. The analysis does not assume changes in consumer behavior as a result of additional information programs, voluntary early reduction programs, partnerships, or regulations; however, consumer behavior does respond to higher energy prices induced by the carbon price.

The imposition of carbon prices before 2005 reduces the demand for energy services and improves the efficiency and carbon intensity of the stock of energy-using equipment. Energy service demand is lower as a result of both the direct effect of energy prices (for example, as consumers adjust thermostats and reduce travel), and the indirect effect of higher energy prices on the economy (reductions in freight and air transport and industrial output). Three primary effects are expected to produce improvements in the efficiency and carbon intensity of the stock of energy-using equipment. First, retirements are accelerated for some equipment, such as less-efficient industrial equipment and oil and coal-fired generating capacity. Second, with higher fossil fuel prices, there is an economic incentive to invest in more efficient equipment, purchase smaller light-duty vehicles⁷ with lower horsepower, and shift to natural gas or renewable generation capacity. Third, higher prices lead to the acceleration of technology improvements prior to 2005.

For most parts of the energy-consuming sectors—residential, commercial, industrial, and transportation—it is assumed that the evolution of technology is dependent on the passage of time as more advanced, more efficient, and/or lower-cost equipment becomes available to consumers. An earlier start date for carbon reductions and higher energy prices prior to 2005 will increase the demand for more energy-efficient equipment. Consequently, efficiency improvements for existing equipment that are included in the reference case are assumed to accelerate in the 2000-2005 time period, as manufacturers respond to consumer demand for efficiency. For example, in the residential and commercial sectors, more energy-efficient models of current technologies that are available in the reference case in 2005 are assumed to be available in 2001 (although the equipment still must be economical to be adopted). The acceleration of technology improvements in the early start cases is not assumed to continue after 2005, because carbon prices decline in the commitment period. In the electricity generation sector, the evolution of technology is influenced by the rate of adoption, with the costs of new generation technologies declining as the technologies penetrate. For example, the earlier adoption of renewable generating technologies reduces the costs of the technologies so that there is a greater economic incentive for their adoption.

Both the earlier EIA study and this analysis use a carbon price as the primary instrument for change in energy markets. Energy consumers in the residential, commercial, industrial, and transportation sectors are assumed to respond to changes in energy prices as they occur without anticipating future price increases. In the electricity generation sector, however, future price increases are anticipated when capacity expansion decisions are made. Similarly, automobile manufactures factor future increases in the price of gasoline into the decision to offer more efficient vehicles. Capacity expansion decisions by electricity generators, refiners, and natural gas pipelines also incorporate anticipated growth in the demand for electricity, petroleum products, and natural gas.

Even if carbon emissions reductions are imposed earlier, the carbon emissions targets do not change after 2008. As a result, carbon emissions in the first commitment period are essentially the same whether the phase-in of reductions begins in 2000 or 2005. Due to changes in energy-using equipment before 2005 in the early start cases, the equipment stock is likely to be less carbon intensive in the commitment period. As a result, the demand for energy services may be higher while still meeting the carbon targets. A key question is whether the early start date encourages sufficient investment in more energy-efficient and less carbon-intensive technologies to reduce the cost of compliance and the overall macroeconomic impacts significantly in the commitment period.

Carbon Prices

In general, an earlier start date for reducing carbon emissions has the immediate impact of improving energy efficiency and encouraging fuel switching, resulting in reductions in energy consumption and carbon emissions before 2005. The longer phase-in results in a more gradual reduction in carbon emissions than projected with a start date of 2005 and an easier transition to the targets within the commitment period. Because of the changes in energy efficiency and fuel mix that occur before 2005 with the earlier start date, the average and marginal costs of compliance are reduced in the commitment period (Figures 7 and 8).

Figure 7. Average Projected Carbon Prices, 2008-2012  [source]

Figure 8. Average Projected Carbon Prices and Annual Carbon Emission Reductions, 2008-2012  [source]

1990+24% Carbon Emissions Target

Because the overall target for reductions in domestic energy-related carbon emissions is relatively low in the 1990+24% case, an earlier phase-in results in relatively small emissions reductions between 2000 and 2004 (Figure 9). Early in the projection period, energy consumption is 2.2 quadrillion British thermal units (Btu) lower in 2005 in the early start case than in the 1990+24% case, in part because of efficiency improvements (Figure 10). By 2010 and beyond, energy consumption is virtually the same in the two cases (Tables 1 and 2). Electricity generators adjust to a slightly lower demand for electricity early in the period by reducing coal-fired generation; however, the overall fuel mix is not significantly changed with the earlier start date.

Figure 9. Projected Carbon Emissions in the 1990+24% and 1990+24% Early Start Cases, 1998-2020  [source]

Figure 10. Projected U.S. Energy Intensity in the 1990+24% and 1990+24% Early Start Cases, 1998-2020  [source]

1990+9% Carbon Emissions Target

Earlier reductions have a larger impact on carbon emissions in the 1990+9% case than in the 1990+24% case (Figure 12). Total energy consumption in 2005 in the early start case is about 3.9 quadrillion Btu lower than in the 1990+9% case, and energy intensity is also lower (Figure 13). The difference in consumption essentially disappears by 2008, the beginning of the first commitment period.

Figure 12. Projected Carbon Emissions in the 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

Figure 13. Projected U.S. Energy Intensity in the 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

In response to the carbon prices, electricity generators use less coal and more natural gas before 2005 in the early start case. Because electricity generators are assumed to anticipate future prices, some changes in capacity additions begin almost immediately in the projection period. In 2005, coal consumption by electricity generators is about 3.0 quadrillion Btu lower in the early start case than in the 1990+9% case, and consumption of natural gas by generators is about 1.0 quadrillion Btu higher. Efficiency improvements reduce the consumption of natural gas by end-use consumers, however, so that total natural gas consumption in 2005 is only about 0.5 quadrillion Btu higher in the early start case. As was seen for the 1990+24% case, earlier carbon reductions in the 1990+9% case have little impact on the economics of nuclear plant life extensions. Also, the use of renewable energy is essentially unchanged early in the projection period, implying that increased use of natural gas is more cost-effective than renewables before 2005, even with higher prices for fossil fuels.

As a result of lower energy consumption and a shift from coal to natural gas before 2005, carbon prices in the 2008 to 2012 period are lower in the early start case. In 2010, the carbon price is $149 per metric ton in the early start case, compared with $163 per metric ton in the 1990+9% case (Figure 14). Carbon prices average $146 per metric ton between 2008 and 2012 in the early start case, compared with $159 in the 1990+9% case. From 2000 through 2020, average carbon prices are about $124 per metric ton in the early start case and $110 per metric ton in the 1990+9% case, with total carbon reductions of 6,259 million metric tons and 5,596 million metric tons, respectively, in the two cases.

Figure 14. Projected Carbon Prices in the 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

1990-7% Carbon Emissions Target

Because the 1990-7% case requires larger reductions in carbon emissions (Figure 15), the impact of an earlier start date is more significant before 2005 than it is in the 1990+24% and 1990+9% cases. In 2005, total energy consumption in the early start case is about 8.4 quadrillion Btu lower than in the 1990-7% case, and energy intensity is also reduced (Figure 16), although total consumption is the same in 2008 when the carbon reductions in the two cases are the same. Total energy consumption is as much as 1.7 quadrillion Btu higher after 2008 in the early start case, with the difference between the two cases narrowing later in the forecast period. Higher use of renewable sources, primarily for electricity generation, changes the fuel mix and allows the carbon reduction targets to be met later in the period, despite higher demand for petroleum in the transportation sector and higher total energy consumption.

Figure 15. Projected Carbon Emissions in the 1990-7% and 1990-7% Early Start Cases, 1998-2020  [source]

Figure 16. Projected U.S. Energy Intensity in the 1990-7% and 1990-7% Early Start Cases, 1998-2020  [source]

Before 2005, electricity generators shift from coal to natural gas and begin to increase the use of renewable energy slightly. In 2005, natural gas consumption by electricity generators is 10.3 quadrillion Btu in the early start case, compared with 7.1 quadrillion Btu in the 1990-7% case, and coal consumption is reduced to 7.5 quadrillion Btu compared to 15.1 quadrillion Btu. By 2008, the differences between coal and gas consumption in the two cases lessen. The use of renewable energy for electricity generation is higher by 0.2 and 0.6 quadrillion Btu in 2005 and 2008, respectively, in the early start case, and the earlier adoption of renewables tends to encourage their continued penetration. As a result, renewable energy consumption for electricity generation is as much as 1.0 quadrillion Btu higher in 2015 in the early start case.

Because of the shift to natural gas for electricity generation, total natural gas consumption is 1.1 quadrillion Btu higher in 2001 in the early start case than in the 1990-7% case. To satisfy demand, domestic natural gas production exceeds 21 quadrillion Btu in 2001 and continues to grow at a rate of nearly 1 quadrillion Btu a year through 2005. Some additional pipeline capacity is required in the early start case, but the pace of capacity expansion is gradual enough to assure that the needed lead times can be met. In 2000, 2001, and 2002, small amounts of additional capacity—0.03, 0.20, and 0.45 trillion cubic feet, respectively—are added. Pipeline capacity increases are greater after 2002, but the industry has the 2- to 3-year lead time necessary for the expansion. In the early years, the utilization of existing capacity is slightly higher in the early start case than in the 1990-7% case.

In the early years, the increases in natural gas production in the early start case—spread over both onshore and offshore sources—appear achievable, inasmuch as similar increases have occurred in the past. In order to meet the higher production levels, wellhead prices are higher in the early start case by as much as $0.33 per thousand cubic feet in 2005; however, the difference in wellhead prices between the two cases diminishes by 2012, and prices are lower in the early start case through the rest of the forecast horizon, in part because natural gas consumption is lower. The higher production in the early years is achieved without a significant increase in drilling, because sufficient reserves are available. Natural gas reserves are lower through 2006 in the early start case but higher later in the forecast as a result of increased drilling and reserve additions. Although such increases could strain the industry, the early start date may prove to be beneficial to consumers in the longer term. End-use prices for natural gas are higher in the early start case than in the 1990-7% case before 2010 but lower after that.

In the 1990-7% early start case, the carbon price in 2010 is reduced by $32 per metric ton to $316 per metric ton from $348 per metric ton in the 1990-7% case with the 2005 start date (Figure 17). Average carbon prices between 2008 and 2012 in the 1990-7% early start case are $310 per metric ton, compared with $349 in the 2005 start case. From 2000 through 2020, average carbon prices are about $254 per metric ton in the early start case and $231 per metric ton in the 2005 start case, and total carbon reductions of 10,113 million metric tons and 8,758 million metric tons, respectively, are achieved in the two cases.

Figure 17. Projected Carbon Prices in the 1990-7% and 1990-7% Early Start Cases, 1998-2020  [source]

The implementation of a carbon permit system will have two effects on the aggregate economy. The carbon permit price will increase the price of energy, resulting in an increase in prices for goods and services. Also, the process of auctioning emissions permits will raise large sums of money, and if permits are also purchased from other countries as assumed in the 1990+9% and 1990+24% cases, there will be both domestic and international payment flows. This analysis assumes that the Federal Government will return the domestic portion of the carbon permit payments to households through a lump sum personal income tax rebate.

Carbon Permit Payments

An earlier start date will affect the time profile of the permit payments but, in general, will not have a significant impact on the total cumulative payments over the entire forecast horizon (Figures 18, 19, and 20). In all cases, the profile of the payments is similar to the carbon price profile. In the commitment period, 2008 through 2012, and beyond to 2020, the carbon emissions targets are identical in the early start and 2005 start cases; consequently, any differences in payments after 2007 are directly attributable to moderations in the carbon price. In all three carbon reduction cases, the early start date changes the carbon price profiles by moderating both the peak prices and the average carbon prices over the 2008 through 2012 period.

Figure 18. Total Projected U.S. Payments for Domestic and International Carbon Permits in the 1990+24% and 1990+24% Early Start Cases, 1998-2020  [source]

Figure 19. Total Projected U.S. Payments for Domestic and International Carbon Permits in the 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

Figure 20. Total Projected U.S. Payments for Domestic and International Carbon Permits in the 1990-7% and 1990-7% Early Start Cases, 1998-2020  [source]

In the 1990+24% early start case, projected payments (in nominal dollars) rise to $67 billion in 2005. By 2010, projected payments rise to $154 billion in the early start case, compared with $164 billion in the case with the 2005 start date. Through 2020, payments in both 1990+24% cases rise steadily, with payments in the early start case slightly below those in the case with the 2005 start date. Total cumulative payments between 2000 and 2020 are similar in the two cases, however, at $3,445 billion in the early start case and $3,292 billion in the 2005 start case—a difference of 4.6 percent. In the early start case, slightly lower payments in the post-2005 period do not offset the additional collections between 2000 and 2005.

In the 1990+9% cases there are larger differences with an early start date in the commitment period and beyond. Payments in the early start case mirror the carbon price profile, and they are consistently below payments in the case with the 2005 start date in the commitment period and beyond. Cumulative payments in the 1990+9% early start case total $6,080 billion, compared with $5,676 billion in the case with the 2005 start date, a difference of 7.1 percent. As was seen in the 1990+24% cases, the lower payments in the post-2005 period in the early start case do not offset the additional payments between 2000 and 2005. In the 1990-7% early start case, projected cumulative payments total $10,558 billion, compared with $10,234 billion in the 1990-7% case with the 2005 start date—a 3.2-percent difference.

Impacts on the Aggregate Economy

In the early start cases, the different carbon price and permit payment profiles change both the timing and the magnitude of the impacts on the aggregate economy for two related reasons. First, the early start cases reflect a slower ramping of both the carbon price and payments relative to the cases with a 2005 start date. Also, in the commitment period of 2008 through 2012 and beyond to 2020, the carbon prices and payments are slightly lower in the early start cases. The 1998 EIA analysis of the Kyoto Protocol focused attention on potential GDP and actual GDP to measure economic impacts. The loss of potential GDP is a measure of the loss in productive capacity of the economy directly attributable to the reduction in energy resources available to the economy. The loss in actual GDP incorporates the adjustment cost to the economy and reflects short-term economic dislocations of capital and labor that may result from higher energy prices. Both are measured in constant 1992 dollars.

In the 1990+9% early case, there is a loss in potential GDP beginning in the year 2000, progressing slowly through 2010 (Figure 21). In the 1990+9% case with the 2005 start date, the movement in potential GDP is more rapid. The difference in time profiles between the two cases is directly attributable to the difference in energy resources available to the economy as the system adjusts to the target level of carbon emissions for the commitment period, 2008 through 2012. Once the carbon emissions target has been reached, however, the two cases merge, and the rates of decline in carbon emissions and energy use are similar. Potential GDP, which is tied to the level of energy availability and use in the economy, then takes on the same path in both the 1990+9% cases. The ultimate impact of the early start on potential GDP is small. By 2010, potential GDP declines by $33 billion in the 2005 start case and by $30 billion in the early start case. By 2020, the loss in potential GDP is $39 billion with the 2005 start date and $42 billion with the 2000 start date. The pattern of results is similar in the 1990+24% and 1990-7% cases (Table 3).

Figure 21. Projected Dollar Losses in Potential and Actual U.S. Gross Domestic Product in the 1990+9% and 1990+9% Early Start Cases Relative to the Reference Case, 1998-2020  [source]

There are also differences between the early start and 2005 start cases in the impacts of the transition on the economy, which are reflected in actual GDP. In the three early start cases, the economy experiences a loss in GDP beginning in 2000; however, the early start date smooths the transition of the economy to the longer run target (Figures 22, 23, and 24). The largest portion of the adjustment loss occurs in roughly the first 5 years after the imposition of the permit system, whether the start date is 2000 or 2005. In general, the loss in actual GDP between 2000 and 2005 in the early start cases is between one-half and nearly three-quarters of the loss between 2005 and 2010 in the cases with the 2005 start date. In the early start cases, actual GDP begins to rebound back toward the reference case level sooner, and the recovery is smoother, than in the cases with a 2005 start. By 2010, the GDP impacts in the 1990+24% early start case are about half those in the case with the 2005 start date. In the 1990+9% and 1990-7% cases, the GDP impacts with the early start date are about one-third of those with the 2005 start date. Ultimately, in all cases, the economy transitions into a long-run path and the losses in actual and potential GDP become very close by 2020.

Figure 22. Projected Dollar Losses in Actual Gross Domestic Product in the 1990+24% and 1990+24% Early Start Cases Relative to the Reference Case, 1998-2020  [source]

Figure 23. Projected Dollar Losses in Actual Gross Domestic Product in the 1990+9% and 1990+9% Early Start Cases Relative to the Reference Case, 1998-2020  [source]

Figure 24. Projected Dollar Losses in Actual Gross Domestic Product in the 1990-7% and 1990-7% Early Start Cases Relative to the Reference Case, 1998-2020  [source]

The effects on the economy can be considered from three perspectives: (1) impacts during the first 5 years, (2) impacts measured in 2010 (in the middle of the commitment period) and 2020 (at the end of the forecast period), and (3) the cumulative and net present value of the impacts from 2000 through 2020.

The First Five Years

In the 1990+24% early start case, the projected loss in actual GDP after 5 years totals $47 billion (constant 1992 dollars). By comparison, in the 1990+24% 2005 start case, the actual GDP loss after 5 years is $97 billion. The difference can also be seen in the growth rate of the economy over the first 5 years (Figure 25). In the 1990+24% 2005 start case, the GDP growth rate between 2005 and 2010 is reduced by 0.2 percentage point relative to the projected growth rate in the reference case. In the early start case, however, the growth rate between 2000 and 2005 is reduced by only 0.1 percentage point relative to the reference case. In the 1990+9% early start case, the loss in actual GDP after 5 years is $136 billion, compared with $189 billion in the 1990+9% 2005 start case. The 5-year average GDP growth rate between 2000 and 2005 is reduced by 0.3 percentage point in the early start case and by 0.4 percentage point in the 2005 start case relative to the reference case. Although the 1990-7% cases have considerably larger impacts on the economy, the same trends are seen in the cases with different start dates. The loss in actual GDP between 2000 and 2005 in the 1990-7% early start case is $253 billion, compared with $398 billion between 2005 and 2010 in the 2005 start case, and the corresponding reductions in GDP growth rate are 0.6 and 0.9 percentage point relative to the reference case.

Figure 25. Projected Five-Year Average GDP Growth Rates in the Early Start Cases, 2000-2005, and in the Kyoto Protocol Analysis Cases, 2005-2010  [source]

Impacts in 2010 and 2020

When the impact is measured in the year 2010, in the middle of the commitment period, the loss in actual GDP for the 1990+24% early start case relative to the reference case is $44 billion, as compared with $97 billion in the 2005 start case (Table 4). Thus, in 2010, the GDP impact in the 1990+24% early start case is 45 percent of the impact in the 2005 start case. For the more stringent carbon reduction cases, the effects of an early start date are even more pronounced: actual GDP losses of $66 billion and $117 billion in 2010 are projected for the 1990+9% and 1990-7% early start cases, respectively, compared with projected losses of $189 billion and $398 billion in 2010 in the corresponding 2005 start cases. The 2010 GDP losses in the two early start cases are therefore only 35 percent and 29 percent of those in the 2005 start cases. Smaller GDP losses are also seen in 2020 for the early start cases, but the differences from the 2005 start cases are not nearly as significant (Table 4).

As indicated above (Figures 22, 23, and 24), the impacts on actual GDP begin to merge with the path of potential GDP by 2020, and the difference between the early start and 2005 start cases narrows. In the 1990+24% cases, the projected loss in actual GDP in the early start case is $45 billion in 2020, compared with $50 billion in the 2005 start case. In the 1990+9% cases, the loss in actual GDP in the early start case is $62 billion, compared with $68 billion in the 2005 start case. In the 1990-7% cases, the projected losses are $52 billion in the early start case and $82 billion in the 2005 start case.

Cumulative and Net Present Value Impacts

For each of the three carbon reduction targets examined in this analysis, the early start and 2005 start cases project roughly the same undiscounted values for the cumulative impact on GDP from 2000 through 2020 (Table 5). In the 1990+24% early start case, the cumulative loss in actual GDP totals $937 billion, compared with $975 billion in the 1990+24% case. Similarly, the projected cumulative GDP losses in the 1990+9% cases are $1,560 (early start) and $1,573 billion (2005 start), and in the 1990-7% cases the projected cumulative losses are $2,475 billion and $2,631 billion, respectively.

Although the cumulative impacts on GDP are similar for the early start and 2005 start cases, the early start cases do involve a tradeoff. The peak impacts are less severe in the early start cases, but they are felt earlier. A net present value calculation takes into consideration the time value of money. Using a discount rate of 7 percent beginning in 2000, the cumulative discounted impacts are larger in the early start cases. In the 1990+24% early start case, the net present value of the cumulative loss in actual GDP is $439 billion, compared with only $404 billion in the 1990+24% 2005 start case. For the 1990+9% cases the projected net present value losses are $846 billion and $750 billion, and for the 1990-7% cases they are $1,430 billion and $1,285 billion in the early start and 2005 start cases, respectively.

Summary

In summary, four primary effects characterize the impacts of the earlier start date:

• The early start date affects the time profile of the carbon permit payments but, in general, does not have a significant impact on cumulative payments over the entire time period.

• With the early start, the peak impact on the economy is smaller because the carbon price profile is more gradual and the peak carbon price is lower.

• Although the peak impact is smaller with the early start, the economy begins the adjustment 5 years earlier and incurs GDP losses earlier in the forecast.

• The economy rebounds more smoothly over time under the early start assumptions.

Early Start Assumptions

In the reference case, more efficient and advanced technologies for the buildings sector (residential and commercial) become available to consumers over time, reflecting the ongoing development of technology. Because an early start date increases the demand for more energy-efficient equipment, the early start cases assume that more efficient models of current technologies, available in the reference case in 2005, will be available in 2001. Efficiency improvements for existing equipment are assumed to accelerate as manufacturers respond to increased consumer demand for more efficient products.⁸

The relatively slow rate of growth in new buildings dampens the growth of energy consumption and carbon emissions; however, it also slows the impact of more efficient equipment on total stock efficiency. For example, the average useful life of commercial natural-gas-fired space heating equipment is at least 20 years (a conservative estimate, inasmuch as the average life of a natural gas boiler is closer to 25 years). Heating systems typically are purchased only for new construction, for major renovations, or when an existing system needs replacement. An earlier start date in the 1990-7% case results in an improvement of 2.4 percent in the 2005 projected average stock efficiency of commercial gas heating equipment relative to the 1990-7% 2005 start case. By 2005, however, about three-fourths of the commercial sector gas heating equipment that was purchased before 2000 is still expected to be in use. As a result, high price signals continue to be required after 2005 to encourage the purchase of more energy-efficient equipment as older equipment is replaced.

An earlier start date for emissions reductions, earlier availability of more energy-efficient appliances, and higher delivered energy prices encourage the more rapid adoption of more efficient technologies for most products through 2004. As the assumed technology menus in the early start and 2005 start cases converge in 2005, the difference in delivered prices between the cases determines the efficiency level of newly purchased equipment. In the residential sector, purchased equipment efficiencies in 2010 are higher in the 2005 start cases than in the early start cases, because delivered energy prices are higher (Table 6). In the commercial sector, the penetration of more efficient equipment in the 2005 start cases approaches that in the early start cases by 2010, although the effects of the pre-2005 purchases in the early start cases are still evident (Table 7).

Early Start Assumptions

Analysis Results

In each of the three early start carbon reduction cases, the 2000 start date for carbon prices results in lower projected cumulative industrial output from 2000 through 2020 than in the corresponding 2005 start case. The impacts of the early carbon prices on output cannot be overcome by the relatively modest reductions in carbon prices in later years.

There are countervailing forces in the industrial sector. Earlier carbon reductions imposed by means of a carbon price have a negative impact on the overall economy, which leads to reduced domestic industrial output and reduces the incentive to invest in more energy-efficient equipment. As a result, although new equipment installed before 2005 in the early start cases tends to be more energy efficient than in the 2005 start cases, less investment takes place because economic growth is lower. In the commitment period, 2008 to 2012, industrial growth is larger in two of the three early start cases because of the lower carbon prices in that period. Therefore, although the accelerated retirements and the addition of more energy-efficient equipment that occur with an earlier start date help to reduce energy consumption, the reduction is at least partially offset by higher economic growth and industrial output.

As an example, in the 1990-7% early start case, higher equipment retirement rates lead to a 5-percent reduction in the stock of existing equipment in the cement industry by 2008 relative to that in the 1990-7% case. The higher rate of stock turnover, combined with the adoption of more energy-efficient equipment, results in a 3.6-percent decrease in energy intensity in the early start case in 2008. However, the output of the cement industry is 6.4 percent higher in 2008 in the early start case because carbon prices are lower. Consequently, energy consumption in 2008 in the cement industry is 2.5 percent higher in the 1990-7% early start case than in the 1990-7% case with the 2005 start date.

1990+24% Carbon Emissions Target

From 2000 through 2005, industrial output is $294 billion lower and energy consumption is 3.5 quadrillion Btu lower in the 1990+24% early start case than in the 1990+24% case. Industrial output and energy consumption from 2008 through 2012 in the early start case do not exceed the levels in the 2005 start case—in contrast to the results for the other carbon reduction cases. As a result, the relative impact on cumulative industrial output from 2000 through 2020 is almost as large for the 1990+24% early start case as for the 1990-7% early start case. Cumulatively, from 2000 through 2020, industrial output is $750 billion lower in the 1990+24% early start case than in the 2005 start case, and energy consumption is 7.4 quadrillion Btu lower. Cumulative expenditures for energy consumption are only $15 billion higher (discounted at 7 percent) in the 1990+24% early start case than in the 1990+24% case.

1990+9% Carbon Emissions Target

From 2000 through 2005, industrial gross output is $372 billion less in the 1990+9% early start case than in the 1990+9% case, and energy consumption is 7.5 quadrillion Btu lower (Figures 26 and 27). In the early start case, industrial output returns to the output levels projected in the 1990+9% case by the end of the forecast period. Industrial energy consumption in the early start case exceeds that in the 1990+9% case for a few years in the 2008-2012 period, due to the slightly higher industrial gross output in those years. As a result, the cumulative loss in output from 2000 through 2020 is only $160 billion lower than in the 1990+9% case with the 2005 start date. For most years, industrial energy consumption in the early start case falls below that in the 1990+9% case. From 2000 through 2020, cumulative energy consumption is 10.7 quadrillion Btu less than in the 1990+9% case with the 2005 start date, and expenditures for industrial energy consumption are $97 billion higher (discounted at 7 percent).

Figure 26. Projections of U.S. Industrial Output in the 1990+9% and 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

Figure 27. Projected Energy Consumption in the U.S. Industrial Sector in the 1990+9% and 1990+9% Early Start Cases, 1998-2020  [source]

1990-7% Carbon Emissions Target

The results for the 1990-7% cases are similar to those for the 1990+9% cases. From 2000 through 2005, industrial output in the 1990-7% early start case is $685 billion lower, and industrial energy consumption is 12.4 quadrillion Btu lower, than in the 1990-7% case. Although industrial output and energy consumption are higher in the early start case from 2008 through 2012, they are lower in the later years. As a result, cumulative industrial output from 2000 through 2020 is $825 billion lower and energy consumption is 15.7 quadrillion lower in the early start case than in the 2005 start case, and cumulative expenditures for energy are $183 billion higher (discounted at 7 percent).

Early Start Assumptions

In the early start cases, fuel prices for transportation are higher before 2005 than they are in the corresponding 2005 start cases, resulting in higher fuel efficiency, lower travel, and lower projected fuel consumption. As for the other end-use consumption sectors, it is assumed that the availability of efficiency improvements is accelerated in the early start cases. The time to introduce technologies for improving the efficiency of light-duty vehicles is assumed to be shortened by 20 percent at an additional incremental cost of 10 percent.

Analysis Results

1990-7% Carbon Emissions Target

In the 1990-7% early start case, gasoline consumption is 1.6 quadrillion Btu lower in 2005 than it is in the 1990-7% case, jet fuel consumption is 0.4 quadrillion Btu lower, and distillate fuel consumption is 0.2 quadrillion Btu lower. More efficient light-duty vehicles penetrate more rapidly and earlier in the projection period in response to higher gasoline prices and slightly (0.9 percent) lower income levels, and the efficiency of new cars is almost 4 miles per gallon higher in 2004. Consumers also respond to higher projected fuel prices by purchasing smaller vehicles and fewer light trucks and by reducing the demand for horsepower. Subcompacts capture a larger share of total car sales in 2005 (23 percent in the early start case, 17 percent in the 2005 start case), the average horsepower of new vehicles is almost 22 percent lower in 2004, and the light truck share of total vehicle sales is lower in 2004 (42.9 percent vs. 46.0 percent). With gasoline prices higher and income levels slightly lower in the early start case, consumers travel less. Light-duty vehicle travel is 6.6 percent lower in 2004, leading to further reductions in fuel consumption (Table 8).

Although new car fuel efficiency is significantly improved in the early start case, the average stock efficiency of vehicles is only 0.7 miles per gallon higher in 2005 because of the slow turnover in the vehicle stock. Stock turnover is slower in the early start case, with 6.7 percent fewer vehicle sales. Therefore, most reductions in fuel consumption in the early start case relative to the 2005 start case are the result of reduced travel. In the air and freight modes, most fuel savings also result from reduced travel due to higher jet fuel prices and lower industrial output. Aircraft and freight trucks have very long useful lives and slow turnover rates. In the early start case, lower travel early in the projection period reduces the opportunity for efficiency improvements in these modes. As a result, the average fuel efficiency projected for these modes in the early start case does not improve even with the higher fuel prices through 2005.

By 2010, projected gasoline fuel prices are 9.1 cents per gallon lower in the 1990-7% early start case than in the 1990-7% case, because the carbon prices are lower. Vehicle travel is almost 4 percent higher in 2010, and new car fuel efficiency is approximately 1.1 miles per gallon lower as consumers shift their purchases to more light trucks, larger vehicles, and higher horsepower. As a result, gasoline consumption in 2010 is similar in the two cases. Air travel is 6.7 percent higher and freight travel is 3.2 percent higher in 2010 in the early start case because the carbon price is lower, with a smaller negative impact on the economy. Total transportation energy consumption is almost 0.5 quadrillion Btu higher in 2010 in the early start case.

The gap in fuel consumption between the two 1990-7% cases narrows slightly by 2020, with a relative increase of only 0.3 quadrillion Btu in the early start case. In 2020, gasoline prices are 11.2 cents per gallon lower in the 1990-7% early start case than in the 1990-7% case. New car fuel economy is the same by 2020 in the two cases; however, light-duty vehicle travel is 3.0 percent higher and air travel is 2.3 percent higher in the early start case, primarily because fuel prices are lower.

1990+9% Carbon Emissions Target

In the 1990+9% early start case, the projected gasoline prices in 2010 are about 3 cents per gallon lower and transportation fuel consumption is nearly 0.2 quadrillion Btu higher than in the 1990+9% case. Lower gasoline prices and higher disposable income in 2010 result in 1.3 percent higher travel for light-duty vehicles and 0.4 miles per gallon lower fuel efficiency for new cars.

1990+24% Carbon Emissions Target

For the 1990+24% cases, transportation energy consumption in 2010 is projected to be the same in the early start and 2005 start cases as a result of the countervailing effects of lower gasoline prices (by 2.1 cents per gallon) and higher income (by 0.7 percent) in the early start case. New car fuel efficiency is only 0.1 miles per gallon higher in the early start case, and the difference is offset by 0.7 percent higher travel.

In the electricity generation sector, the primary impact of an early start date is the acceleration of retirements of carbon-intensive generating capacity, mainly oil- and coal-fired capacity. For example, in the 1990+9% early start case, 11 additional gigawatts of coal-fired capacity and 22 additional gigawatts of other fossil steam capacity, mostly oil-fired, are projected to be retired by 2005, as compared with the 1990+9% case (Table 9). By 2010, even more coal-fired and other fossil steam capacity is retired in the early start case. By 2020, patterns of capacity are similar in the two cases, although slightly less coal-fired capacity is retired, slightly less combustion turbine capacity is added, and there is noticeably more generation from petroleum and less from natural gas in the early start case than in the 2005 start case.

In both of the 1990-7% cases, there are more early retirements than in the less stringent carbon reduction cases. Further, in the early start case, retirements of coal-fired capacity between 2000 and 2010 are nearly triple those in the 2005 start case. There is little additional impact on other fossil steam units in the early start case. After 2010, because of the earlier retirements of coal-fired generating units, there is less opportunity for reducing carbon emissions through retirements of coal-fired units.

To meet electricity demands, additional low-carbon or noncarbon capacity must be built to replace coal-fired capacity. In the 1990+9% cases, the requirements are met primarily with natural-gas-fired combined-cycle capacity. By 2010, an additional 13 gigawatts of combined-cycle capacity are added in the 1990+9% early start case compared with the 2005 start case.

In the 1990-7% cases, the projected differences in coal- and natural-gas-fired capacity are larger, and additional renewable capacity is also built. In the 1990-7% early start case, 26 more gigawatts of combined-cycle capacity and 23 more gigawatts of renewable capacity are built by 2010 than in the 2005 start case. After 2010, however, the differences are reduced or eliminated. By 2020, combined-cycle capacity is slightly lower in the early start case and renewable capacity is 21 gigawatts higher, as the result of two related effects. First, in the early start case, capital costs for new capacity for units beyond a certain threshold (generally 20 percent of the previous year's installed capacity) are reduced, because it is assumed that equipment manufacturers will be able to increase their output at a lower cost in light of earlier and more gradual growth in demand for new capacity. As a result, capital costs are lower in the later years because of earlier market penetration, primarily for wind and biomass technologies. The impact of increased wind capacity is relatively small, however, because of the low capacity utilization rates for wind power stations.

The early start date has little impact in the 1990+24% case. Although 4 additional gigawatts of coal-fired capacity are retired by 2010 in the early start case, generation from coal remains similar to that in the 1990+24% case. By 2020, the primary impact is slightly lower renewable capacity in the early start case, because about 5 gigawatts of other fossil steam units remain in service, due to the slightly lower carbon price later in the projection period. Generation and capacity patterns are similar throughout the entire period in the two 1990+24% cases.

Measured in terms of fuel input¹⁰ per kilowatthour of electrical output, electricity generation is more efficient in the early start cases between 2005 and 2010 than it is in the 2005 start cases. For example, efficiency is 7 percent higher in 2005 in the 1990-7% early start case than in the 1990-7% case, primarily because coal-fired generation is lower and natural-gas-fired generation is higher in the early start case to meet the earlier carbon reduction target. By 2020, however, the differences in generation efficiency are essentially eliminated.

Accelerated retirements of coal-fired and other fossil steam units in the early start cases are primarily the result of the higher carbon prices projected in the early years. Retirement decisions are based on expected operating costs, which are made up mainly of fuel costs. In the early start cases, coal prices are as much as $4 per million Btu higher in 2005 due to the higher carbon price, resulting in early retirements of coal-fired capacity. After 2010, in order to meet electricity demands and carbon reductions that are approximately the same in the early start and 2005 start cases, similar capacity additions and generation are required. As a result, patterns are virtually unchanged in the early start cases. An exception is the noticeable decrease in nuclear and natural-gas-fired generation in the 1990-7% early start case relative to the 1990-7% case, which is offset by increased renewable generation in 2020 in the early start case.