2. Results

This chapter discusses the impacts of the policy assumptions on energy markets, carbon dioxide emissions, and the economy. The initial focus is on the multi-policy cases, Cases 1 and 2, as described in Chapter 1. The chapter begins with a discussion of the overall energy consumption impacts of the multi-policy cases, followed by a discussion of the impacts for each energy consumption sector (electric power, residential, commercial, industrial, and transportation). The impacts of individual policies are presented in the third section. Detailed tables displaying results of the multi-policy cases, compared with the AEO2005 reference case, are provided in Appendix B.

Multi-Policy Case Aggregate Results

Most of the energy reduction in 2025 in Case 1 is for petroleum (2.4 quadrillion Btu, or a 4.5percent reduction), followed by coal (1.0 quadrillion Btu, or 3.4 percent), and natural gas (0.5 quadrillion Btu, or 1.6 percent). The petroleum reductions in Case 1 occur primarily as a result of the more stringent CAFE testing policy, which requires manufacturers to increase the average fuel economy of new light-duty vehicles. In Case 2, which includes the electricity and natural gas voluntary programs, upgraded building code policies, and the national-level EEPS, the reductions are targeted more heavily to the electricity sector. As a result, projected coal use in Case 2 is reduced by 12.7 percent compared to the reference case in 2025, natural gas by 7.9 percent, and oil by 4.8 percent. While some of this reduction occurs because of the voluntary efficiency policy, most of the reductions are from the 9.1-percent reduction in projected electricity demand in Case 2, much of which occurs because of the national-level EEPS. In Case 1, which assumes a five-State EEPS, the reduction in electricity demand in 2025 is 2.4 percent, compared to the reference case.

The impacts on carbon dioxide emissions are greater than for total energy consumption in percentage terms in both cases, because the policies tend to target fossil fuels and have relatively little impact on renewable and nuclear energy. Also, the Case 2 policies result in greater percentage reductions for coal, which has the highest carbon content. Overall carbon dioxide emissions in 2025 are reduced by 3.5 percent in Case 1, relative to the AEO2005 reference case, and by 8.3 percent in Case 2.

The energy reductions under both policy cases grow over time (Figure 1). While most of the policies are specified to begin between 2006 and 2009, the policies are generally phased in over time or target a steady reduction in energy growth. Also, policies that call for improvements in the efficiency of new equipment will impact the market slowly because of gradual equipment turnover. As a result, projected policy impacts on total energy consumption in 2025 are more than six times as great as those in 2010 in both policy cases. The projected energy reductions through 2015 are roughly half of the 2025 reductions.

Because the Case 1 policies reduce oil consumption by 4.5 percent in 2025, primarily through the CAFE testing policy, they reduce dependence on imported oil. The net import share of oil consumption falls from 68.4 percent in the reference case to 67.3 percent in both multi-policy cases. The Case 2 polices, which result in a substantial reduction in natural gas use in 2025 (7.9 percent) compared to the reference case, also reduce dependence on imported natural gas. The projected net import share of natural gas in 2025 falls from 28.2 percent in the reference case to 26.1 percent in Case 2.

Macroeconomic Impacts

The relative importance of reductions in energy use that are attributable to program results provided to EIA as part of the study request, which were incorporated in the analysis without being explicitly modeled in NEMS, presents a significant challenge to the application and interpretation of EIA’s standard macroeconomic analysis.

The reduced demand for natural gas and electricity in the policy cases places downward pressure on prices, because the energy is supplied from lower-cost sources at the margin. For example, the projected wellhead natural gas price in 2025 is 8.5 percent lower in Case 2 than in the reference case. However, given the absence of information regarding the
implementation of key policies included in the two multi-policy cases, EIA was not able to reliably estimate impacts on retail energy prices and energy expenditures. For example, based on experience, it would be reasonable to expect that programs used to meet demand-reduction obligations placed on electricity and gas suppliers subject to an EEPS would be financed through surcharges on consumer bills; however, estimates of the cost of demand reduction in past programs vary widely, and there is no clear basis for using a particular estimate to represent the unspecified programs that might be pursued. The cost of programs used to reduce demand in these policy cases could partially, fully, or more than fully offset the effect of reduced demand on retail energy prices. Without information on changes in delivered prices, or information regarding the impact of these programs on investment, EIA was not in a position to develop reliable estimates of the impacts of the policy suites on actual GDP.

The NEMS macroeconomic model can, however, be used to capture the costs of the energy efficiency policies on the potential output (supply) side.¹³ The energy efficiency policies would force the economy to substitute capital and labor resources for energy, causing a relocation of these productive resources across economic sectors. The adjustment process would cause losses to the potential output of the economy, relative to baseline, due to losses in productivity.

To highlight the distinction between actual and potential GDP, consider an energy tax that induces the same changes in energy consumption as the energy efficiency policies. With such a tax policy, consumers and businesses would react to the higher energy prices that reflect the higher marginal cost. The economic adjustments toward more energy-efficient products would occur through a market mechanism and would be reflected in actual GDP as expenditures are switched, causing idling and dislocation of some resources in the short to medium term. Potential output, on the other hand, reflects full employment of resources. The losses incurred in potential output have more to do with losses in productivity of inputs due to their reallocation and do not include the idling, adjustment, and dislocation costs. Because there is no explicit accounting of the latter types of costs, this analysis focuses only on the impacts on potential GDP or losses in productivity.

Conceptually, it is possible to decompose the impact on potential GDP by looking at three distinct, and possibly opposing, effects:

• The implementation of policies to promote higher energy efficiency forces the economy toward a different mix of resources, using less energy. The adjustment to the new optimal mix entails losses to the potential output from a loss in productivity of capital and labor resources.
• Businesses and consumers lower their energy consumption. The lower level of energy demand itself tends to reduce energy prices, with positive impacts on the economy; however, the costs of implementing the EEPS and other such programs tends to raise delivered energy prices. Without modeling of these programs, the net effect on delivered energy prices is unclear.
• Businesses and consumers alter their investment behavior due to the policies, as discussed above. This affects not only actual GDP but potential GDP as well, through impacts on interest rates, corporate profits, and capital stock formation.

In the present study only the first of these effects, the impact of forced adjustments to factor inputs from the path of the reference case, could be addressed directly. This was done by estimating the cost to the economy of letting lower energy use feed into the potential GDP equation, but keeping energy prices and energy consumption at baseline levels for the demand side of the economy. For Cases 1 and 2, the cumulative losses in potential GDP from 2006 to 2025 due to the loss in productivity were $445 billion (a 0.14-percent reduction) and $864 million (a 0.27-percent reduction), respectively.

To illustrate the effect of possible changes in energy prices, EIA’s analysis also considered how reductions in energy prices, if they occurred, would positively impact potential output, assuming that the lower energy use would not alter the optimal input mix for potential GDP. For this sensitivity analysis, a bounding assumption was that implementation of the EEPS and other such programs does not impact delivered energy costs, which reflect only the effects of lower energy usage. In essence, this controls for the adjustment costs on the production side of the economy and ignores the price implications of implementing the policies. The cumulative positive impact on the economy from 2006 to 2025 from lower energy prices due to lower energy demand is $160 billion (a 0.05-percent increase) for Case 1 and $309 billion (a 0.10-percent increase) for Case 2. To the extent that actual delivered energy prices are affected by implementation costs, these results will overstate the gains in potential GDP due to price changes, which could lead to losses in potential GDP.

Impacts by Sector

The energy policies considered in this analysis reduce energy in all four end-use sectors relative to the reference case (Figure 2). In Case 1, about half of the reductions in energy use in 2025 occur in the transportation sector as a result of the policy to reform CAFE test procedures. In Case 2, the reductions are split more evenly across sectors, with the
transportation sector accounting for the smallest amount of savings among the four sectors in 2025. The Case 2 savings in 2025 attributed to electricity generation account for just over half the total reductions. These savings occur primarily because of the national-level EEPS policy for electricity and the electricity sector voluntary agreement policy. The key impacts in each sector are summarized in the following sections.

Electricity Sector

Case 1 impacts:

• Sales of electricity in 2025 are 2.4 percent less in Case 1 than in the AEO2005 reference case, and power sector energy use falls by 1.6 percent. Because the reduced electricity demand also reduces the need for new, more efficient generating capacity, the average efficiency of the electricity generation is lower in Case 1. As a result, the reductions in energy use for generation are less than proportional to the reductions in electricity demand.
• The projected need for new generating capacity in 2025 is reduced by 19 gigawatts (7.2 percent) in Case 1 compared to the reference case. Cumulative power sector capacity additions from 2004 through 2025 in Case 1 are 244 gigawatts, compared with 263 gigawatts in the reference case. Most of the avoided capacity additions in Case 1 are coalbased (83 percent).
• Carbon dioxide emissions associated with electricity generation are reduced by 2.8 percent in 2025 in Case 1, a greater percentage change than for electricity demand (2.4 percent) or fuel use (1.6 percent). The coal-fired capacity additions avoided in Case 1 account for the higher percentage reduction in carbon dioxide emissions, because the carbon content of coal is greater than that of other fossil fuels.

Case 2 impacts:

• Sales of electricity in 2025 are 9.1 percent lower in Case 2 than in the reference case, and power sector energy use is lower by 9.0 percent. Unlike in Case 1, the percentage reductions for generation are about the same as for electricity demand. Case 2 includes the voluntary agreement policy calling for the industry to reduce energy intensity (or plant heat rates) by 5.0 percent cumulatively between 2006 and 2016. The effect of this policy is partially offset by a reduction in the need for new, more efficient capacity in Case 2.
• The need for new generating capacity through 2025 is reduced by 76 gigawatts (29 percent) in Case 2 compared to the reference case, and total capacity (existing and new additions) is reduced by 7.8 percent. Most of the avoided capacity additions in Case 2 are coal-based (69 percent), and about 5 percent are renewable.
• Projected carbon dioxide emissions associated with electricity generation are reduced by 12.1 percent
  in 2025 in Case 2, compared to the reference case.

Residential Sector

Case 1 impacts:

• In 2025, delivered energy consumption in the residential sector is reduced by 0.4 quadrillion Btu (3.0 percent) relative to the reference case, and primary energy consumption is reduced by 0.7 quadrillion (2.8 percent).
• Electricity savings account for more than half of the cumulative delivered energy savings in Case 1. When conversion losses are factored in, electricity reductions account for 72 percent of the cumulative primary energy savings.
• Residential carbon dioxide emissions, including the emissions associated with the generation of the electricity consumed in the sector, are reduced by 3.6 percent (56 million metric tons) in 2025, a greater percentage reduction than for total energy use, which is reduced by 2.8 percent in 2025.

Case 2 impacts:

• In 2025, delivered energy consumption in the residential sector is reduced by 1.4 quadrillion Btu (9.5 percent), and primary energy is reduced by 2.6 quadrillion (9.7 percent) relative to the reference case.
• Natural gas savings account for more than half of the cumulative delivered energy savings in Case 2, as the addition of building codes for non-HUD homes increases the savings in space heating fuels.
• Carbon dioxide emissions are reduced by 189 million metric tons (12 percent) in 2025, with unspecified EEPS programs contributing most to the decline.

Commercial Sector

Table 6 compares the commercial sector projections in the two multi-policy cases to those in the reference case.

Case 1 impacts:

• In 2025, delivered energy consumption in the commercial sector is reduced by 0.2 quadrillion Btu (1.8 percent) relative to the reference case, and primary energy consumption is reduced by 0.3 quadrillion (1.1 percent).
• Electricity savings account for 55 percent of the cumulative primary energy savings in Case 1 (including conversion losses).
• Commercial sector carbon dioxide emissions, including the emissions associated with the generation of electricity consumed in the sector, are reduced by 2.0 percent (32 million metric tons) in 2025—a greater percentage reduction than for total energy use, which is reduced by 1.1 percent in 2025.

Case 2 impacts:

• In 2025, delivered energy consumption in the commercial sector is reduced by 1.1 quadrillion Btu (8.6 percent), and primary energy consumption is reduced by 2.3 quadrillion Btu (8.7 percent).
• Cumulative energy savings in Case 2—23.3 quadrillion Btu through 2025—are 8 times the savings in Case 1, reflecting the combined energy savings from the policies to upgrade commercial building codes and the national-level EEPS policy.
• Carbon dioxide emissions are reduced by 183 million metric tons (11.3 percent) in 2025, with the assumed impacts of the EEPS policy contributing most to the decline.

Industrial Sector Table 7 compares the industrial sector projections in the two multi-policy cases with those in the reference case.

Key findings:

• In Case 1, industrial petroleum consumption shows the largest reduction, 3.7 percent (0.4 quadrillion Btu) in 2025 relative to the reference case, primarily due to reduced consumption by refineries and oil and natural gas producers as a result of petroleum demand reductions brought about by the CAFE reform policy. Industrial natural gas consumption is reduced by 3.1 percent (0.3 quadrillion Btu) in 2025, due to the combined effects of the voluntary agreements and the five-State EEPS.
• The largest energy consumption impacts in Case 2 are for industrial natural gas and purchased electricity, as a result of the national EEPS policies for natural gas and electricity. In Case 2, the national EEPS policy is the most significant contributor to the 7.9-percent reduction in purchased electricity in 2025 compared with the reference case. Primary energy consumption in 2025 is projected to be 2.4 quadrillion (6.1 percent) lower in Case 2 than in the reference case.
• The net effect of the voluntary programs is to reduce projected combined heat and power (CHP) capacity in both cases, despite the availability of the investment tax credit. The voluntary agreement policy reduces both electricity requirements and steam requirements, thereby reducing the potential for CHP. In Case 2, industrial CHP capacity in 2025 is 2.4 gigawatts (5.9 percent) less than in the reference case.
• Industrial sector carbon dioxide emissions in the reference case increase from 1,664 million metric tons in 2003 to 2,059 million metric tons in 2025, including emissions associated with the generation of the electricity consumed in the sector. In Case 2, carbon dioxide emissions are 160 million metric tons (7.8 percent) lower in 2025 than in the reference case.

Transportation Sector

The policy to revise the CAFE test procedure was included in both multi-policy cases examined in this study. The impacts of the policy are virtually identical in the two cases, and this discussion therefore focuses on the results from Case 1. Results of the reference case and Case 1 are compared in Table 8.

The CAFE reform policy would require a revision in current fuel economy test procedures, with the new procedures phased in from 2008 to 2012. By 2012, manufacturers would have to develop vehicles that provide a 20-percent increase in fuel economy to comply with the more stringent fuel economy test procedures. Initial revisions to the test procedure, to be implemented in 2008, would require a 4-percent increase in fuel economy. An additional 4percent improvement in fuel economy would be required in each following year until 2012.

Key impacts of the CAFE policy:

• Compared with the AEO2005 reference case, the increase in fuel economy for light-duty vehicles projected in Case 1 results in a 3.3-percent reduction (1.1 quadrillion Btu) in petroleum demand in 2015 and a 5.0-percent reduction (1.9 quadrillion Btu) in 2025.
• In 2025, light-duty vehicle travel in Case 1 increases by 2.2 percent (91 billion miles annually) compared to the AEO2005 reference case, because increased vehicle fuel economy reduces the cost of driving.
• The increased penetration of advanced technologies required to meet the more stringent CAFE test procedures increases the average price of a new vehicle in Case 1 compared to the AEO2005 reference case. In 2015, the average price of a new car increases by $530 and the average price of a new light truck increases by $620 (2003 dollars). In 2025, the average incremental price increases are $400 for cars and $480 for light trucks.
• As a result of increased new vehicle prices, projected sales of new light-duty vehicles are lower than in the AEO2005 reference case. In 2015, new light-duty vehicle sales are lower by 2.5 percent (470,000 vehicles), with 61 percent of that reduction attributed to a decrease in light truck sales. In 2025, new light vehicle sales are lower by 2.2 percent (460,000 vehicles), with the decrease in new light truck sales accounting for 67 percent of the total reduction.

The fuel economy projections for new light-duty vehicles in the policy cases are not directly comparable with those in the AEO2005 reference case because the fuel economy projected in the policy case reflects a revised CAFE test procedure, not an increase in the CAFE standard. The projected new car and light truck fuel economy in Case 1 can be adjusted to reflect the tested fuel economy that would have been achieved under the current CAFE test procedure. In Figure 3, “adjusted Case 1” miles per gallon is the new car fuel economy measured under the current test procedure for comparability to the AEO2005 reference case. For example, the tested new car fuel economy in 2015 decreases from 30.3 miles per gallon as reported in the AEO2005 reference case to 28.0 miles per gallon in Case 1; however, with adjustments for fuel economy improvements required under the new CAFE test procedure, new car fuel economy in 2015 in Case 1 is equivalent to 33.6 miles per gallon measured under the current test procedure.

Once the new test procedures have been phased in, there is little additional change in new car fuel economy in Case 1, unlike in the reference case. With the new test procedures, manufacturers implement technology earlier than they would have otherwise to meet the redefined CAFE standard. After 2012, new car fuel economy remains relatively constant through the remainder of forecast period in Case 1, as many of the low-cost, advanced technologies have saturated the market. In the reference case, these technologies penetrate more gradually over time, leading to an upward trend in new car fuel economy over the entire projection period.

Similar to the fuel economy projections for new cars, the direct comparison of the unadjusted fuel economy for new light trucks between Case 1 and the AEO2005 reference case is misleading. For example, in 2015, the unadjusted new light truck fuel economy is projected to be 22.6 miles per gallon in Case 1, compared to 23.4 miles per gallon in the reference case. However, with adjustments for fuel economy improvements required under the new CAFE test procedure, new light truck fuel economy in 2015 is equivalent to 27.1 miles per gallon as measured under the existing CAFE test procedure (Figure 4).

Although there was no increase in the CAFE standards and projections of new light-duty vehicle fuel economy in Case 1 are lower than those projected in the AEO2005 reference case due to the revision of the test procedure, the elimination of the shortfall between tested and on-road fuel econom y results in an increase in the average fuel economy of the stock of light duty vehicles (Figure 5). Compared to the AEO2005 reference case, stock average fuel economy increases by 6.6 percent (1.5 miles per gallon) to 21.8 miles per gallon in 2015 and by 10.8 percent (2.3 miles per gallon) to 23.3 miles per gallon in 202 5. Although Case 1 projections of new light-duty vehicle fuel economy remain relatively flat as new vehicles continue to displace the old vehicle stock, average stock fuel economy would continue to increase beyond 2025.

Impacts of Individual Policies

The approximate energy savings from individual policies that were specifically modeled in NEMS are presented in Table 9, with a comparison to the energy consumption projections of the AEO2005 reference case.¹⁴ The individual policies with the greatest cumulative impa cts on projected energy consumption are the CAFE reform policy and the national-level EEPS for natural gas and for electricity. Together, these three policies account for about 79 percent of the total cumulative energy reduction from 2006 to 2025 of 93.4 quadrillion Btu in Case 2 (a combined energy savings of 3.9 percent of the 20-year total of 2,379 quadrillion Btu). The CAFE reform policy, included in both multi-policy cases, accounts for 50 percent of the annual energy savings projected for Case 1 in 2025 and 21 percent of the Case 2 savings in 2025.

The policies with the smallest cumulative impacts on energy consumption are the tax incentives (tax credits for new and existing homes, residential and commercial tax credits f or efficient equipment and building shells, and an investment tax credit for small CHP plants). Together, these policies, which have a greater impact early in the projection when the incentives are in effect, reduce projected energy consum ption in 2010 by 0.04 quadrillion Btu (less than 0.1 percent) compared to the reference case. The impacts of the tax incentive policies are mitigated by the limited duration of the tax credits and the slow turnover rates of the targeted building stock and related equipment. In addition, the credits tend to be small relative to the marginal cost of adopting the more efficient technology.

The impact of the investment tax credit for CHP is small due to the policy’s limitation of 15 megawatts, a s well as its 3-year duration from 2006 to 2008. In the commercial sector, the tax credit results in between 5 and 6 megawatts of additional commercial CHP capacity in the two multi-policy cases. In the industrial sector, an additional 52 megawatts of eligible CHP capacity is added by 2008 compared to the AEO2005 reference case when the policy is analyzed by itself. When the policy is combined with the other policies, however, the ove rall additions of CHP are reduced. The reduction in electricity and steam demand under the voluntary agreements reduce the potential market for CHP, outweighing the positive effect of the tax incentive.

Of the incentives targeted at the residential sector, the new home tax credits are projected to save the smallest amount of energy cumulatively through 2025, because new construction accounts for a small amount of total residential housing in a given year, and the tax credit is in effect for only 3 years.

Appliance standards are more effective than tax incentives in reducing energy demand, because they limit consumers’ choices to the more efficient technology, thus ensuring the technology adoption and the associated reductions in energy demand. When promulgated, the standards remain in effect over the entire projection period, allowing the impacts to grow over time as more appliances are purchased and/or replaced.

The scope of a particular efficiency standard and its effective date affect the projected energy impact and the efficiency improvement required by the standard. The distribution transformer standard and the reach-in refrigerator efficiency standard both are targeted at very specific segments of commercial energy use. Although these standards save a significant share of energy use per unit, their narrow scope limits their potential effects on total commercial energy use.

The building code policies produce relatively large energy savings, compared to some of the other policies. In the residential sector, building codes achieve an energy savings of 1 percent (0.3 quadrillion Btu) in 2025. In the commercial sector, the building code policy saves 2 percent (0.5 quadrillion Btu) in 2025. However, more than 80 percent of the reduction with the commercial building codes is due to the stipulation of a limit on the watts per square foot allowed for lighting. The proposed code may result in a change in the level of lighting service provided, because the maximum power requirement may be met by reducing the number of lights used in addition to installing lights that use less power.

The impacts of the voluntary agreement and EEPS policies are high relative to the impacts of the numerous policies targeted at specific end uses. Given the inherent uncertainty associated with the outcome of such programs, a variation on the multi-policy Case 1 was examined, and the results are presented in Table 9. In the Case 1 Subset, the effects of the industrial voluntary program and the State-level EEPS policies were removed. This variation on Case 1 provides an estimate of the combined impacts of the “hard” policies targeted at specific end uses, with the exception of the residential and commercial building code policies already omitted from Case 1. The energy savings in the Case 1 Subset are 68 percent of the Case 1 savings in 2025. The Case 1 Subset achieves a 2-percent reduction in total energy consumption in 2025 (2.7 quadrillion Btu). Of these savings, 73 percent (1.9 quadrillion) occur as a result of the CAFE reform policy.

Notes

Results Tables