Analysis of Senate Amendment 2028, the Climate Stewardship Act of 2003
Introduction
In June 2003, the Energy Information Administration (EIA) released an analysis1 of the Climate Stewardship Act of 2003 (S.139) as introduced by Senators McCain and Lieberman in January 2003. S.139 would establish a cap on emissions of greenhouse gases2 from covered sources that would be implemented in two phases beginning in 2010 and 2016 respectively. More recently, in October 2003, Senators McCain and Lieberman proposed an amended version of the bill, SA.2028, that included the first phase of emissions reductions beginning in 2010 but removed references to a second phase of reductions beginning in 2016.
On May 11, 2004, Senator Landrieu asked EIA to evaluate SA.2028. This paper responds to that request, relying on the modeling methodology, data sources, and assumptions used to analyze the original bill, as extensively documented in EIA’s June 2003 report. By using the same modeling system and assumptions, the impacts of SA.2028 can be compared as a sensitivity case to the previously reported results for S.139. However, these results do not reflect updates to EIA’s modeling system and the reference case energy forecast that were included in the Annual Energy Outlook 2004 (AEO2004)3. Given Senator Landrieu’s request for an expedited response, it was not possible to undertake a completely new analysis using the latest updates to the model.
In addition to removing references to a second phase of emission reduction, SA.2028 made several other changes with possible implications for the results. These include:
- SA.2028 omits a provision in S.139 that would have allowed automobile manufacturers to obtain emission allowances in exchange for exceeding the Corporate Average Fuel Economy (CAFÉ) standards by over 20 percent. This change is reflected in EIA’s analysis.
- SA.2028 now states explicitly that emissions from fuel sold for transportation outside the United States (i.e., “bunker fuels”) are not covered sources. Because EIA’s modeling system does not estimate emissions from bunker fuels separately, the exemption for bunker fuels is not reflected in EIA’s analysis. Because carbon dioxide emissions from bunker fuels were 1.6 percent of total energy-related carbon dioxide emissions in 2002, the exclusion of bunker fuels from the cap should not materially affect the results.
- SA.2028 adds a provision entitled “Dedicated Program for Sequestration in Agricultural Soils.” The provision allows an entity to satisfy up to 1.5 percent of its total allowance submission requirements by submitting registered increases in net carbon sequestration in agricultural soils. Entities remain subject to a 15-percent overall limit on offsets. EIA’s analysis methodology incorporates this provision through marginal abatement cost curves for agricultural and forestry combined, but does not separately constrain the proportion of that carbon sequestration from agricultural soils.
For the sake of brevity, the following discussion of the SA.2028 case assumes familiarity with EIA’s previously published analysis of S.139. The SA.2028 case is compared both to the updated reference case from the Annual Energy Outlook 2003, on which the previous analysis was based, and the S.139 case.
Analysis of the SA.2028 Case
Emissions and Allowance Costs
 Figure Data |
 Figure Data |
The most significant change in SA.2028 relative to S.139 is the removal of references to a more restrictive second phase of emission caps beginning in 2016. While this change has its greatest impact after 2016, it also reduces some of the incentive to over-comply and bank allowances during the 2010 to 2015 period. Therefore, the realized level of covered emissions between 2010 and 2015 would tend to be higher in SA.2028, even though the allowance cap over that time period is the same as under S.139 (Figure 1). Eliminating the second phase also means that the cap on the use of offsets remains at 15 percent, instead of being reduced to 10 percent in 2016, as in S.139. This added flexibility helps to reduce the compliance costs of SA.2028 compared to S.139.
With a less restrictive emissions limit under SA.2028, the market for allowances would be expected to clear at a lower price than under S.139 (Figure 2). Estimated allowance prices (in 2001 dollars) grow from $55 per metric ton carbon equivalent in 2010 to $167 in 2025 under SA.2028, compared to a growth of $79 to $221 dollars over the same period under S.139. Thus, on average, emission allowance costs are estimated to be about 30 percent less under SA.2028.
With higher covered emissions and lower allowance costs from 2010 to 2015, the use of emissions offsets to comply is initially reduced under SA.2028. As a result, the limit on offset usage from 2010 to 2015 is not binding. In this situation, the markets for emission offsets and allowances are expected to clear at the same price. By 2016, however, the 15-percent limit on offsets is reached, and competition to supply this constrained demand for offsets causes the offset price to clear below the allowance market price. The allowance price remains higher than the offset price after 2016.
The use of offsets is 75 percent greater beginning in 2016 under SA.2028 than under S.139, since the maximum allowable percentage remains at 15 percent instead of dropping to 10 percent and because the emission cap on which that percentage is applied is higher. This allows the offset market to clear at a higher price after 2015 than in S.139 case, but reduces overall compliance costs since offsets are still cheaper than allowances.
Table 1 compares the emissions-related results of the reference, S.139, and SA.2028 cases for 2010 and 2025.
Energy Sector Results
Under SA.2028, the effective cost of using energy increases compared to the reference case. This occurs because the costs of emission allowances (or their opportunity costs) are passed through to energy consumers. Consumers in the covered sectors will face higher costs for fossil fuels. Electricity consumers in all sectors are expected to face higher prices, as electricity suppliers pass their compliance costs on to customers.
Table 2 presents a summary of the key energy-related results for 2010 and 2025 for the reference, S.139, and SA.2028 cases. In general, the direction of changes in the SA.2028 case is the same as in the S.139 case, but the magnitudes of the changes are reduced, as the SA.2028 case is not as restrictive. In both the S.139 and SA.2028 cases, the use of natural gas, nuclear power, and renewable energy sources is greater than in the reference case, and the use of petroleum and coal is lower.
Energy price increases under SA.2028 are also lower than those under SA.139, resulting in correspondingly lower reductions in energy demand. Impacts of SA.2028 on delivered energy prices vary across sectors and fuels. The variation across sectors depends on whether or not a particular sector is covered under the bill and on the importance of distribution-related costs not impacted by the bill in the overall delivered energy price to each sector. For example, in the residential and commercial sectors, the delivered price of natural gas is virtually unchanged from the reference case level in 2010 and only 4 percent higher than the reference case in 2025. Greater increases occur in the average price of natural gas in the industrial and electric power sectors, 21 percent in 2010 and 58 percent in 2025, because the prices in these sectors include the allowance cost and distribution costs are a smaller component of delivered prices to these sectors.
The increases in gasoline prices projected to occur under SA.2028, 9 percent in 2010 and 19 percent in 2025 relative to the reference case, are expected to result in gradually increasing fuel economy in new passenger vehicles, reaching 27.2 miles per gallon by 2025, an increase of 0.8 miles per gallon over the reference case. Under the S.139 case, projected fuel economy for new vehicles reaches 29 miles per gallon by 2025. SA.2028 eliminates the additional incentive under S.139 that would allow automobile manufacturers to obtain emission allowances in exchange for exceeding the CAFE standards by over 20 percent. Had this incentive been retained in SA.2028, the average fuel economy for new light-duty vehicles in 2025 would be an estimated 0.6 miles per gallon higher, or 27.8 miles per gallon.
In both the SA.2028 and S.139 cases, the electric power sector accounts for about 88 percent of estimated emission reductions. Under SA.2028, however, the reduction in electric-power sector carbon dioxide from the reference case in 2025 is estimated at 404 million metric tons carbon equivalent (47 percent), compared to 663 (76 percent) in the S.139 case. As a result, only 26 gigawatts of nuclear power capacity are added by 2025 under SA.2028, compared to 49 gigawatts in the S.139 case. Relative to the reference case, the price of electricity increases less under SA.2028 (35 percent by 2025) than under S.139 (46 percent by 2025).
The production of coal is not expected to be as severely curtailed under SA.2028 as under S.139. Under SA.2028, coal production is reduced by 8 percent in 2010 and by 59 percent in 2025 relative to their respective reference case levels. Under S.139, the reductions in coal production relative to the reference case are estimated to be 14 percent in 2010 and 78 percent in 2025.
Macroeconomic Results
The estimated macroeconomic impacts of SA.2028 are also significantly less than those estimated for S.139, with the impacts reduced in rough proportion to the corresponding impacts on energy markets. The effects on the economy from higher energy costs result in output losses and shifting of resources.
The measurement of losses in output for the economy, or actual gross domestic product (GDP), incorporates the transitional cost to the aggregate economy as it adjusts to its long-run path. Alternatively, the economic impact of the bill can be measured by its effects on potential GDP, which represents the long-run equilibrium path of the economy in which all resources are fully employed. Table 3 compares the estimated economic losses from SA.2028 and S.139 using these two measures. On an undiscounted basis, the cumulative losses in actual GDP are about $776 billion (1996 dollars) in the SA.2028 case, 43 percent less than in the S.139 case. The peak, single-year impact on actual GDP under SA.2028 occurs in 2025, with a loss of $76 billion (1996 dollars), or about 0.4 percent of GDP. The largest percentage change in actual GDP, 0.5 percent, occurs in 2011, where the estimated loss in actual GDP that year is $57 billion.
Additional Context
As noted in our original S.139 analysis, the assessment of impacts over a 20-year time period is subject to considerable uncertainty. The sensitivity cases presented in the original report illustrate some of the uncertainties, but do not encompass the full range of energy and economic outcomes that might result from the bill’s enactment. The magnitude of the differences across comparable sensitivity cases for SA.2028 would, in most cases, likely be smaller, reflecting the lesser impacts projected under SA.2028.
Another study that has analyzed several variants of S.139 was issued by researchers at the Massachusetts Institute of Technology (MIT) in June 2003.4 This study included two scenarios that maintained the emissions cap at the 2010 level beyond 2015, as contemplated in SA.2028. One of these scenarios (Case 2) did not provide for any offset credits. The other scenario (Case 12) allowed for unlimited offsets for non-carbon dioxide greenhouse gases, notwithstanding the 15-percent limit on offsets imposed under SA.2028. These two scenarios bound a hypothetical case representing SA.2028. Table 4 compares the allowance costs for these two scenarios with those from EIA’s SA.2028 case, with costs from the MIT researchers’ paper converted from 1997 to 2001 dollars. Allowance costs in EIA’s SA.2028 case fall within the range of estimates for the two relevant scenarios in the MIT paper through 2015 and match the Case 2 allowance price in 2020. Other significant differences between the EIA and MIT researchers’ analyses are discussed in EIA’s earlier report, including the much greater responsiveness of oil demand to the introduction of the allowance system in the MIT researchers’ scenarios, which reduces the need for higher allowance prices to encourage adjustments in the electric power and industrial sectors.
Finally, like other EIA analyses, our analysis of SA.2028 focuses on impacts regarding energy choices made by consumers in all sectors and the implications of those decisions for the economy. This focus is consistent with EIA’s statutory mission and expertise. EIA did not quantify, or place any value on, possible health or environmental benefits of curtailing greenhouse gas emissions.
Appendix A: Request Letter from Senator Landrieu 
Appendix B: Comparison Tables for Reference Case, S.139
Case, S.A.2028 Case