significantly delay the building of new plants and change the economics of natural gas replacement of electric generating facilities. Charles This scenario is not idle speculation. MERGE, a model developed by Dr. Alan Manne and Dr. Richard Richels, also describes the process of electricity generation in considerable detail. Their model and others place constraints on the amount of coal capacity that can be replaced by 2010. In short, the SGM assumption that all coal capacity would be replaced by natural gas in such a short time does not appear to be tenable. Energy Efficiency Administration testimony also includes a discussion of the reductions in cost that are assumed to come about from improvements in energy efficiency. More specifically, Dr. Yellen assumes an annual rate of energy efficiency improvement (AEEI) of 0.96%, which is 6% above what the DOE Energy Information Administration used in calculating its 1998 Annual Energy Outlook. CRA's replication also includes this assumed 6 percent higher rate of change in energy efficiency. Kyoto Cost Estimates, Using More Realistic Assumptions Now that the Administration's analysis of the costs of Kyoto to the U.S. has been replicated and the underlying assumptions fully revealed, we seek to ask what might happen if the same underlying analysis is applied but under different assumptions. The first step is to reconcile the CRA model with the results reported by the Administration based on the SGM model. The second step is to focus on the costs of implementing the Kyoto agreement under different underlying assumptions. Reconciling the CRA and SGM Models The CRA model produces nearly identical results to the SGM when the elasticity of substitution between natural gas and coal is increased to produce the same permit price and required emissions reduction, and the rate of autonomous energy efficiency improvement (AEEI) is increased from 0.90/yr to 0.96/yr. Table 4 below (which uses a 1990 stabilization target as the baseline because SGM results for this target have been published) illustrates the possible alignment of assumptions and results among these two models. Note that the SGM model assumes that a $108/metric ton carbon tax would return 2010 greenhouse gas emissions to 1990 levels, largely by converting some coal-fired power plants to natural gas. The CRA model can reproduce this result using the higher elasticity and AEEI assumptions, but using our reference case assumptions, CRA concludes that the permit price would have to be closer to $142/metric ton to achieve the same emissions reduction. Using the $142 carbon price, CRA estimates a direct cost of $28 billion in 2010 and a total GDP loss of $60 billion to purchase carbon permits equal to EIA projected greenhouse gas emissions in 2010. When the carbon price is reduced to $109 by tripling the ease with which natural gas Charles River Associates could replace or substitute for coal in electric utilities and other boilers, the direct cost falls to $20, identical to SGM's results. In this scenario, annual GDP losses would fall to $52 billion." *Adjusted for an increased elasticity of substitution between coal and gas and an increased autonomous energy efficiency improvement rate. Source: CRA Multi-Region Trade Model and Edmonds, et al. We find, like the Administration, that emissions trading would reduce the prices of permits and the economic impacts of emissions limits. However, the CRA model does not estimate cost savings or price reductions from emissions trading as large as those cited by Dr. Yellen. Table 5 below compares the percentage cost reductions estimated by the SGM and CRA models to those assumed by Dr. Yellen. Table 5 Comparison of SGM, Administration, and CRA Savings from Emissions Trading Source: CRA Multi-Region Trade Model, Yellen testimony, and Edmonds, et al. CRA and SGM are very close to each other in estimating the total cost savings of global trading, much closer than are the Administration and the SGM. CRA is somewhat less optimistic than SGM about cost savings from Annex I trading, but more optimistic than SGM about the savings from non-Annex I trading. We believe this is largely due to our assumptions about future For consistency, the reconciliation of CRA and SGM uses a common scenario both models addressed in late 1997. The Administration analysis starts with the Kyoto agreement and requires a larger emissions reduction than estimated above because of updated carbon emissions projections by ELA and a more stringent emissions target. CRA's adjusted results include a somewhat lower required emissions reduction because changing other assumptions to match SGM affects the baseline as well as the policy case. Charles reforms in the Russian economy and our more complete treatment of international trade in energy and other goods that enables us to capture more of the competitive distortions caused by leaving developing countries out of the trading system. How the Administration found larger savings is less clear to us. One of the most interesting conclusions of the SGM is that nearly all of the benefits of Annex I emissions trading disappear if Russia is not allowed to sell its excess emissions permits on the open market. As can be seen in Table 6 below, under Administration assumptions, Annex I trading with no “hot air" from Russia results in less than a 20% reduction in cost, whereas emissions trading provides a 40% reduction in direct cost if Russia is free to sell all of its excess. permits. Our analysis supports this conclusion. Table 6 SGM Estimates of Direct Costs as Percent of GDP in 2010 CRA has replicated the Administration's starting point by adopting arguendo the fuel substitution elasticities and energy efficiencies needed to make the CRA results consistent with those of the Second Generation Model. Dr. Yellen's assumptions do lead logically to her conclusion, and with the same assumptions the CRA model gets very similar results for carbon prices and direct costs. Table 7 below shows that, with these assumptions, the CRA-estimated carbon permit price and direct cost for the case in which no international emissions trading is allowed are virtually identical to those implied by the Administration's testimony. CRA estimates a GDP loss, including both direct and indirect effects of the restriction on carbon emissions, that is about 60% larger than direct costs, or $79 billion in 2010 under the Administration's assumptions. 8 Edmonds, et al. *Adjusted for an increased elasticity of substitution between coal and gas and an increased autonomous energy efficiency improvement rate. All prices are stated in 1997 dollars, which requires an inflation adjustment of 13% to prices reported by the CRA and SGM models in 1992 dollars. The required emissions reduction is smaller in the CRA Adjusted results because adopting the Administration's assumptions changes both the baseline and policy cases. Source: CRA Multi-Region Trade Model and Yellen testimony. CRA Results, Using Less Optimistic Trading Assumptions Less optimistic assumptions about the speed with which utilities can replace coal with natural gas and on the progress of energy efficiency result in a significantly higher cost of reducing U.S. carbon emissions. Table 8 below illustrates the costs associated with other emissions trading scenarios. Table 8 reports results based on CRA's alternative assumptions about cost, namely, that utility replacement of existing coal with new natural gas units is limited during the next 10 years, and that energy efficiency improvement is no greater than that assumed by the Energy Information Administration. Table 8 CRA Results for Limited Emissions Trading Under CRA Assumptions Charles No International Emissions Trading If the nations that ratify the Kyoto Protocol are unable to agree on any trading system, CRA estimates that the cost of emissions reduction would rise to $295 per metric ton and GDP losses would be $128 billion in 2010 under realistic assumptions about utility costs and technology. This is the worst case, but given current divisions of opinion among the negotiating parties, it needs to be kept in mind as a reference. Trading Among Annex I Countries The Kyoto agreement provides explicitly for trading among Annex I countries. If the United States could achieve its goal of unrestricted trading among Annex I countries, the cost of Kyoto under realistic cost assumptions would be about 0.9% of GDP in 2010, or $92 billion, with an international permit price of about $120 per ton. This case includes, in the CRA analysis, about 160 million tons of excess permits offered for sale by Russia. However, European countries and others have objected to Russian sales of permits in excess of projected baseline emissions, and have asked for "concrete limits" on the amount of a country's emissions obligation that can be satisfied through purchase of permits from other countries. A case that reflects the inability of the United States to overcome these objections fully is that of Annex I trading with "no hot air." This case is still optimistic, in that it places no limits on the amount of its obligation that the United States can satisfy through permit purchases internationally, but does assume that other countries succeed in their objective of limiting Russia's ability to sell permits above its projected baseline emissions. It does not include trading with developing countries, because such trading is not permitted under the Kyoto Protocol and is largely opposed by developing countries. Using the CRA model and assumptions, carbon permit prices would be about $171 per metric ton with only Annex B trading and no sales of excess emissions permits from Russia. In that case, GDP losses would be 1.1% of GDP or about $111 billion in 2010. Global Trading CRA concurs with the Administration that full global trading, if achievable, would produce a significant reduction in costs. Under CRA's cost assumptions, global trading would still imply a carbon price of about $50 per ton and a GDP loss of over 0.5% in 2010, or $55 billion. This is a cost reduction of about 57% from the cost of Kyoto with no international emissions trading. 9 Table 8 assumes full Annex I trading but does not include benefits of excluding the EU from bidding for Russian emissions permits. Additional details on this alternate scenario will be available in a forthcoming CRA report, but the disadvantage to the EU of such a policy argues against its acceptance of the rest of the Administration's proposal unless the EU also could share their benefits. |