11.4.3.1 Policy studies for the United States

Both Fischer and Morgenstern (2006) and Lasky (2003) identify treatment of international trade and the disaggregation of the energy sector as important factors leading to differences in the cost of Kyoto for the US economy. Lasky also identifies energy-demand elasticities and sensitivities to higher inflation as important factors. He concludes that the cost of the US joining Kyoto under Annex I permit trading is between -0.5 to -1.2% of GDP by 2010, with a standardized energy-price sensitivity, and including non-CO₂ gases and sinks, but excluding recycling benefits and any ancillary benefits from improved air quality. The cost falls to 0.2% of GDP with global trading of permits. Barker and Ekins (2004) review the large number of modelling studies dealing with the costs of Kyoto for the US economy that were available when the US administration decided to withdraw from the process. These include the World Resources Institute’s meta-analysis (Repetto and Austin, 1997), the EMF-16 studies (Weyant and Hill, 1999) and the US Administration’s own study discussed above (EIA, 1998). The review confirms Lasky’s range of costs but offsets these with benefits from recycling the revenues from permit auctioning and the environmental benefits of lower air pollution. These co-benefits of mitigation are discussed in Section 11.8 below.

Following U.S. rejection of the Kyoto Protocol, there have been a number of policy proposals in the United States focusing on climate change, most notably two proposed during 2005 Congressional debates over comprehensive energy legislation (the Bingaman and McCain-Lieberman proposals, the Regional Greenhouse Gas Initiative, the Pavley Bill in California, and the earlier proposal by the National Commission on Energy Policy). The costs and other consequences of those proposals are summarized in Table 11.10, as compiled by Morgenstern (2005) from studies by the U.S. Energy Information Administration (1998; 2004; 2005). The sectoral implications of (EIA, 2005) are discussed above in Section 11.3.3.

All estimates derive from EIA’s NEMS model, a hybrid top-down, bottom-up model that contains a detailed representation of energy technologies, energy demand, and primary energy supply, coupled with an aggregate model of economic activity (Holte and Kydes, 1997; Kydes, 2000; Gabriel et al., 2001). While the estimates were conducted over a period of seven years, with changes occurring in the baseline, the model produces a remarkably consistent set of estimates, with most physical quantities (including emission reductions) varying more or less linearly with carbon price, and potential absolute GDP impacts varying with the price squared. Real GDP impacts, which include business cycle effects, are less consistent and depend on both policy timing and assumptions about revenue recycling. For example, the real GDP loss of 0.64% shown for ‘Kyoto+9%’ is reduced to 0.3% by 2020 when recycling benefits are taken into account (EIA, 1998).

In addition, EIA (2005) analyses the 2004 scenario of the National Commission on Energy Policy. The estimated cost is 0.4% of the reference case GDP by 2025 and the overall growth of the economy is ‘not materially altered’ (p. 42). However, no costs were included for the implementation of the ‘CAFE’ transportation sector portion of the NCEP programme that produced most of the emission reductions.

As an independent, government statistical agency, EIA’s modelling results tend to be at the centre of most policy debates in the United States. Researchers at MIT (Paltsev et al., 2003) also provided estimates of impacts associated with the McCain-Lieberman proposal that had similar allowance prices but found roughly one-quarter to one-third of the GDP costs reported in the EIA analyses. This is partly explained by the fact that the EIA uses an econometric model to compute GDP costs derived from historic experience in the face of energy price shocks. The MIT and other CGE models assume that, to a large extent, aggregate costs equal the accumulated marginal costs of abatement, typically yielding lower costs than the econometric models (Repetto and Austin, 1997; EIA, 2003).

A threshold question in the McCain-Lieberman discussion has been whether the exclusion of small sources below 10,000 metric tons (e.g. households and agriculture) would alter the efficiency of the program. Pizer et al. (2006) use a CGE model to show that exclusion of these sectors has little impact on costs. However, excluding industry roughly doubles costs while implementing alternative CO₂-reducing policies in the power and transport sectors (a renewable energy standard in the power sector and fuel economy standards for cars) results in costs that are ten times higher.

The states in the U.S. have put forward climate policy proposals. An analysis of a package of eight efficiency measures using a CGE model (Roland-Holst, 2006) indicates that it will reduce GHG emissions by some 30% by 2020 – about half of the Californian target of returning to 1990 CO₂ levels by 2020 – with a net benefit of 2.4% for the state’s output and a small increase in employment (Hanemann et al., 2006). These results, driven by bottom-up estimates of potential savings in the vehicle and building efficiency, remain controversial, as the debate over vehicle fuel economy standards demonstrates (see NHTSA, 2006 for a discussion of bottom-up estimates and issues).