The working group compared the results from 10 different energy-economy models from seven scenarios where all teams tried to use common assumptions for key policy and economic factors. The major findings included:
The US economy will use energy and emit carbon emissions much less intensively (with declining heat content or emissions per dollar of output) in future years than in the past. In addition, the current study expects future trends without any policy changes will already incorporate many promising energy-efficiency opportunities that are often included in estimates of the economic potential for energy savings. Even with these adjustments, however, total emissions without a concerted policy action will hold relatively steady over time and these trends will be contrary to goals of policymakers who want a major transition towards declining emissions in the long run. The study considered a range of policies that could reduce energy use and total emissions: carbon pricing through a tax on emissions, mandated standards requiring more efficient appliances, buildings and automobiles, and subsidies or reductions in the upfront costs of new more energy efficient equipment. Although energy use and emissions can be reduced even more with new programs, the improvements in energy efficiency in this study are more modest than estimated by other groups.
The principal reasons for lower energy demand reductions in this study are attributable to behavioral rather than technical reasons. Other studies have focused on the economic or cost effective potential based solely upon technology performances and costs. The current study extends the analysis to include the rate of adoption of these new technology options as well. The energy and emissions trends in models with explicit technology options for energy efficiency are surprisingly similar to the trends in models that focused more directly on market responses and economic equilibrium. These comments apply to the models as a group and do not describe all the models within a group. This finding suggests that how a model represents technology options does not necessarily determine whether its projections will be higher or lower than other models. Other structural model features, parameter values and assumptions about key conditioning factors appear to be primary contributors to differences in model outcomes. Models differed on whether carbon pricing or mandates were more effective. How consumers spent their energy savings from standards was an important issue. Standards were a less effective policy when consumers chose other goods, services and activities that required more energy use. This offsetting effect could include more purchases of the more energy-efficient activity (the direct "rebound" effect) as well as other activities requiring energy use. Carbon taxes and subsidies for new equipment in most models appear to be additive policy measures that do not detract from each other when combined into one package. Carbon taxes and standards may be more redundant in some models, implying that combining them may reduce their effect to some extent. This issue warrants more analysis, however, because the modeling community is only beginning to evaluate this issue. Although the modeling community is increasingly striving for frameworks that combine richness in technology coverage with realism in market behavioral responses, improvements are required to make the models more useful for policymaking. Many improvements require additional research on the behavior of would-be adopters of new technologies to both economic conditions and policy programs.