of technologies. The costs and emission reduction potential of technical options have been assessed using energy system models that evaluate end-use efficiency improvements, fuel substitution, and new supply technologies in an integrated way. This analysis has been conducted with different degrees of sophistication, depending on the individual country and the particular study. As an example, a summary of the main technical emission reduction options in the UNEP country studies is presented in Table 9.23. The national options are listed in aggregated form and thus represent classes of options rather than individual technologies. The cost curve may be considered in three segments, representing negative/low-cost options, intermediate-cost options, and high-cost options. One general similarity among the country studies is that the least expensive part of the cost curve contains energy end-use savings in households and/or industry. Another is that electricity supply options first appear in the intermediate-cost part of the reduction potential cost curves. On the supply side, most of the studies focused on traditional energy-supply technologies and few included more advanced technologies and/or renewable energy technologies. Consequently, the cost curve either increases very sharply or simply does not include any further reduction after the exhaustion of these options. 9.2.4.2.5 Comparability of the National Studies The emission reduction costing studies for developing countries considered here exhibit similarities with regard to the assessed potential for negative or low-cost emission reduction. In general, these options comprise end-use efficiency improvements, energy supply efficiency improvements, and the introduction of fuels with lower carbon intensity. These technologies cover the first and cheapest part of the emission reduction potential and can, especially in the longer term, be supplemented with renewable energy technologies, more far-reaching end-use savings, and advanced combustion technologies. The long-term emission reduction potential will consequently be extended and is also likely to be cheaper than currently estimated. The individual country studies are difficult to compare quantitatively because of differences in methodological approach and in scenario assumptions for economic growth, energy requirements, and emission reduction costs. An important difference is to be found in the baseline scenario assumptions used in the different studies. In the UNEP studies, the national research teams took official macroeconomic forecasts as the starting point for energy demand projections. In contrast, the LBL studies used a broader-level international perspective to estimate an economic structure and income distribution which could be achieved in a developing country within a given time horizon. Thus, for example, it could be assumed that a country like Brazil would approach an economic structure and income distribution comparable to that of Spain in a given time frame. The advantage of using national macroeconomic projections is that they reflect national views on development. However, national forecasts may be only partially consistent and realistic, while international economic studies may be of help in establishing a consistent data set across countries. Studies using a common, welldocumented background can also be easier to compare than studies that use different national forecasts. The degree of optimism of experts from different countries with regard to the penetration of energy efficiency or of carbon-free energy supply options may differ dramatically. This can lead to different critical assumptions in the baseline as well as in the emission reduction scenarios. One approach is to assume that all possible efficiency improvements will be implemented as part of the baseline scenario, implying that only positive-cost options remain. If the existing energy system - as in many developing countries at present - is relatively inefficient, the above mentioned approach implicitly assumes large investment programmes to be carried out to implement the "efficient options" in parallel to any emission reduction effort. If, instead, it is assumed that major inefficiencies persist in the energy system in the baseline scenario, there will be an interrelationship between emission reduction measures and the general effort to overcome barriers for efficiency improvements in the energy system. In the UNEP study, for example, where country research teams were free to make judgments, the team for Venezuela assumed that all profitable efficiency improvements would be implemented in the reference scenario, while the teams for Brazil and Thailand assumed a relatively inefficient reference scenario. Another difference in reference scenarios between countries relates to difficult assumptions about structural change in the economy, about fuel supply, and about the overall level of development. The striking difference between the UNEP and LBL studies in the case of Brazil is an enlightening example of the consequence of these differing assumptions. Part of the difference between the two studies for Brazil can be explained by different assumptions for the GDP growth rate, but another key difference is a consequence of the low fuel price increase projected in the UNEP study during a period long enough to make the existing alcohol fuels programme unprofitable. Consequently, the reference case assumes replacement of present biomass use, including ethanol, with fossil fuels. The LBL studies, in contrast, defined reference energy scenarios as a continuation of historical trends, implying that the alcohol fuels programme would be sustained in Brazil. A new Brazilian study (La Rovere et al. 1994) has been carried out as a compromise between the assumptions of the UNEP and LBL studies. 9.2.4.2.6 Conclusion The bottom-up CO2 emission reduction costing studies carried out for the energy sector for developing countries exhibit, despite differences in methodological approach, some common empirical results, namely: The 30-40 year reference scenario projections show a tendency to decreasing The potential for a 30-40% emission reduction from baseline over a 40-year time frame has been estimated. However, even after such a reduction, emissions will, on average, be two or three times more than present levels, because of economic growth. The emission reduction potential includes low/low or negative-cost options relating to end-use and conventional supply technologies in the short to medium term. In the 30-40 year time frame, the UNEP country studies have estimated average emission reduction costs to be below $14/tonne of CO2. 9.2.5 Global Studies of the Costs of Reducing CO2 Emissions The review of existing studies up to this point has focused on country or regional analyses. In addition, there are a growing number of studies that attempt to provide a global perspective on the assessment of abatement costs. These studies are important for several reasons. The enhanced greenhouse effect is inherently a global issue. If significant reductions in emissions are required, they can be accomplished only through international accords and cooperation. It will be helpful to have some sense of the overall costs before confronting the difficult issue of burden sharing. A global perspective is also important in assessing the costs to individual countries. Actions taken in one region are apt to have "spillover" effects into other regions. Partial equilibrium analyses ignore potentially significant linkages (e.g., trade in oil, gas, and carbon-intensive basic materials) that could substantially alter the economic impacts of a carbon constraint. Finally, and perhaps most importantly, a global perspective is necessary if we are to identify economically efficient strategies for achieving emission targets. The Framework Convention on Climate Change states that "policies and measures to deal with climate change should be cost-effective so as to ensure global benefits at the lowest possible cost." This means that emission reductions should be carried out where it is cheapest to do so. Analysis on a global scale is needed in order to construct a "least-cost" global abatement supply curve. Of course, all of the caveats expressed above about the limitations of countryspecific or region-specific models should be borne in mind when evaluating the results of global studies. In addition, the global analyses confront aggregation issues which further 23-558 96-30 complicate the interpretation of results. Not surprisingly, most of the global studies employ top-down methodologies. Nevertheless, there have been several bottom-up studies that illustrate the overall technical potential for curbing greenhouse emissions at a world level. 9.2.5.1 A Review of Top-down Studies: the importance of international cooperation Several of the models used for regional analyses provide the capability for analyzing emission abatement costs at a global level. These models typically divide the globe into five or more geopolitical groupings. By necessity, they are highly aggregate in their treatment of macroeconomic and technology issues. In order to represent regional differences and trade effects, local economic and technological detail must be sacrificed. Figure 9.23 summarizes the results of recent studies using these models. Not surprisingly, there is considerable disagreement concerning the costs of emission abatement. Consistent with the regional studies, the top-down global analyses indicate that emission abatement will involve positive costs, but the size of the model-based cost estimates varies from study to study. In this section, we will try to understand why these results differ so widely. In doing so, we hope to gain additional insights into the costs of emission abatement at the regional and global level. 9.2.5.1.1 The costs of stabilizing global emissions A systematic comparison of global models was undertaken by the OECD in its 1992 Model Comparison Project (OECD, 1993). The study included all available global models with the capability of simulating regional carbon tax rates required to achieve specific emission abatement targets and the resulting output losses. The models were calibrated to the same set of input assumptions employed by the parallel study being conducted by the Energy Modeling Forum (see earlier discussion). Results were reported for a business-as-usual scenario and for four scenarios involving various levels of emission reduction. We begin by examining the results from the emission stabilization scenario. For emission stabilization, it was assumed that global emissions would be permanently held at 1990 levels. The global costs of meeting such a target will depend, in part, upon how emission reductions are allocated among regions and whether there is scope for international cooperation. In the analysis that follows, each region is required to hold emissions at 1990 levels without the opportunity of shifting emission abatement from high to low marginal abatement cost regions. That is, there is no trade in carbon emission rights. Later on, we will explore the potential benefits from relaxing this constraint. Figure 9.24 compares output losses entailed by stabilizing emissions in the year 2020. It also shows the reduction in the average annual growth rate of CO2 emissions required to achieve the desired target. Note that losses vary by a factor of nearly threefrom 0.8% to 2.2% of gross world product. Although it is difficult to isolate all the reasons for the differences in abatement cost estimates, the discussion of the preceding sections points to several possible causes. The higher the baseline emission level, the more stringently carbon emissions have to be curtailed to achieve a given target, and hence the higher the overall abatement costs. The global baseline, however, is not the sole determinant of global abatement costs: The GREEN baseline projects the highest emissions for the year 2020, hence it requires the largest annual reductions to stabilize emissions at their 1990 level. Yet this does not lead to the highest cost estimates. This is because GREEN is the most optimistic of the four models concerning the speed at which backstop technologies can be introduced into the energy system. Conversely, ERM projects the lowest baseline emissions for 2020. As a result, less carbon must be removed from the energy system to achieve emission stabilization. Even so, estimated costs (in terms of resulting output losses) are higher than those projected by CRTM. This is because ERM does not include backstop technologies, thus producing higher marginal costs of emission abatement. In a separate experiment, it was determined that ERM and MR project essentially the same baseline emissions when employing identical rates of autonomous energy efficiency improvements. However, for the OECD study, MR adopted an average annual rate of 0.5%, whereas ERM assumed that non-price induced efficiency improvements occur at twice this rate. Standardization for this key parameter brings results of these two models much closer together with respect to their projections of GDP losses from emission abatement. The OECD Model Comparison Project suggests that the principal reason why model-based studies differ with respect to estimates of both baseline emissions and abatement costs is alternative views about the future characteristics of the energy system embodied in the models. In an attempt to place greater reliance on expert knowledge in this area, Manne and Richels (1994) polled a group of individuals on their beliefs about key parameters to which abatement costs are particularly sensitive. For each parameter, expert beliefs were encoded in the form of probability distributions which were, in turn, combined to form a group probability distribution. Using the poll responses to describe the uncertainty surrounding critical parameters, a probability distribution was then constructed for the costs of stabilizing global CO2 emissions at 1990 levels, using the MR model. Figure 9.25 presents the results of this analysis. The spread of the distribution is quite broad, ranging from 0.2% to 6.8% of gross world product (GWP). The median (that is, the fiftieth percentile) is located at approximately 1.0% of GWP. The |