Climate Change 2001:
Working Group III: Mitigation
Other reports in this collection

9.4.2 The Role of Assumptions Baseline

A critical point for the results of any modelling is the definition of the baseline (or reference or business-as-usual) scenario. The SRES (Nakicenovic et al., 2000) explores multiple scenarios using six models and identifies 40 scenarios divided into 6 scenario groups. As OECD (1998) points out, among the key factors and assumptions underlying reference scenarios are:

Differences in the reference scenarios lead to differences in the effects of mitigation policies. Most notably, a reference scenario with a high growth in GHG emissions implies that all the mitigation scenarios associated with that reference case may require much stronger policies to achieve stabilization.

Nevertheless, even if reference scenarios were exactly the same, there are other reasons for changes in model results. Model specification and, more importantly, differences in model parameters also play a significant role in determining the results. Costs and Availability of Technology

If any fuel becomes perfectly elastic in supply at a given price (i.e., the backstop technology), the overall price of energy will be determined independently of the level of demand, which will then become the critical determinant of mitigation costs. Hence, the assumption of a backstop technology strongly determines mitigation costs. Models without a backstop technology will tend to estimate higher economic impacts of a carbon tax, because they rely completely on conventional fuels, so that the tax rate has to rise indefinitely to keep carbon concentrations constant, to offset the effects of economic growth. Endogenous Technological Change

The treatment of technology change is crucial in the macroeconomic modelling of mitigation. The usual means of incorporating technical progress in CGE models is through the use of time trends, as exogenous variables constant across sectors and over time. These trends give the date of the solution. Technical progress usually enters the models via two parameters: (i) autonomous energy efficiency (AEEI) (if technical progress produces savings of energy, then the value share of energy of total costs will be reduced); and (ii) as changes in total factor productivity. The implication of this treatment is that technological progress in the models is assumed to be invariant to the mitigation policies being considered. If in fact the policies lead to improvements in technology, then the costs may be lower then the models suggest. Price Elasticity

In assessing the effects of mitigation, estimations of price-induced substitution possibilities between fuels and between aggregate energy and other inputs can be crucial for model outcomes. All such substitutions become greater as the time for adjustment increases. The problem is that estimates of substitution elasticities are usually highly sensitive to model specification and choice of sample period. There is little agreement on the order of magnitude of some of the substitution elasticities, or even whether they should be positive or negative, e.g., there is debate whether capital and energy are complements or substitutes. If energy and capital are complements, then increasing the price of energy will reduce the demand in production for both energy and capital, reducing both investment and growth. Most CGE models consider very different possibilities of substitution, for example WW, Global 2100, and Nordhaus’s DICE/RICE models assume capital and labour as substitutes, while GREEN assumes capital and energy as direct substitutes. Degree of Aggregation

There are many different products, skills, equipment, and production processes; many important features are missed when they are necessarily lumped into composite variables and functions. A basic difference among models and their results is the level of aggregation. Indeed, in practice, different goods have different energy requirements in production, and therefore any changes in consumption and production patterns will affect them differently. Hence, a highly aggregated model will miss some potentially major interactions between output and energy use, which is precisely the purpose of the analysis. For example, sectoral disaggregation allows the modelling of a shift towards less energy-intensive sectors, which might reduce the share of energy in total inputs. In the same way, when a carbon tax is introduced, it could reduce the estimated costs of abatement by allowing substitution effects of energy-intensive goods by less energy-intensive goods. Treatment of Returns to Scale

Constant returns to scale represent a common assumption on the economic modelling of climate change. However, in practice, economies of scale seem to be the rule rather than the exception. Indeed, there are several reasons for economies of scale, see Pratten (1988), and Buchanan and Yoon (1994). For example, many electricity-generating stations benefit from economies of scale, utilizing a common pool of resources including fuel supply, equipment maintenance, voltage transformers, and connection to the grid. In spite of this fact, the impact of the effects of increasing returns and imperfect competition (IC) in the modelling of climate-change strategies has consistently been neglected in the literature. Most of the global models, if not all, assume explicitly perfect competition, for example, see DICE/RICE, G-Cubed, Global 2100, GREEN, GTEM, WorldScan, and WW.

Other reports in this collection