Energy substitution by sector: technological flexibility and the impact of mitigation policy

Abstract

The impact of mitigation policies on economic activities is a longstanding controversial issue justifying the large strand of literature analysing the effect of climate change policies in terms of environmental and economic costs. From the Kyoto Protocol to the current climate policy agenda approved by the European Union (EU) in October 2014, mitigation of climate change still constitutes a challenging long term objective at the global level, and it is particularly relevant for the EU. Besides the cost-effectiveness issue, and considering the global scope of environmental policies in an open economy, further aspects to carefully account for are the regional distribution of those costs and the impacts on the economic, energy and industrial competitiveness. In this context, it is not surprising the comprehensive use of applied models representing the global economy and the relations between the economic and technological dimension, across countries and the economic sector. In the light of the challenging abatement targets and the assessment of mitigation costs, the current work is organised in three parts. The first part focus on a specific issue under investigation in applied energy economics that is energy substitution and the role of behavioural parameters in influencing cost assessment results. This analysis specifically addresses the computation of energy-output and capital-energy substitution elasticity, which is a measure of the energy-related technological flexibility. Energy-output elasticities are computed for 10 manufacturing sectors for a long time span (1970-2008) for a panel of 21 OECD countries while energy-capital substitution elasticity is estimated at aggregate level for the whole manufacturing industry for the same longitudinal dataset as well for 10 manufacturing sectors, also for separate sub-periods and comparing alternative econometric estimation methods. Results show that the energy long-run elasticity values for specific manufacturing sectors are highly heterogeneous, while the heterogeneity in the energy-capital substitution elasticity shows that the distinction between energy-intensive and non-energy-intensive sectors behaviour is less clear. Consequently, energy intensive sectors may require specific complementary energy conservation policies in order to be compliant with emission targets, hence this heterogeneity in the energy relationships should also be reflected in energy applied models. These results constitute the basis for the second part of the study that is a sensitivity analysis based on a dynamic climate-economy CGE model (GDynE) by computing abatement costs and competitiveness impacts of different climate change policies. In fact, CGE models including energy and CO2 data are particularly suitable to analyse the effect of carbon-abating policies considering they can capture the linkages between regulated and non-abating countries through trade channel and investment dynamics in the long-run. However, these models need to be improved and validated, to increase the reliability of the results, including detailed information on the technological and energy systems, which are represented by elasticities or behavioural parameters. In particular, the energy substitutability is here analysed by taking into account the nest structure and the input mix in the production function, tested for the relationship between capital and energy and among different fuels mix, distinguishing different manufacturing sectors. The simulation exercise reveals that the model produces highly differentiated results when different sets of elasticity parameters are adopted. In general, the reduction in the flexibility of the energy substitution possibilities makes the abatement efforts more expensive at the aggregate level. Moreover, such restriction generates changes in the distribution of the costs associated with the abatement efforts across regions. The direct implication derived by this work is that in order to use energy forecasting models to evaluate costs and feasibility of climate policies, it is necessary to enhance them with empirically estimated behavioural parameters at the highest possible disaggregation level. Finally, the third part focus on the European climate strategy to 2030 recently agreed in October 2014. This framework combines three objectives (GHG abatement, energy efficiency and renewable energy) thus the assessment of the mitigation costs need also to consider the potential trade-offs among simultaneous policies. Hence, in the light of the debate around effectiveness, timing of the abatement targets and optimality of policy mix, the EU approach to climate change represents an interesting case study to be investigated using GDynE model, improved with the sectoral elasticity parameters and including a mechanism to finance technical progress from green investment. Considering the policy mix, the comparison between a pure ETS mechanism and a climate strategy also including RD investment in energy efficiency and renewable energy (financed through a levy on the carbon tax revenue) is analysed. As the results show, in this second case the economic losses are lower and, as the percentage levy increases, reduce up to the point where increasing efficiency gains turn losses into economic gains. Moreover, when focusing on the sectoral differences, manufacturing sectors, and energy-intensive activities in particular, are negatively influenced by the emissions reduction in ETS scenario, however, if a proper policy mix with RD in clean energy technologies is implemented, losses are consistently reduced. When considering different timing of abatement targets, a first evidence is that the choice on whether preferring or not to delay more stringent targets in the future also depend on the selected policy mix. Indeed, if the market-based instrument was the only policy in place, postponing the achievement of more stringent CO2 reduction could seem preferable, however when fostering also energy efficiency and renewable energy support in a well-functioning ETS, the relative suitability of anticipating more challenging target seems to increase.