In the last decade, growing attention has been paid to global warming, its causes and potential impacts. Most recently, the Intergovernmental Panel on Climate Change reported that significant climate change due to the accumulation of anthropogenic greenhouse gases in the atmosphere is very likely to take place1 (IPCC, 2007). Substantial emissions reductions are needed to avoid potentially dangerous interference with the global climate. How such emissions reductions could be achieved and what economic costs and benefits it would imply is a matter of controversy among researchers around the world. However, the importance of innovative technologies in achieving deep cuts of emissions is undisputed. Often, technologies are thought to be the magic bullet for mitigation.
Understanding technological change is essential to generate effective options to curb emissions (Houghton, 2006). The availability, development, nature, and potential of innovations highly influence the effectiveness and economic efficiency of climate change mitigation. Energy innovations that enable or facilitate the reduction of carbon intensive energy use may substantially contribute to mitigation efforts. These include innovations for (1) energy conservation and energy efficiency improvement2; (2) fuel switching towards carbon free or low carbon energy, such as renewable energy; (3) technology switching towards low or no CO2 emitting systems; (4) carbon waste management, such as carbon capture and storage.
Technological change may relate to the development of new technologies or to the accelerated diffusion of already existing technologies. Innovation may be of radical nature, i.e. lead to fundamentally new products or processes, or of incremental nature, i.e. improve the concept or performance of existing sets of technologies. In either case, it implies substantial changes in the character of economic activity (Sue Wing, 2006a).
Many potential innovations do not reach technical maturity or do not pass the diffusion and adoption step and do not gain a market impact. Climate policy can help to bring innovations into the market. However, this has to be done in a cost efficient way with regard to the impact on the whole economy.
Understanding the effectiveness and costs of policy measures, and their success in shifting energy systems toward more environmentally desirable technology paths has been a challenge taken on, among others, by climate and energy-economy-environment policy modelers (Hourcade et al., 2006). In this dissertation, I explore innovative ways of including technological change in the framework of energy-economy models. I illustrate new approaches of treating technology innovations and apply these approaches to analyze the impacts of climate policy and innovation on economic activity, including energy supply and demand, and associated environmental effects. The following sections of this introduction provide important concepts and terminology with respect to modeling approaches and inclusion of technological change and place my research within the literature.
An increasing number of empirical models have been developed to analyze economic and environmental impacts of policy measures (Löschel, 2004), relying on different methodological approaches and addressing different foci. They can be divided into two main types of model approaches, bottom-up and top-down approaches, and differ with respect to the emphasis paid to including detailed, technology-based information of the energy system and including theoretically consistent descriptions of the general economy (Löschel, 2004). A discussion and survey of the specific features, advantages, weaknesses and caveats of each of the two approaches can be found in e.g. Bataille et al. (2006), Hourcade et al. (2006), Löschel (2002, 2004), Weyant and Olavson (1999).
Bottom-up models represent entire energy systems in terms of specific technologies. They simulate (or optimize) the operation of specific energy technologies based on cost and performance characteristics in a partial (equilibrium) framework. They contain detail on current and future technological options and describe competition of these technologies both on the energy supply side and on the energy demand side. Because of their technology focus and the possibility of accounting for fundamentally different technology pathways they can provide detailed information on environmental impacts for each path. However, bottom-up models lack interaction with the rest of the economy and rely on exogenous assumption about the scale of future energy demand (Grubb et al., 2006; Löschel, 2002). They do not include information on producers’ or consumers’ decision-making and, consequently, do not provide information on the behavioral aspects of the technology selection process. In addition, they are not linked to include feedback from macroeconomic variables, such as economic growth, economic structure, energy demand, and international trade. These parameters may change in response to energy and climate policies, which in turn would affect decision-making and technology selection and, subsequently, environmental impacts (Hourcade et al., 2006).
Top-down models, on the other hand, use a broader economic framework. Models of the top-down type are commonly called energy-economy models and include macro-econometric models, optimal growth models and dominantly computable general equilibrium (CGE) models. These models represent economic responsiveness to policies and account for feedbacks in form of for example input substitution, structural change, output adjustment, and trade effects. However, in order to include behavioral and other non-technical factors such as policy instruments, they usually compromise on the level of technology detail, which may be relevant for an appropriate assessment of energy or climate policies (Jaffe et al., 2003; Edmonds et al., 2001). Moreover, technology choice is usually constrained to current practice and substitution elasticities are calibrated to base year information or, in the case of econometric models, estimated based on historical data. These parameters, however, may change in the future in response to the availability of new technologies with their inherent characteristics and in response to new environmental policies. Most top-down models are not able to cope with such radical or even incremental changes, and their simulations into the future (baselines) remain bound to the behavioral and technical structure of the base year or past trends.
The divergence between the two model types became evident when the policy debate shifted towards the economic and technology analysis of reducing greenhouse gas emissions. It turned out that top-down models reveal high costs of greenhouse gas emissions mitigation because they assume that economic markets are in equilibrium and any deviation from this equilibrium imposes costs to the economy. This means they exclude the existence of inefficiencies and, thus, of energy efficiency potentials that could be profitably realized (Hourcade et al., 2006; Bataille et al., 2006). On the other hand, bottom-up models reveal ‘no-regret’ or low cost options to mitigate greenhouse gases because of their technology and efficiency improvement perspective and implicit assumption of the existence of market imperfections. They fail to include (transaction) costs related to removing such market imperfections. Market imperfections may be due to imperfect information, limited financial markets, technology-specific risks, inertia in technology preferences, behavioral change in response to efficiency gains (rebound effects), and more (Hourcarde et al., 2006; Löschel, 2002, 2004). The divergent views on the economics of efficiency improvement potentials are often referred to as the 'efficiency gap' and have been intensely discussed in the literature (Grubb et al., 1993).
An 'ideal' model would couple all sets of information, either in form of hard-linking different model types or in providing a model that incorporates all features, and would perform well in all categories depicted in Figure 1.1. It would be technologically explicit in the full range of activities, consider supporting upstream and downstream technologies, and cover the evolution of technologies and underlying risks and uncertainties (Jacoby et al., 2006; Hourcade et al., 2006; Bataille et al., 2006). Moreover, it would be behaviorally realistic in terms of micro-economic detail and it would include (macro)economic feedbacks, in linking changes in relative costs of goods and services to their supply and demand, as well as balancing budgets and markets.
|Figure 1.1 Energy-economy-environment models|
|Source: Adapted from Hourcade et al. (2006)|
To compensate for the limitations of either of the two approaches, hybrid models have been developed that incorporate features from one model type into the other and aim at combining features of both model types. Bottom-up modelers, usually with a background in engineering, physics or environmental sciences, add macro-economic feedbacks into their models or include micro-economic decision-making. Examples are extensions of the MARKAL optimization model, e.g. MARKAL-MACRO (Manne and Wene, 1992) which adds a growth model and economy-wide production functions to the MARKAL model, or MARKAL-ED (Loulou and Lavigne, 1996) which adds demand elasticities for some key products. A similar approach is followed in MERGE (Manne et al., 1995). In the MESSAGE-MAKRO model, an energy system model is solved in an iterative process with an economy model allowing for feedbacks between the two models (Rao et al., 2006). Another hybrid approach is demonstrated in the CIMS model, which also iterates between energy demand, energy supply, and macroeconomic modules (Bataille et al., 2006; Jaccard et al., 2003).
Top-down modelers, usually with a background in economics, devote efforts to adding explicit technological modules to their models, permitting a choice between these technologies and allowing for shifts in technology characteristics over time towards best practice innovative technologies (Schumacher and Sands, 2006, 2007; Edenhofer et al., 2006; Sands, 2004; McFarland et al., 2004, 2006; Welsch, 1998, 1996). Jacobsen (2000) employs a top-down macro-econometric model to incorporate the diffusion of energy technologies of different vintages associated with different levels of efficiency. Recent efforts devoted to coupling detailed energy models, such as MARKAL, with CGE frameworks include those by Schäfer and Jacoby (2006, 2005) for transport technologies and by Proost and van Regemorter (2000) for energy services. Using advanced mathematical techniques, Böhringer (1998) and Böhringer and Löschel (2006) demonstrate an approach of linking a CGE model with bottom-up activity analysis for electricity generation while other sectors are represented by conventional functional forms used in top-down analysis. Apart from theoretical, analytical as well as computational complexities of combining the two approaches, or features thereof, another important difficulty is to construct an integrated database. Engineering and economic data are most often not consistent and calibration of a model based on both types of datasets remains a challenge (Sue Wing, 2006).
A lesson learned from both model approaches is the importance of technologies, and changes thereof, for the assessment of mitigation costs and options. Independent of the modeling approach the assumptions about technology play a crucial role (Löschel, 2004). Therefore, I turn to a more detailed view on the inclusion of innovation and technological change in energy-economy-environment modeling.
There are different ways of incorporating innovation and technical change in energy-economy-environment models. Up to the late 1990s, most models have been rather weak at this issue (Nemet, 2006). Technologies and technological change were incorporated through exogenous assumptions. In top-down models, changes in technologies were reflected as a result of changes in relative prices through assumptions about elasticity of substitution between input factors.3 In addition, an autonomous energy efficiency improvement (AEEI) parameter was used to reflect an increase in efficiency independent of changes in prices or economic behavior (Grubb et al., 2006).4 Thus, the AEEI subsumes (exogenous) diffusion of new and efficient technologies or 'from heaven' changes in structural relationships. It implies a continuous, steady and incremental improvement and does not allow for radical innovations (Sue Wing, 2006). The concept of an autonomous efficiency improvement indicator is rather limited because the rate and direction of technological change are specified exogenously and are independent of the effects of changes in policy or other model variables. For policy analysis, this implies that substitution between inputs and output reduction are the only ways that input demand can be affected by policy measures. In contrast, penetration of technologies in bottom-up approaches is modeled using cost and performance characteristics of actual technologies. Technological change occurs as one technology is replaced with another, thus allowing for radical changes in technologies but still relying on exogenous assumptions on technology characteristics (Löschel, 2004).
Apart from the development of hybrid modeling approaches (compare section I1), which evolved in response to these criticisms, each community of modelers in itself realized that an enhanced treatment of innovations and technological change was needed to meaningfully evaluate the cost of climate and energy policies. Also, it was acknowledged that the rate and direction of future technological change impose high uncertainties for these evaluations (Edenhofer et al., 2006a). Consequently, modelers attempt to incorporate lessons from the literature on the economics of innovation and endogenous growth theory and seek to endogenize technological change in their models (Köhler et al., 2006; Nemet, 2006).5 Following Clarke et al. (2006a) endogenous technological change refers to technological change that depends – at least in part – on the development of particular socio-economic model variables like prices, investment in research and development, or cumulative production.6 The treatment of technological change became more sophisticated also because of increased computer power and improved algorithms to work with diverse phenomena (such as increasing returns) (Grubb et al., 2002).
A number of survey papers discuss the different treatment of technological change in economic and engineering models used to analyze climate policy. With varying focus they discuss implementation techniques, theoretical background, and implications on energy consumption, costs of environmental policy and timing of abatement measures (Sue Wing, 2006; Clarke et al., 2006a; Vollebergh and Kemfert, 2005; Löschel, 2002, 2004; Goulder, 2004; van der Zwaan et al., 2002; Weyant and Olavson, 1999; Grübler et al., 1999, 1998; Azar and Dowlatabadi, 1999; within the Innovation Modeling Comparison Project see Köhler et al., 2006; Edenhofer et al, 2006a; within the Stanford Energy Modeling Forum project on Technology and Global Climate Change Policies, see Weyant, 2004). A common finding from these efforts is that technology matters and that technology itself is modified by climate policy. Different ways of modeling technological change include introducing (1) backstop technologies, (2) enhanced technology information in hybrid approaches (3) technology learning, e.g. learning-by-doing, (4) R&D based knowledge accumulation (stock of knowledge approach), and (5) spillovers.
(1) Backstop technologies refer to sometimes generic, sometimes specific discrete technologies that are assumed to be exogenously available at some point in time, at specific marginal costs and with fixed characteristics as to emissions or energy intensity. They are often used to represent radical technological change because new production techniques can be explicitly modeled, however, at varying levels of detail (Löschel, 2004; Sue Wing, 2006; Kemfert, 2002). In top-down simulations, a backstop technology is often assumed to be a simple, generic carbon free technology, which becomes economically competitive in future periods in response to rising production costs of conventional technologies due to resources scarcity or policy induced price increases (Popp, 2006, 2006a; Kemfert, 2002; Löschel, 2002). In bottom-up models backstop technologies are usually explicitly represented with complete technology descriptions and expert judgments on typically relatively high production costs (Sue Wing, 2006). Commonly, backstop technologies are assumed to be available to produce any amount of output at constant marginal cost. This may lead to so called bang-bang or flip-flop behavior in models, which means that the backstop technology takes over the entire production once it has become competitive. To alleviate this, modelers often put an ad-hoc constraint on the rate of penetration of the backstop technology, thus imposing imperfect substitutability on the output of the backstop and the conventional technology (Sue Wing, 2006a; Popp, 2006, 2006a). Another profound limitation of this approach is that backstop technologies are discrete technologies with fixed input/output structure and marginal costs. Technological change beyond the assumptions inherent in the backstop technology cannot be accounted for.
(2) Based on the same principle of emphasizing the role of advanced technologies, but much more elaborated in their methodological and technological set-up, hybrid approaches have been developed. As discussed in section I1, they aim to incorporate features from both top-down and bottom-up approaches to reveal a more realistic picture of the energy, environmental and economic effects of climate policy and technological change.
(3) The concept of technology learning is based on the observation that production costs or investment costs of a certain technology or product decrease with cumulated experience of producing it. Experience can be described in terms of cumulated production, output, sales or cumulative installed capacity. Often learning-by-doing is distinguished from learning-by-using or learning-by-researching. Whereas learning-by-doing refers to cost reductions that occur in connection with increasing experience in the production and installation of a specific technology, learning-by-using refers to cost reductions achieved by increased efficiency and experience in using a specific technology. Moreover, learning-by-researching refers to cost reductions that arise as a result of R&D activities (Löschel, 2002). The learning approach is probably the earliest and most popular approach (Messner, 1997; Goulder and Mathai 2000; van der Zwaan et al., 2002). It is typically favored by bottom-up modelers who take advantage of the technological detail inherent to their models and their extensive knowledge of technology characteristics and related costs. Recent bottom-up models include a great number of different technologies for energy production and learning-by-doing for specific, selected technologies (Rao et al., 2006; Hedenus et al, 2006; Barreto, 2001; Seebregts et al., 2000).
Fewer studies so far have implemented learning effects into macroeconomic (top-down) models. They mainly differ with respect to the proxy/indicator for the activity which causes learning: i) cumulative installed capacity of a technology (Gerlagh, 2006; Gerlagh and van der Zwaan, 2003, 2004), ii) sectoral output (Rasmussen, 2001; Carraro and Galeotti, 1997), iii) sectoral capital stock (van Bergeijk et al., 1997), iv) sectoral labor input (Kverndokk et al., 2004), v) technological know-how (learning-by-researching) (Goulder and Mathai, 2000), or vi) a combination of these indicators such as the two-factor experience curve that takes into account cumulative capacity as well as cumulative R&D expenditure (Kouvaritakis et al., 2000, 2000a; Klaassen et al., 2005). Goulder and Mathai (2000) employ a formulation in which cost reduction due to learning is a function of cumulative abatement.
Most studies agree that learning effects are most pronounced for relatively new and fast growing technologies, e.g. non-fossil energy technologies, as an increase of cumulative experience can be more easily achieved (McDonald and Schrattenholzer, 2001). Thus, they separate fossil energy from non-fossil energy and analyze the effects of learning-by-doing in non-fossil energy goods, such as renewable energy (van der Zwaan et al., 2002). When technological progress is induced via learning-by-doing rather than by autonomous efficiency improvement, this has an impact on the costs and optimal timing of environmental policies and of investment, which is the focus of most of those studies.
A wide range of learning rate estimates for renewable energy can be found in the literature (Neij et al., 2004; Papineau, 2006; Junginger et al., 2005; Ibenholt, 2002; IEA, 2000). They differ because of varying assumptions with respect to time periods, cost measures (investment cost, levelized cost of electricity production, electricity or turbine price), experience measures (cumulated installed capacity, cumulative produced capacity, electricity generated), geographical area, system boundaries, data availability and quality, and estimation methods. Given these uncertainties, it comes at no surprise that modeler's conclusions from incorporating learning effects show a broad span of divergence.
In addition, the learning approaches suffer from other important limitations (compare Sue Wing, 2006): 1. In perfect foresight models (or optimization models) non-convexities are introduced by implementing learning effects, which can lead to multiple equilibria; 2. As with backstop technology models, penetration constraints for learning in form of upper bounds on capacity or investment rates need to be included. This implies that the trajectory of cost reduction becomes exogenous; 3. There is lack of transparency of learning rate assumptions, in particular in bottom-up models with large numbers of technologies and activities; 4. To date the approach is still heuristic with no profound theoretical foundation; 5. The simple learning approach implies that technological change results from activity within one and the same industry, it does not take into account spillovers from other industries, upstream or downstream production steps or activities in other countries; 6. The learning-by-doing approach implies that innovation occurs as a costless side effect rather than resulting from costly investment in R&D. It therefore takes the character of a free lunch.
(4) The R&D based knowledge accumulation approach (or stock of knowledge approach) picks up on the latter criticism and presents a learning-by-searching process where technological change is a result of investment in research and development. The approach is based on the idea that there is a stock of 'knowledge', which accumulates in reaction to an economic activity such as R&D. This knowledge influences production possibilities (or sometimes also consumption). The stock of knowledge or human capital is generated through investment into research and development activities. Model parameters, such as price changes induced by policy measures, may lead to increased investment into the stock of knowledge capital with its subsequent effects on substitution possibilities and productivity (Edenhofer et al., 2006; Popp, 2006a, 2004; Kemfert, 2005; Buonanno et al., 2003; Goulder and Mathai, 2000; Nordhaus and Boyer, 2000; Goulder and Schneider, 1999). The approach completely endogenizes technological innovation in treating it as an economic activity, which depends on profit-maximizing decision making from economic agents. It suffers most from a lack of disaggregated data on R&D at the level of individual technologies.
Given the model structure and sectoral and technology detail, macroeconomic (top-down) modelers tend to focus on the R&D approach while the majority of engineering (bottom-up) modelers focus on implementing learning-by-doing. Recently, more and more efforts have been taken to simultaneously model both approaches and reveal effects on economic output, environment and energy based on both costly and costless increase in experience (Bosetti et al., 2006; Goulder and Mathai, 2000; Gerlagh and Lise, 2005; Goulder and Schneider, 1999). This is sometimes referred to as two-factor experience curves (Kouvaritakis et al., 2000; Klaassen et al., 2005).
(5) Another important aspect in modeling technological innovations are spillover effects from R&D investment or technology learning. The existence of spillover effects implies that innovations are not fully appropriable. Spillovers may take the form of positive externalities such as R&D, knowledge, technology, and innovation transfer but also of negative externalities such as the transfer of emissions (carbon leakage) and environmental effects to other regions or countries (Otto et al., 2005; Jaffe et al., 2003; Grubb et al., 2002; Weyant and Olavson, 1999). Weyant and Olavson (1999) define technological spillovers as "any positive externality that results from purposeful investment in technological innovation or development". Such knowledge spillovers and the induced innovation and diffusion of new technologies have been intensively discussed in the literature. See for example Sijm (2004) for a thorough assessment of this issue.
The approaches outlined here are not mutually exclusive, but can be applied independently or in combination. Typically, efforts to an enhanced treatment of technological change do not attempt to make all technological change in the model endogenous but allow certain technologies or industries to change endogenously while others are still treated using an exogenous specification (Clarke et al., 2006a, 2006). For example, emerging, innovative energy sector technologies might be treated endogenously while other technologies and the rate of change in the economy as a whole remain exogenous. Jacoby et al. (2006) call for caution when introducing endogenous technological change in top-down computable general equilibrium (CGE) models, because of potential double counting. Double counting may occur because empirically estimated key elasticities (substitution elasticities, income elasticities) may already reflect a certain degree of endogenous change in technology based on the underlying data. Similarly, some technical change may be incorporated in specific assumptions on technical characteristics and changes thereof when introducing technology information into energy-economy models. Likewise, endogenous change may be included in assumptions on emissions factors.
Based on the approach(es) taken, implementation chosen, and in light of the challenges, uncertainty and limitation in data, parameters, and model solutions issues as well as in light of the raised strengths and limitations of each approach, researchers have found that endogenizing technical change leads to either reduced costs of climate change mitigation or increased costs. Almost all of the above-cited studies conclude that the implication of endogenizing technological change is large for both the optimal timing of mitigation measures and the costs of such policy measures. Clarke et al. (2006a) point out that "models are not meant for prediction but for enhanced understanding. […] different approaches have important insights and stories to tell about how technology might evolve in the future and how it might be influenced by actions to address climate change or other environmental issues. At the same time interpretation of model results and information for decision-making should be taken with care so not to over-extend the implications of modeling exercises". Overall, it is an enormous challenge to incorporate endogenous technological change from different sources, and most importantly complex and complementary interactions thereof (Clarke et al., 2006a).
In this dissertation, I pick up on the issues discussed above and present innovative ways of including innovation and technological change in energy-economy models. I illustrate two new approaches of treating technology innovations, and changes thereof, and provide four different applications of these approaches to analyze the impacts of climate policy and innovation on economic activity, energy transformation and consumption, and associated environmental impacts. This dissertation, thus, covers important methodological issues as well as policy relevant aspects of innovation and climate change mitigation. It reflects on the questions of (1) how to introduce innovation and technological change in a computable general equilibrium (CGE) model as well as (2) what additional and policy relevant information is gained from using these methodologies.
My dissertation follows a cumulative approach and provides four studies that are linked by addressing these questions. Two novel modeling approaches of technological change are developed and applied. First, a hybrid approach of incorporating technology specific information for energy supply in a dynamic multi-sector computable general equilibrium model (CGE) is used to analyze the economic, energy, and environmental consequences of mitigation policies. This approach is then extended to account for specific technology descriptions in energy-intensive production, as well as to provide a detailed comparative economic analysis of a broad range of greenhouse mitigation classes. The second approach implements learning-by-doing effects in upstream production sectors that produce machinery and equipment and deliver capital goods to the energy sector in a multi-region multi-sector dynamic CGE model.
The first approach developed in this dissertation, the hybrid approach, addresses the gap between bottom-up and top-down models. It introduces the richness of engineering characteristics of key technologies to a CGE model, yet allows for a full general equilibrium analysis of energy or climate policies. It works at an intermediate level of technology detail, between the traditional aggregate production functions of top-down models and the extensive technology detail used in bottom-up models. The approach permits a choice between several technologies and allows for shifts in technology characteristics over time towards best practice, innovative technologies. Shifts in energy consumption, in response to changes in energy or CO2 prices, are consistent with shifts between technologies. This is important for both baseline and policy scenarios. Allowing for shifts in discrete technologies provides flexibility for future technology development to be decoupled from the base year structure. Further, improvements in technology characteristics can be based directly on engineering knowledge and projections.
The second approach aims to provide more insights into the effects of technological change, in particular learning-by-doing, in industries that are not immediately affected by climate policy but are responsible for delivering capital goods used in the energy sector. It goes beyond the conventional way of introducing learning-by-doing in energy or electricity producing sectors by separating out the impact of learning-by-doing in economic activities that are located further up in the production chain (such as machinery and equipment that produce renewable energy technologies). Two main effects take place by introducing learning-by-doing in the upstream machinery and equipment industry. Firstly, learning-by-doing leads to a reduction of the unit costs of equipment, which will, via capital goods (investment), translate into reduced costs further down the production chain (e.g. in electricity generation). The second effect relates to international trade. Machinery and equipment technologies are produced for either domestic demand or for exports. Learning-by-doing induced by domestic policies may improve the competitiveness of domestic producers, lead to a higher demand for these technologies and result in higher learning effects with its subsequent effects on costs and prices. This increases the international competitiveness and stimulates national and international demand for this machinery and equipment technology, which then again would induce higher learning. An analysis of learning-by-doing effects in downstream production (e.g. electricity) alone is not able to take account of these international trade effects.
In the chapter II of this dissertation, the technology-based approach is introduced in a multi-sector dynamic computable general equilibrium (CGE) model for Germany, the Second Generation Model (SGM). The focus is placed on advanced electricity technologies and their role within a future German electricity system. This analysis is based on the recognition that substantial mitigation opportunities exist in the electricity sector through the introduction of advanced technologies. Therefore, it models the response of greenhouse gas emissions in Germany to various technology and carbon policy assumptions over the next few decades. In particular, the analysis simulates the potential role of four advanced electricity technologies, advanced pulverized coal (PCA), coal integrated gasification combined cycle (IGCC), natural gas combined cycle (NGCC), with and without the option of carbon capture and storage (CCS), and wind power from the present through 2050. In the baseline scenario, all of the advanced technologies except CCS provide substantial contributions to electricity generation. CO2 policy scenarios are conducted to provide an estimate of the cost of meeting an emissions target, and the share of emissions reductions available from the electricity generation sector.
The second study (chapter III) provides an application of the technology-based approach to an energy-intensive production sector (iron and steel) and explores how this method can improve the realism of energy-intensive industries in top-down economic models. The SGM model is modified by replacing a conventional constant-elasticity-of-substitution (CES) production function with a set of specific technologies. The response of the iron and steel sector to a set of CO2 price scenarios is investigated under the traditional production function approach in CGE models and an approach with separate technologies. The technology-based, integrated approach permits a choice between several technologies for producing iron and steel and allows for shifts in technology characteristics over time towards best practice, innovative technologies. In contrast to technology-based partial-equilibrium models, the general equilibrium framework allows us to analyze interactions between production sectors, for example between electricity generation and iron and steel production, investigate simultaneous economy-wide reactions and capture the main driving forces of greenhouse gas emissions reductions under a climate policy. It can be concluded that technology-specific effects are crucial for the economic assessment of climate policies, in particular the effects relating to process shifts and fuel input structure.
The third study (chapter IV) provides a systematic analysis of options to mitigate greenhouse gas emissions in Germany, across a variety of climate policy scenarios. At least four classes of greenhouse gas mitigation options are available: energy efficiency, fuel switching, CO2 capture and storage, and reductions in emissions of non-CO2 greenhouse gases. These options vary by cost, timing, and our ability to represent them in an economic analysis. The analysis is done with the Second Generation Model (SGM) and embodies energy and other greenhouse gas mitigation possibilities. Policy scenarios are formulated as a change in the levels of the price for greenhouse gas emissions, either applied economy-wide or targeted at energy-intensive sectors of the economy according to the EU emissions trading scheme. The methodology relies on engineering descriptions of electricity generating technologies and how their competitive positions vary with a CO2 price or change in fuel price. Energy efficiency options are represented in the standard CGE format, where non-energy inputs can be substituted for energy inputs within economic production functions or consumer demand equations, as the price of energy increases. The analysis shows that the electric power sector provides substantial opportunities for fuel switching and the deployment of advanced electricity generating technologies, with and without CO2 capture and storage. Furthermore, it accounts for reduction of emissions of non-CO2 gases, which adds a set of mitigation opportunities not usually included in energy-economic modeling efforts.
The fourth study (chapter V) puts a focus on renewable energy and learning. In this chapter, alternative ways of modeling learning-by-doing in the renewable energy sector are analyzed within a top-down multi-region multi-sector CGE model, LEAN_2000. Conventionally, learning-by-doing effects in the renewable energy sector are allocated to the production of renewable based electricity. The study builds on the observation that learning-by-doing also takes place in sectors that deliver capital goods to the renewable electricity sector, in particular in the production of machinery and equipment for renewable energy technologies. Therefore learning-by-doing is implemented alternatively in the renewable energy equipment industry and in renewable electricity production and it is shown why it matters to differentiate between these two approaches. The main differences originate from effects on international trade, since the output of the machinery and equipment sector is intensively traded on international markets unlike renewable electricity. In addition to international trade of a specific good, such as renewable energy equipment, knowledge and technical know-how about this good, which is responsible for learning processes, can spill over from one country to another. Depending on how such spillover effects are treated substantial effects on domestic production and exports patterns can be observed and are analyzed in this chapter.
Although the individual studies are done in collaboration with research partners, I myself am responsible for the main and substantial parts. This relates to methodological research, model implementation, data collection, model runs, interpretation of results, conclusions and write-up. In particular, I set up the German version of the SGM model, implemented the technology set-up for both the electricity and the iron and steel sector, included greenhouse gas mitigation options, conducted the analysis, interpreted the results and wrote and illustrated the papers. For the application in the LEAN_2000 model, I implemented the learning-by-doing approach in the two alternative sectors, collected and adjusted the appropriate data, conducted the analysis and sensitivity runs, compared, analyzed and illustrated the results, and wrote the paper. Any remaining inaccuracies in this thesis are my responsibility.
This work would not have been possible without the support and help of many kind people. Foremost, I would like to thank my advisor, Prof. Dr. Claudia Kemfert, for her ever-supporting feedback, motivation, and inspiration on various drafts and presentations of this work. In addition, I am grateful to my second advisor, Prof. Dr. Heinz Welsch, for making his model, LEAN_2000, available to me, providing helpful model tutoring in the early stages of the analysis, and providing valuable comments and hints throughout the process.
Moreover, I would like to thank my co-authors and research partners, Ron Sands and Michael Kohlhaas, for providing mental, moral and tangible input, support and feedback to the various chapters of this thesis. In particular, I would like to thank Ron for his numerous travels to Berlin, and for our innumerable phone conferences, that moved this work forward and strengthened our combined efforts to turn it into valuable journal publications.
I would very much like to thank the German Ministry for Education and Research (BMBF) and its socio-ecological research framework (SOEF) for providing kind funding for this thesis within the research project Transformation and Innovation in Power Systems (TIPS). Conducting this work as an individual qualification endeavor within the TIPS project, and within its interdisciplinary team of young researchers, has been refreshing and enlightening, motivating, and stimulating, and every now and then slightly puzzling, throughout the whole process. Thanks to the TIPS team.
Many more colleagues, friends and beloved ones contributed to making this thesis a successful project. I would dearly like to thank them all. My highest gratitude goes to my parents who provided me with a background and education needed to accomplish this project, and who were always there for me. In particular, they were available to take great care of my son at all those times when I had to rush to meet deadlines, needed longer hours at the office or went away to attend conferences. This included many, many long train rides for them across Germany, which they took with great easiness and, in some emergency cases, even on a spontaneous basis. In the same way, I would like to thank my son, Jonas, who is as old as this thesis project and who provided a wonderful balance in my life and always managed to put a happy smile on my face.
1 In IPCC terminology, ‘very likely’ means a greater than 90 per cent probability of occurrence.
2 In the literature, the term energy efficiency is often broadly used. It may encompass everything from pure efficiency increase (reduced use of one or all inputs) to fuel switching, output adjustment and sectoral shifts in the economy, i.e. it may include (1) and (2) of the above.
3 In economic terms, this refers to a shift in input use along the unit isoquant.
4 In economic terms, it represents an inward shift of the unit isoquant.
5 To date, there is still a lag between state of the art research in the economics of innovation and in economic modeling of innovation, which may be explained by a number of incompatibilities of the methodological approaches (Köhler et al., 2006; Nemet, 2006). For example, the economics of innovation emphasizes the impact of uncertainty in heterogeneous firms and path dependent technological development, adoption, and diffusion, which are difficult to implement in a more stylized and aggregate applied economic model with its common assumption of a representative producer and consumer. (Köhler et al., 2006; Freeman and Louca, 2001). The literature on innovation is characterized by its richness of description, by case study approaches, and rigorous empirical observation (Nemet, 2006). It provides only a limited set of methods with which to assess changes in technologies. Optimization and simulation models, however, have to deal extensively with uncertainties relating to model parameters and, future development thereof, and require reliable quantitative estimates, which are difficult to arrive at.
6 In contrast, induced technological change refers to an alteration in technological change (additional or different technological change) in response to a (climate) policy or set of policies. Its focus, thus, lies on understanding the effects of specific policy measures.
|DiML DTD Version 4.0||Zertifizierter Dokumentenserver|
der Humboldt-Universität zu Berlin