Skip to main content

Modelling the transition to sustainable economic structures

Risultati finali

Today overwhelming evidence exists that mankind is modifying the Earth s environment and is provoking an increase of the average global atmospheric temperature and the associated detrimental effects of regional and local climate change (IPCC, 2001). In order to minimize the risks induced by substantial climate change (UNFCCC, 1992), carbon dioxide concentrations should be stabilized, preferably during the 21st century and probably at a level not exceeding much more than twice the pre-industrial level (IPCC, 1996). There-fore, anthropogenic greenhouse gas emissions should be reduced substantially below the levels that would be implied by a ¿business-as-usual¿ scenario. This challenge, however large, can be met. Many different measures must be exploited simultaneously to realize it, among which decreasing the levels of carbon dioxide emissions per unit of energy use. Whatever means may contribute to alleviating the global warming problem, a (partial) decarbonisation of energy use seems a necessity. For de-creasing the carbon intensity of energy consumption, no panacea exists. Hence, all non-carbon emitting options should probably, for the moment at least, remain part of an energy mix as diversified as possible. Therefore, expanding the decarbonisation of fossil fuels to options beyond a transition from carbon-intensive fossil fuels (coal and oil) to carbon-poor ones (natural gas) is currently receiving enhanced attention. This paper briefly describes some aspects of carbon capture and storage (CCS) technology, since it has relatively recently entered the scene as promising option to reduce car-bon dioxide emissions, and is at present receiving increasing attention by scientists and policy makers the like. Some of the eleven models of the TranSust project include CCS, as well as renewable technologies, to investigate their potential long-term significance in energy and climate change scenarios. This paper provides a comparison, on the basis of the answers given to questions posed in a questionnaire distributed among the TranSust project members, of how some of the TranSust models simulate CCS and renewable energy technologies. Conclusions and recommendations are given for future model development regarding CCS and renewables. The first major conclusion is that so far only few of the eleven TranSust models include CCS or renewable energy technologies. In view of the ongoing TranSust work on expanding the sustainability features of existing economy-environment-energy models, it is recommended to adapt those models that do not include CCS and renewables to account for these technologies. But also for the few TranSust models that incorporate already CCS and renewables, further research is needed in terms of the refinement and improvement of their simulation of these technologies. An example in case is that today insufficient attention is being paid to possible external environmental effects of carbon storage. Ideally, it should be attempted to include such impacts in the present generation of integrated assessment models.
One key variable in the analysis of climate policies is the marginal cost of reaching a certain emission target. Marginal abatement cost (MAC) curves for CO2 have become a standard tool for analysing the impacts of the Kyoto Protocol and emissions trading. Once such curves are available for different world regions it is very easy to determine permit prices, total abatement costs and regional emissions for different scenarios of international emissions trading. In general, the MAC of reaching a certain CO2 target is defined as the shadow cost of a certain constraint on carbon emissions for a given region and a given time. This shadow cost is equal to the tax that would have to be levied on the emissions to achieve the targeted level or the price of an emission permit in the case of emissions trading. The more severe the constraint, the higher the MACs are. Or, put differently, a higher MAC corresponds to higher emission reductions. MAC curves are obtained either by generating the MACs associated with different levels of reductions or the other way around by levying different levels of a “shadow carbon tax” on emissions that result in different corresponding emission levels. The former approach is most often used in top-down macroeconomic models such as computable general equilibrium (CGE) models. An example are the MAC curves generated from the EPPA model of the MIT (Ellerman & Decaux 1998). The latter approach is most often used by bottom-up energy system models. An example are the MAC curves from the energy systems model POLES (Criqui et al. 1999). The TranSust modelling experiment where the different TranSust Models generated the emission levels associated with different levels of CO2 taxes is also an example of generating MAC curves based on the second approach . The generation and use of MACs and MAC curves is nevertheless not as unproblematic as it seems. One major problem noted by Klepper & Peterson (2003, 2004) is that MACs associated with a certain emission target are not independent of the underlying policy scenario and especially the associated energy prices. There are two important issues in this context. First, restricting emissions leads to a reduced demand of fossil fuels, which in turn drives down the price of fossil fuels or more generally energy prices. Thus, part of the emission tax is set off by lower net energy prices. Second, the prevailing energy prices determine the cost of a certain emission constraint respectively the emission reduction resulting from a certain emission tax. Briefly, the line of argument behind this is as follows. The shadow cost of reaching a certain emission target depends on the unconstraint or business-as-usual (BAU) emission level. For CO2 emissions, this is closely linked to the use of fossil fuels, so that BAU emissions depend on BAU demand for fossil fuels, which in turn reacts to fossil fuel or more generally energy prices. With higher fossil fuel prices, demand for fossil fuels and thus the resulting CO2 emissions are lower than with lower fossil fuel prices that lead to an increased fossil fuel demand and increased CO2 emissions. Thus, the MAC for a certain target depends on the underlying energy prices. To complicate the case even more, energy prices are predominantly determined by world market conditions. However, these conditions are influenced by the different national climate policies through the just mentioned demand effects of emission constraints. Therefore, the MAC curve in one country depends on the mix of climate policies in the rest of the world that co-determines fossil fuel prices. This linkage between energy prices, international climate policies and the level of the MAC for reaching a certain emission target has consequences for the MACs of climate policies resulting from different models. For example, a model with fixed energy prices will show a stronger reaction to a global emission tax than a CGE model with flexible energy prices. In the latter case, energy prices are driven down by the reduced global demand for fossil fuels, which partly offsets the emission tax, so that the gross price for fossil fuels is lower compared to the model with fixed, exogenous energy prices. Also in a multi-regional model with flexible energy prices, the cost of reaching a certain emission target in one country, say Spain, depends on the assumptions on the climate policies in the rest of the world. Thus, a single-country model for Spain with exogenous energy prices will show different emission reductions resulting from a certain emission tax than a regional, but global model that assumes that the tax is levied in all European countries. In this paper, we want to discuss the implications of the link between energy prices and marginal abatement costs for the results of the different types of climate-economy models that are commonly used for climate policy analysis.
The project focused on three issues, which were identified as key elements for improving the capabilities of current models to deal with the concept of sustainability: - Modelling the type and causes for technological change. The transition to sustainable economic structures requires investment activities, which cause the kind of technological change that is considered as being compatible with sustainability. We aim both for a better understanding of the interaction between investment and technological change and the long run consequences for economic activity. - Developing measures of welfare, which are based, not only on flow variables (as GDP or consumption) but also include the relevant stocks. It is the phenomenon of extreme events, as flooding and storms, which reminds that measuring the welfare of an economy just by flow variables misses the losses of capital that are triggered by catastrophic events. In addition we recognize that it is the services generated by flows and stocks that are the actual indicators relevant for measuring economic welfare. - Emphasizing the interaction between flows and stocks and their potential for substitution. After having identified the concept of welfare generating services, e.g. mobility, housing or information, we investigate the drivers that determine the ratio and relationship between the stocks and flow variables. The project approaches these objectives by preparing a set of crosscutting papers. A standardized questionnaire was created to which all partners provided inputs. The questionnaire was organized along the following issues: - Technical progress In the very-long run it is the deliberate choice of type and magnitude of technological innovation that determines economic welfare and climate relevant emissions. - Revenue recycling and the labour market Environmental taxes and charges are investigated in their impact on economic activity and emissions when various tax recycling regimes are applied. - The role of energy prices for assessing the costs of climate policies The linkage between energy prices and international climate policies generates multiple feedbacks via the global markets for fossil fuels. - Measuring the cost of climate policies: Interpretation issues Communication failures about a variety of different cost concepts might be considered as a major reason for disagreements in the design of climate policy. - Carbon capture and storage (CCS) and renewable energy technologies The interaction of technologies for renewable energy and the technology option for carbon capture and storage need a careful integration into the currently available models. - E3-models for sustainability impact analysis: Status quo and prospects The role of currently available models is investigated for providing quantitative indicators suitable for Sustainability Impact Assessment (SIA). - Stocks and flows The longer the time horizon for policy analyses, the more important is the role of a wide range of capital stocks for generating economic welfare and greenhouse gas mitigation. - Welfare measures, consumption and environmental quality For the groups of aggregated neoclassical growth models, disaggregated CGE models and disaggregated macroeconomic models the advantages and shortcomings of measuring economic welfare and environmental quality are compared. - GDP and emissions effects from carbon taxes For a wide range of models their reaction with respect to different carbon tax intensities and tax recycling options are analyzed. The model by model results of this questionnaire are laid down in this cross-cutting paper and made available on the project website.
From the beginning of the project all results were communicated immediately at the project website www.transust.org. The website is open to all people interested in the transition to sustainable economic structures. It serves as a valuable communication tool and will ensure continued dissemination of the project results after finalizing the project.
Since the late eighties, climate policy debates have made an extensive use of modelling results about the costs of meeting a given climate objective. To what extent this endeavour succeeded in putting some rationale in discussions is not so obvious: first, from COP3 (Kyoto) to COP6 (The Hague), it did not manage to create a common understanding between the optimists and the pessimists about the costs of Kyoto targets; second, and per-haps more importantly, it failed, despite the attempts of the 2nd and 3rd reports of the International Panel on Climate Change (IPCC), to clarify what “good use” could be made of results discrepancies in designing a viable climate regime; third, it proved to be non convincing about any robust and consensual qualitative insights that could be derived from modelling exercises regarding the policy mix most likely to minimize welfare costs. Ultimately, negotiations were conducted under pure diplomatic rhetoric with almost no link to well-grounded, even though controversial, economic analysis. Reporting the costs of carbon policies is thus currently made under the threat of a sequence of two sources of bias. The first source of bias is the very nature of the model from which the cost estimate is extracted. Three broad modelling paradigms - the disaggregated technology-rich bottom-up model, the multi-sector top-down model (be it based on micro-economics or econometrics) and the single-agent optimal control model - provide indeed three different types of cost, in close connection with the way in which they ac-count for production and end-use technologies. When attempting to compare the meaningfulness or legitimacy of these types of cost, a manner of trade-off appears between “explicitness” (how easy is it to embrace the explicit expenses summed up in these costs?) and “comprehensiveness” (to what extent do these costs express the full economic burden of a carbon policy?), with the necessary conclusion that all three assessment natures should participate to the careful weighing of any policy proposal. Another source of bias lies in the choice of a reporting template, i.e. in the manner in which modelling results are eventually summed up. The number of dimensions to this choice opens a wide range of possible templates. To any subjectivity, it thus offers the tantalizing possibility to shape a message consistent with its ex-ante conviction, be it optimistic or pessimistic. Clarifying the comparison of two cost assessments expressed in different metrics does not raise conceptual difficulties: it simply requires to focus on one of the two metrics, and to translate the assessment that was not originally expressed following it. However, in most instances, this cannot be done ex-post, i.e. without re-running the model that produced it, an option seldom available. Under this double threat, communicating the costs of climate policies appears quite a challenge, to both the scientific and decision-making communities indeed. On the one hand, experts should bring out rather than hush the complexities exposed in this paper, obviously taking the greatest care in the act to avoid raising more confusion than they dissipate. On the other hand, the decision-making community should accept that the answer to the question they raise is fundamentally more complex and less straightforward than they would want it to be, on the simple ground, ultimately, that it is manifold in itself.
As a second step for obtaining a better understanding of the different model structures and their impact on policy analysis a set of standardized simulations was agreed upon. The following procedure was applied. First for each model a business-as-usual scenario was made up to the year 2050. Second carbon taxes were increased gradually up to 2050 to the final rates of 10, 20, 30, and 50 per ton of CO2. Finally these simulations were carried out with and without recycling assumptions. The results of these joint model simulations are documentated in the working paper “Model Comparison Part 2 - Model Simulations” and are available from the TranSust website.
This paper considers how the treatment of revenues from environmental taxes and charges affects the results from models designed to address environmental problems, such as mitigation of climate change. First, we review the literature on revenue recycling, and then we report the quantitative results from meta-analyses of the costs of mitigating climate change, which show that the use of revenues from carbon taxes is an important influence on the costs. We explain the treatment in the TranSust models and finally discuss the mitigation scenarios and provide some explanation for differences in the results. There are two clear groups of models associated to the way they treat unemployment and the factors that determine employment demand. In the first group, employment and other elements of the labour market are treated more explicitly; in particular the labour market would respond to changes in social security contributions by having an impact on the cost of employment. In the second group of models, employment is given exogenously and often linked to business-as-usual assumptions about population or labour force growth. In some cases, population, employment and labour force are the same In the first group of models, the labour input necessary to satisfy the demand for a given product (or group of products) in a given geographic unit and/or industry category responds to variables such as relative labour costs, technological growth and demand for the products. In addition, wages respond to prices and unemployment and normally prices of products will incorporate wages as one of the production costs. The main research conclusions are: - The meta-analysis of post SRES scenarios indicates that the use of revenue recycling is an important factor in reducing the costs of mitigation in carbon tax scenarios. - Many models in the TranSust project do not including the labour market explicitly, so cannot address the issue of revenue recycling through reductions in employment taxes and charges, such as employers social security contributions. - Those models that do treat the labour market explicitly are also associated with higher product substitution and the modelling of dynamic behaviour of economies. - From the scenario results, the form of recycling (social security or lump-sum) has an impact on the cost of a carbon tax as measured by loss of GDP. - The social security recycling produces the extra benefit of stimulating the labour market thereby reducing overall distortions, providing examples of a small double dividend.
This paper explores the effects of CO2 taxes on CO2 emissions, energy use, and GDP. It relies on a simulation exercise carried out as part of the TranSust network. In which eleven research partners cooperate on developing the next generation of economic models explicitly targeted to sustainability issues. For this paper, we grouped the models in computable general equilibrium (CGE), econometric, growth, and Energy System models. This break up allows us to see common patterns among models of the same kind, when assessing, for example, the marginal abatement costs of emission reductions. The aim of this study is to explore various models employed in Europe for the economic analysis of sustainability. For our information, we largely base our judgment on graphs on replies from the model makers to questions raised during various project meetings and contact by email. A more exhaustive comparison and assessment is an obvious objective of a future project to be defined. We look at the (marginal) abatement costs from an aggregate perspective. More precisely, we define (marginal) abatement costs as the GDP loss (per ton CO2 reduced), following Weitzman and others who argue that GDP is the most proper measure available for welfare when measuring at one instant in time only (Weitzman, 1976). We note that our (marginal) cost estimates deviate from private (marginal) costs of emission reductions, where the latter is better captured by the level of carbon tax or the price of emission permits, since these price signals measure the marginal costs with which the individual is confronted, when deciding to emit more or less carbon. In addition we focus on the mechanisms of emission reductions. We analyze the paths for the carbon intensity of energy and energy intensity of output for different carbon tax scenarios. The issue here is whether emission reductions are achieved through energy savings, that is, an enhanced decrease in energy intensity, or through a switch from high-carbon to low-carbon fuels, that is, an enhanced decrease in carbon intensity, as a reaction to the imposition of a car-bon tax. Whether emission reductions increase or decrease GDP cannot be linked to the model type, but is linked to the assumptions about how the income from the carbon tax is used. Only under tax recycling but not in all recycling schemes sometimes a double dividend is found. On the other hand, no model produced an increase in GDP without a smart tax-recycling scheme. The econometric models stand out as a group in two ways. First, emissions are relatively inelastic. Reduction levels reached through the common carbon tax scenarios are systematically lower for the econometric models, when compared to the other models. Second, and logically following there from, marginal income effects in terms of changes in GDP divided by changes in emissions are systematically higher. As to the mechanisms of emission reduction, most models give an important role to energy savings. Some models also describe fuel switching between carbon-rich and carbon-poor fossil fuels, from coal to gas. Only a few models also describe the substitution of non-carbon fuels for fossil fuels, and in these models, decarbonization plays a major role for deep cuts in emissions.
This working paper deals with a comprehensive overview on the state of modelling sustainable economic structures, including both a comparison of the theoretic frameworks and an analysis of the policy relevance by: - Presenting and comparing the basic theoretical model designs, - Testing the robustness of different models, - Identifying the methodology and scientific achievements Based on a standardised questionnaire a comprehensive comparison of the models involved in TranSust was made which is organized as follows: - General description: Type and primary aim, distinguishing feature - Data and variables: Variables, equations, parameters, projection period - Coverage: Geographical and sectoral coverage - Economic agents and markets: Factors, products, trade - Dynamics and technology: Technological change, interaction between stocks and flows - Policy issues: Economic instruments and policy responses - Policy evaluations: Measuring economic welfare and policy impacts - Additional information about the models: Software, algorithms, references The results of this comparative analysis of model features is documented in this working paper and available from the TranSust website.
How can we meet today’s needs without diminishing the capacity of future generations to meet their own ones? This question characterizes the challenge of sustainability, which during the last decades has become a more and more important guideline for economic, social and environ-mental processes. Indeed, the concept of sustainable development was from its very beginning meant to be relevant for a comprehensive philosophy including apart from environmental aspects a variety of issues. In fact, the pioneering work of the World Council on Environment and Development (WCED, 1987) refers to sustainable development as development that meets the needs of the present without compromising the ability of future generations to meet their own needs. With this famous report Our Common Future, the Brundtland Commission placed sustainability on international political and scientific agendas. Notwithstanding this broad definition, many political discussions initially have adopted a relatively narrow focus, concentrating mainly on areas where sustainability can be defined directly or exclusively in terms of some specific environmental problem (see e.g. discussion in Pezzey, 2001). A reason for this behaviour lies in the possibility of explaining the underlying idea by focusing on a single environmental issue. However, in order to appropriately implement the concept and lead the world towards a sustainable path, the wider notion of sustainability needs to be taken into account, acknowledging thus the original intention of the WCED pioneers. For more than a decade, the European Union (EU) has taken a leading role in the promotion of sustainable development (SD), as is emphasized by various key political decisions starting from the Treaty of Maastricht (1992). At the Lisbon Summit in March 2000, a new strategic goal for the European Union was established. The European Council formulated a ten-year strategy to make the EU the world's most dynamic and competitive economy. Under the strategy, a stronger economy will drive job creation alongside social and environmental policies that ensure sustainable development and social inclusion. The Lisbon Strategy thus touches on most of the EU's economic, social and environmental activities, thereby strengthening the objective of sustainable development with a special focus on competitiveness. In the sequel at the Gothenburg Summit in June 2001, the European Strategy for Sustainable Development (European Commission, 2001) was adopted. This strategy aims at a restructuring of the European economy by means of integrating economic welfare, environmental integrity and social coherence. The transition to these innovative economic structures poses a major challenge to economic policy design. Sustainability is also high on the policy agenda outside of Europe. The World Summit on Sustainable Development in September 2002 emphasized the links to economic development, economic security, dissemination of technologies, and the social issues of health and aging in a world that is growing in population. These goals overlap with the targets put forward by the United Nations Millennium Project. After decades on research on sustainable development, the euphoria for sustainability is thus experiencing a new peak in policy circles. However, recent experiences contemporarily show that the implementation of all these sustainability strategies is difficult. An example in this context is the Lisbon Strategy, which is a commitment to bring about economic, social and environmental renewal in the EU. The European Commission's annual Spring Report examines the strategy in detail. The recent 2004 report acknowledges progress in certain domains, emphasizing however significant problems which hold back the entire strategy. Therefore, the need for an energetic implementation of reform in all the different spheres through integrated strategies is stressed. Indeed, insufficient implementation of the Lisbon strategy could produce significant net costs for Europe, e.g. in terms of reduced economic welfare and a growing gap with some of the large industrial partners in the fields of education and R&D. In order to promote progress towards the Lisbon targets, better ways of incorporating the broader aspects of sustainability are required. For this reason, the objective of this paper is to explore a potentially successful strategy to design sustainability policies, taking sustainability aspects and requirements appropriately into account.
Despite the well-known definition of sustainability in the Brundtland Report (WVED, 1987) it is still controversial how this quality of an economy could be put into a coherent frame-work of economic analysis. We attempt, therefore, a rather pragmatic approach, by making the following propositions: Sustainability is a long-run issue since it is dealing with the state of an economy and a society in the long-run. We are, therefore, interested in the available range of long-run structures of an economy. The long-run structures of an economy are manly described by a list of stocks, e.g. - Human capital of different qualifications - Capital that is reproducible by economic production, as machinery and buildings, - Capital that is reproducible by biological production, as biomass, - Capital that is irreproducible at least in non-geological time scales, as fossil fuels, - Capital that is related to the state of nature, as the atmosphere, water and soil, But also fauna and flora and - Knowledge capital is intimately linked to the use of all other types of capital. Related to the stocks is a list of flows that describe the intensity of economic activ-ity, as production, consumption, investment but also emissions. These flows are linked to stocks either as complements or substitutes. Given the wide range of choices for economic structures that are described by flows and stocks there is a need for evaluating different flow-stock structures. I coin this approach to sustainability the structural approach in contrast to the ecological and the welfare approach, each of which emphasizes only a particular aspect of an economic structure, as the state of the environment or intergenerational distribution of welfare. Daly (1990) and Constanza (1980) are the pioneers of the ecological approach to sustainability, Chichilnisky (1998) and Stavins, Wagner and Wagner (2004) emphasize the welfare implications of different economic structure. Inherent to the structural approach is the understanding of the long-term dynamics that drive the evolution of economic structures. Surprisingly these drivers are almost only decisions about stocks, as investments in human, reproducible and knowledge capital and decisions about the use of exhaustible stocks and the stocks of the biosphere. This overview about the use of stocks in TranSust models reveals a number of insights that can be summarized as follows: - It is surprising that the list of stocks included in the models is rather limited. The ranking of occurrence is reproducible capital for production, human capital, and reproducible capital for consumption. - Human capital is only in a very few models differentiated according to skills. The formation of human capital by reproduction, migration and participation rates is hardly visible but is an obvious pre-requisite for dealing e.g. with the issues of aging. - Reproducible capital for production is the dominating capital stock in most models but would de-serve further disaggregation at least into buildings, machinery, and vehicles in order to obtain better evidence as to the opportunities for substitution against flows, in particular energy. - Only very few models report reproducible capital stocks for consumption. This is a serious drawback since it seriously limits the usefulness of these models for dealing with energy since almost two thirds of energy flows are related to consumption activities. - Natural capital is dealt with only in terms of the concentration of GHG emissions. This limit’s the usability of models for discussing issues of weak and strong sustainability. - The explicit formation of human capital in terms of reproduction, migration, and skills should be given high priority for enabling research on issues as aging and competitiveness.
Technological change is a major force in a country¿s economic growth. Since before the industrial revolution, economies and societies have evolved as a result of technological change. They have moved from a reliance on wind, water, animal power, and wood to reliance first on coal, and then on natural gas and petroleum. Today, many technologies utilize fossil fuels, which have led to the release of large amounts of carbon to the atmosphere, and the scientific consensus is that these releases will cause the earth¿s climate to change. Fortunately, however, technology does not stand still. Technological innovation is increasingly seen as one of the main practical keys for reconciling the current fundamental conflict between economic activity and the environment. No one really believes or is ready to accept that the solution of the climate change problem consists of reducing the pace of economic growth. Instead, it is believed that changes in technology will bring about the longed de-coupling of economic growth from generation of polluting emissions. There is a difference in attitude in this respect, though. Some maintain a faithful view that technological change, having a life of its own, will automatically solve the problem. In contrast, others express the conviction that the process of techno-logical change by and large responds to impulses and incentives, and it has therefore to be fostered by appropriate policy actions. Technological change generally leads to the substitution of obsolete and dirty technologies with cleaner ones. It must be borne in mind however, that technical change is not per se always environment-friendly, as it can lead to the emergence of new sectors and industries with new kinds and degrees of pollution problems, like the generation of new harmful pollutants. There are therefore no substitutes for policy in directing the innovation efforts toward fostering economic growth and helping the environment at the same time. All the above remarks are reflected in climate models, the main quantitative tools de-signed either to depict long run energy and pollution scenarios or to assist in climate change policy analysis. Climate models have traditionally accounted for the presence of technical change, albeit usually evolving in an exogenous fashion. More recently, how-ever, models have been proposed where the technology changes endogenously and/or its change is induced by deliberate choices of agents and government intervention. We have therefore moved or are moving toward to an endogenous and induced formulation of technical change. In particular, both bottom-up and top-down models, a long standing distinction in energy-economy-environment modelling, have been recently modified in order to accommodate forms of endogenous technical change. As it turns out, the bottom-up approach has mostly experimented with the notion of Learning by Doing (LbD hence-forth), while a few top-down models have entertained the notion of a stock of knowledge which accumulates over time via R&D spending (see Galeotti and Carraro, 2003 for a survey).
Sustainability Impact Assessment (SIA) of economic, environmental, and societal effects triggered by governmental policies has become a central requirement for EU policy design. The three dimensions of SIA are inherently intertwined and subject to trade-offs. Quantification of trade-offs for policy decision support requires numerical models in order to assess systematically the interference of complex interacting forces that affect economic performance, environmental quality, and societal conditions. This paper investigates the use of energy-economy-environment (E3) models within the TranSust project for measuring the impacts of policy interference on policy-relevant economic, environmental, and social (institutional) indicators. We find that operational versions of E3-models have a good coverage of central economic indicators. Environmental indicators such as energy-related emissions with simple direct links to economic activities are widely covered, whereas indicators with complex natural science background such as water stress or bio-diversity loss are hardly represented. Societal indicators stand out for very weak cover-age, not at last because they are vaguely defined or incommensurable. Our analysis con-firms prospects for future modelling in the field of integrated assessment that link standard E3-models to theme-specific complementary models with environ-mental and societal focus. The objective of SD needs a comprehensive methodology to perform SIA quantitatively. An issue that cannot be clearly measured will be difficult to improve. In this paper, we have investigated the use of energy-economy-environment (E3) models within the TranSust project for measuring the impacts of policy interference on policy-relevant economic, environmental, and social (institutional) indicators: Operational versions of E3-models have a good coverage of central economic indicators, whereas environmental indicators with complex natural science background and - in particular - social indicators are hardly represented. Our cross-model evaluation confirms the need for future modeling activities in the field of integrated assessment that link standard E3-models to theme-specific complementary models with environmental and societal focus. A final caveat applies: Our focus on quantitative analysis should not exaggerate the role numerical approaches can play in SIA. Policy decisions are the outcome of a broader participatory process where stakeholders and other interested parties communicate a wide range of values, perceptions and judgments to policy makers (Tamborra, 2002). Quantitative analysis if available at all can at best strengthen or weaken policy arguments, putting decision making on a more informed basis.
Most studies take the 'Brundtland definition' (WCED, 1987) as a starting point for the analysis of sustainable development, i.e. " ..development, which meets the needs of the present without compromising the ability of future generations to meet their own needs". This definition bears two direct links to welfare economics. One is covered by the inter-generational allocation problem and the other one refers to definition of 'needs' that might be approached by the microeconomic theory of utility maximization. The discussion about the two concepts of sustainability ('weak' vs. 'strong') that has been established in the literature also shows direct links to the problem of welfare measurement. The main distinguishing feature between these two concepts, namely the existence of a binding resource constraint has straightforward implications for modelling the influence of the state of the environment on welfare. Starting point of this paper is the observation that most existing models can be classified into three broad groups with different concepts regarding welfare measures: - Aggregated neoclassical growth models, - Disaggregated CGE models and - Disaggregated macro-econometric models. Of course this does not represent a clear cut classification for all existing modelling approaches. There is a group of important models following the lines of CGE modelling but only incorporating a partial analytical perspective, for ex-ample of energy relevant activities. The theoretical as well as empirical foundations of welfare measurement in these three groups of models shall then be described in a literature overview. The treatment of welfare in TranSust models is analysed in another section along the lines of this broad classification. The distinguishing features concerning the measurement of welfare can be analysed for the 7 main models in TranSust. This analysis clearly shows strengths as well as shortcomings for an adequate description of sustainable development. That can be shown by extending the analysis to the expected welfare impacts of policy measures that represent 'successful sustainability strategies' in the different models. A standard policy measure of such a 'successful sustainability strategy' could be the introduction of prices for emissions/environmental use via taxes (e.g. road pricing) combined with the recycling of tax revenues to introduce new technologies thereby augmenting capital and leading to higher investment (e.g. public transport system). The single elements of this policy mix induce different adjustment mechanisms in the model types that can be seen as 'responsible' for different welfare impact results.