Tracking embodied environmental factors in the global trade system: the case of CO 2 emissions and material flows

Tracking embodied environmental factors in the global trade system: the case of CO 2 emissions and material flows Stefan Giljum 1, Christian Lutz 2, Martin Bruckner 1, Kirsten Wiebe 2 1 Sustainable Europe Research Institute (SERI), Vienna, Austria 2 Institute for Economic Structures Research (GWS), Osnabrueck, Germany Speaker: Martin Bruckner Corresponding author: Stefan Giljum, stefan.giljum@seri.at Abstract The past years saw an increasing debate on the distribution of the environmental pressures between production and consumption. Whereas production indicators are available from national statistics and, in the case of greenhouse gas emissions, reported under the Kyoto protocol, comprehensive indicators on the environmental responsibility of consumption are still under development. Several so-called multi-regional input-output models have been presented in the literature in the past few years, aiming at calculating this type of environmental consumption indicators in a global perspective. The Global Resource Accounting Model (GRAM) is one of these multi-regional, environmental input-output models and disaggregates 50 countries plus two world regions and their bilateral trade relations. In earlier work, the GRAM model has been applied to investigate the raw materials embodied in international trade. In this paper, we present results from a research project funded by the Austrian Climate and Energy Fund, in which the GRAM model is extended by a global data set on CO 2 emissions, provided by the International Energy Agency. We calculate physical trade balances of embodied CO 2 emissions of different countries and world regions and thus identify the main net-exporters and net-importers of embodied CO 2 emissions in the world economy. As the GRAM model also includes a global data set on material extraction, we can for the first time perform a parallel analysis of both embodied material flows and embodied CO 2 emissions in the global trade system. Focusing in particular on the European economy, we identify the main providers of raw materials for Europe and investigate, whether the main resource suppliers are also those countries with the highest export of embodied CO 2 emissions. The findings have important implications for the design of policies, which aim to increase resource productivity and decrease pressures of consumption on the global climate. Keywords: Consumption indicators, CO 2 emissions, global input-output modelling, international trade, resource flows. 1 Introduction The world is facing global environmental problems, the most urging of which is climate change. These problems are of a complex nature; their impacts do not stop at frontiers. National policy measures addressing the reduction of domestic emissions or resource use may increase environmental pressure in other countries and therefore counteract global efforts. In the case of carbon emissions this phenomenon is known as carbon leakage. This term describes the relocation of production processes and therefore CO 2 emissions as a result of national policy measures which undermines emission reduction efforts and may even offset them. In order to assess the global impacts of an environmental policy international trade has to be considered. Therefore environmental accounting approaches have to be differentiated into 1

territorial accounting (emissions/material use of the domestic production), that is used e.g. under the Kyoto Protocol and the EU Emissions Trading Scheme, and environmental accounting on a consumption base (worldwide emissions/material use associated with domestic consumption). In order to account for ecological rucksacks embodied in domestic consumption, international trade and environmental impacts along international production chains have to be analysed. In cooperation with the German Institute of Economic Structures Research (GWS) SERI constructed the Global Resource Accounting Model (GRAM), a multi-region input-output (MRIO) model (see www.seri.at/gram) which allows calculating the international environmental pressure caused by domestic consumption, including CO 2 emissions and resource use. 2 Approaches to Environmental Accounting Environmental accounting systems are classified according to the concept on which they are based. In the majority of cases environmental accounting nowadays is production based (territorial accounting); i.e. all emissions emitted on national territory are considered. This type of accounting is used e.g. under the Kyoto Protocol and the EU Emissions Trading Scheme. In contrast, a consumption-based approach to carbon accounting aims to account for all carbon emissions released along the production chain of all domestically consumed products. Analogous to this, a consumption-based materials accounting has to include the raw material consumption along the production chain of all domestically consumed products. Apart from this, in the case of carbon accounting another distinction is made regarding the range of included carbon flows. Carbon emissions originate from various sources like the combustion of fossil fuels in industry or households, but also from indirect sources like carbon emissions caused by land use change. Furthermore there are many natural sources and sinks of carbon emissions which, nevertheless, are affected by human interaction. An accounting scheme may include all these sources or only selected ones. According to the scope of an accounting system it can be differentiated into full and partial carbon accounting as shown in table 1. Emission flows can be divided by their sources into emissions from socio-economic sources (emissions from households and industry), from anthropogenic sources (direct and indirect emissions caused by human action), and from natural sources (natural sources and sinks of carbon emissions). Scope Full Partial Concept Production based Consumption based Full territorial carbon accounting Full economy-wide carbon accounting Partial territorial carbon accounting Partial economy-wide carbon accounting Table 1: Classification scheme of approaches to carbon accounting 2

2.1 Territorial vs. Economy-wide Accounting The concept of territorial accounting as used under Kyoto is a straight forward approach to account for environmental impacts. It is characterized by clear system boundaries and good data availability. However, reductions in the levels of territorial environmental impacts can occur due to many reasons. One of them is the relocation of industries to other countries. A national policy designed to reduce e.g. emissions pursuing a territorial approach of environmental accounting thus may well result in an increase of global impacts. Therefore, in order to assess world-wide environmental consequences related to production and consumption of a specific country or world region, it is necessary to take trade aspects fully into account. Only thereby possible shifts of environmental burden can be illustrated, resulting from changing global patterns of production, trade and consumption. A number of studies examined the distribution of environmental pressures between different world regions due to the economic specialisation in the international division of labour, applying methods of physical accounting and environmental-economic modelling. Several studies found empirical evidence for increasing externalisation of environmental burden by industrialised countries through trade and increasing environmental intensity of exports of non-oecd countries (see, for example, Nakano et al., 2009; Peters and Hertwich, 2008; see, for example, Schütz et al., 2004). These findings become particularly relevant, as the externalisation of environmental burden through international trade might be an effective strategy for industrialised countries to maintain high environmental quality within their own borders, while externalising the negative environmental consequences of their consumption processes to other parts of the world. 3 Methods to Calculate Embodied Environmental Factors In order to consider the global dimension, ecological rucksacks of traded products were so far mostly calculated applying a life-cycle assessment (LCA)-oriented approach. Following this approach, direct imports are multiplied by coefficients (or so-called ecological rucksack factors ), reflecting all environmental impacts, e.g. all carbon emissions or RMEs, along the whole production chain of a product. However, due to high efforts in data collection, these rucksacks have so far been calculated for a very limited number of processed products. Another approach tries to calculate ecological rucksacks based on an input-output model. Environmental input-output analysis allows analysing implications of structural changes of the economy for the environment, as well as of changes in technology, trade, investments and consumption and lifestyles. One major advantage of the IO approach compared with LCA-oriented approaches is that it avoids imprecise definitions of system boundaries, as the entire economic system is the scope for the analysis. However, applying the IO approach also entails disadvantages. These refer in particular to the high level of aggregation of economic sectors in the IO tables, which lead to problems of inhomogeneities within (theoretically homogeneous) sectors. In most studies at the national level carried out so far, environmental impacts associated with imports were estimated applying the assumption of an identical production technology of the exporting country and the domestic economy (for example Moll and Acosta, 2006; Weisz, 2006). However, distortions of results can be considerable, if countries show significant differences in technology and economic structure, which is often the case, when trade relations between industrialised and developing countries are investigated (Haukland, 2004). 3

In order to overcome the shortcomings of a single-country model, in particular with regard to environmental consequences of increasing international trade, a number of studies were published in the past few years, which applied multi-regional IO (MRIO) modelling to assess environmental pressures embodied in international trade. These models permit to consider the production technologies and environmental intensities in different countries and world regions by linking national models using international trade data. Several major advantages of the MRIO approach can be identified (see Wiedmann et al., 2006): MRIO models allow for integration of (monetary) trade flows with environmental databases and permit environmental impacts embedded in trade to be accurately and comprehensively evaluated, as variations in production structures and technologies between different countries and world regions are taken into account. Different IO-based analyses on the international level can be undertaken with a MRIO model (e.g. structural path analysis, production layer composition, quantification of shared environmental responsibilities between producers and consumers of goods). With a MRIO model, direct, indirect and induced effects of international trade can be captured. 4 The Global Resource Accounting Model (GRAM) The Global Resource Accounting Model (GRAM) is a multi-region input-output (MRIO) model constructed in the course of the petre project 1, revised and extended in the course of the project Carbon balance of the Austrian foreign trade funded by the Austrian Climate and Energy Fund. The GRAM model has a monetary core for the year 2000 that links OECD IO tables and OECD bilateral trade data (BTD). This monetary core model is extended by global data sets, both on material inputs and on carbon emissions in physical units, which are attached to the IO tables as additional vectors. A detailed description of the initial model, calculating for raw material rucksacks, can be found in the corresponding methodology paper by Giljum et al. (2008). 4.1 Data sources The main data sets required for setting up the GRAM model are input-output tables, trade data, material extraction data, and carbon emission data. For constructing GRAM, international IO data provided by the OECD were used for the reason of reliability, transparency, and comprehensiveness. The latest, third revised (2006) edition of IO tables published by the OECD includes 27 OECD countries (except Iceland, Luxembourg and Mexico) and 9 non-oecd countries (Argentina, Brazil, China, India, Indonesia, Israel, Russia, Singapore and Taiwan). The tables are divided into 48 sectors and industries and based around the year 2000 (Yamano and Ahmad, 2006). An update of the database will offer IO tables for 1995, 2000 and 2005 for most of the countries. 1 Resource productivity, environmental tax reform and sustainable growth in Europe, funded by the Anglo-German Foundation (AGF). For more information about the project see http://www.psi.org.uk/petre. 4

The GRAM model comprises 52 countries and regions, with the OECD dataset providing IO tables for 35 of them. For the remaining countries and regions IO tables were derived under the assumption that the country or region under consideration holds the same production technology as a neighbouring country or a country with a similar economic structure (Giljum et al., 2008). Data on international trade, the modelling of which is the core element of a model calculating all direct and indirect material requirements of countries, were obtained from the OECD too. The bilateral trade data (BTD) of OECD conform to the used IO tables and comprise imports and exports of goods for each OECD country broken down by 61 trading partners and 25 industries. One disadvantage of the BTD data set is that it captures only OECD trade with the rest of the world, while trade between two non-oecd countries is not recorded. Thus, trade between major material consuming countries such as China and India and major material extracting countries such as Brazil, South Africa and Russia was completed by UN COMTRADE data and country by country trade data from the Direction of Trade Statistics from the IMF (2006 edition). With regard to material input data, a large and increasing number of material flow studies are available from national and international statistical offices, environmental agencies and research institutions (see OECD, 2007). The first global dataset in a time series of 1980 to 2002 was compiled by SERI in EU-funded projects (see www.materialflows.net and (Behrens et al., 2007) following the nomenclature and categorisation of materials listed in the handbook for economy-wide material flow accounting published by the Statistical Office of the European Union (EUROSTAT, 2001). This international database (currently available up to 2005) on natural resource extraction is mainly based on international statistics from the International Energy Agency (IEA), the Food and Agricultural Organisation of the United Nations (FAO), British Geological Survey (BGS), United States Geological Survey (USGS), and the German Federal Institute for Geosciences and Natural Resources (BGR). Sectoral data on CO 2 emissions were calculated based on energy statistics (energy balances) of the International Energy Agency. 5 Discussion of Results In the full paper we will present our results, i.e. the CO 2 emissions and the resource extraction embodied in domestic consumption in 52 countries and world regions, in a time series from 1995 to 2005. We will analyse where environmental impacts like resource extraction and CO 2 emissions take place versus where the final products manufactured in these production processes are consumed, and how this context changed over the analysed period. A parallel analysis of CO 2 emissions and raw materials embodied in products will offer insights in the relations of these parameters in a global perspective. Such results are also an important input to the current discussion on producer versus consumer responsibilities in the world economy (see, for example, Lenzen et al., 2006). Whereas most accounting frameworks (e.g. also in the Kyoto protocol) follow a production or territory accounting principle, a consumption-oriented accounting approach is required when discussing concepts such as an allocation of a fair share of world s resources to all inhabitants of the planet (see also Peters, 2008). 5

References Behrens, A., Giljum, S., Kovanda, J., Niza, S. 2007. The material basis of the global economy: Worldwide patterns of natural resource extraction and their implications for sustainable resource use policies. Ecological Economics 64(2), 444-453. EUROSTAT. 2001. Economy-wide material flow accounts and derived indicators. A methodological guide. Statistical Office of the European Union, Luxembourg. Giljum, S., Lutz, C., Jungnitz, A., Bruckner, M., Hinterberger, F. 2008. Global dimensions of European natural resource use. Results from the Global Resource Accounting Model (GRAM). Paper presented at the Paper presented to the OECD/UNEP Conference on Resource Efficiency, Paris. Haukland, E. 2004. Trade and environment: Emissions intensity of Norway's imports and exports. Unpublished Master's thesis, Norwegian University of Science and Technology (NTNU), Trondheim, Norway. Lenzen, M., Murray, J., Sack, F., Wiedmann, T. 2006. Shared producer and consumer responsibility theory and practice. ISA Research Paper 01-06, University of Sydney. Moll, S., Acosta, J. 2006. Environmental implications of resource use. Environmental input-output analyses for Germany. Journal of Industrial Ecology 10(3), 25-40. Nakano, S., Okamura, A., Sakurai, N., Suzuki, M., Tojo, Y., Yamano, N. 2009. The Measurement of CO 2 Embodiements in International Trade: Evidence from the Harmonised Input-Output and Bilateral Trade Database. STI Working Paper 2009/3 (DSTI/DOC(2009)3), Organisation for Economic Co-operation and Development (OECD), Directorate for Science, Technology and Industry, Economic Analysis and Statistics Division, Paris, France. OECD. 2007. Measuring Material Flows and Resource Productivity. Inventory of country activities. ENV/EPOC/SE(2007)2, Organisation for Economic Co-operation and Development, Paris. Peters, G. 2008. From Production-Based to Consumption-Based National Emission Inventories. Ecological Economics 65(1), 13-23. Peters, G.P., Hertwich, E.G. 2008. CO2 Embodied in International Trade with Implications for Global Climate Policy. Environmental Science & Technology 42(5), 1401-1407. Schütz, H., Bringezu, S., Moll, S. 2004. Globalisation and the shifting environmental burden. Material trade flows of the European Union. Wuppertal Institute, Wuppertal. Weisz, H. 2006. Accounting for raw material equivalents of traded goods. A comparison of input-output approaches in physical, monetary, and mixed units. Working Paper 87, IFF/Social Ecology, Vienna. Wiedmann, T., Minx, J., Barrett, J., Wackernagel, M. 2006. Allocating ecological footprints to final consumption categories with input-output analysis. Ecological Economics 56, 28-48. Yamano, N., Ahmad, N. 2006. The OECD's Input-Output Database - 2006 Edition. STI Working Paper 2006/8 (DSTI/DOC(2006)8), Organisation for Economic Co-operation and Development (OECD), Directorate for Science, Technology and Industry, Economic Analysis and Statistics Division, Paris, France. 6