Methods to assess global environmental impacts from Swedish consumption. Synthesis report of methods, studies performed and future development

Transcription

1 Methods to assess global environmental impacts from Swedish consumption Synthesis report of methods, studies performed and future development report 6395 december 2010

2 Methods to assess global environmental impacts from Swedish consumption Synthesis report of methods, studies performed and future development Author: Hanna Brolinson, Louise Sörme, Viveka Palm, Arnold Tukker, Edgar Hertwich and Anders Wadeskog NATURVÅRDSVERKET

3 Orders Phone: + 46 (0) Fax: + 46 (0) natur@cm.se Address: CM Gruppen AB, Box , SE Bromma, Sweden Internet: The Swedish Environmental Protection Agency Phone: +46 (0) Fax: +46 (0) registrator@naturvardsverket.se Address: Naturvårdsverket, SE Stockholm, Sweden Internet: ISBN pdf ISSN Naturvårdsverket 2010 Electronic publication

4 Preface The governmental bill 2010 Swedish environmental objectives for a more effective environmental work stated: The overall goal of environmental policy is to provide the next generation a society in which the major environmental problems in Sweden have been solved. This should be done without causing increased environmental and health problems outside of Sweden. The overall goal requires an ambitious environmental policy in Sweden, EU and international contexts. 1 Swedish Environmental Protection Agency has therefore asked Statistics Sweden to assess methods to assess global environmental impacts caused by consumption in Sweden. The project group consisted of Hanna Brolinson (project manager), Louise Sörme and Viveka Palm, at Statistics Sweden. The authors of the report is Hanna Brolinson, with input from Louise Sörme (especially the section about chemicals), Viveka Palm, Anders Wadeskog (especially the sections concerning input-output analyses), Statistics Sweden, at the Unit of environmental economy and natural resources. The researchers Arnold Tukker, Professor of Sustainable Innovation, NTNU and Edgar Hertwich, Professor, Department of Energy and process engineering, NTNU contributed with the international outlook and input on recent research in the field. Stockholm in December 2010 Swedish Environment Protection Agency 1 Authors own translation from Prop. 2009/10:155 3

5 Contents SPREFACE 3 SUMMARY 5 1 INTRODUCTION Background Purpose 10 2 METHOD 12 3 CONSUMPTION AND ENVIRONMENTAL IMPACTS 13 4 METHODS TO LINK ENVIRONMENTAL IMPACTS TO CONSUMPTION Life cycle assessment Environmentally extended input-output analyses Life cycle assessment in combination with input-output analyses 24 5 AREAS OF ASSESSMENT 26 6 RESULTS AND DISCUSSION Greenhouse gas emissions Other emissions to air Chemicals Water use Land use Biodiversity 56 7 CONCLUSIONS 60 8 ACRONYMS AND ABBREVIATIONS 63 9 REFERENCES 64 4

6 Summary The output of the project is a synthesis of methods to assess environmental impacts in other countries caused by consumption in Sweden. Over the last couple of years a number of studies have been performed testing methods and refining data sets for these types of studies. The following six areas of assessment are taken into account in this study: greenhouse gas emissions, other emissions to air, chemicals, water use, land use and biodiversity. The report includes short introductions to methods used in the different impact areas, references to performed studies and recent research. The data availability and available time series are mentioned as well as the methods strengths and weaknesses. The work is performed by literature studies, references to scientific articles, working papers and ongoing projects in Sweden and internationally. Generally, when calculating environmental impacts caused by consumption in other countries one can either use a bottom-up method or a top down method. In the first case the impacts caused during the production processes for each product that is consumed are summed (LCA based methods). In the other case a consumption view is applied, i.e. data on consumption of a whole nation is combined with environmental data for each industry branch and product categories (input-output based methods). There are also several hybrid methods, for example LCA methods that use input-output based data to fill in data gaps. Some of the methods described in this report are more suitable than others to use in further analyses, for example in designing policies. These methods are the ones which links environmental impact to the total economy in a nation. To get a useful indicator on the environmental pressures linked to consumption we recommend choosing a method which can be used in further analyzing. An input-output based model makes it possible to study what implications changes in consumption, or changes in emission intensities, can have on environmental impacts. For two of the six assessment areas there exist mature methods to assess environmental impacts in other countries caused by consumption in Sweden, namely for greenhouse gas emissions and other emissions to air. The recommended methods to use for these areas are the environmentally extended input-output model. Data and time series are published by Statistics Sweden. Even though these methods are considered mature, they are still under development and enhancements can be foreseen in a near future. To go further with these analyses, one could start looking deeper into the factors that the consumption model can provide. That means comparing results by product groups and also to focus on the improvement of environmental data that are needed to calculate country intensities, i.e. for emissions caused in the countries where the production process takes place. For emissions of chemicals we are suggesting to apply an input-output based approach, which to our knowledge, still has not been tried out for Sweden. We sug- 5

7 gest using the Pollutant Release and Transfer Register (PRTR) which contains data on 90 emissions on substances by industry in an environmentally extended input output analyses. For chemicals there is a possibility to weight the chemicals according to an existing method or to work with specific chemicals. When weighting the chemicals a toxic potential is obtained, which sums up the loads from different substances emissions. Emission intensities for different trade partners must be developed to get good result on emission in other countries. On the input side, i.e. the use of chemicals, there are experiences to use input-output analyses. For the three remaining areas, land use, water use and biodiversity there are no straight forward and ready to use methods or data sets. Instead we suggest initiating studies within the areas and point out possible ways to go forward. The assessment of these areas probably requires the dimension of the state in the environment, as the linkage between environmental impact and pressure is not always obvious. Regarding land use one can add the land use and land cover statistics to the inputoutput model. A prerequisite is to divide these data on industry branches. This has not been performed to a full scale and need to be initiated. Probably some reference to the state in the environment needs to be done, by using for example the HANPP index (human appropriation of net primary production in earth s terrestrial ecosystem), in order to get interpretable results. The HANPP index gives a measure on how much biocapacity that is removed from the available biocapacity and is based on forestry and agriculture statistics, and global data is available. To get relevant land use data in the countries where the products that we import are produced the availability on international data must first be mapped. This method has yet to be tested before we can say how the results can be interpreted. A product view can be pragmatic in this area as a starting point. Critical products could possibly be identified, products known to be linked to negative land use effects, if experts can set up relevant criteria. The Swedish trade with these products can be analyzed, pointing out how Swedish consumption contributes to changes in land use in different countries. A possible way forward would be to utilize land use data for some major categories such as agriculture or forestry in the countries from which Sweden imports goods and gradually develop a database with more environmentally relevant information on the type of land use. Regarding water use one can add water use statistics to the environmentally extended input-output model. The availability on international data must first be mapped in order get relevant water use data in the countries where the products that we import are produced. This could be combined with some kind of index which indicates the state in the environment to be useful, as in the case of land use. Such an index is for example the water scarcity index which shows the global status in water basins. One could also apply a product perspective. It is known that agricultural products are those most related to water problems globally. These products can be identified in the Swedish trade, data can be analyzed and this together with the scarcity index could say something about Swedish consumption s contribution 6

8 to water related problems. These suggested methods have yet to be tested before we can say how the results can be interpreted. Biodiversity is a complex area and is also connected to the other areas to a large extent. There is no biodiversity index ready to use and probably not foreseen in a near future. However, if the five areas above are added to an environmentally extended input-output model they probably together give a relatively good measure on biodiversity related pressures. As a starting point, however, a product view can be useful to apply. In this case critical products must be listed, if this can be done by experts, so that the study of the trade with these can be performed. The outtake of wood and fish should at least be included in such a study. As the six areas of assessment differ in method and data availability there are several options for the scope of further studies and research. Firstly one can make the most mature methods even better, to make more accurate calculations and hence focus on greenhouse gas emissions and other emissions to air. Secondly, one can initiate new studies to increase the number of extensions in the input-output based calculations. Emissions of chemicals, is an extension that seem to be possible to add to the environmentally extended input-output model under some limitations. One can also focus on those diverse areas where no exact methods are available and initiate studies or research so that these can be linked to consumption in Sweden in a near future. Land use and water use can probably also be added to the environmentally extended input-output model, whereas biodiversity seem to be more distant. 7

9 1 Introduction 1.1 Background The environmental impact of goods has become more and more in focus since the 1990s (Swedish EPA, 2002). The look on environmental problems have shifted from local point sources of pollutants to diffuse sources, from local and regional environmental problems to global problems with an increased degree of complexity and a longer time period between cause and effect. The fact that environmental impacts occur not just during production processes but also throughout the entire lifecycle of a product has become obvious. The awareness has risen of producers' ability to reduce impact throughout the lifecycle by analyzing the construction and design of the products. Consumers also have become aware of that they can influence both direct and indirect environmental impact by their choice of product. In later studies it has been shown that the level of income and gender, are factors that are strongly linked to the amount of emissions caused by consumption (Carlsson- Kanyama, 2010). The carbon footprint of a nation is also strongly correlated with per capita consumption expenditure (Hertwich and Peters, 2009). Historically environmental pressure has been viewed upon in a national, production view. This picture has changed during later years, not only due to the fact that some environmental pressures literally moves across borders but also due to a much higher awareness of that a global economy makes consumption patterns in one country linked to environmental impacts in the producer country. Hence, a consumption view is needed to complete the picture. In an open economy, such as Sweden, a part of the environmental pressures are due to exports as well as that a part of the pressures abroad are due to consumption in Sweden. In industrialized countries environmental technology is used to keep point sources of emissions down within the nation. There is still the fact that the consumption contributes to pollutants abroad, sometimes to a lesser extent than if the same goods where produced within the nation, sometimes to a larger extent. Of this follows that changes of consumption patterns might be necessary to a shift towards a more sustainable life style as emission limitation within a nation might not be sufficient. A production view and a consumption view must go hand in hand to complete the picture of environmental impacts, and to avoid that production is moved to countries where the environmental policy is less developed. When viewing environmental pressures and resource use in other countries, caused by consumptions in Sweden, the question is raised: what we can do about it in Sweden? More knowledge is needed to clarify how the consumption perspective can be used to eventually develop instruments to shift to a more sustainable life style. It would also be useful to rank, which consumer goods, that are most dependent on consumer behavior and hence would be fruitful to influence to a more 8

10 sustainable level. For example, transports and leisure items (toys, electronics) are goods that are strongly connected to consumption levels (UNEP, 2009). In the governmental bill Swedish environmental objectives for a more effective environmental work it is stated: The overall goal of environmental policy is to provide the next generation a society in which the major environmental problems in Sweden have been solved. This should be done without causing increased environmental and health problems outside of Sweden. The overall goal requires an ambitious environmental policy in Sweden, EU and international contexts. 2 (Author s emphasizing) As implied above, environmental pressures may not increase in other countries, and hence an international perspective must be applied in the work with the environmental objectives. It has been shown that in some cases, apparent decoupling 3 of CO 2 emissions or primary material use from economic growth is in fact the result of the relocation of material and energy intensive production to other countries (Giljum et al., 2008b; Wiedmann et al., 2008), a situation that must be avoided. In the 2010 assessment of the Swedish environmental objectives, a consumption perspective was implemented as a first step to meet new guidelines from the government. This was published in the de Facto report (Miljömålsrådet, 2010) Hence, the topic is high on the environmental policy agenda and a lot of studies and research have been performed in the latter years. The European Environment Agency has adopted a framework to describe the interactions between society and the environment, DPSIR: Driving forces (activities causing an environmental problem) Pressures (physical activities causing the environmental problem) States (the state in the environment) Impacts (effects or consequences) Responses (measures to decrease the problem) When assessing environmental impacts from consumption the focus is on driving forces and pressures and not the state in the environment. In practice this means that in the case with emissions of carbon dioxide, the focus of assessment is on the activity: the use of fossil fuels and the release of carbon dioxide, not the state in the environment or the impact in the environment, i.e. the concentration of CO 2 in the atmosphere or average temperature. 2 Authors own translation from Prop. 2009/10:155 3 In economics decoupling refers to the ability of an economy to grow without corresponding increases in environmental pressure. 9

11 The responses in the DPSIR model are the measures taken to decrease the environmental problem, contained in the environmental policy in the society. Different environmental policy instruments are used for different environmental impacts. These can be of different types and the most common are: legal, fiscal (e.g. environmental taxes and subsidies), voluntary policy levers (e.g. advice, information agreements). Some of the assessment methods described in this report are more suitable than others to use in further analyses, for example in policy design. These are the ones which link environmental impact to the total economy in a nation, as they enable the evaluation how changed consumption levels implicates changes in environmental pressures. The focus of climate change has put much of the spotlights on greenhouse gas emissions. However, there are other areas which also need to be assessed to get a total view of the pressures caused by consumption. At the same time, many of the environmental pressures are linked to the use of fossil fuels and hence a simplification can be practical and relevant. In the assignment of this study three starting points were given for the choice of areas of assessment: 1. The areas assessed in the consumption part of the 2010 defacto 4 report 2. Biodiversity 3. The Swedish Environmental Objectives These areas should also be complemented if other areas of assessment where found to be of relevance. Hence, as a start an overview of studies of impacts from consumption was performed. 1.2 Purpose The output of the project is a synthesis report containing methods to assess global environmental impacts caused by consumption in Sweden. The areas to be covered are: emissions of greenhouse gases, other emissions to air, chemicals, water use, land use and biodiversity of which the five first were covered in the 2010 assessment of the Swedish environmental objectives. The areas should also be complemented with other areas if found relevant and with the Swedish environmental objectives. The report includes short introductions to methods used in different areas, as references to performed studies and recent research. The methods strengths and weaknesses in its usefulness when assessing global environmental impacts from Swedish consumption will also be made. The report will also point out what important gaps

12 there are to be fulfilled before the global environmental impact from Swedish consumption can be made. The compilation should also be able to be used to evaluate how the consumption perspective can be used in the assessment of the Swedish environmental objectives. The results will also be used to fulfill the tasks given by the governmental bill, Swedish environmental objectives for a more effective environmental work

13 2 Method The synthesis is a compilation of methods, studies performed and suggestions on how to move forward in finding methods to assess global environmental pressures caused by consumption in Sweden. The work has been performed by literature studies with references to scientific articles, working papers and ongoing projects in Sweden and internationally. The recommendations on methods to use or where to focus future work are made on the basis of the gathered expertise of the project members. However, as this is not a static area, recommendations can be subject to changes due to new discoveries of for us unknown research. 12

14 3 Consumption and Environmental impacts Consumption is the final use of goods, which includes products as well as services. To keep track of the economy within a nation an internationally harmonized statistical System of National Accounts (SNA) is used. Statistics is reported to Eurostat 6 at a regular basis and used within the nation for policy making. Within the national accounts the final demand is distributed on private and public consumption, investments and exports and is met by domestically produced and imported goods, schematically shown in Figure 1. All the parts of final demand can be distinguished and handled separately in further analysis. Hence, the consumption within a nation is handled within the national accounts and is a part of final demand. Domestic production Final demand Imports Private consumption Public consumption Investments Exports Figure 1 Goods and services from domestic production and imports is distributed on final demand by private and public consumption, investments and exports. With a consumption perspective applied to environmental impact assessment the consumption of goods within a nation is concerned, whereas the environmental impact is distributed on product categories irrespectively of where in the world or where in the production chain the impact has occurred. To keep track of environmental impacts caused by consumption a system of environmental accounts can be used. The environmental accounts is a internationally harmonized statistical system that is designed to measure contributions from the environment to the economy (for example raw material, water, energy and land) and the impact on the environment from the economy (such as emissions to air, soil and water), as indicated in Figure 2. 6 Eurostat is the statistical bureau of EU 13

15 Environmental accounts Environment Raw material, water, energy and land Emissions to air, soil and water Economy Figure 2 The environmental accounts describe the contributions from the environment to the economy and vice versa. The environmental accounts also distinguish the environmental relevant activities (the environmental goods and services sector) and economic transactions from the national accounts (environmental taxes and subsidies). In the environmental accounts consumption is treated as a component of the final demand (Statistics Sweden, 2003). In year 1993 Statistics Sweden, together with the National Institute for Economic research and the Swedish Environmental Protection Agency, was assigned to create a system to describe the relationship between the economy, environmental and natural resources, in monetary as well as physical terms and the state in the environment. At Statistics Sweden the work with environmental accounts started with the development of an accounting system for physical data linked to the economic activities described in the national accounts. In practice this means environmental and natural resource statistics connected to the industries, products groups and sectors that are used in the national accounts. A system, like the environmental accounts, that is based on the national accounts can also be called a satellite system of the national accounts. In year 2003 UN published a handbook about environmental accounts, Handbook of National Accounting Integrated Environmental and Economic Accounting 2003 (SEEA 2003). According to SEEA there should be four categories of accounts: Flow accounts for pollution, energy and materials. These accounts provide information at the industry level about the use of energy and materials as inputs to production and the generation of pollutants and solid waste

16 Environmental protection and resource management expenditure accounts. These accounts identify expenditures incurred by industry, government and households to protect the environment or to manage natural resources. Natural resource asset accounts. These accounts record stocks and changes in stocks of natural resources such as land, fish, forest, water and minerals. Valuation of non-market flow and environmentally adjusted aggregates. This component presents non-market valuation techniques and their applicability in answering specific policy questions. The environmental accounts at Statistics Sweden consist of several parts and there is some developed statistics within three of the four categories listed above. In the work with this synthesis the following of the statistical production is of specific relevance: Emissions to air by industry sector Water accounts (water usage and emissions to water) Web-tool with input-output based simulation models with links between final demand, production, energy intensities and emissions In further analysis, data on environmental related taxes and subsidies (both environmentally motivated and potentially harmful) can be retrieved from the environmental accounts at Statistics Sweden. The web tool 8 at Statistics Sweden is a platform for publishing the annual environmental accounts data in an analytical and interactive format. Data on twelve pollutants divided on stationary, mobile and process related emissions is presented by industry and in the same format as the annual detailed national accounts. The time series runs from 1993 to Added to this basis are a few simulation modules based on input-output analysis. One of them is a household budget simulation module where the user can either compare emissions of CO 2 by different household types (by region, income or demographics) or create a budget of their own, or changes in a given budget, and have that translated into changes in CO 2 emissions. There are a number of possible development project for the web tool in the coming year. In a near future a linkage to, or inclusion of, environmental accounts data from other member states can be a reality. This opens up for more accurate estimates of the global effects of domestic consumption. With fairly detailed emission data per industry for our major trading partners, it will produce better estimates than has been possible to obtain until now. The environmental accounting system does also allow for further analysis of some types of driving forces, such as increased consumption, changes of energy use or

17 changes in the use and in the demand for products and services over time. It is an international statistical system that allows for analysis of changing trade patterns and other economic driving forces on the environmental pressure. It is designed to use the existing economic statistics and environmental statistics and is in this respect a system that does not entail cumbersome new surveys that increase the burden on society s information providers. Within the EU environmental accounts is developed within several countries and the EU is proposing 9 a regulation so that all member states will report environmental accounts data for some of the categories listed by the UN (SEEA 2003). The regulation is planned to be ready in The proposal: 16

18 4 Methods to link environmental impacts to consumption This chapter contains an introduction to two principal methods that are used when linking environmental impacts to consumption, the LCA and the input-output based model. In the LCA a product view is applied and this can then be scaled up to the consumption. In the input-output based the consumption is linked to emissions and the result can then be divided into product groups or categories. These types are referred to as the bottom-up method or the top-down method respectively. There are also hybrid methods, for example LCAs where missing data is modeled using input-output methods. There are also several kinds of footprint calculations when looking upon environmental impacts caused by consumption. Different footprints can be used to communicate environmental impacts caused by consumption. Common for the different footprints, for example the ecological footprint 10 and the water footprint 11, (their start pages on the Internet are shown in Figure 3) is that they aim to include all impacts related to consumption translated to a symbolic measure. The footprints can be used for assessment, and for communication, when the emissions are known. Within the ecological footprint method the term carbon footprint is also used. However, there is no unambiguous definition of a carbon footprint; it is used as a generic synonym for emissions of carbon dioxide connected to a certain product or region 12. The term carbon footprint of nations has been used by researchers describing a method to assess the emissions of carbon dioxide linked to the economic activities within a nation (Hertwich, Peters, 2009). The method used in that case was a multi-regional input-output model, further described in this report in section There are ISO standardization projects going on both for water footprint and carbon footprints of products more related to the eco-labeling incentive then the assessment of environmental impacts caused by consumption

19 Figure 3. Web pages for water footprints and ecological footprints There are a number of other methods or concepts used when dealing with consumption and environmental impacts. Two of these are for example Planetary Boundaries 13 and the concept fair share in environmental space 14. The main purpose with Planetary Boundaries is to define boundaries within nine areas which may not be trespassed without risking unacceptable changes in the environment. However, the focus is not on Swedish consumption s environmental impact in other countries and will therefore not be referred to in this synthesis. The main purpose with the environmental space concept is that there is a limited amount of resources that can be used in a sustainable way. The resource use is calculated in different areas, which goes hand in hand with the areas in this report. These two methods are more related to setting boundaries for consumption, and do not contribute with methods on how to calculate the environmental impact of Swedish consumption in other countries. 4.1 Life cycle assessment A life cycle assessment, commonly abbreviated as an LCA, is an investigation and evaluation of environmental impacts of a given product or service caused by or necessitated by its existence in a life cycle perspective, illustrated in Figure 4. Information on environmental interventions is collected at a detailed level. The unique feature of an LCA is the focus on products in a life-cycle perspective, and includes raw material acquisition, production, use to waste management (Finnveden et al, 2009). An LCA of a product can answer questions like What are the most important environmental impacts of a given product s life cycle? or Which of these modifications of a production process is preferable in an environmental perspective?. 13 More information can be found at 14 More information can be found at 18

20 Figure 4 Life cycle perspective of products (goods and services). From The method to use when performing an LCA is described in several ISO standards. According to ISO and ISO 14044, the life cycle analyses proceeds through several steps, these steps can be summarized as: Goal and scope In the first phase, the goal and scope of study is formulated and specified in relation to the intended application. The system boundary are also defined they determine which unit processes are included in the LCA and must reflect the goal of the study. Finally the goal and scope phase includes a description of the method applied for assessing potential environmental impacts and which impact categories that are included. Life cycle inventory This second phase 'Inventory' involves data collection and modeling of the product system, as well as description and verification of data. This encompasses all data related to environmental (e.g., CO 2 ) and technical (e.g., intermediate chemicals) quantities for all relevant unit processes within the study boundaries that compose the product system. Examples of inputs and outputs quantities include inputs of materials, energy, chemicals and 'other' and outputs of air emissions, water emissions or solid waste. Life cycle impact assessment The third phase 'Life Cycle Impact Assessment' (commonly referred to as LCIA) is aimed at evaluating the contribution to impact categories such as global warming, acidification, etc. It consists of several steps, characterization, normalization and weighting. Of which the latter are both voluntary according the ISO standard. During characterization each emission or resource consumption is modeled quantitatively. Normalization provides a basis for comparing different types of environmental impact categories (all impacts get the same unit). Weighting implies assigning a weighting factor to each impact category depending on the relative importance. 19

21 Interpretation The phase stage 'interpretation' is an analysis of the major contributions, sensitivity analysis and uncertainty analysis. This stage leads to the conclusion whether the ambitions from the goal and scope can be met. More importantly: what can be learned from the LCA? All conclusions are drafted during this phase. Sometimes an independent critical review is necessary, especially when comparisons are made that are used in the public domain. Figure 5 Basket-of-products in LCA The weighting element has always been controversial as it includes social, political, and ethical values (Finnveden et al, 2009). Still the weighting procedure is widely used and it is therefore important to be aware of the method used. Methods for weighing are described more in detail by Finnveden et al To get a consumption view one can combine knowledge of consumptions patterns with LCA which would give information on the total environmental impact caused by consumption, illustrated in Figure 5. The basket-of-products reflects the environmental impact and the resources used that are associated with the final consumption of an average citizen in a country or region over the entire life cycle of the included products (goods and services). However, the choice of products has been a major obstacle, as it tends to be a rather subjective choice. The EU has initiated some studies that take this basket-of-products into account, for example a proposal on indicators

22 When performing LCAs the extensive data need is met by the use of LCA databases. There is no Swedish LCA database in use. The most used database is Ecoinvent 16, which is developed in Switzerland but include data from many other countries as well. A new version of Ecoinvent is planned to be in use in the end of The Joint Research Center (JRC) 17 is building an LCA database. 4.2 Environmentally extended input-output analyses The foundations of the environmentally extended input-output-model are the inputoutput tables in the national accounts and the emissions by industry sector. In economics an input-output model uses a matrix representation of a nation's (or a region's) economy to predict the effect of changes in one industry on others and by consumers, government, and foreign suppliers on the economy. Input-output tables (usually referred to as IOTs) are commonly maintained by countries to keep track on the flow of products through the economy. The input-output tables give a total table of how much (in monetary units), of which product categories, are supplied to the market within a nation. Figure 6 shows a screenshot of a symmetric input-output table for Sweden, where there are products in the rows and sectors in the columns. In the input-output table the consumption is allocated on input goods, private and public consumption, investments and exports. For example, the first cell in the input-output table in Figure 6 tells how much input of agricultural products the agricultural sector has used during a given year

23 Figure 6 Symmetric input-output table from aspx The consumption within an open economy is dependent on imported goods. The input-output table reflects the situation within the nation and emissions of imported goods caused in other countries must be handled separately, this can be done in several ways using the domestic technology assumption, taking into account bilateral trade or making a multi regional model. These methods are further described in section It is also important to note that the emissions that are calculated from the inputoutput model are the indirect emissions, i.e. emissions caused during the production processes. For some emissions, for example CO 2, the direct emissions, i.e. emissions caused during usage (fuel for transportation or heating purchased by households) are large contributors to the total emissions, and hence must be added separately to get a total view. Indirect emissions are also sometimes referred to as upstream emissions whereas direct emissions are referred to as downstream emissions. Investments in national accounts are for example the consumption of durable goods, such as for example machinery, housing, and livestock. A large part of the investments in a given year depend on the depreciations done in production capital during that year. This means that the investments should probably be seen as inputs instead of a part of final demand. Of this follows, that the environmental pressures must be reallocated to the part of final demand that used the production capital. Practically this means reallocating a part of the environmental pressures from gross investments to private and public consumption and to exports, as well as to the production of the net investment goods. 22

24 The environmental accounts at Statistics Sweden reports data to Eurostat on twelve emissions by industries (this is part Air Emissions Account 18 ), for example CO 2, N 2 O, CH 4, NOx and SOx. The calculations are based on the usage of fossil fuels per industry and applying general or plant specific emission coefficients. These emissions per industry are then used in the input-output based calculations of emissions caused by consumption. Both the emissions by industry and the consumption simulation can be found at Statistics Sweden s web tool 19. Environmentally extended input-output analysis The basic input output relationship developed by Leontief (Leontief and Ford, 1970) gives the relationship between the total output x and final demand y, by the following formula: x = (I - A) -1 y (1) (I-A) -1 is the inverse matrix or Leontief matrix. By applying an intensity matrix F (kg/sek) on equation 1 an extensive quantity of the emissions E (kg) can be calculated. This is shown in equation 2. E = F (I-A) -1 y (2) All types of emissions, or any type of variable, can be linked to consumption using the input-output model, given that the emission data is available per industry sector so that the intensity (i.e. emissions related to the monetary output of the industry) can be calculated (equation given in box above). In a Dutch study (Nijdam and Wilting, 2009), as an example of a less expected variable linked to consumption, noise from road traffic was linked to different consumer goods, showing that it was mostly caused by the consumption categories leisure, labour and food and it was mostly linked to direct environmental pressure (i.e. the environmental pressures occurs when the good is consumed while the indirect load is linked to the process before consumption). Examples of policy questions that can be answered by environmentally extended input-output analysis have been provided by Moll and Watson (2009) and those related to consumption referred to in the final report of the Era-Net SKEP Project EIPOT (EIPOT, 2009). The most relevant questions for this study follows, with examples:

25 Which consumed product groups are most responsible for indirect pressures activated by consumption? These questions have been answered in several studies referred to in this report, for example in the UNEP report (UNEP, 2010) and the ETC/SCP working paper (EEA, 2009). They both point out the categories food and drink, housing and mobility as activities causing large shares of environmental impact. How are indirect environmental pressures distributed across the categories of final use (private household consumption, government consumption, investments, and exports)? Answering this question makes for example the environmental impact from public consumption visible in relation to private consumption. A country can have an export of goods that are produced with a large environmental pressure, but still have a consumption that is not so intensive. Different countries can also be expected to have varying government consumptions and investment patterns. What is the ratio of indirect environmental pressures caused by national consumption which are emitted domestically compared to those taking place in the rest of the world? This is of interest when assessing to which degree environmental impacts caused by Swedish consumption contribute to environmental pressure in other countries. The difference can be an effect of the size of the trade, but also on the environmental legislation in the respective countries, and is also depending on the energy systems in respective countries. How do environmental pressures activated by national consumption compare with environmental pressures activated by national production? Answering this question could increase the awareness of how much of the total emissions take place in other countries in which Sweden s environmental policies are not effective. 4.3 Life cycle assessment in combination with input-output analyses As mentioned in the introduction to this chapter, there are hybrid methods combining LCA and the input-output based method. In this hybrid LCA, or input-output LCA (IO LCA) (Suh, 2009), are the input-output tables connected to LCA studies. There are several developments of the hybrid techniques, such as the tiered-hybrid approach, the IO-based hybrid approach and the integrated hybrid approach (Suh and Huppes, 2005). A benefit when using the hybrid technique is that the need to identify and exclude insignificant processes is reduced or eliminated as the input-output tables account for all upstream processes. 24

26 Another advantage with the hybrid technique is the elimination or decrease of the truncation error. Life cycle analysis and such process-analysis techniques are connected with a truncation error 20. It has been shown that this error can be of considerable size, in some cases as much as 50 %. A study (Lenzen, 2001) shows that this error can be avoided when incorporating the LCA into an input-output model to a hybrid method. 20 Truncation error arises from taking finite number of steps in computation of numerical algorithms 25

27 5 Areas of assessment As a starting point, in the assignment it was given to take the following environmental areas into account in this synthesis: Greenhouse gas emissions Other emissions to air Chemicals Water use Land use Biodiversity It was also given in the assignment to take the Swedish environmental objectives into account when relevant. The Swedish environmental objectives are listed in Figure 7: 1. Reduced Climate Impact 2. Clean Air 3. Natural Acidification Only 4. A Non-Toxic Environment 5. A Protective Ozone Layer 6. A Safe Radiation Environment 7. Zero Eutrophication 8. Flourishing Lakes and Streams 9. Good-Quality Groundwater 10. A Balanced Marine Environment, Flourishing Coastal Areas and Archipelagos 11. Thriving Wetlands 12. Sustainable Forests 13. A Varied Agricultural Landscape 14. A Magnificent Mountain Landscape 15. A Good Built Environment 16. A Rich Diversity of Plant and Animal Life Figure 7 The Swedish environmental objectives from For the completeness of the synthesis report it was given in the assignment to complete with other areas if found relevant. A recent report Assessing the Environmental Impacts of Consumption and Production 21 published by United Nations Environment Program (UNEP, 2010) gives a summary of environmental impacts of consumption and production. The report prioritizes which impact that globally are the most important and which consumption categories that contributes to them the most. As the UNEP report gives a summary of recent research in the field, it has been used in this synthesis as a reference when choosing areas of assessment and impacts

28 According to the UNEP report the prioritization of environmental impacts are those due to: Impacts caused by emissions - Climate change (caused by greenhouse gas emissions) - Eutrophication (overfertilization caused by pollution with nitrogen and phosphorus) - Human and ecotoxic effects caused by human and regional air pollution, indoor air pollution and other toxic emissions Impact related to resource use - Depletion of abiotic resources (fossil energy carriers and metals) - Depletion of biotic resources (most notably fish and wood) - Habitat change and resource competition due to water and land use. The UNEP report also ranked which activities that have the most significant impact on environment globally. For industrialized countries these activities are: Housing Mobility Food Manufactured products. The four areas are in alignment with the Sustainable Consumption and Production (SCP) 22, a program within the European Environmental Agency (EEA) which embraces both a production view and a consumption view on emissions. Within the SCP there is a work ongoing to develop indicators to the reporting and follow-up of the shift towards sustainable production within nations, as well as sustainable consumption patterns for the nations. There is a suggested framework for the indicators within three main elements: Headlines (this element contains overall information on sustainability), Status and Trends in Consumption and Production and The Framework for Change. The element Status and Trends in Consumption and Production contains indicators on resources, production, products, consumption, waste and key consumption cluster (food and drink, housing and buildings and mobility). This synthesis aims to cover the five areas that were assessed in the consumption part of the 2010 assessment of the Swedish environmental objectives report, de Facto 2010 complemented with biodiversity, as well as the Swedish environmental objectives, when relevant. An overview of the areas that are wished to be covered is given in Figure 8 with their most important interconnections. 22 More information on ETC/SCP on and the indicator framework on 27

29 Swedish Environmental Objectives (Miljömålsrådet, 2010) 1. Reduced Climate Impact 2. Clean Air 3. Natural Acidification Only 4. A Non-Toxic Environment 5. A Protective Ozone Layer 6. A Safe Radiation Environment 7. Zero Eutrophication 8. Flourishing Lakes and Streams 9. Good-Quality Groundwater 10. A Balanced Marine Environment, Flourishing Coastal Areas and Archipelagos 11. Thriving Wetlands 12. Sustainable Forests 13. A Varied Agricultural Landscape 14. A Magnificent Mountain Landscape 15. A Good Built Environment 16. A Rich Diversity of Plant and Animal Life de Facto areas (Miljömålsrådet, 2010) Climate (greenhouse gas emissions) Air emissions (sulfur and nitrogen oxides) Chemicals Water use Land use Biodiversity Environmental impacts: (UNEP, 2010) Impacts caused by emissions Climate change (caused by greenhouse gas emissions) Eutrophication (overfertilization caused by pollution with nitrogen and phosphorus) Human and ecotoxic effects caused by human and regional air pollution, indoor air pollution and other toxic emissions Impacts related to resource use Depletion of abiotic resources (fossil energy carriers and metals) Depletion of biotic resources (most notably fish and wood) Habitat change and resource competition due to water and land use. Figure 8 Areas of assessment with their most important interconnections. As can be seen in Figure 8 the areas are well interconnected, indicating that the proposed areas in the assignment are sufficient to assess the most important environmental impacts caused by consumption, according to UNEP (2010) and the Swedish environmental objectives. There are however two environmental objectives that do not have any clear interconnections to the proposed areas for the synthesis: 6. A Safe Radiation Environment and 15. A Good Built Environment. Questions regarding radiation are more related to health aspects than the environment. The target for the built environment includes targets about waste, noise and energy use in buildings and others. While noise is more of a health issue and energy use is related to the objective reduced climate impact, waste needs to be elaborated. The creation of waste is linked to consumption levels and the individuals handling of waste have also an effect on how much that is recycled. Waste connected to consumption is a quite unexplored area. A study performed in Finland 1999 (UNEP, 2009) showed that private consumption caused 17 % of the waste production in Finland, compared to 64 % of exported products. The waste caused by private consumption arises both from the extraction of raw material (in the production country), in the production processes (in the production country) and during use and scrapping of the product (in the country of consumption). As waste is also a product of trade, the trade of waste is also of relevance. There is also an illegal trade flow of waste globally. In Sweden and in the EU, waste is classified as either hazardous or non-hazardous waste. Within the environmental accounts data is published regularly on waste by industry. In theory, the data on waste can be linked to input-output analyses. The data on waste in other countries can be of different quality and not always comparable as the classification of waste might differ from one country to another. Within the Pollutant Release and Transfer Register (PRTR) 28

30 there is some data of the transportation of waste. However, in the PRTR the limits of the amount of waste that must be reported are set high, so there is a large gap of data foreseen. The different environmental problems related to waste is important. The different environmental aspects are linked to transportation, land use and emissions to air, soil and water. The handling of waste is linked to the environmental policy in each country. Some of the waste can be seen as a resource. In Sweden waste is used as an energy resource in district heating plants of which a part can be seen as renewable. The arise of hazardous waste in the production processes in the countries from which Sweden imports goods might be more interesting to study. The amount of waste is not a sufficiently specific measure to be linked to consumption. The area of waste and its impact on environment and health in different countries need more exploration before it can be linked to environmental pressures caused by consumption in Sweden. 29

31 6 Results and discussion This chapter contains short introductions to each assessment area, studies performed, a section about time series and data availability, a discussion about future development and a summary, for each of the areas. greenhouse gas emissions, other emissions to air, chemicals, land use, water use and, biodiversity. 6.1 Greenhouse gas emissions Greenhouse gases (GHG) is the common name for gases that contribute to the greenhouse effect. The gases that are most important are carbon dioxide, methane, nitrous oxide, ozone and, water vapour. Carbon dioxide mainly arises from the use of fossil fuels, ozone from transportation and water vapour from various human activities such as combustion and transportation. Methane mainly arises from agricultural activities and nitrous oxides from agricultural soil management and mobile sources (Suh, 2009). Emissions of greenhouse gases are commonly presented in CO 2 equivalents. Within the environmental accounts there is the Air Emission Accounts 23. The term air emission is used to denote the physical flow of gaseous or particulate materials from the economic system (production or consumption processes) to the atmosphere. Air emissions comprise emissions of greenhouse gases as well as emission of air pollutants such as e.g. SO2, NOx, PM10 etc.12 emissions to air by industry are reported to Eurostat. Of these emissions that are reported are seven climate related, these are shown in Table 1. The emissions to air are calculated by different actors using different models. The allocation to industries is performed by the environmental accounts at Statistics Sweden. 23 Eurostat manual for Air Emission Accounts RA /EN/KS-RA EN.PDF 30

32 Table 1 Emissions of greenhouse gases reported to Eurostat by the environmental Greenhouse CO2 25 Emissions of carbon dioxide (without emissions accounts 24 gases from biomass) by economic activity and households Biomass CO2 26] Emissions of carbon dioxide from biomass by economic activity and households N2O CH4 HFC PFC SF6 Emissions of nitrous oxide by economic activity and households Emissions of methane by economic activity and households Emissions of hydrofluorocarbons (HFCs, given in CO2-equivalents) by economic activity and households Emissions of perfluorocarbons (PFCs, given in CO2-equivalents) by economic activity and households Emissions of sulphur hexafluorids (SF6, given in CO2-equivalents) by economic activity and households In the IPCC report (Naturvårdsverket, 2008) it is stated that one fifth of the global climate gas emissions are caused by a change in land use and forestry. Problematic activities such as deforestation in tropic areas to cultivate for food or bio fuels, or live stock are linked to such emissions. A study has been performed on imports of agricultural and forestry products to Sweden and their effect on carbon leakage globally (Möllersten, 2008). The calculations are performed for palm oil, ethanol, soya beans and beef meat from countries with rain forests. The carbon leakage arises as land use is changed in the country where the products are cultivated, mainly through deforestation and drained peat land Performed studies The assessment of pollution embodied in trade is most usually done by input output analysis 27. Various approaches can be found in literature to make such estimates ions 25 CO2 emissions from biomass are reported separately. 26 CO2 emissions from biomass are reported separately. 27 An alternative is to make LCAs for imported goods, but this is usually time consuming and can often not been done for all individual products in a product group, leading to errors by scaling up, etc. 31

33 The simplest approach is to apply a shortcut, assuming that imported goods and services are made with the same technology as those produced domestically (e.g. Huppes et al., 2006; Palm et al., 2006; Hendrickson, 2006). This is called the domestic technology assumption. This assumption can lead to serious errors (Peters and Hertwich, 2006a and 2006b; Weber and Matthews, 2007; Ghertner and Fripp, 2007; compare also Ahmad, 2003; Ahmad and Wyckhoff, 2004). To take into account trade between countries, researchers started to apply the bilateral trade approach. They identified the main trading partners of a country, made available environmentally extended input-output tables for these countries or country groups, and calculated the embedded pollution and resource use in bilateral trade (see e.g. Weidema et al. (2005, Denmark); Peters and Hertwich (2006b; Norway), Nijdam and Wilting (2005; Netherlands), Weber and Matthews (2007, US) and Norman et al (2007, Canadian-US trade)). A common feature of all these studies is that only bilateral trade was considered and trade with other countries was ignored. The bilateral trade application was for example used in a Swedish study (Naturvårdsverket, 2008) where an assessment of the impact of greenhouse gases from consumption in Sweden was performed, including carbon dioxide, methane and nitrous oxide. The study used data from environmental accounts in different countries and input-output analysis. This was combined with information on trade partners and emission intensities. The overall view was combined with life cycle based assessments of some products used as examples to give a more detailed perspective in some cases. The study showed that transportation, housing and food are the most significant activities giving arise to green house gas emissions inside and outside Sweden. An EEA study (EEA, 2009) was performed where environmental extended inputoutput analysis was discussed to be used for answering the key questions within SCP (sustainable consumption and production). The method was considered being useful for assessing the two SCP perspectives; production and consumption. In the study four key aggregated environmental pressures were examined: greenhouse gas emissions, acidifying emissions, emissions of tropospheric ozone precursors and domestic material input. The economic sectors that dominate direct emissions are agriculture, electricity industry, transport services and some basic manufacturing industry. When analyzing emissions caused by consumption a few groups were identified that contributed significantly to direct and indirect emissions: construction works, foods products (inclusive beverages and alcohol), electricity (together with gas, steam and hot water). Four limitations were found, that limit the method s full potential: time lag in data, level of aggregation of economic sectors, production processes in other countries for imported goods, lack of environmental data. 32

34 In the Resources and Energy Analysis Programme 28 (REAP), which has been developed by Stockholm Environment Institute 29 (SEI), an environmentally extended input-output model makes up the backbone of the model. REAP is a two-region model which distinguishes between products that are produced within the nation and imported products from the rest of the world. The REAP model generates ecological, carbon and greenhouse gas footprints results and baseline data for air emissions, heavy metals and material flows. The REAP is a policy model that contains a scenario module where several variables can be changed, population, consumption and technology. REAP is being developed for Sweden and for EU-27 through the One Planet Economy Network 30. A more comprehensive method would be a truly multi-regional approach, in which economies of the rest of the world are presented together with the country of interest in a multi-regional input-output table with environmental extensions (MR EE IO). One example is Wiedmann et al. (2008), who used Nijdam and Wilting s (2005) data for the EU and other OECD and non-oecd countries and developed them as a true multiregional environmentally extended input-output table. Peters (2007) and Friot et al. (2007) used the GTAP 31 multi-regional input-output database with various individual emissions as a basis for their modeling efforts. Peters (2007) was the first to perform a multi regional study using the complete GTAP database, work which was extended into the calculation of the Carbon footprint of nations by Hertwich and Peters (2009). In this study the life cycle perspective and trade between countries were taken into account. They also argue for including, not only the private consumption part of final demand, but also public consumption and investments. The study presents the carbon footprint of nations in t CO 2 per person. The analysis of the results shows that nutrition, shelter (i.e. operation and maintenance of residence), private mobility, services (public administration, health, and defense) and manufactured products are the product groups with largest impact on the carbon footprint. The study also shows that that food is dominant in poor countries while mobility becomes more important for rich countries. The Global Resource Accounting Model (GRAM) of Giljum et al. (2008b) uses OECD input-output and trade data extended by material extraction in combination with the GINFORS 32 (Global Interindustry Forecasting System) model to calculate indirect material flows of traded products. The use of a complete multiregional environmentally extended input-output model offers the most detailed insights for Information about the GINFORS model on 33

35 determining the specific causes of impacts and allocating impacts to producers or consumers (Lenzen et al., 2007; Peters and Hertwich, 2006a; Peters, 2007) The domestic technology assumption approach usually leads to significant errors in estimates of pollution embodied in trade, particularly for small countries with open economies and whose economic structure differs highly from the countries from which it imports. The multiregional environmentally extended input-output approach is the most comprehensive one, but requires a significant number of data, that are largely lacking today. If a single country would set such a data system up, it would be forced to put more effort in data gathering for other countries, than for its own country. We see hence in practice that the only practical multiregional environmentally extended input-output efforts have used existing databases such as the GTAP 33 multiregional input-output database, and the OECD input-output tables. Still, these sources have important limitations in terms of environmental extensions, sector detail, transparency and quality, and usually need additional work before they are ready for analytical use (e.g. Tukker et al., 2009) 34. Various efforts are under way, however, to improve this situation, mainly driven by the research community. Examples include: Countries such as Sweden, Finland, Denmark, Netherlands, the US, Canada and others have developed environmentally extended input-output tables for their own countries, and in some cases also made environmentally extended input-output tables for countries from which their main imports come, to assess pollution and resources embodied in trade (i.e. the bilateral trade approach mentioned above) In Asia, the Japanese Institute for Developing Economies (IDE JETRO) has developed an Asian multiregional input-output analysis that includes also the US, and is now working on a multiregional input-output model that also includes the BRICS 35 countries. Since this effort does not include extensions, it is less useful for the purpose of Swedish EPA in this project. The EU FP7 36 funds the World Input Output Database (WIOD) project. This project is mainly interested in economic productivity analyses, but as a part of it, builds time series between 1995 and 2006 of a multiregional environmentally extended input-output with about 40 countries, with a sector detail of 30, and a limited number of extensions The OECD tables are for individual countries and need a specific effort to be trade linked. Both GTAP as OECD in principle contain no or just limited extensions, which hence need to be gathered by practitioners. In practice such databases mainly have been used to make analyses with a limited number of pollutants, such as CO2, CH4 and SOx. 35 Brazil, Russia, India and China 36 The seventh framework programme for research and technology development 34

36 The EU FP6 37 funds the development of the first global, detailed harmonized environmentally extended input-output tables (the EXIOPOL project). It covers the 43 major economies in the world plus the Rest of World, with 130 sectors per country and around 150 extensions, but it focuses on a single year (2000) Under FP7 a follow up proposal on EXIOPOL has been selected for funding, which will expand the extensions and build time series until a more recent date. The project will run until With Australian funding, the University of Sydney is developing a tool that is able create a multiregional input-output table from very different source data with relatively limited effort. To some extent this resembles the WIOD and EXIOPOL projects, but the emphasis is on development on automated harmonization and diagnostic tools. Also this project plans to deliver a global multiregional input-output with environmental extensions. Eurostat has launched various projects that in 2-4 years should lead to a multiregional input-output with environmental extensions for the EU27 discerning about 60 sectors and products and some emissions and resources, based on the available national and environmental accounting data. The studies performed take into account emissions of fossil fuels and process emissions, sources and sinks of land use are hence not included. Emissions due to land use have been showed to be a large contributor to overall greenhouse gas emissions, and future enhancements of these methods are hence possible to get a more complete view Time series and data availability As the use of fossil fuels directly causes emissions of CO 2, emissions from fuel based activities (i.e. production of goods) can be calculated as data is available from environmental and economic statistics. Time series are possible to obtain as data is available back to 1993, the current System of national Accounts (SNA) was implemented in 1993, and data is updated on a yearly basis. Statistics Sweden publishes data ( on a regular basis on emissions of CO 2 equivalents connected to consumption. The underlying method is input-output analysis combined with emission data per industry. Emission caused in other countries by imported goods, are calculated with two methods: either as if the production occurs in Sweden (the domestic technology assumption) or with country intensities (bilateral trade). 37 The sixth framework programme for research and technology development 35

37 6.1.3 Suggested way to go forward Swedish EPA could opt for the bilateral trade approach, and link its own environmental extended input-output table to those of its most important trade partners. It is likely that for most of the direct trade partners (Norway, Denmark, some other EU countries, the US) environmentally extended input-output tables are available. If a more comprehensive approach based on multiregional environmentally extended input-output is preferred, the following options exist: Existing databases such as the expanded GTAP databases developed for e.g. the Carbon footprint of nations analysis by Hertwich and Peters (2009) could be used to make estimates of greenhouse gases embodied in imports right away. In around a year from now (2011), both the EXIOPOL as WIOD projects will deliver their multiregional environmentally extended input-output data. As things stand now these data will be public (or available for a minor fee). For environmental analyses, the EXIOPOL data will probably be the best to use in the type of analyses desired by Swedish EPA given sector detail and number of extensions. In a somewhat longer time frame, the EUROSTAT projects are likely to deliver a full multiregional environmentally extended input-output table for the EU27. The advantage here is that as far as trade with the EU27 concerns Sweden can use these data for assessing pollution embodied in trade, using data with an official status 38. Another enhancement of the analyses would be to split industrial sectors and product categories into a finer level in relevant sectors, for example NACE 01, agriculture. If a finer breakdown becomes available it would for example be possible to analyze how emissions would change in a shift to a less meat intensive diet. Something that can be further analyzed is how the result is affected by the fact that production of the same product is connected to different costs and different emissions, depending on producer country. A possible way of forthcoming is to analyze the 50 to 150 goods that are included in the input-output analysis and trade statistics that are weighted with for example WRI (World Resources Institute), complemented with more detailed statistics from the Air Emission Accounts. A quantification of the size of the error that is a result from allocation problems in the trade statistics is wished for. As an example, the EU imports large amounts from China, of which some is processed and traded within EU. This trade is registered as if the origin is EU, and hence environmental impact connected to this trade 38 Note that Swedish ESA95 SUT and related national accounting matrices with environmental accounts will be part of this data set. 36

38 is not registered on the correct country 39. This can be analyzed by choosing a homogeneous product, such as for example t-shirts. For t-shirts an alternative to ton is number in the trade statistics. By analyzing the imports of this product it can be shown that for Sweden 30 % of the import of t-shirts can origin from another country than it has been registered on (personal comment, Statistics Sweden, 2010). This share differs due to which country is studied due to different shares of imported goods. Sweden has a relatively large public consumption. The way of setting up inputoutput tables in Sweden makes a large part of what in other countries would be seen as inputs to public production, instead is viewed as public consumption and thus part of final demand. This means that public procurement in Sweden probably has a greater global effect than in other countries. Swedish public administration is by legislation (SFS 2009: ) forced to implement environmental management systems where environmental aspects, of which procurement is a large contributor, must be taken into consideration. The Swedish Environmental Management Council 41 (SEMCo) support companies and public organizations in this work. The environmental impacts caused by public consumption in Sweden can be studied more in detail Summary for greenhouse gas emissions There are several examples of calculations performed of emissions of greenhouse gases caused by Swedish consumption using input-output analyses with environmental extensions. The results vary depending on the data availability, and the methods chosen to generate the missing data. Emissions intensities for different trade partners have been used, but are not yet reported in sufficient detail by the countries, and thus need to be estimated. Data can be estimated for most countries as the emissions are largely caused by the use of fossil fuels, and also based on the reporting to the IPPC. Input-output data is available for many countries, but there are not always time series available. If input-output data is lacking, data from another country can be used and time series can be modeled. When using Swedish input-output data, time series of emissions caused by consumption can be calculated. By studying time series, trends over time can be evaluated. As the emissions are calculated using the same method every year, the impact of uncertainties is decreased. The result can be divided on different contributors to 39 This is referred to as Rotterdamfällan in Sweden

39 the greenhouse gas emissions, as CO 2 -equivalents or GWP (global warmth potential) which is a measure described by the IPCC. Greenhouse gas emissions and sinks due to land use are not included. Changes in the amount of emissions of greenhouse gases, over time and between countries, are related to: Energy systems Land use Types of goods and the amount of consumer goods For greenhouse gas emissions there exists an input-output based model which can be used to calculate the emissions caused in other countries due to Swedish consumption. There are results published by Statistics Sweden. To go further with these analyses, one could start looking deeper into the factors that the consumption model can provide. That means comparing results by product groups and also to focus on the improvement of environmental data that are needed to calculate country intensities, i.e. for emissions caused in the countries where the production process takes place. 6.2 Other emissions to air Within the environmental accounts there is the Air Emission Accounts 42. The term air emission is used denote the physical flow of gaseous or particulate materials from the economic system (production or consumption processes) to the atmosphere. Air emissions comprise emissions of greenhouse gases as well as emission of air pollutants such as e.g. SO2, NOx, PM10 etc. The emissions that are related to acidification and local air quality are shown in Table 2. These emissions are reported to Eurostat by Statistics Sweden. The emissions are allocated to sectors and industries by the environmental accounts at Statistics Sweden. 42 Eurostat manual for Air Emission Accounts 38

40 Table 2 Air emissions reported to Eurostat by the environmental accounts 43 Acidification NOx Emissions of nitrogen oxides by economic activity and households Local air quality Sox NH3 NMVOC CO PM10 Emissions of sulphur oxides by economic activity and households Emissions of ammonia by economic activity and households Emissions of non-methane volatile organic compounds by economic activity and households Emissions of carbon monoxide by economic activity and households PM10-emissions by economic activity and households Performed studies As mentioned before, all kind of emissions by industry can be linked to consumption with input-output analysis. In a Dutch study (Nijdam and Wilting, 2009) the direct and indirect environmental load of consumption categories was calculated using input-output tables linked with environmental data, covering worldwide impact. The environmental load per Euro turnover is linked to consumer expenditures by the expenditure survey of Statistics Netherland. This was performed for acidifying gases (SO 2 equivalents) among other emissions. At statistics Sweden the emissions of SOx and NOx are related to consumption in the web-tool 44, using input-output analysis Time series and data availability Statistics Sweden publishes data ( on a regular basis on emissions to air linked to consumption. Statistics Sweden uses input-output tables combined with data on emissions by industry. Emission caused in other countries by imported goods, are calculated as if the production occurs in Sweden. A future development would be to apply emission intensities for emissions in other countries, which has been made for CO 2 emissions ions

41 6.2.3 Suggested way to go forward Emissions to air connected to consumption can be treated the same way as the greenhouse gas emissions, i.e. calculated using the input-output model. The results presented by Statistics Sweden are calculated using the domestic technology assumption, and can hence be improved. The analyses can include all the emissions that are reported to Eurostat. Emission intensities for different trade partners must be developed to get a more accurate result. Emission data is not always available for countries outside the EU Summary for emissions to air When using the input-output model to link consumption to emissions to air time series can be produced. By comparing time series trends over time can be studied, as the emissions are calculated using the same method every year, the impact of uncertainties is decreased. For emissions to air the environmental legislation has a greater impact than in the case of emissions of greenhouse gases. These emissions are not as related to the energy system in the country as the CO 2 emissions are, they are more linked to the usage of environmental technology. Data is available in Sweden and the EU, but for countries outside the EU there can be a lack of data foreseen. Changes over time and or between countries can be evaluated. Different amounts of emissions of emissions to air are related to: Environmental legislation Available environmental technology To go further with these analyses, one could start looking deeper into the factors that the consumption model can provide. That means comparing results by product groups and also to focus on the improvement of environmental data that are needed to calculate country intensities, i.e. for emissions caused in the countries where the production process takes place. 6.3 Chemicals Emissions of chemicals take place during production phase and during usage of products containing chemicals. A large part of the chemicals, and products containing chemicals, consumed in Sweden are imported. The production can take place in countries that lack chemical control and that have weak environmental legislation. In such countries the emissions of chemicals can be larger than in countries with more developed rules. The impact on the local environment can be substantial. When assessing the emissions of chemicals caused by consumption in Sweden one should theoretically include both the indirect emissions, i.e. emissions during production and the direct emissions, i.e. during usage. However, in this study we focus on emissions during production as the focus lays on impacts in other countries and due to that the emissions during usage is still, to a large extent, an unveiled area. 40

42 According to a recent report (UNEP, 2010) most studies concerning environmental pressure from consumption and production focus on fossil fuels, energy and greenhouse gas emissions. Few studies have a focus on hazardous chemicals. There are however a few studies that point out which production and consumption activities, products and materials that are linked to releases of toxic substances. A conclusion in the UNEP report is that agriculture and food consumption are identified as one of the most important drivers of toxic emissions as well as other environmental pressures such as habitat change, climate change and water use. The consumption activities that have the largest impact on human toxicity and ecotoxicity are food and beverages, housing and transport (Huppes et al, 2006). A similar result is shown in (Weidema et al, 2005) which states that leisure, housing and food contribute most to human toxicity in the EU. Also the EIPRO project identified products with the largest impact throughout the life-cycle of a product (EIPRO, 2006); the project includes studies with a bottom-up perspective (LCA) and top-down perspective (input-output). In the case of materials, there is one study that focuses on metals and eco-toxicity. That study concludes that the metals that have most impact to terrestrial eco-toxicity is mercury, gold, rhodium, platinum and palladium, in that order (Staal, 2009). These studies can be used as a start and background data on where to focus further research in the area. The environmental accounts at Statistics Sweden publish statistics about the use of chemicals in different industries in Sweden. The chemicals are classified in groups according to the international labeling of hazardous properties 45, the result is called chemical indicators. The data source is the product register at the Swedish Chemical Agency. Important to note is that this data concerns the use of different chemicals, not the emissions. There are currently not enough quantitative data and methods to show the overall chemical pressure in other countries caused by Swedish consumption. Information on the use of hazardous chemicals and the emissions of hazardous chemicals by industries in the countries that export goods to Sweden, are of interest to obtain Performed studies There about chemical substances in use. Keeping track of all of them is not feasible and also not wished for, as not all of them are harmful to health or environment. Instead of studying chemicals one by one a method to weigh them, e.g. according to their inherent properties, has been developed. Still, specific substances may be of interest to study, for example in specific geographic regions, or in specific product categories aspx 41

43 WEIGHTING OF CHEMICALS To our knowledge, two main methods have been proposed to weigh together a number of chemicals. One was mentioned above, and is based on the inherent properties on chemical products and the international labeling of them. (Palm and Carlsson, 2004) This has been used in the Swedish environmental accounts and is can be used in input-output calculations. Data is available at Statistics Sweden (chemical indictors). Internationally there is also method development ongoing in this field, The REACH baseline study (Eurostat, 2009). However, the data concerns the usage of chemicals, and a linkage to emissions in the producer country must be developed. The other main method is concerned with weighting of emissions of certain chemicals is to apply toxicity potentials for each emission in kg, per substance. Methods to do this is found in Guinée et al (2002) and UEStox TM 46. These methods have been used in LCA applications, some of which are input-output-based. Theoretically, the potential impact from all products from chemicals could be assesses using LCA and input-output analysis. The LCA method is usually used to compare two or more different products, to get guidance which product or process is best to use concerning different aspects or which product have the largest toxic impact. In a product LCA you estimate the emission for all products based on the materials that they are composed of, if you have sufficient data. Finnveden (2009) points out that there is a limited knowledge on the use and fate of chemicals and that this is a societal problem and not just a problem for LCA. Often all data is not available in a product LCA, and you have to fill in with input -output data which always includes the whole economy. In input-output analysis the method often is used to see how changed consumption affects for example emission. It can be used for the question If the consumption of a product group increases with Y percent, how much will the use or emission of chemical Z increase? To be able to do this estimate, use or emission data in the production country must be available. You can also make the assumption that the use or emission of a substance is the same per produced million SEK as it is domestically. There are emission data for production sites for some countries through the Pollutant Release and Transfer Register (see below). In the U.S. there is research going on which uses large databases of chemicals emissions by industry. Figure 9 shows a result received by using input-output analysis, the results comes originally from Guinée et al; 2002; Suh The figures are taken from UNEP (2010). In the result the chemicals have been weighted showing human toxicity and freshwater toxicity. The result shows that the production activities that have the most impact on human toxicity in the U.S. are electric services

44 followed by paper and paperboard mills. The production activities that have the largest contribution to freshwater toxicity in the U.S. are cotton and feed grains. Another study using input output analysis which includes toxic impact is Weidema et al, 2005, they included human toxicity and ecotoxicity in their research which concerned Denmark. Contribution by direct emitters to human toxicity in the US. Contribution by direct emitters to freshwater ecotoxicity in the US. Figure 9 The left diagram shows contribution to human ecotoxicity and the right diagram contribution to freshwater ecotoxicity in the US. Figures from UNEP 2010, data originally from Guinée et al 2002; Suh Figure 9 has a national focus, what production activities contributes most within the US, it does not answer to the question how much emission is produced in other countries due to the consumption in US. The authors of the UNEP report points out that only regular emissions are accounted for in the analysis; accidental emissions, illegal dumping and spills are not included. The database that they use to produce Figure 9 is the CEDA database (Suh, 2004). The CEDA database contains data that is reported in TRI (Toxic Release Inventory) in the U.S. which is similar to the reporting of Pollutant Release and Transfer Register (PRTR) in Europe. The Pollutant Release and Transfer Register (PRTR) is a national or regional environmental database or inventory of potentially hazardous substances and pollutants released to air, water and soil 47. It covers 65 economic activities and their amount of pollutant releases to air water and land as well as off-site transfers of waste and of pollutants in waste water. It includes about 90 key pollutants including heavy metals, pesticides, greenhouse gases and dioxins for the year 2007 and onwards. In Europe a register (E-PRTR) provides easily accessible key environmental data from industrial facilities in 27 European Union Member States and in Iceland, Liechtenstein and Norway 48. EU member countries 49 register emissions to the