UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Greenhouse gases embodied in international trade : an input-output analysis for Canada : 2002 Tran, Julie 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_tran_julie.pdf [ 1.21MB ]
Metadata
JSON: 24-1.0072438.json
JSON-LD: 24-1.0072438-ld.json
RDF/XML (Pretty): 24-1.0072438-rdf.xml
RDF/JSON: 24-1.0072438-rdf.json
Turtle: 24-1.0072438-turtle.txt
N-Triples: 24-1.0072438-rdf-ntriples.txt
Original Record: 24-1.0072438-source.json
Full Text
24-1.0072438-fulltext.txt
Citation
24-1.0072438.ris

Full Text

GREENHOUSE GASES EMBODIED IN INTERNATIONAL TRADE: AN INPUT-OUTPUT ANALYSIS FOR CANADA – 2002  by  Julie Tran  B.B.A., HEC Montreal, 1998 M.Sc., HEC Montreal, 2001  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF ARTS  in  THE FACULTY OF GRADUATE STUDIES  (Planning)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2011  © Julie Tran, 2011  Abstract Both climate change and actions to fight it are occurring against rapidly expanding international trade flows, which increasingly lead to the separation of production and consumption patterns. Through the trade of goods in a globally interdependent world, the consumption in each country is linked to greenhouse gases (GHG) emissions in other countries because GHGs are emitted throughout the supply chain involved in producing those goods; a phenomenon referred to as „embodied GHG‟.  In this research, taking a consumption-based approach and using an environmental input-output analysis, I explore the amount of GHGs embodied in Canada‟s imports and export for the year 2002 and determine that Canada has a negative balance of embodied emissions in trade (BEET). This implies that the GHGs emitted in connection with the production of exported goods surpass those emitted in connection with the production of imported goods. In light of Canada‟s large trade surplus in 2002, my finding support the hypothesis that there may be an inherent conflict between a national GHG reduction target for domestic emissions and the aims of improving trade balances or maintaining trade surpluses.  While my negative BEET result holds under different model specifications (i.e., single-country or multi-country model), I also show that it is highly dependent on the exchange rate used to convert the value of Canadian imports from the U.S. into the U.S. currency. Thus, my results demonstrate the weakness of using monetary flows of merchandise trade when trying to estimate physical quantities, in this case, the amount of GHGs embodied in the traded goods. Finally, I also discuss some of the key intersections between climate and trade policies, with a particular focus on climate policies that attempt to link the consumption of goods in a country with the amount of GHGs emitted during their production, whether in that country or elsewhere.  ii  Table of Contents  Abstract....................................................................................................................................... ii Table of Contents....................................................................................................................... iii List of Tables .............................................................................................................................. v List of Figures ............................................................................................................................ vi List of Acronyms ....................................................................................................................... vii 1. Introduction .......................................................................................................................... 1 1.1  Problem Statement ................................................................................................. 1  1.2  Purpose and Research Objectives .......................................................................... 2  1.3  Research Approach................................................................................................. 3  1.4  Scope ...................................................................................................................... 4  1.5  Thesis Structure ...................................................................................................... 4  2. International Framework on Climate Change ....................................................................... 5 2.1  United Nations Framework Convention on Climate Change .................................... 5  2.2  Kyoto Protocol ......................................................................................................... 6  2.3  Toward a Successor Agreement ............................................................................. 9  3. Literature Review ................................................................................................................11 3.1  Theories of Producer and Consumer Responsibility ...............................................11  3.2  Empirical Evidence .................................................................................................12  3.3  Review of Analytical Frameworks ...........................................................................14  3.3.1  Economic Input-output Analysis Framework ...........................................................14  3.3.2  Environmental Input-Output Analysis Framework ...................................................17  3.3.3  Economic Input-Output Life Cycle Assessment Framework ...................................18  3.3.4  Single-Region (or Country) Input-Output Models ....................................................21  3.3.5  Multi-Region (or Country) Input-Output Models ......................................................22  4. Research Approach and Methods .......................................................................................25 4.1  Model Specification ................................................................................................25  4.2  Data Sources .........................................................................................................26  4.2.1  Data Sources of the EIO-LCA Model for Canada ...................................................28  4.2.2  Data Sources of the EIO-LCA Model for the United States .....................................31  4.3  Data Collection Procedures for Import and Export Data .........................................32  iii  4.3.1  Single-Country Approach .......................................................................................33  4.3.2  Multi-Country Approach (Linked Single-Country IO Model) ....................................34  4.4  Assumptions and Limitations ..................................................................................36  4.4.1  Production Processes of Other Countries...............................................................36  4.4.2  Other Assumptions and Limitations ........................................................................37  5. Results ................................................................................................................................38 5.1  Canada‟s Negative Balance of Embodied Emission in Trade .................................39  5.2  Canada‟s Balance of Embodied Emissions in Trade by Industrial Sector ...............44  6. Discussion of Policy Implications .........................................................................................47 6.1  Policy Implications of a Negative BEET ..................................................................47  6.2  Intersections of Climate and Trade Policy ..............................................................48  6.2.1  Border Tax Adjustments .........................................................................................49  6.2.2  Low Carbon Fuel Standard.....................................................................................52  6.2.3  Climate Labelling ....................................................................................................53  6.3  Improving the Climate-Trade Intersections .............................................................54  Bibliography ..............................................................................................................................57 Appendix A: GHG Emissions and Kyoto Targets of Select Countries ........................................61 Appendix B: Canadian Sectors Included in the Analysis ...........................................................62 Appendix C: Example of BEET Calculations for Motor Vehicles Manufacturing.........................64 Appendix D: U.S. Sectors Included in the Analysis ...................................................................68 Appendix E: Canada‟s Sectoral Trade Balances and BEET ......................................................74  iv  List of Tables  Table 1: Quantified Emission Limit or Reduction Commitment under the Kyoto Protocol ........... 7 Table 2: Comparison of GHG Emissions Trends with Kyoto Targets .......................................... 8 Table 3 Example of an Economic Input-Output Table ...............................................................16 Table 4: Recent Studies based on Single-Region Input-Output Analysis...................................22 Table 5: Recent Studies based on Multi-Region Input-Output Analysis .....................................23 Table 6: GHG Embodied in Canadian Merchandise Trade, 2002 ..............................................38 Table 7: GHG Embodied in Canadian Merchandise Trade, by NAICS Sectors, 2002................45 Table 8: Canada's Trade Data, by NAICS Sectors, 2002 ..........................................................45 Table 9: Trade Balance and BEET for Select Industry Groups ..................................................46 Table 10: Agricultural, Resource and Manufacturing Sectors Included in the Analysis ..............62 Table 11: Motor Vehicle Manufacturing: Imports and Exports by Dollar Value, 2002 .................64 Table 12: Imports of Motor Vehicles from the U.S. ....................................................................66 Table 13: Sectors of the U.S. EIO-LCA Model Included in the Analysis ....................................68 Table 14: Canada‟s Balance of Trade and BEET by NAICS Industry Groups, 2002 ..................74  v  List of Figures  Figure 1: Steps in the EIO-LCA Process ...................................................................................20 Figure 2: Illustration of the EIO-LCA Model: Motor Vehicles ......................................................20 Figure 3: SRIO Model and Autonomous Economies .................................................................21 Figure 4: MRIO Models with Unidirectional and Multidirectional Trade ......................................23 Figure 5: Global Warming Pollutants .........................................................................................29 Figure 6: Emission Trends by Gas ............................................................................................31 Figure 7: U.S. Greenhouse Gas Emissions by Gas ...................................................................32 Figure 8: Canada‟s Top Trading Partners, Imports by Dollar Value 2002 ..................................40 Figure 9: Annual Average Exchange Rates of the Canadian dollar to the U.S. dollar ................41 Figure 10: PPP Exchange Rates of the Canadian Dollar to the U.S. Dollar ...............................43 Figure 11: GHG Emissions and Kyoto Targets of Select Countries ...........................................61 Figure 12: Emissions Embodied in Canada‟s Imports of Motor Vehicles ...................................64 Figure 13: Emissions Embodied in Canada‟s Exports of Motor Vehicles ...................................65 Figure 14: Emissions Embodied in Canada‟s Imports of U.S. Automobiles (336111) ................66 Figure 15: Emissions Embodied in Canada‟s Imports of U.S. Heavy-Duty Trucks (336120)......67  vi  List of Acronyms BTA CAD CH4 CO2 CO2e CMU DSB EEE EEI EIO-LCA EPA FCCC GATT GHG GWP HFCs IPCC IO kMT LCA LCFS MER MMT MRIO MT N2O NAICS PFCs PPP SCM SF6 SRIO TBT TDO UN UNFCCC USD WTO  Border Tax Adjustment Canadian Dollar Methane Carbon Dioxide Carbon Dioxide Equivalent Carnegie Mellon University Dispute Settlement Body Emissions Embodied in Exports Emissions Embodied in Imports Economic Input-Output Life Cycle Assessment Environmental Protection Agency Framework Convention on Climate Change General Agreement on Tariffs and Trade Greenhouse Gas(es) Global Warming Potential Hydrofluorocarbons Intergovernmental Panel on Climate Change Input-Output Thousand Metric Tons Life Cycle Assessment Low Carbon Fuel Standard Market Exchange Rate Million Metric Tons Multi-Region Input-Output Model Metric Tons Nitrous Oxide North American Industrial Classification Standard Perfluorocarbons Purchasing Power Parity Subsidies and Countervailing Measures Sulphur Hexafluoride Single-Region Input-Output Model Technical Barriers to Trade Trade Data Online United Nations United Nations Framework Convention on Climate Change United States Dollar World Trade Organization  vii  1. Introduction 1.1  Problem Statement  Human induced climate change is considered by many scientists to be the most serious threat facing the world today. Its impact on ecosystems, economies and local weather has already begun. Extreme weather events (e.g., flooding, drought, and category-5 hurricanes), imperilled ecosystems and threats to biodiversity, global meltdown of glaciers and icecaps, water scarcity, and heightened health and economic risks are all devastating consequences of rising global temperatures. However, both climate change and actions to fight it are occurring against rapidly expanding international trade flows, which increasingly lead to the separation of production and consumption patterns (Rothman, 1998; Muradian et al., 2002; Rees, 2006) and an increased potential to outsource environmentally intensive production processes outside national boundaries. All societies have a responsibility for abating climate change according to, among other criteria, their contribution to global greenhouse gas emissions. Through the trade of goods in a globally interdependent world, the consumption in each country is linked to greenhouse gas emissions in other countries because GHGs are emitted throughout the supply chain involved in producing those goods; this phenomenon is referred to as „embodied GHG‟. Although the „greenhouse gas responsibility‟ of the inhabitants of a country is determined by their consumption of both domestically produced and imported commodities, international negotiations on emission reductions focus solely on territorial emissions. GHGs embodied in international trade are often neglected.  A growing number of researchers argue that the GHG emissions of a country should be calculated as those resulting from the production of the goods consumed within that country, whether the goods are produced domestically or imported. 1 That is, in order to achieve equitable reduction targets, international trade has to be taken into account when assessing nations‟ responsibility for abating climate change. This issue may be particularly important for open economies like Canada, which is, has always been and will always be a trading nation. Taking into account the GHGs embodied in international traded commodities of a country can have a considerable influence on its national GHG balance. In the case of Canada, increased  1  Early studies include Wyckoff and Roop (1994) and Subak (1995). More recent studies, focused on single-country analysis, include Lenzen (1998), Machado (2001), Munskaard and Pedersen (2001), Mongelli et al. (2006), Mäenpää and Siikavirta (2007), while multi-country studies include Ahmad and Wyckoff (2003), Lenzen (2004), Weber and Matthews (2007) and Norman et al. (2007).  1  exports of goods produced in Canada result in an increase of energy consumption and GHG emissions within Canadian borders (in the absence of stringent domestic legislation and regulations to reduce GHGs), while the opposite holds for imports into Canada.  In turn, this situation can have serious implications for the design and success of international environmental treaties such as the Kyoto Protocol and its eventual successor, especially when only a subset of countries take on commitments to reduce GHG emissions. Also, the design of many national policies aimed at reducing GHGs is based on controlling domestic emissions, which are calculated based on the production taking place within national borders. Ignoring the importance of GHGs embodied in international trade flows has been shown by many (Wyckoff and Roop, 1994; Subak, 1995) to be very important, not only if emission reduction schemes are undertaken in a subset of emitting countries, but also in order to achieve equitable reduction targets. 1.2  Purpose and Research Objectives  Given the phenomena described above, this research primarily seeks to compare the amount of GHGs embodied in Canada‟s imports and exports with the view to find out what is Canada‟s balance of embodied emissions in trade (BEET), calculated simply as the subtraction of the GHG emissions embodied in Canada‟s exports from those embodied in imports. This analysis will answer the question as to whether Canada has a GHG trade surplus or deficit. In doing so, I will also test the hypothesis put forward by Munksgaard and Pedersen (2001) in their Danish case study. When taking into account CO2 emissions embodied in international trade, the authors postulate that there may be an inherent conflict between a national target for CO2 emission reductions committed to under the Kyoto Protocol and the objective of improving trade balances or maintaining trade surpluses. They make this hypothesis after they found that Denmark, as a net exporter of CO22 with a trade surplus, has to make an extra effort to reduce its domestic CO2 emissions attributable to large net exports of CO2-intensive goods. In 2002, the year chosen for this study, Canada also had a large surplus in merchandise trade. This research will thus allow me to verify if the same relationship between positive trade balance and negative BEET holds for Canada given the composition of Canadian imports and exports. If so,  2  As a net exporter of CO2, Denmark has more CO2 emissions embodied in its exports than in its imports, meaning than Denmark‟s BEET is negative.  2  this hypothesis would suggest that Canada is a net exporter of CO2, which could make it more difficult for the country to achieve its national GHG emissions reduction targets.  A secondary objective of this research is to discuss some of the policy implication of my findings in relation to the design of national policies aimed at curbing GHGs and to negotiations of a successor agreement to the Kyoto Protocol on climate change, scheduled to continue at the end of 2011 in Durban, South Africa. I also will discuss some of the key intersections between climate and trade policies, with a particular focus on climate policies that seeks to link the consumption of goods in a country with the amount of GHGs emitted during their production , whether in that country or elsewhere (e.g., energy-related border tax adjustment, low carbon fuel standard, climate labelling). This research fills a gap in the current body of empirical work, as it appears that no recent study has looked at the Canadian case. And even those who have included Canada in their studies have only focused on one GHG embodied in imported goods, either carbon dioxide (Wyckoff and Roop, 1994; OECD, 2003) or methane (Subak, 1995), whereas I propose to look at the following three greenhouse gases: carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O). 1.3  Research Approach  The need for a more complete and balanced picture of countries‟ GHG emissions is being recognized and there is a fast growing body of empirical evidence that attempts to quantify the embodiment of GHGs in international trade flows of specific countries. The analytical framework commonly used to analyze GHGs or other pollutants embodied in countries‟ international trade is the environmental input-output analysis, which over the past decade and a half researchers have modified and improved. Environmental input-output analysis consists of the conventional monetary input-output analysis framework expanded to take into account real physical environmental impacts, such as GHG emissions. I also employ this analytical framework for the purpose of my research, a choice that was facilitated by the existence of freely available webbased software developed by researchers at Carnegie Mellon University‟s Green Design Institute, which contained the environmental data on GHG emissions required for my analysis. Section 3 provides a detailed description of the workings of this model.  3  1.4  Scope  Most nations routinely develop monetary input-output tables that track intersectoral relationships among all sectors of the economy. However, these are colossal undertakings and that explains why they are not developed every year. In the EIO-LCA software that I use for my analysis, the most recent input-output model for Canada is from the year 2002. Therefore I chose to focus my investigation into GHGs embodied in Canada‟s trade flows for 2002. It would be very interesting to continue the analysis for years after 2002 and particularly for the years 2009 and 2010 when Canada experienced a reversal of its long-standing trade surpluses into small trade deficits. However that goes beyond the scope of this Master‟s thesis. 1.5  Thesis Structure  This section has provided an introduction to the thesis, laying out its purpose and objectives, and giving a glimpse of the issues I tackle throughout. Section 2 provides a brief overview of the international legal framework within which developed countries have agreed to reduce their GHG emissions and negotiations are ongoing to set new targets for the post-Kyoto period. Section 3 presents a review of the literature on the theories of producer and consumer responsibility, as well as the empirical evidence that has rapidly accumulated in this field in the last decade and a half. Section 3 also delves further into reviewing the analytical frameworks that are commonly used to perform such investigations, i.e., input-output analysis and variants thereof. Section 4 describes the research approach and methods used in this research to quantify Canada‟s BEET, including the data sources, data collection procedures and the model‟s assumptions and limitations. Results are presented in Section 5 and Section 6 concludes with a discussion of some of the policy implications of my findings.  4  2. International Framework on Climate Change 2.1  United Nations Framework Convention on Climate Change  The earth‟s atmosphere is the ultimate open-access resource. Nations will therefore over-exploit it because they gain all the material advantages from activities that generate greenhouse gas emissions but only suffer a fraction of the related environmental damage, since the costs are shared among all nations. As a consequence, nations are reluctant to reduce their emissions unilaterally, as the cost of doing so would be borne solely by them while the gains would be shared by all (Harrison and Sundstrom, 2007). Restraint by all actors (nations, firms, individuals) is thus essential if their joint action is to effectively curb GHG emissions to the level called for by scientists.3 The global nature of this “tragedy of open access” obviously calls for an international response. That response began to emerge almost twenty years ago, at the Rio Earth Summit (1992). Indeed, the foundation for the international climate change regime is the United Nations Framework Convention on Climate Change (UNFCCC or Framework Convention), a treaty that was adopted in 1992 as a basis for a global response to the challenge.4 Its ultimate objective is to stabilize greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous human-caused interference with the climate system. With 194 Parties, the Framework Convention currently enjoys near-universal membership. Although that treaty represented a groundbreaking response to climate change by creating a general framework for action and cooperation, it did not create legally binding commitments for reducing GHGs and has not produced effective results globally.  The Framework Convention sets forth a series of principles that attempts to balance the aim of environmental protection, economic development and the general division of burdens between developed and developing country Parties. For instance, a fundamental and recurring theme in the UNFCCC is that developed and developing countries have “common but differentiated responsibilities and respective capability,” reflecting a view that developed countries bear a greater historical responsibility for the accumulation of GHG emissions in the atmosphere and 3  Broad scientific consensus holds that more than 2oC of average global warming above the pre-industrial level would constitute a dangerous level of climate change. The science indicates that to avoid exceeding the 2 oC limit, industrialized countries need to reduce their combined GHG emissions to 25 to 40 percent below the 1990 level by 2020 (Pembina Institute and David Suzuki Foundation, 2009). 4 Although opened for signature in 1992 at the Earth Summit, it garnered a sufficient number of ratifications to enter into force in 1994.  5  have greater capacity to take action (Danish, 2007). Thus, the Framework Convention divides the Parties into two main groups: the Annex 1 countries5, which comprise primarily developed countries, and the non-Annex 1 countries, which mainly consist of developing countries. Among the general commitments, the UNFCCC requires that all Parties develop and submit national inventories of emissions by sources and removals by sinks, implement national plans to mitigate climate change and report on their progress. However, Annex 1 countries have further obligations under the UNFCCC, which are to adopt national policies to mitigate climate change and to report on the progress of these policies “with the aim of” returning emissions to their 1990 levels. While this goal became a “soft” commitment of the UNFCCC, the 1990 emissions “baseline” became a touchstone for the development of binding emissions limits under the Kyoto Protocol. 2.2  Kyoto Protocol  The Kyoto Protocol is a historic and significant milestone in the international efforts to address climate change as it constitutes the first, and only, legally binding international agreement under which countries have agreed to set national targets to reduce the anthropogenic greenhouse gases that cause climate change. The origins of the Kyoto Protocol can be found in the growing scientific consensus and alarm regarding climate change in the years following the Earth Summit and a determination that a more forceful international response to the threat of climate change was needed, that ought to include binding emission limits. However, the Kyoto Protocol will not be the final word on this issue, as only a subset of industrialized countries have agreed to national GHG reduction targets and those targets expire in 2012.  The Kyoto Protocol was adopted in Kyoto, Japan on 11 December 1997, although many key details were not resolved and negotiations continued until the detailed rules for implementation were adopted in Marrakesh, Morocco in November 2001. Annex B to the Kyoto Protocol outlines binding quantified emission limits for the Framework Convention‟s Annex 1 Parties to collectively reduce their GHG emissions by an average of 5.2 percent below 1990 levels for the initial 2008-2012 commitment period (Table 1). 6 For the Kyoto Protocol to enter into force, at least 55 countries had to ratify it and 55 percent of Annex 1 countries‟ emissions had to be  5  Annex 1 Parties consists of: EU15 (as in 1997), Australia, Bulgaria, Canada, Croatia, Czech Republic, Estonia, Hungary, Iceland , Japan, Latvia, Liechtenstein, Lithuania, Monaco, New Zealand, Norway, Poland, Romania, Russian Federation, Slovakia, Slovenia, Switzerland, Ukraine and the United States. 6 The binding aspect of the Kyoto Protocol targets is discussed later in Section 2.2.  6  covered. The exit of the United States from the Protocol in March 2001, two months after the inauguration of U.S. President George W. Bush, thus created a crisis in the negotiations as it meant that Russia became the keystone for entry into force of the Kyoto Protocol, without which the 55 percent threshold of emissions would not be attained. Canada ratified the Protocol on 17 December 2002, under the leadership of Prime Minister Jean Chrétien, whose own values, personal commitment and imminent retirement played a key role in the ratification process, despite the strong opposition from Canada‟s business community (Harrison, 2007). The Protocol entered into force on 16 February 2005 once Russia ratified. To date, there are 193 Parties to the Protocol and the percentage of Annex 1 Parties‟ emissions is 63.7 percent of the global total.  Table 1: Quantified Emission Limit or Reduction Commitment under the Kyoto Protocol Annex 1 Countries  Target for 2008-2012 (as % of 1990)  EU15 (as in 1997), Bulgaria, Czech Republic, Estonia, Latvia,  -8%  Liechtenstein, Lithuania, Monaco, Romania, Slovakia, Slovenia, Switzerland United States  -7%  Canada, Hungary, Japan, Poland  -6%  Croatia  -5%  New Zealand, Russian Federation, Ukraine  0  Norway  +1%  Australia  +8%  Iceland  +10%  (Source: UNFCCC Kyoto Protocol at http://unfccc.int/kyoto_protocol/items/3145.php)  When ratifying the Kyoto Protocol in 2002, Canada agreed to curb its emissions to 6 percent below its 1990 level during the 2008-2012 commitment period (Table 1). As required by the Framework Convention, Canada annually submits its GHG inventory to the Climate Change Secretariat. Unfortunately, Canada‟s GHG inventory shows an upward trend in GHG emissions since 1990 instead of the required reductions, which has created a significant gap between the 2008 emissions (the last year for which data is available on the UNFCCC website) and Canada‟s Kyoto target (Figure 11a in Appendix A). As of 2008, Canada‟s emissions were 33.6 percent above 1990 levels, as opposed to 6 percent below per Canada‟s Kyoto target (Table 2).  7  In fact, Canada has not produced a plan that will allow it to come close to meeting its Kyoto commitment and is now the world‟s 8th largest emitter of GHG pollution.7 Table 2: Comparison of GHG Emissions Trends with Kyoto Targets Country  Emissions Growth 8  Target for  (with LULUCF)  2008-2012  1990 to 2008  (as % of 1990)  Australia  +33.1%  +8%  Canada  +33.6%  -6%  EU15 (as in 1997  -7.7%  -8%  Japan  -0.2%  -6%  Russian Federation  -50.8%  0  United States  +15.3%  -7%  (Source: UNFCCC GHG Data – Detailed Data by Party at http://unfccc.int/di/DetailedByParty.do)  Table 2 and Figures 11b to 11f in Appendix A show where other countries stand with respect to their Kyoto commitments. Australia‟s situation is similar to that of Canada‟s despite the fact that they had negotiated a +8 percent commitment. Among the parties surveyed, the EU is the closest to meeting its target, thanks to the closure in the early 1990s of inefficient East German facilities following reunification and the replacement of coal by newly exploited reserves of offshore gas in the UK (Harrison and Sundstrom, 2007). And while Japan‟s emissions have trended upward for most of the period since 1990, they have dropped significantly in 2008, presumably because the global financial crisis and a collapse in domestic demand caused the Japanese economy to shrink in 2008. It now remains to be seen whether Japan will continue on that path to meet the commitment by 2012, especially since the disaster at the Fukushima Daiichi nuclear plant in March 2011, which may result in Japan‟s increase reliance on conventional fossil fuels for energy. The one notable exception is Russia, which negotiated an accommodation commitment under the Kyoto Protocol, allowing it to limit its emissions to its 1990 level during the 2008-2012 commitment period. The collapse of the Russian economy in the 1990s caused the country‟s emissions to fall drastically below its 1990 level and they are projected to stay below that level through 2012.  7  http://www.pembina.org/climate Land use, land use change and forestry (LULUCF) is defined by the Climate Change Secretariat as “a greenhouse gas inventory sector that covers emissions and removals of greenhouse gases resulting from direct human-induced land use, land use change and forestry activities. Annex 1 Parties‟ emission reduction targets include LULUCF. 8  8  As noted previously, the central element of the Kyoto Protocol is the Annex I Parties‟ binding quantified emission limitation and reduction commitments, which are established by Article 3 and inscribed in Annex B. The Kyoto Protocol compliance system was designed to be more robust than that of any other international environmental agreement and includes mechanisms to: 1) generate information about performance; 2) facilitate compliance; and 3) deter noncompliance through penalties. The fundamental measure of compliance under the Kyoto Protocol is the obligation of each Annex I Party to hold a sufficient combination of credits at the end of the commitment period, i.e., at the end of 2012, to cover its emissions (Danish, 2007). A Party that fails to meet that obligation is subject to a penalty. Canada, which is clearly not on a path to meet its Kyoto target by the end of 2012, has not been imposed penalties yet since the first commitment period is not over yet. However, Canada, or any other Party that fails to meet its target, will see its target in the second commitment period reduced by the number of credits sufficient to restore it to compliance, plus a penalty „interest rate‟ of 30 percent. This mechanism was initially designed to deter wilful non-compliance; however, its deterrent effect is significantly reduced because the end of the first commitment period is approaching while no binding targets have yet been negotiated and adopted for the second commitment period. The existence of this penalty mechanism and the magnitude of some countries‟ gaps between the Kyoto targets and their GHG emissions may partly explain the Parties‟ reluctance to agree to new and significant binding emission cuts. 2.3  Toward a Successor Agreement  While the international community has been negotiating the second phase of the Kyoto Protocol since 2005, the challenge has become more pressing each year as the expiry of the first commitment period is closing in. The year 2009 marked a critical point for safeguarding the climate as the Copenhagen conference, held in Denmark in December 2009, marked the deadline Parties had agreed to for the completion of these negotiations and an agreement on a framework to fight climate change in the post-2012 period. Unfortunately, the Copenhagen conference fell significantly short of achieving these ambitious objectives, having only produced the Copenhagen Accord, a weak and non-binding agreement hammered out among a handful of countries and not adopted by the United Nations (UN) as a whole. On January 30, 2010, Canada‟s Environment Minister announced new targets for Canada‟s greenhouse gas emissions for 2020. Canada is now committed to reducing its economy-wide GHG emissions by 17 percent below 2005 levels by 2020. According to the Government of Canada, this target was 9  set internationally in the Copenhagen Accord, is aligned with that of the United States and was formally reiterated in the Cancún Agreement (Environment Canada, 2011). Canada‟s new target is equivalent to an increase of 2.5 percent above 1990 levels.9  Another outcome of the Copenhagen summit was a decision to continue the UN process to deliver a global climate deal, including a continuation of the Kyoto Protocol. A new deadline was set for completing a legally-binding agreement at the Cancún climate change summit in Mexico a year later (December 2010). In their overriding desire to reach any deal, the so-called Cancún Agreements were finally reached, but important dates were left out and major issues about the final legal form and the emission cuts all countries will need to make were pushed back yet another year. One may wonder: what is the point of negotiating new binding targets if countries have not even come close to meeting their earlier commitment? In effect, the world is in limbo, at least until the next climate change summit in Durban, South Africa at the end of 2012. The slow progress under the UN „mega-conference‟ and top-down approach has prompted some analysts over the years to propose some other fora as alternatives to the UNFCCC and Kyoto Protocol. For instance, one alternative approach would bring together a limited number of major-emitting and like-minded countries in a more pragmatic and effective bottom-up approach (Victor, 2005).  9  http://www.greenpeace.org/canada/en/recent/ottawa_wrong_climate_change/backgrounder_harper_2010_gas_target/  10  3. Literature Review 3.1  Theories of Producer and Consumer Responsibility  As indicated previously, the UNFCCC requires all Parties to submit annually their national GHG inventories to the Climate Change Secretariat. Such inventories are derived from the production-based approach and assume that a country‟s GHG emissions are exclusively those arising from the production of goods and services taking place within the country‟s borders, with no distinction as to whether those goods and services are produced for export or for domestic consumption.  By contrast, the consumption-based approach excludes emissions associated with exports and includes emissions generated in the production of imports by tracing these back to their place of origin and estimating their emissions based on the production processes used to create them. Muradian et al. (2002), amongst others, postulate that production is conditioned and determined by consumption patterns. If we accept this view, then consumption is a key economic force „steering‟ the production of goods and their related GHG emissions. Rothman (1998) argued early in favour of using a consumption-based approach to measure the full impacts associated with the expenditures taking place in a specific country, with particular attention to international trade. Rothman (1998) cites Rees (1995) and Daly (1996) who argue that “most environmental degradation can be traced to the behaviour of consumers either directly, through activities like the disposal of garbage or the use of cars, or indirectly through the production activities undertaken to satisfy them”. The consumption-based approach is also at the heart of ecological footprint analysis, which is a measure of human demand on the Earth‟s ecosystems. The ecological footprint represents the area of biologically productive land and sea necessary to supply the resources a human population consumes, and to assimilate its carbon emissions. The ecological footprint analysis thus aims to increase the consumers‟ awareness that they are ultimately responsible for the environmental impacts engendered through consumption. Of course, the reality is certainly more complex, as governments offer incentives to develop national industries (with some causing more environmental degradation than others) and firms engage in marketing campaigns to drive consumption.  The principle behind a consumption-based approach is that the GHG emissions related to the production of goods should be allocated to the final consumers of those goods since they gain 11  the utility and to some extent stimulate the production process via their demand. With this approach, the level of GHG emissions attributed to a given country should be the same wherever production occurs (allowing for differences in GHG intensity of different production processes): in effect, the emissions generated at each step of the production process are „embodied‟ in the product sold. Calculating these „embodied‟ emission can prove quite challenging, especially when they need to be traced across all supplying economies. The main difficulty arises from data requirements and availability and various methods have been developed to estimate the „embodied‟ emissions in traded goods (see Section 3.3).  The relevance of the consumption-based approach can be better grasped through the concept of emissions leakage. Leakage occurs when a significant portion of an emitting industry escapes regulation. This problem is of particular relevance for transboundary pollutants, since foreign producers will be outside the jurisdiction of a domestic regulator. In this case, leakage occurs when there is an increase in emissions in country A as a result of an emissions reduction in country B with a stringent emissions policy. In relation to GHGs, this phenomenon is known as carbon leakage and sometimes also refers to the carbon dioxide (or other GHGs) embodied in goods imported from countries without any GHG reduction target under the Kyoto Protocol. These GHGs escape the importing country‟s domestic GHG regulation although they still contribute to climate change through the emissions in the exporting countries (Mongelli et al., 2006; Fischer and Fox, 2007). 3.2  Empirical Evidence  The need for a more complete and balanced picture of countries‟ GHG emissions is being recognized and there is a fast growing body of empirical evidence that attempts to quantify the embodiment of GHGs in international trade flows. In doing so, such studies make a clear distinction between a production-based approach and a consumption-based approach. Over the past decade and a half, researchers have developed and improved various methods to analyze GHGs or other pollutants embodied in the internationally traded goods of an increasing number of countries worldwide. Several single-country studies (Australia – Lenzen, 1998; Brazil – Machado et al., 2001; Denmark – Munskgaard and Pedersen, 2001; Finland – Mäenpää and Siikavirta, 2007; Italy – Mongelli et al., 2006; U.S. – Weber and Matthews, 2007, Ghertner and Fripp, 2007, and Shui and Harriss, 2006), as well as multiple-country studies (Wyckoff and Roop, 1994, Ahmad and Wyckoff, 2003; Subak, 1995) have been completed. 12  Among the earliest studies, Wyckoff and Roop (1994) and Subak (1995) investigated respectively the amount of carbon or methane embodied in a subset of industrial or agricultural goods imported into the six largest OECD countries (U.S., Canada, France, United Kingdom, Germany and Japan) to assess whether the import of carbon- or methane-rich products was an issue worth addressing. Using a monetary input-output (IO) analysis and emissions data to estimate the amount of CO2 emissions embodied in the imports of manufactured goods of these OECD countries, Wyckoff and Roop (1994) found that about 13 percent of the total carbon emissions attributable to these six countries were „embodied‟ in manufactured imports. They conclude that the significant embodiment of carbon in manufactured goods has several implications that should be considered in designing policies aimed at curbing GHGs. For instance, conventional measures of CO2 emissions that rely solely on domestic sources will mislead if these figures have been reduced through increased imports of high-carbon-content products. In fact, if emission control measures are implemented domestically that raise the cost of these products; de facto incentives are created to import these products from countries with no such measures. This in turn increases the environmental burden of the exporting countries to satisfy demand in the importing countries. This evidence supports the need for some form of border tax adjustments on imports that would complement any domestic measures raising the price of carbon. For her part, Subak (1995) found that while the methane embodied in the agricultural imports of the six OECD countries under study was small, the leakage potential was of a sufficient scale to undermine domestic abatement goals. The author concludes that a change in the way GHGs are estimated so that it takes into account emissions in traded goods.  A few recent single-country studies have focused on open economies such as Denmark specifically to assess the impact of a CO2 trade balance10 on the country‟s ability to achieve its national CO2 reduction targets when these are measured based on the conventional productionbased approach. For instance, Munksgaard and Pedersen (2001) found that Denmark, as a net exporter of CO2, has to make an extra effort to reduce its domestic CO2 emissions attributable to exports of CO2-intensive goods. More specifically, Munksgaard and Pedersen (2001) conclude that the case of Denmark illustrates an inherent conflict between a national CO2 target for CO2 emissions reductions and the objective of improving the country‟s trade balances. Empirical findings for Australia, Brazil, Canada and Finland also suggest that these countries  10  A CO2 (or more broadly emissions) trade balance refers to the difference between the CO 2 emissions calculated using a production-based approach and those calculated using a consumption-based approach.  13  are net exporters of CO2 when international trade is taken into account (Lenzen, 1998; Machado et al. 2001; Mäenpää and Siikavirta, 2007; Peters and Hertwich, 2008). Other studies specifically tried to quantify the carbon leakage associated with the emissions embodied in traded goods (Italy – Mongelli et al., 2006; U.S. – Ghertner and Fripp, 2007; Weber and Matthews, 2007 and Shui and Harriss, 2006).  The most comprehensive multi-country study in terms of geographical coverage is the OECD study (Ahmad and Wyckoff, 2003). The purpose of this study was to estimate the CO2 emissions embodied in the international trade of goods for 24 countries (mostly OECD countries, as well as the BRIC11 countries). The authors wanted to create an indicator of CO2 emissions related to domestic demand as a complement to the more common indicator of CO2 emissions associated with domestic production. This research shows that estimates of CO2 emissions generated to satisfy domestic demand in the OECD in 1995 were 5 percent higher than emissions related to domestic production. However, this average figure masks the fact that for many countries, the emissions associated with consumption are often 10 percent greater or less than with domestic production. For Canada, the authors calculated that the CO2 emissions associated with domestic consumption were 10.9 percent lower than those related to domestic production.  Except for the OECD study, all empirical analyses cited above also discuss the policy implications of their findings for the negotiations of international agreements on climate change, as well as for the design of domestic policies to curb GHGs. 3.3  Review of Analytical Frameworks  3.3.1  Economic Input-output Analysis Framework  This section aims to provide a basic conceptual understanding of the various analytical frameworks that have been used to date to deal with the issue of the embodiment of GHGs in international trade flows. As noted above, economic input-output (IO) analysis has been extensively used to assess GHGs embodied in the international trade flows of specific countries (Australia - Lenzen, 1998; Brazil -Machado et al., 2001; Denmark - Munksgaard and Pedersen, 2001; Finland - Mäenpää and Siikavirta, 2007; Italy - Mongelli et al., 2006). 11  Brazil, Russia, India and China.  14  An IO model is a well established linear economic model used to account for economic impacts following a change in the total output produced by the economy. This model was first developed in the United States in the 1930s by Wassily Leontief (an accomplishment for which he received the Nobel Prize in 1973) to represent the various inputs from all sectors of the economy required to produce the output in each economic sector. It is important to note that IO analysis usually uses monetary flows, as opposed to real physical flows of inputs and outputs.  Most nations routinely develop such monetary input-output models, although not to a level of detail that matches that of the United States. The U.S. IO models are created every five years by the U.S. Department of Commerce and represent the transactions among 400+ industry sectors. The most recent input-output model of the U.S. economy is for the year 2002 and divides the economy into 428 sectors. In comparison, Canada‟s most recent IO model, also for the year 2002, has 105 sectors.  Table 3 shows the structure of an input-output model, where all inputs and outputs are measured in dollars. Within the input-output table, each Xij represents input to sector j from sector i in the production process. The column sum Ij represents the total amount of inputs to sector j from all the other sectors. Total output for each sector Xi is the sum across the rows of the intermediate outputs used by the other sectors Oi and the output supplied to final demand by consumers Yi. The gross domestic product (GDP) is the vertical sum of all the final demands Yi or the horizontal sum of all the value added Vj.12 From the input-output table below, a different matrix or table is created A that represents the direct requirements of the intersectoral relationships. The rows of A indicate the dollar amount of output from sector i required to produce one dollar of output from sector j. The resulting aij, or technical coefficients, are considered the direct requirements, i.e., the output from first-level suppliers directly to the industry of interest.  12  It is worth noting that the sectoral relationships shown in Table 3 represent the interactions within a particular economy. With everexpanding trade between countries, attempts at accounting for total supply chain effect are greatly complicated by the predominance of internationally supplies resources and products, including at the intermediate inputs level.  15  Table 3 Example of an Economic Input-Output Table  Output from sectors (i)  Input to sectors (j) n 1 2 3  Intermediate Output O  Final Demand Y  Total Output X  1  X11  X12  X13  X1n  O1  Y1  X1  2  X21  X22  X23  X2n  O2  Y2  X2  3  X31  X32  X33  X3n  O3  Y3  X3  n  Xn1  Xn2  Xn3  Xnn  On  Yn  Xn  Intermediate input I  I1  I2  I3  In  Value added V  V1  V2  V3  Vn  Total input X  X1  X2  X3  Xn  GDP  (Source: Reproduced from Hendrickson et al. (2006), p. 14)  While a more detailed and mathematically-oriented description of economic input-output analysis can be found in Hendrickson et al. (2006), it is sufficient to note here that the required economic purchases in all sectors of the economy (or total output) necessary to produce a vector of desired final demand y can be calculated as:  (1) where x is the vector of total output, I is the identity matrix,13 A is the input-output direct requirements matrix as described above, and y is the vector of desired final demand. In Equation 1, the term (I × y) represents the level of final demand itself and the term (A × y) represents the contributions from the first-level suppliers to satisfy the final demand y. 14 The effects of these initial purchases continue throughout the economy, as the demand for output from the first-level suppliers in turn creates a demand for output from their direct suppliers. These second-level supplier requirements are calculated by further multiplying the direct requirement matrix by the final demand, or A × A × y. The same procedure applies to calculate third, fourth and more levels of suppliers. The infinite series of the supply chain can be replaced by (I – A)-1, which is called the total requirements matrix (for both direct and indirect requirements) or the Leontief inverse (Hendrickson et al., 2006).  The relationship between final demand and total output can then be expressed as:  13  The identity matrix is a matrix of all zeros except for the diagonal elements, which equal to one. The term (I x y) is obtained by multiplying the first term in the parenthesis I by y and the term (A x y) is obtained by multiplying the second term in the parenthesis A by y. 14  16  (2)  or  (3)  Equation 3 indicates that the IO framework can be used to determine changes in total output following an incremental change in final demand. This explains why, throughout the world, IO analysis has been extensively used for planning, in particular to estimate additional output required of all sectors of the economy to support increases in demand in any given sector and why this tool has been especially useful for assessing the potential economic impacts of governments‟ stimulus packages in their concerted efforts towards economic recovery. 3.3.2  Environmental Input-Output Analysis Framework  The application of the monetary input-output framework later evolved to handle the assessment of the environmental repercussions of economic activity (Leontief, 1970). This type of IO analysis is often referred to as “environmental input-output analysis”. The first such analyses were unwieldy and difficult to use in practice due to the complexity of matrices that combined economic data with data on resource requirements and environmental emissions15 (Bjorn et al., 2005). However, with the dramatic increase in computational power seen in the last two decades, the required matrix algebra can now be easily handled, even with personal computers.  In practice, the monetary IO analysis is augmented with additional environmental data, such as sectoral data on resource requirements and environmental emissions that can include energy use, releases of conventional pollutants and greenhouse gas emissions. Environmental IO analysis can then be used to analyse economy-wide environmental impacts, such as emissions or resources and energy used, associated with changes in the output of selected industrial sectors (Leontief, 1970). In the model that I use for this research, the EIO-LCA model, which I will briefly describe in the next sub-section, the vector of total output x is multiplied by the average environmental impact or resource requirement for each sector ri, and the aggregation of these individual impacts represents the total supply chain impact of the final demand y. For greater clarity, using Equation 2, the model can estimate the outputs required throughout the economy to produce a specified set of goods y. The total of these outputs x is referred to as 15  For instance, resource requirements can refer to energy use (measured in terajoules) or water use (measured in thousands of gallons) and environmental emissions can refer to emissions of conventional air pollutants, greenhouse gases or toxic releases (measured in metric tons).  17  the supply chain for the goods in question where the chain is the sequence of all the suppliers. Multiplying the output of each sector (xi) by the environmental impact per dollar of output of sector i (ri), we obtain a vector of direct environmental impacts bi that result from the production of the specified set of goods y.  (4)  where Ri is a matrix with ri on the diagonal and zeros on the off-diagonal. When all elements of the vector bi are summed up, then the total supply chain environmental impact associated with a set level of final demand y is obtained.  Therefore, just as IO analysis estimates the additional production required in all sectors to support an increased output in any given sector, the extended model can evaluate how such an increase in output requires increased use of natural resources or causes increased discharges to the environment. This model is particularly useful for researchers interested in calculating the physical amount of GHGs embodied in the goods produced by a country as they can use monetary IO tables and augment them with GHG intensity factors.16 This model is also being used, with increasing popularity, to estimate the level of GHG emissions embodied in the goods imported and exported by specific countries. In this research, this means that in Equation 2, y is replaced by import and export figures for each sector of the economy and in Equation 4, ri represent the sectoral GHG emission intensity factor. 3.3.3  Economic Input-Output Life Cycle Assessment Framework  The EIO-LCA model (Economic Input-Output Life Cycle Assessment model) referred to above was developed by a group of researchers from the Green Design Institute at Carnegie Mellon University (CMU) who sought to improve upon an existing life cycle assessment17 (LCA) method, which suffered from several major limitations, including being time-consuming, costly, and practically impossible to perform for complicated products with a very high number of components. They developed a novel approach to LCAs by incorporating monetary input-output  16  The GHG intensity factor measures the amount of GHG emissions (in metric tons of CO 2 equivalent) per dollar unit of economic activity, i.e., total GHG emissions divided by total output. It can be calculated at different levels of aggregation, e.g., for a country or an industry sector. 17 Life cycle assessments “study the environmental aspects and potential impacts throughout a product‟s life (i.e., “cradle-to-grave”) from raw material acquisition through production, use and disposal” (Hendrickson et al., 2006).  18  accounts to estimate the physical impacts resulting from the production of goods and services across their entire supply chain (Lave et al., 1995; Hendrickson et al., 1998). What makes the use of this model attractive for the purposes of my research is that the CMU‟s Green Design Institute developed software based on the EIO-LCA model. This software has a user-friendly interface and allows the user to perform analyses in a time-effective manner and at no cost, since it is available free of charge on the Internet at www.eiolca.net. In addition to U.S. databases, datasets for a few other countries are also available on that website. These countries consist of Canada, Spain, China and Germany. A number of researchers have already used this software outside the realm of life cycle assessments per se, to estimate GHG emissions embodied in the United States‟ trade flows (Shui and Harriss, 2006; Ghertner and Fripp, 2007; Weber and Matthews, 2007), as I did in this research. Norman et al. (2007) also used this software but significantly modified the U.S. and Canadian versions of the model to create a bi-national Canada-U.S. EIO-LCA model that implicitly considers trade in goods between Canada and the U.S. to estimate the economy-wide energy intensity and GHG emission intensity for 45 manufacturing and resource sectors in Canada and the U.S. The authors find that overall, U.S. manufacturing and resource industries are about 1.15 times as energy-intensive and 1.3 times as GHG-intensive as Canadian industries, with significant sector-specific discrepancies.  While a detailed description of the EIO-LCA model can be found in Hendrickson et al. (2006), the basic steps are explained in Figure 1 and a case-specific illustration is provided in Figure 2.  First, the purchase of a product is identified, such as the purchase of a motor vehicle. This purchase is used as the desired final demand y for the EIO-LCA model (Step 1). Then, all the supply chain requirements for that product are estimated x and the process of identifying purchases continues until all the stages of the product are represented (Step 2). The Green Design Institute‟s software then simultaneously computes the environmental impacts bi resulting from the initial purchase and the entire supply chain (Steps 3) and sums them up (Step 4) (Hendrickson et al., 2006).  19  Figure 1: Steps in the EIO-LCA Process Step 1: Estimate changes to final demand by sector (y)  Step 2: Assess direct and indirect economic change with I/O model (x)  Step 3: Assess environmental discharges as a result of sector output changes (bi)  Step 4: Sum sector discharges to find overall discharges (  Figure 2: Illustration of the EIO-LCA Model: Motor Vehicles  (Source: Reproduced from Bjorn et al. (2005), p. 21)  20  3.3.4  Single-Region (or Country) Input-Output Models  Environmental IO analysis has long been recognised as a useful method to attribute pollution or resource use to final demand in a consistent framework. As noted earlier, there is a fast growing body of studies based on environmental IO analysis that adopt a consumption approach and investigate environmental pressures, mainly GHG emissions, embodied in international trade flows.  In a single region (or country) input-output (SRIO) model, one key assumption researchers often make is that imported goods are being produced with the same production processes (i.e., same energy efficiency and same fuel type) as they would be in the importing country and thus have the same GHG emission intensities as the domestic goods. This is because the model does not permit the distinction between domestic and foreign production technology. This is illustrated in Figure 3 below where country A‟s imports from countries B, C, D and E are assumed to be produced using country A‟s production processes, energy inputs and GHG intensity factors. Of course, when one is concerned with estimating the GHG embodied in products, this assumption is tenuous and can greatly impact the results. Figure 3: SRIO Model and Autonomous Economies  A  B  E  C  D  (Source: Adapted from Lenzen et al. (2004), p. 395)  Table 4 provides a non-exhaustive list of the recent studies I surveyed that are based on the SRIO model. Wiedmann et al. (2006) and Widemann (2009) provide a more extensive review of the empirical literature on this topic. Most of the studies listed in Table 4 carry out an IO analysis of a closed economy, and subsequently apply multipliers obtained from the direct requirements  21  matrix A to exports and imports. In this approximation, the import structure does not enter the direct requirements matrix, and is hence not reflected in the multipliers (Wiedmann et al., 2007). Table 4: Recent Studies based on Single-Region Input-Output Analysis Reference  Countries  Environmental Impacts  Ghertner and Fripp (2007)  U.S.  GHG, energy, toxics, and air pollutants  Lenzen (1998)  Australia  Primary energy and GHGs  Machado et al. (2001)  Brazil  Energy and CO2  Mäenpää and Siikavirta (2007)  Finland  CO2, CH4, N2O  Mongelli et al. (2006)  Italy  Energy and GHG  Denmark  CO2  U.S.-China  CO2  18  Munksgaard and Pedersen (2001) Shui and Harriss (2006)  12  However, it is generally the case that a country‟s imports are sourced from a number of different countries with different production technologies. And that each of these countries also places import demands on foreign economies. Thus, embodied production factors may continue far upstream in an international supply chain in the same way that intersectoral demands continue far upstream on the domestic level (Wiedmann et al., 2007). SRIO models are limited in that they cannot take into account these differences in production technologies and the international supply chain. This shortcoming thus calls for the use of multi-region input-output (MRIO) models where inter-regional trade flows are internalised within the intermediate demand. 3.3.5  Multi-Region (or Country) Input-Output Models  In recent years, more sophisticated models have emerged based on detailed statistics and multi-region input-output (MRIO) models. Two types of MRIO can be distinguished: 1) linked single-region models; and 2) true multi-region models (Wiedmann et al., 2007). With linked SRIO models, national IO tables are exogenously linked with bilateral trade data for different countries or regions and embodied emissions are calculated for each country or region separately. This approach captures only the last stage of an international supply chain of imports but is nonetheless a first improvement from the simpler SRIO model. Figure 4 further illustrates this approach. Under the unidirectional trade scenario, country A‟s imports are decomposed into imports from countries B, C, D and E. It is assumed that the imports into  18  These studies use the EIO-LCA software developed by Green Design Initiative at Carnegie Mellon University.  22  country A from countries B, C, D and E are produced using each of these countries‟ respective production technology, energy inputs and GHG intensity factors. Figure 4: MRIO Models with Unidirectional and Multidirectional Trade Unidirectional trade  Multidirectional trade  A  A  B  E  C  B  E  C  D  D  (Source: Adapted from Lenzen et al. (2004), p. 395)  By contrast, true MRIO models endogenously combine domestic technical coefficient matrices with import matrices from multiple countries or regions into one large coefficient matrix, thus capturing trade supply chains between all trading partners as well as feedback loops.19 As can be seen in Figure 4 under the multidirectional trade scenario, all trade flows are assumed to be produced using the production technology, energy inputs and GHG intensity factors of the country of origin. Not surprisingly, these models have become increasingly complex from a mathematical perspective. Table 5 provides a non-exhaustive list of the recent studies I surveyed that are based on either MRIO models. Again, Wiedmann et al. (2006) and Wiedmann (2009) provide a more comprehensive review of recent MRIO studies. Table 5: Recent Studies based on Multi-Region Input-Output Analysis Reference  Countries  Environmental Impacts  Ahmad and Wyckoff (2003)  24 countries (mostly OECD)  CO2  Lenzen et al. (2004)  Denmark, Germany, Norway,  CO2  Sweden and rest of world 20  Norman et al. (2007)  Weber and Matthews (2007) Wyckoff and Roop (1994)  14  Canada-U.S.  GHG  U.S. and 7 countries  CO2, SO2, NOx  6 large OECD countries  CO2  19  Feedback loops capture changes in production in one region that result from changes in intermediate demand in another region, which are in turn brought about by demand changes in the first region (Wiedmann et al., 2007). 20 These studies use the EIO-LCA software developed by Green Design Initiative at Carnegie Mellon University.  23  The study by Lenzen et al. (2004) took into account the production methods of Denmark‟s top three trading partners (Germany, Sweden and Norway) and their results show that Denmark‟s CO2 trade surplus of 11 MMT resulting from a single-region model turned into balance when multidirectional trade is considered. They conclude that explicitly taking into account the production processes and energy use structure of key trading partners is important when calculating the CO2 embodied in trade and increases the accuracy of the analysis, despite the fact that more than half of Danish imports had to be lumped together and treated with world average data.  24  4. Research Approach and Methods 4.1  Model Specification  As noted, one of my research objectives is to compare the amount of GHGs emitted in connection with the production of goods in Canada with the quantity of GHGs emitted in connection with the consumption of goods taking place in Canada. In other words, my goal is to evaluate Canada‟s balance of embodied emissions in trade (BEET). Canada‟s BEET is the difference in the level of Canada‟s GHG emissions calculated based on the two distinct approaches described earlier: the consumption and the production-based approaches. Analysts also sometimes refer to the BEET as the emissions trade balance.  In Equation 5, C is a vector referring to the total supply chain associated with goods consumed in Canada, whether produced domestically (Pd) or imported (IM) whereas P is the vector that refers to the total supply chain related to goods produced in Canada, whether consumed domestically (Pd) or exported (EX). It is worth noting that vectors C, Pd, IM and EX are specific cases of the total supply chain output denoted by x in Equation 4 on page 18 and they are all measured in dollars. Rc and Rp respectively represent GHG emission intensities associated with goods consumed in Canada (whether produced domestically or imported) and with goods produced in Canada (whether consumed domestically or exported) and are measured in metric tons of CO2e per dollar. Therefore, the BEET‟s measurement unit is metric tons of CO2e. (5)  (5‟) (6) (7) Rearranging Equation 5 into Equation 5‟, the first two terms cancel out because they each represent the GHG emissions associated with the goods that are both produced and consumed in Canada. Equation 5‟ is then re-written as Equation 6, where vectors Rim and Rex are specific cases of Rc and Rp and represent the GHG emission intensities associated with imported and exported commodities respectively. Thus, using Equation 6 is a simpler way of calculating the BEET, where the embodied emissions in exports (EEE) are subtracted from those in imports 25  (EEI) as in Equation 7. This method eliminates the need to calculate the full consumption- or production-related emissions. Also, as mentioned above, since IM and EX are specific cases of the total supply chain output denoted by x in Equation 4, and using Equation 2 to substitute x, we obtain the following Equation 8, where im and ex are vectors corresponding to imports and exports of goods.  (8)  Given this definition, a country with a positive BEET implies that the GHGs emitted in connection with the production of imported goods surpass those emitted in connection with the production of exported goods, while a negative BEET implies exactly the opposite. The BEET analysis can be an appealing tool to make people aware that their consumption behaviour has environmental consequences beyond their country‟s borders (Muradian et al., 2002). In that sense, it can be compared to the ecological footprint, an indicator that tries to capture a snapshot of humanity‟s demand on natural resources and which also follows the principle of consumer responsibility. 4.2  Data Sources  To estimate Canada‟s BEET, I need data with respect to the four vectors identified in Equation 8 above: imports (im), exports (ex), the GHG emission intensities associated with the imported goods (Rim) and the GHG emission intensities associated with the exported goods (Rex). To obtain Canada‟s import and export data, I use Industry Canada‟s Trade Data Online (TDO) database, which provides data on all merchandise trade clearing customs and is available free of charge on the Internet.21 The data on the import and export GHG intensities is indirectly available through the EIO-LCA software developed by CMU‟s Green Design Institute and accessible on its website.22  As explained in Section 3.3.3, the EIO-LCA method combines physical data on resource use and emissions with traditional monetary input-output tables to calculate the total environmental impacts (measured in physical units) associated with a specified change in the level of output in the economy, which in this case will be imports and exports. The EIO-LCA software provides  21 22  www.ic.gc.ca/tdo http://www.eiolca.net/  26  information on a number of environmental impacts: GHG emissions (in thousands of metric tons of CO2e), conventional air pollutants (in metric tons), energy use from a number of fuel sources (in terajoules) and water use (in thousands of gallons). Given that my research focuses on greenhouse gases, I retrieve only the data on GHG emissions associated with the total supply chain of specified levels of Canadian imports and exports.  For the GHG emission intensity factors associated with exports (Rex), since goods produced in Canada use the same production processes, whether they are destined for export or domestic consumption, their GHG emission intensity is the same and is indirectly available through the EIO-LCA model for Canada. With respect to the GHG emission intensity factors associated with imports (Rim), the situation is more complex. Since goods imported into Canada may be sourced from multiple countries around the world, accurate results require knowing the GHG emission intensity for each of Canada‟s trade partners‟ production processes.  As indicated before, it is often assumed in SRIO models that imported goods are being produced with the same technology as the domestic technology in the same sector. In my first estimates series, I also assume the same GHG emission intensity for Rex and Rim. Making this assumption allows me to use the Canadian EIO-LCA model with respect to the GHG emission intensity associated with both goods imported into and exported from Canada. This approach corresponds to a SRIO model for Canada. However, when one is concerned with the GHGs embodied in products, this assumption is questionable and may significantly affect the results.  As discussed earlier, a number of researchers have dealt with this problem by developing multiregion IO models. Likewise, I have addressed this issue by using a multi-region IO model in a second estimates series, where Canada and the United States (U.S.) are the two countries included in the model. In 2002, the U.S. accounted for 63 percent of Canada‟s imports by value. Thus, by using the EIO-LCA model for the U.S. economy to estimate the GHGs embodied in the goods manufactured in the U.S. and imported into Canada, I am able to use the appropriate vector of GHG emission intensity (Rim) for 63 percent of Canada‟s imports. In 2002, China only accounted for 4 percent of Canada‟s imports. Thus, I decided not to include China in my MRIO analysis because the costs (in time and complexity related to trade data conversion) would have outweighed the benefits of having greater estimates‟ accuracy, given China‟s relatively small share of Canada‟s total imports. For the remaining 37 percent of goods coming from the rest of the world, I still use the Canadian model, which in effect assumes that those imports are 27  produced using the same production processes as in Canada, despite the fact that this produces a wider margin of error. Even in multi-country studies, this assumption still generally applies to a significant share of imports, up to 50 percent sometimes (Lenzen et al., 2004) so my approach is consistent with that of other researchers in this field.  Since the most recent online versions of the EIO-LCA models for Canada and the U.S. both use 2002 input-output data, I performed my analysis for the year 2002. This had a number of benefits: 1) it eliminated data reconciliation problems when combining the two models for the analysis; and 2) it avoided the need to assume the stability of the IO technical coefficients over time. Given the time and data requirements to produce IO tables, statistic offices do not produce them annually. Therefore, if the last IO tables available had been several years older than the year selected for my study, what I would have used is an older snapshot of the economic structure, neglecting structural and technological changes that could have led to improved efficiencies in the meantime (Mongelli et al., 2006). Fortunately, this is not the case here, as I performed the analysis for the year corresponding to the input-output data. Similarly, I did not have to make the assumption that the GHG intensities remained constant over time. In short, by choosing the year 2002 for my analysis, I limited the uncertainty in my emissions estimates. 4.2.1  Data Sources of the EIO-LCA Model for Canada  The first version of the EIO-LCA model for Canada was developed by researchers at the University of Toronto for the year 1998. Bjorn et al., (2005) provide a detailed account of the development of that model, along with the sources of input-output and environmental data. Statistics Canada and the University of Toronto subsequently updated the Canadian EIO-LCA model with 2002 data; this latter model was made available through the Green Design Institute website in March 2008. The 2002 Canadian EIO-LCA model is derived from input-output data from Statistics Canada and includes 105 sectors. Environmental and resource use data are provided by Statistics Canada, Environment Canada and Natural Resources Canada.  The 105 sectors included in the Canadian model encompass both goods-producing sectors and services-oriented ones. Thus, while my research focuses only on the GHG emissions embodied in the goods that cross the Canadian border, those emissions arise from the total supply chain involved in producing those goods. It is this total supply chain that may in turn involve any of the 105 goods- or services-producing sectors. The example of motor vehicles in Figure 2 provides a 28  good illustration. Let‟s assume that one motor vehicle is exported from Canada. In that case, it is a good that is crossing the border but the GHGs embodied in that vehicle come from its total supply chain. As can be seen from the direct and indirect impacts in Figure 2 (p. 20), GHGs are emitted along the supply chain from goods-producing sectors (e.g., motor vehicle parts, primary iron and steel) and services-oriented ones (e.g., business services).  The Canadian EIO-LCA model includes the three primary GHGs regulated under the Kyoto Protocol: carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). The data source for these GHG emissions is the Environment Canada Greenhouse Gas Inventory for 2002, the year of the model (Bjorn et al, 2005).23 Carbon dioxide, produced largely from fossil fuel combustion, is the single largest anthropogenic contributor to global warming. Process emissions of CO2 also occur in the creation of cement, when limestone is broken down, releasing carbon dioxide into the atmosphere. Methane from human activity is the second anthropogenic greenhouse gas in importance, and arises mostly from natural gas exploitation, agriculture and decomposition in landfills. Nitrous oxide, produced largely from the heavy use of fertilizers in agriculture and to a lesser extent from fossil fuel combustion, also has an important greenhouse gas effect, even though to a much lesser degree than the first two (Weaver, 2008). Indeed, recent estimates show that, globally, carbon dioxide is responsible for a significant share of anthropogenic global warming at 43.1 percent, followed by methane at 26.7 percent and far behind, nitrous oxide with a 3.8 percent contribution (Gore, 2009). This is represented graphically in Figure 5. Figure 5: Global Warming Pollutants  Halocarbo n 8% CO, VOCs 7%  Black carbon 12%  Nitrous oxide 4% Carbon dioxide 43%  Methane 26%  23  In 2002, CO2 contributed the largest share of 2002 emissions at 78.8 percent (about 576 MMT), followed by methane with a 12.9percent share (94 MMT) and nitrous oxide with 7.2 percent of emissions (53 MMT), for a total of 98.9 percent of total emission. Perfluorocarbons contributed 0.7 percent (5 MMT) and sulphur hexafluoride and hydrofluorocarbons constituted the remainder.  29  Emissions derived from the Environment Canada‟s inventory and included in the EIO-LCA model are those from fuel combustion, fugitive emissions from coal mining, and cement and lime production. In the 1998 inventory, those emissions accounted for 73 percent of the 689 MMT of CO2-equivalent (CO2e) emitted during the year by all sectors of the Canadian economy. By 2003, they accounted for close to 83 percent of the 740 MMT CO2e emitted by Canada (Environment Canada, 2005). Excluded from the EIO-LCA model are fugitive emissions from oil and gas extraction, solvent use, animal sources, soil management, land cover change, or waste disposal, as their estimates are highly imprecise or difficult to associate with sectors in the model (Bjorn et al., 2005). Also excluded from the Canadian model are emissions of the other three greenhouse gases regulated under the Kyoto Protocol: hydrofluorocarbons (HFCs), perfluorocarbons (PFCs) and sulphur hexafluoride (SF6).24 These industrial compounds25 are the most potent greenhouse gases and many of them have exceptionally long atmospheric lifetimes, resulting essentially in their irreversible accumulation in the atmosphere once emitted (U.S. Environmental Protection Agency).26 However, their current atmospheric concentrations are very low (Weaver, 2008). The emissions‟ shares of the six regulated GHG for Canada in 1990 and 2007 are shown in Figure 6. From 1990 to 2007, the share of the three primary GHGs (CO2, CH4 and N2O) has remained stable at 99 percent. Taken together, GHGs not included in my model account for about 17 percent of Canada‟s total GHG emissions.  Since various GHGs contribute differently to global warming, they need to be expressed in a common measurement unit. The IPCC‟s reference gas is CO2, which explains why GHG emissions are typically expressed in CO2 equivalent (CO2e). The EIO-LCA models I use in this research have used the 100-year global warming potentials, the standard reporting approach under the Kyoto Protocol, to calculate CO2e for the purposes of reporting total GHG emissions. This particular measure of “global warming potential” (GWP) has been defined as the total expected contribution of a given mass of greenhouse gas over 100 years relative to the contribution of the same mass of CO2 over the same time period. Thus, both the radiative forcing27 and the expected lifetime of the compound in the atmosphere are taken into account in 24  The EIO-LCA model for the United States includes HFCs and PFCs emissions, in addition to those of CO2, CH4 and N2O. HFCs are man-made chemicals, many of which have been developed as alternatives to ozone-depleting substances for industrial, commercial and consumer products. Primary aluminum production and semiconductor manufacture are the largest known manmade sources of two perfluorocarbons, tetrafluoromethane and hexafluoroethane. Sulphur hexafluoride is used for insulation and current interruption purposes in many industrial applications. 26 U.S. Environmental Protection Agency: http://www.epa.gov/highgwp/scientific.html 27 The IPCC defines radiative forcing as a measure of the influence that a factor has in altering the balance of incoming and outgoing energy in the Earth-atmosphere system and is an index of the importance of the factor as a potential climate change mechanism (IPCC AR4, 2007). 25  30  GWP calculations (IPCC, 2007). Under this accounting system, 1 ton of methane is equivalent to 23 ton of CO2e and 1 ton of nitrous oxide is equivalent to 296 ton of CO2e (Bjorn et al., 2005). Figure 6: Emission Trends by Gas  CH4 12%  N2O HFCs PFCs SF6 9% 0% 1% 1%  CH4 13%  HFCs PFCs SF6 0% 0% N2O 1% 7%  CO2 79%  CO2 77%  1990  2007  (Source: Environment Canada (2009). The National Inventory Report 1990-2007)  4.2.2  Data Sources of the EIO-LCA Model for the United States  The U.S. models are constructed using the benchmark input-output data produced every five years by the Bureau of Economic Analysis of the U.S. Department of Commerce. Three models are currently available online through CMU‟s Green Design Institute, for the years 1992, 1997 and 2002. The latter is based on the 428 x 428 commodity-by-commodity input-output matrix and represents a higher level of disaggregation, and precision, than that of the Canadian model. This difference needs to be taken into account when retrieving data on Canadian imports from the United States and I will elaborate on this key issue in Section 4.3.2.  With respect to greenhouse gases, data is derived from U.S. governmental sources. The U.S. Environmental Protection Agency (EPA) develops annually the national greenhouse gas inventory to track the trend in emissions and removals since 1990. This GHG inventory is submitted each year to the United Nations in accordance with the UNFCCC. A novelty for the 2002 U.S. EIO-LCA is the inclusion of emissions data for HFCs and PFCs, as well as the distinction between CO2 emissions from fossil fuel combustion and process emissions (eiolca.net). HFC and PFC emissions are combined together in one category given their extremely low emission levels (Figure 7). Data on methane and nitrous oxide emissions are also included. 31  Figure 7: U.S. Greenhouse Gas Emissions by Gas  (Source: U.S. EPA: http://www.epa.gov/climatechange/emissions/usgginventory.html)  4.3  Data Collection Procedures for Import and Export Data  As indicated earlier, I used Industry Canada‟s Trade Data Online database to get Canada‟s import and export data on merchandise trade for the year 2002. Since the classification standard used in the Green Design Institute‟s EIO-LCA software is the North American Industry Classification Standard (NAICS), I chose to retrieve Canada‟s trade data by NAICS codes, rather than by Harmonized System codes, 28 which are part of a universally accepted classification system for goods so countries can administer customs programs and collect compatible trade data.  NAICS is a comprehensive industry classification system encompassing all economic activities in which businesses are engaged. It was jointly adopted in 1997 by Canada, Mexico and the United States, not too long after the January 1, 1994 implementation of the North American Free Trade Agreement. Its goal was to provide common definitions for the North American industrial structure and a common statistical framework to facilitate comparative analyses of the three countries. The NAICS system is hierarchically structured and divides the economy into 20 sectors at the highest level of aggregation (two-digit codes). Then, those sectors are further disaggregated into subsectors (three-digit codes), industry groups (four-digit codes) and 28  The Harmonized Commodity Description and Coding System (HS) of tariff nomenclature is an internationally standardized system of names and numbers for classifying traded products developed and maintained by the World Customs Organization. Given its focus on tariffs, the HS codes only apply to goods but not to services. By contrast, the NAICS industry classification encompasses all economic activities in which businesses are engages, related to both goods and services.  32  industries (five-digit codes). These classification levels are generally comparable across the three countries, with some exceptions. Subsequently, each country may choose to further breakdown industries into national industries (six-digit codes) to capture additional detail. 4.3.1  Single-Country Approach  As mentioned previously, when tracking GHG emissions from various industrial production processes, the more disaggregated the industrial data are the more accurate the estimation of emissions will be. However, the economic and GHG emissions data that can be obtained from the Canadian EIO-LCA model is only available for the 105 sectors included in the model. For the most part, each of these sectors corresponds to a specific NAICS subsector (three-digits) or industry group (four-digits); in some cases, a sector includes a few subsectors or industry groups (Statistics Canada, 2005). Finally, the beverage manufacturing industry group (NAICS 3121) is the only one for which EIO-LCA data comes at a higher level of disaggregation (i.e., at the industry level (five-digits)). All the agricultural, resource and manufacturing sectors considered in my analysis are shown in Appendix B, along with their respective NAICS codes. For each of those 55 sectors, I retrieved Canada‟s import and export data at the corresponding three-, four- or five-digit level for the year 2002, in Canadian 2002 dollars. Since the monetary unit used in the 2002 Canadian EIO-LCA model is also the 2002 Canadian dollar, no monetary conversion is required for this stage of the analysis.  Once I collected all the trade data from the TDO database, I subsequently entered manually into the Canadian EIO-LCA model the dollar amount of imports and exports corresponding to each of the sectors in Appendix B as the desired output y for the EIO-LCA model as per Step 1 in Figure 1 (p. 18). The model identifies all the supply chain requirements x for the production of that output, also in dollar amounts, and the software computes simultaneously the GHG emissions b resulting from those imports and exports and their entire supply chain using the vector Rex that is included in the model. It is at this stage that physical quantities are introduced in the analysis, when sectoral emission intensity factors, measured in thousands of metric tons of CO2e per dollar of output, are used. Appendix C provides a more in-depth demonstration of the working of this model, with snapshots of the output from the EIO-LCA software for motor vehicles trade.  33  When using the EIO-LCA software, one must make sure that the dollar amount to be used as an input into the model is the price paid to producers in current dollars (producer prices) rather than the price paid by consumers (Hendrickson et al., 2006). If there are significant transportation costs or retail mark-ups on goods expenditures, there would be a significant difference between the producer price and the consumer price. For this analysis, I used Statistics Canada‟s import and export data for 2002, in current dollars, from Industry Canada‟s TDO service. The TDO‟s explanatory notes regarding the valuation of imports into Canada indicate that Canadian imports are valued “F.O.B. (free on board), place of direct shipment to Canada.” Thus, the freight and insurance costs incurred in bringing the goods to Canada from the point of direct shipment are excluded. Values are determined using the General Agreement on Tariffs and Trade valuation principles, which, in general, reflect the transaction value or price paid between unrelated sellers and buyers. With regards to export values, the explanatory notes indicate that: 1) Canadian exports to the U.S. are collected by the U.S. as import data from Canada, converted to Canadian dollars using an average monthly rate provided by the Bank of Canada and sent to Statistics Canada for publication as Canadian export;29 and 2) Canadian exports to countries other than the U.S. are recorded at the values declared on export documents which usually reflect the transaction value (e.g., actual selling price). From these explanations, it follows that international transportation costs are excluded from the value of imports and exports (along with insurance costs) and that any mark-ups would be added after the goods have been imported into Canada, when they are sold at retail (and similarly for exports). Therefore, my estimates should not have been inflated through the misuse of prices. 4.3.2  Multi-Country Approach (Linked Single-Country IO Model)  Under the multi-country approach, I assessed the GHG embodied in Canada‟s imports from the United States using the EIO-LCA model for the U.S. Working with both countries‟ data was facilitated by the fact that environmental and economic datasets in both countries are reported by NAICS conventions, enabling broad data comparability across sectors between the countries. The two important issues at this stage are reconciling the different levels of sector aggregation between the two models and converting into U.S. currency the import values in order to be able to input them into the U.S. EIO-LCA model.  29  This method of collecting Canadian exports to the U.S. by using U.S.‟s imports from Canada is more accurate because Canadian exports to the U.S. may otherwise include goods that are only transhipped through the U.S. and are destined to other countries. U.S. imports from Canada are also collected on an F.O.B. basis. In general, countries have more incentives to accurately track their imports rather than their exports for customs collection purposes.  34  As explained in Section 4.2.2, the U.S. EIO-LCA model is based on the 428 x 428 commodityby-commodity input-output matrix, which has a higher level of disaggregation than the Canadian model. Consequently, the 55 sectors included in my analysis (Appendix B) correspond to 308 sectors in the U.S. model (Appendix D). For the most part, each of the sectors in the U.S. model corresponds to a specific NAICS industry (five-digits) or U.S. national industry (six-digits); in some cases, a sector corresponds to a grouping of a few industries (five-digit) or U.S. national industries (www.eiolca.net). I took this difference in aggregation level between the two models into account when retrieving data from the TDO database on Canadian imports originating from the United States. Such data also need to be retrieved with a higher level of disaggregation than was needed to input into the Canadian EIO-LCA. In fact, the level of disaggregation of the import data must match that of the U.S. EIO-LCA to obtain estimates of the GHG embodied in the production of those imported U.S. goods.  When data on the Canadian imports from the U.S. needed to be retrieved at the five-digit level and also were available at that level of detail in the TDO database, the data manipulation was straightforward. The values of Canadian imports from the U.S. were obtained from the TDO and then input into the U.S. EIO-LCA model, after conversion into U.S. currency, to get the estimates of the GHG emitted during their production. However, when data needed to be input into the U.S. EIO-LCA model at the six-digit level to get those GHG estimates,30 or when the required detail level was simply not available from the TDO (even at the five-digit level), the data manipulation involved one more step. In those cases, I retrieved the data on Canadian imports from the U.S. at the next available level in the NAICS hierarchy (going upwards in the hierarchy), which was either the fourth or fifth level. Then, the value of imports was allocated evenly across the number of NAICS codes in the lower hierarchical level to get estimates of imports at those lowest NAICS levels. This is further illustrated in Appendix C. These import estimates were subsequently input into the U.S. EIO-LCA model, after conversion into U.S. currency, to get their corresponding embodied GHG estimates. This procedure resulted in the loss of more accurate GHG estimates at the highest disaggregation level, which were impossible to obtain given the lack of import data at the corresponding level, and is equivalent to having used the average GHG intensity of the more detailed industries.  30  This is because the NAICS six-digit codes correspond to national industries to capture additional specificity; therefore, they are not comparable across the three North American countries, even though the same six-digit code may exist in two or more countries.  35  The second issue revolves around monetary conversion. Because the 2002 U.S. EIO-LCA model uses 2002 U.S. dollars as its monetary unit, I needed to convert the CAD value of the 2002 Canadian imports from the U.S. into 2002 U.S. dollars. From the outset, it is obvious that this is an issue of critical importance as the exchange rate used in the analysis has a significant impact on the monetary values entered into the U.S. EIO-LCA model, which then has a commensurate effect on the corresponding embodied GHG estimates. Some researchers in this field use the market exchange rate (MER) to convert their data (Weber and Matthews, 2007; Ahmed and Wyckoff, 2003) while others favour using purchasing power parity (PPP) conversion rates (Norman et al., 2007) or suggest its use (Ahmed and Wyckoff, 2003; Peters and Hertwich, 2004).  When MER is used, unusual currency strength will falsely suggest relatively low emissions intensity. Some researchers therefore prefer using PPP exchange rates. Those rates equalize the purchasing power of different currencies in their home countries for a given basket of goods and services, thereby eliminating the differences in price levels between countries in the process of conversion. The OECD notes that PPP-based comparisons make more economic sense in specific situations, such as when comparing output levels between countries, while exchange rate comparisons are more appropriate in others.31  In this research, I used both the MER and the PPP exchange rates to convert the value of Canada‟s imports in U.S. currency because I wanted to explore the extent to which an analysis of this type hinges on fluctuations of exchange rates affecting monetary values. Again, Appendix C illustrates the rate conversion procedures discussed here. 4.4  Assumptions and Limitations  4.4.1  Production Processes of Other Countries  Sub-section 4.2 described one of the key assumptions analysts often make respecting the production methods in other countries: that commodities have the same GHG emissions intensity whether they are produced in the country under study or elsewhere. This assumption is prompted by the extreme intricacies of the production-commercialisation-consumption chains. A few researchers have modified their method to deal with this issue, but others have rationalized 31  PPP Frequently Asked Questions webpage on the OECD website.  36  their use of this assumption by indicating – generally – that it is likely to yield conservative results. The use of GHG emission intensities in industrialised countries may in fact underestimate the actual emissions in developing countries given the likely use of cleaner technologies in the former countries. Obviously, this is a broad generalization and one would have to look more carefully into the case at hand to see where imports really come from, and what production processes were used, to better qualify the impacts of such assumption. In Section 5, I further explore this issue. 4.4.2  Other Assumptions and Limitations  It is probable that my model results in a downward bias of the estimates of „embodied emissions‟ in trade goods since emissions related to fuel use in the international transport of goods are ignored.  The IO model operates under the assumption of constant returns to scale, the impact of which is uncertain since there is no „law‟ in economics to predict whether a firm has constant, decreasing or increasing returns to scale (Bjorn et al., 2005). 32 It is worth noting that the notion of returns to scale is different from the notion of economies of scale. Where the latter refers to a firm‟s costs, returns to scale describe the relationship between inputs and outputs in a long-run production function. The relationship between returns to scale and (dis)economies of scale is dependent upon the competition conditions in the input markets.  A final word of caution: in an era of ever-increasing trade liberalization, attempts at accounting for total supply chain effects are greatly complicated by the predominance of internationally supplied resources and products, including at the intermediate inputs level. Thus, it is not clear the extent to which even the most detailed multi-country analyses can adequately assess the emissions embodied in trade flows.  32  Constant returns to scale imply that increasing the output from any sector requires a proportional increase in each input received by that sector from all other sectors. Also, if a 10% increase in all inputs yields more than a 10% increase in output, we are in presence of increasing returns to scale. Conversely, if it yields less than a 10% increase in output, we are in presence of decreasing returns to scale.  37  5. Results My results are summarized in Table 6 below. As expected, the various methods I used to calculate Canada‟s BEET in 2002 gave rise to different estimates for the BEET because each method produced a different estimate of the amount of GHG emissions embodied in Canada‟s imports (EEI). For ease of reference, Equation 7 is copied below.  (7)  The first column shows the results for the single-country approach, where I only used the Canadian EIO-LCA model to quantify the GHG emitted in connection with the production of goods exported from and imported into Canada. In this case, I assumed that production processes in Canada and in the rest of the world share the same emission intensity. The other columns in Table 6 present the estimates resulting from a multi-country approach in which Canada and the United States are the two countries included in the model, and imports other than from the U.S. are assumed to be similar in GHG intensity as goods produced in Canada. As previously explained in Section 4, I first converted the value of Canada‟s imports from the U.S. into U.S. dollars using the MER (column 2) and I then used the PPP to do the conversion once more (column 3). Finally, the fourth column displays the results of the multi-country analysis using PPP for a subset of GHG, namely CO2. Table 6: GHG Embodied in Canadian Merchandise Trade, 2002 Column 1 Single-Country BEET_1 367.7 338.3 29.4 574.4 368.0  Column 2 Using MER BEET_2 367.7 134.2 127.6  Total Exports (billion CAD) Total Imports (billion CAD) Imports from U.S. (billion USD) Imports from ROW (billion CAD) Trade Balance (Ex - Im) (billion CAD) GHG embodied in exports (EEE, MMT CO2e) GHG embodied in imports (EEI, MMT CO2e)a GHG embodied in imports from U.S. (MMT CO2e) GHG embodied in imports from ROW (MMT CO2e) CO2 embodied in exports (MMT) CO2 embodied in imports (MMT) BEET (EEI - EEE) -206.3 a The EEI of columns 2 and 3 is the sum of the next two lines below.  Column 3 Column 4 Multi-Country Using PPP Using PPP BEET_3 BEET_4 367.7 367.7 177.1 177.1 127.6 127.6  574.4 262.6 128.4 134.2  574.4 303.6 169.4 134.2  -311.8  -270.8  462.0 246.3 -215.7  38  5.1  Canada’s Negative Balance of Embodied Emission in Trade  My findings indicate that, regardless of the method used to calculate the BEET, the GHG embodied in Canada‟s exports (EEE) are significantly higher than those embodied in its imports, always yielding a negative BEET. These results imply that a significant portion of Canada‟s GHG emissions can be linked to foreign consumption of Canadian goods, while a lower but still significant GHG amount can be attributed to the consumption of imported goods by Canadians. The fact that the BEET is negative is not surprising and likely reflects a variety of factors, including an overall trade surplus in 2002, large exports of natural resources and an industrial structure oriented towards the production and export of carbon-intensive goods. Furthermore, my results are consistent with those of other recent empirical studies that also found a negative BEET for Canada (Ahmad and Wyckoff, 2003; Peters and Hertwich, 2008). In fact, while Peters and Hertwich (2008) found that, as a group, the 35 Annex B countries they were studying were BEET positive, i.e., they were net importers of CO2 emissions, nine of them were BEET negative, meaning they had more emissions embodied in their exports than in their imports. These nine countries, including Canada, Australia, Poland and the Russian Federation, shared one characteristic: their large exports of natural resources. The Canadian data presented in Appendix E confirm that this is indeed the case for that country.  Before delving into a more detailed analysis of my results by sectors of the Canadian economy, I would like to further reflect on a key result so far, which is the sensitivity of my BEET estimates to the estimation method used. Indeed, when looking at the first three BEET estimates (columns 1 to 3 in Table 6), we see that there are marked differences among them for 2002. However, in the real world, only a set amount of GHG was released into the atmosphere. The question is, which of these estimates is closer to reality?  Using the single-country approach, I had to assume that each category of goods is characterised by the same GHG emission intensity whether the goods are produced in Canada or elsewhere. This means that various production processes, energy efficiencies and fuels, which may combine differently around the world to produce the same good, would result in the same GHG intensity regardless of country of origin. Clearly, when one is concerned about the actual amount of GHG emitted from production, this assumption can bias the results. It is thus important to explore the impacts of this assumption on my first series of estimates. Figure 8 breaks down Canada‟s imports among its top five trading partners in 2002: 63 percent of 39  Canada‟s imports originated from the U.S., followed by China, Japan and Mexico, each with a 4percent share, and the United Kingdom at 3 percent. The rest of the world captured the remaining 22 percent of Canada‟s import market. Figure 8: Canada’s Top Trading Partners, Imports by Dollar Value 2002  3%  United States (U.S.)  22%  4%  China  4%  Japan  4%  Mexico 63%  United Kingdom (U.K.) OTHERS  (Source: Industry Canada, Trade Data Online)  Given that the lion‟s share of Canada‟s imports in 2002 is from the U.S., it is worthwile to ask how the GHG intensity of U.S. goods compares to that of Canadian goods. Norman et al. (2007) have shown that „overall, the U.S. fuel mix is found to be almost 1.3 times as GHG intensive as Canada‟s fuel mix‟.33 Differences in fuel sources for energy supply appear to be the most significant factors influencing the variability in GHG intensity between Canada and the U.S. With few exceptions, the authors found that the U.S. manufacturing sectors exhibit higher GHG intensity than their corresponding Canadian counterparts, a discrepancy that appears to stem mainly from differences between electricity sources. The U.S. generates more than threequarters of its electricity from fossil fuels (primarily coal and natural gas), while Canada generates almost two-thirds of it from hydroelectric sources. Electricity generation in the U.S. is therefore twice as GHG-intensive as it is in Canada, which has significant repercussions throughout the manufacturing sectors of the economy (Norman et al., 2007). Therefore, Canadian exports of manufacturing goods to the U.S. should be less carbon intensive that imports of manufacturing goods from the U.S.  China had one of the highest GHG intensities in the world in 2005 at 1.40 MMT of CO2e per billion U.S. dollars of GDP, compared to 0.59 for the U.S. and 0.78 for the world average (CRS 33  The authors‟ estimates relate to the year 1997.  40  Report for Congress, 2008). This difference is in part explained by the fact that coal provides 70 percent of China‟s total energy, followed by petroleum at 20 percent and gas at 3 percent. Hydroelectric and nuclear sources contribute the remaining 7 percent. Considering that over two-thirds of Canada‟s imports came from the U.S. or China in 2002, it is fair to conclude that my first BEET estimate likely underestimates the amount of GHG embodied in Canada‟s imports. Even though it is not possible to assess the magnitude of this downward bias without a more detailed analysis, it is likely that, given its current magnitude, the BEET would remain negative even if the emissions embodied in imports (EEI) were correctly measured.  To address this problem, I re-estimated the emissions embodied in imports using a two-country approach in which I used the proper U.S. GHG intensity to estimate the emissions associated with the 63 percent of goods imported from the U.S. As noted above, I expected that doing so would have resulted in a higher EEI; however, as Table 6 shows, this is not the case. The main factor causing this result seems to be the Canada-U.S. exchange rate and the historical weakness of the Canadian dollar in 2002 (Figure 9). To use the U.S. EIO-LCA model to estimate the GHG embodied in U.S. manufactured goods imported into Canada, I first had to convert the dollar value of Canadian imports from the U.S. from Canadian to U.S. currency, as illustrated in Appendix C.  Figure 9: Annual Average Exchange Rates of the Canadian dollar to the U.S. dollar 1.00 0.90 0.80 US dollars  0.70 0.60 0.50  1 CAD CDN Dollar  0.40 0.30 0.20 0.10  2009  2008  2007  2006  2005  2004  2003  2002  2001  2000  1999  1998  1997  0.00  (Source: Bank of Canada, Monthly and Annual Average Rates, www.bankofcanada.ca)  41  Under a market-exchange-rate (MER) approach, a strong U.S. currency relative to the Canadian dollar results in apparently lower GHG emissions associated with the production of U.S. goods even though in reality the GHG emission intensity of U.S. goods is not affected by exchange rates. A strong U.S. currency was indeed the case in 2002: when converting the value of Canada‟s Imports from the U.S. from the 2002 Canadian dollar (CAD) into the 2002 U.S. dollar (USD), I multiplied each import value by the annual average exchange rate,34 which was 1.000 CAD for 0.637 USD in 2002 and, as a result, the “numerical” value of Canada‟s imports from the U.S. was reduced by 36.3 percent from CAD 210.7 billion to USD 134.2 billion. Therefore, the use of the 2002 average MER made the value size of the imports appear much smaller than if the Canadian dollar had been stronger, as in more recent years, when in fact, the physical volume of goods crossing the border would have been the same. Clearly, this type of monetary conversion has a considerable impact on the EEI estimates obtained from the U.S. EIO-LCA model. In fact, this conversion problem could not have been more forcefully illustrated than with my selected year, since 2002 witnessed a historical low of the Canadian-U.S. exchange rate. If the same analysis had been performed using the 2008 annual average exchange rate (1.000 CAD = 0.938 USD), when the Canadian dollar was the strongest relative to the U.S. currency, the import values would have been multiplied by a much higher fraction and the resultant EEI estimates would have been proportionately higher.  Therefore, instead of my estimate of the EEI being more accurate and higher with this second method, to reflect a higher U.S. GHG intensity, I must conclude that it is probably not more accurate than with the single-country model, given the heavy reliance on market exchange rate fluctuations, and certainly not higher, as the impact of the monetary conversion trumped that of the GHG-intensity differential.  Exchange rate fluctuations, which are driven more by variables like interest rates differentials between countries and financial flows rather than purchasing power, thus appear to inflate or deflate countries‟ trade flows even though in reality there has been no change in the actual volumes of goods traded. To counter this problem, I used the PPP exchange rate, the advantage of which was explained in greater detail in Section 4.3.2. In contrast to MER, PPP exchange rates oscillate much less over time than market exchange rates (Figure 10): between  34  The Bank of Canada exchange rates are used in this analysis. Bank of Canada exchange rates are nominal quotations – not buying or selling rates – and are intended for statistical and analytical purposes. Rates available from other financial institutions will differ.  42  1997 and 2004,35 the Canadian dollar stayed within a two-cent variation using the PPP while it fluctuated by more than 13 cents using the MER. Figure 10: PPP Exchange Rates of the Canadian Dollar to the U.S. Dollar 1.00 0.90 0.80 US Dollars  0.70 0.60 0.50 0.40  1 CAD  0.30 0.20 0.10  2004  2003  2002  2001  2000  1999  1998  1997  1996  1995  1994  1993  1992  0.00  (Source: OECD at http://www.oecd.org/std/ppp/)  For the third series of estimates, I multiplied each import value by the PPP exchange rate, which was 1.000 CAD for 0.836 USD in 2002 and, as a result, the numerical value of Canada‟s imports from the U.S. was only reduced by 16.4 percent (instead of 36.3 percent in the previous scenario) from CAD 210.7 billion to USD 177.1 billion. Consequently, with this method, the GHG embodied in the U.S. goods bound for Canada were estimated at 169.4 MMT CO2e, which is markedly higher than the 128.4 MMT CO2e estimated using the MER. This specific result is important in itself as it shows how a strong U.S. currency would seemingly result in lower U.S. GHG emissions, when in fact the two are completely unrelated in the physical sphere. However, even when using the PPP, the monetary conversion again trumps the GHG intensity differential so that my third BEET estimate is still greater, in absolute value, than my first BEET estimate (see Table 6).  This analysis shows the weakness of using monetary flows when trying to estimate physical quantities. A far better approach would be to use the physical quantities of traded goods and their known carbon content, which is an area where future research could focus. The use of Industry Canada‟s Trade Data Online database limited my estimates of import data to Canadian 35  The 1997-2004 period was chosen for comparison purposes between the two exchange rate series.  43  currency values, not physical volume. As for the EIO-LCA software, it only allowed me to estimate GHG emissions through the use of monetary flows, since the GHG emission intensity in that software is estimated in thousands of tonnes of CO2e per million dollar of output. Finally, the fourth column in Table 6 displays the results of the multi-country analysis, but only looking at a subset of GHGs, namely CO2, using the PPP conversion method. The BEET went from -270.8 MMT CO2e (column 3) to -215.7 MMT CO2e, indicating that CO2 accounts for 80 percent of the overall BEET, a result that is consistent with the fact that CO2 contributed a 79percent share of Canada‟s emissions in 2003, while CH4 and N2O accounted respectively for 13.0 and 6.7 percent of the emissions (Environment Canada, 2005). This result suggests that other studies that focus exclusively on carbon dioxide tend to considerably underestimate the absolute value of the balance of GHG emissions embodied in trade flows.  This comparative analysis of the different methods for estimating the BEET highlights the importance of being cautious when interpreting the results of this type of analysis. Also, while such analysis led me to think that the third estimate might be closer to reality, it would be highly difficult to conclude by how much. 5.2  Canada’s Balance of Embodied Emissions in Trade by Industrial Sector  As stated at the beginning of this chapter, Canada‟s negative BEET likely reflects a variety of factors such as a trade surplus in 2002, large exports of natural resources and an industrial structure oriented towards the production and export of carbon-intensive goods. Here I examine my results in a more disaggregated fashion.  Table 7 breaks down my three BEET estimates by NAICS sectors, along with presenting those sectors‟ trade balance in 2002. The results show that all sectors display a negative BEET, meaning that the CO2e content of Canadian exports is greater than that of imports. It is worth noting that Canada maintained a trade surplus in all sectors but Manufacturing, and correspondingly, these sectors display a negative BEET, suggesting a negative correlation between trade balances and balances of embodied emissions in trade. Taken as a whole, the manufacturing sector seems to be an exception; however, looking only at this sector‟s trade balance does not depict the whole picture because it hides the magnitude of the two-way trade flows (Table 8). 44  Table 7: GHG Embodied in Canadian Merchandise Trade, by NAICS Sectors, 2002 Sectors Sector Description (NAICS) Agriculture, Forestry, Fishing and 11 Hunting Mining, Quarrying, and Oil & Gas 21 Extraction 22 Utilities 31-33 Manufacturing TOTAL  Trade Balance BEET_1 BEET_2 (Ex-Im) M CAD (Mt CO2e) (Mt CO2e)  BEET_3 (Mt CO2e)  5,107.2  -38.4  -66.0  -63.6  31,182.7 1,304.6 -8,200.4 29,394.1  -63.3 -6.9 -97.8 -206.3  -62.9 -6.6 -176.3 -311.8  -61.5 -5.7 -140.0 -270.8  Table 8: Canada's Trade Data, by NAICS Sectors, 2002 Sectors (NAICS) 11 21 22 31-33  Sector Description Agriculture, Forestry, Fishing and Hunting Mining, Quarrying, and Oil & Gas Extraction Utilities Manufacturing TOTAL  Exports (M CAD)  Imports (M CAD)  Trade Balance (Ex - Im) M CAD  13,017.3  7,910.1  5,107.2  47,447.5 1,814.8 305,371.1 367,650.7  16,264.8 510.2 313,571.5 338,256.6  31,182.7 1,304.6 -8,200.4 29,394.1  Two facts become apparent when we look at Table 8: 1) the dollar value exports and imports of manufacturing goods are significantly larger than those of the other sectors; and 2) these trade flows are almost of the same magnitude. This means that a trade deficit of CAD 8.2 billion, when compared to exports and imports of CAD 305.4 and CAD 313.6 billion respectively, is relatively insignificant, therefore not necessarily contradicting my earlier finding of a negative correlation between trade surpluses and negative BEET. In fact, when the relatively small manufacturing trade deficit is compared with the very large negative BEET of that sector, this result does seem to confirm that the Canadian industrial structure is oriented towards the production and export of carbon-intensive goods. Indeed, it shows that despite a small trade deficit, a lot more GHG are embodied in Canadian exports than in imports of manufacturing goods. A more detailed analysis of this finding is presented below.  Appendix E presents my results at the more disaggregated industry-group level for the multicountry model using PPP. I chose that model for the presentation of these detailed results because I think it resulted in estimates that are closer to reality. A quick glance at the table in Appendix E shows that for 80 percent of the industry groups (four-digit codes), a trade surplus is associated with a negative BEET. In the case of the manufacturing sector, this inverse 45  relationship holds for 77 percent of the industry groups. In particular, when I calculated the correlation coefficient between the „trade balance‟ and „BEET‟ data sets, I found that the correlation coefficient is ρ1 = -0.77 for all industry groups and ρ2 = -0.80 for manufacturing only. This confirms the existence of a strong negative correlation between the two variables.  Table 9 offers a closer look at the few industry groups that contribute the most to the overall negative Canadian BEET, which for the most part relate to natural resources and carbonintensive goods. Table 9: Trade Balance and BEET for Select Industry Groups Sectors EIO-LCA  Description  3361 11A0 2111 3221 3310 3241 3210  Motor vehicle manufacturing Crop and animal production Oil and gas extraction Pulp, paper and paperboard mills Primary metal manufacturing Petroleum and coal products manufacturing Wood product manufacturing  Trade Balance (ex-im) M CAD 28,973.80 4,811.30 23,793.70 17,152.20 6,979.50 7,097.90 15,654.30  BEET_3 MMT CO2e -93.01 -63.80 -59.54 -26.95 -25.59 -19.74 -17.74  Not only did these industry groups all have trade surpluses in 2002, they were also among the largest surpluses. Thus, these results support the hypothesis that a negative BEET likely reflects a variety of factors, including a trade surplus and large exports of natural resources. Also, the fact that the largest provincial contributors to Canada‟s GHG were Alberta and Ontario and that the provinces with the highest GHG intensity factors were Alberta and Saskatchewan (Environment Canada, 2005)36 seems consistent with the result above: motor vehicles are predominantly manufactured in Ontario while both crop and animal production, and oil and gas extraction are most prevalent in the Prairies. Therefore, while in general, Canadian manufacturers are less GHG intensive than their U.S. and Chinese counterparts, this research showed that Canada has net exports of goods that are particularly heavy in GHG emissions.  36  In 2003.  46  6. Discussion of Policy Implications This section briefly discusses the potential policy implications of my research‟s findings and thus addresses my second research objective. 6.1  Policy Implications of a Negative BEET  As discussed in-depth in Section 5, my results show that Canada has a negative balance of embodied emissions in trade, even though there is uncertainty surrounding the exact value of the BEET. A negative BEET implies that the GHGs emitted in connection with the production of exported goods surpass those emitted in connection with the production of imported goods. My findings further show that a trade surplus is most often associated with a negative BEET, a relationship that holds for about 80 percent of the industry groups included in the analysis. This finding supports the hypothesis put forward by Munksgaard and Pedersen (2001) to the effect that there may be an inherent conflict between a national GHG reduction target for domestic emissions and the aim of improving trade balances or maintaining trade surpluses. As a consequence, it may be more difficult to fulfill national GHG reduction commitments if a significant portion of them is caused by foreign demand. Ironically, the recent reversal of Canada‟s trade surpluses into deficits may have improved the odds that Canada meets its Kyoto target before the end of the first commitment period. Indeed, it would be interesting to see how Canada‟s BEET has evolved in the past three years as Canada‟s long-standing trade surpluses converted into trade deficits. Canada‟s exports have decreased sharply in 2009 (exports contracted by over 26 percent in 2009 over 2008) and somewhat rebounded in 2010, although not to the pre-2009 levels, while imports also decreased in 2009 but less sharply (a 16 percent decrease) and also rebounded in 2010. This reversal resulted mainly from a strong Canadian dollar relative to the U.S. dollar, and the U.S. housing and financial crises that prompted a recession in the U.S. starting in 2008. A more detailed analysis of these new circumstances would probably reveal that Canada has recently become a net importer of CO2 as the sharp drop in exports would have resulted in an equally significant reduction of embodied emissions in exports.  47  The conflict identified above could have implications for future multilateral negotiations on GHG reduction strategies, which might call for a reliable methodology for assessing GHGs embodied in international trade. It may be more likely that countries with net GHG exports would push this issue in the hope of ensuring the negotiation of more equitable reduction targets, especially if trade surpluses had a significant influence on the level of national emissions in the selected base year. However, if the calculation of GHG emissions were to be based on consumption and not production, the issue of being able to rely on a robust consumption-based calculation methodology becomes critical.  As my results illustrate, there is an inherent weakness in using monetary flows when trying to estimate physical quantities. Therefore, what is really needed is a shift from monetary to physical flows that would be enabled by a shift to physical input-output tables. These would greatly enhance the estimates of GHG emissions embodied in goods as the carbon content would be calculated based on units of mass, which do not fluctuate with prices and market exchange rates. Still, a word of caution is necessary. With ever-increasing trading between countries, attempts at accounting for total supply chain effects are greatly complicated by the predominance of internationally supplied resources and products, including at the intermediate inputs level. Thus, it is not clear the extent to which even the most detailed multi-country analyses can adequately assess the emissions embodied in trade flows. 6.2  Intersections of Climate and Trade Policy  But until such time as a reliable methodology is developed to calculate those emissions, the question remains regarding how a country can take responsibility for the GHG emissions produced in connection with the goods it imports. Indeed, even though Canada‟s BEET is negative, if one looks only at the amount of GHGs embodied in Canada‟s imports, one finds that these emissions range from 262.6 MMT to 368.0 MMT CO2e in 2002 depending on the method used to estimate them with about half of them related to goods imported from the U.S. (see Table 6). Those emissions are quite large when compared to the 717 MMT of CO2e emissions reported by Canada to the UNFCCC in 2002. Since one tonne of CO2e has the same impact on the climate wherever it is emitted, taking account of CO2e emitted elsewhere but caused by Canadian consumption should be both an ethical duty and in the self-interest of Canadians who will bear the impact of climate change whether the GHGs are emitted in their own backyard or  48  not. I discuss in turn border tax adjustments (BTA), low carbon fuel standards and climate labelling as various policy instruments able to link consumption and embodied GHGs. 6.2.1  Border Tax Adjustments  One avenue that has been suggested early by Wyckoff and Roop (1994) is to impose BTAs on imports that would complement any domestic measures raising the price of carbon. Such combination of measures would raise the price of carbon-rich goods, whether produced at home or abroad, would encourage consumers to buy products that have low or no carbon content and would also theoretically reduce the potential for carbon leakage.  Energy-related BTAs would typically require importers to pay a tax according to the emissions associated with the imported products‟ production, at the same price as faced by domestic producers. This concept has been gaining ground in Europe as the EU is considering options in the absence of a major international agreement to cap GHG emissions (Fischer and Fox, 2009) and out of a concern over loss of productivity relative to the U.S., a key trading partner not party to the Kyoto Protocol (Biermann and Brohm, 2005; Brewer, 2008). The idea has even gained some traction more recently in the U.S., in proposed cap-and-trade legislation such as the Lieberman/Warner bill (S. 2191 “America‟s Climate Security Act”) and the Bingaman/Specter bill (S. 1766 “Low Carbon Economy Act”).37 These legislative proposals included widely debated measures to protect the competitiveness of domestic manufacturing. For instance, the Lieberman/Warner bill would have required importers to purchase GHG allowances at the border to compensate for the difference between the cost of production at home and the cost of production in unregulated jurisdictions. This measure would have been imposed based on country of origin – imports from countries that had not taken sufficient action on climate change would be subject to the border measure, with exception only for least developed countries and de minimis emitters (Fickling, 2010). The measure was a significant concern for Canada and Mexico, whose officials feared that their large volumes of carbon-intensive exports could be put at risk (Schott and Fickling, 2009). They also questioned its WTO legality.  BTAs are not limited to imports. As pointed out in Fischer and Fox (2009), there are conceptually many different unilateral policy options for dealing with the potential carbon leakage resulting from production cost differentials induced by the adoption in one country of 37  Those bills never became law.  49  stringent climate policy and not in the others. As described previously, import taxes (also sometimes referred to as carbon tariffs) level the playing field for domestic consumption, but do nothing abroad. For their part, border rebates for exports level the playing field abroad but still give imports a competitive advantage at home. Full border adjustment policies combine these two measures such that only the emissions arising from domestic consumption are taxed. Hence, if the consumption-based approach is not adopted as the basis to calculate national GHG inventory in future multilateral climate agreements, full border adjustment of domestic climate policies may be the alternative that is the closest to adopting a consumption-based approach to climate change mitigation.  However, many trade law experts have expressed concern that such border adjustment measures may not be compatible with World Trade Organization (WTO) obligations. There are a number of good reviews of the compatibility of GATT/WTO38 law with energy or GHG-related border tax adjustments. Pitschas (1994-1995), Charnovitz (2003) and Biermann and Brohm (2005) adopt mainly a legal perspective while Fischer and Fox (2009) adds an economic perspective. Potential legal hurdles aside, Fischer and Fox (2009) find that while all the different unilateral policy options discussed above have the potential to address the competitiveness issue and support domestic production, none is necessarily effective at reducing global emissions.  On the import side, the combination of treaty and case law is clear in allowing WTO members to apply BTAs, in a manner consistent with the Most Favoured Nation Treatment (MFN) and National Treatment principles,39 to indirect taxes40 (e.g., sales, excise or value-added taxes) levied on imported goods or on inputs physically incorporated into the imported goods (Pitschas, 1994-1995; Biermann and Brohm, 2005). However, what many describe as a murky area of trade law is whether WTO members could impose BTAs for energy taxes levied on the energy used as an input to the production process but which is not physically incorporated into them (Pitschas, 1994-1995; Charnovitz, 2003; Fischer and Fox, 2009). Pitschas (1994-1995) interprets trade law narrowly as meaning only taxes on the inputs physically incorporated into the products can be adjusted. Thus he concludes that since energy taxes are taxes on inputs  38  GATT stands for General Agreement on Tariffs and Trade. The Most Favoured Nation Treatment clause (GATT Article I) prohibits WTO members from discriminating among trading partners. The National Treatment clause (GATT Article III) requires that imported goods be treated no less favourably than “like” domestic products. 40 As opposed to direct taxes such as taxes on wages, profits, interests, rents, royalties and all other form of income and taxes on the ownership of real property. 39  50  not physically incorporated into the products, they cannot be adjusted at the border on the imported products. Fischer and Fox (2009) remain uncertain as to whether specific energy taxes are adjustable on the import side.  Based on a detailed examination of the legality of energy tax adjustments at the border for imported goods and for exported goods, Biermann and Brohm (2005) conclude that despite remaining ambiguity in both the legal provisions and the pertinent case law, BTAs are under certain circumstances compatible with world trade law but that given this persisting legal uncertainty, affected WTO members would likely challenge such energy BTAs before the WTO dispute settlement body (DSB). Nonetheless, they believe that if the EU embarked on a carefully crafted strategy of energy BTAs vis-à-vis non-European industrialized countries, there is a fair chance that it would prevail if these BTAs were challenged before the DSB. Charnovitz (2003) also concludes that whether process-based energy taxes to imported products can pass WTO muster will depend on how carefully they are written to avoid arbitrary discrimination.  Interestingly, Biermann and Brohm (2005) draw a distinction between BTAs vis-à-vis industrialized countries on the one hand and developing countries on the other, both on fairness and legal grounds.41 They suggest that, if justifiable under world trade law, BTAs might be used against industrialized countries that gain trade advantages through persistently lower energy prices owing to insufficient implementation of climate policies but should be avoided against developing countries where other policies to address carbon leakage could be used instead (e.g., financial and technological assistance). A similar distinction also appeared in relation to the GHG allowance measure contained in the Lieberman/Warner bill, although only applicable to least developed countries and de minimis emitters. Canada should at least retain a distinction similar to what was included in the U.S. bill if it were to craft a climate policy with a BTA on imports.  On the export side, while energy-related export rebates would not be construed as constituting an export subsidy42 under the Agreement on Subsidies and Countervailing Measures (SCM) (Pitschas, 1994-1995; Biermann and Brohm, 2005; Fischer and Fox, 2009), some ambiguity still persists on their legality. The GATT SCM Agreement initially specified that taxes on inputs to 41  On fairness grounds, Biermann and Brohm (2005) argue that certain advantages for developing countries are probably justified given the historic overuse of the atmosphere by industrialized nations and their higher-per-capita emissions. On legal grounds, such advantages seem justified, if not required, by the principle of common but differentiated responsibilities and capabilities in the UNFCCC. 42 Export subsidies are prohibited under the SCM Agreement.  51  production are border adjustable only when the goods are “physically incorporated” into the exported products. A revision in the Uruguay Round broadened the category of adjustable taxes to allow rebates for indirect taxes on goods and services if they are consumed in the production of the exported product (Fischer and Fox, 2009). In addition to physically incorporated inputs, export rebates are permitted on „energy, fuels and oil used in the production process‟ (GATT SCM Agreement, Annex II, Footnote 61). However, the U.S. government has been of the view that Footnote 61 to the SCM Agreement should not open the door to broad new border tax adjustments on energy and was intended solely as a technical adjustment for certain countryspecific approaches to taxation (de Coninck, 2008; Biermann and Brohm, 2005). The issue has not yet been clearly settled among legal experts (Fischer and Fox, 2005). The interpretation of this footnote as meaning that energy-related BTAs are permissible on exports is critically important for policies concerning energy or greenhouse gas emissions. Some experts have asked the question, to what degree this interpretation of BTAs on exports extends to imports. For instance, Biermann and Brohm (2005) argue that given the logical connection between BTAs on exports and imports, which are in effect two sides of the same coin, it would be defensible to apply Footnote 61 of Annex II of the SCM Agreement also mutatis mutandis to the interpretation of GATT regarding imports. 6.2.2  Low Carbon Fuel Standard  Low carbon fuel standards (LCFS) represent another policy instrument that jurisdictions can employ to take responsibility for the GHGs embodied in the fuel they consume. The main goal of a low carbon fuel standard is to reduce the carbon intensity of transportation fuels and decrease the CO2 emissions of fuel-powered vehicles in order to reduce the carbon footprint of transportation. The first worldwide LCFS, enacted by California in 2007 and intended to replace the state‟s ethanol blending requirement,43 calls for a reduction of at least 10 percent in the carbon intensity of California‟s transportation fuels by 2020.44 The LCFS required the California Air Resources Board to assign a carbon intensity reference value to each transport fuel sold in California. To do that, CARB quantified the GHGs emitted over the lifecycle of each fuel, from the extraction of the raw material used to produce the fuel to the burning of the fuel in one‟s car (Fickling, 2010). This should have the effect of truly reducing carbon emissions instead of promoting American corn growers. 43  California‟s ethanol blending requirement had been criticized for failing to discriminate between relatively GHG-intensive biofuels such as corn-based ethanol and lower-impact biofuels such as sugarcane and cellulosic ethanol (Flicking, 2010). 44 http://www.arb.ca.gov/fuels/lcfs/lcfs-background.htm  52  The California LCFS has been particularly controversial in Canada because it separates petroleum in two categories (conventional and unconventional) and uses a separate lifecycle analysis to assign the carbon intensity value for unconventional, such as oil sands crude. In light of Canada‟s large exports of oil sands crude to the U.S., it is not surprising that Canada expressed concern over the legality of the standard under WTO rules (Fickling, 2010). In order to win a WTO challenge, California would need to prove that environmental considerations require oil sands crude to be treated differently from conventional crude and Fickling (2010) suggests that a panel‟s decision could hinge upon whether oil sands crude is sufficiently more GHG-intensive to merit a separate category. 6.2.3  Climate Labelling  Climate labelling represents another climate policy alternative that countries can consider implementing to encourage consumers‟ individual responsibility. In order to make informed decisions, consumers need information about the ecological impact of production. Therefore, labelling could be a key instrument of climate policy if it turned out to be permissible under the WTO.  Charnovitz (2003) describes the legal uncertainty around climate labelling. Labels that describe a good‟s characteristics are unlikely to conflict with WTO rules. For instance, a label for new automobiles that would display information about fuel consumption and CO2 emissions would be consistent with GATT and the Technical Barriers to Trade (TBT) Agreement. By contrast, mandatory labels regarding the GHG emitted during the production process of a good could trigger a WTO-based challenge. How TBT obligations would apply to such a label is not settled in WTO law. Charnovitz (2003) reports that, because the TBT Agreement is limited to regulations and standards on product characteristics and their related process, many trade law experts had assumed that so-called unrelated process, such as the type and quantity of energy used in manufacturing, were beyond that Agreement. He notes that others argue that WTO law would almost certainly prohibit a government from requiring a label specifying the level of GHG emitted in the production process. As a case in point, when the Netherlands proposed to require a label identifying whether timber was harvested under sustainable forestry management and notified it to the WTO, several governments raised objections on the grounds that it would  53  violate trade rules. The measure was also criticized within the EU. As a result, the Dutch government, sadly, did not finalize the proposal. 6.3  Improving the Climate-Trade Intersections  While the GATT/WTO rules were clearly not designed to deal with the kind of issues that continue to arise from implementing measures to mitigate climate change, it now seems imperative that climate and trade experts aggressively pursue strategies to clarify areas of legal uncertainty and to remove inconsistencies between the two multilateral systems, such as the one identified above.  Regarding energy/GHG taxes, the optimal solution to avoid potential conflicts between WTO members would be to make the adjustment unnecessary in the first place. Charnovitz (2003) and Biermann and Brohm (2005) suggest that a coordinated approach to national energy/GHG taxes could be an effective way to control emissions without leading to inter-country distortions that in turn calls for BTAs. However, they recognize that this would require intense international co-operation and the prospects for a successful outcome are slim. Nonetheless, Charnovitz (2003) recommend that the climate regime could assume a greater responsibility for promoting a uniform approach to energy/GHG taxation particularly as it relates to imports and exports. This would prevent the development of a hodgepodge of energy/GHG taxes that will confound exporters and lead to trade disputes. In his view, it‟s best for the climate regime to take this issue in its own hands as the trade regime is unlikely to solve this problem.  WTO rules are also unclear with respect to climate labelling. From a trade perspective, assuring that labels do not impede trade through misinformation or unjustified interference is a valid concern. From a climate viewpoint, ensuring that labels can be used to inform the public about the embodied carbon in products to encourage consumers to take responsibility for their purchases is also a valid objective. Thus, Charnovitz (2003) suggest that the two regimes have a basis to work together to ensure that WTO law does not constrain well-designed climate labels.  Even if these inconsistencies are not removed and trade law is not clarified, Fischer and Fox (2009) note that future climate agreements could still provide for the imposition of energyrelated BTAs as long as Parties to the Agreement voluntarily agree to forego their WTO rights. It 54  is perhaps also worth noting that in practice, nothing prevents a country from imposing whatever measures it believes will benefit it as long as it is ready to accept the consequences of its actions. The reality is that countries have both rights and obligations within the WTO framework. Therefore, if Canada were to impose border tax adjustments on imports based on the GHG intensity of production processes, it would simply need to be ready for a WTO challenge from other affected WTO members, including the U.S. This is because these countries have „paid‟ to join the WTO by way of, among others, reducing or eliminating their tariffs on Canadian goods. Thus, it would be considered unfair trade on the part of Canada to unilaterally impose new taxes on their exports.  It is also worth noting that the Canadian case is unique given that Canada and the U.S. share the single largest trading relationship in the world, with a volume of traded goods unsurpassed by any other bilateral trade relation. Therefore, in the highly hypothetical event that Canada were to unilaterally adopt a national cap and trade system or a carbon tax to effectively reduce its GHG emissions (without the U.S. doing the same), to also contemplate imposing border tax adjustments on imports from the U.S. to have an effect equivalent to the domestic climate policies would seem suicidal given the magnitude of these imports and of the bilateral trade relationship itself .This would certainly result in a WTO challenge from the U.S., despite the fact that the U.S. themselves recently considered imposing such import adjustment measures. This is clearly not an avenue that would prove beneficial for the two countries. In this context, implementing an integrated North American cap and trade system that would cover substantially all sectors of the economy may seem to be the preferable alternative. Unfortunately, for the time being, this is off the political agenda in both countries. With regard to Canada‟s trade with developing nations, such as China, possible avenues to effectively take responsibility for GHGs emitted abroad include assisting these other countries in reducing their energy consumption through increased energy efficiency and developing clean coal technologies or renewable energy solutions that can substitute coal and other fossil fuels. Another idea would be to put a price on CO2 emissions embodied in Canadian imports and put this money into a climate fund providing risk capital to new companies focussed on developing low carbon solutions to be deployed both in Canada and these other jurisdictions. However, until there is a price on carbon in Canada or North America, it would be difficult to assess the price of the emissions embodied in Canada‟s imports. Effectively, until Canadians elect a government in Ottawa that cares about climate change and is committed to implement effective 55  measures to reduce GHG emissions, it seems unlikely that the government would want to willingly take responsibility for emissions in other countries caused by Canadian imports.  56  Bibliography Ahmad, N. and A. Wyckoff. 2003. Carbon dioxide emissions embodied in international trade of goods, STI Working Paper 2003/15, OECD, Paris. Biermann, F. and R. Brohm. 2005. Implementing the Kyoto Protocol without the USA: the strategies of energy tax adjustments at the border. Climate Policy. 4: 289-302. Bjorn, A., Declercq-Lopez, L., Spatari, S. and H.L. MacLean. 2005. Decision support for sustainable development using a Canadian economic input-output life-cycle assessment model. Canadian Journal of Civil Engineering. 32: 16-29. Brewer, T. 2008. U.S. Climate change policy and international trade policy intersections: issues needing innovation for a rapidly expanding agenda. Paper prepared for a seminar of the Center for Business and Public Policy, Georgetown University. Carnegie Mellon University Green Design Institute. 2008. Economic input-output life cycle assessment (EIO-LCA) model. <http://www.eiolca.net/> (accessed May 26, 2009). Charnovitz, S. 2003. Trade and climate: potential conflicts and synergies. Report for the Pew Center on Global Climate Change. Congressional Report Service (CRS) Report for Congress. 2008. China’s greenhouse gas emissions and mitigation policies. September 10. <http://www.fas.org/sgp/crs/row/RL34659.pdf> (accessed November 14, 2010). Danish, K.W. 2007. An overview of the international regime addressing climate change. Sustainable Development Law and Policy. 7(2): 10-15, 76-77. de Coninck, H., Fischer, C., Newell, R.G., and T. Ueno. 2008. International technology-oriented agreements to address climate change. Energy Policy. 36(1): 335-356. Department of Foreign Affairs and International Trade (DFAIT). International trade: fast facts on Canada-U.S. commercial relations <http://www.international.gc.ca/commerce/facts-infos/usa2008-eu.aspx> (accessed March 11, 2010). Environment Canada (EC). 2005. Canada’s greenhouse gas inventory: overview 1990-2003. October 2005. < http://www.ec.gc.ca/Publications/EBE51145-7C8C-49C8-A80B40896808CE7E/c68.pdf> (accessed May 7, 2011) Environment Canada (EC). 2009. National Inventory Report 1990-2007: Greenhouse Gas Sources and Sinks in Canada. April 2009. <http://www.ec.gc.ca/Publications/B77E6264-D0E345B5-BE565A395C0B7653/NationalInventoryReport19902007GreenhouseGasSourcesAndSinksInCanada. pdf> (accessed October 30, 2011). Environment Canada (EC). 2011. A Climate Change Plan for the Purposes of the Kyoto Protocol Implementation Act. May 2011. <http://www.climatechange.gc.ca/Content/4/0/4/4044AEA7-3ED0-4897-A73E-  57  D11C62D954FD/COM1410_KPIA%202011_e%20-%20May%2031%20v2.pdf> (accessed October 30, 2011). Fickling, M. 2010. North America‟s uphill battle on climate change and its implications for the North American trading system. CESifo Forum. 4/2010: 45-51. Fischer, C. And A.K. Fox. 2007. Output-based allocation of emissions permits for mitigating tax and trade interactions. Land Economics. 83(4): 575-599. Fischer, C. And A.K. Fox. 2009. Comparing policies to combat emissions leakage. Washington, DC: Resources for the Future. Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland, 2007: Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Ghertner, D.A., and M. Fripp. 2007. Trading away damage: quantifying environmental leakage through consumption-based, life-cycle analysis. Ecological Economics. 63: 563-577. Gore, A. 2009. Our Choice. A Plan to Solve the Climate Crisis. Melcher Media, New York. Harrison, K. 2007. The road not taken: climate change policy in Canada and the United States. Global Environmental Politics. 7(4): 92-108. Harrison, K. and L.M. Sundstrom. 2007. The comparative politics of climate change. Global Environmental Politics. 7(4): 1-18. Hendrickson, C.T., Horvath, A., Joshi, S. and L.B. Lave. 1998. Economic input-output models for environmental life-cycle assessment. Environmental Science and Technology. 32(7): 184A191A. Hendrickson, C.T., Lave, L.B. and H.S. Matthews. 2006. Environmental life-cycle assessment of goods and services: an input-output approach. Resources for the Future. Washington D.C. Industry Canada (IC). Trade data online. <http://www.ic.gc.ca/epic/site/tdo-dcd.nsf/en/Home> (accessed May 25, 2009). Lave, L.B., Cobas-Flores, E. and F.C. McMichael. 1995. Using input-output analysis to estimate economywide discharges. Environmental Science and Technology. 29(9): 420A-426A. Lenzen, M. 1998. Primary energy and greenhouse gases embodied in Australian final consumption: an input-output analysis. Energy Policy. 26: 495-506. Lenzen, M., Pade, L.-L. and J. Munksgaard. 2004. CO2 multipliers in single-region and multiregion input-output models. Economic Systems Research. 16(4): 391-412.  58  Leontief, W. 1970. Environmental repercussions and the economic structure: an input-output approach. The Review of Economic Statistics. 52(3): 262-271. Machado, G., Schaeffer, R. and E. Worrell. 2001. Energy and carbon embodied in the international trade of Brazil: an input-output approach. Ecological Economics. 39: 409-424. Mäenpää, I. and H. Siikavirta. 2007. Greenhouse gases embodied in the international trade and final consumption of Finland: an input-output analysis. Energy Policy. 35: 128-143. Mongelli, I., Tassielli, G. and B. Notarnicola. 2006. Global warming agreements, international trade and energy/carbon embodiments: an input-output approach to the Italian case. Energy Policy. 34: 88-100. Munksgaard, J. and K.A. Pedersen. 2001. CO2 accounts for open economies: producer or consumer responsibility. Energy Policy. 29: 327-334. Muradian, R., O‟Connor, M. and J. Martinez-Alier. 2002. Embodied pollution in trade: estimating the „environmental load displacement‟ of industrialised countries. Ecological Economics. 41: 5167. Norman, J., Charpentier, A.D. and H.L. MacLean. 2007. Economic input-output life-cycle assessment of trade between Canada and the United States. Environmental Science and Technology 41(5): 1523-1532. Organization for Economic Co-Operation and Development (OECD). Purchasing Power Parities, Main Economic Indicators. <http://www.oecd.org/std/ppp> (accessed, March 9, 2010). Pembina Institute, David Suzuki Foundation. 2009. Climate leadership and economic prosperity: final report on an economic study of greenhouse gas targets and policies for Canada. <http://pubs.pembina.org/reports/climate-leadership-report-en.pdf> (accessed, March 31, 2011). Peters, G.P. and E.G. Hertwich. 2008. CO2 embodied in international trade with implications for global climate policy. Environmental Science and Technology 42(5): 1401-1407. Pitschas, C. 1994-1995. GATT/WTO rules for border tax adjusments and the proposed European directive introducing a tax on carbon dioxide emissions and energy. Georgia Journal of International and Comparative Law. 24: 479-500. Rees, W.E. 2006. Globalization, trade and migration: undermining sustainability. Ecological Economics. 59: 220-225. Rothman, D.S. 1998. Environmental Kuznets curves – real progress or passing the buck? A case for consumption-based approaches. Ecological Economics. 25: 177-194. Schott, J.J. and M. Fickling. 2009. Setting the NAFTA agenda on climate change. Policy Brief 09-18, Washington, DC: Peterson Institute for International Economics. Shui, B. and R.C. Harriss. 2006. The role of CO2 embodiment in US-China trade. Energy Policy. 34: 4063-4068.  59  Statistics Canada. 2005. NAICS-Based Small (S), Medium (M), Historical Link (L) and Worksheet (W) Aggregation-Level Industries for Use With the Input-Output Tables, Make and Use Matrix, from 2002 and Onwards Victor, D., J.C. House, and S. Joy. 2005.A Madisonian approach to climate policy. Science 309: 1820-1821. Weaver, A. 2008. Keeping Our Cool: Canada in a Warming World. Penguin Group (Canada), Toronto. Weber, C.L. and H.S. Matthews. 2007. Embodied environmental emissions in US international trade: growth, structures and uncertainties, 1997-2004. Environmental Science Technology 41: 4875-4881. Wiedmann, T., M. Lenzen, K. Turner, and J. Barrett. 2007. Examining the global environmental impact of regional consumption activities – part 2: review of input-output models for the assessment of environmental impacts embodied in trade. Ecological Economics. 61(1): 15-26. Wiedmann, T. 2009. A review of recent multi-region input-output models used for consumptionbased emission and resource accounting. Ecological Economics. 69: 211-222.  60  Appendix A: GHG Emissions and Kyoto Targets of Select Countries Figure 11: GHG Emissions and Kyoto Targets of Select Countries  GHGs (Thousands Mt CO2e)  800 700 600 500 400 300 200 100 0 1990  1995  2000  2005  4,300 4,200 4,100 4,000 3,900 3,800 3,700 1990  2010  1995  2000  Kyoto Target  EU15 GHG Emissions  Canadian GHG Emission  Kyoto Target  Figure 11c Japan GHG Emissions and Kyoto Target 1,350 1,300 1,250 1,200 1,150 1,100 1995  2000  2005  2,500 2,000 1,500 1,000 500 0 1990  1995  2000  Russia GHG Emissions  Kyoto Target  Kyoto Target  500 400 300 200 100 0 1990  3,000  Japan GHG Emissions  600  1995  2000  2005  2010  2010  3,500  2010  Figure 11e Australia GHG Emissions and Kyoto Target  2005  Figure 11d Russia's GHG Emissions and Kyoto Target GHGs (Thousands Mt CO2e)  1,400  1990  GHGs (Thousands Mt CO2e)  Figure 11b EU15 GHG Emissions and Kyoto Target  GHGs (Thousands Mt Co2e)  GHGs (Thousands Mt CO2e)  GHGs (Thousand Mt CO2e)  Figure 11a Canada's GHG Emissions and Kyoto Target  2005  2010  Figure 11f U.S. GHG Emissions and Kyoto Target 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 1990  1995  2000  Australia GHG Emissions  U.S. GHG Emissions  Kyoto Target  Target  2005  2010  (Source: UNFCCC GHG Data – Detailed Data by Party at http://unfccc.int/di/DetailedByParty.do)  61  Appendix B: Canadian Sectors Included in the Analysis Table 10: Agricultural, Resource and Manufacturing Sectors Included in the Analysis Sectors in EIO-LCA 11A0 1130 1140 1150 2111 2121 2122 2123 2131 2211 3111 3113 3114 3115 3116 3117 311A 312A 312B 312C 312D 3122 31A0 3150 3160 3210 3221 3222 3231 3241 3251 3252  Description  NAICS Definition 111, 112 113 114 115 2111 2121 2122 2123 2131 2211 3111 3113 3114 3115 3116 3117 3112, 3118, 3119 31211 31212 31213 31214 3122 313, 314 315 316 321 3221 3222 3231 3241 3251 3252  3253 3254 325A 3261 3262 3273 327A 3310 3320 3330 3341 334A  Crop and animal production Forestry & logging Fishing, hunting & trapping Support activities for agriculture and forestry Oil & gas extraction Coal mining Metal ore mining Non-metallic mineral mining and quarrying Support activities for mining & oil & gas extraction Electric power generation, transmission & distribution Animal food manufacturing Sugar & confectionery product manufacturing Fruit & vegetable preserving & specialty food manufacturing Dairy product manufacturing Meat product manufacturing Seafood product preparation & packaging Miscellaneous food manufacturing Soft-drink & ice manufacturing Breweries Wineries Distilleries Tobacco manufacturing Textile & textile product mill Clothing manufacturing Leather and allied product manufacturing Wood product manufacturing Pulp, paper & paperboard mills Converted paper product manufacturing Printing & related support activities Petroleum & coal products manufacturing Basic chemical manufacturing Resin, synthetic rubber, artificial/synthetic fibres & filaments manufacturing Pesticides, fertilizers & other agricultural chemical manufacturing Pharmaceutical & medicine manufacturing Miscellaneous chemical product manufacturing Plastic product manufacturing Rubber product manufacturing Cement & concrete product manufacturing Miscellaneous non-metallic mineral product manufacturing Primary metal manufacturing Fabricated metal product manufacturing Machinery manufacturing Computer & peripheral equipment manufacturing Electronic product manufacturing  3352 335A 3361 3362 3363 3364 3365 3366 3369  Household appliance manufacturing Electrical equipment & component manufacturing Motor vehicle manufacturing Motor vehicle body & trailer manufacturing Motor vehicle parts manufacturing Aerospace product & parts manufacturing Railroad rolling stock manufacturing Ship & boat building Other transportation equipment manufacturing  3253 3254 3255, 3256, 3259 3261 3262 3273 3271, 3272, 3274, 3279 331 332 333 3341 3342, 3343, 3344, 3345, 3346 3352 3351, 3353, 3359 3361 3362 3363 3364 3365 3366 3369  62  Sectors in EIO-LCA 3370 3390  Description  NAICS Definition  Furniture & related product manufacturing Miscellaneous manufacturing  337 339  63  Appendix C: Example of BEET Calculations for Motor Vehicles Manufacturing Let‟s take the example of motor vehicle manufacturing (NAICS 3361). Table 11 presents the trade flows associated with this sector: Table 11: Motor Vehicle Manufacturing: Imports and Exports by Dollar Value, 2002 Million CAD Motor Vehicle Manufacturing Exports Imports Imports from U.S. Imports from Rest of World  Million USD Using MER Using PPP 64,565.88 n.a. n.a. 35,592.07 n.a. n.a. 24,044.20 15,314.57 20,205.21 11,547.87  For ease of reference, Equations 7 and 8 are copied below: (7) (8) For the SRIO model, using only the EIO-LCA model for Canada, the variables im and ex are replaced by their respective values of CAD 35,592.07 and CAD 64,565.88. These values are the ones I input into the software, which automatically calculates the EEI and EEE according to Equation 8. Figure 12 and Figure 13 provide snapshots of the output produced by the EIO-LCA software. Figure 12: Emissions Embodied in Canada’s Imports of Motor Vehicles  64  Figure 13: Emissions Embodied in Canada’s Exports of Motor Vehicles  From the software‟s outputs above, it is straightforward to calculate the BEET for the sector of motor vehicle manufacturing: (9) For the MRIO model, I still use the EIO-LCA model for Canada for the exports as well as for the imports of motor vehicles from the rest of the world. However, I use the EIO-LCA model for the U.S. for the imports of motor vehicles originating in the U.S. As described in Section 4.3.2, the challenge that arises with using the U.S. model is related to the need to convert the U.S. import data from Canadian dollars to U.S. dollars. This is because the U.S. model is based on inputoutput tables in U.S. currency. In 2002, the market exchange rate (MER) was at a historical low: 1 CAD = 0.636934 USD. However, the purchasing power parity (PPP) rate was 1 CAD = 0.840336 USD. Table 11 presents the conversion results using both rates. My second BEET estimate is calculated when using the MER to convert U.S. import data and my third BEET estimate is calculated when using the PPP. However, I cannot directly input into the U.S. EIO-LCA software either of the following values: USD 15,314.57 (from MER conversion) or USD 20,205.21 (from PPP conversion) given in Table 11. This is because the industrial sectors in the U.S. model are a lot more disaggregated than those of the Canadian model. Thus, the imports of motor vehicles from the U.S., valued at USD 15,314.57, need to be 65  disaggregated further down the NAICS hierarchy (Table 12). This is the other data manipulation challenge related to using a two-country model, as discussed in Section 4.3.2. Table 12: Imports of Motor Vehicles from the U.S. NAICS  Description  3361 33611  Motor Vehicle Manufacturing Automobile and Light-Duty Motor Vehicle Manufacturing Automobile manufacturing Light-truck and Utility Vehicle Manufacturing Heavy-Duty Truck Manufacturing  336111 336112 336120  Million USD Million USD (MER) (PPP) 15,314.57 20,205.21 13,454.68 6,727.34  17,751.37 8,875.69  6,727.34 1,859.89  8,875.69 2,453.84  In Industry Canada‟s Trade Data Online database, the U.S. import data is available at the 5-digit NAICS level at most whereas the U.S. EIO-LCA model sometimes requires data at the 6-digit level. In my example of motor vehicles, the TDO database provides import data for NAICS 33611 (Automobile and Light-Duty Motor Vehicle Manufacturing) and NAICS 33612 (HeavyDuty Truck Manufacturing), but this level of disaggregation is not enough to input the data into the U.S. EIO-LCA model. Therefore, I needed to assume an equal split between the two lower hierarchical levels (336111 and 336112). For NAICS 33612, there was no need to split the imports because there was only one lower-level category. In the end, the values for Canada‟s imports from the U.S. that I input into the U.S. model are those at the six-digit level in Table 12. Figures 14 and 15 provide some snapshots of the output produced by the EIO-LCA software. Figure 14: Emissions Embodied in Canada’s Imports of U.S. Automobiles (336111)  66  Figure 15: Emissions Embodied in Canada’s Imports of U.S. Heavy-Duty Trucks (336120)  67  Appendix D: U.S. Sectors Included in the Analysis Table 13: Sectors of the U.S. EIO-LCA Model Included in the Analysis Sectors in EIO-LCA 11A0  Description Crop and animal production  Sectors in U.S. EIO-LCA 1111A0 1111B0  111200 1113A0  11133 1114 111910 111920 111930 1119B0 1121A0 112120 1123  112A00  1130  Forestry & logging  1140  Fishing, hunting & trapping  1150  Support activities for agriculture and forestry  2111 2121 2122  Oil & gas extraction Coal mining Metal ore mining  2123  Non-metallic mineral mining and quarrying  2131  Support activities for mining & oil & gas extraction  2211 3111  Electric power generation, transmission & distribution Animal food manufacturing  3113  Sugar & confectionery product manufacturing  3114  Fruit & vegetable preserving & specialty food manufacturing Dairy product manufacturing  3115  113300 113A00 114100 114200 115000 211000 212100 212210 212230 2122A0 212310 212320 212390 213111 213112 213113 221100 311111 311119 311313 31131A 31132 31133 31134 311410 311420 311513 311515 31151A 311520  NAICS Definition United States 11111, 11112 11113, 11114 11115, 11116, 11119 11121 11131, 11132, 111331, 111332, 111333, 111334, 111336, 111339, 111335, 111336 11141 11191 11192 11193 11194, 11199 11211, 11213 11212 11231, 11232, 11233, 11234, 11239 11221, 11241, 11242, 11251, 11291, 11292, 11293, 11299 11331 11311, 11321 11411 11421 11511, 11521, 11531 21111 21211 21221 21223 21222, 21229 21231 21232 21239 213111 213112 213113 22111, 22112 311111 311119 311313 311311, 311312 31132 31133 31134 31141 31142 311513 311514 311511, 311512 311520  68  Sectors in EIO-LCA 3116  Description  3117 311A  Seafood product preparation & packaging Miscellaneous food manufacturing  312A 312B 312C 312D 3122 31A0  Soft-drink & ice manufacturing Breweries Wineries Distilleries Tobacco manufacturing Textile & textile product mill  3150  Clothing manufacturing  3160  Leather and allied product manufacturing  3210  Wood product manufacturing  3221  Pulp, paper & paperboard mills  3222  Converted paper product manufacturing  Meat product manufacturing  Sectors in U.S. EIO-LCA 311615 31161A 311700 311210 311221 311225 31122A 311230 311810 311820 311830 311910 311920 311930 311940 311990 312110 312120 312130 312140 3122A0 313100 313210 313220 313230 313240 313310 313320 314110 314120 314910 314990 315100 315210 315220 315230 315290 315900 316110 316210 316990 321100 321219 32121A 32121B 321910 321920 321991  322110 322120 322130 322210 32222A 32222B 322230 322291  NAICS Definition United States 311615 311611, 311612, 311613 31171 31121 311221 311225 311222, 311223 311230 31181 31182 31183 31191 31192 31193 31194 31199 31211 31212 31213 31214 31221, 31222 31311 31321 31322 31323 31324 31331 31332 31411 31412 31491 31499 31511 31521 31522 31523 31529 31599 31611 31621 31699 32111 321219 321211, 321212 321213, 321214 32191 32192 321991 321992 321999 32211 32212 32213 32221 322221, 322222 322223, 322224, 322225, 322226 32223 322291  69  Sectors in EIO-LCA  Description  3231  Printing & related support activities  3241  Petroleum & coal products manufacturing  3251  Basic chemical manufacturing  3252  Resin, synthetic rubber, artificial/synthetic fibres & filaments manufacturing  3253 3254  Pesticides, fertilizers & other agricultural chemical manufacturing Pharmaceutical & medicine manufacturing  325A  Miscellaneous chemical product manufacturing  3261  Plastic product manufacturing  3262  Rubber product manufacturing  3273  Cement & concrete product manufacturing  327A  Miscellaneous non-metallic mineral product manufacturing  Sectors in U.S. EIO-LCA 322299 323110 323120 324110 324121 324122 324191 324199 325110 325120 325130 325181 325182 325188 325190 325211 325212 325220 325310 325320 325411 325412 325413 325414 325510 325520 325610 325620 325910 3259A0 326110 326121 326122 326130 326140 326150 326160 32619A 326210 326220 326290 327310 327320 327330 327390 327110 32712A 32712B 327211 327212 327213 327215 3274A0 327910 327991 327992 327993 327999  NAICS Definition United States 322299 32311 32312 32411 324121 324122 324191 324199 32511 32512 32513 325181 325182 325188 32519 325211 325212 32522 325311 325320 325411 325412 325413 325414 32551 32552 32561 32562 32591 32592, 32599 32611 326121 326122 32613 32614 32615 32616 326191, 326192, 326199 32621 32622 32629 32731 32732 32733 32739 32711 327121, 327122, 327123 327124, 327125 327211 327212 327213 327215 32741, 32742 32791 327991 327992 327993 327999  70  Sectors in EIO-LCA 3310  3320  Description Primary metal manufacturing  Fabricated metal product manufacturing  Sectors in U.S. EIO-LCA 331110 331200 331314 33131A 33131B 331411 331419 331420 331490 331510 331520 332114 33211A 33211B 33221A 33221B 332310 332320 332410 332420 332430 332500 332600 332710 332720 332800 332991 332913 33291A 332996 33299A 33299B 33299C  3330  Machinery manufacturing  333111 333112 333120 333130 333220 333295 33329A  333314 333315 333319 33331A 333414 333415 33341A 333511 333514 333515 33351A 33351B 333611 333612  NAICS Definition United States 33111 33121, 33122 331314 331311, 331312 331315, 331316, 331319 331411 331419 33142 33149 33151 33152 332114 332111, 332112, 332117 332115, 332116 332211, 332214 332212, 332213 33231 33232 33241 33242 33243 33251 33261 33271 33272 33281 332991 332913 33291 332996 332992, 332993 332994, 332995 332997, 332998, 332399 333111 333112 33312 33313 33322 333295 33321, 333291, 333292, 333293, 333294, 333298 333314 333315 333319 333311, 333312, 333313 333414 333415 333411, 333412 333511 333514 333515 333512, 333513 333516, 333518 333611 333612  71  Sectors in EIO-LCA  Description  Sectors in U.S. EIO-LCA 333613 333618 333911 333912 333920 333991 333993 333994 33399A  3341  Computer & peripheral equipment manufacturing  334A  Electronic product manufacturing  3352  Household appliance manufacturing  335A  Electrical equipment & component manufacturing  3361  Motor vehicle manufacturing  3362  Motor vehicle body & trailer manufacturing  33399B 334111 334112 33411A 334210 334220 334290 334300 334411 334412 334413 334417 334418 334419 33441A 334510 334511 334512 334513 334514 334515 334516 334517 33451A 334613 33461A 335210 335221 335222 335224 335228 335110 335120 335311 335312 335313 335314 335911 335912 335920 335930 335991 335999 336111 336112 336120 336211 336212  NAICS Definition United States 333613 333618 333911, 333913 333912 333921, 333922, 333923, 333924 333991 333993 333994 333992, 333997, 333999 333995, 333996 334111 334112 334113, 334119 33421 33422 33429 33431 334411 334412 334413 334417 334418 334419 334414, 334415, 334416 334510 334511 334512 334513 334514 334515 334516 334517 334518, 334519 334613 334611, 334612 33521 335221 335222 335224 335228 33511 33512 335311 335312 335313 335314 335911 335912 33592 33593 335991 335999 336111 336112 33612 336211 336212  72  Sectors in EIO-LCA  Description  3363 3364  Motor vehicle parts manufacturing Aerospace product & parts manufacturing  3365 3366  Railroad rolling stock manufacturing Ship & boat building  3369  Other transportation equipment manufacturing  3370  Furniture & related product manufacturing  3390  Miscellaneous manufacturing  Sectors in U.S. EIO-LCA 336213 336214 336300 336411 336412 336413 336414 33641A 336500 336611 336612 336991 336992 336999 337110 337121 337122 33712A 337127 337212 337215 33721A 337910 337920 339111 339112 339113 339114 339115 339116 339910 339920 339930 339940 339950 339991 339992 339994 33999A  NAICS Definition United States 336213 336214 33631 336411 336412 336413 336414 336415, 336419 33651 336611 336612 336991 336992 336999 33711 337121 337122 337124, 337125, 337129 337127 337212 337215 337211, 337214 33791 33792 339111 339112 339113 339114 339115 339116 33991 33992 33993 33994 33995 339991 339992 339994 339993, 339995, 339999  73  Appendix E: Canada’s Sectoral Trade Balances and BEET Table 14: Canada’s Balance of Trade and BEET by NAICS Industry Groups, 2002 Sectors in EIOLCA 11A0 1130 1140 1150 11 2111 2121 2122 2123 2131 21 2211 22 3111 3113 3114 3115 3116 3117 311A 312A 312B 312C 312D 3122 31A0 3150 3160 3210 3221 3222 3231 3241 3251 3252 3253 3254 325A 3261 3262 3273 327A 3310 3320 3330 3341 334A 3352 335A 3361 3362 3363  Description  Crop and animal production Forestry & logging Fishing, hunting & trapping Support activities for agriculture and forestry Agriculture, Forestry, Fishing and Hunting Oil & gas extraction Coal mining Metal ore mining Non-metallic mineral mining and quarrying Support activities for mining & oil & gas extraction Mining, Quarrying, and Oil & Gas Extraction Electric power generation, transmission & distribution Utilities Animal food manufacturing Sugar & confectionery product manufacturing Fruit & vegetable preserving & specialty food manufacturing Dairy product manufacturing Meat product manufacturing Seafood product preparation & packaging Miscellaneous food manufacturing Soft-drink & ice manufacturing Breweries Wineries Distilleries Tobacco manufacturing Textile & textile product mill Clothing manufacturing Leather and allied product manufacturing Wood product manufacturing Pulp, paper & paperboard mills Converted paper product manufacturing Printing & related support activities Petroleum & coal products manufacturing Basic chemical manufacturing Resin, synthetic rubber, artificial/synthetic fibres & filaments manufacturing Pesticides, fertilizers & other agricultural chemical manufacturing Pharmaceutical & medicine manufacturing Miscellaneous chemical product manufacturing Plastic product manufacturing Rubber product manufacturing Cement & concrete product manufacturing Miscellaneous non-metallic mineral product manufacturing Primary metal manufacturing Fabricated metal product manufacturing Machinery manufacturing Computer & peripheral equipment manufacturing Electronic product manufacturing Household appliance manufacturing Electrical equipment & component manufacturing Motor vehicle manufacturing Motor vehicle body & trailer manufacturing Motor vehicle parts manufacturing  Trade Balance (Ex-Im) Million $ 4,811.3 42.7 504.3 -251.0 5,107.2 23,793.7 598.2 3,185.0 3,605.7 0.0 31,182.7 1,304.6 1,304.6 29.7 24.4 -433.3 -80.2 3,324.1 1,835.8 -152.9 148.3 50.1 -934.7 78.6 59.1 -2,784.9 -3,022.3 -2,076.9 15,654.3 17,152.2 107.0 762.8 7,097.9 -1,185.5 -418.0  BEET_3 (EEI – EEE) kMT CO2e -63,803.4 -95.2 -36.0 279.2 -63,655.4 -59,540.6 907.2 -1,164.9 -1,687.6 0.0 -61,485.9 -5,653.4 -5,653.4 506.1 516.5 933.5 234.3 -7,783.9 -772.8 1,387.1 119.9 22.8 113.7 8.8 4.3 2,823.8 790.7 317.6 -17,738.6 -26,950.7 1,261.9 -114.5 -19,737.1 6,156.8 5,386.5  -252.2 -5,519.1 -5,342.1 2,464.3 -801.2 -2,013.8 683.1 6,979.5 -5,993.8 -10,739.7 -8,972.1 -13,352.2 -911.1 -4,938.9 28,973.8 -431.8 -22,915.1  -817.4 1,670.0 4,713.9 -1,426.0 1,760.7 2,318.7 -1,143.7 -25,588.6 925.38 2,988.0 1,776.9 4,702.3 622.3 2,765.6 -93,012.0 678.2 7,915.0  74  Sectors in EIOLCA 3364 3365 3366 3369 3370 3390 31-33  Description  Aerospace product & parts manufacturing Railroad rolling stock manufacturing Ship & boat building Other transportation equipment manufacturing Furniture & related product manufacturing Miscellaneous manufacturing Manufacturing TOTAL  Trade Balance (Ex-Im) Million $ 3,056.8 20.1 361.3 -848.8 3,758.0 -6,698.0 -8,200.4 29,394.1  BEET_3 (EEI – EEE) kMT CO2e -497.82 153.6 103.4 349.4 -836.9 2,427.4 -139,965.0 -270,759.6  75  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072438/manifest

Comment

Related Items