Reducing Greenhouse Gas Emissions Associated with Consumption ...

Stockholm Environment Institute, Working Paper 2012-05

Reducing Greenhouse Gas Emissions Associated with Consumption: A Methodology for Scenario Analysis Peter Erickson, Chelsea Chandler and Michael Lazarus

Stockholm Environment Institute Kräftriket 2B 106 91 Stockholm Sweden Tel: +46 8 674 7070 Fax: +46 8 674 7020 Web: www.sei-international.org Author contact: Peter Erickson Stockholm Environment Institute-U.S. Centre 1402 Third Avenue, Suite 900 Seattle, WA 98101, USA [email protected] Head of Communications: Robert Watt Publications Manager: Erik Willis Cover Photo: © Markles55/Flickr This publication may be reproduced in whole or in part and in any form for educational or non-profit purposes, without special permission from the copyright holder(s) provided acknowledgement of the source is made. No use of this publication may be made for resale or other commercial purpose, without the written permission of the copyright holder(s). About SEI Working Papers: The SEI working paper series aims to expand and accelerate the availability of our research, stimulate discussion, and elicit feedback. SEI working papers are work in progress and typically contain preliminary research, analysis, findings, and recommendations. Many SEI working papers are drafts that will be subsequently revised for a refereed journal or book. Other papers share timely and innovative knowledge that we consider valuable and policyrelevant, but which may not be intended for later publication. Copyright © August 2012 by Stockholm Environment Institute

STOCKHOLM ENVIRONMENT INSTITUTE WORKING PAPER NO. 2012-05

Reducing Greenhouse Gas Emissions Associated with Consumption: A Methodology for Scenario Analysis

Peter Erickson, Chelsea Chandler and Michael Lazarus Stockholm Environment Institute

ABSTRACT In recent years, climate policy analysts have explored the links between consumption patterns and greenhouse gases (GHGs) by developing methods to estimate life-cycle emissions associated with different categories of consumption, e.g., a carbon “footprint” or a “consumption-based” GHG inventory. Implicit in many of the studies is the notion that shifts in consumption patterns could lead to reductions in global emissions. For example, if consumers were to shift their purchases from particularly GHG-intensive goods and services (e.g. red meat) to less GHG-intensive goods and services (e.g., grains and legumes) global emissions may decline. However, surprisingly few studies have attempted to construct long-term scenarios for how shifts in consumption patterns and behavior could reduce emissions. In this paper, we develop and document a methodology for constructing long-term scenarios of a transition to low-GHG consumption. We then apply it to a major U.S. city, Seattle, Wash., which has been active in climate action planning and helped organize over one thousand U.S. mayors to adopt GHG-reduction goals.

REDUCING GREENHOUSE GAS EMISSIONS ASSOCIATED WITH CONSUMPTION

SEI WP 2012-05

TABLE OF CONTENTS 1. Introduction and context ............................................................................................. 3 2. Consumption-based GHG accounting ........................................................................ 3 3. From consumption-based accounting to scenario analysis ........................................ 5 Evaluation of methods ....................................................................................................... 8 4. Proposed scenario methodology .............................................................................. 11 Forecasting activity ......................................................................................................... 11 Forecasting intensity ........................................................................................................ 12 5. Catalogue of low-GHG measures and policies ........................................................ 13 Typology of behaviors and measures ............................................................................... 14 6. Piloting the approach for a major U.S. city............................................................... 17 GHGs associated with current consumption in Seattle ...................................................... 17 Baseline forecast of GHGs associated with consumption .................................................. 18 Low-GHG consumption mitigation measures ................................................................... 20 Low-GHG consumption scenario results .......................................................................... 22 7. Key limitations and uncertainties ............................................................................. 24 8. Next steps and conclusions....................................................................................... 25 Acknowledgments ........................................................................................................ 26 References .................................................................................................................... 27 Appendix A. Methodology for forecasting consumption activity ................................... 33 Appendix B: Testing the approach for the U.S. ............................................................. 35 GHGs associated with current consumption in the U.S. .................................................... 35 Low-GHG consumption mitigation measures and scenario results .................................... 36 Conclusions of U.S. analysis ............................................................................................ 37

2


SEI WP 2012-05

1. INTRODUCTION AND CONTEXT In recent years, climate policy analysts have explored the links between consumption patterns and greenhouse gases (GHGs) by developing methods to estimate life-cycle emissions associated with different categories of consumption, e.g., a carbon “footprint” or a “consumption-based” GHG inventory (Minx et al. 2009; Dawkins et al. 2010; Wiedmann 2008; Jones and Kammen 2011; Weber et al. 2010; Davis and Caldeira 2010; Peters et al. 2011). Implicit in many of the studies is the notion that shifts in consumption patterns could lead to reductions in global emissions. For example, if consumers were to shift their purchases from particularly GHG-intensive goods and services (e.g. red meat) to less GHGintensive goods and services (e.g., grains and legumes), global emissions might decline. However, surprisingly few studies have attempted to construct long-term scenarios for how shifts in consumption patterns and behavior could reduce emissions. For example, emissions scenarios constructed by the Intergovernmental Panel on Climate Change (Metz et al. 2007), the International Energy Agency (IEA 2010), McKinsey & Company (2009), and Ecofys (WWF 2011) primarily focus on the production practices of key sectors (e.g., power, industry, buildings, transportation, agriculture), rather than on the consumption behaviors associated with emissions-intensive products. Accordingly, these scenarios pay relatively little attention to strategies associated with changes in consumption patterns (e.g., diet shift) and other behaviors that might lead to reduced emissions. The goal of this paper is to document and pilot a methodology for constructing long-term emissions scenarios associated with a transition to low-GHG consumption. In six main sections, the paper: •

Reviews the consumption-based approach to emissions accounting and its advantages and disadvantages relative to traditional production-based approaches;

•

Outlines an overall framework for developing emissions scenarios from a consumption-based perspective, as well as a series of key methodological issues, and documents how they have been addressed in existing consumption-based scenario approaches;

•

Develops a methodology for creating and evaluating baseline and mitigation emissions scenarios from a consumption-based perspective;

•

Presents a catalogue of policy options to reduce consumption-based emissions;

•

Tests the approach for a major U.S. city, Seattle, Wash. (the approach is piloted for the entire United States in an appendix); and

•

Identifies key limitations, uncertainties, and strengths of this approach.

We end with some conclusions and suggested next steps for consumption-based scenario analysis. 2. CONSUMPTION-BASED GHG ACCOUNTING Over the past two decades, a relatively standard method has evolved to account for GHG emissions by geographic scale. Widely referred to as the production-based method, it involves quantifying the emissions produced within a regional boundary, such as a nation (IPCC 1996a), sub-national territory (EPA 2008), or community (ICLEI 2009; C40 and ICLEI 2012). Most production-based emission estimates draw from methods first set forth in 1994 by the IPCC for use by nations in compiling their official GHG emission inventories (IPCC 1996a). These methods serve as the official methods for nations to submit inventories and 3


SEI WP 2012-05

track progress towards emission-reduction goals, such as those established in the United Nations Framework Convention on Climate Change (UNFCCC) and its Kyoto Protocol. Over time, analysts have pointed out limitations in production-based GHG accounting. One prominent critique is that it may give a misleading impression of national progress in curbing GHG emissions, as decreases (or increases) in national emissions may be due to the shift in activities (such as manufacturing) to (or from) another country, without decreasing (or increasing) total global emissions (Wiedmann 2008; Peters and Hertwich 2008; Peters et al. 2011; Carbon Trust 2011). As an alternative, several authors have suggested consumptionbased emissions accounting, which includes the emissions embodied in trade and would therefore, in principle, account for such carbon leakage effects, whether due to climate policy or instead to broader economic trends (Peters 2008). Consumption-based accounting can also highlight the GHGs associated with different product types, ranging from food and beverages, to clothing and appliances. Highlighting the emissions associated with products can help build support for new opportunities to reduce global GHGs, such as consuming fewer high-GHG items and other strategies collectively referred to as “sustainable consumption” (Oregon DEQ 2007). At the level of the entire globe, consumption and production-based emissions would be identical. But for most countries, the difference between production- and consumption-based emissions is significant. For example, for countries where consumption-based emissions exceed production-based emissions (due, in most cases, to these countries relying significantly on imports), the difference averages 29% (Peters et al. 2011). As shown in Figure 1, production and consumption-based approaches share some sources of emissions. For example, fuels used directly by households, or in-region emissions from producing goods that are consumed within the region, are included in both approaches. Figure 1: Production and consumption-based emissions accounting

Most methods for conducting consumption-based GHG inventories have used input-output (IO) analysis, including assessments for nations (Davis and Caldeira 2010; Hertwich and Peters 2009; Weber and Matthews 2007; Lenzen et al. 2012) as well as sub-national territories (Ramaswami et al. 2008; Stanton, Bueno, Ackerman, et al. 2011). Organization of economic data into input-output tables allows for expenditures in one sector of the economy to be tracked to all other sectors (Miller and Blair 2009). If the emissions intensity (as measured on a per-dollar basis) of each of these industries is known, the entire GHG emissions released to produce the car can be estimated by totaling the emissions associated with each individual unit of activity in the product chain.

4


SEI WP 2012-05

The way in which consumption-based GHG inventories are calculated presents some disadvantages for fully integrating them into national or sub-national climate policy. For one, the complex modeling calculations depend on less certain national and international data sources, making them inherently less reliable than relatively straightforward production-based inventories that can be calculated based largely on direct fuel-use data within each territory. Consumption-based emissions accounting also suffers from flaws that limit its ability to track progress against the very types of goals it might inspire. For example, at the sub-national level, few consistent, regular data sources are available that would enable monitoring of changes in consumption behaviors (e.g., residents’ diets) over time. These limitations are not fatal, and will be important to overcome to track progress on the types of measures discussed in this paper. 1 3. FROM CONSUMPTION-BASED ACCOUNTING TO SCENARIO ANALYSIS A consumption-based emissions inventory provides a snapshot of emissions at a given point in time. While such an inventory can help identify categories of consumption responsible for a significant share of the overall emissions footprint, an inventory cannot, on its own, assess or demonstrate possible pathways to reduce those emissions. For that, analysts turn to scenario analyses that identify promising courses of action, evaluate specific possible measures, and assemble a cohesive storyline about how a region could affect its footprint over time. International organizations such as the IPCC, the IEA, and numerous academic and private-sector research institutions have conducted numerous scenarios showing how to transition to low-carbon energy and a low-carbon economy. However, as noted in the introduction, few have specifically addressed the role of consumption. To begin constructing a method for conducting scenario analysis of low-GHG consumption, it is useful to identify the underlying drivers of impacts, in this case, greenhouse gases. At the most basic level, an economy is people who produce and consume goods and services. The number of people, their consumption behaviors, and the means in which the consumed goods and services are produced all affect GHG emissions. If written as an equation, these drivers would look as follows and be called a “Kaya Identity” (Kaya 1990): 2 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝐸𝑛𝑒𝑟𝑔𝑦 𝐺𝐻𝐺𝑠 �×� �×� �. 𝑃𝑒𝑟𝑠𝑜𝑛 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝐸𝑛𝑒𝑟𝑔𝑦

𝐺𝐻𝐺𝑠 = (𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛) × �

In the case of environmentally extended input-output modeling, used for most consumptionbased GHG inventories, the last two terms often reduce to yield a simplified relationship, as follows, where activity is measured in dollar-value of consumption and the final term is a measure of GHG intensity of that activity: 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝐺𝐻𝐺𝑠 �×� �. 3 𝑃𝑒𝑟𝑠𝑜𝑛 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦

𝐺𝐻𝐺𝑠 = (𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛) × �

To construct a forecast of emissions, an analyst must consider changes in each of these variables. For example, to create a business-as-usual (BAU) baseline, or reference scenario, an analyst would forecast changes in the terms in the equations above in the absence of policies specifically designed to reduce GHGs. In a mitigation scenario, the analyst would 1

For additional limitations, see Hertwich (2005). This is similar to the drivers population, affluence, and technology, or the “IPAT” equation developed by Ehlrich and Holdren in the early 1970s, which has since been expressed in a number of similar iterations (Chertow 2000; Hertwich 2005). 3 Note that this second version of the equation is quite similar to the IPAT equation, where activity per person is a function of affluence, and GHGs per activity is a function of the technologies employed. 2

5


SEI WP 2012-05

pose changes in some subset of these drivers to explore an alternate pathway: one where measures are taken to reduce GHG emissions. Using a simple relationship such as the Kaya identity brings transparency, and an ability to easily combine measures that affect one term or another without double-counting. (For example, reducing vehicle travel and increasing vehicle efficiency both reduce GHGs from personal transportation, but the impact of either one is muted by the other.) The downside to this approach is that it cannot (on its own) account for second-order impacts, such as if improving the fuel efficiency of freight trucks increases truck traffic at the expense of rail traffic, or take into account broader macroeconomic, or price, effects (Greene and Plotkin 2011). Using a relationship like a Kaya identity to create a forecast methodology is the first step towards creating a scenario of future low-GHG consumption. The method for forecasting emissions is also subject to several other, more detailed considerations, such as how to model changes in consumption activity and production energy and GHG-intensity. We will discuss these factors more in the next section, where we discuss our model. Apart from the forecast methodology, analysts also make several other important choices when constructing scenarios of GHGs associated with consumption. These include: 

Types of consumers, or “final demand”, to consider. Most national economic accounts track consumption or final demand from three sources: consumer expenditures, government spending, and business capital investment (plus net exports). Household demand is the largest component of demand (about two-thirds in the U.S.) and is most often the focus, but can exclude significant sources of emissions. In some cases, emissions associated with capital investment can be treated as an input to production and therefore included indirectly (Hertwich 2005).



Categories of consumption addressed. Consumers purchase a wide variety of goods and services. Some studies address all broad categories, whereas others focus on a particular subset (such as food).



Modeling in monetary or functional units. Most input-output models work in monetary units, and track expenditures in one sector to all other sectors. However, GHG emissions may correlate more strongly with “functional” units, such as kilograms (or calories) of food or number of pairs of blue jeans.



Quantification of rebound. Income freed up by reduced spending on certain goods and services is often spent instead on other goods and services, or the “rebound effect” (Hertwich 2008; Ornetzeder et al. 2008; Girod et al. 2010). The study must articulate whether it will redirect freed-up income to other types of consumption or instead assumes that any freed-up income is retained by consumers (or not earned in the first place).



Identification of mitigation measures. Construction of the mitigation scenario often involves identifying specific measures to include in the scenario, measures that in aggregate tell a plausible and internally consistent storyline. This is called a “bottomup” analysis. Alternatively some studies identify a particular end-state lifestyle as the goal and model the impacts of such a lifestyle. Regardless, studies need a clear rationale for what measures are selected.



Time scale. Scenario analyses can explore transformations over relatively short term (10 years or less) to long term (30 years or more) time scales. The longer the time scale, the more speculative the scenario becomes, but the bigger the transformations that can be envisioned. Alternatively, some scenarios simply look at a static (base 6


SEI WP 2012-05

year) snapshot of consumption behaviors and imagine how that particular consumption pattern could be altered, without articulating how the specific changes might evolve over time. There is no widespread agreement about how to address these factors in scenario analyses of consumption, although several trends do emerge. For example, most analyses to date focus on household demand only but do not quantify the rebound effect. However, other particulars vary considerably, such as what measures are analyzed and the timescale of the analysis. Table 1 compares how several existing analyses addressed the factors discussed above. Table 1: Comparison of scenarios from a consumption perspective EU – European Commission (Brohmann and Barth 2011)

UK – London (Bioregional and the London Sustainable Development Commission 2009)

UK – University of Surrey (Druckman and Jackson 2010)

UK – WRAP (Scott et al. 2009)

US – Michigan State and others (Dietz et al. 2009)

California – LBNL (Wei et al. 2011)

Forecast

Forecasts both demandand supplyside trends 4

Forecasts pop. but holds other demand and supply-side trends constant

None (Static)

Econometric forecast of demandand supplyside trends

None (Static)

None (Static)

Types of “final demand” considered

HH

HH, Gov

HH

HH, Gov

HH

HH

Units

Unclear 5

Monetary

Monetary

Monetary

Functional

Functional

Quantification of rebound

No

No

No

Yes

No

No

Identification of mitigation measures

Bottom-up

Bottom-up

End-state: models “minimum income standard”

Bottom-up

Bottom-up

Bottom-up

Time scale

Through 2030

Through 2050

None

Through 2050

Through 2015

Through 2050

X

X

X

X

Goods

X

X

X

X

Services

X

X

X

Construction

X

X

X

Personal transportation

X

X

X

X

X

X

X

X

Scenario

Methodology

Categories of consumption addressed Food

Home energy

X

X

4

Based on PRIMES model of the European Commission DG Energy; and CAPRI, an agriculture model of European Commission DG Agriculture 5 Forecasts use the CAPRI model, but not clear what units are, per http://www.eupopp.net/docs/ wp4.2_bau_sc_scenar.pdf.

7


SEI WP 2012-05

GHG reductions due to consumption-based measures 6 Food,

Goods,

N/A 7

~14% 8

~16%

10-14% 9

N/A7

5-7%

~20%

~21%

N/A

N/A

5-8%

~34%

~37%

10-14%

N/A

10-15% 10

London

UK

UK

US

California

Services, Construction Personal Transport, Home Energy Total Other factors Geographic

EU-27

Focus

Table 1 summarizes studies that have looked at a variety of measures across commodities and sectors, but other studies have dug deeper into single products, especially food (Duchin 2008; Stehfest et al. 2009; Audsley et al. 2010; Audsley et al. 2009). These studies display similar variations in methodology and help illustrate the implications of different choices. For example, of the four food-specific studies we reviewed, three assume that, in the baseline, future diets do not change. The fourth (Stehfest et al. 2009) assumes that increasing incomes (especially in developing countries) will lead to increased consumption of meat and dairy and thus that average per-capita GHG emissions will be higher. One of the studies (Duchin 2008) identifies an additional factor to consider: climate change may affect yields of agricultural crops in different world regions and, in turn, the types of diets possible. Evaluation of methods

Given the diverse approaches taken in existing scenarios of consumption, it is clear that no standard method is yet being employed. While expecting future analysts to follow a single common standard is probably not realistic (nor necessarily desirable), it is worth considering what traits lead to a robust and realistic scenario. We suggest that scenario analyses strive to meet the following criteria: 

Complete and systematic, such that all relevant sources of demand, categories of consumption, economic trends and drivers, and mitigation measures are included. Some studies may conduct in-depth investigations of particular factors, but even in doing so, should generally discuss the existence of the factors not considered. For example, a study that focuses only on household demand should mention the existence of other types of demand: government and capital investment. 11

6

At maximum implementation of measures considered, regardless of what year that implementation occurs. Source did not report a total BAU against which to assess the impact of the measures assessed. 8 Approximated based on identification of consumption-based measures in the study’s appendix. Reductions would likely be greater if the study did not also apply aggressive efficiency improvements in most sectors. Range reflects measures that only apply to goods (including food) and services (on the low end) to also including measures that address personal transportation and home energy (on the high end). 9 Depending on degree of rebound. 10 A related study, Jones and Kammen (2011) found potential of 20% based on maximum implementation of the same measures identified. 11 In some cases, capital investment may be modeled as in input to production, and so would not need to be considered as a separate type of demand. 7

8


SEI WP 2012-05



Dynamic, such that the study considers the interplay of key factors over time. Most critical is the effect of changing production practices on the GHG-intensity of consumption. For example, improving industrial production practices may lower the GHG intensity of goods, meaning that each unit of behavior change has less of an effect. How consumers spend income freed up by changing consumption patterns (the rebound effect) is also important. A third factor – price, or the market effects of increasing or decreasing consumption – is more difficult to assess.



Cohesive, such that the overall storyline is internally consistent and plays out over a time scale that is long enough to allow for significant changes but short enough that economic structures can still be anticipated. The time scale of consumption-based scenarios may be limited by the reliance on input-output models, which assume particular patterns of industry inter-dependence and are not easily adjusted to reflect changing conditions over time. On the other hand, at least one scenario (Scott et al 2009) has attempted to forecast how the production structure (the input-output matrix itself) would change over time.

For each of the key methodological factors discussed above (e.g., forecast methodology, categories of consumption), Table 2 recommends an approach that strives to satisfy the above criteria. However, note that these criteria are relatively broad, and may not be attainable in all circumstances.

9


SEI WP 2012-05

Table 2: Recommended approach to consumption-based scenarios Methodology element

Recommendation

Rationale

Known limitations

Example scenarios

Forecast of baseline GHG emissions

Include trends in population, demand (affluence and consumer behavior), and supply (technology and production practices).

Allows the scenario to include important relevant trends, dynamically adjust to these trends, and tell a cohesive story about the evolution of the economy and consumer behavior.

Can be difficult for sub-national territories where forecasts in demand and production may not exist.

European Commission (Brohmann and Barth 2011) and UK WRAP (Scott et al. 2009), though both could benefit from better documentation.

Types of final demand considered

Include household consumption, government spending, and capital investment. Include capital investment as an input to production if possible.

Capital investment is a large fraction of final demand in many countries and should be included. However, since it is largely in service to production of other goods and services (and, in turn, to household demand), the best practice would be to endogenize it.

Data on government spending and capital may be limited; Including capital investment as an input to production (endogenizing it) has proven difficult, and is not consistent with national economic accounts.

None known that assess all types of final demand, suggesting this is an area for future research.

Categories of consumption (e.g., food, goods, services)

Include all significant categories wherever possible.

Including all categories contributes to completeness.

Local data on some categories may not exist.

UK – London study (Bioregional and the London Sustainable Development Commission 2009) addresses all categories.

Modeling in monetary or functional units

Use functional units whenever possible.

Functional units correlate more strongly with GHGs than do monetary units.

Few data sources to support trends in functional units.

Stehfest et al (2009), for food.

Quantification of rebound

Quantify rebound.

The rebound effect is a well-known phenomenon.

Generally requires broad assumptions as relatively little empirical research exists.

UK WRAP (Scott et al. 2009), though other, more specific assessments (Ornetzeder et al. 2008; Girod et al. 2010) explore emerging best practices.

Identification of mitigation measures

Include all measures shown to have significant GHG abatement potential, ideally with discussion (at least) or assessment of potential consumer uptake.

Including all significant measures contributes to completeness, and an assessment of plausible consume uptake may help increase the cohesion (and plausibility) of the scenario.

Relatively little research to characterize plausible magnitudes or rates of change.

The UK – London study was the most comprehensive study reviewed, but two U.S. studies (Dietz et al. 2009; Wei et al. 2011) offered best assessment of consumer uptake.

Time scale

No specific recommendation.

Longer timelines are more speculative; shorter timelines limit the transformative potential.

10


SEI WP 2012-05

4. PROPOSED SCENARIO METHODOLOGY In this section, we propose a methodology for scenario analysis for use in the United States and for sub-national governments therein. We design this method to meet most of the recommended practices discussed in the previous section. However, given limitations in available data, not all recommendations can be realized. First, a scenario analysis of future emissions associated with consumption must start with a snapshot of emissions in the base year: the consumption-based GHG inventory. Reviews of methods for consumption-based GHG inventories have been published elsewhere (Davis and Caldeira 2010; Peters 2008; Chavez and Ramaswami 2011) and will not be reviewed in detail here, except to reiterate that input-output analysis is the de facto technique for such analyses, at least for GHGs associated with goods and services, and as described in Section 2. GHGs associated with home energy use and personal vehicle travel are better assessed with locally available data on home energy use (e.g., MBTU of natural gas use) and vehicle travel (e.g., vehicle miles traveled). Our discussion below focuses on how to forecast the consumptionbased GHG inventory and develop scenarios from it. As discussed above, a forecast methodology starts with consideration of the underlying drivers, per the equation: 𝐺𝐻𝐺𝑠 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 �×� �. 𝐺𝐻𝐺𝑠 = (𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛) × � 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝑃𝑒𝑟𝑠𝑜𝑛

Forecasting activity

Forecasting changes in the first term, activity, requires making predictions about consumer behavior. Entire fields of study are devoted to consumer behavior, and the U.S. and many other countries have rich data on historical consumer spending, such as the Consumer Expenditure Survey in the U.S. These data can be used to make broad forecasts of how consumer behavior might evolve over time. For example, analysis of consumer spending data over time has shown that as average incomes increase, consumers spend disproportionately more or less on different types of goods and services. Based on these data, researchers have developed parameters (e.g., economic elasticities) to forecast expenditures on particular product types based on changes in income or overall expenditure levels (Taylor and Houthakker 2009). However, spending more on particular categories of goods and services does not necessarily translate to higher GHGs. For example, research has found that increasing incomes may lead consumers to buy higher priced goods, but not necessarily more of them (Girod and de Haan 2010). Forecasts of consumer expenditures must therefore consider whether a unit increase in spending may translate to less than one unit increase in GHGs. To address these issues, we propose to forecast consumption activity using a two-step process. The first step is to forecast total consumer expenditures using the most relevant forecast of economic activity (i.e., expenditures, or income). Forecasts of expenditures or income are often available from local planning agencies. For example, in the U.S., local metropolitan planning organizations and the national Energy Information Administration (EIA) create demographic and econometric models to forecast economic activity and other parameters, such as vehicle travel, housing (and construction) activity, and population growth. The second step is to use economic elasticities to convert overall spending levels into forecasts of changing consumption of functional units (e.g. kilograms of food or hours of entertainment, instead of dollars) for each product and service category. Such elasticities are

11


SEI WP 2012-05

available at the national level from analyses of the U.S. Consumer Expenditure Survey and could (in theory) be similarly developed using local expenditure trend data, where available. Applying this process to a base year consumption-based GHG inventory (with associated expenditure data) yields a forecast of both spending and GHG emissions associated with each category of consumption. (For further details of this two-step approach, including sources of the elasticities used, please see Appendix A). The exception to this two-step approach is for personal transport and home energy activity. For these activities, more specific forecasts are often available from planning agencies, such as for vehicle travel (in miles, or gallons of fuel) and home energy (households, household size, or energy use), and should be used where available. Forecasting intensity

Forecasting the GHGs per unit of activity is equally complex, if not more. Due to the complex, global supply chains for many types of products, assessing changes in the GHG intensity would, ideally, involve considering trends in energy supplies and production technologies globally and how these trends in different world regions would contribute to the emissions intensities of dozens or hundreds of product types. For starters, analysts can use existing national (e.g. U.S. EIA) or international (e.g., IEA) forecasts of practices for different primary materials (e.g., steel, aluminum, chemicals) to get a rough sense of expected improvements. However, approximating how these changes would translate into changes in the GHG-intensity of specific products (e.g., a car) requires further analysis to understand the input structure (e.g., an input-output table), or the pattern of interindustry relationships. A simplified approach could be to alter the GHG intensities of these primary, producing sectors, and re-run the input-output model to generate new GHG intensities for the consumer products (e.g., cars, clothing, furniture, appliances). Ideally, an additional step would be to adjust the input-output table itself, as current input-output tables cannot adequately model changes in sectors where future supply chains may differ considerably (e.g., if future buildings were to use much more of one material and much less of another). Although we do not explore methods to do so here, this remains an important area for further research. In the meantime, we suggest using a static input-output table, and do not see this as a significant limitation in shorter- to medium-term analyses, since structural decomposition analysis has shown (at least in the recent past) that input structure has had a much smaller effect on GHGs associated with consumption than have other factors, especially consumption levels and the direct GHG-intensity of production (Baiocchi and Minx 2010). Lastly, forecasts for vehicle efficiencies and household energy can generally be taken from national or (if available) local projections, which are commonly available (e.g., the U.S. EIA produces forecasts of vehicle fuel economies and building energy demands that can be used to construct a baseline). Together, by forecasting changes in the activity and intensity drivers, the base year consumption-based GHG inventory can be translated into a forecast of the emissions associated with consumption. The approach described above applies directly for constructing the baseline scenario, but the approach for constructing the mitigation scenario is largely similar, but for specific changes that are posed. For example, instead of forecasting the baseline level of meat consumption (an activity driver), the analyst would suppose that meat consumption shifted in favor of (or at the expense of) another food (e.g., legumes). To identify what such changes might be, the next section discusses low-GHG measures and policies. 12


SEI WP 2012-05

5. CATALOGUE OF LOW-GHG MEASURES AND POLICIES Historically, assessments of GHG abatement potential have focused primarily on technical options to reduce GHGs in the world regions emitting them. Options are categorized by sector (e.g., energy supply, transport, industry, buildings) and include technologies such as renewable energy generation (e.g., wind power), efficient vehicles, and efficient appliances. This categorization, or typology, of options has remained relatively constant since the foundational work of the IPCC (1996b), including recent global assessments of GHG abatement potential (Metz et al. 2007; McKinsey & Company 2010). In most cases, this sector-based typology is well-suited to discussions of global climate policy that (rightly) focus on making a substantial transition to a low-carbon economy through national-scale policies, such as emissions trading systems. However, efforts to avoid dangerous climate change and limit warming to 2° or 1.5°C may require additional measures beyond those normally considered in existing assessments. In particular, systemic behavior change may be needed – not only to implement the technologyfocused solutions already well-established (e.g., electric cars) – but also to shift away from consumption behaviors where technological solutions are less effective or where underlying resource constraints (e.g., land and water availability) may limit continued growth in global consumption. A focus on consumption has also returned to the international agenda, as parties to the UNFCCC climate agreement in Cancun, Mexico, in 2010 called for “more sustainable production and consumption and lifestyles” as a key component of a “paradigm shift towards building a low-carbon society” (UNFCCC 2011). We take a cue from prior efforts (IPCC 1996b) to expand the existing typology of GHG abatement measures to document those focused on consumption (and which, on balance, include a greater focus on consumer behavior). These measures might help support a transition to a low-carbon society and enhance the effectiveness of existing policy proposals, such as emissions trading systems. 12 We do not list the numerous individual, production-, sector-, and technology-oriented policies already well-chronicled by the IPCC (Metz et al. 2007) and others (McKinsey & Company 2010), such as electric cars or efficient appliances, even if uptake of those policies requires consumer behavior changes. Other researchers have documented and/or quantified a suite of consumption changes that would reduce GHGs (Jones and Kammen 2011; Tukker et al. 2010; Lebel and Lorek 2008; Dietz et al. 2009; Scott et al. 2009), but we have not yet seen a standard set of consumption-related measures such as that which has evolved on the production side from the IPCC and related efforts and which can be considered in scenario modeling exercises.

12

We recognize that this sole focus on consumption-based policies here brings some peril. For one, pursuing these strategies in isolation could lead to significant rebound effects that could significantly diminish their effectiveness (Alcott 2010; Ornetzeder et al. 2008). And too much of a focus on consumption could distract from the needed focus on economy-wide policies (e.g., emissions trading systems or other carbon pricing). Nevertheless, given that many of these measures are rarely considered, we find it important to document these as part of a broader effort to consider transitions to sustainable production-consumption systems (Lebel and Lorek 2008; Tukker et al. 2010; Scott 2009).

13


SEI WP 2012-05

Typology of behaviors and measures

Table 3 lists behavior changes and measures to reduce GHGs associated with consumption. Table 3: Behaviors and measures to reduce GHGs associated with consumption Category

Behavior

Behavior description

Examples of supporting policy measures that have been adopted

Other possible policy measures or resources

Shift diets

Shift consumption from high-GHG to low-GHG foods Reduce overall food consumption

City of San Francisco resolution adopting meatless Mondays 13

Voluntary or mandated product labeling 14, education, 15 and public health efforts 16

Reduce food waste

Reduce waste of edible food in homes and businesses to reduce need for new food

Education campaigns 17

Regulations or outreach to support packaging to extend product life; re-use of catering waste for animal feed 18

Recycle or convert food waste

Recycle food into compost or convert food into energy via methane capture

Programs or regulations (e.g., EU Landfill Directive) that divert food waste to composting or anaerobic digestion facilities

US EPA maintains a tool to quantify GHG benefits of composting 19

Increase product longevity

Increase care and maintenance of products or purchase more durable products 20

Bans and fees adopted on some disposable goods (e.g., shopping bags) to encourage longer-lived alternatives

Extended Producer Responsibility (Product stewardship) efforts could be expanded to focus on durability

Increase use of product-service systems

Intensify the use of goods by shifting consumption from individually owned goods to shared goods

Government support for tool-sharing in Portland, OR 21

Government support through tax policy, pilot projects 22

Food

Goods

13

See http://www.sfbos.org/ftp/uploadedfiles/bdsupvrs/bosagendas/materials/bag040610_100413.pdf. For example, the Carbon Trust (2008) has developed a voluntary carbon footprinting standard in partnership with the UK Government. The World Resources Institute and World Business Council for Sustainable Development have developed a Product Accounting and Reporting Standard (WBCSD and WRI 2010). The Oregon Global Warming Commission (2011) has recommended standards, incentives, and/or mandates for product footprinting/labeling. 15 For example, the Oregon Global Warming Commission (2011) has recommended that the State “develop and disseminate information: easy-to-use life cycle metrics for different food types.” 16 For example, see Willett and Skerrett (2001) for diet recommendations that would also be lower GHGs (Stehfest et al. 2009). 17 For example, see http://www.lovefoodhatewaste.com/. 18 For an assessment of low-GHG food behaviors, including several related to waste, see Garnett (2011). 19 For example, see http://www.epa.gov/lmop/ and EPA (2010). The EPA is reportedly expanding this tool to address other organic waste conversion technologies (e.g., anaerobic digestion). 20 For an extended discussion of product lifespans, see Cooper (2008). For a case study of longevity in the clothing sector, see Carbon Trust (2011). For a discussion of some of the challenges and emissions or ecological “payback” periods, see van Nes and Cramer (2006) 21 See http://www.portlandoregon.gov/bps/article/402810. 22 For a discussion of product-service systems, including challenges thereof, see Mont and Lindhqvist (2003) and Tukker and Tischner (2006). 14

14


Category

SEI WP 2012-05

Behavior




Purchase lowGHG goods

Shift consumption from high-GHG to low-GHG goods

National, state and local governments have implemented environmentalpurchasing criteria, though few focused on GHGs

Voluntary product labeling efforts;14 partnerships with manufacturers to assess and reduce GHG-intensity; 23 business tax credits 24

Purchase fewer goods

Purchase less “stuff”

Several educational programs 25

Smaller homes may have a side benefit of needing less stuff. 26

Recycle goods

Recycle goods at end of life

Hundreds of recycling programs, including efforts to work with manufacturers and/or require product takeback.

Require government or private sector efforts to implement infrastructure

Purchase lowGHG services

 Shift consumption from high-GHG to low-GHG services, or shift from direct consumption of products to “productservice” systems, as discussed above.  Service-sector companies implement the same behaviors listed above under “Goods” in the business setting for their purchases of “intermediate goods”

[None identified]

 Efforts to reduce building-sector emissions would also reduce emissions intensity of services. Many studies document these technologies and measures. 27  For intermediate goods, same “Other Possible Policy Measures” apply as under Goods, above

Choose smaller homes and buildings

Residences (and businesses) choose homes and buildings that are, on average, smaller.

Building codes that favor multi-family development implicitly favor smaller homes

Increase longevity of homes and buildings

Retrofit older structures for functional and energy performance. Choose and design new structures to be adaptable to a variety of uses.

 Policies focused on building preservation and longevity may help reduce embodied GHGs but don’t necessarily reduce life-cycle GHGs 28  Pay as you Save schemes to fund retrofits (a number of successful schemes in Europe)

Services

Construction

23

Several European countries have experimented with covenants with industry to reduce energy consumption, but not generally life-cycle emissions. In addition, a number of voluntary government-industry partnerships exist in the U.S. (e.g., EPA’s Climate Leaders and EnergySTAR programs), but these too generally do not focus on lifecycle emissions. Walmart has a sustainable packaging scorecard. 24 For example, the State of Oregon has a business energy tax credit to benefit businesses that lower energy use, but it is not focused on life-cycle emissions. 25 For example, King County, Wash., has had a program that encouraged consumers to “give experiences instead of stuff” at the holidays. 26 For discussion, see Oregon DEQ (2010). 27 For example, see Metz et al (2007), IEA (2010) or McKinsey (2009). 28 Trade-offs exist between the embodied GHGs in older buildings and the opportunities for increased energy and location efficiency of new construction.

15


Category

Behavior


SEI WP 2012-05



 Land use planning  Public transportation infrastructure  Road and parking pricing

Public health policy to encourage walking /cycling

Personal transportation Shift modes of transport

Shift from singleoccupant vehicle travel to carpool, public transit, or bike/pedestrian travel

Reduce number and/or length of vehicle trips

Take fewer or shorter trips in vehicles

Reduce air travel

Reduce personal or business air travel by substituting videoconferencing, train travel, or local travel

[None identified]

May require government and/or private sector support to develop infrastructure for highdefinition videoconferencing.

Consumers operate appliances with low intensity (e.g., limiting use of hot-water washing and drying) 29

[None identified]

Efforts to promote and facilitate lower-intensity practices by coordinating consumers and manufacturers 30

Home energy Implement behaviors to reduce GHG intensity of appliance use

We find it striking how few policies have supported the behaviors listed in Table 3. In part, this is because regional, national, and international approaches to reducing GHG emissions have focused primarily on measures that affect the production, or supply of energy. Still, the line between production and consumption-based policies is not as sharp as perhaps we have implied above, as some policies may address both. For example, comprehensive climate policy, whether an economy-wide emissions trading system or a carbon tax, would result in a price on carbon that would not only incent changes in energy production and supply but also encourage consumers, through changes in prices, to adopt many of the behaviors above. Indeed, most analysts agree that an economy-wide policy (emissions trading system or carbon tax) is far more efficient and practical than a large suite of policies acting on individual producers and/or consumers. Accordingly, carbon pricing is a key overarching policy that extends across all of the categories of consumption above. 31 Still, under such an economywide policy, many of the behaviors in Table 3 would help, and in some cases, be necessary, to reduce GHGs. Besides pricing, other overarching policies should perhaps also be considered. For example, measures under which affluent consumers forgo personal incomes (and, by extension, expenditures) for other benefits, such as increased leisure time, should perhaps also be considered (Druckman and Jackson 2010).

29

For example, washing clothing on lower heat settings or line-drying clothing saves significant energy (Carbon Trust 2011). 30 For example, achieving significant uptake of low-temperature clothes washing would require coordination of appliance and detergent manufacturers (to enable low-temperature washing), retailers (to provide information), and consumers (to adopt the practice) (Carbon Trust 2011). 31 The Oregon Global Warming Commission (2011) has recommended a carbon tax to address emissions associated with consumption.

16


SEI WP 2012-05

6. PILOTING THE APPROACH FOR A MAJOR U.S. CITY To test our approach, we developed a scenario of low-GHG consumption for a major U.S. city: Seattle, Wash. The City of Seattle has long been a leader in city-level action and planning to address climate change. In 2005, Seattle led the effort to convince over 1,000 U.S. mayors to commit their cities to reducing their GHG emissions by 7% from 1990 levels by 2012. The City is now in the midst of an ambitious update to its own climate action plan, driven by a new goal for the community to be net carbon-neutral by 2050. Because of its national leadership, ambitious long-term goals, and strong foundation of existing emissionreducing activities in buildings, transportation, and waste, Seattle is a good candidate to demonstrate how low-GHG consumption could further contribute to GHG abatement. Furthermore, the Seattle area has recently been the subject of two related efforts. The first, a complete consumption-based GHG inventory for the region (Erickson et al. 2012), provides a snapshot of the emissions associated with consumption in the region in 2008. The second, a deep GHG-reduction scenario in Seattle’s buildings, transport, and waste sectors, provides a long-term scenario for these sectors, plus additional demographic projections that can support forecasts of consumption behaviors (Lazarus et al. 2011). Together, these studies provide a solid foundation on which to construct a long-term scenario of low-GHG consumption. GHGs associated with current consumption in Seattle

To estimate the GHGs associated with current consumption in Seattle, we build from a recent effort in the Seattle region to conduct a consumption-based GHG inventory for all goods and services consumed in the King County, including Seattle, in 2008 (Stanton, Bueno, Cegan, et al. 2011; Erickson et al. 2012). To compile this estimate, we scaled King County’s consumption-based inventory to Seattle according to the ratio of the two regions’ populations, but we substituted Seattle’s figures for emissions from use of energy in the home and in personal vehicles. 32 King County’s consumption-based inventory, and therefore also our estimate here, defines consumption as the sum of consumer expenditures, government purchasing, and business capital investment (or net accumulations to business inventory). Under this definition, which is also termed “final demand” and is based on common practice for national economic accounts, emissions associated with business operation (e.g., energy, intermediate goods, food) are only attributed to Seattle to the extent those businesses’ services are consumed by Seattle residents or government. 33 As displayed in Figure 2, we estimate that Seattle’s consumption-based emissions are about 26 tons CO2e per resident, several times more than the per capita emissions released from within Seattle itself (City of Seattle 2009).

32

To match King County’s method, we also should have included emissions associated with operation of government buildings and vehicles in Seattle. Since these figures were not readily available, we did not include them here, but they are not expected to contribute substantially to our estimate of Seattle’s consumption-based emissions. 33 For example, emissions associated with the headquarters of Starbucks or Amazon.com would largely (though not entirely) be attributed to consumption outside Seattle and not included in Seattle’s consumption-based inventory.

17


SEI WP 2012-05

Figure 2: GHG emissions associated with consumption in Seattle, 2010 (tCO2e/person) 3.3 6.4

Personal Transportation

2.0

Home Energy Freight Food 1.2

3.6

Other Goods Services Construction

2.5

Other 4.2

3.4

Total = 26 tons CO2e / person

Seattle is unique in one key respect: it has a very low-carbon electricity supply, provided almost entirely by hydroelectricity. Accordingly, emissions associated with home energy, at 1.2 tCO2e per person, are much smaller than in the rest of the nation, where they are at least 3.9 tCO2e per person (EPA 2011). 34 Baseline forecast of GHGs associated with consumption

As described in Section 3, a baseline scenario is a forecast of emissions in the absence of policies specifically designed to reduce GHGs. The baseline scenario, and its underlying drivers, form the basis for the mitigation scenario where an alternate, low-GHG path is taken. As recommended in Table 2, we construct the baseline by considering both demand- and supply-side trends. We forecast changing consumption (demand) of goods and services based on projected income growth in the Seattle region (Conway 2006), an assumption about how much of that extra income is saved versus spent on goods and services (Taylor and Houthakker 2009), and the expenditure elasticities of Girod and de Haan (2010) to characterize how the consumption “basket” will change over time, as discussed earlier in this report. Projected housing and travel forecasts are taken from regional studies conducted by the Puget Sound Regional Council (PSRC 2010; 2008). We forecast changing baseline production practices (supply) of goods and services by using EIA forecasts for the emissions intensity from buildings (CO2e per square foot) and specific industry sectors (CO2e per unit of output), which we then process in the underlying inputoutput model that generated that consumption-based GHG inventory (Stanton, Bueno, Cegan, et al. 2011) to yield estimates of the future emissions-intensities of products and services.35 Trends in building and vehicle efficiency are also taken from EIA forecasts (EIA 2010).

34

We say “at least” because the EPA’s national inventory assigns to households only the emissions associated with burning the fossil fuels (e.g., coal, natural gas) used to provide energy in the home. A consumption-based inventory also counts the pre-combustion emissions, those associated with extracting and processing the fuels involved. 35 We assume that the production structure, or the pattern of intermediate consumption of goods and services approximated by the input-output tables, does not change.

18


SEI WP 2012-05

We report most results in per-person terms, a practice that helps make these results comparable to other cities. Reporting results on a per-person basis also helps avoid penalizing cities for increasing density, a practice that may reduce global GHGs if new residents otherwise would have lived in a more GHG-intensive area, such as (in most cases) a lowerdensity suburb (Kennedy et al. 2009). At the same time, we also use existing forecasts of population in the region to report absolute GHGs. Figure 3 shows the forecast of baseline GHGs between 2010 and 2030. Overall, per capita emissions are expected to rise modestly from about 26 tCO2e/person in 2010 to 29 tCO2e/person in 2030, as forecast increases in consumption outweigh forecast improvements in GHG-intensity of production. Figure 3: Forecast GHG emissions from consumption in Seattle, baseline scenario 35 30 Other

25

Construction Services

GHGs per person, 20 tCO2e (Consumption 15 Perspective)

Other Goods Food Freight

10

Home Energy Personal Transportation

5 0 2010

2015

2020

2025

2030

However, not all categories of consumption contribute equally to this rising trend. Figure 4 displays the change within each category between 2010 and 2030. Figure 4: Change in baseline GHGs per capita, 2010-2030 1.0 0.8 0.6 0.4

Change in BAU GHGs per 0.2 person, 2010 to 2030 (tCO2e) (0.2)

Personal Home Energy Transportation

Food

Freight

(0.4) (0.6)

19

Other Goods

Services

Construction

Other


SEI WP 2012-05

Two categories, services and other goods, are expected to contribute most of the growth in GHGs between now and 2030. (Trends in the Other category are more speculative, as discussed further below.) This is largely because, as incomes increase, people tend to devote more of their new income to goods (such as clothing and home furnishings) and services (such as entertainment and education) than to home energy or gasoline for personal transportation. Personal transportation represents a special case: personal vehicle travel in the Seattle area has reached a saturation point, such that per capita vehicle travel is actually declining, and vehicles are forecast to become significantly more efficient by 2030, even in the absence of new climate policy. 36 Despite the significant emissions associated with goods and services, few local climate initiatives focus on them. Instead, local governments more commonly focus on transportation, building energy, and waste on which they have more direct influence (Arup 2011). In the mitigation scenario discussed next, we will explore the potential GHG reductions in these other categories due to behavior shifts. Note that all of these estimates are uncertain and driven by assumptions about future income growth and that consumers will spend new income in the future in similar ways as they have in the past. Particularly uncertain, however, are the forecasts of emissions changes for construction and other, which are driven substantially by business capital investment for which we had neither adequate stand-alone forecasts nor elasticities to construct estimates based on changing incomes in the Seattle area. As a result, we largely assumed that expenditures in these sectors would scale with incomes and would not be subject to as much increasing “quality” (higher price per functional unit) as would consumer goods such as clothing and home furnishings. Further research into trends of construction and business capital investment would be beneficial. Low-GHG consumption mitigation measures

In the mitigation scenario, Seattle (and, as will be described below, the world) departs from “business-as-usual” to pursue a dramatically different path. Table 4 lists the measures applied for both shifts in activity. Whenever possible, the measures were selected to represent aggressive, but plausible, uptake of behavior change, technology shifts, and fuel shifts. However, such judgments are highly subjective, as there is little historical or empirical precedent for the type of transition envisioned, especially for behavioral measures such as extreme diet shifts and significant (50%) reduction in purchasing of some items (e.g., clothing).

36

The decline in personal transportation emissions would be even greater if not for a forecast increase in GHGs associated with personal air travel, due to increased air travel associated with rising incomes.

20


SEI WP 2012-05

Table 4: Key measures applied in the mitigation scenario for 2030 Food

Goods

Intensity 37 (production, or supply)

Activity (behavior, consumption or demand)

 GHG intensity of food declines by 0 to 20% from the present, depending on the food type  Grazing land no longer needed for ruminants reverts to grassland, sequestering about 1.8 tCO2e annually per hectare 38  The GHG intensity of goods declines 25% to 50% from the present, depending on the product

 Reductions in beef, pork, poultry, and dairy of 85%, 81%, 61%, and 79%, respectively 39  6% overall reduction in food consumption compared to present

Services

 GHG intensity of services declines 20% to over 50% from the present, depending on the service

Construction

 GHG intensity of construction declines over 30% from the present

Personal transportation

 Light-duty vehicles use about 55% less energy per mile than in the present (Lazarus et al. 2011) and about 30% less GHGs per unit of energy due to assumed 50% penetration of low-GHG biofuels  Air travel 39% less energy-intensive per passenger-mile (Greene and Plotkin 2011), 35% less GHGs per unit of energy due to low-GHG biofuels  Seattle’s electricity remains zero-carbon

Home energy

Other Freight

 GHG intensity of these items declines 47% from the present  Freight transport uses about 38% less energy per mile than in the present and about 28% less GHGs per unit of energy, due mainly to biofuels (Lazarus et al. 2011)

 Home furnishings, clothing, and home and office supplies last twice as long (reducing purchases by half relative to baseline)  Healthy lifestyles enable purchases of healthcare supplies to hold constant per person (rather than increase)  Consumption of services increased with income (same as in baseline)  Consuming of entertainment and “other” services increases, as income freed up by reduced expenditures on goods and services is redirected to these services  Residential construction activity continues at present rate, does not climb with income (a 9% decline relative to baseline)  Per-resident light duty vehicle use declines 30% from present (Lazarus et al. 2011)  Per-resident air travel stops rising, declining 25% from BAU in 2030 (IEA 2010)

 Aggressive building retrofits as described in Lazarus et al (2011) and based on assumptions of an LBNL analysis (Adelaar et al. 2008)  Living area per person stays constant, instead of rising gradually in the baseline  No change from baseline 40  Reduces in proportion to reduction in purchase of goods and food

37 All changes in GHG-intensity of production of food, goods, services, and construction are based on application of an input –output model (Stanton, Bueno, Cegan, et al. 2011) to improvements forecast in the U.S. economy under an economy-wide emissions limit (EIA 2009) . 38 Bouman (2005) provides land area requirements for feed production and grazing for different animals. We assume that land no longer needed for grazing reverts to grassland at the average U.S. rates in EPA (2011). In addition, given that we assume that not all meat consumption is replaced by plant-based sources, additional land would be freed up from production of feed crops. We do not quantify the potential GHG benefits of this freed-up land. 39 Developed by assuming that diets change to the Harvard Medical School’s Healthy diet (Willett 2001), using the method of Stehfest et al (2009). Reduction in consumption of these animal products is substituted, on a calorie-forcalorie basis, up to a 2,000 calorie diet. This results in a 6% decrease in overall food consumption. 40 Much of the Other category is industrial equipment. Arguably, some of this could be avoided if product consumption decreases. On the other hand, the transition to low-carbon transport infrastructures or buildings may require additional investments. We assume no change from baseline.

21


SEI WP 2012-05

Low-GHG consumption scenario results

Together, the mitigation measures described in Table 4 dramatically reduce the GHG emissions associated with Seattle’s consumption to about 15 t CO2e per person in 2030, 47% less than the forecast baseline emissions of 29 t CO2e per person. Figure 5: Emissions in the baseline and mitigation scenarios, Seattle 35 30 25 GHGs per person, 20 tCO2e (Consumption 15 Perspective)

Baseline Scenario Mitigation Scenario

10 5 2010

2015

2020

2025

2030

Figure 6 displays emissions in the mitigation scenario by sector of consumption. The category with the greatest (absolute) reductions is personal transportation, where over 3 tCO2e reductions are driven by decreases in the energy-intensity (increasing fuel economy) of transportation vehicles (especially cars, but also buses and airplanes), decreasing GHGintensity of the fuels (greater use of low-GHG biofuels, or increasing penetration of hydrogen fuel cells), and reduced vehicle activity (due to reduced trip length and mode shifting). The second biggest category of reductions, with a decrease just under 3 tCO2e, is “other goods”, with reductions due both to decreased GHG-intensity of production as well as reduced consumption of several goods, such as home furnishings, clothing, and household supplies.

22


SEI WP 2012-05

Figure 6: Emissions in the mitigation scenario for Seattle, by category of consumption 30

25 Other 20


GHGs per person tCO2e 15 (Consumption Perspective)


10


5

0 2010

2015

2020

2025

2030

Broadly speaking, GHG reductions are due either to shifts in intensity or shifts in activity (Table 4). Reductions in intensity are due largely to production practices driven by national (or international) trends, such as the GHG-intensity of factories and energy supplies. Reductions in activity are due largely to shifts in behavior or consumption, such as reduced meat consumption, or declining purchases of home furnishings. Some measures, especially those related to home energy and personal transportation, could in some cases be considered either an activity or an intensity measure (or both). For example, retrofitting a home for improved energy performance is an action that (in most cases) can only be undertaken by a homeowner (and might therefore be considered a behavior), but can only be done if technology (such as efficient appliances, lighting, and heating systems) is available. To help understand the relative contribution of intensity- and activity-based measures, Figure 7 displays the successive effects of several measures (or bundles of measures). (Note that because these measures are applied successively (to avoid double-counting), the contribution of measures applied later, such as consumption shifts, will appear somewhat smaller than if they had been applied earlier.) This analysis suggests: 

Behavioral measures can significantly reduce GHGs associated with consumption. Diet shift, extended product lifespans, and other shifts in purchasing reduce emissions by about 2.7 tCO2e/capita, or 10% of emissions associated with an average Seattle resident. Home energy retrofits and shifts in vehicle travel (reduced personal vehicle travel, air travel) could reduce emissions by another 0.9 tCO2e/capita (after accounting for increased efficiency of these travel modes).



Nevertheless, the collective potential of these actions is still less than the potential shifts in GHG-intensity of production due to national or international action. This finding holds true even if the behavior-related actions are considered first, before action at these broader scales. This finding suggests that behavioral measures can supplement, but not replace, larger-scale action.



The rebound effect, while potentially significant, can be minimized if redirected to the lowest GHG-intensity services, such as entertainment services, with relatively 23


SEI WP 2012-05

much lower GHGs per dollar of expenditure. Redirecting spending from the avoided goods, home energy, vehicle fuels, and airplane tickets to low-GHG services leads to an increase in emissions of 1.1 tCO2e/person, or about 17% of the 6.1 tCO2e/person savings associated with these behaviors. Figure 7: Contribution of intensity- and activity-based measures to GHG abatement in the mitigation scenario in 2030 30

25

20 GHGs per person, tCO2e (Consumption 15 Perspective) 10

5

0 Baseline

Decreased intensity (goods and services)

Efficient vehicles

VMT reduction Home energy Consumption retrofits shifts (activity)

Rebound

Net (Mitigation Scenario)

Even with this aggressive set of measures that affect both local consumption and national (and international) production, however, per-person emissions associated with consumption by Seattle residents in 2030 would still be 15 tCO2e per capita. This is still far above the needed global average of about 1 tCO2e per person needed this century to maintain a “likely” chance of limiting warming to 2° C and averting the worst impacts of climate change (Rogelj et al. 2011). This suggests that more ambitious actions will be needed on both the production and consumption sides. 7. KEY LIMITATIONS AND UNCERTAINTIES We have described a particular approach to forecasting consumption-based emissions and creating scenarios of GHG emissions reduction based on behavior change. Our approach relies strongly on existing economic forecasts (e.g., of income, or vehicle travel) as well as on continuation of recent trends (e.g., how consumers will spend their money as average incomes increase, or how the GHG-intensity of production will change). This approach allows for a jurisdiction with a consumption-based GHG inventory to conduct a scenario analysis of future pathways of emissions. The key innovation of this approach is that it incorporates a dynamic, scenario planning framework that includes a baseline forecast of both consumer behavior and national and international production practices. Measuring GHG abatement potential against a baseline forecast is standard practice in production-based abatement studies (Metz et al. 2007), yet few assessments of the contribution of behavior change to GHG mitigation yet take this approach. Despite this innovation, the approach used here has several limitations related to the forecast of baseline emissions; addressing these limitations could be the subject of future research:

24


SEI WP 2012-05



It assumes that consumer behavior will continue evolving in the future as it has in the past, relying on income elasticities from the economic literature (Taylor and Houthakker 2009) and conversion of these elasticities to consider increasing quality (price) (Girod and de Haan 2010). Such an approach cannot foresee changing trends in consumer behavior that are unrelated to income, such as rapid shifts in technologies (e.g. mobile devices).



It assumes that the production structure – the supply chains of producing industries, and sometimes called the “production recipe” (Lenzen et al. 2012) – is fixed over time. Methods to assess changes in the input-output structure over time do exist, whether in a baseline or alternative scenario, but are complicated, speculative, and not widely employed (Scott et al. 2009).



It assumes that international GHG intensities or production will evolve just as national GHG intensities (EIA 2010), which themselves are quite uncertain.

The approach is also limited by other factors not directly related to the baseline forecast, such as errors and uncertainties in adapting a consumption-based GHG inventory. Consumptionbased inventories at the community scale are not yet widely employed, and methods are still emerging. Many (including the one adapted here) are built from national economic consumption data scaled to local circumstances, which may not accurately reflect local consumption practices (Erickson et al. 2012). Furthermore, economic units may not correlate as well with emissions as physical units (Girod and de Haan 2010). Finally, we assume that dramatic behavior change (e.g., 85% reduction in beef consumption) could be mobilized across an entire city’s population. Such a change may not be feasible – at the least, little attention has been devoted to the mechanisms and processes to bring about many types of behavior change, even as research is increasingly focusing on the topic (Allaway et al. 2012; Timmer et al. 2009). 8. NEXT STEPS AND CONCLUSIONS In this paper, we have demonstrated an approach to constructing long-term scenarios of a shift to low-GHG consumption. In particular, we have documented, evaluated, and piloted methods to constructing baseline and mitigation scenarios for GHGs associated with all consumption in a community. A shift to low-GHG consumption may be needed to help limit warming to 1.5 or 2.0° C and avert the worst impacts of climate change. In addition to the development of a methodology, several interesting findings have emerged: 

GHGs from goods and services may grow faster than GHGs from personal transportation and home energy. Driven by increasing incomes, growth in expenditures on goods and services is projected to lead to GHGs from these categories increasing more rapidly than from personal transportation and home energy, where emissions are expected to remain flat (or decline) on a per capita basis.



Behavior change can contribute substantially to GHG abatement through 2030. Our pilot analysis for the City of Seattle shows that shifts in purchasing of goods, services, and food can contribute about 10% reduction from baseline in 2030, on par with the effect of more-efficient personal vehicles. Implementing these shifts would require broad societal shifts and collective action for which few mechanisms and policies have been demonstrated.

25




SEI WP 2012-05

Despite this aggressive scenario of measures, GHGs in 2030 would only drop to 15 tCO2e per person for Seattle, a long way from the needed global goal of about 1 t CO2e per person by 2050.

Further research is needed in several areas, especially on policies to encourage collective action. Achieving the levels of behavior change modeled in this analysis would require extensive efforts, likely far beyond what has been undertaken previously.

ACKNOWLEDGMENTS This paper was prepared with SEI’s core project funding. We thank former SEI colleagues Elizabeth A. Stanton and Katy Roelich for discussions on the model logic and the catalogue of low-GHG policies, respectively. A version of this paper was presented at the Behavior, Energy, and Climate Change (BECC) conference in Washington, D.C., in December 2011.

26


SEI WP 2012-05

REFERENCES Adelaar, M., Pasini, M., de Buen, O. and Selkowitz, S., 2008. Green Building in North America: Paper 1: Green Building Energy Scenarios for 2030. Commission for Environmental Cooperation. http://www.cec.org/Page.asp?PageID=30101&ContentID=17347&SiteNodeID=468. Alcott, B., 2010. ‘Impact caps: why population, affluence and technology strategies should be abandoned’. Journal of Cleaner Production, 18(6). 552–60. doi:10.1016/j.jclepro.2009.08.001. Allaway, D., Kochan, L. and Grabham, C., 2012. Literature Review: Alternatives for Advancing Sustainable Production and Consumption Through Government Programs and Policies. Oregon Department of Environmental Quality, Portland, OR. http://www.deq.state.or.us/ lq/pubs/docs/sw/2050vision/LitReviewAlternativesAdvancingSustainableProduction.pdf. Arup, 2011. Climate Action in Megacities: C40 Cities Baseline and Opportunities: Version 1.0. Arup for C40 Cities. Audsley, E., Brander, M., Chatterton, J., Murphy-Bokern, D., Webster, C. and Williams, A., 2009. How Low Can We Go? An Assessment of Greenhouse Gas Emissions from the UK Food System and the Scope to Reduce Them by 2050. Cranfield University for WWF-UK and the Food Climate Research Network, Bedford, UK. http://www.wwf.org.uk/research_centre/ research_centre_results.cfm?uNewsID=3678. Audsley, E., Chatterton, J., Graves, A., Morris, J., Murphy-Bokern, D., Pearn, K., Sandars, D. and Williams, A., 2010. Food, Land and Greenhouse Gases: The Effect of Changes in UK Food Consumption on Land Requirements and Greenhouse Gas Emissions. Cranfield University for the United Kingdom’s Government’s Committee on Climate Change, Bedford, UK. Baiocchi, G. and Minx, J. C., 2010. ‘Understanding Changes in the UK’s CO2 Emissions: A Global Perspective’. Environmental Science & Technology, 44(4). 1177–84. doi:10.1021/es902662h. Bioregional and the London Sustainable Development Commission, 2009. Capital Consumption: The Transition to Sustainable Consumption and Production in London. London, UK. http://www.bioregional.com/news-views/publications/capital-consumption/. Bouwman, A. F., Van der Hoek, K. W., Eickhout, B. and Soenario, I., 2005. ‘Exploring changes in world ruminant production systems’. Agricultural Systems, 84(2). 121–53. doi:10.1016/j.agsy.2004.05.006. Brohmann, B. and Barth, R., 2011. Policies to Promote Sustainable Consumption Patterns in Europe: How Effective Are Sustainable Consumption Policies in EU-27? EUPOPP. Öko-Institut e.V. - Institute for Applied Ecology; published by EUPOPP Partners, Darmstadt, Germany. http://www.oeko.de/oekodoc/1163/2011-323-en.pdf. C40 and ICLEI - Local Governments for Sustainability, 2012. Global Protocol for CommunityScale GHG Emissions (GPC): Version 1.0. C40 Cities Climate Leadership Group and ICLEI Local Governments for Sustainability In collaboration with: World Bank, UNEP, UNHABITAT, World Resources Institute. http://live.c40cities.org. Carbon Trust, 2011. International Carbon Flows. CTC793. The Carbon Trust, London, UK. http://www.carbontrust.co.uk/policy-legislation/international-carbon-flows/. Carbon Trust, 2008. Product Carbon Footprinting: The New Business Opportunity Pack. CTC744. London, UK. http://www.carbontrust.co.uk/Publications/pages/publicationdetail.aspx? id=CTC744.

27


SEI WP 2012-05

Chavez, A. and Ramaswami, A., 2011. ‘Progress toward low carbon cities: approaches for transboundary GHG emissions’ footprinting’. Carbon Management, 2(4). 471–82. doi:10.4155/cmt.11.38. Chertow, M. R., 2000. ‘The IPAT Equation and Its Variants’. Journal of Industrial Ecology, 4(4). 13–29. doi:10.1162/10881980052541927. City of Seattle, 2009. 2008 Seattle Community Greenhouse Gas Inventory. Office of Sustainability and Environment. https://www.seattle.gov/environment/documents/2008-community-inventoryfullreport.pdf. Conway, D., 2006. Puget Sound Economic and Demographic Forecast: Detailed Forecasts and Methodology. Dick Conway & Associates for the Puget Sound Regional Council. Cooper, T., 2008. ‘Slower Consumption Reflections on Product Life Spans and the “Throwaway Society”’. Journal of Industrial Ecology, 9(1-2). 51–67. doi:10.1162/1088198054084671. Davis, S. J. and Caldeira, K., 2010. ‘Consumption-based accounting of CO2 emissions’. Proceedings of the National Academy of Sciences, 107(12). 5687–92. doi:10.1073/pnas.0906974107. Dawkins, E., Roelich, K. and Owen, A., 2010. A Consumption Approach for Emissions Accounting - the REAP Tool and REAP Data for 2006. Stockholm Environment Institute, York, UK. http://www.sei-international.org/publications?pid=1916. Dietz, T., Gardner, G. T., Gilligan, J., Stern, P. C. and Vandenbergh, M. P., 2009. ‘Household actions can provide a behavioral wedge to rapidly reduce US carbon emissions’. Proceedings of the National Academy of Sciences, 106(44). 18452 –18456. doi:10.1073/pnas.0908738106. Druckman, A. and Jackson, T., 2010. ‘The bare necessities: How much household carbon do we really need?’ Ecological Economics, 69(9). 1794–1804. doi:10.1016/j.ecolecon.2010.04.018. Duchin, F., 2008. ‘Sustainable Consumption of Food: A Framework for Analyzing Scenarios about Changes in Diets’. Journal of Industrial Ecology, 9(1-2). 99–114. doi:10.1162/1088198054084707. Erickson, P., Stanton, E. A., Chandler, C., Lazarus, M., Bueno, R., Munitz, C., Cegan, J., Daudon, M. and Donegan, S., 2012. Greenhouse Gas Emissions in King County. Report commissioned by King County, Wash. Stockholm Environment Institute-U.S. Center, Seattle, WA. http://www.sei-international.org/publications?pid=2026. Garnett, T., 2011. ‘Where are the best opportunities for reducing greenhouse gas emissions in the food system (including the food chain)?’ Food Policy, 36(Supplement 1). S23–S32. doi:10.1016/j.foodpol.2010.10.010. Girod, B. and de Haan, P., 2010. ‘More or Better? A Model for Changes in Household Greenhouse Gas Emissions due to Higher Income’. Journal of Industrial Ecology, 14(1). 31– 49. doi:10.1111/j.1530-9290.2009.00202.x. Girod, B., de Haan, P. and Scholz, R. W., 2010. ‘Consumption-as-usual instead of ceteris paribus assumption for demand’. The International Journal of Life Cycle Assessment, 16(1). 3–11. doi:10.1007/s11367-010-0240-z. Greene, D. L. and Plotkin, S. E., 2011. Reducing Greenhouse Gas Emissions from U.S. Transportation. Pew Center on Global Climate Change. http://www.c2es.org/publications/reducing-ghg-emissions-from-transportation. Hertwich, E. G., 2008. ‘Consumption and the Rebound Effect: An Industrial Ecology Perspective’. Journal of Industrial Ecology, 9(1-2). 85–98. doi:10.1162/1088198054084635.

28


SEI WP 2012-05

Hertwich, E. G., 2005. ‘Life Cycle Approaches to Sustainable Consumption: A Critical Review’. Environmental Science & Technology, 39(13). 4673–84. doi:10.1021/es0497375. Hertwich, E. G. and Peters, G. P., 2009. ‘Carbon Footprint of Nations: A Global, Trade-Linked Analysis’. Environmental Science & Technology, 43(16). 6414–20. doi:10.1021/es803496a. ICLEI - Local Governments for Sustainability, 2009. ‘International Local Government Greenhouse Gas (GHG) Emissions Analysis Protocol (IEAP) - Version 1.0’. Intergovernmental Panel on Climate Change, 1996a. Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories (3 Volumes). Task Force on National Greenhouse Gas Inventories. http://www.ipcc-nggip.iges.or.jp/public/gl/invs1.html. Intergovernmental Panel on Climate Change, 1996b. Technologies, Policies and Measures for Mitigating Climate Change. R. T. Watson, M. C. Zinyowera, and R. H. Moss (eds.). IPCC Technical Paper 1, Working Group II. International Energy Agency, 2010. Energy Technology Perspectives 2010: Scenarios & Strategies to 2050. OECD Publishing, Paris, France. http://www.iea.org/publications/free_new_Desc.asp? PUBS_ID=2100. Jones, C. M. and Kammen, D. M., 2011. ‘Quantifying Carbon Footprint Reduction Opportunities for U.S. Households and Communities’. Environmental Science & Technology, 45(9). 4088–95. doi:10.1021/es102221h. Kaya, Y., 1990. Impact of Carbon Dioxide Emission Control on GNP Growth: Interpretation of Proposed Scenarios. IPCC Energy and Industry Subgroup, Response Strategies Working Group, Paris. Kennedy, C., Steinberger, J., Gasson, B., Hansen, Y., Hillman, T., Havránek, M., Pataki, D., Phdungsilp, A., Ramaswami, A. and Mendez, G. V., 2009. ‘Greenhouse Gas Emissions from Global Cities’. Environmental Science & Technology, 43(19). 7297–7302. doi:10.1021/es900213p. Lazarus, M., Erickson, P., Chandler, C., Daudon, M., Donegan, S., Gallivan, F. and Ang-Olson, J., 2011. Getting to Zero: A Pathway to a Carbon Neutral Seattle. Report for the City of Seattle Office of Sustainability and Environment. Stockholm Environment Institute-U.S. Center, with support from Cascadia Consulting Group and ICF International, Seattle, WA. http://www.seiinternational.org/publications?pid=1903. Lebel, L. and Lorek, S., 2008. ‘Enabling Sustainable Production-Consumption Systems’. Annual Review of Environment and Resources, 33(1). 241–75. doi:10.1146/annurev.environ.33.022007.145734. Lenzen, M., Kanemoto, K., Moran, D. and Geschke, A., 2012. ‘Mapping the Structure of the World Economy’. Environmental Science & Technology, . doi:10.1021/es300171x. McKinsey & Company, 2010. Impact of the Financial Crisis on Carbon Economics: Version 2.1 of the Global Greenhouse Gas Abatement Cost Curve. http://www.mckinsey.com/clientservice/sustainability/pdf/Impact_Financial_Crisis_Carbon_Ec onomics_GHGcostcurveV2.1.pdf. McKinsey & Company, 2009. Pathways to a Low-Carbon Economy: Version 2 of the Global Greenhouse Gas Abatement Cost Curve. McKinsey & Company. Metz, B., Davidson, O., Bosch, P., Dave, R. and Meyer, L., 2007. Climate Change 2007: Mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK and New York, NY. Retrieved June.

29


SEI WP 2012-05

Miller, R. E. and Blair, P. D., 2009. Input-output Analysis: Foundations and Extensions. 2nd ed. Cambridge University Press, Cambridge UK. Minx, J., Wiedmann, T., Wood, R., Peters, G. P., Lenzen, M., Owen, A., Scott, K., Barrett, J., Hubacek, K., Baiocchi, G., Paul, A., Dawkins, E., Briggs, J., Guan, D., Suh, S. and Ackerman, F., 2009. ‘Input-output analysis and carbon footprinting: An overview of applications’. Economic Systems Research, 21(3). 187–216. doi:10.1080/09535310903541298. Mont, O. and Lindhqvist, T., 2003. ‘The role of public policy in advancement of product service systems’. Journal of Cleaner Production, 11(8). 905–14. doi:10.1016/S0959-6526(02)00152-X. Oregon Department of Environmental Quality, 2010. A Life Cycle Approach to Prioritizing Methods of Preventing Waste from the Residential Construction Sector in the State of Oregon: Phase II Report. Quantis, Earth Advantage, and Oregon Home Builders Association for the Oregon DEQ, Portland. Oregon Department of Environmental Quality, 2007. Oregon DEQ Waste Prevention Strategy: Ten-year Framework and Short-term Plan. file://C:\Users\Victoria Clark\Documents\Zotero\References for Upload\EndNote_Library_Pete\My EndNote Library.Data\PDF\WastePreventionStrategy-1729550865/WastePreventionStrategy.pdf. Oregon Global Warming Commission, 2011. Report to the Legislature 2011. Oregon Department of Energy on behalf of the Oregon Global Warming Commission, Portland, OR, US. http://www.keeporegoncool.org/content/report-legislature-2011. Ornetzeder, M., Hertwich, E. G., Hubacek, K., Korytarova, K. and Haas, W., 2008. ‘The environmental effect of car-free housing: A case in Vienna’. Ecological Economics, 65(3). 516– 30. doi:10.1016/j.ecolecon.2007.07.022. Peters, G. P., 2008. ‘From production-based to consumption-based national emissions inventories’. Ecological Economics, 65(1). 13–23. doi:10.1016/j.ecolecon.2007.10.014. Peters, G. P. and Hertwich, E. G., 2008. ‘CO2 Embodied in International Trade with Implications for Global Climate Policy’. Environmental Science & Technology, 42(5). 1401–7. doi:10.1021/es072023k. Peters, G. P., Minx, J. C., Weber, C. L. and Edenhofer, O., 2011. ‘Growth in emission transfers via international trade from 1990 to 2008’. Proceedings of the National Academy of Sciences of the United States of America, 108(21). 8903–8. doi:10.1073/pnas.1006388108. Puget Sound Regional Council, 2010. Transportation 2040: Toward a Sustainable Transportation System. http://www.psrc.org/transportation/t2040. Puget Sound Regional Council, 2008. Vision 2040: The Growth Management, Environmental, Economic, and Transportation Strategy for the Central Puget Sound Region. Seattle, WA. Ramaswami, A., Hillman, T., Janson, B., Reiner, M. and Thomas, G., 2008. ‘A Demand-Centered, Hybrid Life-Cycle Methodology for City-Scale Greenhouse Gas Inventories’. Environmental Science & Technology, 42(17). 6455–61. doi:10.1021/es702992q. Rogelj, J., Hare, W., Lowe, J., van Vuuren, D. P., Riahi, K., Matthews, B., Hanaoka, T., Jiang, K. and Meinshausen, M., 2011. ‘Emission pathways consistent with a 2[thinsp][deg]C global temperature limit’. Nature Clim. Change, 1(8). 413–18. doi:10.1038/nclimate1258. Scott, K., 2009. A Literature Review on Sustainable Lifestyles and Recommendations for Further Research. Stockholm Environment Institute, Stockholm. http://www.seiinternational.org/publications?pid=1219. Scott, K., Barrett, J., Baiochhi, G. and Minx, J., 2009. Meeting the UK Climate Change Challenge: The Contribution of Resource Efficiency. Report commissioned by WRAP (Waste & Resources Action Programme). Stockholm Environment Institute and the University of Durham 30


SEI WP 2012-05

Business School, Banbury, OX, UK. http://www.wrap.org.uk/wrap_corporate/ annual_conference/resource_efficiency.html. Stanton, E. A., Bueno, R., Ackerman, F., Erickson, P., Hammerschlag, R. and Cegan, J., 2011. Greenhouse Gas Impacts of Oregon’s Consumption: Technical Report. Stockholm Environment Institute (U.S.) for the Oregon Department of Environmental Quality. http://www.deq.state.or.us/lq/pubs/docs/ConsumptionBasedGHGEmissionsInventoryORTechni calReport.pdf. Stanton, E. A., Bueno, R., Cegan, J. and Munitz, C., 2011. King County Community Greenhouse Gas Emissions Inventory – Consumption Methodology: Technical Report. Stockholm Environment Institute-U.S. Center for the King County Department of Natural Resources and Parks, Somerville, MA. http://sei-us.org/publications/id/432. Stehfest, E., Bouwman, L., van Vuuren, D., den Elzen, M., Eickhout, B. and Kabat, P., 2009. ‘Climate benefits of changing diet’. Climatic Change, 95(1-2). 83–102. doi:10.1007/s10584008-9534-6. Taylor, L. D. and Houthakker, H. S., 2009. Consumer Demand in the United States: Prices, Income, and Consumption Behavior. Springer Verlag. Timmer, V., Prinet, E. and Timmer, D., 2009. Sustainable Household Consumption: Key Considerations and Elements of a Canadian Strategy. OneEarth for the Consumers Council of Canada and the Office of Consumer Affairs, Industry Canada, Toronto, Canada. http://www.consumerscouncil.com/index.cfm?pagePath=Research/Recent_Reports/Request_for _information&id=18258. Tukker, A., Cohen, M. J., Hubacek, K. and Mont, O., 2010. ‘The Impacts of Household Consumption and Options for Change’. Journal of Industrial Ecology, 14(1). 13–30. doi:10.1111/j.1530-9290.2009.00208.x. Tukker, A. and Tischner, U., 2006. ‘Product-services as a research field: past, present and future. Reflections from a decade of research’. Journal of Cleaner Production, 14(17). 1552–56. doi:10.1016/j.jclepro.2006.01.022. U.S. Energy Information Administration, 2010. Annual Energy Outlook 2010 with Projections to 2035. DOE/EIA-0383(2010). Washington, DC. http://www.eia.gov/oiaf/aeo/pdf/ 0383(2010).pdf. U.S. Energy Information Administration, 2009. Energy Market and Economic Impacts of H.R. 2454, the American Clean Energy and Security Act of 2009. http://www.eia.gov/oiaf/servicerpt/hr2454/index.html?featureclicked=2&. U.S. Environmental Protection Agency, 2012. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2010. U.S. Environmental Protection Agency, 2011. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2009. U.S. Environmental Protection Agency, 2010. Solid Waste Management and Greenhouse Gases: Documentation for Greenhouse Gas Emission and Energy Factors Used in the Waste Reduction Model (WARM). http://www.epa.gov/climatechange/wycd/waste/SWMGHGreport.html. U.S. Environmental Protection Agency, 2008. State Inventory Tool (SIT). ICF International for the EPA, Washington D.C. United Nations Framework Convention on Climate Change, 2011. Decision 1/CP.16: The Cancun Agreements: Outcome of the Work of the Ad Hoc Working Group on Longterm Cooperative Action Under the Convention. Cancun, Mexico. http://unfccc.int/resource/docs/2010/ cop16/eng/07a01.pdf. 31


SEI WP 2012-05

van Nes, N. and Cramer, J., 2006. ‘Product lifetime optimization: a challenging strategy towards more sustainable consumption patterns’. Journal of Cleaner Production, 14(15-16). 1307–18. doi:10.1016/j.jclepro.2005.04.006. Weber, C. A., Matthews, D., Venkatesh, A., Costello, C. and Matthews, H. S., 2010. The 2002 Benchmark Version of the Economic Input-output Life Cycle Assessment (EIO-LCA) Model. http://www.eiolca.net. Weber, C. L. and Matthews, H. S., 2007. ‘Embodied Environmental Emissions in U.S. International Trade’. Environmental Science & Technology, 41. 4875–81. Wei, M., Jones, C., Greenblatt, J. and MacMahon, J., 2011. ‘Carbon Reduction Potential from Behavior Change in Future Energy Systems’. Wiedmann, T., 2008. Development of an Embedded Carbon Emissions Indicator: Producing a Time Series of Input-Output Tables and Embedded Carbon Dioxide Emissions for the UK by Using a MRIO Data Optimisation System. Defra (Department for Environment, Food and Rural Affairs), London, UK. Willett, W., 2001. Eat, Drink, and Be Healthy: The Harvard Medical School Guide to Healthy Eating. Simon & Schuster Source, New York. World Business Council for Sustainable Development and World Resources Institute, 2010. Product Accounting & Reporting Standard: Draft for Stakeholder Review. World Business Council for Sustainable Development and the World Resources Institute. WWF, 2011. The Energy Report: 100% Renewable Energy by 2050. Ecofys and Office for Metropolitan Architecture for WWF International, Gland, Switzerland. http://wwf.panda.org/energyreport/. Yang, C., McCollum, D., McCarthy, R. and Leighty, W., 2008. Identifying Options for Deep Reductions in Greenhouse Gas Emissions from California Transportation: Meeting an 80% Reduction Goal in 2050. UCD-ITS-RR-08-23. UC Davis Institute ofTransportation Studies.

32


SEI WP 2012-05

APPENDIX A. METHODOLOGY FOR FORECASTING CONSUMPTION ACTIVITY As described in the main text of this report, a common approach to constructing GHG forecasts is employ a method based on the Kaya Identity (Kaya 1990; Yang et al. 2008), a variation of which can be expressed as: 𝐸𝑛𝑒𝑟𝑔𝑦 𝐺𝐻𝐺𝑠 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 �×� �×� � 𝐺𝐻𝐺𝑠 = (𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛) × � 𝐴𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝐸𝑛𝑒𝑟𝑔𝑦 𝑃𝑒𝑟𝑠𝑜𝑛

In consumption-based GHG inventories and the input-output models they rely upon (Peters 2008; Davis and Caldeira 2010), activity is denominated as monetary units (e.g., U.S. dollars). Yet changes in expenditures over time do not necessarily relate linearly to changes in GHG emissions. For example, as incomes grow and expenditures increase, some fraction of the increased expenditures is devoted to increased quality (or price) of items, not to more items (Girod and de Haan 2010). Accordingly, instead of forecasting consumption activity as a one-to-one relationship with expenditures, we use a relationship derived from elasticities the consumer expenditure literature. In particular, our method to forecast activity per person (the second term in the equation above) is to: 1. Start with a forecast of total (real) expenditures, as are often generated by government agencies (for example, as in the EIA’s Annual Energy Outlook). If a forecast of expenditures is not available (e.g., at a local level), we instead use a local forecast of income and an assumed change in the savings rate to estimate what fraction of income is converted to expenditures. 2. Use expenditure elasticities to forecast activity by category of consumption. An expenditure elasticity is the ratio of the percent change in spending on a given item (e.g., clothing) to the percent change in total expenditure (Taylor and Houthakker 2009). Increases in spending, however, do not necessarily translate to increased quantity of goods or services. For example, there is little evidence that households that spend more on food eat more kilograms or calories of food (Jones and Kammen 2011). Accordingly, we also employ an alternate definition of expenditure elasticity– one based on functional units for each type of good or service rather than expenditures. 41 These elasticities relate the percent change in consumption (in terms of functional unit, such as kg of goods or hours of entertainment) to the percent change in total expenditure (Girod and de Haan 2010). For example, if the elasticity was 0.8 for furniture, then a 10% increase in overall expenditure would translate into an 8% increase in furniture purchases. 42

41

Defining functional units is standard practice in process-based life-cycle assessments. We define functional units by type of goods as in Girod and de Haan (2010) – e.g. the functional unit for goods and food is kg, the functional unit for services is either time (e.g. for education and entertainment) or currency (e.g., for financial planning services). 42 We are aware of only one set of elasticities based on functional units, derived in Switzerland (Girod and de Haan 2010). Because patterns may differ between the U.S. (the subject of our analysis) and Switzerland, and because the Swiss elasticities are available only for a limited set of highly aggregated categories (e.g., goods), we therefore start with traditional, currency-based expenditure elasticities for the U.S. (Taylor and Houthakker 2009) and adjust them by the ratio of the functional-unit and currency-based elasticities of Girod and de Haan (2010). For example, the U.S. elasticity for expenditure on apparel is 0.87, the Swiss elasticity is 1.3, and the Swiss elasticity (on a functional unit basis) is 0.88. We therefore approximate an elasticity for apparel (on a functional unit basis) for the U.S. of 0.87 * 0.88/1.3 = 0.59.

33


SEI WP 2012-05

There are other reasons, too, why changes in expenditures may not relate linearly to changes in environmental impact. For example, since input-output analysis uses average emissions intensities of producing sectors (including electricity), it cannot usually accurately assess the changes at the margin in energy supplies or production practices. For example, in several world regions, the source of baseload electricity is low-carbon hydropower, but the marginal source (the source that would be affected by a change in demand) is fossil fuels (often natural gas). Similar differences could exist for other factors of production, including capital (if new technologies were replacing old technologies, for example) and intermediate goods (such as cotton sources used to make clothing).

34


SEI WP 2012-05

APPENDIX B: TESTING THE APPROACH FOR THE U.S. In addition to piloting the scenario methodology for Seattle, we applied it to the United States as a whole. We estimated a consumption-based GHG inventory for the U.S. using the same model used to estimate such an inventory for the City of Seattle, SEI’s CBEI model (Stanton, Bueno, Cegan, et al. 2011). This appendix presents results of the exercise for the U.S. and discusses differences – in method and in results – between the U.S. and Seattle case studies. GHGs associated with current consumption in the U.S.

We estimate that consumption-based emissions for the U.S. as a while are about 26 tons of CO2e per resident, just as for Seattle. However, the distribution of emissions across product categories is different for the U.S. as for Seattle. Emissions associated with home energy are much higher in the U.S. as a whole than in Seattle, due mostly to Seattle’s very low-carbon electricity supply. On the other hand, emissions associated with all other categories of consumption are lower, on average, in the U.S. as a whole than in Seattle, due to the lower relative wealth and expenditures by the average U.S. resident than Seattlelites. Figure 8: GHG emissions associated with consumption in U.S., 2010 (tCO2e/person) 2.7 6.2

1.7

Personal Transportation Home Energy Freight

2.9

Food Other Goods Services 4.5

2.6

Construction Other

2.5

2.6

Total = 26 tons CO2e / person

We forecast future, baseline emissions much the same as for Seattle, but use forecasts of consumer expenditures, transportation activity, and home energy at the national level from the EIA’s Annual Energy Outlook (EIA 2010). We use the same expenditure elasticities as for Seattle, drawn from analysis of the national Consumer Expenditure Survey (Taylor and Houthakker 2009), the same method for estimating changes in the emissions-intensities of producing goods and services, and the same trends in building and vehicle efficiency. Figure 8 displays the baseline forecast of GHG emissions associated with U.S. consumption. While the starting point of about 26 t CO2e per person is close to Seattle’s, a slightly slower rate of forecast growth in income and expenditure leads to emissions in 2030 being 27 tCO2e per person, lower than our forecast for Seattle of 29 tCO2e per person.

35


SEI WP 2012-05

Figure 9: Forecast GHG emissions associated with consumption in the U.S., baseline scenario 30

25 Other 20


GHGs per person, tCO2e 15 (Consumption Perspective)


10


5

0 2010

2015

2020

2025

2030

Low-GHG consumption mitigation measures and scenario results

Measures considered for the U.S. are the same as for Seattle (Table 4) with the exception of the home energy category, where we use the U.S. “Deep Green” scenario from the Green Building North America study (Adelaar et al. 2008) and personal transportation, where we adopt the national scenario of the Pew Center (Greene and Plotkin 2011). Together, these mitigation measures reduce GHG emissions associated with U.S. consumption to about 14 tCO2e per person, a 47% reduction from baseline, which is the same relative reduction as for Seattle. Figure 10: Emissions in the baseline and mitigation scenarios, U.S. 30

25

20 GHGs per person, tCO2e 15 (Consumption Perspective)

Baseline Scenario Mitigation Scenario

10

5

2010

2015

2020

36

2025

2030


SEI WP 2012-05

Figure 10 displays the results of the mitigation scenario for the U.S., by category of consumption. Compared with Seattle, reductions in national emissions from home energy were somewhat greater (because Seattle’s electricity was already zero carbon), whereas reductions in personal transportation were not as great (because we assumed greater potential to reduce vehicle travel in urbanized Seattle and somewhat greater uptake of low-GHG biofuels). Figure 11: Emissions in the mitigation scenario for the U.S., by category of consumption 30

25 Other 20


GHGs per person tCO2e 15 (Consumption Perspective)


10


5

0 2010

2015

2020

2025

2030

Conclusions of U.S. analysis

We demonstrated the feasibility of conducting a scenario analysis of GHG emissions associated with U.S. consumption. The analysis starts with an consumption-based GHG inventory for the U.S., forecasts this through 2020 using forecasts of consumer expenditures, vehicle travel, home energy, and the energy- and GHG-intensities of producing goods and services, most of which were drawn from the EIA’s Annual Energy Outlook (EIA 2010). Several conclusions emerge from this analysis: 

Consumption-based GHGs of 26 tCO2e per person in 2010 imply a national consumption-based GHG inventory of 8.0 billion tons CO2e, about 17% greater than the emissions released within the U.S. borders (production-based GHG inventory) of 6.8 billion tons CO2e (EPA 2012).



Emissions associated with U.S. consumption are expected to grow modestly (per person) between 2010 and 2030, though this implies a growth in absolute emissions of roughly 2 billion tons CO2e, or 1.2% per year, roughly double the EIA’s forecast rate of increase in production-based emissions of 0.6% per year.



Implementation of production- and consumption-based GHG mitigation measures could reduce consumption-based GHGs to about 14 tCO2e per person, or 5.3 billion tCO2e, in 2030, a 47% reduction from baseline. On an absolute basis, these reductions are 16% below 2010 levels in 2020 and 33% below 2010 levels in 2030.

37




SEI WP 2012-05

Behavioral GHG abatement measures can play a role in reducing GHGs associated with consumption. Diet shift, extended product lifespans, and other shifts in purchasing reduce emissions by about 6% from baseline in 2030. This reduction is somewhat less (on a percentage basis) than for Seattle, given that the share of GHG abatement from goods and services is somewhat less for the U.S. than for Seattle.

38

SEI - Africa Institute of Resource Assessment University of Dar es Salaam P.O. Box 35097, Dar es Salaam Tanzania Tel: +255-(0)766079061 SEI - Asia 15th Floor Witthyakit Building 254 Chulalongkorn University Chulalongkorn Soi 64 Phyathai Road Pathumwan Bangkok 10330 Thailand Tel: +(66) 22514415 SEI - Oxford Suite 193 266 Banbury Road, Oxford, OX2 7DL UK Tel: +44 1865 426316 SEI - Stockholm Kräftriket 2B SE -106 91 Stockholm Sweden Tel: +46 8 674 7070

SEI - Tallinn Lai 34, Tallinn, 10133 Estonia Tel: +372 6 276 100 SEI - U.S. 11 Curtis Avenue Somerville, MA 02144 USA Tel: +1 617 627-3786 SEI - York University of York Heslington York YO10 5DD UK Tel: +44 1904 43 2897

The Stockholm Environment Institute SEI is an independent, international research institute. It has been engaged in environment and development issues at local, national, regional and global policy levels for more than a quarter of a century. SEI supports decision making for sustainable development by bridging science and policy.

sei-international.org