Policy Analysis Are Services Better for Climate Change? SANGWON SUH* Department of Bioproducts and Biosystems Engineering, College of Food, Agricultural and Natural Resource Sciences, University of Minnesota, 2004 Folwell Avenue, Saint Paul, Minnesota 55108 and Institute of Environmental Sciences (CML), Leiden University, P.O. Box 9518, 2300 RA, Leiden, the Netherlands
Embodied greenhouse gas (GHG) emissions and their structure of inducement by the supply-chain networks of 480 goods and services in the United States are analyzed for 44 GHGs. Producing a dollar of a product or service generates an average of 0.36 kg of CO2 equivalent GHGs onsite, increasing to 0.83 kg when supply-chain-induced emissions are taken into account. Services produce less than 5% of total U.S. GHG emissions directly, and their direct GHG emission intensities per dollar output are much less (0.04 kg CO2 equiv/$) than those of physical products, even when supply-chain-induced emissions are included (0.47 kg CO2 equiv/$). When both supply-chain effects and the volume of household expenditures are taken into account, however, household consumption of services excluding electric utilities and transportation services proves to be responsible for 37.6% of total industrial GHG emissions in the United States, almost twice the amount due to household consumption of electric utility and transportation services. Given the current structure of GHG emissions, a shift to a service-oriented economy is shown to entail a decrease in GHG emission intensity per unit GDP but an increase, by necessity, in overall GHG emissions in absolute terms. The results are discussed in the context of U.S. climate change policy.
Introduction Despite some skepticism, recent scientific evidence indicates that the current level of ambient greenhouse gas (GHG) concentration is well above the level of natural variability and is driven by anthropogenic interventions (1-3). For the past 100 years global annual anthropogenic CO2 emissions due to fossil fuel combustion have increased by approximately a factor 12, and responsibility for around a quarter of the total accumulative global emissions during the period has been ascribed to the United States (4-6). In the U.S., major industrial GHG emission sources include electric power production (subsequently referred to as “electric utility”), transportation, and several manufacturing industries, including petroleum refining, iron and steel manufacturing, and cement production, which together generate around 80% of the total (7). In contrast, the service segment (excluding electric utility and transportation services), comprising banking, hospitals, and the retail trade, for example, as well as computer and data processing services, accounts for less * Phone: (612) 624 5307; fax: (612) 625 6286; e-mail: sangwon@ umn.edu. 10.1021/es0609351 CCC: $33.50 Published on Web 09/27/2006
2006 American Chemical Society
FIGURE 1. The share of durable goods such as housing and furniture in total consumption expenditure, in constant prices, has been remarkably stable over the four decades since 1960, whereas services have encroached mainly on the share of non-durable goods, which are major manufacturing goods. In the United States, the break-even point in total consumer expenditure between services and non-durable goods, in constant prices, was reached in 1967 (3). than 5% of total U.S. GHG emissions (7). Economically, however, the services sector is the largest and the fastest growing sector in the United States. The relative share of services in personal consumption expenditure has been steadily rising, mainly by encroaching on the share of manufactured products (Figure 1). Over the past 10 years, the size of services in aggregate gross domestic product (GDP) has grown by an average of 6.3% a year in the United States, thus doubling every 11-12 years (8). Today, the services sector contributes around 60% of total U.S. GDP (8). Given their economic importance and negligible contribution to GHG emission, it is not surprising to see that growth of the service economy is recognized for many as a “wedge” to mitigate climate change (9). The objective of the present paper is to gain a deeper understanding of climate change implications of services. Throughout the analysis, global warming potential (GWP) 100 by Houghton et al. (10) is used to aggregate GHG emission data into CO2 equivalents, and all prices are in 1998 producer prices. “Services” in this paper are defined in a narrow sense to exclude the categories of electric utility, steam supply, and transportation services. Whenever relevant, the 480 goods and services are grouped into four segments, (1) utility and transportation, (2) primary, (3) secondary, and (4) tertiary sectors, for an easier comparison. Using the classification used by the Bureau of Economic Analysis (BEA), primary sector includes 10100-110000; secondary sector is 110101641200; tertiary sector is 650100-820000, excluding electric utility and transportation (“railroads and related services”, “local and suburban transit and interurban highway passenger transportation”, “trucking and courier services excluding air”, “water transportation”, “air transportation”, “freight forwarders and other transportation services”, “electric utility”, and “sanitary services and steam supply”). Analyses are made generally at both the detailed and the aggregated levels.
Methods and Data In a recently completed database project, major U.S. environmental emission inventories, including the national VOL. 40, NO. 21, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6555
Greenhouse Gas Inventory, Toxics Releases Inventory (TRI), National Toxics Inventory (NTI), and National Environmental Trends (NET) database were linked with the supply-chain networks of 480 products and services, based, wherever possible, on the most detailed 6-digit Standard Industrial Classification (SIC) codes. The resulting database contains a total of 1344 environmental interventions, including emissions of 44 GHGs, and their inducement structure as described by 1998 detailed U.S. input-output tables (1113). The 44 GHGs that are included in the database are carbon dioxide, methane, nitrous oxide, trichloromethane, sulfur hexafluoride, tetrachloromethane, perfluorobutane, perfluorocyclobutane, perfluoroethane, perfluorohexane, perfluoromethane, perfluoropentane, perfluoropropane, methylbromide, methyl cyclohexane, halon-1211, halon-1301, 7 different HCFCs, 13 different HFCs, 6 different CFCs, and dichloromethane. For details on the data sources and calculation methods used to construct the database, see ref 11. Using this database, the current study analyzes the structure of GHG emissions by the U.S. economy using four different metrics: (1) direct GHG emission per dollar production (kg CO2 equiv/$), (2) total direct GHG emission (kg CO2 equiv), (3) embodied GHG emission per dollar of a product (kg CO2 equiv/$), and (4) total embodied GHG emission (kg CO2 equiv). Calculus of each of these four metrics is shown below. The third and the fourth metrics involve calculation of both direct and indirect GHG emissions following the concept of “embodiment” introduced in the field of energy analysis in the 1970s (14-15). Not included in these figures are, however, direct GHG emissions from household activities. Let c ) 1, ... , C index commodities including goods and services, and let p ) 1, ... , P index GHGs. The direct GHG emission intensity of a good or service c is calculated by P
mc )
∑
p)1
( ) fp ×
Bpc qc
(1)
TABLE 1. Top Ten Products and Services with Respect to Direct GHG Emission Intensity product/service category
GWP intensity (kg CO2 equiv/$)
Lime Electric utility Sanitary services and steam supply Chemical and fertilizer minerals Food grains Other structural clay products Hydraulic cement Tobacco Fruits Coal
20.7 9.5 7.5 7.4 4.4 4.3 4.1 4.0 3.7 3.6
rank 1 2 3 4 5 6 7 8 9 10
FIGURE 2. Total domestic production (q) in million U.S. dollars vs direct GHG emissions in kg of CO2 equivalents (m2). Plots in a vertical cross-section at any q indicate the amount of GHG directly generated in producing q amount of products or services, plots in a horizontal cross-section at any m2 indicte the amount of production possible for a given level of direct GHG emissions.
where fp is GWP of GHG p, Bpc is annual direct emission of GHG p to produce commodity c, and qc is total annual output of commodity c. In a matrix formula eq 1 can be noted as
Total embodied GHG emission, the fourth metric, is derived by
m1 ) fBqˆ -1
m4 ) m3yˆ
(2)
where m1 is a 1 × C vector of direct GHG emission per a dollar of goods and services, which is the first metric previously described, f is a 1 × P vector of characterization factors, B is a P × C matrix of direct GHG emission by goods and services, and q is a C × 1 vector of total annual commodity outputs, and hat (∧) diagonalizes a vector. By removing q in the right side of eq 2, it becomes
m2 ) fB
(3)
where m2 is an 1 × C vector of total direct GHG emission by goods and services, which is the second metric. The third metric, embodied GHG emission, that is direct and indirect GHG emission induced, per a dollar of consumption (cf. 14) is calculated by
m3 ) m1(I - A)-1
(4)
where A is a C × C technology matrix with endogenized imports and I is a C × C identity matrix (11). 6556
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 21, 2006
(5)
where y is a C × 1 vector of U.S. household consumption. With these four vectors calculated by the eqs 2-5, the four metrics are prepared for each of the 480 goods and services consumed in the United States, which will help better understand the intricate networks of GHG inducement structure of the U.S. economy. These quantities form the basis of the present analysis, while relevant other analytical tools such as input or output contribution analyses (11) are utilized when feasible.
Results The results show that production of a dollar of good or service generates, on average, 0.36 kg of CO2 equivalent GHGs directly (metric 1). By far the largest on-site GHG emitter per dollar is “lime production”, generating 20.7 kg of CO2 equivalent global warming impact per dollar (Table 1). Lime production is an important non-combustion-related CO2 emission source. So-called quicklime or burnt lime is produced from limestone, liberating CO2 from calcium carbonate by the reaction CaCO3 f CaO + CO2. Stoichiometric production of 1 kg of quicklime (CaO) generates 0.78 kg of CO2. For the
TABLE 2. Top Ten Products and Services with Respect to Total Direct GHG Emissions
product/service category
GWP (Tg CO2 equiv)
share in total industrial GHG emissions
Electric utility Sanitary services and steam supply Air transportation Crude petroleum and natural gas Blast furnaces and steel mills Petroleum refining Trucking and courier services, excl. air Feed grains Meat animals Coal subtotal
2175 262 190 173 164 162 154 135 115 83 3613
38.7% 4.7% 3.4% 3.1% 2.9% 2.9% 2.7% 2.4% 2.1% 1.5% 64.4%
rank 1 2 3 4 5 6 7 8 9 10
TABLE 3. Top Ten Products and Services with Respect to Direct and Supply Chain GHG Emission Intensity rank 1 2 3 4 5 6 7 8 9 10
product/service category
GWP intensity (kg CO2 equiv/$)
Lime Electric utility Sanitary services and steam supply Chemical and fertilizer minerals Miscellaneous livestock Meat animals Hydraulic cement Food grains Other structural clay products Tobacco
22.1 10.2 8.5 8.4 5.8 5.6 5.5 5.4 5.3 4.9
same reason, structural clay and hydraulic cement are included in Table 1. Next in intensity is the “electric utility”, with its major CO2 emissions due to coal and other fossil fuel combustion. “Sanitary services and steam supply” as well as “coal production” are calculated to be high in direct GHG emission intensity, owing mainly to methane (CH4) emissions from landfill and to mining and coal processing, respectively. The landfill activities classified under “sanitary services and steam supply”’ are the largest methane source in the United States, generating 9.6 million tons of methane in 1998. CO2 emissions from waste incineration are an additional source of global warming impact in this category. Methane is stored in coal and/or its earth matrix, and is freed during mining, crushing, and pulverization processes, thereby contributing to the high direct GHG emission intensity of mining. One of the main reasons for “chemical and fertilizer minerals” being ranked within the top ten are CO2 emissions from ammonia (NH3) production. Ammonia production is another major non-combustion-related source, thanks to the production process of the hydrogen required to form ammonia together with nitrogen. Hydrogen is generally extracted from natural gas by means of chemical reactions known as steam reforming and watergas shift: CH4 + H2O f CO + 3H2 and CO (g) + H2O f CO2 + H2. Stoichiometrically, 1 kg of ammonia produced using these reactions as a hydrogen source generates 0.97 kg CO2, a figure rising to 1.29 kg CO2 if only the steam reforming process is used. Agricultural products are also found to cause large on-site GHG emissions per dollar output. The inclusion of “food grains”, “tobacco”, and “fruits” in Table 1 can be attributed largely to methane and nitrous oxide emissions from rice cultivation, burning of agricultural residues, and use of synthetic fertilizers and manure, the latter additionally associated with CO2 emissions from off-road agricultural vehicles. In terms of direct emission (Table 1), none of the services were found to be GHG-intensive. After aggregation, the direct GHG emissions intensity of the primary sector is calculated
to be 2.05 kg of CO2 equivalents per dollar; this figure is far lower for the secondary and the tertiary sectors, which generate only 0.30 and 0.04 kg of CO2 equivalent GHGs per dollar, respectively. Multiplying direct GHG emission intensities by annual production volumes yields total direct GHG emissions (metric 2) (Figure 2). As the figure shows, electric utility and transportation and the primary sector are among the largest direct GHG emitters, occupying the upper part of the graph (see also Table 2). The direct GHG emission intensities of different product groups can be clearly distinguished, as most of the plots of the primary and tertiary sectors are aligned around the upper and lower parts, respectively, of the linear regression line of the secondary sector. Nonetheless, in terms of the overall size of economy, services are among the largest, occupying the right side of the graph. In general, services occupy the lower-right part of Figure 2, indicating their low GHG emission intensity per dollar output with respect to direct emissions. For a given level of economic production, the difference in direct GHG emissions between services and other products, reflected in a vertical cross-section of the graph of Figure 2, may be as high as a factor 103. Including emissions from the entire supply-chain network led to only minor changes in the ranking of the most GHG emission intensive products (Table 3), indicating only a weak link with other products and services of these products via upstream supply chains (metric 3). Some livestock products came into the picture as they involve consumption of “feed grains” and “prepared feeds”, etc., that are already high in GHG emission intensity. On average, 1 dollar of products or services in the U.S. generates 0.83 kg of CO2 equivalents directly and indirectly through the supply chain. In general, then, indirect GHG emissions in the upstream supply chain exceed the average intensity of direct emissions (0.36 kg/$). However, the relative magnitude of indirect emissions compared with direct emissions varies substantially from sector to sector (Figure 3). In particular, the supply-chain GHG emissions of the tertiary sector induces are, on average, over 16 times greater than its direct GHG emissions. Nevertheless, the GHG emission intensities of services are still lower than those of other products. On average, the total direct and supply-chain GHG emissions induced per dollar output decrease as follows: electric utility and transportation (5.3 kg CO2 equiv/$), primary sector (3.1 kg CO2 equiv/$), secondary sector (1.0 kg CO2 equiv/$), and tertiary sector (0.5 kg CO2 equiv/$). Linking total direct and supply chain GHG emission intensities with household consumption expenditure data yields Table 4 (metric 4) (10). The composition of the 30 largest GHG emission-inducing products and services in Table 4 differs substantially from that of the other three tables. First, the list includes the most basic necessities of energy, shelter, mobility, health care, food, etc. Second, except for VOL. 40, NO. 21, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6557
FIGURE 3. Total GHG emissions as a multiple of direct emissions by electric utility and transportation (U), and primary (P), secondary (S), and tertiary (T) sectors. The gray bar indicates the 20-80% range of the accumulative frequency distribution, neglecting one extreme value on each side. a few items such as electric utility, the products and services listed in Table 4 do not have high GHG emission intensities. The total direct and supply-chain GHG emission intensity of “motor vehicles and passenger cars”, for instance, is 1.08 kg CO2 equiv, only slightly higher than the average, indicating in turn the high consumption volume of these products and services. Third, most of the products and services in Table 4 are associated with supply-chain GHG emissions rather than direct emissions. For instance, the total direct and
supply-chain GHG emissions induced by household consumption of “motor vehicles and passenger cars” occur in far-removed upstream processes such as “blast furnaces and steel mills” (15.8%), various organic and inorganic chemical processes (10.8%), various mining (5.1%), “electric utility” (21.4%), and so on. Another striking difference is that around half the items on the list are now services. A total of 37.6% of overall GHG emissions are induced through household consumption of services (Figure 4). Nearly half the GHG emissions induced by services are associated with electric utility and transportation (45.1%). Adding on-site GHG emissions from the primary and secondary sectors upstream of the services, 84.9% of the total emissions due to the household consumption of services take place outside the sector itself (Figure 4). Ranked within the top 30 are even services that do not generally supply tangible materials, like “hospitals”’ (5th), “banking” (17th), and “insurance carriers” (24th). These services induce indirect GHG emissions at various industries, including, in particular, electric utility, transportation, and construction. Hospitals, for instance, rely on direct GHG emissions from “electric utility” (37.0%), “sanitary services and steam supply” (7.7%), various agricultural products (4.5%), “crude petroleum and natural gas” (3.8%), “blast furnaces and steel mills” (3.0%), “air transportation” (2.8%), “platemaking and related services” (2.7%), various construction and its maintenance (2.5%), etc. These indirect emissions are induced primarily by the direct consumption of “electric utility” (25.5%), “sanitary services and steam supply” (5.1%), “real estate agents” (6.4%), “industrial inorganic and organic
TABLE 4. Top Thirty Products and Services with Respect to Total GHG Emission Embodiments, Allowing for Household Consumption Expenditure
rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 a
product/service category Electric utility Retail trade, excl. eating and drinking*a Motor vehicles and passenger cars Eating and drinking places* Hospitals* Petroleum refining Non-farm residential 1 unit structures Owner-occupied dwellings* Real estate agents* Meat packing plants Air transportation Wholesale trade* Industrial and commercial building Apparel Other construction Natural gas distribution* Banking* Automotive repair shops and services* Poultry slaughtering and processing Sanitary services and steam supply* Doctors and dentists* Trucking and courier services Alterations of non-farm construction Insurance carriers* Other State and local government* enterprises Nursing and personal care facilities* Water supply and sewerage systems* Computer and data processing services* Telephone and other communication* Other amusement and recreation* Subtotal
* Designates services.
6558
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 21, 2006
GWP (Tg CO2 equiv)
share in total industrial GHG emissions
984.4 326.8 310.0 301.3 268.3 183.5 161.2 148.7 124.5 116.7 112.4 111.8 100.5 95.2 94.7 67.6 65.7 64.8 59.8 58.8 58.5 57.4 53.4 50.2 47.2
16.2% 5.4% 5.1% 5.0% 4.4% 3.0% 2.7% 2.4% 2.1% 1.9% 1.9% 1.8% 1.7% 1.6% 1.6% 1.1% 1.1% 1.1% 1.0% 1.0% 0.96% 0.95% 0.88% 0.83% 0.78%
46.1 45.9 43.2 43.2 42.3 4244.1
0.76% 0.76% 0.71% 0.71% 0.70% 69.9%
FIGURE 4. Total direct and supply-chain GHG emissions induced by household consumption expenditure on electric utility and transportation (U), and primary (P), secondary (S), and tertiary (T) sectors, broken down according to on-site emission sources. chemicals” (5.0%), “industrial and commercial buildings” (4.4%), “drugs” (3.5%), “surgical and medical instruments and apparatus” (2.1%), “surgical appliances and supplies” (1.9%), etc. Likewise, the aggregate GHG emissions of “banking” are induced through direct consumption of “electric utility” (12.0%), “security and commodity brokers” (9.0%), “sanitary services and steam supply” (7.2%), “industrial and commercial buildings” (4.7%), “U.S. postal service” (4.0%), “computer and data processing services” (2.7%), “motor vehicles and passenger cars” (2.2%), “advertising” (1.7%), “computer peripheral equipment” (1.5%), “warehousing and storage” (1.2%), “‘blankbooks, looseleaf binders and devices” (1.0%), etc. The mechanism by which services emerge once the supply chain and total consumption volume are taken into due account can be approached by comparing Figure 2 and Figure 5. In Figure 5 the relationship between household consumption expenditure and total direct and supply-chain GHG emissions are plotted on a log-scale. Compared with Figure 2, distribution has shifted upward in Figure 5 as supply-chain GHG emissions are added to direct emissions. Furthermore, the distribution is denser in Figure 5 than in Figure 2. However, while the plots for the primary and tertiary sectors asymptotically approach the secondary sector, they do not generally overlap. Another interesting observation is that the plots for the primary and secondary sectors have undergone a shift to the lower left, while this is not generally the case for the tertiary sector. The mechanism underlying these shifts sheds light on the basic structure of GHG emission inducement. First, the shift of the primary and secondary sectors to the lower left of Figure 5 indicates that a substantial fraction of the output of these sectors is not consumed directly by household consumers. Excluding the fraction exported, the difference between total production and total household consumption yields the amount consumed by the subsequent downstream supply chain prior to delivery to final consumers. U.S. households consume 86.0% of primary sector outputs and 43.8% of secondary sector outputs indirectly in the form of other industry outputs, a substantial part of which are services. In this regard, services act as an interface between primary and secondary products and household consumers. This explains how the plots in Figure 5 approach one another only asymptotically. As tertiary sector services rely on input of primary and secondary products, their total direct and indirect GHG emission intensities increase substantially when supply chains are taken into account. However, these intensities do not exceed those of the inputs, as additional value-added is created, “diluting” overall intensities per dollar (Figure 6). By combining these two effects, the plots for the tertiary sector move closer to the level of its inputs, but not
FIGURE 5. Total household consumption expenditure (y) in million U.S. dollars vs total GHG emission embodiment in kg CO2 equivalents (m4).
FIGURE 6. Simplified linear supply chain comprising a primary (P), secondary (S), and tertiary (T) sector produces p, s, and t, respectively, with t being delivered to a household consumer (C). At each stage, value added (VA) is created and GHGs are emitted. The total direct and supply-chain GHG emission intensity of t is calculated as (GHGP + GHGS + GHGT)/(VAP + VAS +VAT), less than that of s if and only if (VAT/GAGT) > [(VAP + VAS)/(GHGP + GHGS)], while the total GHG induced by t is invariably larger than that of s for a nonzero GHGT. With their low direct GHG emissions and high value added, services generally satisfy these conditions. beyond. The value-added effect also helps explain how less GHG emission intensive services rise in ranking when total household expenditure is taken into account (Table 4). As these products and services are located at the near-toconsumer side of the supply chain, they will have undergone longer supply paths and corresponding value-added processes before being delivered to household consumers. These services therefore generally have a much larger value-added moiety accumulated in their price and the overall volume of consumption for the same material contents will consequently be higher. In the same light, the high GHG emission intensities of basic materials and agricultural products in Tables 1 to 3 reflect their relatively low prices, which can be ascribed to the fact that the processes in question are generally located at the start of the series of value-adding processes along the supply chain. Under these circumstances, although the GHG emission intensities of services are, by their very nature, lower than those of other sectors, aggregate GHG emissions in absolute terms will not be automatically reduced merely by engendering a structural shift toward a more service-oriented economy if the same or higher material welfare is to be maintained (Figure 6).
Discussions and Outlook Over the past decade, theoretical as well as empirical grounds for the existence of a negative relationship between income VOL. 40, NO. 21, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6559
and environmental degradation, known as the Environmental Kuznets Curve (EKC), have attracted considerable scientific interest (16-21). Some have identified the spontaneous shift from material-intensive industry to less material-intensive services in the course of economic growth as a key factor driving the decoupling economic prosperity from environmental degradation (22-24). Nevertheless, for CO2 emissions Shanfik and Bandyopadhay (25) found an opposing trend. The present analysis contributes to these findings, by explaining why services are less GHG emission intensive, and necessarily so, so that a shift to services will not, in itself, reduce aggregate GHG emissions. It is certainly true that a shift to a more service-oriented economy will reduce the GHG emission intensity per unit GDP and is desirable, especially in the context of U.S. climate change policy, with which the U.S. President committed the country to a strategy to reduce greenhouse gas intensity per dollar production by 18% over the next 10 years, slightly more than the rate of reduction since 1990, achieved in the absence of substantial policy coordination (26). What is often neglected is that services are deeply anchored to manufacturing outputs, and growth in services sector also lifts, by necessity, manufacturing outputs. In 2004, a dollar of consumption on seemingly material-free services, which exclude utility and transportation services, requires about a quarter worth of outputs from manufacturing, utility, and transportation services sectors in the United States (27). In fact, for over four decades, U.S. production of manufactured goods has generally followed an upward trend in absolute term, although its relative share in GDP has been gradually surpassed by services. Therefore, mitigation of climate change, which requires actual reduction of GHG emissions in absolute terms, is not achieved automatically in the course of economic growth and associated structural change unless the services become independent of embedded GHG emission intensive products. Further study may be focused on identification of key structural paths and priority areas that are relevant for reducing GHG emissions of the United States. Time series and scenario analyses incorporating future structural changes will further scrutinize the role of services for climate change.
Literature Cited (1) Pacala, S. W.; Bulte, E.; List, J. A.; Levin, S. A. Environmental economics. False alarm over environmental false alarms. Science 2003, 301, 1187-1188. (2) Karl, T. R.; Trenberth, K. E. Modern Global Climate Change. Science 2003, 302, 1719-1723. (3) Crowley, T. J. Causes of Climate Change Over the Past 1000 Years. Science 2000, 289, 270-277. (4) Marland, G.; Boden, T.; Andres, R. J. Global CO2 Emissions from Fossil-Fuel Burning, Cement Manufacture, and Gas Flaring: 1751-2000; Carbon Dioxide Information Analysis Center; Oak Ridge National Laboratory: Oak Ridge, TN, 2002. (5) Baumert, K. A.; Kete, N. Climate Protection in a Disparate World: World Resources Institute (WRI): Washington, DC, 2002.
6560
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 21, 2006
(6) Baer, P.; Harte, J.; Haya, B.; Herzog, A. V.; Holdren, J.; Hultman, N. E.; Kammen, D. M.; Norgaard, R. B.; Raymond, L. Equity and Greenhouse Gas Responsibility. Science 2001, 289, 2287. (7) U.S. Environmental Protection Agency. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 -2001; EPA: Washington, DC, 2003. (8) Bureau of Economic Analysis. National Income and Product Accounts (NIPA). http://www.bea.gov. (9) Pacala, S.; Socolow, R. Stabilization Wedges: Solving the Climate Problem for the Next 50 Years with Current Technologies. Science 2004, 305, 968-972. (10) Houghton, J. T.; Ding, Y.; Griggs, D. J.; Noguer, M.; van der Linden, P. J.; Xiaosu, D. Climate Change 2001: The Scientific Basis; Cambridge University Press: Cambridge, 2001. (11) Suh, S. Developing a sectoral environmental database for inputoutput analysis: the comprehensive environmental data archive of the US. Econ. Syst. Res. 2005, 17, 449-469. (12) Comprehensive Environmental Data Archive (CEDA) ver 3.0; EnviroInformatica: Seoul, South Korea, 2004. http://www.enviroinformatica.com. (13) Bureau of Economic Analysis. U.S. Make and Use Matrices for 1998; Bureau of Economic Analysis: Washington, DC, 2003. (14) Bullard, C.; Herendeen, R. Energy costs of goods and services. Energy Policy 1975, 3, 268-278. (15) Costanza, R. Embodied energy and economic evaluation. Science 1980, 210, 1219-1224. (16) World Bank. World Development Report 1992; World Bank: Washington, DC, 1992. (17) Beckerman, W. Economic growth and the environment: Whose growth? whose environment? World Dev. 1992, 20, 481-496. (18) Grossman, G. M.; Krueger, A. B. Economic Growth and the Environment. Q. J. Econ. 1995, 112, 353-377. (19) Myers, N. Consumption Challenge. Sustainable Sci. 1997, 276, 53-55. (20) Vincent, J. R.; Panayotou, T. Consumption: Challenge to sustainable development or distraction? Science 1997, 276, 5557. (21) Bruvoll, A.; Medin, H. Factors Behind the Environmental Kuznets Curve. A Decomposition of the Changes in Air Pollution. Environ. Resour. Econ. 2003, 24, 27-48. (22) Bernardini, O.; Galli, R. Dematerialization: long-term trends in the intensity of use of materials and energy. Futures 1993, 25, 431-448. (23) Ja¨nicke, M.; Binder, M.; Monch, H. ‘Dirty industries’: Patterns of change in industrial countries. Environ. Resour. Econ. 1997, 9, 467-491. (24) Panayotou, T. Economic Growth and the Environment. United Nations Economic Commission for Europe: Geneva, 2003. (25) Shafik, N.; Bandyopadhay, S. Economic Growth and Environmental Quality: Time Series and Cross Country Evidence; background paper for the World Development Report; World Bank: Washington, DC, 1992. (26) U.S. Climate Change Policy: Fact Sheet. White House, Office of the Press Secretary, Washington, DC, 2003. (27) Bureau of Economic Analysis. 2004 U.S. Annual Input-Output Table; Bureau of Economic Analysis: Washington, DC, 2006.
Received for review April 18, 2006. Revised manuscript received August 23, 2006. Accepted August 28, 2006. ES0609351
