APPENDIX C
An Explanation of Our Research and
Analysis Methodology
Precisely calculating the global warming emissions associated with household consumption in the United States is a complex process.* Direct emissions from a car's tailpipe or a lawnmower's motor are relatively obvious and easy to quantify, but these direct “tailpipe” assessments fail to account for what are often called “upstream” or “indirect” impacts, which can be considerable. In the case of the lawnmower, for instance, upstream impacts include emissions resulting from manufacturing, emissions resulting from the retail process (including from the wholesaler who sells the mower to the retailer), and emissions resulting from shipping raw materials to the manufacturer and transporting the finished product to the wholesaler and then to the retailer. A similar pattern exists for all consumer goods.
Quantifying emissions that result from purchasing a service, such as an appointment with the dentist, or from consuming food is even more difficult. To do it effectively, one has to incorporate a fairly detailed understanding of emissions from electricity use and production, as well as the emissions that result every step of the way from purchases of materials and parts and from fuel use, transportation, and a host of other potential sources.
To accomplish this task in a detailed and comprehensive manner, we worked with a team of modelers at the Stockholm Environment Institute's U.S. Center, including Frank Ackerman, Elizabeth A. Stanton, and Ramón Bueno.
The team began by gathering and analyzing a significant amount of scientific data (primarily source emissions estimates) as well as detailed economic information on consumer and industry spending patterns.
The team analyzed emissions from all sources in 2006, the most recent year available when we initiated the project. We found that 7,075 million metric tons of carbon dioxide equivalent (CO2e)* had been emitted that year in the United States, of which 5,418 million metric tons were attributed to specific types of consumer expenditures and uses and included in our industry analysis.† Although we reviewed and analyzed numerous sources, we relied on sources including the U.S. Department of Energy's Emissions of Greenhouse Gases in the United States 2006‡ and “Official Energy Statistics from the U.S. Government: Households, Buildings, Industry, and Vehicles, 2001”§ and the U.S. Department of Transportation's “U.S. Government Energy Consumption by Agency and Source (Trillion Btu).”** All emissions data were entered into a spreadsheet model to derive national industry shares of consumption-based emissions and industry emissions coefficients for use in the analysis.*
To make this economic information fit together, the team needed to include analysis of industry output and consumer demand by industry sector for 2006, as well as to incorporate an understanding of the inter-industry linkages to determine upstream impacts. For these we relied on an economic modeling tool called IMPLAN, using information from the IMPLAN 2006 U.S. Data File.†
IMPLAN was originally developed in the 1970s by the USDA Forest Service to assist in land and resource management and planning. In the early 1990s, a group working on IMPLAN databases at the University of Minnesota formed the Minnesota IMPLAN Group (MIG) to privatize the development of IMPLAN data and software to make it more user friendly and available to a broader range of users. Since the release of the first publicly available version in 1996, subsequent versions of the IMPLAN software have continued to evolve and expand its flexibility for use in economic impact modeling. The current analysis utilized IMPLAN Professional™ Version 2.0.‡
The IMPLAN Version 2.0 modeling system is one of the most widely used and accepted regional economic analysis tools for predicting economic impacts in the United States. IMPLAN is currently used by over 2,000 public and private institutions, including government agencies, colleges and universities, nonprofit organizations, corporations, and business development and community planning organizations.§ It is designed to perform economic impact analysis and assist in planning and policy decisions on a broad range of issues. Among others, these include natural resource issues; plant openings, closings, and relocations; and economic related policy-driven scenarios.
IMPLAN is an “input-output” economic modeling tool. Input-output models, and the IMPLAN tool in particular, allow users to analyze how an economy responds to changes. Using inter-industry relationships (also known as supply linkages) within a particular region, input-output models show how an increase in demand for a product or service affects or causes changes in other industries. The increase in demand for a good or service affects the producer of the good or service and its employees and suppliers, as well as the supplier's employees and suppliers, among others. These linkages ultimately generate a total effect in the economy that is greater than the initial change in demand, commonly referred to as a multiplier effect.
In its most basic form, the IMPLAN system is composed of software and a database. The software provides a mechanism for data retrieval, model development, and the resulting impact analysis. The database consists of national and regional data in 528 separate sectors about employment, industry output, and institutional demand and transfers.* The sectors closely follow the accounting conventions used by the U.S. Bureau of Economic Analysis (BEA) and link to the standard industrial classifications (known as SIC codes). The database includes information derived from numerous published and unpublished sources, including surveys and reports from the U.S. Department of Agriculture, the U.S. Department of Commerce, the U.S. Department of Health and Human Services, and the U.S. Department of the Interior, among others.†
We started with the IMPLAN model by designating the entire United States as the study area and using the model to identify the total U.S. industry output (in dollars). According to IMPLAN, total industry output in 2006 was just under $27.8 trillion for all industry sectors. We then used IMPLAN to identify total household expenditures (in dollars) by each industry sector. According to IMPLAN, total household demand in 2006 was just under $10.0 trillion,*with national household expenditures estimated for nine separate income classes using data from the BEA's Benchmark Input-Output Study and the Consumer Expenditure Survey (CES).†
IMPLAN reporting and analysis, similar to all input-output models, is based on producer prices and industry demand. To allocate consumer spending as accurately as possible, IMPLAN uses so-called margins. Margins represent the differences between producer and purchaser prices and are used to allocate consumer expenditures to the correct input-output sector. For example, when the IMPLAN model analyzes a consumer's purchase of a light fixture at a lighting store, it allocates a portion of the consumer's outlay to the retailer, a portion to the wholesaler, and a portion to the transportation and manufacturing sectors. Each sector, in other words, benefits to some extent from the consumer purchase. Household expenditures are thus allocated across the 509 industry sectors that we analyzed in IMPLAN.
Margins are applied only to retail purchases of manufactured goods that have markups. Purchases of services do not have margins applied because the service is produced at the same time it is purchased and has no markup by other sectors. Just as margins are used to allocate consumer purchases, IMPLAN also quantifies industry linkages for each producer sector. For example, when the portion of the purchase price for a light fixture is allocated to the manufacturing sector, the model in turn subtracts from the manufacturer a portion of that money it spent to purchase the raw materials, equipment, and services needed to manufacture the lamp.
By identifying the use of commodities by a given industry's production process, the model ultimately derives multipliers that are triggered by each $1 of consumption for goods or services produced by a specific industry.*
With the emissions and economic data in hand, and the industry framework established, the modeling team then matched the emissions for each industry sector with the appropriate IMPLAN sectors. In this step of the analysis we were interested in determining how much CO2e is emitted per dollar of industry output, both for the direct effects and for the indirect effects. In most instances, the emissions data provided breakouts by industry classifications that were easily matched to IMPLAN's sectors or to the North American Industry Classification System (NAICS) codes.† In those instances where the classification wasn't clear, we determined how to attribute the emissions on the basis of a variety of specific factors, such as fuel type or end-use category. This careful matching had two purposes: first, it provided a basis for identifying which sectors produce the most emissions; second, it allowed us to derive industry-specific intensities.
The next step in the process involved matching the total CO2e emissions for each industry with the respective dollar outputs for each industry sector. By dividing each industry's total emissions by the total output, the team developed the CO2e coefficients (metric tons CO2e/$1 output)—the key factor for completing the impact analysis.‡
To better understand how the analysis and modeling actually works to calculate CO2e emissions impacts, we can look at a simplified example. Let's assume a consumer goes to a local furniture store and purchases a lamp for $100. Since the store is a retailer, it doesn't keep the whole $100, as we noted earlier. Based on the margins derived from IMPLAN, the $100 is actually split between the retailer, truck and air transportation, a wholesaler, and the lighting fixture manufacturer. In fact, as the table indicates, less than half of the cost of the lamp stays with the store. The remainder, approximately 53 percent, is funneled to other sectors to pay the suppliers and transportation costs. On the basis of the direct and indirect emissions coefficients for the industries noted, we see that the $100 expenditure for a lamp generates 0.03 metric ton of CO2e emissions throughout the economy. The table offers a snapshot of how the model works in practice.
Table C.1. CO2e Emissions for a $100 Lamp
*The terms “household consumption,” “consumer expenditures,” and “consumer or household demand” are all used in this report and refer to the same activities.
*Carbon dioxide equivalent (CO2e) is a measure used to incorporate the contribution of a variety of greenhouse gas emissions on the basis of their global warming potential relative to carbon dioxide.
†Please note that these raw data have subsequently been converted to conventional pounds and tons (2,000 pounds per ton) in the examples and tables throughout the book. Initial data are from U.S. Department of Energy, Energy Information Administration (EIA), “Distribution of Total U.S. Greenhouse Gas Emissions by End-Use Sector, 2006,” p. 5 in Emissions of Greenhouse Gases in the United States 2006, DOE/EIA-0573 (Washington, DC: EIA, November 2007). Approximately 24 percent of the total reported gases (mostly from industrial sources) were excluded from the analysis, either because they were far removed from consumer activity or to avoid double counting in the model.
‡Ibid.
§U.S. Department of Energy, Energy Information Administration (EIA), “Official Energy Statistics from the U.S. Government: Households, Buildings, Industry, and Vehicles, 2001.”
**U.S. Department of Transportation, Research and Innovative Technology Administration, Bureau of Transportation Statistics, “U.S. Government Energy Consumption by Agency and Source (Trillion Btu),” table 4-19 in “National Transportation Statistics 2008.” Available at http://www.bts.gov/publications/national_transportation_statistics. Although most of the data incorporated in the analysis are from 2006, the results of this study also reflect emissions coefficients based in part on data from 2002 and 2003 for the transportation sector.
*For our purposes, emissions coefficients refer to ratios of CO2 emissions per dollar of industry output.
†For a more comprehensive discussion of the IMPLAN model, see the IMPLAN discussion that follows in this appendix.
‡The most current version of the software is IMPLAN Version 3. It was released in November 2009, after the current analysis was complete. For additional information, see www.implan.com.
§ For a current listing of registered IMPLAN users, see www.implan.com.
* Our analysis accounts for 509 separate industry sectors, as 19 sectors are not considered part of the traditional industry analysis. This includes sectors such as scrap, federal and state enterprises (that are part of other sectors), the U.S. Postal Service, domestic services, and used and secondhand goods.
† For a more comprehensive listing of data sources used to develop IMPLAN data files, see www.implan.com.
*Total household demand (consumer expenditures) in 2006 was actually just over $10 trillion; however, adjustments were made using margins to distribute wholesale, transportation, and retail, to avoid double counting.
†It should also be noted that IMPLAN is a so-called closed model of the U.S. economy; in other words, it assumes that expenditures from outside the United States have the same emissions profile as those within the United States.
*For our purposes we utilize Type I multipliers, which incorporate direct and indirect impacts only. Our analysis does not include induced impacts, which are from increased household spending caused by the employment and income gains due to the direct and indirect impacts.
†NAICS replaced the Standard Industrial Classification (SIC) system used by the federal government for classifying industries by type.
‡ To derive the direct emissions impacts, each dollar of consumer spending (margined) was matched directly with the emissions coefficient for each industry sector. To derive the indirect impacts, we again used the industry coefficients and the consumer expenditures (margined) but applied them to the Type I multipliers (noted earlier) developed with IMPLAN. Each dollar of consumer spending was multiplied by an industry-specific Type I multiplier to calculate the indirect output. This total was then multiplied by the emissions coefficient for each industry. The direct and indirect emissions impacts were summed to calculate a total impact for each expenditure.