In 2009, an interagency workgroup composed of members from six federal agencies and various White House offices was convened to develop estimates of the “social cost of carbon” for regulatory impact analysis. This process represents a step forward in increasing transparency and consistency in benefit–cost analyses of federal regulatory actions and can serve as a model for future revisions of the social cost of carbon estimates to keep pace with the evolving state of the science.
Estimating the “Social Cost of Carbon” for Regulatory Impact Analysis
Charles Griffiths, Elizabeth Kopits, Alex Marten, Chris Moore, Steve Newbold, and Ann Wolverton
November 8, 2010
The Need for Social Cost of Carbon Estimates
Under Executive Order 12866, agencies in the executive branch of the U.S. federal government are required “to assess both the costs and the benefits of the intended regulation and, recognizing that some costs and benefits are difficult to quantify, propose or adopt a regulation only upon a reasoned determination that the benefits of the intended regulation justify its costs.” As the government begins regulating carbon emissions, estimates of foregone damages are needed to quantify benefits. One way to accomplish this is by assessing the “social cost of carbon” (SCC). The SCC is the present value of the stream of future economic damages associated with an incremental increase (by convention, one metric ton) in carbon dioxide (CO2) emissions in a particular year. It is intended to be a comprehensive measure, including economic losses due to changes in agricultural productivity, human health risks, property damages from increased flood frequencies, the loss of ecosystem services, and so on. The SCC is a marginal value so it may not be accurate for valuing large changes in emissions. Many U.S. government regulations will lead to relatively small reductions in cumulative global emissions, however, so for these regulations the SCC is the appropriate shadow value for estimating the economic benefits of CO2 reductions.
Most published estimates of the SCC have been derived from “integrated assessment models” (IAMs) that combine climate processes, economic growth, and feedbacks between the two in a single modeling framework. IAMs used to estimate the SCC rely on highly simplified representations of the potential economic damages from climate change and are necessarily limited by the current state of the climate economics literature, which continues to expand rapidly. Despite the inherent uncertainties in models such as these, they are the best tools currently available for estimating the SCC.
The Interagency SCC Workgroup Process
The 2009–2010 interagency process to develop SCC estimates was the first U.S. federal government effort to promote consistency in the way that agencies calculate the social benefits of reducing CO2 emissions in regulatory impact analyses. Prior to 2008, reductions in CO2 emissions were not valued in federal benefit–cost analyses. From 2008 to 2009, SCC estimates were used in some regulatory analyses, but the values employed varied substantially among agencies. By convening the interagency workgroup, the Obama administration sought to improve the accuracy and consistency of how agencies value reductions in CO2 emissions through the development of a transparent methodology for selecting a set of SCC estimates. Recognizing the immediate need for SCC estimates and considering the time that would be required to develop original estimates, the workgroup issued a set of preliminary or interim SCC estimates in July of 2009 based on an assessment of published estimates. Most of the published SCC estimates are derived from one of three IAMS: William Nordhaus’s “Dynamic Integrated Climate Economy” (DICE) model, Richard Tol’s “Climate Framework for Uncertainty, Negotiation, and Distribution” (FUND) model, and Chris Hope’s “Policy Analysis for the Greenhouse Effect” (PAGE) model. The results of this preliminary effort were offered for public comment in the joint U.S. Environmental Protection Agency (EPA) and Department of Transportation (DOT) fuel economy and CO2 tailpipe emission proposed rulemaking.
After the release of the interim values, the group developed improved SCC estimates that better reflect key uncertainties and model differences, using a defensible set of input assumptions firmly grounded in the existing literature. The resulting range is based on estimates from DICE, FUND, and PAGE as applied to five socioeconomic and emissions scenarios drawn from the Stanford Energy Modeling Forum, all given equal weight. In all models, no weighting was applied to damages in various regions or countries. That is, a $1 loss in GDP to a poor country was valued the same as a $1 loss to a rich country. Because there is disagreement among experts about the appropriate way to transform the stream of economic damages over time into a single value, the present value of the time path of global damages in each model-scenario combination was calculated using three discount rates—5, 3, and 2.5 percent per year. This range encompasses values consistent with some arguments for using market rates, for reflecting uncertainty in rates over time, or for using below market rates to avoid unfairly penalizing future generations. Finally, in a step toward more formal uncertainty analysis, all model runs employed a probabilistic representation of climate sensitivity (in addition to other parameters in PAGE and FUND). Other features of the IAMs were left unchanged, relying on the developers’ own best estimates.
From these model runs the workgroup selected four SCC estimates to reflect the global damages caused by CO2 emissions: $5, $21 (central value), $35, and $65 for 2010 emissions reductions (in 2007 U.S. dollars). The first three estimates are based on the average SCC across the three models and five socioeconomic and emissions scenarios for the 5, 3, and 2.5 percent discount rates, respectively. The $65 value—the 95th percentile of the SCC distribution at a 3 percent discount rate—was chosen to represent potential higher-than-expected impacts from temperature change. The SCC estimates grow over time at rates endogenously determined by the models. For instance, with a discount rate of 3 percent, the mean SCC estimate increases from $21 per ton of CO2 in 2010 to $24 in 2015 and $26 in 2020. To put these numbers in perspective, an SCC of $21 suggests that the cost to society of burning one additional ton of coal or gallon of gasoline is roughly $40 and 18 cents, respectively.
To date, the SCC values developed by the group have been used to quantify the benefits of reducing CO2 emissions in several final rules, including multiple U.S. Department of Energy (DOE) energy efficiency standards and the joint EPA-DOT fuel economy and CO2 tailpipe emission final rule.
The interagency workgroup presented these SCC estimates with a clear acknowledgement of the many uncertainties involved. It recommended the estimates be revisited on a regular basis and revised as needed to reflect the growing body of scientific knowledge regarding climate change impacts and their potential economic damages. The process also helped to highlight several key areas in need of additional research. First, there are a number of important gaps in our understanding of how the anticipated physical impacts of climate change can be translated into economic damages. For example, the IAMs used by the workgroup make substantially different assumptions about the potential economic damages of temperature changes on agricultural production, human health, and other key categories, and none of the models accounts for the effect of CO2 emissions on ocean acidification and the subsequent harm to fisheries and other marine resources. It also is clear that these models do not adequately represent the potential catastrophic damages from extreme climate risks, and how to better those representations remains an important area of research. Second, there is a need for a more complete representation of the numerous sources of uncertainty within the physical and economic components of the models. Third, some aspects of these IAMs are based on older climate science and economic studies that may have been superseded by more recent research. Increasing the frequency by which modeling improvements and new research on climate damages is incorporated into IAMs should be a high priority in future efforts. Fourth, there is a need to capture and more transparently represent a wider range of defensible assumptions regarding economic behavior (for example, adaptation and technology diffusion) in economic damage estimates than is typically captured in any one IAM. Fifth, while approaches to discounting the future benefits and costs of policies with very long time horizons has received much attention by scholars in recent decades, it is not clear that the key theoretical findings from this research are adequately incorporated into real-world policy evaluations. And finally, most existing IAMs are designed to estimate the SCC alone, so model extensions are needed to cover other greenhouse gases.
EPA and DOE are beginning to respond to some of these identified needs through a pair of upcoming workshops focused on critical modeling issues in integrated assessment modeling and research on climate impacts and damages in specific sectors. EPA researchers are also exploring the development of a more generalized modeling framework for climate change assessment and valuation.
Charles Griffiths, Elizabeth Kopits, Alex Marten, Chris Moore, Steve Newbold, and Ann Wolverton are economists in EPA’s National Center for Environmental Economics. All the authors actively participated in the interagency SCC discussion, and Elizabeth Kopits and Ann Wolverton served as senior economists on the President’s Council of Economic Advisers during 2009 and/or 2010. The views expressed in this paper are those of the authors and do not necessarily represent those of the U.S Environmental Protection Agency.