Emissions Trading in the U.S.: Experience, Lessons, and Considerations for Greenhouse Gases

In recent years, emissions trading has become an important element of programs to control air pollution. Experience indicates that an emissions trading program, if designed and implemented effectively, can achieve environmental goals faster and at lower costs than traditional command-and-control alternatives. Under such a program, emissions are capped but sources have the flexibility to find and apply the lowest-cost methods for reducing pollution. A cap-and-trade program is especially attractive for controlling global pollutants such as greenhouse gases because their warming effects are the same regardless of where they are emitted, the costs of reducing emissions vary widely by source, and the cap ensures that the environmental goal is attained.

Report authors Denny Ellerman and Paul Joskow of the Massachusetts Institute of Technology and David Harrison of National Economic Research Associates, Inc. review six diverse U.S. emissions trading programs, drawing general lessons for future applications and discussing considerations for controlling greenhouse gas emissions. The authors derive five key lessons from this experience. First, emissions trading has been successful in its major objective of lowering the cost of meeting emission reduction goals. Second, the use of emissions trading has enhanced—not compromised—the achievement of environmental goals. Third, emissions trading has worked best when the allowances or credits being traded are clearly defined and tradable without case-by-case certification. Fourth, banking has played an important role in improving the economic and environmental performance of emissions trading programs. Finally, while the initial allocation of allowances in cap-and-trade programs is important from a distributional perspective, the method of allocation generally does not impair the program’s potential cost savings or environmental performance.

With growing Congressional interest in programs to address climate change—including the recent introduction of economy-wide cap-and-trade legislation controlling greenhouse gas emissions—the application of lessons learned from previous emissions trading programs is timely. In addition to this review, the Pew Center is simultaneously releasing a complementary report, Designing a Mandatory Greenhouse Gas Reduction Program for the U.S., which examines additional options for designing a domestic climate change program.

The authors and the Pew Center are grateful to Dallas Burtraw and Tom Tietenberg for reviewing a previous draft of this report. The authors also wish to acknowledge Henry Jacoby, Juan-Pablo Montero, Daniel Radov, and Eric Haxthausen for their contributions to various parts of the report, and James Patchett and Warren Herold for their research assistance.

Executive Summary

Emissions trading has emerged over the last two decades as a popular policy tool for controlling air pollution. Indeed, most major air quality improvement initiatives in the United States now include emissions trading as a component of emissions control programs. The primary attraction of emissions trading is that a properly designed program provides a framework to meet emissions reduction goals at the lowest possible cost. It does so by giving emissions sources the flexibility to find and apply the lowest-cost methods for reducing pollution. Emission sources with low-cost compliance options have an incentive to reduce emissions more than they would under command-and-control regulation. By trading emission credits and allowances to high-cost compliance sources, which can then reduce emissions less, cost-effective emission reductions are achieved by both parties. When inter-temporal trading is allowed, sources can also reduce emissions early, accumulating credits or allowances that can be used for compliance in future periods if this reduces cumulative compliance costs. Accordingly, cap-and-trade programs achieve the greatest cost savings when the costs of controlling emissions vary widely across sources or over time. In practice, well-designed emissions trading programs also have achieved environmental goals more quickly and with greater confidence than more costly command-and-control alternatives.

Emissions trading has achieved prominence beyond the United States largely in the context of discussions regarding implementation of the Kyoto Protocol, a proposed international agreement to control emissions of carbon dioxide (CO2) and other greenhouse gases. The Kyoto Protocol provides for the use of various emissions trading mechanisms at the international level. Some countries already are developing emissions trading programs while the process of ratifying the Protocol moves forward. Both the United Kingdom and Denmark have instituted greenhouse gas (GHG) emissions trading programs, and, in December 2002, the European environment ministers agreed on the ground rules for a European Union trading program that would begin in 2005 for large sources of CO2 emissions (and later for other GHG emissions). Indeed, proposals to control GHG emissions in the United States also include the use of emissions trading.

The theoretical virtues of emissions trading have been recognized for many decades—the basic elements were outlined in Coase (1960) and elaborated in Dales (1968)—but actual emissions trading programs have been brought from the textbook to the policy arena mostly in the last decade. It is important to recognize, however, that while properly designed emissions trading programs can reduce the cost of meeting environmental goals, experience does not indicate that significant emissions reductions can be obtained without costs. Emissions trading can be an effective mechanism for controlling emissions by providing sources with the flexibility to select the lowest-cost opportunities for abatement, but it does not make costs disappear. Moreover, emissions trading programs must be designed properly in order to realize their potential cost-reduction and environmental compliance goals. As with any emissions control program, poor design is likely to lead to disappointing results.

Experience with emissions trading, including both the design and operation of trading programs, provides a number of general lessons for future applications. This report reviews the experience with six emissions trading programs with which one or more of the authors have considerable experience:

  • The early Environmental Protection Agency (EPA) Emissions Trading programs that began in the late 1970s;
  • The Lead Trading program for gasoline that was implemented in the 1980s;
  • The Acid Rain program for electric industry sulfur dioxide (SO2) emissions and the Los Angeles air basin (RECLAIM) programs for both nitrogen oxides (NOx) and SO2 emissions, all of which went into operation in the mid-1990s;
  • The federal mobile source averaging, banking, and trading (ABT) programs that began in the early 1990s; and
  • The Northeast NOx Budget trading program, which began operations in the late 1990s.

Based on this experience, this report identifies and discusses five general lessons concerning the design and implementation of emissions trading programs, and two considerations of particular relevance for GHG applications.

General Lessons from Experience with Emissions Trading

Emissions trading has been successful in its major objective of lowering the cost of meeting emission reduction goals. Experience shows that properly designed emissions trading programs can reduce compliance costs significantly compared to command-and-control alternatives. While it is impossible to provide precise measures of cost savings compared to hypothetical control approaches that might have been applied, the available evidence suggests that the increased compliance flexibility of emissions trading yields costs savings of as much as 50 percent.

The use of emissions trading has enhanced—not compromised—the achievement of environmental goals. While some skeptics have suggested that emissions trading is a way of evading environmental requirements, experience to date with well-designed trading programs indicates that emissions trading helps achieve environmental goals in several ways.

For one thing, the achievement of required emission reductions has been accelerated when emission reduction requirements are phased-in and firms are able to bank emissions reduction credits. The Lead Trading program for gasoline, the Acid Rain program for the electric industry, the federal mobile source ABT programs, and the Northeast NOx Budget programs each achieved environmental goals more quickly through these program design features. Moreover, giving firms with high abatement costs the flexibility to meet their compliance obligations by buying emissions allowances eliminates the rationale underlying requests for special exemptions from emissions regulations based on “hardship” and “high cost.” The reduction of compliance costs has also led to instances of tighter emissions targets, in keeping with efforts to balance the costs and benefits of emissions reductions. Finally, properly designed emissions trading programs appear to provide other efficiency gains, such as greater incentives for innovation and improved emissions monitoring.

Emissions trading has worked best when allowances or credits being traded are clearly defined and tradable without case-by-case pre-certification. Several different types of emissions trading mechanisms have been implemented. Their performance has varied widely, and these variations illuminate the key features of emissions trading programs that are most likely to lead to significant cost savings while maintaining (or exceeding) environmental goals.

The term “emissions trading” is used, often very loosely, to refer to three different types of trading programs: (1) reduction credit trading, in which credits for emission reductions must be pre-certified relative to an emission standard before they can be traded; (2) emission rate averaging, in which credits and debits are certified automatically according to a set average emission rate; and (3) cap-and-trade programs, in which an overall cap is set, allowances (i.e., rights to emit a unit) equal to the cap are distributed, and sources subject to the cap are required to surrender an allowance for every unit (e.g., ton) they emit.

The turnaround in perception of emissions trading over the last decade—from a reputation as a theoretically attractive but largely impractical approach to its acceptance as a practical framework for meeting air quality goals in a cost-effective manner—largely reflects the increased use of averaging and cap-and-trade type programs. The performance of the early EPA reduction credit programs was very poor and gave “emissions trading” a bad name. These early EPA programs emphasized case-by-case pre-certification of emission reductions and were characterized by burdensome and time-consuming administrative approval processes that made trading difficult. The averaging and cap-and-trade programs have been much more successful. While the use of cap-and-trade or averaging does not guarantee success, and the problems with the reduction credit-based approach can be reduced by good design, avoiding high transaction costs associated with trade-by-trade administrative certification is critical to the success of an emissions trading program. The success of any emissions trading program also requires several additional elements: emissions levels must be readily measured, legal emissions rates or caps must be clearly specified, and compliance must be verified and enforced aggressively.

Banking has played an important role in improving the economic and environmental performance of emissions trading programs. Early advocates of emissions trading tended to emphasize gains from trading among participants (i.e., low-cost compliance sources selling credits and allowances to high-cost compliance sources) in the same time period. The experience with the programs reviewed here indicates that inter-temporal trading also has been important. The form that inter-temporal trading most often takes is credit or allowance banking, i.e., reducing emissions early and accumulating credits or allowances that can be used for compliance in future periods. Banking improves environmental performance and reduces cumulative compliance costs. Moreover, it has been particularly important in providing flexibility to deal with many uncertainties associated with an emissions trading market—production levels, compliance costs, and the many other factors that influence demand for credits or allowances. Indeed, the one major program without a substantial banking provision, the Los Angeles RECLAIM program, appears to have suffered because of its absence.

The initial allocation of allowances in cap-and-trade programs has shown that equity and political concerns can be addressed without impairing the cost savings from trading or the environmental performance of these programs. Because emissions allowances in cap-and-trade programs are valuable, their allocation has been perhaps the single most contentious issue in establishing the existing cap-and-trade programs. However, the ability to allocate this valuable commodity and thereby account for the economic impacts of new regulatory requirements has been an important means of attaining political support for more stringent emissions caps. Moreover, despite all the jockeying for allowance allotments through the political process, the allocations of allowances to firms in the major programs have not compromised environmental goals or cost savings. The three cap-and-trade programs that have been observed so far all have relied upon “grandfathering,” i.e., distributing allowances without charge to sources based upon historical emissions information, which generally does not affect firms’ choices regarding cost-effective emission reductions and thus the overall cost savings from emissions trading. There are other methods of allocating initial allowances—such as auctioning by the government and distributing on the basis of future information—that can affect cost savings and other overall impacts; but the major effects of the initial allocation are to distribute valuable assets in some manner and to provide effective compensation for the financial impacts of capping emissions on participating sources.

Considerations for Greenhouse Gas Control Programs

Emissions trading seems especially well-suited to be part of a program to control greenhouse gas emissions. The emissions trading programs reviewed for this report generally have spatial or temporal limitations because sources of the pollutants included in these programs—such as lead, SO2, and NOx—may have different environmental impacts depending on the sources’ locations (e.g., upwind or downwind from population centers) and the time of the emissions (e.g., summer or winter). The concerns of trading programs associated with climate change are different because greenhouse gases are both uniformly mixed in the earth’s atmosphere and long-lived. The effects of GHG emissions thus are the same regardless of where the source is located and when the emissions occur (within a broad time band). This means that emissions trading can be global in scope as well as inter-temporal, creating an opportunity for the banking of emission credits, which allows emissions to vary from year to year as long as an aggregate inter-temporal cap is achieved.

Emissions trading is also well suited for GHG emissions control because the costs of reducing emissions vary widely between individual greenhouse gases, sectors, and countries, and thus there are large potential gains from trade. While other market-based approaches, such as emissions taxes, also would provide for these cost savings, the cap-and-trade version of emissions trading has the further advantage of providing greater certainty that an emission target will be met. Moreover, GHG emissions generally can be measured using relatively inexpensive methods (e.g., fuel consumption and emission factors), rather than the expensive continuous emissions monitoring required for some existing trading programs.

Furthermore, emissions trading provides important incentives for low-cost compliance sources initially outside the program to find ways to participate, and thereby further reduce costs. This opt-in feature is useful because an environmentally and cost-effective solution for reducing concentrations of greenhouse gases should be comprehensive and global, whereas initial controls on GHG emissions will—for political reasons—likely be limited, if not to certain sectors and greenhouse gases, then almost certainly to a restricted number of countries. Therefore, an important criterion for initial measures is that they be able to induce participation by sources not yet controlled. The markets created by cap-and-trade programs provide incentives for sources outside the trading program to enter if they can provide reductions more cheaply than the market prices—a common feature of any market. Although, as discussed below, the voluntary nature of these incentives can create some problems, the ability to induce further participation is an important reason to include a market-based approach initially. Indeed, it is hard to imagine how command-and-control regulations or emissions taxes could provide similar incentives to non-participants to adopt new measures to reduce greenhouse gas emissions.

Opt-in or voluntary features have a strategic role that is likely to warrant their inclusion despite the potential problems associated with them. Experience with allowing sources not covered by mandatory emissions trading programs to “opt-in,” i.e., to voluntarily assume emissions control obligations and to participate in the emissions market, has revealed a trade-off. Setting clear baselines for opting-in lowers transactions costs and thus encourages participation; but some of this participation consists of credits for calculated “reductions” that are unrelated to the trading program and actually lead to increased emissions. For example, in the Acid Rain Program, evidence indicates that many of the voluntary participants received credits for having emissions below the pre-specified baseline even though they took no abatement actions. The simple emissions baseline had been set higher than these facilities’ actual emissions, so at least some of the credits they received did not represent real emissions reductions.

This experience suggests that the decision whether or not to include opt-in provisions should be determined by weighing the cost-saving benefits against the emissions-increasing potential. For greenhouse gases, the potential cost-savings benefits of including a voluntary element in the mandatory program are large because initial efforts are not likely to be comprehensive and global, as they must be eventually to achieve their environmental goals and be cost-effective. Opt-in provisions also have value in improving measurement and monitoring techniques, in familiarizing participants with the requirements of emissions trading, and more generally with inducing participation of sources outside the trading program that can offer cheaper abatement. As a result, allowing participants outside the mandatory GHG emissions control program to opt-in has a strategic value that has not been prominent in other opt-in programs. Indeed, it should be possible to learn from existing experience with opt-in programs how to reduce adverse effects while achieving cost-savings.

Viewed from a broad historical perspective, emissions trading has come a long way since the first theoretical insights forty years ago and the first tentative application almost a quarter of a century ago. Although still not the dominant form of controlling pollution in the United States or elsewhere, emissions trading is being included in an increasing number of programs and proposals throughout the world, and its role seems likely to expand in the future.


Emissions trading has emerged as a practical framework for introducing cost-reducing flexibility into environmental control programs and reducing the costs associated with conventional command-and-control regulation of air pollution emissions. Over the last two decades considerable experience with various forms of emissions trading has been gained, and today nearly all proposals for new initiatives to control air emissions include some form of emissions trading. This report has attempted to summarize that experience and to draw appropriate lessons that may apply to proposals to limit GHG emissions. In doing so, we hope that the reader has gained a better understanding of emissions trading and the reasons for its increasing importance as an instrument for addressing environmental problems.

Six diverse programs constitute the primary U.S. experience with air emissions trading. The EPA’s early attempts starting in the late 1970s to introduce flexibility into the Clean Air Act through netting, offsets, bubbles, and banking were not particularly encouraging. Most of the potential trades, and economic gains from trading, in these early systems were frustrated by the high transaction costs of certifying emission reductions. The first really successful use of emissions trading occurred in the mid-1980s when the lead content in gasoline was reduced by 90 percent in a program that allowed refiners to automatically earn credits for exceeding the mandated reductions in lead content and to sell those credits to others or bank them for later use.

The Acid Rain or SO2 allowance trading program for electricity generators, which has become by far the most prominent experiment in emissions trading, was adopted in 1990 and implemented beginning in 1995. This innovative program introduced a significantly different form of emissions trading, known as cap-and-trade, in which participants traded a fixed number of allowances—or rights to emit—equal in aggregate number to the cap, instead of trading on the differences from some pre-existing or external standard as had been the case in the early EPA trading programs and the lead phase-down program.

Another cap-and-trade program, the RECLAIM program for both SO2 and NOx emissions, was developed and implemented at the same time as the Acid Rain program by the regulatory authority in the Los Angeles Basin as part of its efforts to bring that area into attainment with National Ambient Air Quality Standards. The RECLAIM program is the first instance of emissions trading both supplementing and supplanting a pre-existing command-and-control structure that theoretically was capable of achieving the same environmental objective. The standards of the pre-existing command-and-control system largely determined the level of the cap, and the program’s ten-year phase-in design and trading provided the flexibility that led to the achievement of environmental goals that had been previously elusive. RECLAIM also introduced trading among different sectors.

The 1990 Amendments to the Clean Air Act also provided enabling legislation for two other emissions trading programs. Emissions from mobile sources were more effectively and efficiently controlled by the introduction of mobile source averaging, banking, and trading programs. The mobile source programs followed the example of the lead phase-down program by allowing firms to create credits automatically for any reductions beyond a required uniform emission standard and to use these credits in lieu of more costly reductions elsewhere or later within the company and to sell them. The 1990 Amendments also provided the mechanism that encouraged states in the Northeastern United States to adopt cap-and-trade programs to control NOx emissions that contributed to ozone non-attainment in that region of the country. As was the case in the RECLAIM program, emissions trading was adopted as a means to attain environmental objectives more quickly and cost-effectively than had proved possible through conventional command-and-control regulation.

There are many lessons to be gained from the experience with these six programs, but the five most important lessons can be summarized as follows. First, the major objective of emissions trading, lowering the cost of meeting emission reduction goals, has been achieved in most of these programs. Second, emissions trading has not compromised the achievement of the environmental goals embodied in these programs. If anything, and this is perhaps surprising, the achievement of those goals has been enhanced by emissions trading. Third, emissions trading has worked best in reducing costs and achieving environmental goals when the credits being traded are clearly defined and readily tradable without case-by-case certification. Fourth, temporal flexibility, i.e., the ability to bank allowances, has been more important than generally expected, and the ability to bank has contributed significantly to accelerating emission reductions and dampening price fluctuations. Fifth, the initial allocation of allowances in cap-and-trade programs has shown that equitable and political concerns can be met without impairing either the cost savings from trading or the environmental performance of these programs. In addition, the success of any emissions trading program requires that emissions levels can be readily measured and compliance verified and enforced.

All of these five lessons are relevant when considering the use of emissions trading in a program aimed at reducing GHG emissions. In fact, emissions trading seems especially appropriate for this environmental problem. Greenhouse gas emissions mix uniformly and remain in the atmosphere for a long time. Thus, it matters little where or when the emissions are reduced, as long as the required cumulative reductions are made. These specific characteristics of GHG emissions eliminate two of the concerns that have limited the scope of emissions trading in many other programs.

Although an effective GHG mitigation program must eventually be global in scope and comprehensive in its coverage of pollutants and economic sectors, the likelihood that control efforts will be limited initially to the richer countries, the more easily measurable gases, and perhaps to certain sectors of the economy introduces another consideration. The ability to induce initially uncapped sources to participate voluntarily in the early efforts will reduce costs and prepare the way for extending the caps. Thus, providing opportunities to opt-in for uncapped sources that can reduce emissions at lower cost than those within the cap has a strategic value beyond the potential cost savings. Although some existing programs with voluntary provisions have revealed opportunities for misuse, these problems can be managed more successfully now with the benefit of experience. The strategic value of opt-in provisions in any GHG emission control program makes their inclusion highly desirable.

Emissions trading has come a long way since the first theoretical insights forty years ago and the first tentative application almost a quarter of a century ago. Since then, the use of emissions trading has expanded steadily and significant experience has been gained. Although not the dominant form of controlling pollution in the United States or elsewhere, emissions trading now seems firmly established as a valuable instrument and its future use seems sure to increase. Our review of experience over the past quarter century suggests that this trend toward greater use of emissions trading will improve the performance of environmental regulation, including efforts to control GHG emissions.