Grantee Research Project Results
Final Report: Efficiency and Distributional Consequences of the Allocation of Tradable Emission Allowances
EPA Grant Number: R828628Title: Efficiency and Distributional Consequences of the Allocation of Tradable Emission Allowances
Investigators: Burtraw, Dallas , Pizer, Billy , Fischer, Carolyn , Carlson, Curtis , Palmer, Karen
Institution: Resources for the Future
EPA Project Officer: Chung, Serena
Project Period: October 1, 2000 through September 30, 2002 (Extended to September 30, 2003)
Project Amount: $251,000
RFA: Market Mechanisms and Incentives for Environmental Management (2000) RFA Text | Recipients Lists
Research Category: Environmental Justice
Objective:
The objective of this research project was to address a set of issues relevant to the design of market-based programs for environmental protection. The previous generation of research highlighted potential cost savings and the feasibility of market-based programs compared to technology standards applied uniformly to a sector or industry. The predicted improvement in cost effectiveness stems from expanded flexibility in determining the means of compliance as well as the consequent reallocation of compliance efforts to minimize costs while obtaining environmental goals. Experience with these programs has been positive in the sense that anticipated cost savings from emission trading in most instances have been realized, and environmental integrity has been maintained.
This research project addressed a second generation of issues that have surfaced as economists and policymakers have gained experience with market-based approaches in general and tradable permit programs in particular. These issues involve analytical and practical questions about how emission permits are allocated and how the decision about allocation affects the social cost and environmental impact of regulation. Equally important and related is the distribution of costs and potential transfer of wealth that is associated with permit allocation methods.
This research project addressed these new issues using analytical and numerical methods within a general equilibrium model and a detailed partial equilibrium model. This research project resulted in several academic papers currently under review at journals. This research project also led to many presentations before academic audiences and general audiences and to many meetings with agency and legislative staff and stakeholder groups who were interested in the implications of this work for public policy. The research also contributed to the development of modeling capability that has been employed recently by the U.S. Environmental Protection Agency (EPA) in a closely related research project and that currently is being used in projects with the New York State Energy Research and Development Authority and the State of Maryland Power Plant Research Program.
In summary, this research project provided an analysis that can help guide policymakers on approaches to the allocation of emission allowances, and the research suggests that decisions in this matter have tremendous significance. Furthermore, normative guidance is achieved that has attracted the interest of a diverse community of interest groups.
Summary/Accomplishments (Outputs/Outcomes):
Almost all of the existing emission trading programs have allocated emission permits to affected facilities free of charge. The most popular approach has been grandfathering allowances (e.g., allocating allowances according to some formula based on historic emissions or historic production, as in the case of SO2 emission allowance trading).
Another approach would be to distribute emission allowances through a revenue-raising auction. An auction is used to allocate a small fraction of the allowances in the SO2 program, but revenues are returned to industry according to the proportion of the original allocation. A revenue-raising auction rarely has been used in practice, primarily because of political considerations, but has received strong support in recent economic studies. To our knowledge, the first use of a revenue-raising auction to distribute emission allowances will occur in the State of Virginia's distribution of allowances for new sources within the NOx trading program. The Bush Administration's original Clear Skies Proposal (S. 1844) also proposed an auction as part of its approach to distributing emission allowances, although that was removed from the proposed legislation in the fall of 2003. The Jeffords bill (S. 366) also proposed an auction, with revenues returned to households.
A third approach currently being used by some States in the Ozone Transport Commission NOx Trading Program, which will be used by others in the State Implementation Plan (SIP) Call Region Program, is an updating of allocations based on recent input, emissions, or production levels. This approach also is used in the Carper bill (S. 843). A similar approach is tradable performance standards, which would allocate permits based on output multiplied by a fixed average emission rate (rather than shares in a fixed emissions cap). Firms would buy (sell) permits for their emissions to the extent their emission rates were below (above) the standard. An explicit program of tradable performance standards was implemented for the phasedown of lead in the United States. A portfolio standard for renewable-resource generation would create similar incentives. Another similar approach has been used in the NOx program in Sweden, where revenues from a tax on emissions are recycled to the affected segments of the industry. (In 1982, the U.S. EPA set an interrefinery average for lead usage among importers and refineries producing leaded gasoline. Refineries using less lead per unit than the standard could sell these credits to others using more than average. The standard was lowered in a series of steps, and permit banking was allowed from 1985 until the end of 1987.)
The Efficiency of Approaches To Allocating Emission Allowances in a Static Setting Can Vary Widely
Two separate methodologies used in this research project indicate that there may be important efficiency consequences stemming from the way that emission allowances are initially distributed. Both methods identify efficiency issues associated with preexisting distortions away from economic efficiency because of differences between price and marginal cost. The sources of the distortions, however, differ (Burtraw, 2003; Burtraw and Palmer, 2003).
One method explores the role of preexisting distortions away from economic efficiency in factor markets (labor, capital) because of the presence of taxes on labor or capital income. The inefficiency stems from the fact that taxes cause product prices to be above the marginal opportunity cost to society of using those resources. The analytical and computable general equilibrium simulation models estimate the potential efficiency consequences of different approaches to allocation, taking account of preexisting taxes in factor markets.
The second method explores the role of preexisting differences between price and marginal cost in product markets, especially in the electricity market where the difference between price and marginal cost is endemic. The method used in examining the electricity market relies on a detailed partial equilibrium simulation model that characterizes regulatory institutions within the industry to estimate both efficiency and distributional consequences of approaches to allocation.
In addition to efficiency issues stemming from the difference between price and marginal cost, the method of distributing allowances can have efficiency consequences when the method introduces an incentive to change output. Under output-based allocation, the combination of an emissions price with an output subsidy shifts emissions control efforts toward a lower emission rate and shifts control away from contraction in output. The resulting industry equilibrium is expected to have higher marginal costs of control and lower output prices compared to the social optimum given any targeted level of abatement. In the case of a cap-and-trade system, allocating by output leads to higher permit prices. In the case of a performance standard, to achieve a particular emissions level, the standard must be stricter than would be required in an optimal policy (i.e., the emission rate that would prevail under a Pigouvian emissions tax). Alternatively, setting the standard at that Pigouvian level will result in more emissions than in the optimal case because total output is greater.
Fischer (2001) addresses these issues by analyzing schemes that rebate revenues based on output shares: tradable performance standards, an emissions tax with market-share rebates, and tradable permits with output-based allocation. Political pressure often exists to earmark environmental tax revenues or profits from permits (rents) to the industry affected by the regulation. All three policies accomplish this by effectively combining a tax on emissions with a subsidy to output. The result is a shifting of emissions control efforts toward greater emissions rate reduction and less output contraction, with higher marginal costs of control and lower output prices compared to the social optimum, given any targeted level of abatement. These welfare costs depend on the degree of output substitutability and likely are to be much larger in the long run. The research concludes that some political and market-failure justifications may exist, but policymakers should carefully consider industry characteristics before engaging in output-based rebating.
Fischer (2001) laid out the fundamental conceptual issues of output-based allocation and analyzed the effects under perfect competition. It is complemented by numerical modeling in Fischer and Fox (working paper, 2002). That paper addresses the question: What are the efficiency tradeoffs of output-based allocation systems? The output subsidy implicit in these policies shifts emissions control efforts toward more emissions rate reduction and less output substitution, raising overall abatement costs compared to the cost-effective outcome. The subsidy, however, also may mitigate the effects of interactions with preexisting market distortions such as taxes. With multiple polluting sectors, different subsidies implicit in output-based allocation also affect the distribution of abatement effort across sectors as well as the terms of trade with foreign sectors. The relative magnitude of these distortions is affected by the rule for distributing permit endowments across sectors, each of which subsequently allocates among its firms according to output.
Fischer and Fox (working paper, 2002) use a general equilibrium model (Global Trade Analysis Project) to estimate the welfare impacts of four allocation scenarios of a national policy of CO2 reduction: (1) auctioned emission permits with revenues recycled to lower preexisting taxes; (2) auctioned emission permits with revenues distributed in lump sum to households; (3) output-based allocation with sector permit endowments based on historical emissions; and (4) output-based allocation with sector permit endowments based on historical shares of gross domestic product. Welfare comparisons are made with respect to scenario 1, which is the optimal policy. We find that output-based allocation may or may not improve upon lump-sum distribution (scenario 2) in a second-best setting. The results show that scenario 3 is worse than scenario 2, as the greater subsidy to more pollution-intensive sectors dramatically shifts the abatement burden. Scenario 4, however, being closer to a value-added rebate and thus more like the optimal tax rebate, outperforms scenario 2. Fischer and Fox (working paper, 2002) thus illustrate the situations where output-based allocation is likely to impose a smaller welfare cost than auctioned permits and then discuss the distributional effects on sectors and factors of likely interest to policymakers.
Burtraw and Palmer (2003) provided an overview of the choice of modeling framework for examining this issue, reviewing results from the computable general equilibrium literature and contrasting them with results from the partial equilibrium literature. The results from the partial equilibrium literature emphasize market structure and economic behavior within markets, and these results are summarized below. Burtraw (2003) provides a more narrow characterization of the issue, pertinent for the electricity sector and intended for an audience of lawyers.
Market Structure is Expected to Play a Large Role in the Relative Performance of Instruments
To investigate the issue of market structure and deviations from competitive market assumptions, we use two different methods and two modeling platforms. One is an analytical modeling framework and the second is a detailed simulation model of the electricity sector.
Regulated firms may have significant market shares in the sense that their output represents a nonnegligible fraction of the total output of all participants in the regulatory program. Policies no longer operate identically across all firms if some firms are regulated and some firms are not. For large, potentially regulated firms, the firm realizes that its own production behavior can affect the value of all of its emission allocations. For example, larger firms have more slices of the allocation pie and thus less incentive to expand output because adding to overall production reduces the value of each of its existing slices. As a result, the output share effect can eliminate a key advantage of market-based policies: marginal cost equalization. Furthermore, in the case of imperfect competition, preliminary analytical results show that a byproduct of this result can be inefficient production shifting. To the extent that firms are large because they have lower costs, production will be shifted to smaller, higher cost firms, which face larger implicit subsidies.
To analyze the role of imperfect competition and its effect on the performance of instruments, Fischer (2003a) used an analytical model of Cournot imperfect competition. By comparing output-based allocations to fixed output subsidies (rather than no subsidies), the output share effects were isolated from the general efficiency issues discussed previously. Two arguments for including an output subsidy are imperfect competition, in which an environmental regulation alone could exacerbate underprovision of goods, and imperfect participation, in which imposing a regulation on a subset of polluters could cause output to shift to exempt firms. Both these scenarios, however, imply that output shares among program participants likely are to be significant. In this situation, output-based allocation of permits offers less of a subsidy than a fixed rebate, and it can lead to inefficient shifting of production among participants. Rebating an emission tax reduces the incentive to abate, and it will not equalize marginal abatement costs if costs differ among emitters. These results hold in a Cournot duopoly model whether emission rates are determined simultaneously or strategically in a two-stage model.
The strongest results lie with the refunded emissions tax, so Fischer (2003c) explored this further. Output-based refunding of environmental policy revenues combines a tax on emissions with a subsidy to output. Subsidies can alleviate the underprovision of output, which is the classic characteristic of imperfect competition. When market shares are significant, however, endogenous refunding reduces the incentive to abatement and the marginal net tax or subsidy. If market shares differ, marginal abatement costs will not be equalized, and production is shifted among participants. In an asymmetric Cournot duopoly, endogenous refunding of revenues leads to higher output, emissions, and overall costs compared with a fixed rebate program targeting the same emissions intensity. These results hold whether emissions rates are determined simultaneously or strategically in a two-stage model.
Detailed Modeling in the Electricity Sector Shows That Allocation Can Have a Large Effect on the Cost of Market-Based Emission Policies
At this juncture, roughly one-half of the U.S. population lives in States that have committed to opening up electricity markets to competition originally unleashed by the 1992 Energy Policy Act. The other half, however, resides in regions practicing cost of service regulation, and the trend toward competition appears to have stalled. The cost recovery rules under traditional regulation vary greatly from those under competition, significantly altering the incentives for compliance. Under traditional average cost pricing, permits usually are valued at "original cost" for the purpose of cost recovery. The market price of permits allocated for free would not be reflected in electricity prices (except when the permits are purchased). With competition and marginal cost pricing, however, the market price of permits would be reflected in electricity generation without regard to their original costs, subject to the caveat that the emitting technology is the marginal production technology and determines marginal cost. One of the justifications for allocating emission permits without charge is that they serve as a form of compensation to the affected industry. Clearly, the nature of electricity regulation has a direct bearing on the magnitude of compensation. It is particularly relevant to note that different regulatory practice is in place in different states, especially throughout the 23-jurisdiction NOx SIP Call region. This mix of cases requires detailed modeling capability to assess the efficiency costs of alternative allocations.
Burtraw, et al. (2001) examined the cost effectiveness and distributional effects of alternative approaches to distributing carbon emission allowances. They found that the cost depends on the deviation of price from marginal cost. The auction is roughly one-half of the cost of the other approaches, as measured by changes in consumer and producer surplus. This robust finding is supplementary to earlier findings (Fischer) regarding the performance of various instruments in a general equilibrium setting, and it offers strong policy implications for the design of policy to regulate carbon emissions in the electricity sector. It also offers a hypothesis regarding the regulation of other pollutants that we subsequently examined.
The Efficiency Consequences of the Approach to Distributing Allowances Differs Among Pollutants
Drs. Burtraw and Palmer, with significant help from David Evans (who joined the project), have extended Resources for the Future's electricity sector model to address the question of whether the dramatic findings regarding the cost effectiveness of an auction hold up when different pollutants are regulated. This analytical work fed into stand-alone publications and also into work funded by a program office at the U.S. EPA. We find that in every case, the auction is the most efficient. In summarizing a broad effort, however, Palmer and Burtraw (2004) find that in the case of sulfur dioxide, nitrogen oxides, and mercury emission allowances, allocation has a relatively small effect on the social costs of imposing the emissions caps and on the introduction of clean technologies. This finding comes in stark contrast to the dramatic effect that allocation has in the case of carbon dioxide.
Allowance Distribution is a Mechanism for Compensation
The initial distribution of emission allowances has important consequences on who bears the incidence of cost of environmental regulation. Distribution of allowances based on electricity output yields the smallest increase in electricity price, but greatest increase in natural gas price. This is especially dramatic in the case of carbon policy (Burtraw, et al., 2001-). In the case of carbon policy, the auction does better than the output-based allocation at protecting households and at preserving asset values for incumbent producers.
Grandfathering yields the greatest rewards for producers, but it imposes a substantial cost on consumers. Producer preferences for the policy vary across fuel types as measured by the different effects on the asset values of existing generating capacity. Owners of coal and gas-fired units prefer the historic approach, and nonemitting units (including nuclear and hydro) prefer an allowance auction. Taken as a group, all existing generators nationwide prefer historic to the other two approaches, but are nearly indifferent between updating and the auction in our central case when considering regulation of conventional pollutants other than carbon (Palmer and Burtraw, 2004).
There Are Differences in Environmental Performance of the Policies
Emission rate intensity of economic activity is one measure that has been suggested to promote environmental improvement while promoting environmental growth. Rate-based emissions policies (such as tradable performance standards) fix average emissions intensity, whereas cap-and-trade policies fix total emissions. Fischer (2003b) shows that unfettered trade between rate-based and cap-and-trade programs always raises combined emissions, except when product markets are related in particular ways. Gains from trade are passed on fully to consumers in the rate-based sector, resulting in more output and greater emissions allocations. The paper considers several policy options to offset the expansion, including a tax, an "exchange rate" to adjust for relative permit values, output-based allocation for the rate-based sector, and tightening the cap. A range of combinations of tighter allocations could improve situations in both sectors with trade while holding emissions constant.
Under a cap-and-trade program, one should expect the aggregate amount of emissions to be invariant with respect to how emission allowances initially are distributed. We examined three ways, however, that environmental performance could vary: leakage out of the sector, increases in other pollutants, and leakage out of the region that is regulated.
Burtraw, et al. (2001) examined leakage of emissions out of the sector and other side effects that may occur because of fuel switching away from electricity to direct combustion of fuels. In the context of a carbon policy applied just to the electricity sector, the leakage occurs because of increased use of natural gas in the electricity sector, which drives up the gas price. As a consequence, and also because of lower prices for coal resulting from decreased consumption in the electricity sector, there is greater use of coal by industry. The most important finding is that the amount of leakage of this sort varies greatly with the policy. Output-based allocation leads to greater electricity consumption in general and greater levels of pollution of emissions such as mercury or NOx that are not capped.
On another track, the research takes note that political pressure often exists for rebating environmental levies, particularly when incomplete environmental regulation allegedly creates an "unlevel playing field" with other firms or industries that are not regulated. Fischer, together with colleagues in France (Marc Vielle, CEA/IDEI, University of Toulouse; Alain Bernard, Ministry of Equipment, Transports and Housing), used computable general equilibrium modeling to find that output-based allocation could be used as a response to leakage across jurisdictional borders. In the general equilibrium setting, Bernard, et al. (2001) used a computable general equilibrium model incorporating an international trade model to address issues of distortions in factor markets and distortions introduced by the output subsidy. The project used the GEMINI-E3/GemWTraP model with seven individual countries/regions (including the United States) to simulate detailed, sector-specific evaluations of marginal abatement costs for carbon dioxide emissions. The model produces estimates of the relative cost effectiveness of instruments, including tax interactions and trade leakages, as well as the differences in the distribution of abatement efforts across sectors. It also compares different potential rules for determining industry-level allotments (which, if subsequently allocated by output shares, imply different output subsidies) such as historic emission shares, value-added shares, or other compensation standards. The authors assess the conditions under which rebating environmental levies is justified for the sector subject to environmental regulation. It combines a theoretical approach based on second-best modeling with numerical simulations aimed at determining the most sensitive parameters. The authors find that if an adequate tax on production can be levied in the sector that is not subject to environmental regulation, no rebate is justified for the regulated sector. Moreover, even in the case of constrained taxation in the sector not subject to the environmental regulation, a tax rebate or a subsidy in the regulated sector is not necessarily a welfare-increasing policy. The exception occurs when the goods of the competing sectors are close substitutes. We find that these kinds of policy constraints can be quite costly in terms of overall economic efficiency.
Strategic Incentives Stemming From Interjurisdictional Relations Affect the Choice of Allocation Mechanisms
Another thematic area has to do with interests of local and regional government. An important and basic result of the literature on fiscal federalism is that the optimal level of government to provide a given public service is the smallest level of government that encompasses all of the appreciable benefits and costs. Although NOx is a transboundary air pollutant, the implementation of a NOx trading program is left basically to the States. Currently, unfolding policies and procedures governing regional NOx trading in the United States place important authority at the state level, including the authority to determine the method of allocating emission allowances to individual sources. States also have authority over the extent and form of electricity regulation relevant to NOx and SO2 trading. There are also important differences among the states in the character of generation technology and in the tax treatment of electricity. All of these differences suggest the possibility that states will adopt a patchwork of different policies for implementing NOx trading.
This research project examines the possibility for strategic behavior among jurisdictions in permit allocations and the efficiency and distributional effects of decisions that are made in this multijurisdictional context. Evans (in preparation, 2004) uses the enhancements to the electricity model to explore the role of interjurisdictional strategic incentives to influence mobility across state borders of goods and services (including electricity), factor inputs such as labor and capital (including the location of new generation capacity), and emission permits.
How Do Dynamic Issues in the Economy or Regulation Affect the Performance of Instruments?
In a short-run static model, many variables such as abatement cost structures, technologies, and numbers of participants are assumed to be exogenous. In the long run, however, investment planning, innovation, capacity expansion, and entry all can be affected by output allocation schemes. Thus, with these decisions endogenous to the choice of allocation method, the relative magnitude of the efficiency losses compared to other distribution mechanisms could be quite different in a dynamic model. Furthermore, the directions of the effects are not always transparent.
Burtraw, et al. (2001) find that output-based allocation of carbon emission allowances leads to a dramatic increase in natural gas use in the electricity sector. Also, output-based allocation leads to greater electricity consumption in general and greater levels of pollution of emissions that are not capped.
The question of technological change, however, is different when considering regulation of conventional pollutants such as SO2, NOx, and mercury. In this case, Palmer and Burtraw (2004) find that the influence on the pace of introduction of new technology for electricity generation, and especially the expanded use of natural gas, does not hinge strongly on the method of distributing allowances.
Selected Media Contacts
Burtraw and/or Palmer were interviewed by a number of news organizations about this research project and its extensions. These interviews included: The Economist, Natural Gas Week, ABC.com, KCPW, National Journal, Inside EPA, Bloomberg News Service, Air Daily, Inside Washington, Chemical and Engineering News, UPI Newswire, Air Daily, Greenwire News Service, New York Times, Dow Jones News Wire, Resource News Service, and IEEE Spectrum.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 75 publications | 2 publications in selected types | All 2 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Burtraw D. Carbon emission trading costs and allowance allocations:evaluating the options. Resources, Fall, No. 145, 2001, pp. 13-16. |
R828628 (2001) R828628 (Final) |
Exit |
|
Palmer K, Paul A, Bharvirkar R. The effect on asset values of the allocation of carbon dioxide emission allowances. The Electricity Journal 2002;15(5):51-62. |
R828628 (2001) R828628 (Final) |
Exit Exit Exit |
Supplemental Keywords:
air, global climate, nitrogen oxides, sulfates, acid rain, public policy, decisionmaking, cost benefit, public good, compensation, conservation, modeling, analytical, northeast, electricity sector., RFA, Economic, Social, & Behavioral Science Research Program, Scientific Discipline, Chemistry, Economics, Market mechanisms, Social Science, electricity generation plants, equilibrium analysis, auctioning permits, decision making, air pollution, cost benefit, socioeconomics, cap and trade systems, environmental Compliance, emissions trading, pollution fees, tradable pollution permits, public policy, allowance allocation, cost effective, pollution allowance trading, econometricsProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.