Ask the Scholar

Document scope · 1 page
doc
Scholar
Ask about this object, its catalog metadata, its source description, or the page inventory. For page-specific OCR and visual context, open one of the page chats.

Scholar Source Context

Document identity
localId
225358686
label
Trading Papers Sent to Mexico Energy Ministry [Global Climate Change] [binder]
core
doc
dtoType
document
pageCount
1
Source metadata
id
225358686
contentType
document
title
Trading Papers Sent to Mexico Energy Ministry [Global Climate Change] [binder]
collections
Records of the Council of Economic Advisers (Clinton Administration)
Vivian Wu's Files
imageCount
1
hasImages
yes
source
import
hasTranscription
no
Source extras
naId
225358686
levelOfDescription
fileUnit
otherTitles
42-t-7422601-20171095F-045-001-2019
recordType
description
ocrSource
nara-archive
Single page context
seq
1
pageIndex
0
type
document
mediaId
430f6e1d3c6b6006
ocrText
FOIA Number: 2017-1095-F FOIA MARKER This is not a textual record. This is used as an administrative marker by the William J. Clinton Presidential Library Staff. Collection/Record Group: Clinton Presidential Records Subgroup/Office of Origin: Council of Economic Advisers Series/Staff Member: Subject Files Subseries: OA/ID Number: 21609 FolderID: Folder Title: Trading Papers Sent to Mexico Energy Ministry [Global Climate Change] [binder] Stack: Row: Section: Shelf: Position: S 21 4 11 1 EXECUTIVE OFFICE OF THE PRESIDENT COUNCIL OF ECONOMIC ADVISERS WASHINGTON, D.C. 20502 May 10, 1999 Manuel Betancourt Secretaria de Energia Subsecretaria de Politica y Desarrollo de Energeticos Direccion General de Politica y Desarrollo Energeticos Insurgentes Sur No. 890, Piso 14 Col. Del Valle C.P. 03100 Mexico, D.F. Dear Lic. Betancourt: Please find enclosed a set of papers describing the U.S. sulfur dioxide emissions trading program. These papers are from the U.S. Environmental Protection Agency, Resources for the Future (a non-partisan think-tank focused on environmental and natural resource issues), and academic journals. I hope that you will find these papers useful in understanding the design and performance of our emissions trading system. If you would like to follow-up with more discussion of this program, please don't hesitate to contact me. I would be more than happy to speak with you and your colleagues about the program. In addition, please feel free to contact Stephanie Benkovic of the U.S. Environmental Protection Agency at 202-564-9142 and [email protected]. I will also compile some papers that may provide insights on the potential design of a greenhouse gas emissions trading system and send them to you in the near future. We look forward to continued work with you and your secretariat on sustainable energy, climate change, and other related issues. Sincerely, 7.4l Joe Aldy Senior Adviser enclosures cc: Stephanie Benkovic, U.S. Environmental Protection Agency Mark Mazur, U.S. Department of Energy United States Office of Air and Radiation EPA Revised August 1998 Environmental Protection Agency Acid Rain Division (6204]) Emissions Trading of Sulfur Dioxide: ACID RAIN PROGRAM The U.S. Experience Introduction mandatory ceiling, on total SO₂ emissions from utilities. After full implementation of The overall goal of the United States' Acid the program, annual emissions will be Rain Program is to achieve significant capped at 8.95 million tons, which environmental and public health benefits represents a 50 percent decrease in SO₂ through reductions in emissions of sulfur emissions from electric utilities from 1980 dioxide (SO₂) and nitrogen oxides (NOx), levels. Enforcing this cap is a rigorous the primary causes of acid rain. To achieve monitoring system that allows EPA to this goal at the lowest cost to society, the ensure accuracy and verifiable emissions program employs both traditional and reductions. innovative, market-based approaches for controlling air pollution. In a dramatic departure from traditional command and control regulatory At the core of the SO₂ program is a cap, or structures, this program utilizes a market- based system of tradeable emissions 8.95 Million Tons; 49% below allowances. The system allows regulated utilities 1980 level flexibility in how they 30 attain compliance, so that they can select the 25 most cost effective Without Acid Rain Program strategy for their 20 particular facilities. Million tons 15 Allowable Emissions The program was 10 established in the 1990 amendments to the With Acid Rain Program 5 Clean Air Act, the nation's air pollution 0 control law. As stated 1980 1985 1990 1995 2000 2005 2010 in Title IV of the Clean Air Act, the primary goal of the Acid Rain Figure 3 The Clean Air Act will result in a 10 million ton reduction Program is to reduce in SO₂ emissions from 1980 levels by the year 2010. annual SO₂ emissions by 10 million tons below 1980 levels. To year. achieve these reductions, the law requires a two-phase tightening of restrictions placed The allowance trading system gives utilities on fossil fuel-fired power plants, which flexibility in devising compliance strategies. accounted for two thirds of national SO₂ This flexibility is a result of the fact that a emissions in 1980. regulated unit need only to hold enough allowances at year end to account for its Phase I began in 1995. It affects 263 emissions. Unlike traditional regulatory boilers, or "units," at 110 mostly coal- programs, EPA does not dictate the burning electric utility plants that were compliance options. Options might identified in the law as the largest SO₂ include employing energy conservation emitters. Additional units may voluntarily measures, increasing reliance on renewable join Phase I of the program early as energy, reducing usage, employing substitution or compensating units. In 1997 pollution control technologies, switching there were a total of 423 affected units. to lower sulfur fuel, or developing other alternate strategies to reduce emissions. Phase II, which begins in the year 2000, tightens the annual emissions limits Because allowances have a value and can be imposed on these large, higher emitting purchased and owned, they are sometimes plants. In addition, restrictions come into misconstrued as "rights" to a certain effect on smaller, cleaner plants fired by amount of pollution. However, the law coal, oil, and gas. This phase affects makes clear that allowances are merely existing utility units serving generators authorizations to emit, since at all times with an output capacity of greater than 25 the government retains the authority to megawatts and all new utility units. require affected sources to comply with the Clean Air Act limits. The concept of The Clean Air Act also calls for a 2 million authorizing sources to emit is common to ton reduction in NOx emissions by the all air pollution regulations; no regulations year 2000 through a more conventional dictate zero air emissions. With more regulatory program. A significant portion traditional forms of regulations, sources of this reduction will be achieved by coal- are authorized to emit through emissions fired utility boilers that will be required to limits or through requirements to install install low NOx burner technologies and specific technologies. to meet new emissions standards. The financial value and enduring validity The Allowance System of excess allowances encourage SO₂ reductions beyond what is required by the Allowance trading is the innovative tool present year's allotment. Allowances are for achieving SO₂ emissions reductions fully marketable commodities: units that required by the cap. Allowances are the reduce their emissions below the number currency with which compliance with the of allowances they hold may trade their SO₂ emissions requirements is achieved; excess allowances with other units in their one allowance authorizes a regulated unit system, sell them on the open market to to emit one ton of SO₂ during a specific other utilities or through EPA auctions, or year or any year thereafter. At the end of bank them to cover emissions in future each year, the unit must hold an equal or years. greater number of allowances as the tons of SO₂ emitted that year, i.e., a unit that emits Banking 5,000 tons of SO₂ must hold at least 5,000 Banking encourages early emissions allowances that are eligible to be used that reductions by allowing participants to store their excess allowances and by assuring that existing rates greater than 2.5 those allowances will retain their pounds/mmBtu and was multiplied by the compliance value in future years. Utilities unit's average fuel use from 1985 through may use banked allowances to help comply 1987. These allocations are listed in the in later years, or they may sell them. In Clean Air Act and are codified in the the first year of the Program's Allowance System Regulations. implementation, Phase I affected units collectively achieved 40 percent greater In Phase II, the limits imposed on Phase I emissions reductions than was required by plants will be tightened, and emissions law. This dramatic result demonstrates limits will be imposed on smaller, cleaner that banking leads to SO₂ reductions earlier units that were not included in Phase I. than if banking were not permitted. These EPA allocated allowances to each unit at extra reductions result in environmental the lesser of its existing emission rate or 1.2 and health benefits that begin sooner. pounds of SO₂/mmBtu, multiplied by the unit's baseline fuel consumption. Banking also reduces compliance costs by allowing utilities more flexibility in the Equity issues in allowance allocations timing of their pollution control In addition to the standard formulas for investments. allocating emission allowances, the Clean Air Act Amendments set forth special Phase I affected sources are likely to be provisions to address equity concerns banking allowances in anticipation of the raised by some states. For example, in more stringent emission limits and higher some cases, states that had already reduced costs in Phase II. It is expected that rather the emissions of their electric utilities to than a sharp drop in emissions at the start levels well below the national average were of Phase II, reductions will move toward given extra allowances. Similarly, a state the ultimate cap more gradually as banked with high population growth in the 1980s allowances from previous years are used for was given bonus allowances for its electric compliance. Although the reduction rate utilities to compensate for this growth. In may slow down after Phase II, the all cases, these redistributions of allowances environmental and health benefits sparked were done without increasing the overall by the initial overcompliance will have level of national emissions. In other been accruing for several years. words, increases in allowance allocations in certain states or at certain electric utilities Regardless of how many allowances a unit were offset by decreases in the allowance holds, it is never entitled to exceed the allocations to other states or emissions source-specific limits set under Title I of sources. the Clean Air Act to protect public health. Permits Allocations Permitting under the Acid Rain Program is EPA allots allowances to each affected unit simple and flexible. For example, the Phase for each year, beginning with Phase I units II permitting form consists primarily of a in 1995. These allocations were statement of standard legal requirements in determined prior to the start of the trading the Clean Air Act that companies certify program, based on selected emission rates they will meet. Utilities may select from and each unit's representative fuel several compliance options, devise the most utilization level. In Phase I, the specified cost-effective compliance plan, and revise emission rate was 2.5 pounds of plans easily without needing government SO₂/mmBtu (million British thermal review or approval. The standard permit units). This rate was applied to units with forms ensure consistency on a nationwide basis. Source makes Then mail trade Allowance allowance trade information to EPA with brokers, Transfers environmental EPA is responsible groups,or other for recording the sources transfer of allowances that are used for compliance and for confirming that utilities hold at ATS least as many allowances as tons of SO₂ emitted by EPA sends a the end of the year. transfer EPA enters trade confirmation information into the To fulfil this role, report to each of Allowance Tracking the traders System EPA designed and maintains the Figure 4 The allowance transfer process Allowance Tracking System (ATS), a computer program that is the official record of ATS also makes the movement of allowances allowance holdings and transfers. Every easy to track. ATS records the issuance of all utility unit, corporation, group, or allowances, the holdings of allowances in individual holding allowances has an account accounts, the deduction of allowances for with ATS, held in the name of a designated compliance purposes, the transfer of individual. EPA established accounts for allowances between accounts, and the utility units affected by both Phase I and number of allowances held in EPA reserves. Phase II. In addition, any person or group Information on ATS accounts and wishing to purchase allowances may open an transactions are available to the public via ATS account. the Internet. The structure of the Acid Rain Program Determining Compliance simplifies the transfer of allowances, thereby At the end of the year, units must hold in lowering the cost of transactions. The that year's compliance subaccount an transfer process is comprised of four steps for amount of allowances equal to or greater utilities. A utility makes an allowance trade than the amount of SO₂ emitted during that with another utility, a broker, or year. By January 30 following the environmental group. They enter compliance year, units must finalize and information about the trade on a one-page report to EPA allowance transactions that form and mail it to EPA. Next, EPA enters were used in attaining quantities of this information into ATS and then sends a allowances necessary for compliance. The transfer confirmation report to each of the amount of emissions is contained in the traders. Because there is no need for case-by- Emission Tracking System (ETS), which is case approval of trades, allowance transfers operated by the Acid Rain Program and are straightforward and fast. Only transfers records each hour of each unit's hourly of allowances to be used for compliance emissions throughout the year. require EPA notification; notification of other transfers is voluntary. After the January 30 deadline, EPA deducts Rain Program has witnessed large reductions effects of the program. Industry's role is to in SO₂ emissions. Since the start of the reduce emissions in the most cost effective program in 1995, all Phase I affected units manner. Brokers provide information on have been in compliance, and utilities the market and facilitate allowance transfers reduced emissions significantly more than for utilities and other parties, thus reducing required by the cap. Annual emissions from transaction costs. Finally, environmental these Phase I units dropped by more than organizations monitor the overall half between 1980 and 1995, from 10.9 performance through access to emissions and million tons to 5.3 million tons. allowance data. In addition, these groups may purchase and retire allowances, a small Environmental effects corresponding to but direct action individuals can take to these early reductions are starting to emerge. reduce pollution. A United States Geological Survey study determined that 1995 wet sulfate deposition Just as each participant has a role in ensuring declined in the eastern United States by 10 to the integrity of the Program, each element of 25 percent. Other anticipated benefits the Program works in concert to bring about include avoided health costs of $12 to $40 efficient emissions reductions. The allowance billion per year by 2010; improvements in trading system capitalizes on the power of visibility amounting to $3.5 billion by 2010; the marketplace to reduce SO₂ emissions in fewer acidic lakes and streams; and reduced the most cost effective manner possible. The damage to buildings and outdoor cultural permitting program allows sources the artifacts. flexibility to tailor and update their compliance strategy based on their individual At the same time, the Acid Rain program circumstances. The continuous emissions has cost significantly less than the benefits, monitoring and reporting systems provide and estimates for the program's long term the accurate and standardized accounting of costs continue to drop. As of 1994, the emissions necessary to make the program estimated cost of the program was $2 to $2.5 work, and the excess emissions penalties billion per year by 2010. This cost is half of provide strong incentives for compliance. what a command and control regulatory Each of these separate components program would cost. contributes to the effective working of an integrated program that harnesses market The cost of reducing a ton of SO₂ from the incentives for the benefit of the utility sector has been much lower than environment. expected: scrubber costs have dropped, removal efficiencies have improved, and For More Information expected increases in costs associated with the increased use of low sulfur coal have not For more information on the Acid Rain materialized. These reductions in cost have Program, please visit the Acid Rain been reflected in allowance prices. In just Program's web site at two years, allowance prices have dropped www.epa.gov/acidrain or call the Acid from $150/ton to less than $100/ton. Rain Hotline at 202-564-9620. Important Roles Played by Participants Successful implementation of the Acid Rain Program hinges on straightforward and well- defined roles for each participant. EPA monitors utilities' emissions to ensure compliance, and measures the environmental allowances from each unit's compliance Strong enforcement measures deter subaccount in an amount equal to its SO₂ noncompliance. If annual emissions exceed emissions for that year. If the unit's the number of allowances held, the owners emissions do not exceed its allowances, the or operators of delinquent units must pay a remaining allowances are carried forward, or penalty of $2,000 (adjusted for inflation) per banked, into the next year's subaccount. excess ton of SO₂ or NOx emissions. This fee is substantially higher than the cost of Ensuring Environmental and Health compliance, e.g., an allowance. In addition, Benefits through Stringent Monitoring the number of exceeded allowances is deducted from the violating utility's account Tradable allowances are intended to achieve for the next year to fully offset the environmental goals at the lowest cost to environmental impact. society. However, even if no trades were to occur, these environmental goals would still The emissions monitoring and reporting be met by the Acid Rain Program. The systems are critical to the program. Program uses several mechanisms to ensure Monitoring ensures, through stringent that its environmental and health goals are accounting, that the SO₂ and NOx emissions met. During Phase II, the Clean Air Act reduction goals are met. They also instill places a cap of 8.95 million on the total confidence in allowance transactions by number of allowances issued to units each certifying the existence and quantity of the year. This effectively caps emissions at 8.95 commodity being traded. million tons annually and ensures that the mandated emissions reductions are Emissions data is available to the public in maintained over time. This reduction several formats, including over the Internet. represents a 50 percent decrease in SO₂ Such public access provides the program emissions from 1980 levels. transparency that assures integrity and public trust in the system. To enforce the integrity of the emissions limit, EPA tracks on a continuous basis each Results unit's emissions of SO2, NOx and CO2, as well as volumetric flow and opacity. In most The Acid Rain Program's environmental cases, a continuous emission monitoring goals are already starting to be met. After (CEM) system must be installed. Through the first year of implementation, the Acid the CEM system, units report their hourly SO2 Emissions from emissions data to EPA 263 Phase I Units on a quarterly basis. The reporting process is 12 evolving from the 10 9.4 9.3 submittal of diskettes to 8.7 Allowable Emissions the use of modems, 8 7.4 7.1 which has greatly 6.0 increased the efficiency 6 of reporting. The data SO2 (million tons) 4.5 4.8 4.8 is then recorded in the 4 Emissions Tracking System, which serves as 2 a repository of 0 emissions data for the utility industry. 1980 1985 1990 1995 1996 1997 United States Office of \ir and Radiation Revised February 1999 EPA Environmental Protection Agency Acid Rain Division 62041) EPA 430-F-92-015 Environmental Benefits ACID RAIN of the Acid Rain Program PROGRAM Introduction enjoyment of scenic vistas across our country, particularly in National Parks. Stress to our Acidic deposition, or acid rain as it IS forests that populate the ridges of mountains commonly known, occurs when emissions of from Maine to Georgia will be reduced. sulfur dioxide (SO₂) and oxides of nitrogen Deterioration of our historic buildings and (NOx) react in the atmosphere with water, monuments will be slowed. Finally, reductions in oxygen, and oxidants to form various acidic SO₂ and NOx will reduce sulfates, nitrates, and compounds. These compounds then fall to ground level ozone (smog), leading to the earth in either drv form (such as gas and improvements in public health. particles) or wet form (such as rain, snow, and fog). Prevailing winds transport the compounds, sometimes hundreds of miles, Surface Waters across state and national borders. Acid rain primarily affects sensitive bodies of Electric unlity plants account for about 70 water, that is, those that rest atop soil with a percent of annual SO₂ emissions and 30 limited ability to neutralize acidic compounds percent of NOx emissions in the United (called "buffering capacity"). Many lakes and States. Mobile sources (transportation) also streams examined in a National Surface Water contribute significantly to NOx emissions. Survey (NSWS) suffer from chronic acidity, a Overall. over 20 million tons of SO₂ and condition in which water has a constant low pH NOx are emitted into the atmosphere each level. The survey investigated the effects of acidic year. deposition in over 1,000 lakes larger than 10 acres and in thousands of miles of streams believed to Acid rain causes acidification of lakes and be sensitive to acidification. Of the lakes and streams and contributes to damage of trees at streams surveved in the NSWS, acid rain has high elevations (for example, red spruce trees been determined to cause acidity in 75 percent above 2,000 feet in elevation). In addition, of the acidic lakes and about 50 percent of the acid rain accelerates the decav of building acidic streams. Several regions in the U.S. were materials and paints, including irreplaceable identified as containing many of the surface buildings, statues, and sculptures that are part waters sensitive to acidification. They include, of our nation's cultural heritage. Prior to but are not limited to, the Adirondacks, the falling to the earth, SO₂ and NOx gases and mid-Appalachian highlands, the upper Midwest their particulate matter derivatives, sulfates and the high elevation West. and nitrates, contribute to visibility degradation and impact public health. In some sensitive lakes and streams, acidification has completely eradicated fish species, such as the Implementation of the Acid Rain Program brook trout, leaving these bodies of water barren. under the 1990 Clean Air Act Amendments In fact, hundreds of the lakes in the Adirondacks will confer significant benefits on the nation. surveved in the NSWS have acidity levels By reducing SO₂ and NOx, many acidified indicative of chemical conditions unsuitable for lakes and streams will improve substantially the survival of sensitive fish species. so that they can once again support fish life. lisibility will improve, allowing for increased Critical pH Levels for Selected Aquatic Organisms 65 6.0 55 5.0 45 40 Yeilcw Perch Brook Trout Lake Trout Smailmouth Bass Rainbow Trout Common Shiner American Toad* Wood Frog* Leopard Frog* Spotted Salamander* Crayfish** Mayfly** Clam** Snall** "embryonic life stages "selected species Note: Solid symbols for each type of organism are placed in favorable pH ranges. Shaded symbols are placed in less favorable ranges. No symbol is placed in pH ranges that generally do not support populations of a particular type of organism. Source: National Acid Precipitation Assessment Program 1990 Integrated Assessment Report, November 1991 Emissions from U.S. sources also contribute Acidification is also a problem in surface water to acidic deposition in eastern Canada, where populations that were not surveyed in federal the soil is very similar to the soil of the research projects. For example, although lakes Adirondack Mountains, and the lakes are smaller than 10 acres were not included in the consequently extremely vulnerable to chronic NSWS, there are from one to four times as many acidification problems. The Canadian of these small lakes as there are larger lakes. In government has estimated that 14,000 lakes in the Adirondacks, the percentage of acidic lakes is eastern Canada are acidic. significantly higher when it includes smaller lakes (26 percent) than when it includes only the target Streams flowing over soil with low buffering size lakes (14 percent). capacity are equally as susceptible to damage from acid rain as lakes are. Approximately The acidification problem in both the United 580 of the streams in the Mid-Atlantic Coastal States and Canada grows in magnitude if Plain are acidic primarily due to acidic "episodic acidification" (brief periods of low pH deposition. The New Jersey Pine Barrens levels from snowmelt or heavy downpours) is area endures the highest rate of acidic streams taken into account. Lakes and streams in the nation with over 90 percent of the throughout the United States, including high streams acidic. Over 1,350 of the streams in elevation western lakes, are sensitive to episodic the Mid-Atlantic Highlands (mid-Appalachia) acidification. In the Mid-Appalachians, the are acidic, primarily due to acidic deposition. Mid-Atlantic Coastal Plain, and the Adirondack Many streams. in that area have already Mountains, many additional lakes and streams experienced trout losses due to the rising become temporarily acidic during storms and acidity. snowmelt. Episodic acidification can cause large scale "fish kills." 2 Approximately 70 percent of sensitive lakes in in several ways; for example, acidic cloud water at the Adirondacks are at risk of episodic high elevations may increase the susceptibility of acidification. This amount is over three times the red spruce to winter injury. the amount of chronically acidic lakes. In the mid- Appalachians, approximately 30 percent There also is a concern about the impact of acid of sensitive streams are likely to become rain on forest soils. There is good reason to acidic during an episode. This level is seven believe that long-term changes in the chemistry of times the number of chronically acidic some sensitive soils may have alreadv occurred as streams in that area. a result of acid rain. As acid rain moves through the soils, it can strip away vital plant nutrients Acid rain control will produce significant through chemical reactions, thus posing a benefits in terms of lowered surface water potential threat to future forest productivity. acidity. If acidic deposition levels were to remain constant over the next 50 years (the time frame used for projection models), the Visibility acidification rate of lakes in the Adirondacks that are larger than 10 acres would rise by 50 Sulfur dioxide emissions lead to the formation of percent or more. Scientists predict, however, sulfate particles in the atmosphere. Sulfate that the decrease in SO₂ emissions required particles account for more than 50 percent of the by the Acid Rain Program will significantly visibility reduction in the eastern part of the reduce acidification due to atmospheric United States, affecting our enjoyment of national sulfur. Without the reductions in SO₂ parks, such as the Shenandoah and the Great emissions, the proportions of acidic aquatic Smokv Mountains. The Acid Rain Program is systems in sensitive ecosystems would remain expected to improve the visual range in the high or dramatically worsen. eastern U.S. by 30 percent. Based on a study of the value national park visitors place on visibility, The impact of nitrogen on surface waters is the visual range improvements expected at also critical. Nitrogen plays a significant role national parks of the eastern United States due to in episodic acidification. and new research the Acid Rain Program's SO₂ reductions will be recognizes the importance of nitrogen in worth a billion dollars by the year 2010. In the long-term chronic acidification as well. western part of the United States, nitrogen and Furthermore, the adverse impact of carbon also play roles, but sulfur has been atmospheric nitrogen deposition on estuaries implicated as an important source of visibility and other large water bodies may be impairment in many of the Colorado River significant. For example, 30 to 40 percent of Plateau national parks, including the Grand the nitrogen in the Chesapeake Bav comes Canvon, Canvonlands, and Bryce Canyon. from atmospheric deposition. Nitrogen is an important factor in causing eutrophication (oxygen depletion) of water bodies. Materials Acid rain and the dry deposition of acidic Forests particles are known to contribute to the corrosion of metals and deterioration of stone Acid rain has been implicated in contributing and paint on buildings, cultural objects, and cars. to forest degradation, especially in The corrosion seriously depreciates the objects' high-elevation spruce trees that populate the value to society. Dry deposition of acidic ridges of the Appalachian Mountains from compounds can also dirty buildings and other Maine to Georgia, including national park structures, leading to increased maintenance costs. areas such as the Shenandoah and Great To reduce damage to automotive paint caused by Smokev Mountain national parks. Acidic acid rain and acidic dry deposition, some deposition seems to impair the trees' growth manufacturers use acid-resistant paints, at an 3 average cost of $5 for each new vehicle (or a Clean Air for Better Life total of $61 million per year for all new cars and trucks sold in the U.S.) The Acid Rain Bv significantly reducing SO₂ emissions, the Program will reduce damage to materials by Clean Air Act will confer numerous benefits on limiting SO, emissions. The benefits of the the nation. Scientists project that the 10 Acid Rain Program are measured, in part, by million-ton reduction in SO₂ emissions required the costs now paid to repair or prevent by the Acid Rain Program should significantly damage--the costs of repairing buildings. decrease or slow down the acidification of water using acid-resistant paints on new vehicles, bodies and will reduce stress to forests. In fact, a plus the value that society places on the United States Geological Survey study determined details of a statue lost forever to acid rain. that after the first year of the Acid Rain Program's implementation, 1995 wet sulfate deposition in the eastern United States declined bv 10 to 25 Health percent. Anticipated benefits also include significantly improved visibility and longer life Based on health concerns, SO, has spans of building materials and structures of historically been regulated under the Clean cultural importance. Finally, the reductions in Air Act. Sulfur dioxide interacts in the emissions will help to protect public health. atmosphere to form sulfate aerosols, which may be transported long distances through the air. Most sulfate aerosols are particles that For More Information can be inhaled. In the eastern United States, sulfate aerosols make up about 25 percent of For more information, please visit the Acid Rain the inhalable particles. According to recent Program's World Wide Web site at studies at Harvard and New York www.epa.gov/acidrain, or contact the Acid Rain Universities, higher levels of sulfate aerosols Hotline at 202-564-9620. are associated with increased morbidity (sickness) and mortality from lung disorders such as asthma and bronchitis. By lowering sulfate aerosol levels, the Acid Rain Program will reduce the incidence and the severity of asthma and bronchitis. When fully implemented by the year 2010 the public health benefits of the Acid Rain Program will be significant due to decreased mortality, hospital admissions, and emergency room visits. Decreases in nitrogen oxide emissions are also expected to have a beneficial impact on health effects by reducing the nitrate component of inhalable particulates and reducing the nitrogen oxides available to react with volatile organic compounds and form ozone. Ozone impacts on human health include a number of morbidity and mortality risks associated with lung disorders. 4 EPA United States Office of Air and Radiation Revised February 1999 Environmental Protection Agency Acid Rain Division (6204J) The Environmental Impacts ACID RAIN PROGRAM of SO₂ Allowance Trading SO2 allowance trading allows utilities flexibility in the timing and location of their emissions reductions. EPA is assessing how trading affects emission patterns around the country. The first three years of the program have produced extra SO2 reductions in almost every affected state. Most allowances were used in the same state as they were allocated, indicating little geographic shifting of emissions due to Introduction Existing utility units were allocated allowances The Environmental Protection Agency's Acid for each future compliance year. At the end of Rain Program, established under Title IV of the each year, each unit must surrender, or "retire," 1990 Clean Air Act Amendments, calls for major as many allowances as tons of SO2 emitted that reductions of sulfur dioxide (SO2) and nitrogen year. Allowances have a specific use year, oxides (NOx), the pollutants that cause acid rain. making them valid for surrender that year or any In an innovative approach to environmental year thereafter. protection, the program uses market incentives to achieve a nationwide limit on SO2 emissions How does EPA verify compliance? more cost effectively than traditional regulatory In the year-end compliance process, EPA methods. The Acid Rain Program requires a two- reconciles its data on allowance holdings and phase tightening of restrictions on fossil fuel- utility emissions. Accurate accounting of utility fired power plants, resulting in a permanent cap emissions is essential to the trading system's on SO2 of 8.95 million tons nationwide, which integrity. EPA tracks on a continuous basis each is half the amount emitted by utilities in 1980. unit's allowance holdings and emissions of SO2, Phase I runs from 1995 through 1999 and affects NOX, and CO2, and other parameters. In most roughly 440 of the larger, higher-emitting utility cases, units install a continuous emissions boilers, or units, in the country. Phase II begins monitoring (CEM) system that reports their in 2000 and extends to over 2,000 units. hourly emissions data to EPA. This market-based program uses a system of Because stringent monitoring and reporting tradable emissions allowances. Each SO2 certifies that environmental requirements are allowance is an authorization to emit one ton of being met, EPA can give utilities flexibility in SO2 in a given year. Allowances may be bought, the methods and timing by which they achieve sold, or banked by utilities, brokers, or anyone compliance. This flexibility allows utilities to else interested in holding them. The creation of pursue their most cost effective strategies, which a market for emissions allowances attaches a could include installing flue gas desulfurization clear monetary value to emissions reductions, ("scrubbers"), switching to lower sulfur fuels, thereby providing utilities an economic incentive incorporating more renewable energy, or to cut their SO2 emissions. Unneeded purchasing allowances. allowances can be sold for profit. Results Utility SO2 Emissions In 1997, Phase I affected units emitted well under 30 their collective emissions limit. These units were allocated 7.1 million allowances for the 1997 25 compliance year and only emitted a total of 5.5 million tons of SO2.¹ 20 Without Acid Rain Program Environmental data reflect improvements begun in 1995. Large areas of the eastern U.S. saw SO2 Emissions (million tons) 15 Allowances Allocated accompanying the downward emissions trend 10 With Acid Rain Program up to a 25 percent decrease in sulfur concentration in rain in 1995 and 1996. Sulfur concentration in 5 dry deposition decreased by 30 percent in the eastern U.S. between 1989 and 1995. Ambient, or 0 airborne, concentrations of sulfur dioxide also 1980 1985 1990 1995 2000 2005 2010 declined by 17 percent nationally between 1994 and 1995 and remained flat from 1995 to 1996. Figure 1: By 2010, utility SO2 emissions will decrease by 8.5 million tons from 1980 levels. Banking encourages early reductions, but gives utilities flexibility in timing these cuts. What role does banking play? These dramatic emissions reductions help Even so, Illinois, Kentucky, and New demonstrate that banking leads to SO2 reductions Hampshire emitted 29, 31, and 36 percent less, earlier than if banking of allowances were not respectively, than their 1990 emissions levels. permitted. Banking encourages early reductions by allowing participants to garner the benefits of over Moreover, there have been significant emission control through the storage of unused allowances reductions in some of the highest emitting areas and by assuring that those allowances will retain of the country. For example, electric utilities in their compliance value in future years. Phase I units Ohio and Indiana reduced SO2 emissions by 44 are likely to be banking allowances in anticipation of percent and 50 percent respectively from 1990 the more stringent emissions limits and higher costs levels. This outcome is important because of Phase II. Although banked allowances will delay emissions reductions in these high emitting states the full effect of the emissions cap on SO2 until 2010, support a fundamental premise of Title IV's the early reductions of emissions may provide greater market based approach, i.e., that the highest health and environmental benefits than would have emitting plants have an incentive to make occurred without banking (see Figure 1). substantial reductions in emissions because they Where did extra S02 reductions occur? Allowance Utilization, Relative to 1997 Allocation Nearly every state containing Phase I affected 150000 units enjoyed overall emissions reductions. 100000 Figure 2 illustrates the tons of sulfur dioxide 50000 each state emitted below or in excess of its initial 0 NH KY IL allocation. In all but three states, utilities Allowances -50000 IA KS MS MA MD MN FL MI 100000 AL emitted less SO2 than originally provided for in OH WI 150000 NY their allowance allocation. Only Illinois, PA -200000 TN W Kentucky, and New Hampshire emitted more -250000 MO/ than originally allocated, but in amounts -300000 350000 GA overshadowed by large, extra reductions in the other states. Figure 2: The differences between allowances initially allocated and allowances retired show that almost every state (IL, KY, NH excepted) made greater SO₂ reductions than required. can achieve these reductions at lower cost per ton than most lower-emitting plants. Concerns that Real and Potential SO2 Emissions the biggest emitters of SO2 would simply buy in Ohio, 1997 allowances and continue to emit at their historical 2,500,000 2,099,653 levels have thus far proven to be unwarranted. 2,000,000 Tons S02 1,500,000 1,288,197 1,184,641 Is trading shifting the location of emissions? 1,000,000 Other An analysis of the allowances retired in 1996 and allowances 500,000 Ohio 1997 reveals that the overwhelming majority of allowances 0 allowances were retired in the same state in which 1990 1997 1997 they were initially allocated (see figure 3). For Emissions Allocation missions example, affected units in Ohio were allocated 1,288,197 allowances, in aggregate, for the 1997 Figure 3: Typical of Phase I states, Ohio retired mostly its own allowances. The band represents out-of-state allowances compliance year. At year end, these units emitted retired in 1997. 1,184,641 tons of emissions. The figure depicts the difference: affected utilities in Ohio emitted 103,556 fewer tons of SO2 than allowed by their allowances were obtained from other units allocation. This pattern of retiring same-state within their own companies. This trend is allowances shows that allowance trading has not led illustrated by a geographic information system to emissions shifting from one state to another in (GIS) analysis of utility units that retired more any significant amount. The extra emissions allowances than initially allocated in 1996. The reductions achieved in almost every affected state analysis covered only the allowances retired by further confirms that trading has not led to major these units since the units' environmental impact inflows of emissions into the vast majority of states. was greater than provided for in the 1996 allocations. From where have utilities obtained "extra" allowances needed for compliance? The locations of both the units that retired more In cases where electric utility units required more allowances than allocated and the units to which allowances than their allocation, most of these those allowances were initially allocated were GEOGRAPHIC CENTERS OF TRADING, 1996 Buying Activity Selling Activity Mean Center of Buying Activity. Mean Center of Selling Activity. Weighted by Allowances Bought Weighted by Allowances Sold Figure 4: Our study revealed that allowance transfer activity was highest in the Midwest. 3 mapped (figure 4). For each map, the geographic are not expected to impact sensitive ecosystems. mean was calculated to determine the center of In fact, in many sensitive areas such as the trading activity. The means were weighted by the Adirondacks and mid-Appalachians, there is number of allowances bought or sold. projected to be virtually no change in deposition due to trading. EPA will continue to analyze the The geographic centers were very close. This impacts of allowance trading to assess how this proximity indicates that there was no significant flow innovative market-based program affects of allowances from one region to another among environmental results. these units. In addition, the analysis reveals that the high emitters were more concentrated in the For More Information Midwest, as were the units that sold them the excess For more information on the Acid Rain allowances. The tight geographic correlation Program, please visit the Acid Rain Program's corroborates EPA's observations that units tended to web site at www.epa.gov/acidrain, or call the acquire their excess allowances from within their Acid Rain Hotline at 202-564-9620. own company. 1. U.S. EPA, Acid Rain Program Compliance When will utilities use the allowances they Report 1997, EPA 430/R-98-012, August 1998. buy? Many utilities will acquire allowances now to ensure their ability to comply later. Thus, even though a utility is buying allowances, it does not necessarily intend to use the allowances immediately. In some cases allowances being traded today bear vintages that make them valid only in future years. Furthermore, allowances often pass through several owners before being retired. The most important transfer in an allowance's life, from the environmental perspective, is the retirement of the allowance, because only then does the allowance correspond with one ton of real pollution. Where an allowance is held today does not necessarily indicate where it will be retired. In spite of these uncertainties surrounding how much and when utilities will emit, the emissions cap ensures that pollution reduction goals will be met. For now, utilities will continue to adapt their trading strategies as the Acid Rain Program matures. Currently almost all of the allowances used for compliance come from sources within the same utility company and by extension, the same geographic region. In the future, companies are likely to use more allowances acquired from other companies across greater geographic distances. However, even if all economically advantageous trades are made, EPA models project that trading will lead to only small changes in deposition which 4 Acid Deposition Sulfur and nitrogen oxides are emit- ranges from 20 to 60 percent of total meteorological measurements that ted into the atmosphere primarily deposition. are used to estimate rate of the actual from the burning of fossil fuels. The United States Environmen- deposition, or "flux." Data representing These emissions react in the atmo- tal Protection agency (EPA) is required total deposition loadings (e.g., total sphere to form compounds that are by several Congressional and other sulfate or nitrate) are what many envi- transported long distances and are mandates to assess the effectiveness ronmental scientists use for integrated subsequently deposited in the form of of air pollution control efforts. These ecological assessments. pollutants such as particulate matter mandates include Title IX of the Clean (sulfates and nitrates), SO₂, NO₂, Air Act Amendments (CAAA), the PRIMARY ATMOSPHERIC nitric acid and when reacted with National Acid Precipitation Assess- DEPOSITION MONITORING volatile organic compounds (VOCs) ment Program (NAPAP), the Govern- NETWORKS form ozone. The effects of atmo- ment Performance and Results Act, spheric deposition include acidifica- and the U.S. Canada Air Qualitv The National Atmospheric Deposi- ion of lakes and streams, nutrient Agreement. One measure of effective- tion Program (NADP) and the Clean enrichment of coastal waters and ness of these efforts is whether sus- Air Status and Trends Network large river basins, soil nutrient deple- tained reductions in the amount of (CASTNet), described in detail below, tion and decline of sensitive forests, atmospheric deposition over broad were developed to monitor wet and agricultural crop damage, and im- geographic regions are occurring. dry acid deposition, respectively. pacts on ecosystem biodiversity. However, changes in the atmosphere Monitoring site locations are pre- Toxic pollutants and metals also can happen very slowly and trends are dominantly rural by design to assess be transported and deposited through often obscured by the wide variability the relationship between regional atmospheric processes. (See Chapter of measurements and climate. Nu- pollution and changes in regional 5: Air Toxics.) merous years of continuous and con- patterns in deposition. CASTNet also Both local and long-range sistent data are required to overcome includes measurements of rural emission sources contribute to atmo- this variability, making long-term ozone and the chemical constituents spheric deposition. Total atmospheric monitoring networks especially critical of PM₂ Rural monitoring sites of deposition is determined using both for characterizing deposition levels NADP and CASTNet provide data wet and dry deposition measure- and identifying relationships among where sensitive ecosystems are lo- ments. Wet deposition is the portion emissions, atmospheric loadings, and cated and provide insight into natural dissolved in cloud droplets and is effects on human health and the envi- background levels of pollutants deposited during rain or other forms ronment. where urban influences are minimal. of precipitation. Drv deposition is the For wet and dry deposition, These data provide needed informa- portion deposited on dry surfaces these studies typically include mea- tion to scientists and policy analysts during periods of no precipitation as surement of concentration levels of to studv and evaluate numerous particles or in a gaseous form. Al- key chemical components as well as environmental effects, particularly though the term "acid rain" is widely precipitation amounts. For drv depo- those caused by regional sources of recognized, the dry deposition portion sition, analyses also must include emissions for which long range trans- CHAPTER 7: ACID DEPOSITION 101 NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT, 1997 Figure 7-1. The NADP/NTN Network. port plays an important role. Mea- that time, the network became known TRENDS ANALYSES FOR SULFATE surements from these networks are as the NADP/NTN (National Trends AND NITRATE CONCENTRATIONS IN also important for understanding Network). By the mid-1980s, the NADP WET DEPOSITION non-ecological impacts of air pollu- had grown to nearly 200 sites where it tion such as visibility impairment stands today as the longest running Sulfate concentrations in precipita- and damage to materials, particu- national deposition monitoring net- tion have decreased over the past two decades.¹ The reductions were larly those of cultural and historical work (see Figure 7-1). importance. The NADP analyzes the constitu- relatively large in the early 1980s ents important in precipitation chem- followed by more moderate declines National Atmospheric Deposition istry, including those affecting rainfall until 1995. These reductions in sul- Network acidity and those that may have eco- fates are similar to changes in SO₂ The NADP was initiated in the late logical effects. The Network measures emissions. In 1995, however, con- centrations of sulfates in precipita- 1970s as a cooperative program be- sulfate, nitrate, hydrogen ion (measure tween federal and state agencies, of acidity), ammonia, chloride, and tion over a large area of the Eastern United States exhibited a dramatic universities, electric utilities, and other base cations (calcium, magnesium, industries to determine geographical potassium). To ensure comparability and unprecedented reduction.² In of results, laboratory analyses for all 1995 and continued in 1996, sul- patterns and trends in precipitation chemistry in the United States. Col- samples are conducted by the NADP's fates have been estimated to be 10- Central Analytical Lab at the Illinois 25 percent lower than levels ex- lection of weekly wet deposition State Water Survev. A new subnet- pected with a continuation of samples began in 1978. The size of work of the NADP, the Mercurv Deposi- 1983-1994 trends (see Figure 7-2). the NADP Network grew rapidly in the tion Network (MDN) measures mer- This important reduction in acid pre- early 1980s when the major research effort by the NAPAP called for charac- cury in precipitation. The MDN is cipitation is directly related to the terization of acid deposition levels. At discussed in Chapter 5 of this report. large regional decreases in SO₂ emis- 102 CHAPTER 7: ACID DEPOSITION NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT. 1997 Figure 7-2. Percent differences in mean annual measured sulfate concentrations as compared to projected concentrations for 1995-1996 for the Eastern United States (from NADP/NTN). Small squares on the map show locations of electric utility plants affected under Phase I of the acid rain program. Areas on the map depicting higher sulfate concentrations (e.g., south and east of Lake Michigan and the southwestern portion of map) appear to be due to below average precipitation volumes, which are associated with higher concentrations of sulfate. In addition, these results may have been affected by SO₂ emission increases at some Phase II emissions sources that will not be controlled by the acid rain program until the year 2000. -20 -15 -12 -b 5 10 20 sions resulting from Phase [ of the The dense network of NADP/ it also can contribute to higher lev- acid rain program (see the SO₂ sec- NTN sites facilitate the development els of wet deposition. Figures 7-5 tion in Chapter 2). The largest re- of concentration and wet deposition and 7-6 present estimates for total ductions in sulfate concentrations maps to describe the trends and wet deposition of sulfates and ni- occurred along the Ohio River Val- spatial patterns in the constituents trates respectively, by multiplying ley and in states immediately down- of acid precipitation. Figures 7-3 concentration by the total amount of wind of this region. For example, and 7-4 show sulfate and nitrate precipitation. During 1997, the the average reduction in sulfate concentrations in precipitation lev- highest sulfate wet deposition oc- concentrations in Ohio was approxi- els for 1997. Sulfate concentrations curred in western New York State mately 21 percent, in Marvland, 27 in precipitation are highest in the extending southward through the percent, and in Pennsvlvania, 15 per- Great Lake States and areas extend- Ohio Valley and along the Appala- cent. The largest decrease (32 per- ing eastward. Nitrates in precipita- chian ridge. Nitrate deposition cent) occurred in the northern portion tion are more regionally uniform. shows a similar pattern. of West Virginia. Reductions in hvdro- The highest nitrate levels in precipi- gen ion concentrations (H1) in the tation are in the vicinity of the Great Clean Air Status and Trends East. the primary indicator of precipi- Lakes, with relatively high concen- Network tation acidity, were very similar to trations extending from the Plains The CASTNet provides atmospheric those of sulfate concentrations, both States to the Northeast. data on the dry deposition compo- in magnitude and location. Nitrate Reported concentrations and nent of total acid deposition, ground- concentrations at NADP/NTN sites total wet deposition are both depen- level ozone and other forms of atmos- were not appreciably different in dent on the amount of precipitation in pheric pollution. CASTNet is 1995-1996 from historical levels.³ a particular year. While larger considered the nation's primary Analyses based on the 1997 data are amounts of precipitation can dilute source for atmospheric data to esti- not yet available. the measured pollutant concentration, mate dry acidic deposition and to CHAPTER 7: ACID DEPOSITION 103 NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT. 1997 Figure 7-3. Sulfate concentration in precipitation. 1997. e0.2 -0.3 .10 a0.4 -0.5 606 0.3 DRIVE 01.1 a0.3 no.4 e0.8 a e0.1 -0.2 -0.8.1 TOA e0.2 0.3 0.8 THE =0.2 01.0 -2.1 -0.3 - -0.3 $1.5 e0.8 $1.8 -2.0 0.5 0.8 a0.8 #1.4 .1. -0.4 422 .3 $0.6 a17 a2.0 =0.1 90.7 -1.6 a1.9 25 0.7 -0.3 0.2 -0.0 IVD a0.7 #111 +1.5 .212 2.0 e1.2 22 a25 a0.5 a218 -0.1 -0.7 "0.7 a2 #1.2 e0.4 2.2 e0.8 .1.1 2.2 -1.9 =0.1 e0.7 40.8 -26 .1 0.5 a0.7 $1.6 1.6 =0.8 -0.5 e1.1 are $1.6 CO.7 #1.0 $1.4 *1.4 &8 all -20 .1.3 -1.8 a0.4 -0.8 e0.9 =1.2 $1.42 1.5 SO. 2. a1.1 0.41.1 (mg/L) $1.1 an5 #1.1 e1.3 e1.2 $1.3 #1.1 *1.6 <0.5 0.8 -0.8 e1.1 *1.0*1.0 #1.2 $1.5 -0.9 $12 $10 #0.8 & $1.2 1.4 et.1% 1.4 2.5 -1.1 a0.7 Figure 7-4. Nitrate concentration in precipitation. 1997. a0.3 =0.5 =0.3 e0.8 607 -0.4 08 ORDER *1.0 a0.2 -0.8 -0.9 -0.3 $1 =0.1 1.5 -1.2. 0.4 =0.8 1.0 $1.1 =0.3 .1.2 0.5 a0.7 e1.6 $1.9 word .1.2 0.8 -0.8 $1.1 e1.7 e1.6 $1.6 -0.6 4 1.2 on 17 .02 1.3 $1.4 1.8 0.8 *0.7 -0.8 -0.9 1.4 .1.2 *114 a1.4 +1.4 #18 1.7 m0.8 =1.6 -18 =0.3 a1.1 *1.1 -1.8 1.5 $1.6 e0.6 0.0 $1.3 .1.3 *1.4 $1.4 a0.4 -0.8 $12 #1.5 a1.4 2808 $1.2 $1.1 1.2 $1,4 a0.8 e1.1 gos st.3 *1.2 +1.0 12 a15 BIR -11 -0.7 *1.1 .00 $1.1 a0.8 -0.9 $0.87 91.1 a0.8 0.90.7 NO, (mg/L) a1.1 -0.9 a0.7 e1.1 e0.8 #1.8 a0.8 -0.7 <0.6 0.9 e0.8 =1.0 e0.7 -0.0 -0.8 -0.700 -0.8 e0.7 40.8 40.5 -0.8 1.0 -0.7y 0.0 >1.8 -0.7 -0.5 104 CHAPTER 7: ACID DEPOSITION NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT. 1997 Figure 7-5. Wet deposition of sulfate. 1997. & a2 & e1 #10 a2 04 01 17 of 01 & a0 6 of of & -25 a5 #13 04 e16 e1 e10 -20 #18 .1 a18 #18 01 +10 #20 g21 -20 01 *3.7 03 % o4 07 -28 19 #10 a13 #16 -20 a2 01 = -17 20 222 $18 04 08 23 -22 is 04 .25 19 a20 13 at of -13 atd #18 *12 $14 .3 622 -18 04 # $10 #18 e15 #16 =10 SO, e10 (kg/ha) =24 .15 $ a7 e17 05 -17 <3 $ all $16.18 #14 e13 .19 $18 #15 a11 07 22 e15 23 27 off, -12 Figure 7-6. Wet deposition of nitrate. 1997. a8 a4. .10 a2 e7 as 16) =7 80 #1 07 .15 & e13 #11 $18 a10 .12 a18 e17 e10 $15 #14 6 of e1 16 e11 #11 #12 e16 01 & *14 04 13 .12 #10 e16 15 #18 44 014 of e13 a15 913 013 +15 #10 e16 1612 e12 & & #12 e10 $12 a)2 NO, OF e10 (kg/ha) +14 of =12 #12 54 *10 at 011 .12 & of #12 673 #7 05 15 #10 16 20 e7 #10 CHAPTER 7: ACID DEPOSITION 105 NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT. 1997 Figure 7-7. The CASTNet Network. provide data on rural ozone levels. weekly average atmospheric databases provide the necessary Used in conjunction with other na- concentrations of sulfate, nitrate, data to estimate trends and spatial tional monitoring networks, ammonium, sulfur dioxide, and patterns in total atmospheric deposi- CASTNet is used to determine the nitric acid. tion. National Oceanic and Atmo- effectiveness of national emission hourly concentrations of ambient spheric Administration (NOAA) also control programs. Established in ozone levels. operates a smaller dry deposition 1987, CASTNet now comprises 72 meteorological conditions required network called Atmospheric Inte- monitoring stations across the United for calculating dry deposition grated Assessment Monitoring Net- States, as shown in Figure 7-7. The rates. work (AIRMoN) focused on ad- longest data records are primarily at dressing research issues specifically eastern sites. The majority of the Dry Deposition related to dry deposition measure- monitoring stations are operated by Dry deposition rates are calculated ment. EPA's Office of Air and Radiation; using atmospheric concentrations, however, 19 stations are operated by Concentration Trends Analysis at meteorological data, and information the National Park Service in coopera- CASTNet Sites on land use, vegetation, and surface tion with EPA. Of the total number of conditions. CASTNet complements CASTNet data were analyzed for the sites, 67 measure dry-deposition; 18 the database compiled by NADP. period 1989-1995. During this 7-year measure wet-deposition; 68 measure Because of the interdependence of period, atmospheric concentrations of ozone; and 8 measure aerosols for wet and dry deposition, CASTNet sulfur dioxide and sulfate at 34 east- visibility assessment. also collects wet deposition data at ern CASTNet sites showed statisti- Each CASTNet dry deposition the 18 sites where there are no cally-significant declining trends. station measures: NADP/NTN stations within a 50 km The average reduction in sulfur diox- radius. Together, these two long-term ide concentrations for all sites was 35 106 CHAPTER 7: ACID DEPOSITION NATIONAL AIR QUALITY AND EMISSIONS TRENDS REPORT. 1997 Figure 7-8. Trends in annual mean aerosol sulfate concentrations at Whiteface Monitoring Stations (NAMS) net- Mountain and Mayville, 1978-1996. works. Most air quality samples at SLAMS/NAMS sites are located in urban areas, while CASTNet sites SO₂ metric tons) 11 Midwest emissions are in rural locations. Hourly ozone measurements are taken at each of 9 the 50 sites operated by EPA. Data from these sites provide information 7 to help characterize ozone transport issues and ozone exposure levels. 5 Future trend reports will present information on rural O3 concentra- tions measured at CASTNet sites. SO4 (µg/m³) 5 Mayville References 3 1. Lynch, J.A., Grim, J.W., and Bowersox, V.C. 1995 Trends in Whiteface Mt. Precipitation Chemistry in the United 1 States: A National Perspective, 1980- 1978 1981 1984 1987 1990 1993 1996 1992. Atmospheric Environment Vol Also shown are annual SO₂ emissions for the Midwest as explained in the text. Lines 29, No. 11. through the points are the result of multiple regression and smoothing and are only 2. Lynch, J.A., Bowersox, V.C., and added to aid the eye. Grim, J.W. 1996. Trends in Precipita- tion Chemistry in the United States. percent and in sulfate concentra- air quality fluctuates from year to An Analysis of the Effects in 1995 of tions was 26 percent. Trends in year, the underlying trend in annual Phase I of the Clean Air Act Amend total nitrate concentrations (nitrates sulfate concentrations tracks emis- ments of 1990, Title IV. U.S. Geologi plus nitric acid) were not as pro- sions with both trends exhibiting a cal Survey. Open-file Report 96-0346. nounced, with an average reduction small decrease. Then, sulfates de- 3. Lynch, J.A., Bowersox, V.C., and Grim, of S percent. A regional estimate for clined sharply in 1995, correspond- J.W. Acid Rain Reduced in Eastern U.S.A. a cluster of sites in the Ohio River ing to a 36 percent reduction in re- Submitted to Environmental Science Valley showed a close correspon- gional emissions. During 1995, Technology. In Review. dence between declining sulfur emissions from this region ac- 4. Holland, D. M., Principe, P., and dioxide concentration (35 percent) counted for 38 percent of the na- Sickles, J.E., II. Trends in Atmospher- and declining sulfur dioxide emis- tional emission inventory. The air ic Concentration of Sulfur and Nitro- sions (32 percent) in this region.4 quality improvement (~1 µg/m²) was gen Species in the Eastern United The relationship between re- approximately 30 percent for Mavville States. Atmospheric Environment, Vol. 33, 37-49 gional SO₂ emissions and sulfate air and an impressive 47 percent for the quality is graphically illustrated in more distant Whiteface Mountain 5. Husain, L., Dutkiewicz, V.A., and Figure 7-8. This recently published location. 4 Dass, M. 1998. Evidence for Decrease graph compares long-term trends in Atmospheric Sulfer Burden in the (1978-1996) in annual mean aerosol Rural Ozone Eastern United States Caused by Re- sulfate concentrations at two rural duction in SO, Emissions. Geophysical Ozone data collected by CASTNet Research Letters Vol. 25 No. 7. locations in New York State with up- are complementary to the larger wind SO₂ emissions for the Midwest ozone data sets gathered by the region (MN, WI, IL, MI, IN, OH, WV, KY State and Local Air Monitoring Sta- and Western PA). Although average tions (SLAMS) and National Air CHAPTER 7: ACID DEPOSITION 107 EPA United States Office of Air and Radiation Updated September 1998 Environmental Protection Agency Acid Rain Division (62041) EPA +30-1-92-019 Overview of the ACID RAIN PROGRAM Acid Rain Program Introduction in 1997 there were a total of 423 Phase I affected units. The overall goal of the Acid Rain Program is to achieve significant environmental and public Phase II, which begins in the year 2000, tightens health benefits through reductions in emissions the annual emissions limits imposed on these of sulfur dioxide (SO₂) and nitrogen oxides large, higher emitting plants and also sets (NOx), the primary causes of acid rain. To restrictions on smaller, cleaner plants fired by achieve this goal at the lowest cost to society, coal, oil, and gas. The program affects existing the program employs both traditional and utility units serving generators with an output innovative, market-based approaches for capacity of greater than 25 megawatts and all controlling air pollution. In addition, the new utility units. program encourages energy efficiency and pollution prevention. The Act also calls for a 2 million ton reduction in NOx emissions by the year 2000. A Title IV of the Clean Air Act sets as its primary significant portion of this reduction will be goal the reduction of annual SO₂ emissions by achieved by coal-fired utility boilers that will be 10 million tons below 1980 levels. To achieve required to install low NOx burner technologies these reductions, the law requires a two-phase and to meet new emissions standards. tightening of the restrictions placed on fossil fuel-fired power plants. Operating Principles: Feasible, Flexible, Accountable Phase I began in 1995 and affects 263 units at 110 mostly coal-burning electric utility plants The Acid Rain Program is implemented located in 21 eastern and midwestern states. through an integrated set of rules and guidance Additional units may join Phase I of the designed to accomplish three primary program as substitution or compensating units; objectives: Achieve environmental benefits through reductions in SO2 and NOx emissions S02 Allowance Program Facilitate active trading of allowances and Emission Reductions use of other compliance options to 30 minimize compliance costs, maximize Utility SO2 Emissions (million tons) 25 Without Acid Rain Program economic efficiency, and permit strong economic growth 20 Promote pollution prevention and energy 15 efficient strategies and technologies 10 Each individual component fulfills a vital 5 With Acid Rain Program function in the larger program: the allowance trading system creates 0 low-cost rules of exchange that minimize 1980 1985 1990 1995 2000 2005 2010 government intrusion and make allowance trading a viable compliance strategy for Figure I The Acid Rain Program will result in a 10 reducing SO₂; million ton reduction in SO2 emissions from 1980 the opt-in program allows nonaffected levels by the year 2010. industrial and small utility units to rain accelerates the decay of building materials, Acid Rain Formation paints, and cultural artifacts, including irreplaceable buildings, statues, and sculptures that are part of our nation's cultural heritage. Prior to falling to the earth, SO₂ and NOx SO₂ NOx Acid Rain gases and their particulate matter derivatives, sulfates and nitrates, contribute to visibility degradation and impact public health. Implementation of the Acid Rain Program under the 1990 Clean Air Act Amendments will confer significant benefits on the nation. Coal-fired electric utilities By reducing SO₂ and NOx, many acidified and other sources that burn fossil fuels smit lakes and streams will significantly improve so sulfur dioxide and nitrogen exides that they can once again support fish life. Visibility will improve, allowing for increased Figure 2 Sulfur dioxide and nitrogen oxide emissions react enjoyment of scenic vistas across our country, with water vapor and oxidants in the atmosphere and are particularly in National Parks. chemically transformed into acid compounds. These compounds are deposited in rain or snow (wet deposition); Stress to our forests that populate the ridges of the compounds also join airborne particles and fall to the earth as dry deposition. mountains from Maine to Georgia will be reduced. Deterioration of our historic buildings and monuments will be slowed. participate in allowance trading; Finally, reductions in SO₂ and NOx will reduce the NOx emissions reduction rule sets new sulfates, nitrates, and ground level ozone NOx emissions standards for existing (smog), leading to improvements in public coal-fired utility boilers and allows health. emissions averaging to reduce costs; the permitting process affords sources Allowance Trading maximum flexibility in selecting the most cost-effective approach to reducing The Acid Rain Program represents a dramatic emissions; departure from traditional command and the continuous emission monitoring (CEM) control regulatory methods that establish requirements provide credible accounting specific, inflexible emissions limitations with of emissions to ensure the integrity of the which all affected sources must comply. market-based allowance system and the Instead, the program introduces an allowance achievement of the reduction goals; trading system that harnesses the incentives of the free market to reduce pollution. the excess emissions provision provides incentives to ensure self-enforcement, Under this system, affected utility units were greatly reducing the need for government allocated allowances based on their historic fuel intervention; and, finally, consumption and a specific emissions rate. the appeals procedures allow the regulated Each allowance permits a unit to emit 1 ton of community to appeal decisions with which SO₂ during or after a specified year. For each it may disagree. ton of SO₂ discharged in a given year, one allowance is retired, that is, it can no longer be Together these measures ensure the used. achievement of environmental benefits at the least cost to society. Allowances may be bought, sold, or banked. Any person may acquire allowances and Environmental Benefits participate in the trading system. However, regardless of the number of allowances a source Acid rain causes acidification of lakes and holds, it may not emit at levels that would streams and contributes to damage of trees at violate federal or state limits set under Title I of high elevations (for example, red spruce trees the Clean Air Act to protect public health. above 2,000 feet in elevation). In addition, acid 2 National SO2 Emissions by Source National NOx Emissions by Source Non-Road Engines Metals Processing and Vehicles 2% 3% On-Road Vehicles Fuel Combustion- 13% 17% Fuel Combustion- 30% Industrial Industrial Fuel Combustion- 5% 4% Fuel Combustion- Other Other 5% All Other 7% All Other Fuel Combustion Fuel Combustion- Electric Utility Non-Road Engines Electric Utility 67% 28% and Vehicles 19% Figure 3 Electric utilities contribute 67 percent of SO2 emissions and 28 percent of NOx emissions nationally. Source: National Air Pollutant Emission Trends, 1900-1996, US EPA, Office of Air Quality Planning and Standards, December 1997 (EPA-454/R-97-011). During Phase II of the program, the Act sets a by which EPA determines compliance with the permanent ceiling (or cap) of 8.95 million emissions limitations. Any party interested in allowances for total annual allowance participating in the trading system may open an allocations to utilities. This cap firmly restricts ATS account by submitting an application to emissions and ensures that environmental EPA. Accounts contain information on unit benefits will be achieved and maintained. account balances, account representatives (which must be appointed by each trading Annual Reconciliation party), and serial numbers for each allowance. ATS is computerized to expedite the flow of Annual Reconciliation refers to EPA's task of data and to assist in the development of a viable reconciling the annual emissions of a unit with market for allowances. its allowance holdings. At the end of the year, utilities are granted a 30-day true-up or grace Auctions period, during which SO₂ allowances may be purchased, if necessary, to cover each unit's Because the availability of allowances is crucial emissions for the year. At the end of the grace to ensure both the economic efficiency of the period, the allowances a unit holds in its emissions limitation program and the addition compliance account must equal or exceed the of new electric-generating capacity, Title IV of annual SO₂ emissions recorded by the unit's the Clean Air Act Amendments mandates that monitoring system. Any remaining allowances EPA hold or sponsor yearly auctions of may be sold or banked for use in future years. allowances for a small portion of the total allowances allocated each year. The auctions The EPA must consider a number of issues help ensure that units have a public source of related to the operation of the unit over the past allowances beyond those allocated initially to year, including the unit's SO₂ emissions, existing units. Moreover, the auctions provided whether or not the unit was "underutilized," price information to the allowance market early and whether the unit is subject to special in the regulatory program. provisions (e.g., a state law limiting emissions on an system-wide basis). Voluntary Entry: The Opt-in Program The Allowance Tracking System The Opt-in Program expands EPA's Acid Rain Program to include additional SO₂-emitting EPA has instituted an electronic record keeping sources. Recognizing that there are additional and notification system called the Allowance emission reduction opportunities in the Tracking System (ATS) to track allowance industrial sector, Congress established the transactions and the status of allowance Opt-in Program under section 410 of the Clean accounts. ATS is the official tally of allowances Air Act Amendments of 1990. The Opt-in 3 Program allows sources that are not required to The NOx program embodies many of the same participate in the Acid Rain Program the principles of the SO₂ trading program in its opportunity to enter the program on a design: a results-orientation, flexibility in the voluntary basis, reduce their SO₂ emissions, and method to achieve emission reductions, and receive their own acid rain allowances. program integrity through measurement of the emissions. However, it does not "cap" NOx The participation of these additional sources emissions as the SO₂ program does, nor does it will reduce the cost of achieving the 10 million utilize an allowance trading system. ton reduction in SO₂ emissions mandated under the Clean Air Act. As participating sources Emission limitations for the NOx boilers reduce their SO₂ emissions at a relatively low provide flexibility for utilities by focusing on the cost, their reductions -- in the form of emission rate to be achieved (expressed in allowances -- can be transferred to electric pounds of NOx per million Btu of heat input). utilities where emission reductions are more Two options for compliance with the emission expensive. limitations are provided: compliance with an individual emission rate for a boiler or The Opt-in Program offers a combustion averaging of emission rates over two or more source a financial incentive to voluntarily units to meet an overall emission rate limitation. reduce its SO₂ emissions. By reducing (In the latter case, units must have the same emissions below its allowance allocation, an owner or operator.) opt-in source will create unused allowances, which it can sell in the SO₂ allowance market. These options give utilities flexibility to meet Opting in will be profitable if the revenue from the emission limitations in the most allowances exceeds the combined cost of the cost-effective way and allow for the further emissions reduction and the cost of development of technologies to reduce the cost participating in the Opt-in Program. of compliance. Pollution Prevention If a utility properly installs and maintains the appropriate control equipment designed to The allowance trading system contains an meet the emission limitation established in the inherent incentive for utilities to prevent regulations, but is still unable to meet the pollution, since for each ton of SO₂ that a utility limitation, the NOx program allows the utility avoids emitting, one fewer allowance must be to apply for an alternative emission limitation retired. Utilities that reduce emissions through (AEL) that corresponds to the level that the energy efficiency and renewable energy are able utility demonstrates is achievable. to sell, use, or bank their surplus allowances. As also provided in the Act, EPA has set aside Phase I of the program, which was delayed a a reserve of 300,000 allowances to stimulate year due to litigation, began on January 1, 1996, energy efficiency and renewable energy and affects two types of boilers (which are generation. Those utilities that either implement among those already targeted for Phase I SO₂ demand-side energy conservation programs to reductions): dry-bottom wall-fired boilers and curtail emissions or install renewable energy tangentially fired boilers. Dry-bottom wall-fired generation facilities may be eligible to receive boilers must meet a limitation of 0.50 lbs of extra allowances from this reserve. NOx per mmBtu averaged over the year, and tangentially fired boilers must achieve a The NOx Program limitation of 0.45 lbs of NOx per mmBtu, again, averaged over the year. Approximately The Clean Air Act Amendments of 1990 set a 170 boilers must comply with these NOx goal of reducing NOx by 2 million tons from performance standards during Phase I. 1980 levels. The Acid Rain program focuses on one set of sources that emit NOx, coal-fired The regulations to govern the Phase II portion electric utility boilers. As with the SO₂ emission of the program, which begins in 2000, were reduction requirements, the NOx program is published in December 1996. These implemented in two phases, beginning in 1996 regulations set lower emission limits for Group and 2000. 1 boilers first subject to regulations in Phase II. In addition, these regulations would establish 4 initial NOx emission limitations for Group 2 Representative, to represent the owners and boilers, which include boilers applying operators of the source in all matters relating to cell-burner technology, cyclone boilers, wet the holding and disposal of allowances for its bottom boilers, and other types of coal-fired units that are affected by the Clean Air Act. The boilers. Designated Representative is also responsible for all submissions pertaining to permits, Emissions Monitoring and Reporting compliance plans, emission monitoring reports, offset plans, compliance certification, and other Under the Acid Rain Program, each unit must necessary information. A source may appoint continuously measure and record its emissions an Alternate Designated Representative to act of SO₂, NOx, and CO2, as well as volumetric on behalf of the Designated Representative. flow and opacity. In most cases, a continuous emission monitoring (CEM) system must be Permitting used. There are provisions for initial equipment certification procedures, periodic quality The Designated Representative for each source assurance and quality control procedures, is required to file a permit application for the record keeping and reporting, and procedures source and a compliance plan for each affected for filling in missing data periods. Units report unit at the source. The Acid Rain permits and hourly emissions data to EPA on a quarterly compliance plans are simple, allow sources to basis. This data is then recorded in the fashion a compliance strategy tailored to their Emissions Tracking System, which serves as a individual needs, and foster trading. For repository of emissions data for the utility example, they allow sources to make real-time industry. The emissions monitoring and allowance trading decisions through the use of reporting systems are critical to the program. automatic permit amendments. They instill confidence in allowance transactions by certifying the existence and The permits stipulate the initial allowance quantity of the commodity being traded. They allocation for each affected unit at a source. ensure that NOx averaging plans are working. Monitoring also ensures, through accurate accounting, that the SO₂ and NOx emissions reduction Phase 1 Sources goals are met. Excess Emissions The owners or operators of delinquent units must pay a penalty of $2,000, adjusted for inflation, per excess ton of SO₂ (if annual emissions exceed the number of allowances held) or NOx emissions. In addition, Phase II Sources violating utilities must offset the excess SO₂ emissions with allowances in an amount equivalent to the excess. A utility may either have allowances deducted immediately or submit an excess emissions offset plan to EPA that outlines how these cutbacks will be achieved. Designated Representatives Figure 4 Phase I and Phase II affected units. In Phase II the number of Each source appoints one affected units will increase to 2,220 as smaller, cleaner plants are included in individual, the Designated the regulatory network. 5 Permit applications must certify that each SO2 Emissions from unit account will hold a sufficient number of allowances to cover the unit's SO₂ 263 Phase I Units emissions for the year, will comply with 12 the applicable NOx limit, and will 10 9.4 9.3 monitor and report emissions. Permits 8.7 Allowable Emissions are subject to public comment before 8 7.4 7.1 approval. 6.0 6 SO2 (million tons) 4.5 4.8 .8.4. 4.8 4 Compliance Options: Freedom to Choose 2 0 The Acid Rain Program allows sources to select their own compliance strategy. For 1980 1985 1990 1995 1996 1997 example, to reduce SO₂ an affected source may repower its units, use cleaner Figure 5 Emissions at the 263 Phase I units have been well below burning fuel, or reassign some of its their required level. energy production capacity from dirtier units to cleaner ones. Sources also may decide effects of the program through economic and to reduce electricity generation by adopting environmental studies. conservation or efficiency measures. Some of the options afford the unit special treatment, The Acid Rain Program is already being viewed such as a compliance extension or extra around the world as a prototype for tackling allowances. Most options, like fuel switching, emerging environmental issues. The allowance require no special prior approval, allowing the trading system capitalizes on the power of the source to respond quickly to market conditions marketplace to reduce SO₂ emissions in the without needing government approval. For most cost-effective manner possible. The NOx, the source may meet the performance permitting program allows sources the flexibility standard on a utility-unit basis, enter into an to tailor and update their compliance strategy emissions averaging plan, or apply for an based on their individual circumstances. The alternative emissions limitation. continuous emissions monitoring and reporting systems provide the accurate accounting of In either case, the program allows affected emissions necessary to make the program work, utilities to combine these and other options in and the excess emissions penalties provide ways they fit in order to tailor their compliance strong incentives for self-enforcement. Each of plans to the unique needs of each unit or these separate components contributes to the system. effective working of an integrated program that lets market incentives do the work to achieve A Model Program cost-effective emissions reductions. The General Accounting Office confirmed the EPA gained broad input into the development benefits of this approach, projecting that the of the Acid Rain Program by consulting with allowance trading system could save as much as representatives from various stakeholder $3 billion per year over 50 percent groups, including utilities, coal and gas compared with a command and control companies, emissions control equipment approach typical of previous environmental vendors, labor, academia, Public Utility protection programs. Commissions, state pollution control agencies, and environmental groups. For More Information EPA is maintaining this open door policy as it implements the program, and it continues to For more information on the Acid Rain solicit ideas from the numerous and diverse Program, please visit the Acid Rain Program's individuals, and groups interested in acid rain World Wide Web site at http:// control. Inaddition, EPA is collaborating with www.epa.gov/acidrain or contact the Acid groups who wish to evaluate the benefits and Rain Hotline at 202-564-9620. 6 Allowance Trading System Fact Sheet http://www.epa.gov/acidrain/allsys.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Allowance Trading System Fact Sheet The innovative, market-based SO2 allowance trading component of the Acid Rain Program allows utilities to adopt the most cost-effective strategy to reduce SO2 emissions at units in their systems. The Acid Rain Program operating permit outlines the specific program requirements and compliance options chosen by each source. Affected utilities are also required to install systems that continuously monitor emissions of SO2, NOx, and other related pollutants in order to track progress, ensure compliance, and provide credibility to the trading component of the program. In any year that compliance is not achieved, excess emissions penalties will apply, and sources either will have allowances deducted immediately from their accounts or may submit a plan to EPA that specifies how the excess SO2 emissions will be offset. Introduction Allowance trading is the centerpiece of EPA's Acid Rain Program, and allowances are the currency with which compliance with the SO2 emissions requirements is achieved. Through the market-based allowance trading system, utilities regulated under the program, rather than a governing agency, decide the most cost-effective way to use available resources to comply with the acid rain requirements of the Clean Air Act. Utilities can reduce emissions by employing energy conservation measures, increasing reliance on renewable energy, reducing usage, employing pollution control technologies, switching to lower sulfur fuel, or developing other alternate strategies. Units that reduce their emissions below the number of allowances they hold may trade allowances with other units in their system, sell them to other utilities on the open market or through EPA auctions, or bank them to cover emissions in future years. Allowance trading provides incentives for energy conservation and technology innovation that can both lower the cost of compliance and yield pollution prevention benefits. The market-based allowance trading system capitalizes on the power of the marketplace to reduce SO2 emissions cost-effectively and uses economic incentives to promote conservation and the development of innovative technology. The Acid Rain Program may establish a precedent for solving other environmental problems in a way that minimizes the costs to society and promotes new technologies. Frequently Asked Questions About the Allowance System 1. What Are Allowances? 2. How Are Allowances Allocated? 3. How Else Can Allowances Be Obtained? 4. Who May Participate in Allowance Trading? 5. What Is the System for Keeping Track of Allowances? 6. What Information is Contained in ATS Accounts? 7. How Are Allowance Transfers Submitted? 1 of 5 4/22/99 4:35 PM Allowance Trading System Fact Sheet http://www.epa.gov/acidrain/allsys.html 8. How Will Compliance Be Determined? 9. How Can Allowance Allocation, Transfer, Sale, or Deduction Decisions Be Appealed? What Are Allowances? An allowance authorizes a unit within a utility or industrial source to emit one ton of SO2 during a given year or any year thereafter. At the end of each year, the unit must hold an amount of allowances at least equal to its annual emissions, i.e., a unit that emits 5,000 tons of S02 must hold at least 5,000 allowances that are usable in that year. However, regardless of how many allowances a unit holds, it is never entitled to exceed the limits set under Title I of the Act to protect public health. Allowances are fully marketable commodities. Once allocated, allowances may be bought, sold, traded, or banked for use in future years. Allowances may not be used for compliance prior to the calendar year for which they are allocated. How Are Allowances Allocated? Allowances are allocated for each year beginning in 1995. In Phase I, EPA allocates allowances to each unit at an emission rate of 2.5 pounds of SO2/mmBtu (million British thermal units) of heat input, multiplied by the unit's baseline mmBtu (the average fossil fuel consumed from 1985 through 1987). These allowance allocations are listed in Table A of the Clean Air Act and codified in the Allowance System Regulations (Part 73, Table 1). Alternative or additional allowance allocations are made for various units, including affected units in Illinois, Indiana, and Ohio, which will be allocated a pro rata share of 200,000 additional allowances each year from 1995 to 1999. In Phase II, which begins in the year 2000, the limits imposed on Phase I plants are tightened, and emissions limits are also imposed on smaller, cleaner units. Allowance allocation calculations are made for various types of units, such as coal- and gas-fired units with low and high emissions rates or low fuel consumption. EPA allocates allowances to each unit at an emission rate of 1.2 pounds of SO2/mmBtu of heat input, multiplied by the unit's baseline. During Phase II, the Act places a cap at 8.95 million on the number of allowances issued to units each year. This effectively caps emissions at 8.95 million tons annually and ensures that the mandated emissions reductions are maintained over time. How Else Can Allowances Be Obtained? In addition to annual allocations, allowances are also available upon aplication to three EPA reserves. In Phase I, units can apply for and receive additional allowances by installing 2 of 5 4/22/99 4:35 PM Allowance Trading System Fact Sheet http://www.epa.gov/acidrain/allsys.html qualifying Phase I technology (a technology that can be demonstrated to remove at least 90 percent of the unit's SO2 emissions) or by reassigning their reduction requirements among other units employing such technology. A second reserve provides allowances as incentives for units achieving SO2 emissions reductions through customer-oriented conservation measures or renewable energy generation. The third reserve contains allowances set aside for auctions, which are sponsored yearly by EPA. In addition, allowances are given as incentives for utilities that replace boilers with new, cleaner and more efficient technologies. The incentives also apply to small diesel fuel refiners that have exceeded the Clean Air Act requirements to remove sulfur from fuels. Units that began operating in 1996 or later will not be allocated allowances. Instead, they will have to purchase allowances from the market or from the EPA auctions and direct sales to cover their SO2 emissions. Who May Participate in Allowance Trading? Allowances may be bought, sold, and traded by any individual, corporation, or governing body, including brokers, municipalities, environmental groups, and private citizens. The primary participants in allowance trading are officials designated and authorized to represent the owners and operators of electric utility plants that emit SO2. Other potential participants are utility power pools, or groups of units choosing to aggregate some or all of the allowances held by the individual units within the pool. The parties involved in the pool determine the details of these allowance-pooling arrangements. What Is the System for Keeping Track of Allowances? EPA's role in allowance trading is to record allowance transfers that are used for compliance and to ensure at the end of the year that a unit's emissions do not exceed the number of allowances it holds. To accomplish this, EPA maintains an Allowance Tracking System (ATS). Each affected utility unit, corporation, group, or individual holding allowances has an account in the ATS. Parties must notify EPA to have transfers recorded in their ATS account, but it is not necessary to record all transfers with EPA until such time that the allowances are to be used to meet a unit's SO2 emissions limitation requirement. ATS accounts are, however, the official records for allowance holdings and transfers used for compliance purposes. To facilitate tracking and recording, EPA assigns every account an identification number and every allowance a serial number. EPA established accounts for utility units affected by both Phase I and Phase II. Each unit account consists of a compliance subaccount for allowances that may be used for compliance in the current year and future year subaccounts for allowances to be used in years to come. Any person or group, including brokers and investors, wishing to purchase allowances may open an ATS account. To open a general ATS account, the interested party submits the Account Information Form to EPA. 3 of 5 4/22/99 4:35 PM Allowance Trading System Fact Sheet http://www.epa.gov/acidrain/allsys.html What Information is Contained in ATS Accounts? ATS accounts tracks the issuance of all allowances; the holdings of allowances in accounts; the holdings of allowances in various allowance reserves, such as the EPA Auction and Sale Reserve and the Conservation and Renewable Energy Reserve; the deduction of allowances for compliance purposes; and the transfer of allowances between accounts. Information in the ATS accounts is available to the public. How Are Allowance Transfers Submitted? Allowance transfer requests and all correspondence with EPA concerning compliance with the Acid Rain Program must be performed by authorized account representatives. For a unit account, the Designated Representative, who represents the owners and operators of that unit, performs this function. For a general account, the Authorized Account Representative is the person who represents the parties with an ownership interest in the allowances, and who signs the Account Information Form to open the account. To request the transfer of allowances from one account to another, the Authorized Account Representative submits to EPA an Allowance Transfer Form. The Authorized Account Representatives of both the transferor and transferee must sign the form. How Will Compliance Be Determined? At the end of the year, units must hold in their compliance subaccounts a quantity of allowances equal to or greater than the amount of SO2 emitted during that year. To cover their emissions for the previous year, units must finalize allowance transactions and submit them to EPA by January 30 to be recorded in their unit accounts. The amount of emissions is determined in accordance with the monitoring and reporting requirements described in the Continuous Emission Monitoring Rule. After the Januarv 30 deadline and the final submitted transfers are recorded, EPA deducts allowances from each unit's compliance subaccount in an amount equal to its SO2 emissions for that year. If the unit's emissions do not exceed its allowances, the remaining allowances are carried forward, or banked, into the next year's subaccount, which then becomes the current compliance subaccount. If a unit's emissions exceed its allowances, the unit must pay a penalty and surrender allowances for the following year to EPA as excess emission offsets. Unless otherwise provided in an offset plan, EPA deducts allowances from the compliance subaccount in an amount equal to the excess emissions. 4 of 5 4/22/99 4:35 PM Allowance Trading System Fact Sheet http://www.epa.gov/acidrain/allsys.html How Can Allowance Allocation, Transfer, Sale or Deduction Decisions Be Appealed? A person challenging an allowance, transfer, sale, or deduction decision may file a petition for review with the Environmental Appeals Board. Only the authorized account representative or certifying official involved may appeal such a decision. If EPA makes an error in the recordation of an allowance transfer, the authorized account representative may file a claim of error requesting EPA to correct the mistake. Where the claim of error procedure is applicable, the decision may not be appealed to the Environmental Appeals Board if a claim of error notification was not first submitted. Final agency actions may be appealed to the federal courts. Acid Rain Program Home | Trading and Trends I Contact Us http://www.epa.gov/acidrain/allsys.html Last Modified October 1997 5 of 5 4/22/99 4:35 PM Trading Activity Breakdown http://www.epa.gov/acidrain/ats/qlyupd.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Trading Activity Breakdown: Cumulative Graphics Definitions for all of the trading categories listed in the charts below are available. Through December of 1998, over 4,900 transfers moving 62.8 million allowances were reporte the Allowance Tracking System. Approximately 60 percent of these allowances (38.1 million) W transferred within organizations, as delineated in Figure 1. The remaining 24.7 million allowance transferred through December 1998 were transferred between organizations. Figure 2 displays t breakdown of these transfers between distinct entities, which represent "true" market activity to observers of the program. The growing magnitude of trades between organizations is depicted i Figure 3. FIGURE 1 Allowances Transferred Within Organizations 3/94 - 12/98 38.1 Million Allowances Reallocations 91.3% Intra-utility transfers 8.7% FIGURE 2 Allowances Transferred Between Organizations 3/94 - 12/ 24.7 Million Allowances Between Utilities Between Utilities and Brokers 13.8% 1 of 2 4/22/99 4:36 PM Trading Activity Breakdown http://www.epa.gov/acidrain/ats/qlyupd.html 66.8% Other* (betweenb rokers, indep endent investors, etc.) 15.3% Between Utilities 4.0% and Fuel Cos. *This category is prim arily composed of transfers between and am ong allow ance brokers FIGURE 3 Allowances Transferred Between Organizations, by categ 12,000,000 Util to Fuel Fuel to Util 10,000,000 Other Util to Broke Broker to Ut: 8,000,000 Inter-utility 6,000,000 4,000,000 2,000,000 0 1994 1995 1996 1997 1998 Acid Rain Program Home I Trading and Trends I Contact Us http://www.epa.gov/acidrain/ats/qlyupd.html Last Modified March 1999 2 of 2 4/22/99 4:36 PM Private Allowance Transfers Reported to ATS http://www.epa.gov/acidrain/ats/defs.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation EPA's Classification Methodology for Private Allowance Transfers Reported to ATS EPA has produced a straightforward classification methodology and applied it to all trades reported to the Allowance Tracking System through the conclusion of the most recent quarter. This represents EPA's effort to provide a general indicator of market activity to interested parties. Definition Examples (1st Category and 2nd quarter, 1996) Any transfer of allowances from one utility operating Inter-Utility company's account to a different utility operating company's account, provided LILCO General Account to the operating companies are Detroit Edison Unit Account not controlled by the same parent company. Both transferor and transferee accounts can be either general or unit accounts. Any transfer from one unit account to another unit account within the same Intra-Utility operating company. - Within one operating Within operating company- NYSEG Unit Account to company Any transfer from one different NYSEG Unit operating company's unit Account - Between operating account to another's unit companies within the same account within the same parent Between operating parent company company. companies within same parent Allowances transferred from company - Ohio Power Unit the general account of one Account to Appalachian Power Unit Account (both operating company to a unit are within AEP) account of the same operating company or parent company may also be intra-utility trades. * Any transfer from a unit or general account of one 1 of 3 4/22/99 4:36 PM Private Allowance Transfers Reported to ATS http://www.epa.gov/acidrain/ats/defs.html operating company to a general Unit to general account Reallocation account of the same operating within same company - company or parent company Wisconsin Public Service (some may become intra-utility Unit Account to Wisconsin in the future), any pooling Public Service General activity (e.g. Phase I extension Account redistributions), or any transfer in which the transferor is a Pooling - Phase I Extension partial owner of the transferee General Account to Georgia account, or vice versa. Power Unit Account Allowances transferred from Among co-owners - the general account of one Jacksonville Electric Unit operating company to a unit Account to Florida Power account of the same operating and Light General Account company or parent company (FP&L is a co-owner of may also be reallocations.* Jacksonville's Unit Account) Cantor Fitzgerald to Big Any transfer from an allowance Broker/Trader to Utility Rivers Electric Corp General broker or trader to a utility. Account Any transfer from a utility to PG&E Unit Account to Utility to Broker/Trader an allowance broker or trader. Enron Power Marketing Any transfer from a fuel Peabody Coal Company to Fuel Company to Utility supplier (e.g., coal, gas) to a PSI General Account utility. PECO Energy Company Unit Any transfer from a utility to a Utility to Fuel Company Account to Canterbury Coal fuel supplier. Co. Any transfer that does not fit into any of the above categories. This includes Emissions Trading, LLC to transfers involving AIG Trading Other environmental groups, non-utility accounts, or Paramount Petroleum to individuals. Also included are Cantor Fitzgerald broker to broker transfers, fuel supplier to broker transfers, etc. *Each of these transactions from a general account to a unit account in the same operating company or parent company is investigated to determine the prior location of the allowances traded from the general account. If the allowances originated from a different unit account of the same operating or parent company, the transaction is classified as intra-utility. If the allowances originated from the same unit account of the same operating or parent company into which they are now being transferred, the transaction is classified as a reallocation. If the allowances originated from a unit or general account of a different operating or parent company, the transaction is classified a reallocation (because the transaction into the general account would have already been classified as inter-utility, and the latter trade in question here involves only a reallocation of allowances among accounts). 2 of 3 4/22/99 4:36 PM Private Allowance Transfers Reported to ATS http://www.epa.gov/acidrain/ats/defs.html Acid Rain Program Home I Trading and Trends I Contact Us http://www.epa.gov/acidrain/ats/defs.html 3 of 3 4/22/99 4:36 PM rading Activity Summary Table http://www.epa.gov/acidrain/ats/cumchart.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Cumulative Trading Activity Table Total Private Transfer Activity Second Fourth First Quarter Third Quarter Annual Total Quarter Quarter 1994 NA 48 transactions 15 transactions 152 transactions 215 transactions 1.8 M 1.7 M 5.8 M 9.2 M NA allowances allowances allowances allowances 1995 181 transactions 118 transactions 137 transactions 177 transactions 613 transactions 10.6 M 2.2 M 0.8 M 3.1M 16.7 M allowances allowances allowances allowances allowances 1996 1,074 447 transactions 112 transactions 196 transactions 319 transactions transactions 3.4 M 0.5 M 1.1 M 3.2 M 8.2 M allowances allowances allowances allowances allowances 1997 1,429 514 transactions 263 transactions 253 transactions 399 transactions transactions 4.1 M 4.2 M 2.0 M 4.9 M 15.2 M allowances allowances allowances allowances allowances 1998 1,584 624 transactions 264 transactions 275 transactions 421 transactions transactions 4.6 M 2.8 M 3.4 M 2.7 M 13.5 M allowances allowances allowances allowances allowances Private Transfers Between Economically Distinct Organizations 1 of 2 4/22/99 4:36 PM Trading Activity Summary Table http://www.epa.gov/acidrain/ats/cumchart.htm Second Fourth First Quarter Third Quarter Annual Total Quarter Quarter 1994 NA 41 transactions 5 transactions 20 transactions 66 transactions 0.8 M 7,000 90,000 0.9 M NA allowances allowances allowances allowances 1995 55 transactions 88 transactions 90 transactions 96 transactions 329 transactions 0.5 M 0.2 M 0.4 M 0.8 M 1.9 M allowances allowances allowances allowances allowances 1996 148 transactions 98 transactions 137 transactions 195 transactions 578 transactions 0.9 M 0.5 M 0.9 M 2.1M 4.4 M allowances allowances allowances allowances allowances 1997 166 transactions 208 transactions 174 transactions 262 transactions 810 transactions 1.7M 1.9 M 1.6 M 2.7M 7.9 M allowances allowances allowances allowances allowances 1998 188 transactions 218 transactions 200 transactions 336 transactions 942 transactions 2.0 M 2.3 M 2.7 M 2.5 M 9.5 M allowances allowances allowances allowances allowances Acid Rain Program Home I Trading and Trends I Contact Us http://www.epa.gov/acidrain/ats/cumchart.html 2 of 2 4/22/99 4:36 PM Allowance Prices http://www.epa.gov/acidrain/ats/prices.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects I SO2 Program 1 NOx Program Emissions Conservation Monthly Average Price of Sulfur Dioxide Allowances Under the Acid Rain Program Cantor Fitzgerald Allowance Price in dollars Fieldston Pub lications 250 250 200 200 150 150 100 100 50 50 9/94 12/94 3/95 6/95 9/95 12/95 3/96 6/96 9/96 12/96 3/97 6/97 9/97 12/97 3/98 6/98 9/98 12/98 Month / Year Though the EPA does not officially track allowance prices, the monthly average price of a current vintage year allowance, as reported by a brokerage firm and Fieldston Publications' market survey, is recorded here for informational purposes. View table with supporting price data. Acid Rain Program Home I Trading and Trends I Contact Us http://www.epa.gov/acidrain/ats/prices.html Last Modified March 1999 1 of 1 4/22/99 4:37 PM Allowance Price Data Table http://www.epa.gov/acidrain/ats/pricetbl.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation The table below displays the supporting data for the SO2 Allowance Price Chart. Prices reported by the brokerage firms and the Fieldston Publications' market survey are rounded to the nearest whole dollar. Allowance Price in Dollars Month/Year Emissions Exchange Cantor Fitzgerald EBS Fieldston Publications 8/94 150 145 150 9/94 150 147 150 10/94 145 145 150 11/94 145 144 145 12/94 142 135 140 1/95 141 138 137 2/95 136 135 135 3/95 133 133 133 4/95 132 132 132 5/95 132 132 132 6/95 132 131 130 7/95 130 130 130 8/95 130 130 130 9/95 127 126 128 10/95 125 122 128 11/95 119 117 120 12/95 105 109 111 1/96 92 95 98 2/96 74 79 81 3/96 70 69 83 4/96 81 76 85 5/96 79 79 84 6/96 81 80 83 7/96 82 81 82 8/96 83 82 82 9/96 88 87 86 10/96 92 90 86 11/96 91 92 92 12/96 91 90 94 1 of 2 4/22/99 4:37 PM Allowance Price Data Table http://www.epa.gov/acidrain/ats/pricetbl.html 1/97 97 97 96 2/97 106 102 99 3/97 115 110 111 4/97 113 115 115 5/97 94 98 100 6/97 89 90 93 7/97 87 88 90 8/97 89 91 91 9/97 101 104 92 10/97 105 104 102 11/97 104 107 110 12/97 98 100 102 1/98 96 98 98 2/98 101 101 101 3/98 104 113 105 4/98 133 139 134 5/98 140 148 138 6/98 -- 189 193 7/98 -- 197 188 8/98 -- 189 208 9/98 -- 169 177 Acid Rain Program Home I Price Chart I Contact Us http://www.epa.gov/acidrain/ats/pricetbl.html 2 of 2 4/22/99 4:37 PM Cumulative Allowance Transfers in the Allowance Tracking System http://www.epa.gov/acidrain/ats/cumtrans.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Cumulative Allowances Transferred Under the Acid Rain Program Allowances (in Millions) 100 90 $0 70 EPA Transfers 60 50 40 30 20 Private 10 0 3/94 9/94 3/95 9/95 3/96 9/96 3/97 9/97 3/98 9/98 Quarter (cumulative) * These numbers represent only those transfers of allowances which are reported to the Allowance Tracking System. Additional transfers may be taking place in the allowance market. Private Transfers These transfers were submitted by authorized account representatives for market accounts. (EPA does not attempt to determine what constitutes an actual trade where money is exchanged). EPA / Market Transfers Most of these transfers involved movement of allowances from EPA accounts to market accounts (e.g., auctions, Phase I extension allowances, substitution allowances, etc.) Acid Rain Program Home I Trading and Trends I Contact Us http://www.epa.gov/acidrain/ats/cumtrans.html Last Modified March 1999 1 of 1 4/22/99 4:37 PM The Allowance Tracking System http://www.epa.gov/acidrain/ats/atsintro.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation The Allowance Tracking System: Accounting for SO2 Allowances Under the Acid Rain Program What is the Allowance Tracking System? What is the Role of the ATS in the Allowance Market? How is the ATS Organized? Who is Authorized to Manage a Unit or General Account? How Should AARs Report Allowance Transfers to the ATS? Reporting Allowance Transfers Electronically NEW What is the Allowance Tracking System? Functioning much like a bank, the ATS is an automated system used to track the allowances held by utilities, other affected companies, and other organizations or individuals. These allowances may be bought, sold, or transferred at any time. Specifically, the ATS tracks: The issuance of all allowances. The holdings of allowances in accounts. The holdings of allowances in various allowance reserves, such as the EPA Auction and Sale Reserve and the Conservation and Renewable Energy Reserve. The deduction of allowances for compliance purposes. The transfer of allowances between accounts. Each allowance within the ATS is identified by a 12-digit serial number consisting of 4 digits signifying the first year in which the allowance can be used for compliance and a unique 8-digit identifier. For example, an allowance that becomes eligible for use in compliance during 1995 could be numbered 1995-04875234. Only EPA can create emission allowances, and the ATS is the only official record of their creation, transfer, and use for compliance purposes. What is the role of ATS in the allowance market? The primary role of the ATS is to provide an efficient, automated means of monitoring compliance with the Acid Rain Program. The ATS also provides the allowance market with a record of who is holding allowances, the date of allowance transfers, and the allowances transferred. ATS information is available on the Internet. The ATS does not, however, record the price or other terms associated with allowance trades; such information is better collected and reported by the private sector through established exchanges or other trade information 1 of 3 4/22/99 4:38 PM The Allowance Tracking System http://www.epa.gov/acidrain/ats/atsintro.html brokers. Transfers to be executed at some future date also are not recorded in the ATS. Although submitting allowance transfers to EPA is voluntary, EPA expects most transfers to be recorded in the ATS. How is the ATS organized? The ATS contains two types of accounts: unit accounts and general accounts. As part of the implementation of the Acid Rain Program, EPA established unit accounts for all utilities governed by the Acid Rain Program. EPA uses unit accounts in determining compliance with the Acid Rain Program. General accounts are used to hold or trade allowances but are not subject to allowance deductions to cover emissions. Any individual or group, including a utility, can open a general account by submitting an Allowance Account Information form. General accounts can be used for a variety of purposes: Utilities may pool their emission allowances in general accounts. Brokers may use general accounts to hold allowances that they buy or sell for customers. Investors may use general accounts to hold allowances they have acquired for eventual resale. Public interest groups wishing to remove a portion of the available allowances from the market may purchase allowances and place them in general accounts. Who is authorized to manage a unit or general account? Whenever a unit or general account is opened, the owners or operators of the affected utility unit, or persons or companies that have an ownership interest in the allowances held in a general account, must select an authorized account representative (AAR) to act on their behalf. Designated representatives that are responsible for a utility's permitting and monitoring requirements automatically become AARs for unit accounts within the ATS. AARs and their appointed alternates are the official contact persons for EPA's Acid Rain Program. They are the only persons able to authorize the transfer of allowances, surrender allowances for compliance purposes, and change information associated with their particular unit or general accounts. AARs may represent more than one account, including both unit and general accounts. How should AARs report allowance transfers to the ATS? Allowance transfers are reported to the ATS through the submittal of an Allowance Transfer form. This form must list the serial numbers of the allowances to be transferred and include the signatures of both the transferor (the party selling or transferring the allowances) and the transferee (the party buying or receiving the allowances). EPA will record the trade within 5 business days of receiving the form and will notify both AARs of the transaction within 5 2 of 3 4/22/99 4:38 PM The Allowance Tracking System http://www.epa.gov/acidrain/ats/atsintro.html business days of recording the trade. Note: see the rule revision of December 11, 1998 regarding single signature requirement. Electronic transfers of allowances The Acid Rain Division has introduced the option of submitting transfer information via Electronic Data Interchange (EDI). Details Acid Rain Program Home I SO2 Emissions Trading I Contact Us http://www.epa.gov/acidrain/ats/atsintro.html Last Modified December 1998 3 of 3 4/22/99 4:38 PM Auctions Fact Sheet http://www.epa.gov/acidrain/auctions/auctions.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Allowance Auctions Factsheet Because the availability of allowances is crucial to ensure both the economic efficiency of the emissions limitation program and the addition of new electric-generating capacity, Title IV of the Clean Air Act Amendments mandates that EPA hold or sponsor yearly auctions of allowances for a small portion of the total allowances allocated each year. The auctions help ensure that new units with a public source of allowances beyond those allocated initially to existing units. Moreover, the auctions help to signal price information to the allowance market in the early stages of the regulatory program. Frequently Asked Questions on Allowance Auctions 1. Where Do Allowances Come From? 2. Who Administers the EPA Auctions? 3. How are the Auctions Conducted? Where Do Allowances Come From? To supply the auctions with allowances, EPA set aside a Special Allowance Reserve of approximately 2.8 percent of the total annual allowances allocated to all units. During Phase I, when the allocated allowances total 5.7 million allowances annually, 150,000 allowances are available every year for auctions. During Phase II, when allowance allocations total 8.95 million allowances annually, 250,000 allowances are earmarked annually for auctions. Private allowance holders (such as utilities or brokers) also may offer their allowances for sale at the EPA auctions, provided that the allowances are dated for the year in which they are offered, for any previous year, or for 7 years in the future. Authorized account representatives must notify the administrator of the EPA auctions of their intent to sell at least 15 business days prior to the auctions. The account representatives must specify the number of allowances they are offering and their minimum price requirements. Table: Allowances Offered at Auctions Year of Auction Spot Auction Advance Auction* 1998 150,000 125,000 1999 150,000 125,000 2000 and after 125,000 125,000 * Not useable until 7 years after purchase. 1 of 3 4/22/99 4:38 PM Auctions Fact Sheet http://www.epa.gov/acidrain/auctions/auctions.html Who Administers the EPA Auctions? The auctions are currently conducted for EPA by the Chicago Board of Trade (CBOT). This authorization is made possible by the Clean Air Act Amendments that gave EPA the authority to delegate the administration of the auctions. After an objective selection process, EPA chose CBOT to run the auctions because of its demonstrated ability in handling and processing financial instruments and using transactional information systems. Because EPA delegates to CBOT (as opposed to contracting with CBOT) to administer the auctions, CBOT is not compensated by EPA for its services nor allowed to charge fees. CBOT is not allowed to bid for allowances in the auctions nor transfer allowances in the EPA Allowance Tracking System. Only the administrative functions of the auction program have been delegated to CBOT; all other aspects of the auctions remain with EPA, as do all allowance transfer functions. How Are the Auctions Conducted? The auctions began in 1993 and are held annually, usually on the last Monday of March. Auctions are divided into two segments: (1) a spot allowance auction, in which allowances are sold that can be used in that same year for compliance purposes, and (2) an advance auction for the sale of allowances that will become usable for compliance 7 years after the transaction date, although they can be traded earlier. Bidders must send sealed offers containing information on the number and type (spot or advance) of allowances desired and the purchase price to CBOT, no later than 3 business days prior to the auctions. Each bid must also include a certified check or letter of credit for the total bid cost. (Other forms of payment may be permitted by EPA upon public notice.) The auctions sell allowances from the Special Allowance Reserve on the basis of bid price, starting with the highest priced bid and continuing until all allowances have been sold or the number of bids is exhausted. EPA may not set a minimum price for allowances from the Special Allowance Reserve. Allowances are sold from the Special Allowance Reserve before allowances offered by private holders are sold. Offered allowances are sold in ascending order, starting with the allowances for which private holders have set the lowest minimum price requirements. Offered allowances are sold until the allowance supply is depleted, bids are used up, or the minimum price for the next set of offered allowances exceeds the purchase price of the next bid. EPA returns proceeds and unsold allowances from the auctioning of reserve allowances on a pro rata basis to those units from which EPA originally withheld allowances to create the Special Allowance Reserve. Proceeds from the sale of offered allowances are returned to private allowance holders that contributed the allowances to the auctions. EPA likewise returns payment from unsuccessful bids and allowances from unsuccessful offers. 2 of 3 4/22/99 4:38 PM Auetions Fact Sheet http://www.epa.gov/acidrain/auctions/auctions.html Acid Rain Program Home I Auctions I Contact Us http://www.epa.gov/acidrain/auctions/auction.html Last Modified February 1998 3 of 3 4/22/99 4:38 PM 1999 Acid Rain Ailowance Auction http://www.epa.gov/acidrain/auctions/99result.htm EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation 1999 Acid Rain Allowance Auction Results I. Allowances Available for Auction Apot Auction 7 Year Advance Auction Origin of Allowances (First Usable in 1999) (First Usable in 2006) EPA 150,000 125,000 Privately Offered 2,510 0 Total 152,510 125,000 II. Spot Auction Results Allowances Number of Bids Number of Bidders Bid Price Bid For: 500,567 Successful: 27 Successful: 11 Highest: $230.00 Sold: 150,010 Unsuccessful: 50 Unsuccessful: 12 Clearing: $200.55 Total: 77 Total: 23 Lowest: $41.16 Average (weighted): $207.03 III. 7-Year Advance Auction Results Allowances Number of Bids Number of Bidders Bid Price Bid For: 253,055 Successful: 21 Successful: 11 Highest: $220.51 Sold: 125,000 Unsuccessful: 11 Unsuccessful: 1 Clearing: $167.55 Total: 32 Total: 12 Lowest: $110.31 Average (weighted): $179.79 Spot Auction Winners 1 of 2 4/22/99 4:39 PM 1999 Acid Rain Allowance Auction http://www.epa.gov/acidrain/auctions/99result.htm Percentage of Bidder's Name Quantity Amount Paid Total Allowances American Electric Power 74,424 49.61% $15,100,733.20 Cinergy Services, Inc. 30,000 20.00% $6,466,050.00 Cantor Fitzgerald EBS 21,250 14.17% $4,402,812.50 Potomac Electric Power Company 10,000 6.67% $2,120,100.00 Milton R. Young Station 7,467 4.98% $1,553,210.67 Wisconsin Electric Power Company 5,000 3.33% $1,016,980.00 Baltimore Gas and Electric Company 1,804 1.20% $382,448.00 Sacramento Municipal Utility District 40 0.03% $8,700.00 Acid Rain Retirement Fund 13 0.01% $2,990.00 Maryland Environmental Law Society 11 0.01% $2,436.00 K-Marx Capital 1 <0.01% $225.00 TOTAL 150,010 100% $31,056,685.37 7 Year Advance Auction Winners Bidder's Name Percentage of Quantity Amount Paid Total Allowances American Electric Power 121,945 97.56% $21,869,384.75 Cantor Fitzgerald EBS 3,000 2.40% $592,530.00 The Clean Air Conservancy 40 0.03% $8,420.40 St. Edmonds Academy, Delaware 6 <0.01% $1,278.00 Birney Elementary 2 <0.01% $441.02 Democrats of Amherst, NH 2 <0.01% $425.02 Summit Academy North 1 <0.01% $220.51 Snape/Bower Clean Air Preserve 1 <0.01% $212.51 Combined Schools Clean Air Preserve 1 <0.01% $210.51 Pepperdine Environmental Law Society 1 <0.01% $200.00 Matthew L. Floyd 1 <0.01% $185.00 TOTAL 125,000 100% $22,473,507.72 TOTAL AUCTION PROCEEDS: $53,530,193.09 Acid Rain Program Home I Auctions I Contact Us http://www.epa.gov/acidrain/auctions/99result.htm March 24, 1999 2 of 2 4/22/99 4:39 PM 1997 Allowances Available for Auction http://www.epa.gov/acidrain/auctions/98result.htm EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation 1998 EPA SO2 Allowance Auction Results I. Allowances Available for Auction Spot Auction 7 Year Advance Auction Origin of Allowances (first usable in 1998) (first usable in 2005) EPA 150,000 125,000 II. Spot Auction Results Allowances Number of Bids Number of Bidders Bid Price Bid For: 767,097 Successful: 21 Successful: 11 Highest: $228.92 Sold: 150,000 Unsuccessful: 88 Unsuccessful: 25 Clearing: $115.01 Total: 109 Total: 36 Lowest: $56.91 Average (weighted): $116.96 III. 7 Year AdvanceAuction Results Allowances Number of Bids Number of Bidders Bid Price Bid For: 509,009 Successful: 20 Successful: 3 Highest: $115.01 Sold: 125,000 Unsuccessful: 34 Unsuccessful: 9 Clearing: $108.30 Total: 54 Total: 12 Lowest: $87.00 Average (weighted): $111.05 IV. Spot Auction Winners 1 of 2 4/22/99 4:39 PM 1997 Allowances Available for Auction http://www.epa.gov/acidrain/auctions/98result.htm Percentage of Total Bidder's Name Quantity Amount Paid Allowances Cantor Fitzgerald 110,000 73.33 $12,943,200.00 Allowance Holdings Corporation 39,978 26.65 $4,597,869.78 Vermont Center 10 0.01 $1,155.00 Drake Environmental Law Society 3 <0.1 $346.53 CUNY Environmental Law Students 2 <0.1 $250.00 Association McGeorge Environmental Law 2 <0.1 $250.00 Forum Michael J. Walsh 1 <0.1 $228.92 Carolina Clean Skies 1 <0.1 $150.00 Maria Damon/ Cornell Univ. 1 <0.1 $131.51 Catholic Univ. Environmental Law 1 <0.1 $126.90 Society Univ. of Wisconsin- Madison 1 <0.1 $120.00 Total 150,000 100 $17,543,828.64 V.7 Year Advance Auction Winners Percentage of Total Bidder's Name Quantity Amount Paid Allowances Cantor Fitzgerald 124,995 99.9 $13,880,908.50 Maryland Environmental Law 4 <0.1 $446.08 Society Hatena Environmental Network 1 <0.1 $110.00 Japan Total 125,000 100 $13,881,464.58 Total Auction Proceeds: $31,425,293.22 Acid Rain Program Home I Auctions I Contact Us http://www.epa.gov/acidrain/auctions/98result.htm March 25, 1998 2 of 2 4/22/99 4:39 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Opt-in Program Fact Sheet The Opt-in Program expands EPA's Acid Rain Program to include additional sulfur dioxide (SO2) emitting sources. Recognizing that there are additional emission reduction opportunities i the industrial sector, Congress established the Opt-in Program under section 410 of the Clean A Act Amendments of 1990. The Opt-in Program allows sources not required to participate in the Acid Rain Program the opportunity to enter the program on a voluntary basis, reduce their SO2 emissions, and receive their own acid rain allowances. The participation of these additional sources will reduce the cost of achieving the 10 million ton reduction in SO2 emissions mandated under the Clean Air Act. As participating sources reduce their SO2 emissions at a relatively low cost, their reductions -- in the form of allowances -- can be transferred to electric utilities where emission reductions are more expensive. The Opt-in Program offers a combustion source a financial incentive to voluntarily reduce its SO2 emissions. By reducing emissions below its allowance allocation, an opt-in source will crea unused allowances, which it can sell in the SO2 allowance market. Opting in will be profitable if the revenue from allowances exceeds the combined cost of the emissions reduction and the cost of participating in the Opt-in Program. An opt-in source must comply with the same or similar provisions as utility units affected under the mandatory Acid Rain Program. These provisions relate to allowance trading, permitting, excess emissions, monitoring, end-of-year compliance and enforcement. Most basic to the program is the requirement that each year the opt-in source must hold enough allowances to cover its annual SO2 emissions. Frequently Asked Questions about the Opt-in Program 1. Who Can Opt-in? 2. How Does a Source Opt-in? 3. How Is the Number of Opt-in Allowances Calculated? 4. Are There Restrictions on Opt-in Allowances? 5. Why is EPA Concerned with Reductions in Utilitization? 6. Questions About the Thermal Energy Exception What is the Thermal Energy Exception? What is a Thermal Energy Plan? How are Allowances Transferred Under a Thermal Energy Plan? Must a Thermal Energy Plan Be Renewed? May a Retiring Source Participate Under the Thermal Energy Exception? 7. May a Source Withdraw From the Opt-in Program? 1 of 7 4/22/99 4:40 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html Who Can Opt-in? All operating stationary combustion sources that emit SO2 but are not otherwise required to me the mandatory SO2 emissions limitations of Title IV are eligible to opt into the Acid Rain Program. Combustion sources are defined as fossil fuel-fired boilers, turbines or internal combustion engines. Examples of Eligible Combustion Sources Existing utility units serving generators less than 25 megawatts Simple combustion turbines built before November, 1990 Industrial boilers Non-affected municipal waste combustors burning some amount of fossil fuel Examples of Ineligible Sources Utility units affected under 40 CFR part 72.6 New units exempt under 40 CFR part 72.7 Retired units exempt under 40 CFR part 72.8 Mobile sources The Acid Rain Division has published a guidance document to help sources determine whether not they are affected by the mandatory utility program (see "Do the Acid Rain SO2 Regulations Apply to You?" EPA 430-R-94-002). How Does A Source Opt-in? In order to enter the Opt-in Program, a combustion source must submit an opt-in permit application and monitoring plan to its permitting authority, receive an opt-in permit, and install and certify its emission monitors. After receipt of a complete opt-in permit application and after the monitoring plan is determined to be sufficient, the permitting authority will issue a draft opt-in permit for the source to review. The draft permit is then made available for public comment and eventually issued or denied with 12 or 18 months of receipt of a complete application 12 months if the EPA is the permitting authority or 18 months if the State is the permitting authority. The opt-in source must renew the opt-in permit before it expires, and, in most cases, the permit will last for 5 years. The opt-in permit application must clearly identify the combustion source and contain all of the information needed to calculate the combustion source's allowance allocation, including its fuel input and emissions data as well as historic and current emission limits. The certification of monitoring systems for the combustion source will follow the same procedures and requirements as for affected units in the mandatory utility program (see 40 CFR part 75 for more specific requirements). However, there will be no provisional nor automatic certification of monitoring systems for combustion sources. In addition, an approved opt-in 2 of 7 4/22/99 4:40 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html permit for an applying combustion source will expire if the combustion source fails to complete the certification of its monitoring systems within 180 days after the permit's approval. EPA expects combustion sources to pursue the resolution of the permitting process towards a final opt-in permit and the certification of the combustion source's monitors simultaneously. The renewal for an opt-in permit must be requested at least 6 months prior to the expiration of an existing opt-in permit, if EPA is the permitting authority, and at least 18 months prior to the expiration of the existing opt-in permit, if the State is the permitting authority. The renewal, however, should be much more straightforward than the initial opt-in permit application, becaus the allowance allocation, once established, cannot be altered in the renewal. How is the Number of Opt-in Allowances Calculated? Opt-in allowances are allocated upon entry into the Opt-in Program. The number of allowances an opt-in source receives is based on the product of its average heat input for all fuel consumed during 1985 - 1987 (known as its "baseline") and the lesser of three emissions rates: its 1985 actual emissions rate, its 1985 allowable emissions rate, and its allowable emissions rate at the time the combustion source submits an opt-in permit application. If the source began operating after 1985, then EPA will accept an "alternative baseline" which is the average heat input for all fuel consumed during the first three consecutive calendar years for which the combustion source operated after December 31, 1985. The emissions rates used for t allowance calculation will be the actual and allowable emissions rate for the first year of this thr year period as well as the combustion source's current allowable SO2 emissions rate. ANNUAL THE 955 ACTUAL 502 EMISSIONS RATE OPT-IN = [EASELINE X THE LESSER OF THE 1985 ALLOWABLE 502 EMISSIONS RATE THE UPPENT ALLCA&BLE 502 EMISSIONS PATE ALLOWANCES Are There Restrictions on Opt-in Allowances? There are restrictions on opt-in allowances that do not apply to allowances allocated to affected units under the mandatory utility program. They are: 1. Only opt-in allowances dated for the current or previous years can be transferred to other accounts in the Allowance Tracking System. 2. Only opt-in allowances for past years may be offered for sale in the spot auction, and no opt-in allowances may be offered for sale in the advance auction. 3. Opt-in allowances must be surrendered to the EPA if the opt-in source experiences reduc utilization relative to its baseline. The number of allowances surrendered will be proportional to the reduction in the opt-in source's historic operations (i.e., its baseline). 3 of 7 4/22/99 4:40 PM Opt-in Prograin Fact Sheet http://www.epa.gov/acidrain/optin/optin.html The only exception to the surrender of allowances for reduced utilization is known as the thermal energy exception. Why is EPA Concerned with Reductions in Utilization? Because participation in the opt-in program is voluntary and does not include all sources, the statute recognized that shifts in utilization from opt-in sources to sources not participating in the program would free up allowances without reducing emissions. EPA, in accordance with sectio 410(f)of the Act, will deduct allowances from an opt-in source when it shuts down or reduces utilization. For most years, utilization is calculated as the three year rolling average of heat input for the opt-in source. An opt-in source is considered to have reduced its utilization if its average utilization is below its baseline (i.e., its average historic fuel input). If an opt-in source has reduced its utilization, EPA will deduct allowances from the opt-in source as described in the equation below. EPA will consider documented claims of demand side efficiency improvements as well as improvements in the efficiency of electricity or steam production at the opt-in source when determining reduced utilization. EPA will give credit and not deduct allowances for any portion of the opt-in source's reduced utilization provided by efficiency improvements. ALLOWANCES DEDUCTED AVERAGE UTILIZATION II ANNUAL OFT-IN ALLOWANCES X FOR REDUCED EASELINE UTILIZATION Questions About the Thermal Energy Exception 1. 2. What is the Thermal Energy Exception? Title IV allows for the transfer of allowances tha otherwise must be deducted to account for reduced utilization or shutdown, but the transfer may only be made to a unit that is replacing the thermal energy previously supplie by the opt-in source. Thermal energy is the thermal output or steam produced by a combustion source that is used directly as part of a manufacturing process, but not used t produce electricity. A replacement unit must be an affected unit under the Acid Rain Program and prove that it actually replaces the opt-in source's thermal energy. 3. What is a Thermal Energy Plan? In order for an opt-in source to transfer allowances to a replacement unit under the thermal energy exception, the opt-in source must submit a 4 of 7 4/22/99 4:40 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html thermal energy plan and amend its opt-in permit accordingly. A thermal energy plan is entered into jointly by the opt-in source and its replacement units The plan estimates the amount of thermal energy replaced and requires additional information on the operations of both the opt-in source and its replacement units while governed by the plan. A thermal energy plan has a fixed duration and must be renewed. The term of the plan extends over full calendar years, so interested opt-in sources and replacement units could have a plan last as many as five years, if the opt-in permit is on a calendar year basis, or three years, if the opt-in permit begins on a date other than January 1. 4. How Are Allowances Transferred Under a Thermal Energy Plan? The number of allowances transferred from the opt-in source to the replacement unit is calculated using the following equation: The allowance transfer from the opt-in source to the replacement unit will take place annually. Such transfer will follow EPA's deduction of allowances to offset emissions for the previous compliance year and before any other allowance transfers are recorded in the Allowance Tracking System for that year. Since the number of allowances that will be transferred each year is contained in the approved thermal energy plan, the opt-in source and each replacement unit will know exactly how many allowances will be transferred. At the end of each year, EPA will adjust this allowance transfer to reflect the actual therm energy replaced. In the event that the amount of replacement is altered by an opt-in source's confirmation report, a report submitted after allowance reconciliation that documents efficiency improvements, the EPA will make adjustments in the following manner: EPA will consider the number of allowances transferred to replacement units fixe after the reconciliation process has ended. The Agency will rely on the opt-in source to surrender any additional allowances needed to make the accounting consistent with both the confirmed efficiency estimates and the number of opt-in allowances available for transfer. When a thermal energy plan is terminated, the allowance transfer for the current year would be reversed, essentially restoring the opt-in source's original allocation for the current year. Future year opt-in source allowances would remain unchanged. THERMAL ENERGY X ALLOWABLE EMISSIONS RATE AT REPLACEMENT UNI ALLOWANCES EFFICIENCY CONSTANT TRANSFERRED 2000 5. Must A Thermal Energy Plan Be Renewed? A thermal energy plan will, in most circumstances, be renewed along with the renewal of the opt-in permit. At other times, th renewal of the plan will be considered a revision of the opt-in permit and will follow the procedures established for all Acid Rain permit revisions. One important aspect of renewing a thermal energy plan is the possible change in the allowable emissions rates at 5 of 7 4/22/99 4:40 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html the replacement units. The Agency will rely on the current allowable rate in effect at the time of renewal for the allowance transfer under the thermal energy plan. 6. May a Retiring Source Participate Under the Thermal Energy Exception? A retiring source may enter the Opt-in Program and participate in a thermal energy plan, but must b operating at the time it submits its opt-in permit application. Operating is defined for this purpose as having documented fuel consumption for more than 876 hours in the 6 months immediately preceding application (i.e., operating at 20 percent capacity factor or greater) EPA places this requirement on retiring combustion sources, to prevent sources shut dow long ago from claiming allowances and increasing overall emissions. Retiring combustion sources seeking to become opt-in sources can receive an exemption from the Opt-in Program's monitoring requirements, consistent with 40 CFR part 75.67. The designated representative of such a source must petition the Administrator for such a exemption. May a Source Withdraw From the Opt-In Program? An opt-in source can withdraw from the program provided it meets certain conditions: 1. The opt-in source must submit its annual compliance certification report by January 30 of the first calendar year in which the withdrawal is to be effective (rather than March 1); 2. The opt-in source must immediately provide additional allowances if it has excess emissions; 3. The opt-in source must surrender all allowances allocated to it for the year in which the withdrawal is to take effect and for all years thereafter. If the opt-in source does not meet these conditions to withdraw, the opt-in source shall remain i the Opt-in Program and remain subject to all requirements of the program. Withdrawal will take effect on January 1. For opt-in sources that withdraw from the program, they cannot reapply to opt in until the year before the original opt-in permit needed to be renewed. Additional Details A monitoring plan is sufficient if the plan appears to contain information demonstrating that all emissions are monitored and reported in accordance with 40 CFR part 75. A determination of sufficiency shall not be construed as the approval or disapproval of the combustion source's monitoring systems. The term "permitting authority" is used to designate the entity responsible for issuing and administering permits. Initially, the EPA is the permitting authority. As states, and in som 6 of 7 4/22/99 4:40 PM Opt-in Program Fact Sheet http://www.epa.gov/acidrain/optin/optin.html cases, local air pollution control agencies are authorized to issue permits under Titles IV and V, they would become the permitting authority. The 1985 allowable emissions rate is the most stringent federally enforceable limitation f SO2 (in lb/mmBtu) applicable to the combustion source for 1985. The allowable emissions rate at the time of application if it exists, is a new lower emissions limit that is finalized but not yet in effect for the applying opt-in source. EPA W consider this new limit (known as "the current promulgated SO2 emissions rate") and will adjust the combustion source's allowance allocation for the year and all years after such limit takes effect. The Allowance Tracking System (ATS) is an automated system operated by EPA's Acid Rain Division and used to track the allowances held by utilities, opt-in sources and other organizations and individuals. More on the Allowance Tracking System. EPA holds annual auctions from a special allowance reserve and from offers of allowanc from private holders. There are two types of auctions: (1) a spot allowance auction, in which allowances are sold that can be used in that same year for compliance purposes, an (2) an advance auction for the sale of allowances that will become usable in the future. More on Auctions. Demand side efficiency improvements include demand side measures that improve the efficiency of electricity or steam consumption. Qualified demand side measures applicable to the calculation of utilization for opt-in sources are listed in Appendix A, Section 1 of 4 CFR part 73. More on energy efficiency improvements under the Clean Air Act. Acid Rain Program Home I SO2 Emissions Trading Contact Us http://www.epa.gov/acidrain/optin/optin.html 7 of 7 4/22/99 4:40 PM Early Reduction Credits http://www.epa.gov/acidrain/erc.html EPA United States Environmental Protection Agency Acid Rain Program Overview Effects SO2 Program NOx Program Emissions Conservation Early Reduction Credits Units may receive additional allowances under Section 404(e) for SO2 emissions reductions accomplished in years prior to the implementation of the Acid Rain Title. Only a few Phase I and Phase II utility systems are eligible for these early reduction credits. Eligible Phase I units receive allowances for voluntary emissions reductions made after enactment and before January 1, 1995. Eligible Phase II units receive allowances for voluntary SO2 emissions reductions made between January 1, 1995 and January 1, 2000. Acid Rain Program Home I SO2 Emissions Trading I Contact Us http://www.epa.gov/acidrain/erc.html 1 of 1 4/22/99 4:40 PM Cost Savings, Market Performance, and Economic Benefits of the U.S. Acid Rain Program Dallas Burtraw Discussion Paper 98-28-REV Revised September 1998 RESOURCES FOR THE FUTURE 1616 P Street, NW Washington, DC 20036 Telephone 202-328-5000 Fax 202-939-3460 © 1998 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the author. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not undergone formal peer review or the editorial treatment accorded RFF books and other publications. Cost Savings, Market Performance, and Economic Benefits of the U.S. Acid Rain Program Dallas Burtraw Abstract This paper reports on four areas of research concerning Title IV of the 1990 Clean Air Act Amendments that regulates emissions of SO₂ from electricity generation. The first is the costs of the program over the long-run as estimated from the current perspective taking into account recent changes in fuel markets and technology. We compare projected costs with potential cost savings that can be attributable to formal trading of emission allowances. The second area is an evaluation of how well allowance trading has worked to date. The third area is the relationship between compliance costs and economic costs from a general equilibrium perspective. The fourth area is a comparison of benefits and costs for the program. Key Words: acid rain, benefit-cost analysis, air pollution, permit trading, Clean Air Act JEL Classification Numbers: H43, Q2, Q4 ii Table of Contents Introduction 1 Costs of the SO₂ Program and Cost Savings from Allowance Trading 2 Performance of the Program to Date 7 Compliance Costs versus Economic Costs 10 Comparing Benefits and Costs 13 Conclusion 15 References 17 List of Figures and Tables Figure 1. Annual 90 Percent Confidence Intervals for Total Health Benefits Compared with Expected Annualized Costs 15 Table 1. Estimates of Long-run (2010) Annual and Marginal Cost 3 Table 2. The Contribution of Price and Technological Change to Compliance Costs 5 Table 3. Cost Estimates for Compliance in 1995 8 Table 4. Benefits and Costs Expressed as "Per Affected Capita" in 2010 13 iii C Dallas Burtraw* Title IV of the 1990 US Clean Air Act Amendments regulates emissions of SO₂ from electric utility facilities and instituted two important innovations in US environmental policy. The more widely acknowledged of these is the SO₂ emission trading program, which is designed to encourage the electricity industry to minimize the cost of reducing emissions. The industry is allocated a fixed number of total allowances and firms are required to hold one allowance for each ton of sulfur dioxide they emit. 1 Firms are allowed to transfer allowances among facilities or to other firms, or to bank them for use in future years. This approach enables firms operating at high marginal pollution abatement costs to purchase SO₂ emission allowances from firms operating at low marginal abatement costs, thereby lowering the cost of compliance. The second and less widely acknowledged innovation is the annual cap on average aggregate emissions by electric utilities, set at about one-half of the amount emitted in 1980. The cap accommodates an allowance bank, so that in any one year aggregate industry emissions must be less than the number of allowances allocated for that year plus the surplus that has accrued from previous years. Unlike most previous regulations in the US, including technology standards or emission rate standards, the emissions cap represents a guarantee that emissions will not increase with economic growth. 2 This paper begins with a summary of recent projections of the costs of compliance when the program is fully implemented in the next decade, and estimates of stemming from the trading program in the long run. Second, I evaluate how well the allowance trading has worked to date, and what one can expect to happen in the future. * Fellow, Quality of the Environment Division, Resources for the Future. An earlier version of this paper was presented April 9, 1997 at the Institute of Development Studies, University of Sussex. The paper will appear in , S. Sorrell and J. Skea, eds., forthcoming (1998) from Edward Elgar Publishing. 1 Allowances are allocated to individual facilities in proportion to fuel consumption multiplied by an emission factor during the 1985-1987 period. About 2.8 percent of the annual allowance allocations are withheld by the EPA and distributed to buyers through an annual auction run by the Chicago Board of Trade. The revenues are returned to the utilities that were the original owners of the allowances. 2 Title IV also used a more traditional approach in setting NOₓ emission rate limitations for coal-fired electric utility units, although this approach has been modified to allow emission rate averaging among commonly owned and operated facilities. Hence, there is no cap on NOₓ emissions, but Title IV is expected to result in a 27 percent reduction from 1990 emissions for electric utilities. 1 Burtraw RFF 98-28-REV Most attention has been focused on the reduction in "out-of-pocket" compliance costs that may result from allowance trading. However, compliance costs are not the same thing as economic costs. A third focus of this paper is to explain how hidden economic costs inflate the total cost of the program in an important way, and how recognition of these costs provides an important lesson for the design of future programs. Fourth, I return to the primary purpose of the program, which after all is to reduce acidification of the environment and associated effects on human well-being. An economic assessment of benefits, though extremely uncertain, indicates that benefits are an order of magnitude greater than costs under the program. An important component of this favorable assessment rests with the design of the tradable permit program which has helped to reduce costs substantially. To make sense of the wide variety of claims about the costs of the SO₂ program one has to put these claims in historic perspective. In the 1980s there were over seventy proposed pieces of legislation that suggested a variety of regulatory approaches aimed at the problem of acidification. One prominent proposal was the Sikorski/Waxman bill in 1983 that sought to rollback emissions, by about the same amount as eventually required under Title IV, by requiring the installation of scrubbers (flue gas desulfurization equipment) at the fifty dirtiest plants. 3 The estimated levelized cost of this proposal ranged from about $7.9 billion per year according to government studies (OTA, 1983) to $11.5 billion per year according to an industry study (TBS, 1983; 1995 dollars). Another bill (H.R. 4567) in 1986 was aimed at similar environmental gains but promoted cost reductions by applying a target average emission rate for each utility company. Taking account of changes in fuel and other input prices between 1983 and 1986, an industry study (TBS, 1986) found costs would be $7.5 billion per year compared with estimates of $3.5 to $6.2 billion by ICF in a study for the EPA, and $3.4 to $4.3 billion by the OTA (1995 dollars). 4 Though ultimate regulation did not involve either of these approaches the estimates provide an indication of what the program might have cost under alternative regulatory approaches, and have often been used as a basis for comparison. One of the earliest studies of the cost under an allowance trading system was Elman et al. (1990) who estimated the marginal cost of compliance and inferred this would be the 3 These plants represented 89 percent of the nation's pre-New Source Performance Standard coal-fired capacity. Fuel switching to low-sulfur fuel and other facility improvements would have been required at other facilities. Scrubbing would have been applied to about half of the affected capacity and accounted for 70 percent of the SO₂ reduction. 4 It is noteworthy that an industry study (TBS, 1986) suggested the costs would be higher as a result because although "intra-firm trading" would reduce the need for scrubbing, it would increase reliance on low sulfur coals resulting in an increased premium on low sulfur coal that would raise costs for units already using low sulfur coal. This prediction is contradicted by the turn of events under Title IV, when the cost of lost sulfur coal fell with its expanded use. 2 Burtraw RFF 98-28-REV value of emission allowances. Under a perfect trading market, this study predicted marginal costs (presumed to equal allowance prices under perfect trading) of $742-1032 (1995 dollars) and costs of $1935 for imperfect trading, ranging up to $2580 or $5160 at a number of utilities. However, for evaluation of the program compared to prior expectations, the most useful study is ICF (1990), which was done for the EPA and available prior to enactment of the legislation. This study captured more accurately the ultimate design of the regulation, and projected marginal costs of $579-760 (1995 dollars) for full compliance under the program. This and a number of other studies are summarized in Table 1.5 (billion 1995 dollars) (1995 dollars) (1995 dollars) 1.0 291 174 0.9 239 [EPRI] 436 [EPA] 2.3 532 252 [EPRI] 1.4-2.9 543 286-334 2.2-3.3 230-374 [EPRI] 2.4-3.3 520 314-405 [EPA] 1.9-5.5 579-760 280-467 An important feature of the studies summarized in Table 1 is that, as a group, they have successively estimated a sequence of declining projections of annual and marginal costs of compliance. There are several contributing reasons for this. One is that the trading program ignited a search for ways to reduce emissions at less cost, as theory suggests is likely to occur with this type of regulation, and the fruitful results of this enterprise are measured by latter studies (Burtraw, 1996). It is also the case that advantageous trends in fuel markets contributed to a decline in emission rates, making it easier for utilities to attain the goals of the program thereby reducing program costs (Ellerman and Montero, 1998; Burtraw, 1996). Indeed, the right-hand column in Table 1 reporting average cost per ton estimated by the various studies reflects this decline. The differences in average cost estimates are due to differences in annual cost reported in the 5 These estimates describe long-run costs expected to obtain when the allowance bank, which is expected to build up to about 11 million tons by the end of Phase 1 (in the year 2000), is drawn down and net contributions to the bank are zero. 3 Burtraw RFF 98-28-REV first column, but also they are due to differences in the estimate of tons reduced under the program compared to various specified baseline projections of emissions that would occur in the absence of the program. To the extent these emission reductions would have occurred even in the absence of the SO₂ program then the program should not be given credit; however, it appears the trading program encouraged utilities to capitalize on these advantageous trends while other regulatory approaches might not have. A third reason for declining projections of cost is that the market structure for industries offering compliance services to utilities was dramatically changed under the trading program. What were previously independent factor markets supplying services to utilities (coal mining, rail transport, and scrubber manufacturing) were thrown into a competition with each other by the program's flexible implementation. This unleashed competitive pressure to find ways to reduce costs in all these markets. All of the explanations listed appear to contribute to some degree to the decline in estimates of compliance costs. The Carlson et al. (1998) model can help to sort out these factors because it provides a framework in which we can vary factors one at a time and explore their significance. While most of the studies in Table 1 rely on engineering-based models of compliance options and their costs, Carlson et al. uses a simulation model based on marginal abatement cost functions derived from an econometrically estimated long-run total cost function for electricity generation for a sample of over 800 generating units over the period 1985-1994.⁶ From an economist's perspective, this study is superior because it takes into account behavioral responses to changes in relative input prices. These behavioral responses generally take the form of substitution among inputs to reduce jointly the costs of generating electricity and of complying with emission reduction requirements. The econometric approach also affords a method for measuring the role of technological change in reducing the costs of SO₂ abatement over time, and to develop forecasts of future compliance costs and gains from trade that implicitly incorporate future behavioral changes including future responses to changes in technology. Table 2 presents several estimates using this model, varying assumptions about fuel prices and technological change. The columns in the table represent the annual cost of a command and control approach (uniform emission rate standard), the annual cost of efficient trading, its associated marginal abatement cost and finally the estimated gains from trade that are available from efficient trading. 6 The cost function they estimate treats fuel type (high-sulfur and low-sulfur coal), labor and generating capital as fully variable inputs. The econometric model consists of the cost function plus two share equations that specify the share of total costs attributed to capital and labor, and an equation for the firm's mean annual emission rate. The study uses a translog form for the cost function, adding dummy variables for each plant in the database to measure fixed effects that vary among the plants. Costs for units with scrubbers are taken directly from reported data. The model does not investigate whether early commitments to build scrubbers was economical, but several studies have suggested that several of these investments were not. 4 Burtraw RFF 98-28-REV (million 1995 (million 1995 (1995 (million 1995 dollars) dollars) dollars) dollars) The first row in Table 2 reports numbers for a benchmark scenario that uses assumptions comparable to those in previous studies, including assumptions that relative fuel prices remain stable at 1995 levels, technology including the utilization rate of scrubbers is characterized at 1995 levels and the historic method for measuring emissions based on sampling of coals and engineering formulas remains in effect. The retirement rates for coal facilities, and replacement with scrubbed coal technology, is taken from projections by the Energy Information Administration (EIA, 1996). The estimates assume that no additional retrofit scrubbers are constructed after Phase I. This benchmark predicts long-run marginal abatement costs will be $436 (1995 dollars). The second row presents an estimate under most benchmark assumptions but with prices held to their 1989 level (implying a higher price for low sulfur coal relative to high sulfur coal than obtained in 1995) and the time trend for technological change (factor productivity) also held at 1989 levels.⁷ From this vantage point, marginal abatement costs rise to $560, or 28 percent greater. One should note that this is not far from that offered by ICF (1990) calculated with comparable information reported in Table 1. The last row in Table 2 presents the Carlson et al. (1998) preferred estimate, reproduced for convenience from Table 1. Compared to the benchmark, this scenario adopts 1995 prices and 2010 technology. It assumes utilization rates and performance of in-place scrubbers continues to improve. It assumes a slower retirement rate of coal-fired facilities, and half of retired facilities are replaced by gas. Also it assumes that the use of continuous emission monitoring systems in place of the historic measure of emissions will raise emission 7 Technological change here captures both exogenous efficiency improvements at the power plant and improvements induced by the program, but it does not capture improvements in scrubber technology and performance. 5 Burtraw RFF 98-28-REV estimates and necessitate a greater level of control. 8 Explicitly missing from consideration in all of these studies is the influence of other potential regulatory actions, such as further control on particulates or NOₓ emissions or actions to meet global warming goals. Based on this information and model, the savings from allowance trading that are expected from an efficient allowance market are about $780 million annually, or about 42 percent of the costs under command and control. Previous studies (Van Horn Consulting et al., 1993; GAO, 1994) suggested an efficient allowance market would result in cost savings of about twice this much in percentage terms. In all these studies, the command and control approach that is modeled is "enlightened" in that it is a performance standard (emission rate or emission tonnage standard) applied to each facility, calibrated to achieve the same level of total emissions.⁹ This approach already encompasses many of the beneficial incentives of the SO₂ trading program compared to a technology forcing approach by providing individual facilities with flexibility in achieving the standard. Other command and control approaches that were seriously considered in the US, such as forced scrubbing at larger facilities, could have cost substantially more. Forced scrubbing is also the approach embodied in the New Source Performance Standards for SO₂ from power plants. 10 In summary, estimates of both a command and control approach and allowance trading have fallen over time due to a number of factors. Allowance trading is expected to result in cost savings relative to a command and control approach, but the absolute magnitude of these savings is expected to be somewhat less than previously envisioned due to changes in fuel markets. However, a more rigid approach that forced firms to adopt specified technologies would have precluded them from taking advantage of some of these trends in the industry. Compared to this, allowance trading could be argued to constitute a greater savings than we estimate. Though they are substantial by any accounting, cost savings have been exaggerated in many accounts of the program. Advocates of ambitious climate change policies have suggested that SO₂ allowance prices are "so low" and that economists and engineers got it "so wrong" that policy makers should virtually ignore cost projections when developing new regulations such as a carbon permit trading program (or trading of nitrogen oxide (NOx) permits). These exaggerated comparisons have often used inconsistent estimates of cost, 8 In 1995 the continuous emission monitors estimated seven percent higher emissions than did the historic approach on average, although there was considerable variability among facilities. 9 In their command and control scenario, Carlson et al. (1998) apply a uniform emission rate standard to all facilities. GAO (1994) and Van Horn Consulting et al. (1993) allow intra-utility trading, but no trading between utilities. GAO (1994) also models a scenario that requires each facility to achieve its SO₂ allowance allocation without trading, and finds cost savings more than double that in the case in which internal-trading is allowed in the command and control baseline. 10 1978 Clean Air Act regulation of sulfur emissions from newly constructed fossil fuel-fired electricity generating facilities imposes a rate-based standard that requires a 90 percent reduction in a smokestack's SO₂ emissions, or 70 percent if the facility used low-sulfur coal. Although nominally a performance standard, it effectively dictates technological choices and precludes compliance through the use of process changes or demand reduction. The only available technology to achieve such reductions is scrubbing, and the use of low- sulfur coal is not a permissible way to avoid the threshold for the strict standard. 6 Burtraw RFF 98-28-REV leading to erroneous conclusions (Bohi and Burtraw, 1997). The most obvious of these errors is to compare allowance prices today directly with long-run marginal cost estimates. This is erroneous because long-run marginal cost estimates describe costs that will not be incurred until late in the next decade, when the allowance bank is drawn down and net contributions to the bank are zero. Allowance prices today should be proximate to short-run marginal cost and these are related to long-run marginal costs by the rate of interest. Discounting long-run marginal costs for the year 2010 to the present reduces them by about two-thirds! A proper comparison of program costs indicates they are expected to be less by almost one-half that indicated in information available to legislators in 1990. Further, program costs are perhaps half again the cost of an inflexible technology standard. However, they are not reduced by fifteen fold, as some have claimed. 11 Furthermore, as described previously, a significant portion of the actual decline is due to factors exogenous to the program. Nonetheless, after accounting for these factors the remaining decline in costs is impressive, and in general it points toward success for the program. The previous section reported on estimates of the cost of compliance when the program is fully implemented. The second phase of the program begins in the year 2000, when an expanded set of facilities will be brought into the program and the average emission rate for all facilities will reduce to about 1.2 lb. per million Btu. In the first phase of the program, which began in 1995, firms have been over-complying in order to save allowances in the bank and to ease the transition into Phase II. In the second phase, utilities will begin to draw down their allowance banks until net contributions to the bank are zero and annual emissions equal annual allocations for an average year. This is projected to occur by the year 2010. A specific economic relationship is expected to govern use of banked allowances, and the magnitude of the marginal costs in a given year. As previously mentioned, allowance prices in a given year should reflect marginal abatement costs in that year. To appreciate why this is so, imagine instead that allowance prices were less than a firm's marginal costs. Then the firm could decrease its compliance activities and purchase allowances in the market as an alternative means of compliance, earning positive net revenues. A similar reasoning governs the relationship between marginal costs at different points in time. Marginal costs in any given year should equal the present discounted value of marginal abatement costs in the future. Imagine instead that allowance prices were less than the present discounted value of marginal costs in the future. Then again the firm could purchase allowances in the current year and "bank them" for use in place of compliance 11 For instance, March 10, 1997 EPA Administrator Carol Browner argued: During the 1990 debate on the acid rain program, industry initially projected the cost of an emission allowance to be $1500 per ton of sulfur dioxide. Today, those allowances are selling for less than $100." ("New Initiatives in Environmental Protection," (newsletter), March 31, 1997, Commonwealth Club of California. See also "Economists" Cold Forecast; Assumptions: Expect their dire predictions about the impact of the global warming treaty on the United States. Ignore all of them," by Elaine Karmarck, , December 28, 1997. 7 Burtraw RFF 98-28-REV activities in the future, generating net revenues. Of course, the converse would hold as well, if allowance prices were greater than these respective measures of marginal costs. Hence, the banking provision in the allowance market serves an intertemporal arbitrage function that allows firms to identify the least cost means of compliance over time just as allowance trading allows firms to identify the least cost means of compliance over geography. Using this relationship from economic theory provides one way to check on the performance of the market and the likely accuracy of estimated costs. If cost estimates for the long-run are correct, and using a discount rate of 8 percent reflecting the opportunity cost of capital for firms in the industry, current allowance prices should be about one-third of projected long-run costs. The present discounted value of long-run marginal costs of $291 in 2010 should be about $95 in 1997. Allowance prices hovered between $100 and $110 for most of 1997, suggesting that the model is roughly consistent with current activities, and that intertemporal arbitrage is working to an important degree. The explicit omission of other potential regulatory actions (NOx, CO₂) in the models may explain much of the remaining difference. Another way to measure the performance of the market to date is to look at the allocation of compliance activities among firms in the current period. Economic theory suggests that the marginal cost of compliance activities should be the same at all facilities (except as may be constrained by local ambient air quality restrictions). To investigate this Carlson et al. (1998) evaluated their estimated marginal abatement cost functions at the level of emissions in the industry in 1995. These results are reported in Table 3. The cost of efficient trading is projected to be $552 million, with a marginal cost for fuel switching activities of $101. This compares with a cost under a command and control emission rate standard of $802 million, representing a savings of 30 percent. 12 Further, the projected marginal cost of $101 is remarkably close to allowance prices in this period (around $90 in 1995), reinforcing the notion that intertemporal arbitrage is working as it should. (million 1995 dollars) (1995 dollars) (1995 dollars) 552 101 194 832 180 291 726 153 210 * Includes scrubbing costs 12 As noted previously, in their command and control scenario, Carlson et al. (1998) apply a uniform emission rate standard to all facilities. 8 Burtraw RFF 98-28-REV However, the second row of the table reports the estimated actual cost of compliance when the estimated abatement cost functions are evaluated using observed emissions at individual facilities. The cost to the industry in 1995 is estimated to be $832 million, or 50 percent more expensive than a least cost solution (reported in row 1) according to the model. The only other estimate for 1995 that we are aware of is Ellerman et al. (1997) who found costs in 1995 to be $726 million. The notion of an industry marginal cost does not apply in this context, since costs presumably differed among firms. However, Carlson et al. calculate the marginal cost at each facility weighted by that facilities portion of total generation and summed for the industry results in an estimate of $180. Ellerman et al. do not report marginal cost but they find an average cost from fuel switching activities of $153. Both of these figures compare poorly with observed allowance prices in 1995. We consider the proximity of the Carlson et al. econometric estimates to the Ellerman et al. survey estimates of actual costs, and the distinction between these estimates and the estimated cost of efficient compliance as evidence that in the first year of the program there were unrealized potential gains from trade suggesting that important imperfections characterized the allowance market in the first year of the program. Moreover, the difference between the observed price of allowances (approximately $90 in 1995) and both Ellerman et al. and Carlson et al. estimates of actual compliance cost provides further evidence that the new program of allowance trading was not yet a mature institution in 1995. What appears to have occurred in 1995 is that different patterns of compliance behavior co-existed in the industry. Many utilities appear to have taken advantage of the flexibility afforded by the allowance program to find ways to reduce costs of compliance, including taking advantage of allowance trading per se However, many other utilities appear to have pursued a solitary strategy, rationalizing to some degree the cost of emission reductions within the firm, but not taking advantage of the allowance market to rationalize costs with other firms in the industry (Bohi and Burtraw, 1997). This glance at the first year of performance in the market is disconcerting, but it may be a poor indicator of the rich long-run prospects for the program. One should expect that an industry that has heretofore been subject to cost recovery rules in a regulated setting would take some time in adjusting to a new incentive-based approach to environmental regulation. In fact, one can be relatively confident that the future holds better things in store for the program. One reason is that trading activity is increasing. Trades can be recorded with the EPA at any time prior to the use of an allowance for compliance, and recorded trades are monitored in the EPA's electronic, on-line Allowance Tracking System (ATS). The EPA has developed an algorithm for classifying trades as "economically significant" if they are transfers between independent firms, and the majority of trades that are recorded have been transfers for accounting convenience or other reasons within firms. However, the number of economically significant trades has virtually doubled each year through 1997. A second reason for optimism about the efficiency of the market is that the industry- wide level of emissions increased slightly in 1996. An adjustment was necessary because the weighted average costs in 1995 were greater than the present discounted value of marginal 9 Burtraw RFF 98-28-REV costs in the future and greater than allowance prices that reflect the latter measure. Behavior at the outset of the program appears to have been risk averse, with firms making sure they could achieve compliance through a "go-it-alone" strategy. Subsequently, there appears to be a growing level of comfort with the use of the market as a way to ensure compliance. A third reason for optimism has to do with over-arching trends in the electric utility industry in the US. The industry is in a fundamental period of realignment as competition at the wholesale level, and perhaps ultimately at the retail level, is beginning to emerge. This type of competition is placing pressure on the industry to find ways to reduce costs in all segments of its business. One should not expect the industry to absorb $300 million in unnecessary costs in the future when allowance trading provides a fairly simple means of reducing those costs. Though widely known, it is not often acknowledged that compliance costs and economic costs are not the same thing. Compliance costs describe the out-of-pocket expenses by a firm or industry to comply with regulations. Economic costs describe the value of goods and services that were lost to the economy due to the regulation. This can include so-called hidden costs or benefits, such as costs incurred but not reported as compliance costs or indirect productivity changes that result from environmental compliance. One type of important hidden cost stems from the interaction of the program with the pre-existing tax system. Important distortions away from economic efficiency stem from pre- existing taxes. Labor income taxes are particularly important because of their magnitude in the economy. Labor taxes impose a difference between the before-tax wage (or the value of the marginal product of labor to firms) and the after-tax wage (or the opportunity cost of labor from the worker's perspective). This difference causes workers to substitute away from labor to leisure compared to an efficient outcome were these two measures equal. Any regulation that raises product prices potentially imposes a hidden cost on the economy by lowering the real wage of workers. This can be viewed as a "virtual tax" magnifying the significance of previous taxes, with losses in productivity as a consequence. If there were no pre-existing distortions in the economy, the impact of regulatory costs would be of little concern. However, the cost of distortions associated with taxes grow more than proportionally with the size of the tax, and hence the hidden cost of regulation can be of great importance when pre-existing taxes are taken into account. This hidden cost has been termed the tax-interaction effect (Parry, 1995) and it tends to erode the usual efficiency benefits identified with setting prices to include external costs. The tax interaction effect is particularly important in the allowance trading system, relative to a command and control approach, because the price of the final product in a competitive market should reflect not only compliance cost for emission reductions but also the opportunity cost (or price) of allowances used for compliance. As the electric utility industry in the US moves toward competition, this effect of allowances on electricity prices is expected to emerge. The trading program is expected to result in significant savings in compliance 10 Burtraw RFF 98-28-REV costs, but commensurately it internalizes additional costs in the way of allowance prices into electricity prices. 13 Goulder et al. (1997) have investigated the magnitude of the tax-interaction effect in the context of the SO₂ program using both analytical and numerical general equilibrium models. They find that this effect will cost the economy about $1.06 billion per year (1995 dollars) in Phase II of the program, adding an additional 70 percent to their estimated compliance costs for the program. There are important lessons here for the design of other environmental programs that may refer to the SO₂ program as a model. First Goulder et al. (1997) find that the tax interaction effect is more significant, relative to the magnitude of compliance costs, for programs that are aimed at small emission reductions such as may describe possible policies for CO₂ reductions. 14 Further, since the tax-interaction effect is positive even for a small emission reduction, then it is possible for the economic costs of regulation to be greater than the benefits even when compliance costs start at zero. Second, the cost of the tax-interaction effect can be largely (but not entirely) offset by policies that raise revenues for the government, because these revenues can (in principle) be used to offset preexisting taxes and correct distortions in the labor market resulting from these taxes. The authors find that over half of the economic cost of the tax interaction effect, or $622 million, could be avoided if emission allowances were auctioned rather than grandfathered and the revenue was used to reduce the marginal tax rate on labor income. Unfortunately, the SO₂ program does not raise revenues since allowances are distributed for zero cost. Consequently, the current program design imposes a hidden cost on the economy that could be avoided if allowances were auctioned instead of allocated without charge. The recommendation that allowances be auctioned comes with significant political liability. The endowment of allowances without charge, so-called "grandfathering" of allowances, is an important form of compensation to the electric utility industry. The industry's attitude toward the SO₂ program would have been considerably more negative had this compensation not existed. Hence policy makers face a trade-off between efficiency and compensation in the design of the program. There is an equity aspect to this issue that counter-balances the concerns of industry as well. At the time legislators adopted the SO₂ program in 1990, state public utility commissions were in the business of regulating the industry and setting electricity prices, and they were the safeguard to ensure utilities could not charge customers for something the utilities received for free. Hence, endowing allowances at zero cost was not controversial to 13 The effect on product prices should occur without regard to how the firm acquired allowances originally, if prices reflect marginal costs and allowances are valued at their opportunity cost. 14 Small emission reductions require small compliance expenditures, leaving a larger quantity of emissions to be included in a permit trading scheme. Allowance prices would be low, relative to a case with greater emission reductions, but they apply to a larger quantity of emissions. Hence, the portion of costs associated with permit use relative to the portion of costs associated with compliance cost is greater. 11 Burtraw RFF 98-28-REV the design of the SO₂ program. However, in the near future we expect regulators to exit the business of setting prices, at least with respect to electricity generation, and electricity prices will be set in a competitive market. 15 In the textbook and presumably in the market, SO₂ allowances take on the value of their opportunity cost, and this will be passed on through marginal cost pricing in a competitive electricity market, regardless of how the utility acquired the allowances originally. The value of this endowment, coming out of the hides of electricity consumers and accruing to the industry, is potentially very large. A potentially important qualification concerns whether coal-fired electric generating facilities which are the source of SO₂ emissions are likely to be the marginal unit that is the highest cost unit that is operated to meet electricity demand and hence will be determining electricity price. For example, at times when a gas-fired unit is the marginal unit, the price of electricity should reflect the marginal costs of that unit and will not reflect allowance costs for coal-fired generation that may be running at costs below the margin. On the other hand, when coal-fired units are the marginal unit then all infra-marginal technologies that operate at less cost than the marginal unit will receive the same price of electricity, which reflects the costs of allowances. To investigate this issue, we exercised an electricity dispatch model aggregated by season, time block and region, that is maintained at Resources for the Future. The model is intended to characterize the industry under competitive conditions expected to prevail after deregulation. We find that coal-fired facilities can be expected to generate about one-half of all electricity generated in the US, just as they do under current regulatory practice. Further, we find that about one-half of all electricity would be generated when a coal facility is the marginal unit. Hence, on a per kilowatt-hour basis, the effect on electricity prices from emission allowances appears to be about the same as it would be if we were to assume generation from coal-fired units always occurred when these were the only units operating and their marginal costs were to determine electricity prices when they were in operation. Under the assumption that the program works efficiently and firms value SO₂ allowances at their opportunity cost (the price of an allowance, equal to the marginal abatement cost) then the magnitude of the compensation under the program (the grandfathering of allowances valued at their opportunity cost) far outweighs the compliance cost incurred by firms. I estimate that the present discounted value of this difference is between $10 and $20 billion (1995 dollars). This represents a tremendous transfer of wealth from electric utility customers to the industry, and this transfer should be of interest in the design of other programs. In particular, the design of a CO₂ permit program would involve smaller emission reductions in percentage terms and therefore a much greater transfer of wealth relative to the cost of compliance and also in absolute terms. Both the equity and the efficiency aspects of 15 This will occur if state or federal policy-makers move to retail competition. At the time of this writing, fifteen states have committed to retail competition to be implemented at some future date, and virtually all states are actively considering doing so (Ando and Palmer, 1998). Even if retail competition is slow to develop, wholesale competition is likely to emerge in the near future. 12 Burtraw RFF 98-28-REV the tax-interaction effect can be largely remedied by a carbon tax or auctioned carbon permits. Recognizing that there may political obstacles to a revenue-raising carbon policy, a hybrid system involving an auction of some portion of the permits and grandfathering the remaining may be a useful compromise. The primary measure of success of the SO₂ program, from the perspective of economics, is the comparison of benefits and costs. Burtraw et al. (1997) describe an integrated assessment of the of benefits and costs through the year 2030, with benefits quantified for health, visibility and aquatics. 16 The accounting of benefits is incomplete because several environmental pathways could not be modeled completely. The cost estimate and estimate of health benefits are calculated for the entire nation, but visibility and aquatic benefits are calculated only to illustrate the potential relative magnitude of these benefits; and, the estimates do not necessarily apply to the entire nation. Midpoint estimates of the benefits and costs per affected capita for the year 2010 are summarized in Table 4. (1995 dollars) 4.09 69.25 0.72 3.90 6.79 6.19 Source: Burtraw et al. (1997) The study finds that benefits of the SO₂ program are an order of magnitude greater than costs, a result that contrasts sharply with estimates in 1990 that pegged benefits about equal to costs (Portney, 1990). What explains the difference between the earlier and recent estimates? 16 The integrated assessment involved nearly thirty researchers at a dozen institutions The assessment is based on reduced-form models that were calibrated to several larger models of utility emissions and costs, atmospheric transport of pollution, visibility impairment, effects on aquatic systems and human health, and valuation of effects. Economic assessment of costs are calculated with an engineering model constructed for the assessment. Economic valuation of damage to aquatic systems relied on random utility models of recreational use. Health mortality valuation relied on compensating wage and contingent valuation studies and morbidity valuation relied on a number of studies and methods. Valuation of visibility effects at national parks used contingent valuation methods, and valuation of visibility in residential areas used a combination of contingent valuation combined with hedonic property value studies. 13 Burtraw RFF 98-28-REV The lion's share of benefits result from reduced risk of premature mortality, especially through reduced exposure to sulfates, and these expected benefits measure several times the expected costs of the program. Significant benefits are also estimated for improvements in health morbidity, recreational visibility and residential visibility, each of which measures approximately equal to costs. These areas, namely human health and visibility, were not the focus of acid rain research in the 1980s, and new information suggests these benefits are greater than were previously anticipated. In contrast, areas that were the focus of attention in the 1980s including effects to soils, forests and aquatic systems still have not been modeled comprehensively, but evidence suggests benefits in these areas to be relatively small, at least with respect to "use values" for the environmental assets that are affected. In the 1980s public attention to air pollution from SO₂ and NOₓ emissions largely centered on the problem of acidification ("acid rain"), with particular concern for its affect on water and soil chemistry and ultimately ecological systems. It is surprising to many that relatively low benefits are estimated by economists for effects on aquatics (in Burtraw et al., 1997) or are expected to result from effects on forests and agriculture. One reason is that willingness to pay for environmental improvement depends on the availability of substitute assets. Economists would not expect changes in quality at one site to elicit large benefits if there are many sites available for comparable recreational opportunities. In contrast, individuals do not have the same kind of substitution possibilities with respect to health and visibility, which may help explain the relatively larger benefit estimates for these endpoints. Furthermore, one should note that the low values for aquatics stem from an assessment of values, or values in the case of agriculture. Environmental changes may also yield values, and estimates for nonuse values are not available. Nonetheless, the evidence, based on a small number of relatively narrow studies, suggests these values may be significant. On the cost side, compliance has cost one-half or less what it was anticipated to cost in 1990. Trends in both directions, that is, increased appreciation of benefits and decreased estimate of costs have qualified the program as a tremendous success from an economic perspective. There are huge uncertainties, especially on the benefits side of the ledger, especially in valuation of mortality. Recent economic critiques have argued that the use of the value of a statistical life as the basis for valuing health risks from air pollution, instead of a more appropriate measure of quality adjusted life years lost, could grossly overestimate mortality benefits. In addition, economists have questioned the appropriateness of using labor studies of prime age men to value changes in life expectancy that occur among an older population. On the other hand, we note that because environmental exposures are involuntary, compared with studies of labor market behavior, the latter may underestimate willingness to pay to avoid environmental exposures. Burtraw et al. (1997) used Monte Carlo analysis and a parametric one-sided sensitivity analysis to investigate some of these sources of uncertainties. Their analysis finds that there is no year in which health benefits alone at the 5 percent confidence interval are less than the levelized expected costs. These results are illustrated in 14 Burtraw RFF 98-28-REV Figure 1, with millions of (1990) dollars on the vertical axis and years of the program on the horizontal axis. As noted, significant benefits are also estimated for improvements in health morbidity, recreational visibility and residential visibility, each of which measures approximately equal to costs. Despite tremendous uncertainties about benefits and costs, the main conclusion that benefits soundly outweigh costs appears to be robust. 45,000 40,000 35,000 30,000 25,000 Benefits 20,000 Cost 15,000 10,000 5,000 1995 2000 2005 2010 2015 2020 2025 2030 Source: Burtraw et al. (1997) Although tremendous attention has been addressed to the innovative design of the SO₂ emission allowance trading program, an economic approach to policy analysis should perhaps focus first on the comparison of benefits and costs of the program. Evidence to date suggests that benefits are expected to greatly outweigh the costs of the program. Surprises on the benefits side that contribute to this favorable assessment are in benefit categories of human health and visibility that were not the primary focus of research on acid rain in the 1980s. Favorable events on the cost side have to do with the expected successful implementation of the allowance trading program, and with advantageous trends in markets that supply the electric utility sector. Developments in these input markets, which include scrubber manufacturing, coal mining and rail transport of coal, are at least partially related to the SO₂ program. Due to advantageous trends in the industry leading to increased use of low sulfur coal, the quantity of emissions that had to be reduced is less than anticipated. This led to lower costs than many envisioned, and commensurately SO₂ trading could provide a smaller measure of cost savings. However, the SO₂ program deserves considerable credit for 15 Burtraw RFF 98-28-REV providing firms with the flexibility to capitalize on these advantageous trends as they strategize their compliance activities. There is some reason for concern about the ability of the SO₂ market to function efficiently and for firms to capture possible cost savings based on evidence for 1995, the first year in which the program was binding. However, there are several reasons one can expect the market to become more efficient over time and the long-run prospects for the program are bright. As scholars and policy analysts attempt to draw lessons from the SO₂ program, one area that should receive significant attention is the manner in which emission allowances are allocated to the industry. Evidence suggests that grandfathering of allowances can impose significant efficiency costs. Furthermore, this approach represents a tremendous transfer of wealth that raises equity issues as well. Rarely in economics do efficiency and equity issues point in the same direction, but in this case they do. The recommendation that follows is that emission allowances should be auctioned or allocated in some means that raises revenue for government that can be used to reduce other distortionary taxes. If the allocation of allowances is to serve as compensation for industry, this function should be weighed carefully against the benefits of raising revenue. A hybrid program, in which some portion of allowances are grandfathered and the rest auctioned by the government could offer a compromise that would improve programs of this type in future applications. 16 Burtraw RFF 98-28-REV Ando, Amy, and Karen Palmer. 1998. , Discussion Paper 98-19, Resources for the Future, Washington, D.C., March. Bohi, Douglas R., and Dallas Burtraw. 1997. "SO₂ Allowance Trading: How Do Expectations and Experience Measure Up?," , vol. 10, no. 7, pp. 67-77. Burtraw, Dallas. 1996. "The SO₂ Emissions Trading Program: Cost Savings Without Allowance Trades," , xiv (April), pp. 79-94. Burtraw, Dallas, Alan J. Krupnick, Erin Mansur, David Austin, and Deirdre Farrell. 1997. , Discussion Paper 97-31-REV, Resources for the Future, Washington, D.C., September. Carlson, Curtis, Dallas Burtraw, Maureen Cropper, and Karen Palmer. 1998. , Discussion Paper 98-44, Resources for the Future, Washington, D.C., July. Ellerman, A. Denny, and Juan Pablo Montero. 1998. "Why Are Allowance Prices so Low? An Analysis of SO₂ Emissions Trading Program," (forthcoming). Ellerman, A. Denny, Richard Schmalensee, Paul L. Joskow, Juan Pablo Montero, and Elizabeth M. Bailey. 1997. (Cambridge, Mass.: MIT Center for Energy and Environmental Policy Research). Elman, Barry, Bruce Braine, and Richard Stuebi. 1990. "Acid Rain Emission Allowances and Future Capacity Growth in the Electric Utility Industry," , vol. 40, no. 7, pp. 979-986. Goulder, Lawrence H., Ian W.H. Parry, and Dallas Burtraw. 1997. "Revenue-Raising vs. Other Approaches to Environmental Protection: The Critical Significance of Pre-Existing Tax Distortions," , vol. 28, no. 4 (Winter), pp. 708-731. ICF Resources Incorporated (ICF). 1990. , prepared for U.S. Environmental Protection Agency, Washington, D.C., July. ICF. 1995. , prepared for the U.S. Environmental Protection Agency, September. Parry, Ian W.H. 1995. "Pollution Taxes and Revenue Recycling," , vol. 29, no. 3 (supplement), pp. S-64-S-77. Portney, Paul R. 1990. "Economics and the Clean Air Act," , vol. 4, no. 4, pp. 173-181. 17 Burtraw RFF 98-28-REV Temple, Barker, and Sloane, Inc. (TBS). 1983. , prepared for the Edison Electric Institute, September 20. Temple, Barker, and Sloane, Inc. (TBS). 1986. prepared for the Edison Electric Institute, April 14. U.S. Energy Information Administration (EIA). 1996. DOE/EIA-0383(97), Washington, D.C., December. U.S. Government Accounting Office (GAO). 1994. , GAO/RCED-95-30, Washington, D.C. U.S. Office of Technology Assessment (OTA). 1983. , Staff Memorandum (revised July 12). Van Horn Consulting, Energy Ventures Analysis, Inc., and Keith D. White. 1993. , prepared for the Electric Power Research Institute, EPRI TR-102510 (August). White, Keith. 1997. prepared for the Electric Power Research Institute (EPRI), EPRI TR-107897 (April). White, Keith, Energy Ventures Analysis, Inc., and Van Horn Consulting. 1995. prepared for the Electric Power Research Institute (EPRI), TR-105490, Palo Alto, Calif. (Final Report: September). 18 The Costs and Benefits of Reducing Acid Rain Dallas Burtraw, Alan Krupnick, Erin Mansur, David Austin, and Deirdre Farrell Discussion Paper 97-31-REV Revised September 1997 Original paper issued July 1997 RESOURCES FOR THE FUTURE 1616 P Street, NW Washington, DC 20036 Telephone 202-328-5000 Fax 202-939-3460 © 1997 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the authors. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not undergone formal peer review or the editorial treatment accorded RFF books and other publications. The Costs and Benefits of Reducing Acid Rain Dallas Burtraw, Alan Krupnick, Erin Mansur, David Austin, and Deirdre Farrell* Abstract Title IV of the 1990 Clean Air Act Amendments initiated a dramatic reduction in emissions of sulfur dioxide and nitrogen oxides by electric power plants. This paper presents the results of an integrated assessment of the benefits and costs of the program, using the Tracking and Analysis Framework (TAF) developed for the National Acid Precipitation Assessment Program (NAPAP). Although dramatic uncertainties characterize our estimates especially with respect to the benefits of the program, many of which we have modeled explicitly, we find that the benefits can be expected to substantially outweigh the costs of the emission reductions. The lion's share of benefits result from reduced risk of premature mortality, especially through reduced exposure to sulfates, and these expected benefits measure several times the expected costs of the program. Significant benefits are also estimated for improvements in health morbidity, recreational visibility and residential visibility, each of which measures approximately equal to costs. In contrast, areas that were the focus of attention in the 1980s including effects to soils, forests and aquatic systems still have not been modeled comprehensively, but evidence suggests benefits in these areas to be relatively small, at least with respect to "use values" for the environmental assets that are affected. Key Words: acid rain, sulfur dioxide, nitrogen oxides, cost-benefit analysis, Clean Air Act, Title IV JEL Classification Nos.: H43, Q2, Q4 * Resources for the Future, 1616 P Street, Washington DC, 20036; [email protected]. The authors acknowledge the contribution of over 30 researchers at nearly a dozen research institutions around the country who helped develop the Tracking and Analysis Framework, and especially the intellectual contribution of Cary Bloyd, Max Henrion, John Molburg, Jack Shannon and Rich Sonnenblick, Mitch Small, Tim Sullivan and many others, and the members of the model peer review panel organized by Oak Ridge National Laboratory. We also acknowledge the financial and intellectual contribution of NAPAP and its member agencies, especially the Department of Energy, the Environmental Protection Agency, and the National Oceanic and Atmospheric Administration. We appreciate comments from Jane Hall on a previous version of this paper. ii Table of Contents I. Introduction 1 II. Description of TAF 2 The Benefits Valuation Module 4 Health Effects 4 Visibility 6 Recreational Lake Fishing 7 Costs and Emissions 8 Other Effects Not Modeled 8 III. Baselines and Scenarios 9 The Hagler Bailly Study of Health Benefits from Sulfate Reductions 12 The EPA's Regulatory Impact Assessment for Particulates 12 Alternative Cost and Emission Estimates 12 IV. Results 13 Comparison with HB Study 19 Comparison with the RIA 22 Cost Comparisons 22 Unmodeled Pathways and Research Priorities 24 V. Conclusions 26 References 30 List of Tables and Figures Table 1. Options for Assessing Mortality Effects 6 Table 2. Per Capita Benefits in 2010 for Affected Population 13 Table 3. Expected Total Health Benefits for 2010 and Percent of National SO2 Emission Reductions by State 19 Table 4. Comparison of HB and TAF Mortality Sulfate Benefits for Eastern U.S. with Percent Changes over Previous Scenario, Year 2010 20 Table 5. Long-run (Phase II, year 2010) Cost Estimates for SO2 Reduction 23 Table 6. Long-run (Phase II, year 2010) Cost Estimates for NOx Reduction 24 Table 7. Qualitative Evaluation of Expected Benefits and Value of Additional Information for Modeled and Nonmodeled Pathways 25 Table 8. Major Uncertainties and Omissions and Direction of Bias in TAF 28 Figure 1. Costs and Benefits for Modeled Pathways for Affected Populations 14 Figure 2. Benefit-cost Ratio for Health Benefits under Alternative Assumptions about Electricity Industry 16 Figure 3. Annual Mean Total Health Benefits with 90% Confidence Intervals Compared with Expected Annualized Costs 17 Figure 4. Annual Mean Total Health Benefits and Levelized Costs by Scenario 18 iii The Costs and Benefits of Reducing Acid Rain Dallas Burtraw, Alan Krupnick, Erin Mansur, David Austin, and Deirdre Farrell* I. INTRODUCTION This paper presents the first contemporary analysis of the prospective benefits and the costs of Title IV's Allowance Trading System for reducing sulfur dioxide (SO₂) emissions and Title IV's mandated reductions in emissions of nitrogen oxides (NOx). 1 This benefit-cost assessment is conducted using the Tracking Analysis Framework (TAF) that was developed to support the activities of the National Acid Precipitation Assessment Program (NAPAP). Control of SO₂ emissions under the 1990 Clean Air Act Amendments instituted two important innovations in U.S. environmental policy. The more widely acknowledged of these is the SO₂ emissions trading program. Firms are allowed to transfer allowances among facilities, or to bank them for use in future years. Less widely acknowledged is the average annual cap on aggregate emissions by electric utilities, set at about one-half of the amount emitted in 1980. The emissions cap represents a guarantee that emissions will not increase with economic growth. Title IV also used a more traditional approach in setting NOx emission rate limitations for coal-fired electric utility units, although this approach has been modified to allow emission rate averaging among commonly owned and operated facilities. Hence, there is no cap on NOx emissions, but Title IV is expected to result in a 27 percent reduction from their 1990 emissions. Below, we describe TAF and its components, and then turn to the development and description of baseline assumptions, and default and sensitivity case scenarios. Next we report the results for our default scenario. Then, three sets of sensitivity analyses are reported. One set uses assumptions other than our default assumptions. We explore the sensitivity of results to changes in baselines and compare levelized costs with expected health mortality benefits only. Confidence intervals are constructed out of the statistical uncertainties associated with the health effects and monetary values. Subsequently we compare the TAF mortality benefits with estimates developed by Hagler Bailly (HB, 1995) for EPA's Acid Rain Division, and the * Resources for the Future, 1616 P Street, Washington DC, 20036; [email protected]. The authors acknowledge the contribution of over 30 researchers at nearly a dozen research institutions around the country who helped develop the Tracking and Analysis Framework, and especially the intellectual contribution of Cary Bloyd, Max Henrion, John Molburg, Jack Shannon and Rich Sonnenblick, and many others, and the members of the model peer review panel organized by Oak Ridge National Laboratory. We also acknowledge the financial and intellectual contribution of NAPAP and its member agencies, especially the Department of Energy, the Environmental Protection Agency, and the National Oceanic and Atmospheric Administration. We appreciate comments from Jane Hall on a previous version of this paper. 1 There exist analyses of benefits and analyses of the costs, some of which we describe below. This is the first analysis to compare benefits and costs under uniform assumptions. 1 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV EPA's Regulatory Impact Analysis (RIA) for the proposed new particulate standard (USEPA, 1996). Finally we compare cost estimates in TAF with other recent studies. We find that expected health mortality benefits alone far exceed expected costs in all but one of the variations in assumptions we tried, and that even 5th percentile estimates of benefits do not dip below costs for any of these scenarios (but one) in any years after 2010. Our mortality benefits estimates are substantially lower than HB's, though we predict larger changes in sulfate concentrations for a given change in emissions. Our lower mortality benefit estimates can be attributed primarily to our estimates of smaller emission reductions (lower emissions in the baseline against which we compare the policy), and smaller reductions in mortality risks for a given change in sulfates and a smaller value of statistical life. In addition, we find that mean values for three other modeled pathways--health morbidity, recreational visibility, and residential visibility--each approximately equal mean levelized costs of the program. Recreational lake fishing benefits appear to be of much lower magnitude. We compare these estimates in a qualitative and relative ranking with our informed conjecture about the likely magnitude and uncertainty of several other benefit categories that are not modeled quantitatively. We compare the TAF default cost estimate with other scenarios and other recent estimates within TAF. The TAF default estimate is on the low end of other estimates with respect to SO₂. In part, this reflects a downward trend in the estimates and secular trends in the industries that affect compliance. The TAF estimate as well as these others assumes that a well-functioning SO₂ allowance market will allow the industry to achieve emission reductions at minimum feasible cost. This is not strictly evident in the first two years of compliance with the program, and so the short-run cost estimates may be low. However, we expect least cost compliance to be a reasonable characterization of the future, as the electricity industry becomes increasingly competitive. The estimates of NOₓ control costs are proximate to those of other recent estimates; however, the estimates do not reflect the opportunity for emission rate averaging which are expected to lower costs. At the same time, the TAF default estimate and all other estimates are low because they fail to account for the costs of the regulation within a general equilibrium framework with pre-existing taxes and distortions away from efficiency. Nonetheless, with all the omissions, caveats and uncertainties taken into account, we find the benefits of Title IV exceed the costs by a significant margin. II. DESCRIPTION OF TAF The Tracking and Analysis Framework (TAF) is an integrated assessment model of acid precipitation damages and the effects of Title IV of the 1990 Clean Air Act Amendments. 2 TAF integrates models of electric utility emissions and costs, pollutant transport and deposition (including formation of secondary particulates but excluding ozone), visibility effects, effects on recreational lake fishing through changes in soil and aquatic 2 TAF is documented in Bloyd et al. 1996. 2 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV chemistry, human health effects, and valuation of benefits. The model also can be readily extended to include other benefit pathways as modeling capability is developed. To ensure that each component represented the state of the science in its respective modeling domain, each module was constructed and refined by a group of experts in that field, and draws primarily on peer reviewed literature to construct the integrated model. Thus, TAF is the work of a team of over 30 modelers and scientists from institutions all over the country. As the framework integrating these literatures, TAF itself was subject to an extensive peer review in December 1995, which concluded that "TAF represent(s) a major advancement in our ability to perform integrated assessments" and that the model was ready for use by NAPAP.³ Considerable uncertainty in parameter and model form exists in each of our modeled domains and in the underlying scientific and economic literature. We selected Analytica™ as the modeling platform for TAF in part because of its capability to propagate model uncertainties, and we adopted a process designed to identify and characterize those uncertainties. We chose an integrated assessment framework, as opposed to a suite of related but unlinked models, because it met the following needs of the TAF project: To provide comparable results across a variety of effects (visibility, recreational lake fishing, human health), for a common region (continental US), and over a single time horizon (1995-2030); To provide an integrated analysis of costs and benefits based on common assumptions, and to provide insight about model assumptions and components which contribute significantly to overall results; To suggest productive areas for future research and additional modeling based on an assessment of the current model's critical uncertainties and omissions. TAF characterizes emissions, emission transport, atmospheric concentrations of pollutants and health effects at the state level. This level of aggregation introduces some uncertainty into the analysis, but it is not evident that a bias is introduced. The estimation of effects also is amenable to modeling at a less centralized level, and we use probabilistic methods to represent variations in sources of emissions, geography and population density within states. TAF omits benefits that occur in Canada and Mexico. Recreational lake effects are characterized for a distribution of lakes in the Adirondacks. Recreational visibility effects are characterized at two parks and valued nationally, and residential visibility effects are characterized and valued for five metropolitan areas. These results are most usefully considered on a per capita basis. In this paper we do not try to assess regional issues in a thorough way, but we do display the regional pattern of health benefits. 3 ORNL, 1995. TAF is being used to provide supplementary analysis to NAPAP in drafting their 1996 Integrated Assessment, but it is not a centerpiece. 3 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV The Benefits Valuation Module Benefit valuation is essential for comparison of various physical effects with each other and with costs. From an economic perspective, values are measured by how much of one asset or service individuals in society are willing to sacrifice in order to obtain or preserve another. Economics refers to this as an "opportunity cost approach" to valuation. Values are expressed in monetary terms, although, in principle, they can be expressed in other metrics. The value or opportunity cost of goods and services that are readily traded in markets is reflected in their prices. For goods that are not traded in markets, the economics literature on monetizing benefits and costs is more developed in certain areas than in others, which is reflected in the characterization of uncertainty in the benefit models. The Benefits Valuation Module provides an accounting of pathways and benefit endpoints considered in TAF. The module values effects on visibility (recreational and residential), Adirondack lake sport fish populations, and human health. These effects are valued only where physical effects have been modeled in TAF, so comprehensive geographic coverage is not provided. Other kinds of effects, such as forest, stream, and material damages, are not valued at this time, but they are represented in TAF in a qualitative manner. Health Effects The Health Effects Module is designed to estimate the health impacts of changes in air pollution concentrations. Impacts are expressed in terms of the number of days of acute morbidity effects of various types, the number of chronic disease cases, and the number of statistical lives lost to premature death. The change in the annual number of impacts of each health endpoint is the output of this module. Inputs consist of changes in ambient concentrations of SO₂ and NOₓ, demographic information on the population of interest, and miscellaneous additional information such as background PM₁₀ levels for analysis of thresholds. The module is based on concentration-response (C-R) functions found in the peer- reviewed literature. The C-R functions are taken, for the most part, from articles reviewed in the U.S. Environmental Protection Agency (EPA) Criteria Documents (see, for example, USEPA 1995). These documents are outcomes of a recurring comprehensive process initiated by the Clean Air Act and its Amendments for reviewing what is known about the health effects of the so-called "criteria" air pollutants.4 Such information, and judgments about its quality, eventually help the Administrator of the EPA make decisions about National Ambient Air Quality Standards (NAAQS) that would "protect the public against adverse health effects with a margin of safety." These Criteria Documents contain thousands of pages evaluating toxicological, clinical, and epidemiological studies that relate particular criteria pollutants to a variety of health endpoints, including primarily acute cardiopulmonary and respiratory effects, chronic effects and prevalence of chronic illness, and premature mortality. The TAF 4 The Criteria Pollutants include ozone [O₃], nitrogen dioxide [NO₂], sulfur dioxide [SO2], particulate matter less than 10 microns in diameter [PM₁₀], lead [Pb], and carbon monoxide [CO]. 4 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Health Effects Module contains C-R functions for PM₁₀, total suspended particulates (TSP), SO₂, sulfates (SO₄), NO₂, and nitrates (NO3). 5 The Health Effects Module calculates morbidity impacts resulting from sulfates and nitrates, which are particulates created from emissions of SO₂ and NOx, respectively, and SO₂ and NOx as gases. Mortality impacts are only represented as resulting from the particulates. The C-R functions found in the literature for these endpoints are documented within the software model. The top level of the Health Effects Module is structured according to an influence diagram that visually depicts the fact that, in a C-R function, concentration changes and demographic data determine the number of morbidity and mortality impacts experienced in a population. Within the morbidity and mortality submodules there is great flexibility to structure the model so as to test a range of assumptions about the relationship between pollutant concentrations and health effects. For both the morbidity and mortality endpoints, the Health Effects Module contains a comprehensive library of C-R functions found in the peer reviewed literature, in total consisting of more than fifty studies linking air pollution to premature death, chronic disease, hospitalizations and other symptoms. The user may select from among any of the studies in the library available for a given health endpoint, or may decide to weight coefficients from a number of studies. For the mortality endpoint, in addition to the choice of a C-R function, the user may decide to treat the various components of particulate matter separately. For example, some evidence suggests that the fine fraction of the mass may have more of an effect than the coarser components. Four plausible interpretations of the evidence on this subject are offered to the user as options (Table 1). Within each of these options, the user may choose from among the available studies in the library for each pollutant and endpoint, or use a combination of other C-R functions weighted to reflect the user's judgment. Options 1 and 2 assume that sulfates and nitrates have the equivalent potency in causing health effects as any other particle 10 microns or less in diameter (PM₁₀), but option 2 allows the user to look at the age-disaggregated effects of air pollution on mortality. These reflect the fact that the over 65 population is more likely to die as a result of high particulate levels than is the under 65 population. Option 3 treats sulfates as distinct and associates them with relatively greater potency than other constituents of PM10. Option 4 treats both sulfates and nitrates as relatively more potent than other components of PM₁₀. We focus on the third option as most plausibly representing the evidence at the time the work was completed. The morbidity submodule allows the user a choice of either aggregating SO₂, PM₁₀ and sulfate effects according to a scheme designed to avoid double-counting, such as symptom days and restricted activity days, or of using SO₄ effects as a proxy for particulate and SO₂ effects. NOx is included for eye irritation and phlegm days. As with the mortality 5 Since nitrates are particulates, and no independent effect of nitrates on health has been established, they are treated as a component of PM₁₀. 5 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV submodule, default studies have been identified for each endpoint, but where other studies exist in the literature they may be substituted for the defaults. Table 1. Options for Assessing Mortality Effects Option 1 Sulfates and nitrates treated as PM10 Option 2 Sulfates and nitrates treated as PM10, disaggregated by age Option 3 (default) Nitrates treated as PM10, sulfates distinct and more potent Option 4 Sulfates and nitrates treated as sulfates The Health Valuation Submodule of the Benefits Valuation Module assigns monetary values taken from the environmental economics literature (e.g., Lee et al., 1994) to the health effects estimates produced by the Health Effects Module. The benefits are totaled to obtain annual health benefits for each year modeled. The Health Valuation Submodule also contains a comprehensive library, based on the environmental economics literature, of values associated with morbidity and mortality endpoints. As with the Morbidity and Mortality Submodules, defaults have been selected, but the user may test the effects of assigning alternative values to the various health endpoints, consistent with the valuation literature. Visibility The Visibility Effects module calculates changes in visual range for five cities (Albany, NY; Atlantic City, NJ; Charlottesville, VA; Knoxville, TN; and Washington, DC), and two national parks (the Grand Canyon, and Shenandoah). Seasonal distributions of mid- day visual range are based on estimated atmospheric sulfate and nitrate concentrations from the Atmospheric Pathways module a reduced-form model of the Advanced Statistical Trajectory Regional Air Pollution (ASTRAP) model. Calculation of change in visual range is based on the Visibility Assessment Scoping Model (VASM), which uses Monte Carlo techniques to produce short-term variations of visual impairment based on seasonal lognormal distribution parameters of the six important particulate species (sulfate, nitrate, elemental carbon, organic carbon, fine-particle dust, and coarse-particle dust), relative humidity distribution statistics from climatology, and modeled changes in the seasonal means of the sulfate and nitrate concentrations. The Visibility Valuation submodules examine both recreational and residential benefits. Chestnut and Rowe (1990) proposed a functional form to value both recreational and residential visibility that takes into account the nonlinearity of willingness to pay (WTP) for a given change in visual range (i.e., the diminishing marginal utility for visibility enhancement). WTP for improvements in recreational visibility were drawn from contingent valuation (CV) studies and involve both use and nonuse values for residents living in either park's state or another state ("out-of-state" residents). To value residential visibility improvements we employ a range of WTP coefficients from the Brookshire et al. (1979) Los Angeles study, and the McClelland et al. (1991) study of Atlanta and Chicago. We assume residential WTP is 6 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV positive only for local residents (e.g. only use values matter), so we adjust values for "in-state" recreational visibility to avoid double counting with improvements in residential visibility. Recreational Lake Fishing The Recreational Lake Fishing module predicts changes in lake chemistry and soil chemistry caused by acid deposition. Using a set of "acid stress indexes" (ASIs) that describe the responses of specific species of fish to varying levels of acidity (pH) in the water, the module estimates economic benefits resulting from improvements in recreational fishing due to decreased acidification. Future surface-water and soil chemistry conditions in the watersheds are projected by reduced-form models based on the Model of Acidification of Groundwater in Catchments (MAGIC). MAGIC is a lumped-parameter model that uses chemical equilibrium and mass balance equations to predict changes in lake and soil chemistry. The reduced-form models have been applied to lakes in New York's Adirondack region, using a set of 33 lakes chosen to be representative of the target population of lakes in the region. The Recreational Lake Fishing Benefits module allows the user to specify whether benefits are to be estimated on the basis of benefits to recreational anglers, or alternatively as avoided lake liming costs. The Recreational Lake Fishing submodule estimates changes in the catch rates (catch per unit effort, or CPUE) of anglers fishing for three species of fish in Adirondack Park. Values are assigned to these changes through the use of a "random utility" travel cost model. Benefits are calculated for the change in value of a single-day fishing trip (as opposed to an overnight or multi-day outing) as a result of changes in CPUE. The submodule also estimates the change in the annual number of single-day fishing trips the average Adirondack Park angler will take in the park, as a function of changes in CPUEs and other factors. 6 The submodule does not attempt to account for benefits enjoyed by new anglers attracted by improved conditions, or for angler benefits other than improvements in catch rates.⁷ There are two reasons for this simplification. First, no more than 10 percent of the population fishes for recreation, and the use benefits enjoyed by non-anglers are probably of second order to the anglers' values for improved catch rates. Second, there are no reliable estimates on which to base valuations of non-fishing uses--in part because relative aesthetic effects are much smaller than the effects on fish populations. TAF does not include estimates of the nonuse or existence benefits that may be enjoyed by persons not visiting the affected lakes. Other parts of the country are not modeled currently. However, we use the effects in 6 The aquatics valuation literature focuses on single-day trips because it is thought that valuations for multi-day trips, of which there are far fewer, are intrinsically different. For instance, it would be necessary to better control for lake amenities such as lodging and camping facilities, which would presumably be important determinants of lake choice. These "use values", for multi-day trips, are not represented in the TAF analysis. 7 Changes in fish populations could be correlated with changes in lake amenities such as the health of the lakeside flora and fauna. 7 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV the Adirondack region to illustrate the potential magnitude of benefits on a broader scale by presenting benefits in per capita terms for the affected population. Costs and Emissions Embedded in TAF are estimates of costs and an algorithm for determining compliance activities at different facilities, developed by Argonne National Laboratory, and based on their unit inventory called GECOT. Compliance options for SO₂ reductions include scrubbing, fuel switching (including plant modifications), retirement and replacement of plants. Decisions by utilities to install retrofit desulfurization equipment (scrubbers) at 21 units for compliance in Phase I of the SO₂ trading program are taken as given. The module ranks further compliance options on a unit cost ($/ton reduction) basis, with the most-cost-effective units being implemented first, until the emission reduction requirements are satisfied. Many units are found to achieve cost savings through fuel switching and/or blending, consistent with other studies (Burtraw, 1996; Ellerman and Montero, 1996). In these cases the emission reductions are not included in the analysis of benefits because we assume the baseline (without Title IV) scenario also should reflect these emission reductions. However, we observe that the flexibility of the emission allowance trading program has allowed firms to take advantage of advantageous trends in fuel markets and to realize cost savings, while conventional regulatory approaches such as technology standards may have prohibited firms from doing so. Emission allowance trading is modeled implicitly by allocating compliance in a cost-effective way. NOx compliance is modeled to achieve emission rate reductions sufficient to meet the emission reduction goals of the program. Emission rates and costs are equivalent to low NOₓ burners absent further flexibility for compliance that characterizes the SO₂ program. This description differs somewhat from actual implementation which has allowed firms to average emission rates among commonly owned and operated facilities, and hence our estimate of costs can be viewed as conservative (high). The total cost of compliance is calculated as the present value of revenue requirements to cover compliance costs summed over all units, and this quantity is levelized (equal annual costs spread over the lesser of 35 years or the remaining life of the facility) for comparison with benefits. For SO₂ reductions, the module predicts the industry will rely on fuel switching and blending as the primary means of compliance, and that much of this switching will be implemented at low cost or cost savings to the affected firms. Scrubbing is also implemented, to a limited degree. This scenario appears robust to recent developments in the coal industry, and hence we use these estimates as a benchmark for compliance costs over the long-run. We explore the robustness of the module through scenario analysis about plant lifetimes and future electricity demand, and through comparison with other recent studies. Other Effects Not Modeled There are numerous other effects of Title IV that TAF does not model quantitatively because of a lack of proper scientific and/or economic data and models. These include effects 8 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV to material and cultural resources, nonuse of ecosystem health, recreational forests, agriculture and commercial forestry, and radiative forcing. Material and cultural resource valuation lacks a complete inventory of affected assets, data about the economic lives of affected assets, and information on behavioral responses. While nonuse values of ecosystem health are expected to be large, there is no characterization of ecosystem changes associated with Title IV or of a valuation framework for assessing benefits from improvements in ecological indicators, especially given the temporal aspects of ecological dynamics. Similarly, the link between primary pollutants and forest recreation effects that people care most about is not established. Exposure to ambient ozone is likely to be the most significant air pollutant causing significant effects on crops, but the studies examining these effects fail to account for behavioral responses in an adequate way, and the data on changes in ozone as a result of Title IV are not currently available. Lastly, atmospheric models predict changes in particulates and their effect on radiative forcing, but the economic methods for modeling damages of climate change are very uncertain, and data for valuation of local effects are not available. III. BASELINES AND SCENARIOS The analysis requires an estimate of the time path of emissions of SO₂ and NOx (plus associated abatement costs) from 1995 to 2030 in the absence of Title IV--termed the baseline--and estimates of the emissions (and costs) associated with Title IV. Subtracting the emissions for the scenario from the baseline emissions provides emissions changes (which are fed into the atmospheric transport module) to estimate benefits of Title IV. These benefit estimates are compared with costs under a consistent set of assumptions, as well as "off-line" comparisons with alternative cost estimates. We developed three baselines and picked one as the default. The baselines differ according to an estimate of plant lifetimes (60 versus 70 years) and the growth in electricity demand over the period (3 percent, termed "high growth", and 1 percent, termed "low growth"). Growth rates in electricity demand are weighted by state population growth. We think that the 70 year-low growth baseline is the most likely, but also examine the effects of a 60 year-low growth and 70 year-high growth baseline. The scenarios all involve Title IV with SO₂ trading and NOₓ reductions mandates. Specifically, the first phase of SO₂ reductions implemented in 1995 require average emission rates to be about 2.5 lb. sulfur per million Btu heat input. This rate applies to 431 units, including nearly 200 so-called "substitution and compensation" units that were voluntarily brought into Phase I to ease the cost of compliance on average. The second phase, taking effect in 2000, will lower the average emission rates to about 1.2, and will affect over 2,000 units. The first phase of NOx controls took effect in 1996 and reduced emission rates to .45 or .50 lb. per million Btu, affecting 239 units, all but 16 of which were also affected by Phase I SO₂ rules. The second phase of NOx controls expand the set of affected facilities and go into effect in 2000 and are not yet final, but are expected to take effect in 2000. Since health benefits emerge as by far the most important of the benefits we quantify, we focus several analyses toward an exploration of the sensitivity of those benefits to various 9 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV sets of assumptions. First we test the sensitivity of health benefits to two alternative baseline scenarios involving different life expectancies for power plants and different projections for growth in electricity demand. Next, under our default baseline for power plant life expectancy and growth in electricity demand, we explore four scenarios involving different assumptions involved in estimating health benefits, which we compare with our default assumptions for the health benefits case. The default case health benefits estimates resulted from our best judgment about the epidemiological and valuation literature at the time the work was completed. Our most important choice concerns the C-R functions for the mortality effects of reductions in sulfates and nitrates. For sulfates we use a weighted mean of the coefficient estimate of two benefit studies, giving both studies equal weight. This coefficient predicts the change in the number of incidents of mortality annually resulting from changes in total PM₁₀ (sulfate and nitrate) concentrations. The low estimate (0.1 percent), based on Plagiannakos and Parker (1988), assumes that sulfates are equally as potent as any PM₁₀ particle class, and estimates only daily mortality. The high estimate (0.7 percent), based on Pope et al. (1995), addresses the effects of cumulative exposure to fine particles, and probably captures much of the daily mortality risk. The high estimate implies that sulfates, which fall into the fine fraction of the particulate mass, are a relatively potent constituent of PM₁₀. We depart from some of the benefit literature (Hagler Bailly, 1995) by ignoring the higher estimates of the particulate mortality coefficient (1.4 percent) found in the Dockery et al. (1993) study because it only examines mortality effects in 6 cities and a sample of 8,111 people versus the 151 cities and 552,000 people covered by the Pope et al. (1995) study.⁸ For nitrates, we assume that they are no different in potency from any constituent of PM₁₀ based on Schwartz and Dockery (1992). Taken together, these choices imply that nitrates are, overall, less potent than sulfates, an assumption that reasonably reflects the state of the literature. For both functions we assume that there are no thresholds, meaning health benefits from emissions reductions can be expected to occur irrespective of the baseline concentration of particulates. The other key choice is the estimate of WTP for mortality risk reductions. In the base case, we use a lognormal distribution with mean of $3.1 million per statistical life (in $1990), and a 90 percent confidence interval of $1.6 and $6 million. This distribution generally accords with the valuation literature, but is somewhat on the low side because we give less weight to the labor market studies relative to the contingent valuation studies, the latter being marginally more appropriate for valuing mortality risks in the environmental health context and also capture age effects, based on Jones-Lee et al. (1985). The Jones-Lee study finds that 8 A fourth study, Evans et al. (1984), with a mid-range particulate mortality coefficient (0.3 percent) was also considered by Hagler Bailly (1995). We implicitly endorse use of this study as it falls within the range of low and high estimates we use in the TAF default case. 10 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV the value of a statistical life for the 65 years and older group is about 75 percent of that of the average (40 year old) participant in their study. As seen below, for the default case, benefits of reduced risk of health mortality alone far exceed the costs of emission reductions. For this reason, we only explore downside sensitivities of the health benefits estimates which are designed to test whether there are plausible assumptions under which benefits no longer exceed costs. We explore three options and compare them each individually to the default assumptions. Then we explore a combined case. Impose threshold for PM₁₀. In this case we assume that there is a threshold in effects at a 24-hour average concentration of 30 µg/m³ PM₁₀ (Lee, et al., 1995). Days in which the baseline concentration of PM₁₀ in a county is below this amount will not register benefits of sulfate or nitrate reductions. Treat sulfates as PM₁₀. In this case we assume that nitrates have no effect on mortality rates, in line with the lack of any direct epidemiological evidence linking nitrates with such effects. We assume that sulfates are no more potent than any other PM₁₀ constituent and use the daily mortality studies only (equivalent to applying the base case mortality assumptions for nitrates to sulfates). Mortality Risk Valuation. Even using the Jones-Lee et al. study to adjust the value of a statistical life (VSL) estimates for age probably overestimates benefits of mortality risk reductions from PM₁₀ because this study (and the rest of the VSL literature) provides estimates for reducing risks of accidental and immediate death, such as in a car accident. Particulate matter exposure, on the other hand, may lead to higher probabilities of death for individuals only when they are already quite old. For most of the population, then, the mortality benefits of today's PM₁₀ reduction may be zero or very small. It may contribute to a higher probability of developing chronic respiratory disease which, in turn, may reduce life expectancy. Said another way, the WTP for a risk reduction realized in the future is likely to be much lower if one has to pay today versus in the future. Unfortunately, we cannot take this effect into account directly in the sensitivity analysis. Instead we use an approach to adjust the VSL downwards, based on life-years remaining, that probably provides a lower bound to the VSL. 9-10 9 This age disaggregated estimate, which is based on a procedure that assumes each year of life is worth the default VSL divided by the life expectancy of a 40 year old, and that those over 65 are willing only to pay by the year for the number of years they can be expected to live, results in the assignment of a VSL to the over 65 population of $0.9 million, about 1/3 that of the under 65 population, for which the VSL is assumed to be the default. 10 It is worth noting that there are additional reasons why WTP estimates in the "auto-death"-type context may over or underestimate risks in the PM-mortality context. The former may overestimate the latter if the older people at risk from PM have compromised health. The former may underestimate the latter if air pollution is thought to be an involuntary risk and auto-death risk is thought to be voluntary. 11 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Combined Case. In this case we assume that there is a 30 µg/m³ PM₁₀ threshold and that sulfates only have the potency of the average PM₁₀ particle. The Hagler Bailly Study of Health Benefits from Sulfate Reductions EPA commissioned a study by Hagler Bailly (HB, 1995) of the health benefits of reductions in emissions associated with Title IV reductions in SO₂ emissions. This study began with estimates of emissions changes made by ICF Resources using their Coal and Electric Utilities Model (CEUM) on behalf of the EPA's Acid Rain Division. The EPA tied those changes in emissions to their Regional Acid Deposition Model (RADM) to obtain changes in sulfate (fine particle) concentrations in the Eastern U.S., and used a particular set of health concentration-response and valuation functions to estimate the monetized health benefits of the emissions changes. The HB study places an expected value of a statistical life at $3.2 million (1990$) compared to an expected value of $3.1 in TAF. Expected benefits are higher from the HB study than under our default health benefit estimates. To reconcile the differences between these results, we incorporated the HB study into TAF to compare results for 2010. Since EPA population estimates could not be easily obtained, TAF population projections were used in all scenarios. 11 We calculated mortality benefits per ton for SO₂ emission reduction. We did not have access to RADM for direct comparison with ASTRAP in TAF. However, by explicitly comparing the state-aggregated emission reductions forecast by TAF and HB, and the predicted health effects and valuation functions, we were able to impute the influence that the different atmospheric models had on benefit estimates. The EPA's Regulatory Impact Assessment for Particulates As part of the Regulatory Impact Assessment (RIA) for the new proposed standards for particulates, the EPA has developed new health effect and valuation functions. The RIA examines mortality effects from PM₂.5 (which includes both sulfates and nitrates), using a value of a statistical life of $4.8 million. We incorporate, these new functions as an option in TAF and compare them with our default assumptions. Alternative Cost and Emission Estimates To examine the sensitivity of our findings to changes in costs, we compare our default cost estimates for SO₂ reductions with several others from the literature including White et al. (1995), compiled for the Electric Power Research Institute, ICF (1995), compiled for the EPA, and GAO (1994). Burtraw et al. (1997) provides econometric estimates of short-run 11 HB used different population estimates than are modeled in TAF. The estimates for 2010 differ from TAF both by state and nationally, with average difference leaving HB estimate about 3.6 percent above that of TAF, though most of the difference is in areas without a large change in SO4 in the benefits module. Hence, their benefits estimate increase only 1.5 percent as a result of population differences. We adjust HB's population estimates to fit ours; i.e., all results presented in this paper for HB reflect a small downward population adjustment. 12 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV and long-run costs. We compare TAF default estimates of the cost of NOx control with an ICF (1996) and E.H. Pechan (1996). With respect to emissions, our default estimates lead to emission estimates that are proximate to empirical measures based on the first two years of the program. In reporting our results, we focus primarily on long-term estimates for the year 2010 and beyond, when the program will be in full swing. To consider an alternative emission scenario and its effects on benefits, we compare the Argonne model with emissions and forecast by HB. IV. RESULTS Figure 1 and Table 2 summarize the mean expected costs and benefits in per capita terms for the included benefit pathways, for our main run of 50 realizations of the Monte Carlo simulation model. We emphasize that the exact same results will not obtain in running the model two different times because of random aspects of the sampling procedure, and because we may vary the choice of sampling procedure. The virtue of this approach is that it avoids a false sense of precision in the estimates, and allows us to focus on the likely distribution of outcomes and identify qualitative results. Table 2: Per Capita Benefits in 2010 for Affected Population. Effect Benefits per Capita (1990$) Morbidity 3.50 Mortality 59.29 Aquatic 0.62 Rec. Visibility 3.34 Resid. Visibility 5.81 Costs 5.30 Estimates in Table 2 are projected for the year 2010, when the second phase of the SO₂ program and the NOx programs are expected to be in full effect. (Note the vertical axis in Figure 1 is a log scale.) The dominant source of benefits is reduced human mortality risk, and taken singularly it results in a mean benefit estimate in 2010 that is nearly an order of magnitude greater than costs. Expected benefits from human morbidity, recreational visibility and residential visibility each individually are approximately equal to the annualized expected cost per capita in 2010. Health and recreational visibility benefits are presented as the average per capita benefits for all U.S. residents. Recreational visibility represents an estimate of average willingness-to- pay for modeled visibility improvements at just two parks--Grand Canyon and Shenandoah. Although there would be improvements at other park locations, problems of embedding benefit endpoints in the application of contingent valuation techniques to estimation of nonuse benefits suggest that measures of WTP at other locations would not be additive to these, and indeed we 13 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV may capture most of the WTP for improvements across the entire nation with these locations (Chestnut and Rowe, 1990). Costs and Benefits per Capita $100.00 $10.00 Mortality Resid. Visibility Costs Rec. Visibility Morbidity $1.00 Rec. Lake Fishing $0.10 1995 2000 2005 2010 2015 2020 2025 2030 Year Figure 1: Costs and Benefits for Modeled Pathways for Affected Populations (Log-scale) A virtue of a per capita comparison is that we can include benefit pathways that are not modeled for the entire U.S. The residential visibility benefits are those benefits that obtain for all residents in the five modeled cities of Washington, Atlantic City, Knoxville, Charlottesville, and Albany. The aquatic benefits are those that obtain for the portion of the population that is engaged in recreational fishing in Adirondack lakes. We express these benefits in per capita terms for each affected population in order to obtain a measure of the potential magnitude of such benefits at a national level. In the case of residential visibility, an extrapolation to the national level would likely overstate benefits because changes in sulfate and nitrate concentrations would be less in other parts of the country. In the case of aquatic effects, an extrapolation to the national level also would likely overstate benefits because a large portion of the population does not pursue recreational fishing, and again because the changes in lake chemistry would be less in most, if not all other parts of the country. A potential point of confusion is the measure of tons reduced under the program, which depends importantly on the characterization of what would have happened to emissions in the absence of the program. To avoid confusion over the baseline emissions (we discuss the issue again below), we draw attention to benefits calculated per ton of emission reductions. The location of emission reductions still matters importantly to the calculation of benefits per ton, and this is modeled explicitly in TAF. Measured in this way, health still plays a dominant role in the assessment of benefits. Median mortality benefits for the entire U.S. per ton SO₂ reduction under TAF's default scenario are $3,102. We find the 90 percent confidence interval around TAF's reference case estimate for SO₂ mortality benefits to range from $1743 to $9,649. The median value of human morbidity effects for TAF are $193 per ton of SO₂ 14 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV reduction. The median estimate of benefits resulting from changes in NOx emissions in 2010 are $463 per ton for mortality (through the change in nitrate concentrations) and $137 for morbidity. These do not include the effects from changes in ozone concentrations. In contrast, annualized costs in 2010 are estimated to average $271 per ton of SO₂ emission reduction, and $382 per ton for NOx emission reduction. Figure 2 displays the benefit-cost ratio, using health benefits only, under alternative baseline assumptions about plant lifetimes and growth in electricity demand. In the year 2020, when the benefit-cost ratio is greatest, annual costs (1990$) are $1.56 billion in the default case (low growth-70 year retirements), while benefits are $19.9 billion. Costs drop to $1.19 billion in the low growth-60 year retirements case, while benefits drop to $13.2 billion. Costs are $1.63 billion in the high growth-70 year retirements case, while benefits are $21 billion. For the high growth-70 year retirement case, the benefit-cost ratio of Title IV is even larger than for the default case. The reasons are that changes in assumptions affect both benefits and costs in the same direction but to differing degrees. For instance, lower growth in electricity demand implies that there is a lower opportunity cost to retiring older plants. It also suggests that emissions in the baseline would be lower, and hence emission reductions and program benefits would be lower. It is also interesting that benefits in the low growth-60 year retirement case are less than or equal to the default case in every year. The benefit-cost ratios are not much different among the cases. Perhaps the most interesting aspect of this figure is that the benefit-cost ratio does not vary by a huge amount under the different assumptions, even though the measure of benefits or the measure of costs taken separately does vary significantly. This points out a virtue of TAF in that it allows us to explore benefits and costs under a consistent set of assumptions and sensitivity cases. Because of the dominant role of health, we devote a considerable part of our sensitivity analysis to whether the mortality and morbidity benefit estimates are robust. Figure 3 displays the annual health benefits alone for the default scenario, with associated uncertainty bars, in comparison with our default annualized expected cost estimates, in millions of dollars. Annualized costs for SO₂ and NOx reductions are about $761 million per year in 1995, increasing to $1.51 billion in 2000 and $1.56 billion in 2020. Expected benefits in the default scenario rise from $5.1 billion in 1995 to $19.9 billion in 2020, dropping back to $15.5 billion by 2030. The ramp up of benefits is attributable to meeting Title IV year 2000 goals as well as to population and income growth, while the drop after 2020 is attributable to plant retirements that occur in the baseline. Our main observation in Figure 3 is that the uncertainty bounds around the benefit estimates show that there is no year in which benefits (at the 5 percent confidence level) are less than expected annualized costs. Uncertainty in the cost estimates is explored through the three scenarios involving alternative assumptions about plant lifetimes and electricity demand growth described in Figure 2. However, these alternatives generate such a small range in costs, compared to uncertainty in benefits, that it does not display in Figure 3. About 94 percent of total health benefits result from mortality benefits in 2010. Only about 11 percent of total 15 Benefit (Health only): Cost Ratio Under Alternative Assumptions about Electricity Industry 15.0 13.0 11.0 70 high 9.0 70 low 60 low 7.0 5.0 3.0 1995 2000 2005 2010 2015 2020 2025 2030 Year Figure 2: Benefit-cost ratio for health benefits under alternative assumptions about electricity industry 16 45,000 40,000 35,000 30,000 25,000 Benefits 20,000 Cost 15,000 10,000 5,000 1995 2000 2005 2010 2015 2020 2025 2030 Year Figure 3: Annual mean total health benefits with 90% confidence intervals compared with expected annualized costs 17 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV benefits are attributable to NOₓ reductions (the rest are attributable to sulfate reductions). Of morbidity benefits, NOₓ reductions account for closer to 27 percent of the benefits, according to our analysis. Figure 4 reports the annual mean total health benefits and annualized costs for the TAF default case, the HB case for sulfates only, each of the three separate sensitivity analyses which are designed to reduce benefits, and for the combination case. The three separate analyses do not eliminate the gap between expected benefits and levelized costs in any year when taken separately. The most dramatic reduction in benefits occurs when we use very conservative assumptions to value statistical life. For instance, in 2020, expected total health benefits are $19.9 billion in the default scenario, but only $5.4 billion in this sensitivity case. Uncertainty analysis on the three sensitivity cases reveals that 5th percentile benefits are less than annualized costs up to, but not beyond, either 2000 or 2005, depending on the case, but that in no case do total costs exceed total health benefits. Only when we do a combined sensitivity analysis--where we assume that sulfates affect mortality with the potency of the average component of PM₁₀ and that there is an effects threshold of 30 µg/m³ PM₁₀--do we find total expected costs proximate to total expected benefits, though it is still less than benefits. Annual Mean Total Health Benefits and Annualized Costs by Scenario 100,000 10,000 HB (Sulfates) Default Threshold PM10 Age Disag Cost PM10 and Threshold 1,000 100 1995 2000 2005 2010 2015 2020 2025 2030 Year Figure 4: Annual mean total health benefits and levelized costs (log scale) by scenario Legislative debates about acid rain in the 1980s had a sharp regional character. Since acid deposition typically occurs far from the source of emissions, which were largely concentrated in the Ohio Valley, many observers claimed that emissions from these power plants were contributing to environmental degradation in the Northeast. The regional decomposition of health benefits from reduced emissions is less parochial because atmospheric concentrations are 18 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV affected closer to the source of emissions. Table 3 illustrates that, expressed in per capita terms, the greatest health benefits accrue to the regions with the greatest changes in emissions. Table 3: Expected total health benefits for 2010 and percent of national SO₂ emission reductions by state Per Capita Health Benefits Percent of National Percent of National SO₂ State (1990$) Health Benefits Emission Reductions WV $171.38 1.8% 12.0% OH $159.85 10.2% 23.3% DC $158.81 0.5% 0.0% PA $157.84 11.0% 9.8% KY $148,29 3.3% 11.0% VA $135.41 5.5% 0.4% MD $131.77 4.1% 0.4% IN $131.46 4.4% 16.0% DE $131.21 0.6% 0.0% NJ $130.70 6.2% 0.4% NY $115.42 11.6% 2.2% Other $36.93 40.8% 24.5% Comparison with HB study Table 4 provides sulfate mortality benefit estimates comparing the HB model and our default values in TAF. This comparison was obtained through a different run of the model than the results reported previously and consequently the results for our default assumptions vary due to use of a different sample drawn with a different sampling procedure. A common sample was drawn for all examples in this comparison, however. Because we use identical census population projections for both estimates, there are three margins along which TAF and HB estimates may differ: (i) the quantities and locations of emissions changes differ; (ii) the "source-receptor matrices" linking emissions to concentrations over space differ; and (iii) the concentration-mortality risk estimates and the estimates of the value of a statistical life differ. Each scenario in Table 4 is identified by the source for emission changes (EPA for the HB study, or TAF's default values), the atmospheric model (RADM for HB, or ASTRAP for TAF) and the health effects and valuation functions (HB or TAF). 19 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Table 4: Comparison of HB and TAF Mortality Sulfate Benefits (billions $1990) for Eastern U.S. with percent changes over previous scenario, Year 2010 New Health Scenario: EPA/HB New Transport New Emissions TAF Emissions EPA EPA TAF TAF Transport RADM ASTRAP-TAF ASTRAP-TAF ASTRAP-TAF Model Health/ HB HB HB TAF Valuation Mean Benefits 30.3 56.9 25.1 15.4 (billion $) Percent +87.8% -55.9% -38.6% Change Benefits ($) 3,289 6,179 6,296 3,852 per ton Median Benefits (billion $) 19.0 34.5 15.3 12.64 Percent +81.6% -55.7% -17.4% Change Benefits ($) 2,061 3,749 3,828 3,153 per ton Expected (mean) mortality benefits are higher from HB than TAF. Before any adjustments, HB found sulfate mortality health benefits of $31 billion in 2010 (1990$) in the Eastern U.S. while TAF estimates benefits of $15.4 billion in this region. 12 We reconcile differences in population estimates by using TAF's estimates in both scenarios. Under these assumptions, adjusted HB estimates are $30.3 billion as reported in Table 4. Although we do not focus on uncertainty in this reconciliation, we note that the adjusted HB estimates range from $5 to $67 billion for the 20th and 80th percentiles around the mean. TAF has tighter uncertainty bands, at $7.6 to $24 billion for the same confidence interval. This uncertainty difference is driven primarily by our use of a narrower range of PM-mortality studies than those used by Hagler Bailly. 12 When including morbidity, total health benefits for HB are $35.5 billion, and for TAF they are $17.8 billion. Benefits in the Eastern U.S. make up the 98% of the benefits in TAF for the entire U.S. Also note that in using RADM for atmospheric modeling, HB is using the median of several runs of the model rather than the mean. 20 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV To reveal the effect of each of these differences in underlying assumptions, we begin with the HB analysis, and gradually replace HB assumptions with default assumptions in TAF. We begin by substituting the ASTRAP source-receptor coefficients used in the TAF analysis for RADM, which results in a large (87.8 percent) increase in the benefit estimates (from $30.3 billion to $56.9 billion). Viewing the results state-by-state explains what is happening. The HB version with RADM shows a highly variable pattern of benefits, ranging from $360 per capita in Tennessee to $13 per capita in Minnesota. The highest benefits are in the south (particularly the southeast and southern Appalachians); Mid-Atlantic and New England states have very low benefits per capita. In contrast, the TAF-ASTRAP version yields larger benefits that are also less variable. Further, ASTRAP shows much larger effects in the northeast, followed by the Mid-Atlantic states, and then some southern states, with smaller benefits in several others. These differences translate to per capita benefits with ASTRAP that are highest in the Mid-Atlantic, about equal in the Midwest and Northeast, and smallest in the South. What could account for such differences between ASTRAP and RADM? Shannon, et al. (1997) found the two models' predictions reasonably in agreement for predicting atmospheric sulfate concentrations in the eastern U.S., though RADM actually predicts greater sulfate reductions in the more populated regions including the mid-Atlantic. Weather patterns appear to be handled differently by RADM and ASTRAP in a way that could account for much of the difference in benefits. In the HB application, the median episode taken over 30 episodes is used rather than a weighted average of episodes. In contrast, ASTRAP uses 11 years of daily meteorology to develop its source-receptor (S-R) matrices, which are constructed to represent average meteorology for each season. Given the lognormal distribution of meteorology, the median could be far below the mean. Substituting EPA emissions forecasts with the TAF emissions forecasts decreases mortality benefits (which drop 56 percent from $56.9 billion to $25.1 billion). Although approximately equal average annual emissions should obtain in the long run, the EPA forecast suggests a higher baseline level of emissions and hence greater emission reductions under the program. EPA's higher baseline projects fewer units switching to coals with lower sulfur content than does the TAF model. We complete the reconciliation by substituting TAF mortality coefficients and values of a statistical life for those in HB (recall that we are only considering mortality effects). This switch decreases our mean benefit estimate by 38.6 percent (from $25.1 to $15.4 billion). This change is primarily a consequence of the inclusion by HB of the Dockery et al. (1993) "six city" study relating annual PM2.5 concentrations to the probability an individual in the cities will die during the study period. While HB assigns this study a weight of 25 percent, we give it no weight because it is dominated, in our opinion, by the Pope et al. (1995) study, which uses a similar approach but is applied to 552,000 individuals over 151 cities. HB also uses somewhat higher values of a statistical life than we do (they use an estimate with a expected value of $3.2 million, the expected value of the TAF estimate is $3.1 million). 21 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Comparison with the RIA For further analysis of heath effects we substitute coefficients from the EPA's draft Regulatory Impact Assessment (RIA) for particulates into the health effects and health valuation modules. Compared with the EPA/HB analysis in Table 4, which reported mean annual mortality benefits from sulfate reductions in 2010 of $30.3 billion in the Eastern U.S., the RIA (using EPA emissions and RADM for atmospheric modeling) approach yields $25.6 billion. The RIA uses a higher value of a statistical life ($4.8 million) than HB, but predicts a smaller change in mortality for the same change in sulfate concentrations, despite including long-term mortality effects. The EPA/RADM/RIA analysis estimates are still larger than the mean estimates for TAF of $15.4 billion. When we substitute the RIA for TAF in measuring health effects and valuation (TAF/ASTRAP/RIA), the expected benefits fall to $21.3 billion. Cost Comparisons The costs of SO₂ reductions under Title IV have attracted considerable attention because of the innovative allowance trading program. Cost projections from the middle 1980s based on command and control approaches, and projections of marginal costs under a market with an inadequate level of trading, ranged as high as $1500 per ton (Bohi and Burtraw, 1997). At the time of Title IV's enactment the EPA projected costs in 2010 of $450- $620 per ton (ICF, 1990). Cost estimates have continued to decline, in large part because the program gives utilities the flexibility to exploit advantageous trends in coal markets and the cost of rail transport that have led to a drop in the cost of switching to lower sulfur coal. Table 5 reports a series of estimates for average costs (which are expected to be lower than marginal costs in Phase II), illustrating that various projections have continued to decrease as allowance trading has taken hold. Nonetheless, the TAF default costs are on the low end of this range. The ICF (1995) estimates are the final in a series of declining estimates provided for the EPA by ICF since 1989. ICF (1995) estimates were reported in the EPA's Regulatory Impact Assessment for Title IV. These estimates describe a considerably greater emissions reduction because of higher projected emissions in the baseline than assumed in TAF. The greater annual costs spread over greater emission reductions yield comparable average costs. It makes sense that the average costs per ton are greater in the TAF estimates since it assumes more switching to low sulfur coal for economic reasons in the baseline; a greater portion of this switching is accounted for as part of Title IV by ICF and this brings down the average cost per ton in that study. Based on recent econometric estimates (Burtraw et al., 1997) and the recent trend in fuel markets, and also due to current trends toward increasing competition in the electric utility industry, we believe the TAF estimates can be taken as central estimates. ICF (1995) suggests annual costs about 2.5 times those included in the TAF default case; however, the estimated cost per ton reduction is just about equal to that for TAF. 22 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Table 5: Long run (Phase II, year 2010) cost estimates for SO2 reduction Annual Cost Average Cost per Ton SO2 Study (billion 1990 dollars) (1990 dollars) TAF Default 0.8 205 ICF (1995) 2.0 216 White, et al. (1995) 1.2-2.5 245-286 GAO (1994) 1.8-2.9 197-320 Van Horn Consulting et al. (1993) 2.0-3.2 289-373 Other reported estimates include White et al. (1995) and Van Horn Consulting et al. (1993) which were compiled for the Electric Power Research Institute. The range of estimates in White et al. is associated with the level of plant utilization, comparable to TAF's low and high electricity demand cases. Van Horn Consulting was also the contractor for the GAO (1994) estimates. The range of estimates for GAO pertain to variations in the liquidity of the allowance market, and the range in the Van Horn Consulting estimates cover a mix of scenarios. Another aspect of regulatory costs that has only recently been investigated and estimated is the hidden social cost of imposing additional regulations in a second-best setting characterized by pre-existing regulations and taxes that already distort the economy away from economic efficiency. This issue has ignited colorful debate with respect to policies to address climate change. Goulder et al. (1997) addressed this issue in an analytical and computable general equilibrium model of the SO₂ program to estimate hidden social costs due to the second-best setting for Title IV. They estimated that the social costs stemming from interactions between the trading program and pre-existing taxes in the economy were $533 million per year. This social cost stems from the fact that the SO₂ program, like any regulation, imposes a cost that reduces the real wage of workers. This cost can be viewed as a virtual tax, and when imposed on top of pre-existing taxes, has large consequences for economic efficiency. Unfortunately, as far as this issue is concerned, the SO₂ trading program imposes particularly large costs because it encourages firms to internalize not only their abatement costs, but also the cost of residual emissions through the opportunity cost of SO₂ allowances. Were the program to raise revenues through the auction of permits, it could use these revenues to offset this tax-interaction effect by reducing other distortionary taxes. However, the SO₂ allowances are allocated without charge, so there is no revenue available for this purpose, and consequently the tax-interaction effect and resulting social cost is substantial. Table 6 reports alternative cost estimates for the NOx portion of Title IV. The E.H. Pechan (1996) estimate may be high because it reflects average costs for 3.7 million tons per year in NOx emission reductions. This is greater than the other estimates, and reflect reductions as a result of Title IV requirements coupled with requirements on electric utilities 23 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV stemming from other parts of the Clean Air Act Amendments. Also, like the TAF estimates, E.H. Pechan does not allow for averaging of emission rates among commonly owned and operated facilities, which is a feature in the actual implementation of the regulation. In contrast, the TAF emission reductions are 1.97 million tons per year and the ICF (1996) reductions are forecast to be 2.06 million tons per year, both reflecting only the specific requirements of Title IV. Among these three sets of estimates, only ICF reflects the reduction in costs that can be expected through emissions averaging, which helps explain why it is lower than the others. Table 6: Long run (Phase II, year 2010) cost estimates for NOx reduction Annual Cost Average Cost per Ton NOx Study (billion 1990 dollars) (1990 dollars) TAF Default 0.8 382 ICF (1996) 0.5 229 E.H. Pechan (1996) 1.6 438 Considering the alternative cost estimates, and also recognizing that costs stemming from the second-best setting of environmental regulation are excluded, we argue that TAF's more conservative default estimate is a reasonable midpoint. We feel especially justified in this view because of the apparent magnitude of benefits compared to costs. If one were to double TAF's estimate, this difference would have an important effect on the benefit-cost comparisons illustrated in our previous examples; however, it would not by itself change the qualitative finding that benefits appear to outweigh costs by a significant margin. Unmodeled Pathways and Research Priorities To varying degrees, members of the team of scientists and economists that contributed to construction of TAF initiated review and modeling of environmental pathways that were not part of our quantitative analysis. Based in part on these efforts, we have constructed a qualitative review of pathways that are not modeled, including a relative ranking of their expected magnitude, and a prioritization for further research according to our assessment of the value of additional information for each. This evaluation is reported in Table 7. Short run and long run research needs vary among the modeled and unmodeled pathways. Estimates of health and visibility benefits remain uncertain; however, the cost of reducing uncertainty appears to be relatively lessithan many other areas. To evaluate Title IV on the basis of a comparison of benefits and costs, it may be sufficient to focus efforts at assessing benefits from health and visibility, because these benefits alone appear to outweigh the costs. Environmental areas including aquatics and forests stand to benefit in addition. 24 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Table 7: Qualitative evaluation of expected benefits and value of additional information for modeled and nonmodeled pathways Categories Expected Benefit: Value of Additional Information: high high-mid Are expected benefits large? With the goal of improving mid benefit estimates, what is low-mid the relative short-term low return on investment? Health: Mortality Health: Morbidity Visibility Materials and Cultural Resources Nonuse Values: Ecosystem Health Aquatics: Recreation Forests: Recreation Agriculture and Commercial Forestry Radiative Forcing While there are many issues facing the health scientists and epidemiologists, economists should work to improve the basis for the valuation of small risks to mortality health due to environmental changes. Also, economists need to develop estimates for WTP to avoid these risks that depend on the age and health status of the affected individual. For visibility, valuation needs to be more precise with respect to the endpoints that are important for assessment of benefits, and particular attention should be paid to the nature of preferences for changes in visibility, such as the trade-off between changes in the mean and extreme values of visual range. Benefits to materials and cultural resources is another area where benefits may be sizable. Rapid progress could be made through further work on the valuation of cultural resources, which should concentrate on the identification of the resources and the attributes of those resources that are meaningful endpoints to individuals. Assessment of benefits to commercial materials requires an improved inventory of affected materials, and improved estimates of their economically useful lives. Over the longer time frame, we suggest assessment of nonuse values for ecosystem health should be afforded high priority. However, a research emphasis in this area would require sustained levels of funding over several years to yield results that would be reliable. 25 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Also, agriculture and commercial forestry would receive a somewhat higher ranking in Table 7 were a sustained level of funding to be committed. One reason sustained funding over time would be necessary is that agriculture is undergoing structural change due to reforms passed by Congress in 1996 that may not be fully attained until the next decade. In addition, estimating rural ozone effects may be costly and time consuming, though such modeling would also contribute to an understanding of human health benefits and forest recreation. The most important of the uncertainties and omissions in this analysis are summarized in Table 8 which appears as an appendix to this paper. In this table we indicate our qualitative assessment of the direction of the bias for each of these shortcomings. A plus sign (+) indicates the uncertainty or omission imposes an upward bias in our benefit or cost estimate; a negative sign (-) indicates otherwise. Additional information about the uncertainties and limitations at each step in the TAF model is provided in the documentation for TAF (Bloyd et al., 1996). V. CONCLUSIONS Although important limitations, caveats, and major uncertainties inhibit the comprehensiveness of this benefit-cost analysis, the clear conclusion that emerges from the array of scenarios we explore is that the benefits of Title IV exceed the costs by a substantial margin. This assessment differs from the information that was available to policy makers at the time the program was enacted in 1990. At that time, Portney (1990) ventured to offer a comprehensive assessment of the Clean Air Act Amendments. Portney wrote that the expected benefits and costs appeared to be about equal for Title IV, in part because of the cost savings that were expected to result from the innovative allowance trading program. Since that time it appears that costs have fallen significantly compared to prior expectations, and benefits are now thought to be greater than expected. Expected benefits tend to be high in some areas that were not a primary focus of benefits assessment in the 1980s, particularly health and visibility. The dominant category of benefits is mortality, which we expect to be several times the costs of the program. We find mortality values that are less than previous estimates for the EPA. Still, in our analysis there is no year in which health benefits alone at the 5 percent confidence level are less than the levelized expected costs. About 89 percent of the total health benefits are attributable to changes in SO₂ and 11 percent attributable to changes in NOx emissions. We emphasize that there are tremendous uncertainties in measuring and valuing mortality. Recent economic critiques have argued that the use of the value of a statistical life as the basis for valuing health risks from air pollution, instead of a more appropriate measure of quality adjusted life years lost, could grossly overestimate mortality benefits. In addition, economists have questioned the appropriateness of using labor studies of prime age men to value changes in life expectancy that occur among an older population. In the future we expect these critiques to gain in credibility as more is learned about how to measure benefits. On the other hand, we note that because environmental exposures are involuntary, compared with studies of labor market behavior, the latter may underestimate willingness to pay to avoid environmental exposures. 26 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Morbidity, recreational visibility, and residential visibility benefits each separately appear to measure at comparable magnitude with costs. About 73 percent of morbidity effects are attributable to SO₂, and 27 percent are attributable to NOx. (We do not model the contribution of NOₓ to ozone.) The visibility estimates illustrate their potential magnitude, but we note that the literature is narrow and should be subject to closer scrutiny. Public attention in the 1980s to air pollution from SO₂ and NOx emissions largely centered on the problem of acidification ("acid rain"), with particular concern for its affect on water and soil chemistry and ultimately ecological systems. It is surprising to many that relatively low benefits are estimated by economists for effects on aquatics (in this study) or are expected to result from effects on forests and agriculture. One reason is that willingness to pay for environmental improvement depends on the availability of substitute assets. Economists would not expect changes in quality at one site to elicit large benefits if there are many sites available for comparable recreational opportunities. In contrast, individuals do not have the same kind of substitution possibilities with respect to health and visibility, which may help explain the relatively larger benefit estimates for these endpoints. Furthermore, one should note that the low values for aquatics stem from an assessment of use values, or commodity values in the case of agriculture. Environmental changes may also yield nonuse values, and estimates for nonuse values are not available. Nonetheless, the evidence, based on a small number of relatively narrow studies, suggests these values may be significant. The costs of compliance under Title IV have attracted attention because of the innovative allowance trading program. Many recent estimates find costs to be lower than anticipated for SO₂ emission reductions, in large part because of the flexibility the program gives firms to find least-cost ways to reduce emissions and to take advantage of advantageous trends in fuel and factor markets. Nonetheless, the TAF default costs are on the low end of previous estimates for SO₂ and somewhat high for NOx control, and they do not take into account hidden social costs stemming from the second-best setting for environmental policy. These factors impart uncertainty around estimates of costs in this study. The strength of this analysis using TAF is the flexibility it gives us to explore uncertainties in the measurement of benefits and costs, and to employ consistent assumptions in the comparison of benefits and costs. We acknowledge important gaps and uncertainties in this analysis. Nonetheless, in spite of, and in some cases because of these important caveats, our exploration of the relevant uncertainties leads us to find compelling evidence that benefits of Title IV substantially exceed costs. 27 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Table 8: Major uncertainties and omissions and direction of bias in TAF Uncertainties and Omissions Bias Description BENEFITS AGGREGATION TO STATE LEVEL ? Emissions, atmospheric transport and effects are modeled at state level. Probability distributions are used to represent variability within states in the simulations. ATMOSPHERIC MODEL DOES NOT + Ammonia may be a limiting factor in formation of CAPTURE ROLE OF AMMONIA secondary particulates. Reductions in one (e.g. sulfates) may allow increases in the other (e.g. nitrates). AQUATIC EFFECTS CAPTURE ONLY - The measure does not capture effects on other LIMITED RECREATIONAL USE, recreational uses. ONLY AT LAKES AQUATIC EFFECTS LIMITED TO +/? The Adirondack region has high participation rates ADIRONDACKS compared to nation. Calculation of effects on "per affected capita" basis yields inflated values when extrapolated to nation. RECREATIONAL VISIBILITY +/? Only two parks included, but this may capture majority of benefits. Contingent valuation methods uncertain. Valuation is not precise with respect to the distribution of visibility improvements over time. RESIDENTIAL VISIBILITY ? Only five cities evaluated; benefits represented on "affected per capita" basis. MORBIDITY MEASURES - Reduced workplace productivity for small effects not captured. MORTALITY COEFFICIENT + Use of mortality coefficients treats all mortality effects equally. A preferable approach would be life-years lost. VALUE OF STATISTICAL LIFE +/? The VOSL approach does not value appropriately small changes in life expectancy realized late in life (+). Health status is not included. (+) However, VOSL ignores involuntary nature of exposure (-). OMITTED ENVIRONMENTAL - Magnitude of use values for omitted pathways may ENDPOINTS AND NONUSE VALUES be small as indicated by included aquatic endpoint. LISTED IN TABLE 6 However, nonuse measures are not explored and may be significant. BENEFITS OUTSIDE U.S. EXCLUDED - The analysis is limited to the continental U.S. COSTS SO₂ PROGRAM MODELED AS PERFECT - Regulatory incentives may hamper allowance TRADING trading. 28 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Table 8: Major uncertainties and omissions and direction of bias in TAF (continued) Uncertainties and Omissions Bias Description NOx ABATEMENT MODEL DOES NOT + Implementation of NOx rules allows emission rate REFLECT EMISSIONS AVERAGING averaging among commonly owned and operated units which lowers costs. ELECTRICITY DEMAND GROWTH - Previous analysis has indicated electricity demand growth and plant lifetime to be the most important variables in costs. Both variables explored in sensitivity analysis. Our analysis is conservative (low) on projected demand growth. PLANT LIFETIME ? Plant lifetime is treated parametrically. PARTIAL EQUILIBRIUM ANALYSIS - General equilibrium effects indicate hidden efficiency costs from regulations that raise product costs. Also, failure of program to raise revenue. 29 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV REFERENCES Bloyd, Cary, et al. 1996. Tracking and Analysis Framework (TAF) Model Documentation and User's Guide, ANL/DIS/TM-36, Argonne National Laboratory (December). Bohi, Douglas R., and Dallas Burtraw. 1997. "SO₂ Allowance Trading: How Do Expectations and Experience Measure Up?" Electricity Journal, (August/September) forthcoming. Brookshire, D., et al. 1979. Methods Development for Assessing Air Pollution Control Benefits. Volume 2: Experiments in Valuing Nonmarket Goods. A Case Study of Alternative Benefit Measures of Air Pollution in the South Coast Air Basin of Southern California, EPA-600/6-79-0016, U.S. Environmental Protection Agency, Washington, D.C. (February). Burtraw, Dallas, Curtis Carlson, Maureen Cropper, and Karen Palmer. 1997. "SO₂ Control by Electric Utilities: What are the Gains from Trade?" unpublished mimeo, Resources for the Future, Washington, D.C. Burtraw, Dallas. 1996. "The SO₂ Emissions Trading Program: Cost Savings Without Allowance Trades," Contemporary Economic Policy, 14 (April), pp. 79-94. Chestnut, L. G., and R. D. Rowe. 1990. Preservation Values for Visibility Protection at the National Parks (Research Triangle, N.C., U.S. Environmental Protection Agency). Dockery, D. W., F. E. Speizer, and D. O. Stram. 1989. "Effects of Inhalable Particles on Respiratory Health of Children," American Review of Respiratory Diseases, 139, pp. 587- 594. Dockery, D. W., C. A. Pope III, X. Xu, John D. Spengler, James H. Ware, Martha E. Fay, Benjamin G. Ferris, Jr., and Frank E. Speizer. 1993. "An Association between Air Pollution and Mortality in Six U.S. Cities," The New England Journal of Medicine, vol. 329, no. 24, pp. 1753-1759. E. H. Pechan and Associates. 1996. "Emission Reduction and Cost Analysis Model for NOx (ERCAM-NOX)," Report No. 96.09.002/1763, September. Ellerman, A. Denny, and Juan-Pablo Montero. 1996. "Why are Allowance Prices so Low? An Analysis of the SO2 Emissions Trading Program," MIT-CEEPR Working paper 96- 001 (February), Massachusetts Institute of Technology. Evans, J. S., T. Tosteson, and P. L. Kinney. 1984. "Cross-Sectional Mortality Studies and Air Pollution Risk Assessment," Environment International, 10, pp. 55-83. Goulder, Lawrence H., Ian W.H. Parry and Dallas Burtraw. 1997. "Revenue-Raising VS. Other Approaches to Environmental Protection: The Critical Significance of Pre-Existing Tax Distortions," RAND Journal of Economics, forthcoming. 30 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV Government Accounting Office (GAO). 1994. "Air Pollution: Allowance Trading Offers an Opportunity to Reduce Emissions at Less Cost," GAO/RCED-95-30, Washington, D.C. Hagler Bailly. 1995. Human Health Benefits Assessment of the Acid Rain Provisions of the 1990 Clean Air Act Amendments. Final report prepared by Hagler Bailly Consulting, Inc., Boulder, Colorado, under subcontract to ICF Incorporated, Fairfax, Virginia, for U.S. Environmental Protection Agency, Acid Rain Division. ICF. 1990. "Comparison of the Economic Impacts of the Acid Rain Provisions of the Senate Bill (S.1630) and the House Bill (S.1630)," prepared for the U.S. Environmental Protection Agency (July). ICF. 1995. "Economic Analysis of Title IV Requirements of the 1990 Clean Air Act Amendments," prepared for the U.S. Environmental Protection Agency (September). ICF. 1996. "Regulatory Impact Analysis of NOₓ Regulations," prepared for the U.S. Environmental Protection Agency (October 24). Jones-Lee, M. W., M. Hammerton, and P. R. Philips. 1985. "The Value of Safety: Results of a National Sample Survey," The Economic Journal, 95 (March), pp. 49-72. Lee, R., A. J. Krupnick, D. Burtraw, et al. 1995. Estimating Externalities of Electric Fuel Cycles: Analytical Methods and Issues, and Estimating Externalities of Coal Fuel Cycles (Washington, D.C., McGraw-Hill/Utility Data Institute). McClelland, G., et al. 1991. Valuing Eastern Visibility: A Field Test of the Contingent Valuation, draft, U.S. Environmental Protection Agency, Washington, D.C. Oak Ridge National Laboratory. 1995. "Peer Review Of The Tracking And Analysis Framework (TAF) For Use In The 1996 NAPAP Integrated Assessment," Oak Ridge, Tennessee (ORNL/M-4994). Plagiannakos, T., and J. Parker. 1988. "An Assessment of Air Pollution Effects on Human Health in Ontario," Ontario Hydro, March. Pope, C. A. III, M. J. Thun, M. M. Namboodiri, D. W. Dockery, J.S. Evans, F. E. Speizer, and C. W. Heath, Jr. 1995. "Particulate Air Pollution as a Predictor of Mortality in a Prospective Study of U.S. Adults," American Journal of Respiratory and Critical Care Medicine, 151, pp. 669-674. Portney, Paul R. 1990. "Economics and the Clean Air Act," Journal of Economic Perspectives, vol. 4, no. 4, pp. 173-181. Schwartz, J., and D. W. Dockery. 1992. "Increased Mortality in Philadelphia Associated with Daily Air Pollution Concentrations," American Review of Respiratory Disease, 145, pp. 600-604. Shannon, J. D., E. C. Trexler, and R. Sonnenblick. 1997. "Modeling Visibility for Assessment," Atmospheric Environment, forthcoming. 31 Burtraw, Krupnick, Mansur, Austin, and Farrell RFF 97-31-REV U.S. Environmental Protection Agency (USEPA). 1995. "Air Quality Criteria for Particulate Matter (Draft)," Environmental Criteria and Assessments Office, Research Triangle Park, N.C. U.S. Environmental Protection Agency (USEPA). 1996. "Regulatory Impact Analysis for Proposed Particulate Matter National Ambient Air Quality Standard (Draft)," prepared by Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards, USEPA, Research Triangle Park, N.C. (December.) Van Horn Consulting, Energy Ventures Analysis, Inc., and Keith D. White. 1993. Integrated Analysis of Fuel, Technology and Emission Allowance Markets, prepared for the Electric Power Research Institute, EPRI TR-102510 (August). White, Keith D., Energy Ventures Analysis, Inc., and Van Horn Consulting. 1995. The Emission Allowance Market and Electric Utility SO2 Compliance in a Competitive and Uncertain Future, prepared for the Electric Power Research Institute, EPRI TR-105490s (September). 32 RESOURCES FOR THE FUTURE SO₂ Allowance Trading: How Experience and Expectations Measure Up Douglas R. Bohi Dallas Burtraw Discussion Paper 97-24 February 1997 Resources for the Future 1616 P Street, NW Washington, DC 20036 Telephone 202-328-5000 Fax 202-939-3460 © 1997 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the authors. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not undergone formal peer review or the editorial treatment accorded RFF books and other publications. SO₂ Allowance Trading: How Experience and Expectations Measure Up Douglas R. Bohi and Dallas Burtraw Abstract The SO2 trading program has achieved reductions in emissions ahead of schedule, with allowance prices below the marginal costs that were anticipated for the program. This paper explores the experience with the program and proposes a taxonomy of reasons why allowance prices are low. The overarching reason is that the most costly investments to accommodate full emission reductions have been successfully delayed. Application of a discount rate to these long run marginal costs yields an estimate of allowance price close to that observed today. Several factors have contributed to the delay in bearing these costs, and helped to reduce their magnitude. One group of factors stems from market fundamentals, especially the cost of rail transport of low sulfur coal. A second group includes the influences of state and federal regulators. A third group includes distinctions from the "imagined" program compared to that which was actually been enacted. -ii- SO₂ Allowance Trading: How Experience and Expectations Measure Up Douglas R. Bohi and Dallas Burtraw¹ 1. INTRODUCTION The sulfur dioxide (SO2) emission allowance trading program enacted through Title IV of the 1990 Clean Air Act Amendments (CAAA) ushered in the first nationwide effort to use a market in tradable "emission allowances" as the way to encourage the electricity industry to minimize the cost of reducing emissions. The electricity industry is allocated a fixed number of total allowances and firms are required to hold one allowance for each ton of sulfur dioxide they emit. Firms may choose to buy allowances to meet their requirement rather than reduce emissions, or to sell excess allowances not required for compliance. Since the first phase of the program began in January, 1995, the experience has been full of pleasant surprises. Firms are reducing emissions in excess of statutory requirements in Phase I and accumulating an unexpectedly large bank of allowances for use in the second phase of the program that begins in January 2000. Probably the most surprising development has been the decline in the price for allowances, compared with the expectations of many analysts at the inception of the program. Also the cost of emission reductions using low sulfur coal or flue gas scrubbing equipment has been lower than expected. At the same time, it appears that the volume of allowance trading to date, while substantial, has been less than analysts had first believed necessary to take full advantage of allowance trading. Hence, there is an apparent puzzle: greater than expected cost savings have been achieved under the allowance trading program with a lower than expected level of trading (Conrad and Kohn, 1996, offer another perspective on a similar question). 1 Douglas R. Bohi, Charles River Associates, Inc.; Dallas Burtraw, Quality of the Environment Division, Resources for the Future. The authors are grateful for research assistance from Brian Kropp, Ron Lile and Erin Mansur, and comments from Curtis Carlson. -1- Bohi and Burtraw RFF 97-24 An explanation for these observations is called for not only for evaluation of the current program, but also to arrive at lessons that might guide the use of tradable permit programs in other applications in the future. This is especially important with respect to the differences between expectations of allowance prices and their current value, because some analysts assert that engineers and economists so overstated the cost of meeting SO2 reductions that their cost projections for future environmental regulations such as carbon dioxide controls should be discounted significantly. This paper will evaluate the experience with the emission trading program to date and explain why the surprises indicated above came to pass. We offer a taxonomy of reasons why allowance prices are below the expectations of some. The lesson is not so much that prior estimates were wrong (at least the more sophisticated and unbiased ones) but that they were static snapshots taken in a dynamic policy environment. We argue the program should be given credit for the attainment of cost savings in emissions reductions because it has provided incentives for firms to innovate and to exploit advantageous trends in fuel markets and in the industry. Nonetheless, there have been impediments to trading allowances and other idiosyncrasies in program design that have promoted uneconomic investments and prevented firms from realizing additional savings. The experience to date illustrates that the potential savings from incentive based environmental policies such as the allowance trading program are enormous. However, improvements in trading behavior will be increasingly important for the industry to capture these savings as it approaches the more stringent second phase of the program. Due to increasingly competitive pressures to reduce costs in electricity generation, we are optimistic that a boost in allowance trading will be realized, but this is far from certain. 2. AN EXPERIMENT IN REGULATORY REFORM In contrast to traditional command and control environmental regulation, where firms typically adopt a common technology to reduce emissions by a specific amount, the emission -2- Bohi and Burtraw RFF 97-24 trading program allows each firm to choose its own individual emission rate and the means to achieve that rate. This flexibility allows firms to identify the cheapest way to reduce its compliance costs. Allowance trading takes advantage of the fact that emission control costs vary across different generating plants, and encourages firms with the cheapest control costs to undertake the most emission reduction. To work in this idealized way, however, each firm must have the incentive to minimize its costs. One should be careful in assuming that electric utilities will seek to, or even will be allowed to, minimize the cost of reducing emissions. The problem is that electric utilities are not allowed to maximize their profits. Instead, their prices and profits are regulated by the states in which they operate. The state regulators allow the firms to charge prices for electricity that will recover only those costs of production that are deemed to be prudently incurred. In fact, many state public utility regulators have imposed cost recovery rules that distort the firms' incentive to minimize costs (see Rose, 1997; Bohi, 1994; Rose, 1992; and Bohi and Burtraw, 1992). For example, regulators typically allow more favorable cost recovery on scrubbing and fuel switching than on allowance purchases. Several states have granted preapproval for the recovery of scrubber costs. Investments in capital costs such as scrubbers or fuel handling equipment in the case of fuel switching can be depreciated or expensed beginning as soon as facilities are in operation, and in some states scrubber costs can be recovered even before the units are in operation. Fuel cost increases are typically passed through to customers in the year the expense is incurred. In contrast, though allowance acquisitions may be necessary for investment and planning purposes years in advance of their use, their cost cannot be recovered until after the allowances are used for compliance. If the price of an allowance falls over the period of time between acquisition and use the state regulator may deny full recovery of cost on the grounds that a particular transaction was imprudent. If the price rises and the utility decides to sell the allowances and adopt a different -3- Bohi and Burtraw RFF 97-24 compliance strategy, the utility is unlikely to be able to capture the capital gains. Hence, the utility faces a downside risk for which there is no upside compensation. In summary, the potential savings from allowance trading are tremendous but for a variety of reasons firms may have an incentive to pursue abatement options over allowances transactions even when they appear to be more expensive. Several studies have estimated a high cost associated with impediments to trading (Fullerton et al., 1996; Winnebrake et al., 1995; USGAO, 1994). For a dissenting view on whether state regulators have impeded trading, see Bailey (1996). In the last year, apparently despite these impediments, there has been a tremendous boost in allowance trading (Dean and Kruger, 1997). One reason may be that regulators have corrected the early signals given to utilities, but evidence of this is lacking. It is more likely the case that the potentially achievable cost savings from trading overwhelm the restrictive signals issued by some regulators, especially as utilities come to recognize the oncoming pressures of competition set in motion by the 1992 Energy Policy Act. While many have complained that regulators stopped paying attention to emission trading in 1992 (Rose, 1997), it may turn out to be a good thing.² 3. OBSERVATIONS OF THE EARLY PERFORMANCE Four interrelated aspects of the experience with the emission trading program deserve to be highlighted and they are listed in Table 1. First, the volume of sulfur dioxide actually emitted since the law was passed in 1990, and before it took effect in 1995, is significantly lower than expected.³ Utilities have been aggressive in taking advantage of the opportunity to 2 One utility employee told us that the utility had stopped filing the burdensome paperwork that the public utility commission required to be filed for every allowance transaction. So far, he reported, the regulators did not notice or care. 3 Projections by both ICF (1989), p. 24, and Molburg et al. (1991) Table 5.1, p. 47, showed emissions rising steadily from 1985 to 1995 when the new Clean Air Act requirements went into effect. Instead, emissions rose only to 1990, the year the Clean Air Act was enacted, and dropped steadily thereafter. For a comparison of these studies see Reid, Bohi and Burtraw (1994). -4- Bohi and Burtraw RFF 97-24 save allowances earned through emission reduction actions, and the volume of "banked" allowances that will be available for use in Phase II of the program is larger than expected. The bank represents a "win-win" outcome for the environment and for industry. The early reductions provide an opportunity for environmental recovery and improved public health at an earlier point in time than would occur otherwise, while the bank provides an opportunity for industry to lower its overall costs of compliance and the ability to ease into the more stringent Phase II. Estimates of the volume of banked allowances expected to accumulate during Phase I range between 7 and 15 million. White et al. (1995) estimate the allowance bank will reach 9.4 million tons in the year 2000. RDI (1995) estimates the bank will reach 12-15 million. In 1991 the estimate of total Phase I banked allowances made for the Department of Energy (DOE) was only 4.35 million (Molburg et al., 1991, Table 4.10, p. 44). The more recent estimates of the Phase I bank are equivalent to 25% to 50% of the allocation of allowances during the first five-year phase of the program. Table 1: Four observations of SO2 allowance trading to date. 1. Over-compliance in Phase I will lead to a substantial bank for Phase II. 2. The volume of allowance trading has been less than anticipated. 3. The price of allowances is low. 4. The cost of emission reductions is low. A second observation is that the volume of allowance trading has been less than many anticipated. The overwhelming portion of the approximately 53 million private allowances transferred in 3,300 transactions that have been recorded through the EPA's Allowance Tracking System were accounting transfers within firms. A relatively small percentage, roughly about six to ten percent, were economic transfers between independent parties. The actual number of economic transfers is not easily deciphered from the allowance data. Three efforts to identify economic transfers that reach somewhat different conclusions are reported in -5- Bohi and Burtraw RFF 97-24 Table 1. As of the middle of 1995, White et al. (1995) find less than 1.6 million allowances have been traded among utilities. Joskow et al. (1996) find that as of the first quarter of 1996 about 6.5 million allowances have been traded among utilities, between utilities and third parties and between non-utility parties. In addition, 775,000 allowances have been purchased at EPA auctions (excluding allowances offered for sale by private parties through the auction). Dean and Kruger (1997) report that economically meaningful transactions (defined as allowances acquisitions by utilities) involving 4.8 million allowances had occurred as of the end of 1996. Lile et al. (1996) review the issues involved in these assessments and promote the perspective developed in Dean and Kruger. Montero, Ellerman and Schmalensee (1996) find that for compliance in 1995, "spatial" trading (as opposed to "inter-temporal" trading, or banking) account for about 500,000 out of 5.3 million tons from facilities covered by the program in Phase I. About two-thirds of this trading was intra-utility, and only one-third (involving primarily six utilities) was inter-utility, representing about 3% of allowances used for compliance. The data suggest that a good number of utilities have failed to take advantage of the allowance market to reduce their costs of compliance, though most have taken advantage of the flexibility in the allowance trading program to bank allowances for Phase II, and to pursue fuel switching. Eighty-five percent of the base allowances issued in the first phase (excluding facilities in the Substitution and Compensation Program) go to utilities located in only eleven states (see columns 1 and 2 of Table 2). All of the states may be said to actively reduce the incentive to trade. For example, all of the states proscribe that the benefits that may result from allowance trading will be passed on to customers. The State of Connecticut provides the one exception to this rule, wherein it allowed a utility to retain 15% of the revenue from the sale of allowances (Rose, 1997). In addition, six states provide a regulatory bias in favor of scrubbers as a compliance option (column 3). All six states have significant indigenous coal industries and -6- Bohi and Burtraw RFF 97-24 have acted to protect their markets against encroachment from lower sulfur coal. Out of 110 affected generation facilities, 26 generating units at 14 facilities opted for scrubbing and all of these units are located in states with a regulatory bias in favor of scrubbing. Table 2: Allowance trading and compliance plans by state. (1) (2) (3) (4) (5) (6) State % of Phase I Regulatory Bias For Plans to Expected Phase I Allowances "Other" Compliance Trade Status Toward Compliance Options Allowances Allowances Plans Mainly Ohio 16.9 Yes Intrafirm Buy Switching; Scrub Gavin Mainly Indiana 12.6 Yes Sell Buy Switching; Scrub 7 Units Bank or Georgia 10.2 No Intrafirm Sell Switching Only Scrub 5 Units; Penns 9.4 Yes Buy Sell Switching Scrub 4 units; West Va 8.7 Yes Intrastate Sell Switching Intrastate Switching and Illinois 6.9 Yes and Buy Sell Buying Allowances Missouri 6.2 No No Sell Switching Only Kentucky 4.9 Yes Intrastate Buy Mainly and Sell Switching; Scrub 6 Units Alabama 4.1 No Intrafirm Sell Switching Only New York 2.7 No No Sell Switching; Scrub 2 units Florida 2.4 No No Buy Switching Only TOTAL 85.0 SOURCES: Column (3) and (4) from interviews with public utility commission staff; Column (5) from ICF model with "Low-Flexible" assumptions, reported in National Acid Precipitation Assessment Program, 1990 Integrated Assessment Report, p. 425; and Column (6) from U.S. Environmental Protection Agency, "SO2 Phase I and II Boiler Compliance Methods," Office of Air and Radiation, June 14, 1993. -7- Bohi and Burtraw RFF 97-24 One indication of the unexpected nature of the amount of allowance trading so far is reflected in the discrepancies between the expected status of the utilities in each state as a buyer or seller of allowances (column 5) and the actual compliance plans (column 4). The lack of correspondence between plans and estimates is of interest because the ICF optimization model was the primary source of estimates of the efficiency gains from a system of allowance trading that was used in the political debate over the acid rain provisions in the CAAA. Though the costs of the program have been low, experience so far suggests that the promised benefits of allowance trading are not being realized in the way they were initially expected. Another indication of the amount of trading is illustrated in Figure 1. Arrayed along the horizontal axis of this figure are utilities (at the holding company level) according to the number of total allowances allocations they are to receive between 1995 and 2026. The vertical axis is a ratio of the amount of "economically distinct" transactions each utility has been involved in to their total allowance allocation. We include intra-utility trades, which could be more important for the larger firms and holding companies. The ratio captures our intuition that utilities with more allowances should be more involved in trading, so one might expect the ratios across firms to be approximately equal if all firms attempted to take equal advantage of the market. The variation in these ratios suggests that utilities vary greatly in the amount of activity in the market, even when we control for their size and allowance allocations. A third and particularly important observation is that the price of allowances has been lower than expected. Table 3 reports expected and actual allowance prices and the volume of inter-utility trades. Prior to the enactment of the Clean Air Act Amendments in 1990, the estimated price of allowances ran as high as $1500, a number that was enshrined in the Act as the fixed price of direct sales by EPA. During 1990, the EPA cited a price estimate of $750 as the best guess of what allowances (and emission reductions at the margin) would cost. -8- 0.0000% 0.2000% 0.4000% 0.6000% 0.8000% 1.0000% 1.2000% 1.4000% 1.6000% Southern Company American Electric Power Tennessee Valley Authority TU Electric* Cinergy Corp. GPU Generation Allegheny Power System Detroit Edison Company* -9- Illinois Power Ohio Edison Company Pennsylvania Power & Light as % of Initial Allowances Commonwealth Edison Virginia Power Total Number of Economically Distinct Allowances Bought/Sold Duke Power Company* Central & South West Services* Carolina Power & Light* PacifiCorp* Union Electric Company Florida Power & Light* ENTERGY* Bohi and Burtraw RFF 97-24 Instead, actual prices for inter-utility trades have been considerably lower, starting out in the range of $250-300 in 1992, they fell steadily to $110-140 in 1995, and to around $70 in 1996 before rebounding to the low $90s at the time of this writing. Table 3: Allowance prices and volumes. Allowance EPA Auction Quantity of Inter-Utility Trades Price Clearing Price (thousands) (Spot) White et al.* Joskow et al.** Dean & Kruger Expected 1989 $1500 1990 $750 Actual 1992 $250-$300 70 130 1993 $130-$300 $157 639 226 1994 $130-$140 $159 562 1,467 1995 $110-$140 $132 320*** 4,918 1996 $90 $66 4,800**** Notes: The expected price in 1989 is from ICF (1989), and in 1990 it is from White et al. (1995). Actual prices and trading volumes are from White et al. (1995), Joskow et al. (1996) and the industry press. White et al. (1995) report inter-utlity trades. Joskow et al. (1996) report trades between April and March of the subsequent year, and include inter-utility trades, trades between utilities and third parties, and trades between two non-utility parties. Data for 1995 is for first half of year only. Dean and Kruger (1997) report all allowance acquisitions by utilities through the end of 1996. Table 4: Projected annual costs under alternative implementations for Phase II. Command and Constrainted trading Flexible interutility billion dollars control baseline (internal transfers) trading ICF (1989) 3.3 4.7 2.7 4.0 Van Horn Consulting et al. (1993) 5.1 3.4 2.2 GAO (1994) 4.3 2.5 1.4 Burtraw, et al. (1997) 2.5 1.0 A fourth observation, also related to the price of allowances, is that the cost of emission reductions has been dramatically less than industry projections before adoption of the program. Table 4 presents four sets of estimates compiled at various times of the relative -10- Bohi and Burtraw RFF 97-24 annual costs in Phase II for three implementation scenarios: command and control, constrained trading (only among facilities within a firm), and interutility trading. The estimates pertain to annual cost when the net contributions to the allowance bank are zero sometime in Phase II. The estimates illustrate an evolution as utility compliance plans have become known and set into motion, and new information has surfaced about compliance options and their costs. According to estimates made by ICF (1989) for the U.S. Environmental Protection Agency (EPA) before the CAAA was enacted, the levelized annual cost of the program was expected to be $2.7 - 4.0 billion. These costs reflect a cost savings of 40% over a command and control approach. The GAO study in 1994 also estimated that compliance costs in the year 2001 would be almost 40 percent less than they would have been under a command and control approach of emission rate limits applied to specific facilities, but costs in both cases were less than forecast in 1989. Specific technology requirements, which were also considered in various proposals in the 1980s, would have been even more expensive. The Burtraw et al. (1997) study differs from the others, which were primarily engineering based studies, by employing an econometric analysis. They find that permit trading would reduce the costs of the program by 60% compared with a command and control approach. 4. WHY ARE ALLOWANCE PRICES so LOW? In this section we reconcile the most widely discussed of the discrepancies between experience and expectations: Why allowance prices are so low? The answer, in this case at least, is not that the experts are wrong. In fact, the explanations for the divergence between expectations and observed allowance prices are several, and can be organized into three groups, as listed in Table 5. First are market fundamentals, which is analogous to the term used in financial economics that describe changes that have driven down marginal costs of compliance. Second are regulatory influences that also work to reduce observed allowance -11- Bohi and Burtraw RFF 97-24 prices to date. Third are the distinctions between the imagined program that was the subject of many analyses, and the actual program as set forth in the Clean Air Act. Table 5: Nine reasons why allowance prices are so low. Market Fundamentals: 1. Discounting of future costs. 2. Widespread availability of low sulfur coal. 3. Competition and innovation. 4. General equilibrium effects. Regulatory Influences: 5. Sunk "uneconomic" investments in scrubbers. 6. Annual auction invites strategic under-bidding. The Imagined versus Real Program: 7. Bonus allowances subsidies for scrubbing delay future costs. 8. Two phases of program segregate sellers and buyers. 9. Substitution and Compensation units delay future costs. Market Fundamentals 1. Marginal cost estimates that will be incurred in the future should be discounted to the present to be relevant for current allowance prices. The single most important factor in comparing the marginal costs of the program with observed allowance prices to date is the role of discounting. An investment of $126 that is to be made in the year 2000 for the purpose of compliance has a present value of (1-d)³ 128 in 1997, where d represents a discount rate. For instance if d=.08, then the investment in the year 2000 has a present value in 1997 of $100. Furthermore, anything that happens to delay this investment beyond the year 2000 serves to lower its present value. Most analyses of long run marginal costs which were used for allowance price projections were built on models that accommodated full implementation of Phase II of Title IV, which will be felt when the bank from Phase I is drawn down and net contributions to the bank are zero. The bank is expected to be exhausted toward the end of the next decade. Table 6 reports several estimates along with the net present value (NPV) of the long run -12- Bohi and Burtraw RFF 97-24 marginal costs obtained by discounting these estimates at a rate of 8% per year (real), assuming these marginal costs were incurred in the year 2010. White et al. (1995) suggest an 8% real discount rate as a benchmark, above the after-tax weighted average cost of capital (5% real), to account for increased risk as the industry enters a more competitive era, and uncertainty around the allowance market. White et al. (1995) and ICF (1995) are engineering studies of marginal costs. The RDI (1995) estimates are projections of allowance prices, rather than marginal costs per se, so the comparison with the engineering estimates is somewhat misleading. Several of the complementary explanations we discuss below have helped to delay the time at which these investments would be needed, thereby lowering the price of allowances in the present. Table 6. Estimates of the long run marginal cost of emission reductions, and the role of discounting future costs for present allowance transactions. Long Run Marginal Cost Estimates Net Present Value White et al. (1995) $528 $179 ($335-$614) ($114-$209) ICF (1995) $516 $175 RDI (1995)* $220 $75 ($100-$360) ($34-$122) Burtraw et al. (1997) $284 $97 Notes: * RDI (1995) estimates are for allowance prices rather than marginal costs. 2. Low price and widespread availability of low sulfur coal. Low sulfur fuel appears to account for about half of the emission reductions in Phase I, but more importantly it accounts for emission reductions at the margin, serving to set marginal costs and allowance prices. The most important aspect in the trend toward lower compliance costs is that, while the demand for low sulfur coal has gone up, the price has gone down due to dynamic changes in coal markets. Also, there is roughly equal access to low-priced coal to all utilities affected by Phase I regardless of their location. -13- Bohi and Burtraw RFF 97-24 The relevant factor for determining the marginal cost of fuel switching or blending as a compliance option is the difference between low and high sulfur fuel, termed the "sulfur premium." This premium widened in 1990, but then stabilized and narrowed in most states since then. Surprisingly, the shift in demand toward low sulfur coal did not cause the price of low sulfur coal or the sulfur premium to rise. Thus, switching to low sulfur coal did not entail a major economic burden, as some analysts for electricity industry had predicted it would in the 1980s.⁴ After Phase II requirements are in place in 2000, the range of sulfur content in coal that achieves compliance is cut to about half that in Phase I. Focusing demand on this narrower range of coals will test the capability of the supply system to deliver low sulfur coal at stable prices. Two questions are central to the determination of future compliance costs. Is it possible to further increase the infrastructure and efficiency in transporting coal from Wyoming to the east? Is the low sulfur coal resource base east of the Mississippi large enough to continue to be an important source of supply to eastern utilities? Scrubbers may be more important for meeting compliance requirements in the future. For example, plants that were able to meet Phase I requirements by blending low and high sulfur coals are likely to be faced with a choice between a switch to low sulfur coal or the installation of a scrubber. The capital costs associated with switching from high to low sulfur coal is about 35% of the capital costs of a scrubber per kW of capacity. In addition, if the answer to either of the two questions raised above about the supply of low sulfur coal turns out negative, and low sulfur coal is not as competitive in Phase II, scrubbers may be the cheapest option for meeting the emissions cap. 4 In fact, one industry study suggested the costs of sulfur abatement would be higher if fuel switching played an important role, because although "intra-firm trading" would reduce expensive scrubbing it would increase reliance on low sulfur coals resulting in an increased sulfur premium that would raise costs for units already using low sulfur coal. (TBS, 1986). -14- Bohi and Burtraw RFF 97-24 3. Competition and innovation. The minemouth price of low sulfur coal located in the Powder River Basin has always been cheaper than eastern alternatives, but the geographic extent of the market has been limited by transportation costs. As noted by Ellerman and Montero (1996), Fieldston Company (1996) and Burtraw (1996), the cost of rail transportation from the Powder River Basin to delivery points east has fallen recently. Because of competition created by the Staggers Rail Act of 1980, the cost of transportation from the Powder River Basin has declined from a rate of over two mills per ton-mile in the early to mid 1980s to less than one mill per ton-mile today (Ellerman and Montero, 1996), p 5). Railroad competition has also reduced rail rates for central Appalachian and midwestern coals, but by less than the reduction in Powder River Basin rates for delivery to the midwest. As reported by Burtraw (1996, p. 13), transportation rates dropped from 20-26 mills per ton mile to 10-14 mills over the same period. Another area of innovation has been coal-blending at the power plant. The blending of fuels with different sulfur contents was thought difficult before 1990. Since that time engineers have learned they can blend up to 40% low sulfur subbituminous coals with higher sulfur bituminous coal, without incurring major capital expense or a decline in performance. Yet another leading compliance option is scrubbing. The cost of installing and operating scrubbers has fallen in recent years, no doubt in response to competition from lower sulfur coal while removal efficiencies have improved. In addition, the fact that "a ton of emissions saved is an allowance earned" gives an incentive for utilities and scrubber suppliers to find ways to improve the efficiency of existing scrubbers. 4. General equilibrium effects One source of systematic bias in engineering models and many economic models is the failure to take into account all the opportunities that individuals in the economy have to alter the investment and consumption behavior in response to changes in prices. For instance, since -15- Bohi and Burtraw RFF 97-24 electricity prices are affected by the SO2 program, consumers may decide to switch to other sources of energy services at less cost. Goulder et al. (1996) embed a ballpark estimate of marginal technological abatement cost within a computable general equilibrium model. This model allows producers and consumers to substitute among alternatives as the price of electricity changes reflecting expenditures on technological abatement. These substitutions tend to amplify emission reductions in their model; consequently, the environmental goal of a 10 million ton reduction is achieved with less investment in technological abatement than would be anticipated without such opportunities to substitute, and at a lower cost. They find the general equilibrium context to lower the private marginal abatement costs by 5 to 15 percent. Regulatory Influences 5. Regulatory incentives promoted "uneconomic" scrubber investments that, once in place, have low marginal cost. Scrubbing accounts for about 50 percent of the emissions reductions in Phase I. This compliance option has been heavily promoted in some states by the legislatures and state public utility commissions. Scrubbing involves large up-front capital costs, an anathema in an industry already burdened by excess generating capacity and facing the prospect that a significant part of their capital costs cannot be fully recovered in a more competitive market.⁵ Although scrubbers have been installed on 14 plants (26 generating units) to meet Phase I compliance requirements, their choice may not reflect an underlying economic advantage as much as favorable regulatory treatment, as described previously. In addition, 19 units qualified for a significant bonus of emission allowances (3.5 million tons, or over 10 percent of the allocation in the five years of Phase I) awarded by the CAAA to utilities that add scrubbers in Phase I. Even with these bonuses, these investments appear uneconomic ex post. Although from the current perspective capital cost of investments in retrofit scrubbers appear 5 Chao and Wilson (1993) estimate that the option value provided by holding allowances rather than scrubbing is worth $85 per ton, although their calculations assume that the price of allowances is $400 per ton and the elasticity of low sulfur coal is very low. -16- Bohi and Burtraw RFF 97-24 uneconomic, once in place their operating costs are low, perhaps $60-$90 per ton removed, so they will be run to full capacity. The operating costs of scrubbing do not determine the marginal abatement cost, since they are fully utilized. However, they reduce the need and lower the marginal cost for other compliance options, thereby affecting allowance prices. 6. The annual auction provides incentives for buyers and sellers to understate their opportunity costs, which may have an influence on allowance prices. The auction design set forth in the statute is a discriminating price, sealed bid auction that provides strategic incentives for bidders and sellers to understate their reservation prices (Cason, 1993, 1995). Whether this incentive has led to average auction prices that are biased relative to opportunity costs is an open question. Joskow, Schmalensee and Bailey (1996) argue that the auction has been unimportant with regard to the performance of the allowance trading program generally. Furthermore, some utilities have argued that they could do better if they were allowed to retain ownership of allowances that are allocated through the auction. Ellerman and Montero (1996) suggest, however, that the auction played a useful role in the early stages of the program in signaling the relatively low compliance costs that faced many utilities. Further, some brokers and utilities report that future allowance transactions are hinged to the auction price, so that any change in the auction would disrupt the evolution of the market. Whether or not the auction design is amended in the future, evidence is consistent with the notion that the annual auction is poorly designed and provides a price signal below true opportunity costs. Whether the influence of the auction is or is not important has not been determined convincingly The Imagined versus Real Program 7. Bonus allowances add to the surplus of allowances in the early stage of the program, with downward pressure on price. One aspect of Title IV that was not included in many early models of the program is the role of bonus allowances for utilities that chose to meet emission reductions through -17- Bohi and Burtraw RFF 97-24 scrubbing. These allowances expand the amount of Phase I allowances available by 3.5 million, over 10 percent of allocations in Phase I. This increase in supply in the near term places downward pressure on costs and on allowance prices by delaying the necessity of abatement expenditures. 8. The two phases of the program segregate most potential allowance sellers into Phase I and buyers into Phase II, resulting in a near term downward trend in the price of allowances. Burtraw et al. (1997) estimate the mean marginal abatement cost weighted by emissions at various facilities in 1994. They find the cost for Phase I facilities in 1994 to be $60, and for Phase II facilities to be $384. This difference suggests that a substantial bank would accumulate to delay the relatively more expensive abatement efforts that otherwise would be required from Phase II facilities after 2000. If the allowance market performed perfectly, the accumulation of an allowance bank would not necessarily have an impact on the relationship between observable allowance prices and marginal costs. However, to the extent there are market imperfections, less than perfectly rational planning, or regulatory intervention, the asymmetry between Phase I and Phase II may exacerbate trends in the allowance market that are unrelated to technological costs pushing down current allowance prices. 9. The opportunity to rely on Substitution and Compensation units further reduces the costs of complying with Phase I emission reductions. Substitution and Compensation units voluntarily brought into Phase I typically have the lowest marginal costs of units affected in Phase II. By opting into Phase I, managers at these units can begin emission reductions at these units and delay reductions at other units by building up their allowance bank. Hence, to the extent that the surplus of allowances in Phase I leads allowance prices to diverge from naive estimates of marginal abatement cost, this trend would be exacerbated by the opportunity to use Substitution and Compensation units in Phase I. -18- Bohi and Burtraw RFF 97-24 5. CONCLUSION The puzzle that has emerged from the early experience with the SO2 allowance trading program is: Why are allowance prices so low, especially when trading volume appears to be lower than expected? This paper offers several answers to this question. The conclusion that emerges is that the institution of allowance trading and the flexibility it gives to firms has enabled electric utilities to achieve environmental goal at dramatic reductions in costs, compared to previous regulatory experience. However, one must also consider the particular aspects of this program that delay significantly the time at which future investments will be incurred. This often overlooked factor also contributes significantly to the low allowance prices that have been observed in the first years of the program. -19- Bohi and Burtraw RFF 97-24 REFERENCES Bailey, E. M. 1996. Allowance Trading Activity and State Regulatory Rulings: Evidence from the U.S. Acid Rain Program, MIT-CEEPR 96-002WP, Massachusetts Institute of Technology. Bohi, D. R. 1994. The Electricity Journal, 7, pp. 20-27. Bohi, D. R., and D. Burtraw. 1992. Resources and Energy, 14, pp. 129-153. Burtraw, D. 1996. Contemporary Economic Policy, 14, pp. 79-94. Burtraw, D., C. Carlson, M. Cropper, and K. Palmer. 1997. "Econometric Estimates of SO2 Abatement Costs Under Title IV," in Proceedings of AWMA/Acid Rain Electric Utilities Conference. Cason, T.N. 1993. Journal of Environmental Economics and Management, 25, pp. 177- 195. Cason, T.N. 1995. American Economic Review, 85, pp. 905-922. Chao, H.-P., and R. Wilson. 1993. Journal of Regulatory Economics, 5, pp. 233-249. Conrad, K., and R. E. Kohn. 1996. Energy Policy, 24, pp. 1051-1059. Dean, M., and J. Kruger. 1997. "Using EPA's Allowance Tracking System to Assess the Allowance Market," in Proceedings of AWMA/Acid Rain Electric Utilities Conference. Ellerman, A. D., and Juan-Pablo Montero. 1996. "Why are Allowance Prices so Low? An Analysis of the SO2 Emissions Trading Program," MIT-CEEPR working paper. Fieldston Company, Inc. 1996. Coal Supply and Transportation Markets During Phase One: Change, Risk and Opportunity, prepared for Electric Power Research Institute, TR- 105916. Fullerton, D., S.P. McDermott, and J.P. Caulkinset. 1996. Sulfur Dioxide Compliance of a Regulated Utility (Austin, Texas, University of Texas at Austin, Department of Economics). Goulder, L. H., I. W.H. Parry, and D. Burtraw. 1996. Revenue-Raising VS. Other Approaches to Environmental Protection: The Critical Significance of Pre-Existing Tax Distortions, Discussion Paper 96-14, Resources for the Future, Washington D.C., April. ICF. 1989. Economic Analysis of Title V (sic) (Acid Rain Provision) of the Administration's Proposed Clean Air Act Amendments (HR 3030/S 1490), prepared for the U.S. Environmental Protection Agency. ICF. 1995. Economic Analysis of Title IV Requirements of the 1990 Clean Air Act Amendments, prepared for the U.S. Environmental Protection Agency. -20- Bohi and Burtraw RFF 97-24 Joskow, P.L., R. Schmalensee, and E. M. Bailey. 1996. "Auction Design and the Market for Sulfur Dioxide Emissions," mimeo, MIT Center for Energy and Environmental Policy Research. Lile, R. D., D. R. Bohi, and D. Burtraw. 1996. An Assessment of the EPA's SO₂ Emission Allowance Tracking System (Washington, D.C., Resources for the Future). Molburg, J. C., J. A. Fox, G. Pandola, and C. M. Cilek. 1991. Analysis of the Clean Air Act Amendments of 1990: A Forecast of the Electric Utility Industry Response to Title IV, Acid Deposition Control, ANL/EAIS/TM-81, Argonne National Laboratory. Montero, J-P., A. D. Ellerman, and R. Schmalensee. 1996. "The U.S. Allowance Trading Program for Sulfur Dioxide: An Update After the First Year of Compliance," unpublished mimeo (draft), Massachusetts Institute of Technology. Reid, J., D. R. Bohi, and D. Burtraw. 1994. Recommendations to NAPAP Regarding SO2 Emission Projections (Washington, D.C., Resources for the Future). Resource Data International, Inc. (RDI). 1995. RDI's Phase I 1995 Databook (Boulder, Colorado, RDI). Rose, K. 1992. Public Utility Commission Implementation of the Clean Air Act's Allowance Trading Program (The National Regulatory Research Institute, Ohio State University). Rose, K. 1997. Market Based Approaches to Environmental Policy: Regulatory Innovations to the Fore, R. F. Kosobud and J. M. Zimmerman, eds. (New York, N.Y., Van Nostrand Reinhold), forthcoming. Temple, B., and Sloane, Inc. 1986. Evaluation of H.R. 4567: The 'Acid Deposition Control Act of 1986', prepared for the Edison Electric Institute. U.S. General Accounting Office (USGAO). 1994. Air Pollution: Allowance Trading Offers an Opportunity to Reduce Emissions at Less Cost, GAO/RCED-95-30. U.S. National Acid Precipitation Assessment Program (NAPAP). 1991. 1990 Integrated Assessment Report, Washington D.C. Van Horn Consulting, Energy Ventures Analysis, Inc., and K. D. White. 1993. Integrated Analysis of Fuel, Technology and Emission Allowance Markets, EPRI TR-102510, prepared for the Electric Power Research Institute. White, K. D., Energy Ventures Analysis, Inc., and Van Horn Consulting. 1995. The Emission Allowance Market and Electric Utility SO₂ Compliance in a Competitive and Uncertain Future, prepared for the Electric Power Research Institute, EPRI TR-105490s. Winebrake, J., M. A. Bernstein, and A. E. Farrell. 1995. The Electricity Journal, 8, pp. 50-54. -21- RESOURCES FOR THE FUTURE An Assessment of the EPA's SO₂ Emission Allowance Tracking System Ronald D. Lile Douglas R. Bohi Dallas Burtraw Discussion Paper 97-21 November 1996 Resources for the Future 1616 P Street, NW Washington, DC 20036 Telephone 202-328-5000 Fax 202-939-3460 © 1997 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the authors. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not undergone formal peer review or the editorial treatment accorded RFF books and other publications. An Assessment of the EPA's SO₂ Emission Allowance Tracking System Ronald D. Lile, Douglas R. Bohi, and Dallas Burtraw Abstract On November 8, 1996, various Environmental Protection Agency (EPA) officials, scholars and industry representatives gathered at Resources for the Future (RFF) to examine the EPA's method for classifying private SO₂ allowance transactions by the Allowance Tracking System (ATS). The one-day workshop at RFF was designed to evaluate how well the EPA's classification scheme within the ATS currently meets the needs of constituencies with a vested interest in the allowance trading system, and to determine if other classifications would be more beneficial. The EPA has limited its collection of information to that which is necessary to ensure compliance with environmental goals. In particular, the EPA has interpreted its mission to be one of minimal interference in guiding the development of the allowance market and that its primary purpose is emission compliance and not the monitoring of transactions. Therefore, the goal of the ATS is to provide a central registry of recorded allowance transfers for the purpose of emission compliance. As a result, the ATS is unusual as a mechanism for monitoring market activity because it provides information about the buyer and seller of an allowance but does not provide price information. Furthermore, the EPA has limited its role so as not to exercise approval of individual allowance trades, and has excluded from consideration options for expanding the EPA's data collection effort. However, the EPA recognizes that the interests of Congress and the public extend beyond compliance with the environmental goals to include the development of allowance trading to help achieve these goals at the lowest possible cost. In addition, there is widespread interest in the development of SO₂ emission allowance trading as a prototype for other potential trading programs, and the ATS provides a potential template for the oversight role of the environmental regulator in programs such as these. Therefore, another goal of the workshop at RFF was to assess how well the ATS performs in promoting the development of allowance trading in general, and with respect to the interests and needs of each of the constituencies interested in the SO₂ allowance trading program. This discussion paper incorporates observations, suggestions and concerns expressed during this workshop. Furthermore, this discussion paper concludes with recommendations regarding the EPA's current classification methodology. Key Words: transaction costs, regulated industries, electric utilities, emissions JEL Classification Nos.: D23, D49, H70, K23, L94, Q25 -ii- Table of Contents Introduction 1 The Organization of ATS 3 Classifying Allowance Transactions and Capturing Transactions Costs 5 Inter-utility 7 Reallocation 7 Intra-utility 9 Broker to utility, utility to broker/trader, fuel company to utility and utility to fuel company 10 Other 10 An Illustration of the Current Taxonomy 10 Strategic Behavior in Trading Activity 13 What Needs to be Done to Improve the Market, and Is There a Role for ATS in Doing So? 14 Observations About the Future 15 Summary and Recommendations 16 List of Participants at the November 8 Meeting 18 List of Figures and Tables Figure 1. Current methodology 6 Figure 2. Previous methodology 6 Table 1. Examples of Inter-Utility and Intra-Utility Trades 11 Table 2. Examples of Reallocations and Intra-Utility Trades 11 Table 3. Examples of Trades Involving Co-owners 12 -iii- An Assessment of the EPA's SO₂ Emission Allowance Tracking System Ronald D. Lile, Douglas R. Bohi, and Dallas Burtraw¹ INTRODUCTION Allowance trading is an innovative feature of Title IV of the 1990 Clean Air Act Amendments, intended to reduce the costs of compliance with the goals of the statute. Title IV is a precedent-setting approach to environmental legislation that places government in the role of setting standards of performance, while leaving to the private sector the task of finding the best way to meet these standards. This approach provides the firm with the flexibility to find the most cost effective way of achieving the standard. The allowance trading program has the potential of dramatically lowering the costs of attaining the environmental goal of a national average cap on SO₂ emissions. Title IV requires the Environmental Protection Agency (EPA) to establish a system for the collection of information on allowance transfers, primarily for the purpose of monitoring compliance. The system that has been established is known as the Allowance Tracking System (ATS). Although its primary function is to facilitate regulatory oversight of compliance, the information that is available in the ATS is of interest to a number of constituencies for various purposes. Allowance brokers and electric utilities potentially could rely on the ATS to provide information about trading activity. Regulators could rely on the ATS to provide information that may be relevant to oversight of utility compliance activities and cost recovery. Policy analysts and Congress could rely on the ATS to provide an indication of the performance of the market and its effect on the costs of implementing emission reductions under Title IV. In 1 The authors are, respectively: Research Assistant, Energy and Natural Resources Division, Resources for the Future; Senior Fellow and Division Director, Energy and Natural Resources Division, Resources for the Future; and Fellow, Quality of the Environment Division, Resources for the Future. This research was supported in part by funding from the EPA. We would like to thank the participants at RFF's November 8, 1996 workshop on the EPA's Allowance Tracking System for their constructive contributions to this report. Furthermore, we received helpful comments from EPRI's Keith White. In addition, we are especially grateful to Melanie Dean, Joe Kruger and Alex Salpeter of the EPA's Acid Rain Office for their generous assistance. All remaining errors are the responsibility of the authors. -1- Lile, Bohi, and Burtraw RFF 97-21 addition, environmental advocates are interested in the ATS for information on the influence of allowance trading on the geographic or temporal pattern of emissions. On November 8, 1996, various EPA officials, scholars and industry representatives gathered at Resources for the Future (RFF) to examine the EPA's method for classifying private SO₂ allowance transactions by the Allowance Tracking System (ATS). The one-day workshop at RFF was designed to evaluate how well the EPA's classification scheme utilizes information in ATS to meet the needs of constituencies with a vested interest in the allowance trading system, and to determine if other classifications would be more beneficial. The EPA has limited its collection of information to that which is necessary to ensure compliance with environmental goals. In particular, the EPA has interpreted its mission to be one of minimal interference in guiding the development of the allowance market and that its primary purpose is emission compliance and not the monitoring of transactions. Therefore, the goal of the ATS is to provide a central registry of recorded allowance transfers for the purpose of emission compliance. As a result, the ATS is unusual as a mechanism for monitoring market activity because it provides information about the buyer and seller of an allowance but does not provide price information. Furthermore, the EPA has limited its role so as not to exercise approval of individual allowance trades, and has excluded from consideration options for expanding the EPA's data collection effort. It is the EPA's contention that the private sector should fill the information void, to the extent that the market needs a clear indication of allowance prices in order to be able to function. Such information is widely available from several sources and is reasonably accurate. However, the EPA recognizes that the interests of Congress and the public extend beyond compliance with the environmental goals to include the development of allowance trading to help achieve these goals at the lowest possible cost. In addition, there is widespread interest in the development of SO₂ emission allowance trading as a prototype for other potential trading programs, and the ATS provides a potential template for the oversight role of the environmental regulator in programs such as these. Therefore, another goal of the workshop at RFF was to assess how well the ATS performs as an evaluation mechanism for allowance trading activity, with respect to the interests and needs of each of the constituencies interested in the SO₂ allowance trading program. This paper is an assessment of the ATS in -2- Lile, Bohi, and Burtraw RFF 97-21 this regard. Observations, suggestions and concerns expressed during the workshop have been incorporated into this assessment. THE ORGANIZATION OF ATS The ATS is a database operated and maintained on EPA's National Computer Center. While ATS itself cannot currently be directly accessed, EPA provides weekly extracts of the most relevant information. These extracts, i.e., electronic files, can be assessed electronically by the public through the World Wide Web (address: http://www.epa.gov/acidrain/atsdata.html). As discussed below, expanded capabilities that will allow the public interactive access to these data files are in development at the EPA and are expected to be available in the summer of 1998. The following files derivable from ATS are the most relevant to the trading analysis: ACCOUNT -- contains information such Account Number, Account Name, Account Rep ID, Alternate Rep ID, Plant ID/ORISPL, Unit/Boiler ID and various flags denoting type of account. REPRESENTATIVES -- contains information such Account Rep/Alternate ID, Name, Address, Phone Number. OWNERS -- contains Account Number, Owner ID, Owner Name, Binding Parties. TRANSACTIONS -- contains information such Transaction Number, Transaction Type, Date Received, Date Recorded, Transferee Account, Transferor Account, Account Rep for Transferee, Account Rep for Transferor, Amount of Allowances Transferred. ALLOW_IN_TRANSACT -- contains information on allowances involved in a given transaction such as Transaction Number, Starting Serial Number, Ending Serial Number. ALLOW_HELD_BY_ACCOUNT: contains information on Account, Allowance Use Year, and Serial Number. UTILITY -- information on the utility company that operates (responsible for dispatching) a given plant. PLANT -- contains information on the State where Plant Located. PLANT_UTILITY_XREF -- a file combining the PLANT and UTILITY information, for effective retrieval/query. These files provide various ways to view current allowance data. These are very large files. For example, the TRANSACTION file includes over 9,500 transactions involving over -3- Lile, Bohi, and Burtraw RFF 97-21 41 million allowances. The TRANSACTIONS file may include transactions among unit accounts, which are accounts associated with generating units, and general accounts, which can be set up by any person, group or corporation.² The general account differs from the unit account because it is not associated with an individual generating unit. There are a variety of reasons that the data in these files may not provide information in a form that is useful to potential users. For instance, a transfer from a unit account to a general account and another transfer back to the original unit account would show up as two separate transactions that actually cancel each other out. Such a sequence of transactions is not unusual. To understand the data, the user needs to conduct some analysis. However, the website does not offer the option of merging files or conducting searches yet. However, files can easily be downloaded and manipulated by interested parties.³ To facilitate monitoring of the allowance market for the various purposes suggested above, the EPA's Acid Rain Division has devised a routine to categorize allowance transfers. EPA runs several queries on ATS data to produce a file with most of the information necessary to classify a trade. Using this consolidated information, the EPA can organize allowance transfers according to various categories that are viewed as meaningful for economic and environmental measures of the program. The EPA places allowance transactions in the following categories: Intra-Utility, Inter-Utility, Utility to Broker, Broker to Utility, Utility to Fuel Company, Fuel Company to Utility, Reallocation and "Other." These categories represent all private trades reported to ATS. Since the ultimate goal is emissions compliance by utilities, allowance transfers viewed as meaningful for economic and environmental measures of the program involve those allowances acquired by utilities. As noted, the Acid Rain Division is in the process of converting the ATS from a mainframe system to a Windows based system. The new windows based system should allow for interactive data filtering and manipulation. Currently, users must rely on analysis already conducted and posted by the EPA, or download the entire database to another computer to perform independent analysis. In addition, the Acid 2 A unit may be owned by more than one operating company. A transfer from a unit account to a general account of any of the owners is classified a reallocation. For further explanation, see the discussion of Reallocations in the text. 3 Another feature utilizing the above ATS data, but not available on the website, is a "history search." This feature enables the EPA to track the entire history of an allowance by its serial number. -4- Lile, Bohi, and Burtraw RFF 97-21 Rain Division is in the process of developing an electronic transfer program, which converts the ATS into a paper-less system. A pilot program is planned for next year with the Southern Company. However, there are some obstacles that will need to be resolved before the electronic transfer program is fully implemented, such as an amendment to the current requirement that a transfer include both the seller's and buyer's signature. CLASSIFYING ALLOWANCE TRANSACTIONS AND CAPTURING TRANSACTIONS COSTS The primary interest of most users of the ATS involves a measure of "meaningful transactions." A transfer between units owned by the same utility might be meaningful for analysis of the environmental effects of sulfur dioxide emissions. Such a transfer also may be meaningful for an assessment of cost savings from flexibility resulting from transfers within a firm. However, the greatest interest among users of the ATS hinges around economically meaningful transfers between economically separate organizations, which provides evidence of the role of allowance trading for compliance. It also provides information about the value and availability of allowances that are of direct interest to utilities and brokers. The Acid Rain Division's current taxonomy for categorizing allowance transactions and its previous taxonomy concentrate on organizing transactions to identify economically meaningful transfers among separate organizations. The main distinction between the previous classification scheme and the current scheme is the definition of inter-utility transactions. These approaches to the classification are presented in full in Dean and Kruger (1997).4 The categories that are used to organize transfers along with observations and potential problems are listed below. Although the categories themselves are stable, the categorizations of a transaction between two particular entities may become dated due to mergers and acquisitions. 4 "Using EPA's Allowance Tracking System to Assess the Allowance Market," Melanie Dean and Joe Kruger, Proceedings of AWMA/Acid Rain Electric Utilities Conference, January 1997. -5- Lile, Bohi, and Burtraw RFF 97-21 Figure 1. Current methodology: 1,918,979 Allowances Acquired by Utilities through Private Transactions Reported to ATS from 1/96 through 9/96 Broker to Utility 27.0% Intra-Utility 51.8% Fuel Co to Utility 1.9% Within Operating Co. 65% Within Holding Co. 35% Inter-Utility 19.4% Figure 2. Previous 1,895,061 Allowances Acquired by Utilties through Private Transactions Reported to ATS from 1/96 through 9/96 Broker to Utility 19.6% Intra-Utility 34.2% Fuel Co to 1.6% Inter-Utility 44.7% "Two Main distinctions from the current methodology - No breakdowns within the Intra-Utility trades and trades between two different operating companies are classified as Inter-utility trades (even if they are within the same holding co.) -6- Lile, Bohi, and Burtraw RFF 97-21 Inter-utility⁵ The current scheme defines inter-utility transfers as any transfer of allowances from one utility operating company's account to a different operating company's account, provided the operating companies are not controlled by the same holding company. These are viewed as economically distinct transactions. If the transaction involves a trade between two units affiliated with the same holding company, it is classified as an intra-utility transaction under the current scheme. The previous scheme classified all transactions between operating companies as inter- utility transactions, without a clear differentiation about their status with regard to a holding company. As a result, this scheme overstated inter-utility transactions compared to the current approach. Figures 1 and 2 illustrate the overstatement for the period of 1 January 1996 through 31 October 1996, although the magnitude of the overstatement is inflated in this example because January had the largest number of intra-utility transfers in the history of the trading program. To improve the accuracy of the classification, the current classification algorithm determines the unit's holding company affiliation before classifying the transaction. 6 One reason for emphasizing the difference among these type of trades in the current scheme is that these type of trades have different transactions costs. Reallocation Reallocation transfers are defined as any transfer from a unit account to a general account of the same operating company or holding company (and, in given circumstances, vice versa), any pooling activity, or any transfer in which the transferor is the partial owner of the transferee account (and vice versa). Since the statute requires each unit (not plant) to be in compliance at the end of each year, the initial allocation of allowances goes to the generating unit account. Firms may want to aggregate allowances from various unit accounts into one or more general accounts to more 5 At the workshop, the EPA presented two sub-categories for the Inter-utility classification: economically distinct and "among co-owners." As a result of the workshop, the EPA has moved the "among co-owners" distinction to the Reallocation classification. 6 Holding company information is not part of the ATS, but can be found in publicly available sources such as The Electrical World's Directory of Electric Power Producers. -7- Lile, Bohi, and Burtraw RFF 97-21 easily assess allowance holdings and needs for the future, and potentially to facilitate allowance transactions with other firms. However, when an allowance is used for compliance at a specific generating unit, the allowances must first be reallocated to the unit account associated with that generating unit. Furthermore, when a generating unit is owned jointly by multiple operating companies, the allocation of allowances to that unit also goes to the unit's account which, as in all cases, has an authorized account representative. That representative is responsible for reallocating the allowances among co-owners according to mutually agreed upon guidelines. Some co- owners view the allowances as assets and these owners want it reflected on their books, which provides one of many possible motivations for reallocation among co-owners. The transfer of allowances to a general account obfuscates the question of compliance at a unit until the time for truing up comes at the end of the year. Although the ATS data indicates when there are co-owners of a unit, the ATS does not indicate how an account for that unit will be managed among co-owners. Some state PUCs prevent automatic separation of allowances among co-owners. Consequently it is not possible for the EPA to employ a consistent algorithm for categorizing allowance reallocations among co-owners. One possibility would be for the agency to gather information from the manager of the unit account about the rules for reallocation, but this takes the agency in a direction other than its principle mission which concerns compliance activities. The goal of the EPA's categorization algorithm is to indicate these as reallocations rather than economically meaningful inter-utility transfers.⁷ The EPA currently evaluates the "ownership" on both sides of each transfer. This is done by determining the binding parties of both the transferee and transferor of allowances. If there is overlap of any degree, the EPA classifies this transaction as a Reallocation. As a result, this approach captures all transfers among co-owners as reallocations. This approach would be imperfect because in some cases a co-owner may reallocate an allowance originally allocated to a different unit to the co-owned generating unit for compliance purposes, while allocating allowances from the co-owned unit to other purposes. This reallocation may in fact be an Intra- utility transfer. However, for the sake of the methodology's validity, the EPA has decided to err 7 At the workshop, the EPA presented transactions among co-owners as Inter-utility transfers. However, as a result of the workshop, the EPA now considers all transactions among co-owners as Reallocations. -8- Lile, Bohi, and Burtraw RFF 97-21 on the conservative side by under reporting transactions that may be economically significant. Although the EPA has acknowledged that this methodology will underestimate Intra-utility transfers, this methodology is the best plan that has been suggested to date. Intra-utility Within the intra-utility transactions, there are two sub-categories: transactions between units within the same operating company (we label these "Intra-utility Class 1") and transactions between units owned by different operating companies within the same holding company (we label these "Intra-utility Class 2"). The sub-category distinction is useful because each sub-category has different transactions costs. An Intra-utility Class 1 transaction is defined as either a transfer from one unit account to another unit account within the same operating company or, in some cases, a transfer from a general account of one operating company to a unit account of the same operating company. An Intra-utility Class 2 transaction is defined as either a transfer from a unit account of one operating company to another unit account of a different operating company within the same holding company or, in some cases, a transfer from a general account of one operating company to a unit account of a different operating unit within the same operating company. Trades within the same operating company are likely to have lower transaction costs than trades between operating companies within the same holding company. In some cases trades among operating companies within the same holding company can have transaction costs that are higher than inter-utility trades. One reason is that the Public Utilities Holding Company Act (PUHCA) requires special reporting requirements for holding companies (Parent companies that are not holding companies under PUHCA are not required to follow these requirements.). These reporting requirements have an impact on transactions costs. 8 Any kind of reallocation or transaction between operating companies (under the PUHCA) must be documented at the market price. Since it is difficult to adhere to these reporting requirements, utilities turn to the market. This is one of the reasons for treating these trades differently. 8 Transaction costs for trades among operating companies within the same holding company may have be large when the operating companies are in different states and thus have different regulatory rules. -9- Lile, Bohi, and Burtraw RFF 97-21 Broker/trader to utility, utility to broker/trader, fuel company to utility and utility to fuel company These classifications are relatively self-explanatory. The EPA created these categories specifically to report on the data and to limit its interpretive role. However, there is the possibility of confusion in the case where some companies perform dual roles. To clarify the role of these "dual role" market participants, EPA determines from the participant its predominant role, and then EPA classifies them accordingly. One important concern in registering trading activity involving brokers is the desire to avoid double-counting brokered transactions. This could occur if a broker took possession of an allowance and registered its possession with the ATS. If this was the vehicle for managing a transfer between utilities, two transactions would be reported when only one meaningful transaction really occurred. An important factor concerns the type of brokers involved in a trade, and there are basically two types of brokers. One type takes title to an allowance and the other that doesn't take title. As a result, it is difficult to determine the incremental transaction costs associated with brokered transactions. Transactions involving brokers that do not take title will not be captured by the ATS, so double-counting will be avoided. Other All transactions that do not fall into the previous categories are classified as "other." So far the "other" category involves an insignificant number of transactions and allowances. AN ILLUSTRATION OF THE CURRENT TAXONOMY The current taxonomy will successfully organize the majority of private allowance transfers registered with the EPA into useful and descriptive categories. Tables 1 and 2 illustrate the distinction between reallocations, inter-utility and intra-utility trades. However, as the previous section indicates, there are still ample opportunities to mis-classify market activity. As noted above, the intra-utility trades involve two sub-categories in which we have called "Intra-utility Class 1" and "Intra-utility Class 2" trades. Furthermore, each sub-category had two possibilities: one strictly involving unit accounts and the other involving general accounts. Table 1 illustrates Intra-utility trades, both Class 1 and Class 2, involving only unit accounts. Table 1 also includes an example of an inter-utility trade. In contrast, Table 2 -10- Lile, Bohi, and Burtraw RFF 97-21 illustrates Intra-utility trades, both Class 1 and Class 2, involving general accounts. Furthermore, Table 2 combines intra-utility trades with reallocations to illustrate the similarity between these types of trades. Table 1. Examples of Inter-Utility and Intra-Utility Trades Inter-Utility unit account unit account operating company A operating company B Intra-Utility Class 1 unit 1 account unit 2 account operating company A operating company A Intra-Utility Class 2 unit 1 account unit 2 account operating company A operating company B subsidiary of holding company subsidiary of holding company Z Z Table 2. Examples of Reallocations and Intra-Utility Trades Reallocation Reallocation unit 1 account general account unit 1 account operating company operating company A operating company A A Reallocation Intra-Utility Class 1 unit 1 account general account unit 2 account operating company A operating company A operating company A Reallocation Intra-utility Class 2 unit 1 account, general account, unit 2 account, operating company A operating company A operating company B subsidiary of holding subsidiary of holding subsidiary of holding company Z company Z company Z -11- Lile, Bohi, and Burtraw RFF 97-21 Table 3. Examples of Trades involving Co-owners Reallocation Inter-Utility unit 1 account general account unit account co-owned by operating company A operating company operating companies D A, B, C Reallocation Reallocation* unit 1 account general account unit 2 account co-owned by operating company A co-owned by operating companies operating companies A, B, C A, B, C Reallocation Reallocation* unit 1 account general account unit 2 account co-owned by operating company A co-owned by operating companies operating companies A, B C,D A: subsidiary of C: subsidiary of holding company Z holding company Z B: subsidiary of D: subsidiary of holding company Y holding company X Reallocation Reallocation unit 1 account general account unit 1 account co-owned by operating company A co-owned by operating companies operating companies A, B, C A, B, C Inter-Utility unit account unit account co-owned by operating company C operating companies A, B Reallocation* unit 1 account unit 2 account co-owned by co-owned by operating companies A,B operating companies A,C,D Reallocation* unit 1 account unit 2 account co-owned by operating company c operating companies A, B subsidiary of A: subsidiary of holding company Z holding company Z B: subsidiary of holding company Y * This is an example that could be classified as an Intra-utility Class 1 transfer and as a result would not be captured as such by the ATS. -12- Lile, Bohi, and Burtraw RFF 97-21 Table 3 captures the issues involving co-ownership of generating units. There is the possibility of confusion in classifying Reallocations, Intra-utility or Inter-utility where a unit has multiple owners. As explained above, the EPA's categorization algorithm is to determine the binding parties (co-owners of generating unit accounts or general accounts) of both the transferee and the transferor of a transaction. If there is any degree of overlap the trade is classified as a Reallocation.⁹ If ownership differs on both sides of the trade, the trade is considered an Inter-utility trade. This approach is imperfect because in some cases a co-owner may reallocate an allowance originally allocated to a different unit to the co-owned generating unit for compliance purposes, while allocating allowances from the co-owned unit to other purposes. This reallocation may in fact be an Intra-utility Class 1 transfer. Due to this classification methodology, the EPA's current taxonomy will underestimate Intra-utility transfers. Nonetheless, since the Reallocation methodology errs in a conservative manner, this categorization algorithm is the best option short of expanding the EPA's data collection. It is evident that, as the number of co-owners increase, it is more difficult to categorize the trades. Furthermore, as noted before, as operating companies merge, this problem is exacerbated. STRATEGIC BEHAVIOR IN TRADING ACTIVITY Several hypothetical or imagined behaviors by participants in the allowance market have led to questions about whether there is strategic behavior that will obscure important allowance trading activities. One suggestion is that brokers or utility companies might "churn" the market in order to register a false level of trading activity. However, according to participants at the RFF workshop, brokers do not perform extra transactions just to "show" activity. Furthermore, the ATS does not capture transactions unless they are reported to the ATS, which involves a further level of accounting effort that would discourage "churning" the market. An obvious motivation for strategic behavior stems from the current regulated nature of the utility industry, both for purposes of environmental compliance and for cost recovery. There are numerous interested parties including environmental groups and rate payers who would like to have a say in the decisions of utilities. In this light, maneuvers to obfuscate 9 As noted before, this categorization algorithm may miss some economically distinct transactions. -13- Lile, Bohi, and Burtraw RFF 97-21 trading activity might be viewed as a fundamental contradiction between the interests of the public and the private. Some observers have suggested that utilities have occasionally used forward contracts to postpone taking possession of allowances and thereby avoid or postpone either regulatory interference or the attention of environmental watchdog groups. The impending restructuring of the electricity industry, accompanied by some degree of deregulation or "re-regulation," also has strategic implications for participants in the allowance market. Allowance trading activity can be a signal of future investment plans. Further, since allowances constitute a sizable asset themselves, a utility may want to veil its activity in the allowance market, for competitive reasons. Since trades do not have to be recorded with the ATS until they become relevant for compliance, there may be less information content about actual trading volume in the ATS in the future than there has been to date. In the future, the motivation for the use of instruments such as forward contracts may follow from the competitive pressures of the industry. For these reasons there is an apparent conflict between the interests of analysts who wish to obtain detailed information about the market, presumably through the ATS, and the interests of participants in the market who wish to keep information about their own plans and operations private. The paradox is that to the extent the interest for public disclosure prevails, providing a better measure of market performance, the market may do less well, as participants retreat from public scrutiny or the scrutiny of their competitors. The advice from many participants seems to be that the market will work best if it is left alone. However, if the SO₂ trading program and the ATS are to serve as a basis for other regulatory experiments at the national or state level, there will have to be some meaningful way to gauge the performance of these institutions and to garner lessons for the design of new ones. WHAT NEEDS TO BE DONE TO IMPROVE THE MARKET, AND IS THERE A ROLE FOR ATS IN DOING SO? This question invoked opposing views from participants at the RFF workshop. One view argued there is nothing wrong with the ATS or the market. The market is currently thin and in time will become more active. Furthermore, the only existing problem is "bad" information. From this perspective, when the EPA puts out incomplete information it leads to confusion. Moreover, the market doesn't need the EPA to analyze the data because private -14- Lile, Bohi, and Burtraw RFF 97-21 market analysts can do it. Private sources are supplying market information that is valued higher than that supplied by EPA. A contrasting view is there are inadequacies in the market and that private sources of market information that may cause confusion. In particular, private sources of information may not be reliable for policy evaluation purposes. The EPA can provide a source of information that in principle would be free of any bias. This view prompted the following rejoinder: "What makes the EPA's interpretation right and all of the others wrong?" If the information that EPA supplied was transparent and objective, then there would be no opposition to the EPA supplying it. However, if the information was transparent and objective, then why would the market need it? However, one might answer that the EPA is the primary institution with perspective, incentive and responsibility to identify shortcomings in the market if they exist. The organization of data in the ATS is the way such potential shortcomings can be identified. In many cases, the proprietary interests of the private parties may limit their disclosure of data or it may render data too expensive for "public interest" groups seeking access. There is also a difference of opinion over how well the market is working. Many market observers note that some potentially important traders have yet to get into the market, with the result that sizable potential cost savings are unrealized. The ATS is the source of information that would allow an analysis of this issue. Nonetheless, to say the market is partly broken is not to say that it needs to be fixed, but indeed the infant market may do well if left alone, especially under increasing competitive pressures in the industry to find ways to reduce costs. In any event, there is little to suggest that manipulation of information in the ATS or changes in the collection of data are ways to get players in the market. OBSERVATIONS ABOUT THE FUTURE Several changes are occurring that signal a maturation of the allowance market. One is in the public attitude with respect to the role of allowance trading as a means to reduce the cost of pollution control. Many observers suggest that environmental advocacy groups are no longer interested in looking at each trade from an environmental perspective. These groups have finally moved over to the EPA's stance that, for environmental purposes, aggregate trades and emissions are important, not individual trades and emissions. This shift in attitude seems -15- Lile, Bohi, and Burtraw RFF 97-21 to indicate that one constraint on active participation in the market that has deterred some potential trading activity, the harsh scrutiny of advocacy groups and negative characterizations in the media, may be of diminished importance in the future. The second trend that bodes well for the allowance market is increasing competition in the electric industry and the resulting pressure to reduce costs. In the future this pressure may override other obstacles that may have hindered the allowance trading activities of some utilities to date. Ironically, though, the move toward a more competitive industry environment will also make it harder for analysts to gather information from market participants, as it will be in the interests of those participants to veil their activities for strategic competitive purposes. In Phase I of the SO2 trading program, which began in 1995, the goal of most utilities has been to over-comply with necessary emission reductions in order to bank allowances for Phase II. In doing so, utilities have been able to put off capital costs as long as possible, including installing scrubbers. Since there is an abundant supply of Powder River Basin (PRB) coal (low sulfur coal) at relatively low prices, compared to the costs of capital investments for compliance, there may be further delays in the capital investments compared to what was expected when the program was adopted in 1990. The opportunity to delay large investments is consistent with the value of waiting in making capital investments in an uncertainty environment. The implication is that the flexibility implicit in the allowance trading program will be put to great use. One area where market participants may desire greater information concerns a timely report of emissions data. Although the brokers and the utilities in general currently do not look at the emissions data, but improved access to emissions data would be useful in determining what companies are actually doing, such as, for example, a decision to burn low or high sulfur coal. SUMMARY AND RECOMMENDATIONS The fundamental test for the EPA's current taxonomy is whether it provides information that is potentially "wrong" or easily "misinterpreted." Workshop participants agree that the current classification scheme passed this test and offer widespread support for the EPA's current methodology for categorizing allowance trading activity. In summary, the -16- Lile, Bohi, and Burtraw RFF 97-21 current classification provides information that is of widespread interest and runs little risk of providing false information. Our analysis and the discussion at the workshop lead us to recommend the EPA proceed with its current classification methodology. However, we offer the following suggestions to clarify the methodology. 1. Documentation of the definition and limitation for the methodologies should be available with the data. 2. Replace all references to "holding company" with "parent company." In addition, it is suggested that a footnote describing why the term "holding company" was avoided (i.e., Public Utility Holding Company Act) be incorporated into the online documentation. 3. The documentation should incorporate visual examples to delineate the differences between a Reallocation, an Inter-utility trade and an Intra-utility trade. 4. The documentation should incorporate a general note indicating that the ATS is designed primarily for compliance purposes and has limitations when used for evaluation purposes. Further, this note should indicate the ATS only captures trade activity that is recorded with the EPA. Some trades may not be captured by the ATS until those allowances are used for compliance. Hence, the ATS transaction date may not be the actual trade date. Other limitations on the ATS data as a measure of allowance trading activity should be made explicit. 5. The EPA should provide its data in an interactive file format. The interactive format should allow the user to conduct various searches such as the number of inter-utility transactions, number of interstate transactions, number of transactions within a state, transaction type, history search (as indicated above) and so on. 6. Due to the concerns regarding transactions costs and market analysis, it is recommended that the EPA incorporate titles for the Intra-utility sub-categories. We utilized "Intra-utility Class 1" and "Intra-utility Class 2" as titles to make the discussion regarding intra-utility trades as transparent as possible. -17- Lile, Bohi, and Burtraw RFF 97-21 Participants at the November 8 meeting included: Carlton W. Bartels, Director of the Environmental Brokerage at Cantor Fitzgerald; Elizabeth M. Bailey and A. Denny Ellerman, from MIT's Center for Energy and Economic Policy Research; Doug Bohi, Dallas Burtraw and Ron Lile, from Resources for the Future; Daniel Chartier, Manager of Emissions Trading at Wisconsin Electric Power Company; Melanie Dean, Joe Kruger, Brian McLean, Sharon Saile, Alex Salpeter, Claire Schary, Mary Shellabarger, Janice Wagner, all from the EPA's Acid Rain Division; Andrew Ertel, Manager of the Emissions Brokerage Desk at Natural Resources Group, Inc.; Gary Hart of Southern Company Services, Inc.; Ken Rose, from the National Regulatory Research Institute at Ohio State University; George Spencer, Editor of the Air Daily. -18-