Sulfur dioxide control by electric utilities : what are the gains from trade?

TLDR: In this paper, the authors investigated whether the much-heralded fall in the cost of abating SO2, compared to original estimates, can be attributed to allowance trading and demonstrated that, for plants that use low-sulfur coal to reduce SO2 emissions, technical change and the fall in prices of low sulfur coal have lowered marginal abatement cost curves by over 50 percent since 1985.

Abstract: Title IV of the 1990 Clean Air Act Amendments (CAAA) established a market for transferable sulfur dioxide (SO2) emission allowances among electric utilities. This market offers firms facing high marginal abatement costs the opportunity to purchase the right to emit SO2 from firms with lower costs, and this is expected to yield cost savings compared to a command‐and‐control approach to environmental regulation. This paper uses econometrically estimated marginal abatement cost functions for power plants affected by Title IV of the CAAA to evaluate the performance of the SO2 allowance market. Specifically, we investigate whether the much‐heralded fall in the cost of abating SO2, compared to original estimates, can be attributed to allowance trading. We demonstrate that, for plants that use low‐sulfur coal to reduce SO2 emissions, technical change and the fall in prices of low‐sulfur coal have lowered marginal abatement cost curves by over 50 percent since 1985. The flexibility to take advantage of these chan...

Figures

TABLE 3: PHASE I (1995 AND 1996) COST ESTIMATES

TABLE 2: LONG-RUN (PHASE II, YEAR 2010) COST ESTIMATES

FIGURE 2: THE EFFECT OF CHANGES IN FUEL PRICES AND TECHNICAL CHANGE ON MARGINAL ABATEMENT COST FUNCTIONS

FIGURE 1: LOW-SULFUR AND HIGH-SULFUR COAL PRICES

TABLE 1: 1985 AND 1994 WEIGHTED AVERAGE MARGINAL ABATEMENT COSTS AND

Full text

Sulfur Dioxide Control by Electric Utilities:
What Are the Gains from Trade?
Curtis Carlson, Dallas Burtraw,
Maureen Cropper, and Karen Palmer
Discussion Paper 98-44-REV
April 2000
RESOURCES
for the future
1616 P Street, NW
Washington, DC 20036
Telephone 202-328-5000
Fax 202-939-3460
Internet: http://www.rff.org
©2000 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.

SULFUR DIOXIDE CONTROL BY ELECTRIC UTILITIES:
WHAT ARE THE GAINS FROM TRADE?
Abstract
Title IV of the 1990 Clean Air Act Amendments (CAAA) established a market for
transferable sulfur dioxide (SO
2
) emission allowances among electric utilities. This market offers
firms facing high marginal abatement costs the opportunity to purchase the right to emit SO
2
from firms with lower costs, and is expected to yield cost savings compared to a command and
control approach to environmental regulation. This paper uses econometrically estimated
marginal abatement cost functions for power plants affected by Title IV of the CAAA to
evaluate the performance of the SO
2
allowance market. Specifically, we investigate whether the
much-heralded fall in the cost of abating SO
2
, compared to original estimates, can be attributed to
allowance trading. We demonstrate that, for plants using low-sulfur coal to reduce SO
2
emissions, technical changes and the fall in low-sulfur coal prices have lowered marginal
abatement cost curves by over 50% since 1985. The flexibility to take advantage of these changes
is the main source of cost reductions, rather than trading per se. In the long run, allowance trading
may achieve cost savings of $700-$800 million per year compared to an “enlightened” command
and control program characterized by a uniform emission rate standard. The cost savings would
be twice as great if the alternative to trading were forced scrubbing. However, a comparison of
potential cost savings in 1995 and 1996 with actual emissions costs suggests that most trading
gains were unrealized in the first two years of the program.
Key Words: acid rain, sulfur dioxide, air pollution, Clean Air Act, Title IV, permit trading
JEL Classification Nos.: H43, Q2, Q4
This paper is also published in the Journal of Political Economy, Vol. 108, No. 6, 1292-1326.

I. Introduction ..................................................................................................................................2
II. Institutions...............................................................................................................................4
III. Methodology..............................................................................................................................6
A. Calculation of the Gains from Allowance Trading..................................................................6
The Role of Scrubbing vs. Fuel-switching................................................................................7
Computation of the Gains from Trade.....................................................................................8
B. Estimation of Marginal Abatement Cost Curves ....................................................................8
Econometric Model..................................................................................................................9
Data........................................................................................................................................12
Results of the Estimation.......................................................................................................13
Table 1: 1985 and 1994 Weighted Average Marginal Abatement Costs and SO2 Emission Rates
for Coal-Fired Units without Scrubbers.........................................................................................14
Figure 1: Low-Sulfur and High-Sulfur Coal Prices.........................................................................15
Figure 2: The Effect of Changes in Fuel Prices and Technical Change on Marginal Abatement
Cost Functions...............................................................................................................................16
IV. The Least-Cost Solution and Potential Gains from Trade in the Long Run.....................17
A. Preferred Estimates of the Least-Cost Solution....................................................................17
Electricity Output..................................................................................................................17
Input Prices and Technical Change ........................................................................................17
Retirement of Coal-Fired Power Plants.................................................................................18
Baseline Emissions.................................................................................................................18
Continuous Emissions Monitoring Data................................................................................18
Minimum Compliance Costs in the Preferred Case...............................................................18
B. Comparisons and Sensitivity Analyses.................................................................................19
Table 2: Long-Run (Phase II, year 2010) Cost Estimates..............................................................20
C. Potential Gains from Trade ...................................................................................................21
V. The Performance of the Allowance Market in 1995 and 1996..................................................23
Table 3: Phase I (1995 and 1996) Cost Estimates .........................................................................23
VI. Conclusions..............................................................................................................................25
Appendix A....................................................................................................................................27
Appendix B....................................................................................................................................30
Emission Equation and Cost Function Parameter Estimates.....................................................30
References ......................................................................................................................................32

2
SO
2
CONTROL BY ELECTRIC UTILITIES:
WHAT ARE THE GAINS FROM TRADE?
*
I. INTRODUCTION
For years economists have urged policy makers to use market-based approaches to
control pollution (taxes or tradable permits) rather than relying on uniform emission standards or
uniform technology mandates (command and control). This advice was largely ignored until the
1990 Clean Air Act Amendments (CAAA) established a market for sulfur dioxide (SO
2
)
allowances. Coupled with a cap on overall annual emissions, the SO
2
allowance market gives
electric utilities the opportunity to trade rights to emit SO
2
rather than forcing them to install
SO
2
abatement technology or emit at a uniform rate. By equalizing marginal abatement costs
among power plants, trading should limit SO
2
emissions at a lower cost than the traditional
command and control approach.
The SO
2
allowance market presents the first real test of the wisdom of economists'
advice, and therefore merits careful evaluation. Has the allowance market significantly lowered
the costs of abating SO
2
, as economists claimed it would? An answer in the affirmative would
strengthen the case for marketable permits to control other pollutants, such as greenhouse gases.
Conversely, if cost savings are small, this would have implications for the design (or even the
adoption) of market-based approaches to controlling pollution in the future.
The purpose of this paper is to evaluate the performance of the SO
2
allowance market.
Specifically, we ask two questions: (1) How much can the trading of permits reduce the costs of
controlling SO
2
, compared to command and control; i.e., what are the potential gains from trade?
(2) Were these trading gains realized in the first years of the allowance market? The answers
require that we estimate marginal abatement cost functions for fuel switching at all generating
units that do not scrub their emissions, calculate the expected cost of post-combustion abatement
(scrubbers), and compute the least-cost solution to achieving the cap on SO
2
emissions. The
difference between the least-cost solution and the cost under our counter-factual command and
control policy represents the potential static efficiency gains from allowance trading. We
compute these gains for 1995 and 1996, the first two years of the allowance market, and the
expected savings in 2010, when the emissions cap will be stricter and applied more broadly and
when the allowance market should be functioning as a mature market.
*
* This paper has benefited from comments and suggestions from Robert Chambers, Denny Ellerman, Suzi Kerr,
Richard Newell, Paul Portney and Byron Swift. We would especially like to thank Don Fullerton for his
exceptionally detailed comments which have significantly improved the paper. Carlson's work on this paper was not
part of his official duties at the National Oceanic and Atmospheric Administration (NOAA) and his employment
with NOAA does not constitute an endorsement by the agency of the views expressed in this paper. This paper does
not represent the views of the World Bank, its Executive Directors, or the countries they represent. The authors are
grateful to the Environmental Protection Agency for partial support of this study.

Carlson, Burtraw, Cropper, and Palmer RFF Discussion Paper 98-44-REV
3
The command and control policy against which we measure gains from allowance trading
is key to the analysis. A policy that would have imposed end-of-stack abatement technology
(scrubbing) would have been significantly more expensive than an emission rate standard applied
uniformly to all facilities.
1
A uniform emission rate standard provides firms with considerable
flexibility, including the opportunity to take advantage of technical change that is precluded under
a more rigid technology standard; hence it is a favorable characterization of a command and
control approach. In our analysis we evaluate the gains from allowance trading compared to each
of two command and control alternatives–forced scrubbing and a uniform emission rate
standard.
2
Our approach to evaluating the allowance market is very different from the approach used
by other observers to assess market performance. Both the Administrator of the USEPA and the
chair of the Council of Economic Advisors proclaimed the success of the allowance market by
comparing allowance prices (circa $100 per ton in 1997) with estimates of marginal abatement
costs produced at the time the CAAA were written (as high as $1500).
3
Since the former are
much lower than the latter, they concluded that the trading of SO
2
allowances has greatly reduced
the cost of curbing SO
2
emissions.
This argument is flawed for two reasons. First, it is inappropriate to judge how well the
allowance market is performing simply by comparing current allowance prices with ex ante
estimates of marginal abatement costs in the least-cost solution. Price can equal marginal
abatement cost even if many utilities that might benefit from trading fail to participate in the
market. Second, comparing current allowance prices with ex ante estimates of marginal abatement
costs shows only that the latter were too high; it does not mean that the allowance market was
responsible for the fall in marginal abatement costs.
4
Our analysis suggests that the above claims for the allowance market are misleading—
especially the suggestion that formal trading has lowered the cost of SO
2
abatement several fold.
In contrast, we reach the following conclusions:
(1) Marginal abatement costs for SO
2
are much lower today than were estimated in 1990.
Technical improvements including advances in the ability to burn low-sulfur coal at existing
generators, as well as improvements in overall generating efficiency, lowered the typical unit's
1. In 1983 the Sikorski/Waxman bill sought to reduce SO
2
emissions by requiring the installation of scrubbers (flue
gas desulfurization equipment) at the fifty dirtiest plants. Studies estimate that the annual cost of this proposal
would have ranged from $7.9 billion (OTA, 1983) to $11.5 billion (TBS, 1983), in 1995 dollars.
2. One justification for the use of an emission rate standard is that it is the approach used to regulate NOx
emissions under Title IV of the 1990 Clean Air Act Amendments.
3. On 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,"
The Commonwealth, March 31, 1997.) Likewise, in testimony before Congress, CEA Chair Janet Yellen noted,
"Emission permit prices, currently at approximately $100 per ton of SO
2
are well below earlier estimates . . ..
Trading programs may not always bring cost savings as large as those achieved by the SO
2
program . . . . . "
(Yellen 1998).
4. It should also be noted that the ex ante estimates of marginal abatement costs were generally for the second phase
of the program, and therefore cannot be compared with current allowance prices unless they are discounted to the
present.

Frequently asked questions

Q1. What have the authors contributed in "Sulfur dioxide control by electric utilities: what are the gains from trade?" ?

This paper uses econometrically estimated marginal abatement cost functions for power plants affected by Title IV of the CAAA to evaluate the performance of the SO2 allowance market. Specifically, the authors investigate whether the much-heralded fall in the cost of abating SO2, compared to original estimates, can be attributed to allowance trading. The authors demonstrate that, for plants using low-sulfur coal to reduce SO2 emissions, technical changes and the fall in low-sulfur coal prices have lowered marginal abatement cost curves by over 50 % since 1985. However, a comparison of potential cost savings in 1995 and 1996 with actual emissions costs suggests that most trading gains were unrealized in the first two years of the program. 

Q2. What future works have the authors mentioned in the paper "Sulfur dioxide control by electric utilities: what are the gains from trade?" ?

This suggests that attempts to estimate the future costs of other pollution control programs may be similarly flawed, especially given the difficulty in forecasting future trends in technological change. Second, their results suggest that, in designing an allowance market, it is important for policymakers to consider the source of trading gains and how these gains might change over time. As these price differences have diminished, so have potential trading gains. Lastly, their results suggest that it will take time for allowance markets to mature and, therefore, for the potential gains from trade to be realized. 

Q3. Why is the emissions of scrubbed coal relatively insensitive to sulfur content?

Because scrubbers remove about 95% of the sulfur content of coal, emissions are relatively insensitive to the sulfur content of coal burned. 

Q4. Why is the market for CO2 likely to generate large trading gains?

The market for CO2 is initially likely to generate large trading gains because coal-fired power plants, by converting to natural gas, can reduce their CO2 emissions at a lower cost than oil- and gas-fired plants. 

Q5. What is the important assumption that might bias their cost estimates downward?

The one assumption that might bias their cost estimates downward is their assumption that technical progress will continue from 1995 until 2010 at the same rate as between 1985 and 1994. 

Q6. What is the reason why utilities are less eager to trade allowances?

If potential gains from trade are small and transaction costs of using the market are substantial, utilities will be less eager to trade allowances. 

Q7. How much of the change in the MAC function is due to technological improvements?

The effect of technological improvements, represented by the vertical distance between curves (a) and (b) accounts for about 20 percent of the change of the MAC function, or a decline of about $50 per ton between 1985 and 1995. 

Q8. What is the alternative approach to modeling the sulfur content of coal?

24 An alternative approach to modeling the sulfur content of coal, used by Kolstad and Turnovsky (1998), is to allow plants to select sulfur content as a continuous attribute, given a hedonic price function for coal. 

Q9. What are the two factors that have lowered the trading gains?

The relatively lower trading gains that the authors predict for the allowance market in the long run are largely the result of two factors—declines in the price of low sulfur coal and improvements in technology that have lowered the cost of fuel switching. 

Q10. How much of the fall in the marginal abatement cost function is due to technological progress?

The effect of changes in fuel prices, represented by the vertical distance between curves (b) and (d) accounts for the remaining 80 percent of the fall in the marginal abatement cost function, a decline of about $200 per ton.average high sulfur coal price in their dataset is lower than the national average. 

Q11. What is the reason for the failure to take advantage of cost-saving opportunities to switch fuel?

As noted above, this failure to take advantage of cost-saving opportunities to switch fuel may be the result of inability to escape from long-term fuel contracts or insufficient incentives to find the lowest priced fuel as a result of regulatory fuel adjustment clauses (i.e. non-cost-minimizing behavior). 

Q12. What are the conditions to ensure the cost function is linearly homogeneous in input prices?

The following conditions are imposed to insure the cost function is linearly homogeneous in input prices: ∑ ∑ ∑∑ ∑ =α=α=α=α=α j j j jtjejqi j iji ,0 and 1 i, j = k, l, fls, fhs. 19. 

Q13. How do the authors assign emissions to units that share a boiler?

For those units that share boilers and/or stacks, the authors assign emissions based on the percentage of total heat input consumed by each boiler. 

Q14. What is the common way to calculate the marginal abatement cost curve?

Assuming that a premium must be paid per million Btu for low-sulfur coal, this implies that the marginal abatement cost curve slopes down as emissions of SO2 increase. 

Q15. What is the cost of abating an additional ton of SO2 emissions?

The marginal cost of abating an additional ton of SO2 emissions, may be derived from the fact that e=SO2/mmBtu, where mmBtu is millions of BTUs of heat input. 

Q16. What are the assumptions that result in a lower cost of scrubbing than were we?

These assumptions result in a lower cost of scrubbing than were the authors to assume that some of these scrubbers would be built after 1995, when a higher cost of capital for the industry would be applicable. 

Q17. How much will the average unit's marginal abatement cost fall?

If the current trend in technological improvements continues until the year 2010, this average unit's marginal abatement cost will fall by an additional $100 per ton ((d) to (e)).