Initial Draft – Please Do Not Quote Without Permission
Minimizing Carbon Leakage under Open Trade:
Strategies for the Allocation of Pollution Permits
Carolyn Fischer
fischer@rff.org
Resources for the Future
1616 P Street, NW
Washington, DC 20036
Alan K. Fox
Alan.Fox@usitc.gov
U.S. International Trade Commission
500 E Street, SW
Washington, DC 20436
Key Words: emissions trading, output-based allocation, tax interaction, carbon leakage
JEL Classification Numbers: Q2, Q43, H2, D58, D61
Initial Draft – Please Do Not Quote Without Permission
Minimizing Carbon Leakage under Open Trade:
Strategies for the Allocation of Pollution Permits
Carolyn Fischer and Alan Fox*
Introduction
Among US policy makers, an important criticism of the Kyoto Protocol, the
international agreement to reduce emissions of the greenhouse gases that cause global
climate change, is the lack of binding emissions targets for major emitters among
developing countries. In particular, energy-intensive industries worry that a policy (like a
cap-and-trade program for CO2) that levies a price on domestic emissions alone will
distort the playing field with their competitors in nonparticipating countries.
Policymakers express concern that such trade impacts will result in a partial undoing of
their efforts to reduce emissions, known as “carbon leakage.” Nor is leakage strictly an
international phenomenon. Distortions also can arise within the domestic economy if the
environmental policy is unevenly applied, due to technical, administrative, or other
concerns. Sectors covered by the policy may fear a reallocation of resources toward the
uncovered sectors, which would then also require additional or more stringent
regulations.
Bernard et al. (2001) show that such inefficient reallocations can be prevented by
taxing the unregulated sectors, according to the emissions embodied in their output.
However, when unregulated sectors reside abroad, such a tax may be prohibited by WTO
rules and, in any case, would only apply to imports, not to all the unregulated production.
Among domestic unregulated sectors, such a tax would be as difficult to implement as a
full downstream emissions trading program. Barring such a tax, they show that the next
best policy is to subsidize the output of the regulated sectors. The optimal subsidy then
reflects the value of the emissions crowded out by additional output in that regulated
sector.
One way of implementing such a subsidy in a traditional emissions cap-and-trade
program is by allocating the permits to the affected sectors based on their output. The
value of additional permits represents an incentive to produce more, offsetting part of the
*Alan
Fox is with the Office of Economics of the United States International Trade Commission,
Washington, DC. The opinions and ideas expressed herein are solely those of the authors and do not
represent in any way the views of the U.S. International Trade Commission or any of its individual
Commissioners.
Initial Draft – Please Do Not Quote Without Permission
price increase induced by the emissions regulation itself. Another way is by setting
performance standards; in this case, each sector must meet an average emissions
requirement. The effect is in theory identical—each firm must surrender permits
according to its emissions and receives an allocation according to its output. However, in
practice it is difficult to set performance standards such that marginal cost equalization is
met, unless the permits are also tradable across sectors, and it is more difficult to ensure
that a particular emissions target is met.
These mechanisms have been of interest in researchers concerned about preexisting tax distortions. Labor taxes distort the consumption-leisure tradeoff, and
environmental regulation that further raises consumer prices exacerbates those costs (a
collection of the literature is available in Goulder 2002). Since the output subsidy
embodied in performance standards limits those price increases, these mechanisms can
outperform a system of grandfathered emissions permits (Goulder et al. 1999, Parry and
Williams 1999, Fullerton and Metcalf 2001).
In an international context, trade distortions can be as important as tax distortions:
while all domestic sectors are covered by the emissions trading program, trade partners
are not, causing carbon leakage. Fischer and Fox (2004) consider the effects of different
allocation mechanisms, including output-based allocation, on the efficiency and
distributional effects of a carbon emissions trading program in the US, when both tax and
trade distortions are taken into account. They find that the output subsidies implicit in
OBA mitigate tax interactions, which can lead to higher welfare than grandfathering.
However, the rule for determining the sector allotments, which determines the effective
output subsidy, matters. OBA with sectoral distributions based on value added generates
effective subsidies similar to a broad-based tax reduction, performing nearly like
auctioning with revenue recycling, which generates the highest welfare. OBA based on
historical emissions supports the output of more polluting industries, which more
effectively counteracts carbon leakage, but is more costly in welfare terms. Importantly,
they also find that industry production and trade impacts among less energy-intensive
sectors are also shown to be quite sensitive to allocation rules.
This paper analyzes tradeoffs in choosing different allocation mechanisms in
order to achieve overall emissions reductions goals while minimizing intra- and
international leakage and minimizing the traditional measures of welfare loss associated
with emissions reductions. It expands on the preceding literature in two important ways.
First, it allows for leakage among domestic sectors, by dividing the economy into sectors
that are more likely to be covered under an emissions trading program and those that
2
Initial Draft – Please Do Not Quote Without Permission
would remain uncovered. Second, it assumes the policy goal is to meet a net emissions
constraint—that is, to ensure true reductions, net of leakage—rather than a simply
domestic emissions constraint. This assumption helps judge the true welfare costs and
sectoral impacts of reducing global emissions.
To lend greater currency to the analysis, reductions will be considered in the
framework outlined by the Climate Stewardship Act of 2003, co-sponsored by Senators
John McCain and Joseph Lieberman. The goal of the paper is to uncover certain rules for
identifying sectors and mechanisms for minimizing the costs of broader emissions
reduction when considering the interaction of one country's emissions policies with
unregulated emissions and labor market distortions in other countries.
The analysis is carried out in a modified GTAP framework augmented with a
labor-leisure choice. Such a framework allows for the consideration of how distortionary
tax instruments may be offset by revenues generated from permit auctions or may be
exacerbated by policies that raise prices. The current implementation employs the GTAP
version 6 (Release Candidate) database.
Description of Model
The model and simulations in this paper are based on version 5 of the
GTAPinGAMS package developed by Thomas Rutherford and documented for version 4
of the dataset and model in Rutherford and Paltsev (2000). The GTAP-EG model serves
as the platform for the model outlined here. Fischer and Fox (2004) was based on the
same model. This paper extends that model, adapting the framework to employ the latest
available GTAP database, Version 6.0 Release Candidate. This allows us to update the
analysis to 2001, the base year of the latest GTAP database.
Two features are added to the GTAP-EG structure to allow us to model the impact
of an output-based permit allocation scheme. First, the appropriate structure for
simulating an output-based allocation scheme must be incorporated into the model. Next,
the household is given a labor-leisure choice so that labor taxes are distorting. This
distortionary tax allows us to conduct simulations recycling revenue from pollution
permits to offset the distorting tax instrument. Tax instruments are also selectively
incorporated into the framework of the model, so that the transport sectors may be treated
differently from other heavy users of energy. This allows for a modeling approach more
consistent with that proposed in the McCain-Lieberman legislation.
3
Initial Draft – Please Do Not Quote Without Permission
Incorporating Output-Based Allocation
Several changes need to be made to the GTAP-EG code to incorporate outputbased allocation of pollution permits. The profit function is not directly accessible in the
MPSGE framework. Instead, we incorporate output-based permit allocation through the
production function as a tax/subsidy, combining it with side constraints on the values of
 and e to duplicate the effect on the profit function above. Additionally, we create an
additional composite fossil fuel nest to production. This allows us to incorporate the
pollution permit as a Leontief technology, allowing us to track pollution permits through
the model.
In the original GTAP-EG model, the treatment of energy goods does not allow for
tracking of permits by sector. In order to track pollution permits, we need to ensure that
one permit is demanded for each unit of carbon that enters into production. This is
accomplished by separating out the energy goods into a separate activity, a Leontief
technology combining the polluting inputs with permits, into a new composite good
(labeled ffi in the code for fossil fuel input). The composite of permit and energy input is
then included in the production block for the output good (y).
The next step is to incorporate a distortion in the form of an endogenous tax into
the sector’s production function. This tax, z, will allow us to mimic the impact on the
firm of an output-based subsidy. This is accomplished by including constraints to
establish values for  , e , and the relationship between z,  , and e .
The sectoral unit demand for pollution permits,  , is defined as

i ,r     fe,i ,r 
fe

 fe,i ,r 

y i ,r 
where  fe,i ,r is the carbon coefficient for final energy good fe {coal , oil , gas} in sector i
of country r. The variable  fe,i ,r represents the demand of the fossil fuel input. That is,
the unit demand for permits is equal to the unit demand for carbon.
The allocation of pollution permits is defined simply as the baseline unit demand
for carbon multiplied by the percentage cap (  r ) on emissions:
 

ei ,r   r  fe,i ,r 
 fe yi ,r 
4
Initial Draft – Please Do Not Quote Without Permission
Note that the parameter y i ,r does not appear in the MPSGE representation in
Appendix A. This is because the baseline value of output in the program is calibrated to
1.
The net demand for permits and the tax instrument in the production function are
then tied together through a final constraint:
zi ,r piy,r   r i ,r  ei ,r 
That is, the unit tax/subsidy on output is equal to the value of the net unit demand
for pollution permits. The MPSGE representation of these equations can be found in
Appendix A.
Incorporating Labor-Leisure Choice
The GTAP-EG model has also been extended by incorporating a labor-leisure
choice into the household’s decision. The procedure is documented in Fox (2002).
Incorporating a labor-leisure choice allows us to treat the labor tax as a distorting tax,
hence giving us a distorting policy instrument to offset with auction permit revenues.
Since we have no data on labor taxes within the GTAP-EG database, we assume a labor
tax rate of 40 percent within Annex B countries and a 20 percent tax rate within all other
countries and the Rest of World.
Emissions Permit Program Structure
To lend greater currency to the simulation results, we choose to follow the general
outline of the cap-and-trade scheme found in the Climate Stewardship Act of 2003, S.139,
hereafter referred to by the names of its co-sponsors, Senators John McCain and Joseph
Lieberman.1 McCain Lieberman treats a much broader spectrum of airborne pollutants,
but this paper will concern itself only with CO2 emissions. All references to pollution in
this paper are to the carbon equivalent of these CO2 emissions. McCain-Lieberman
foresees two classes of fossil fuel consumers who are directly affected by a cap-and-trade
scheme. First, major point source emitters (those responsible for more than 10,000
metric tons of CO2 , or about 2,700 metric tons of carbon) are subject to a cap-and-trade
system. We model this as a simple carbon permit requirement on the use of all final
energy goods in the model—coal, refined petroleum products, or natural gas.
Unfortunately, our database only provides information on the size of sectors, not on the
1
For a more complete analysis of McCain-Lieberman, see Pizer and Kopp (2003).
5
Initial Draft – Please Do Not Quote Without Permission
size of individual firms with those sectors. We therefore assume that the covered sectors
are covered in their entirety. The sectors subject to the cap are

petroleum and coal products (refined);

natural gas;

iron and steel industry;

chemical industry;

non-ferrous metals;

non-metallic minerals; and

paper, pulp, and print.
This overstates somewhat the magnitude of covered pollution, depending on the
distribution of firm size within each of these sectors.
Carbon emissions originating from transportation fuels such as gasoline, diesel,
and jet fuel also generate a permit liability. Our model and database lack the sectoral and
commodity detail to permit analysis by fuel type. Instead, we require that major users of
transportation fuels (air transport, water transport, other transport, and household
demand) pay a price for refined petroleum products that reflects the cost of obtaining the
necessary pollution permits. This both understates and overstates the coverage of
emitters. It overstates coverage by including all consumption of petroleum and coal
products, whether or not the product is a transport fuel. However, it also understates
coverage, because it excludes transport fuels used by sectors that are not covered by the
cap or are not transport sectors (e.g., gasoline and diesel consumed by the construction
industry or agriculture). However, since final demand and the transport sectors are
covered, we believe that the substantial majority of such emissions are covered.
Policy Scenarios
Defining the Cap
To focus on the costs of leakage, we assume the emissions target is equal to a 14
percent reduction in benchmark emissions in covered sectors, 186.15 million metric tons
6
Initial Draft – Please Do Not Quote Without Permission
in the 2001 data set. This differs from Fischer and Fox (2004) in that the target
reductions must be achieved net of carbon leakage to other sectors and other countries.
Permit Allocation
The legislation calls for gratis allocation of permits to the covered industries, while
providing for the formation of a Climate Change Credit Corporation (CCCC) to receive
pollution permits on behalf of final demand and the transport sectors (air, water, and
other). We have interpreted this as applying a tax on purchases made by households and
transport sectors of petroleum and coal products, with permit revenue being rebated back
to the representative agent in the model through a lump-sum tax rebate. The covered
industrial sectors, listed above, are treated differently, as we describe below. Fiscal
balance in the model is maintained by adjusting the labor tax rate (set at 40 percent in the
benchmark).
A total of four allocation mechanisms are employed to explore the implications of
allocation mechanisms in a program such as the McCain-Lieberman proposal. In
addition to a simulation designed to match the legislation as closely as possible, we
consider several alternative scenarios with slightly different mechanisms and
implications. These scenarios vary the allocation rules for the covered sectors only; in all
cases, permits assigned to the CCCC are treated as in the grandfathering scheme, with
revenue rebated lump-sum to the representative agent.
Grandfathering: McCain-Lieberman calls for grandfathering permits to the major
emitters in proportion to their historical emissions. This method implies a lump-sum
rebate to firms equal to the value of the permits granted.
Auctioned Industrial Permits: Permits for industrial users are auctioned, and the
revenue is used to reduce the distorting labor tax.
Output-Based Allocation Based on Historical Usage: Permits are allocated to
firms based on output shares in covered sectors, sector shares are based on historical
emissions, and government revenue is held constant through a labor tax.
Output-Based Allocation Based on Value-Added: Permits are allocated to firms
based on output shares in covered sectors, sector shares are based on historical shares of
value added, and government revenue is held constant through a labor tax.
7
Initial Draft – Please Do Not Quote Without Permission
Results
Summary
When maximizing welfare subject to a net emissions constraint in a situation with
incomplete coverage and carbon leakage, we find a striking result: even with labor tax
distortions, auctioning permits to the covered sector and recycling the revenues is not a
second-best optimal strategy. Rather, output-based allocations with a value-added rule
for sectoral distributions proves less costly. It minimizes changes in overall production,
employment and the real wage.
Table 1: Percentage Change in Summary Indicators
Indicator for United States
Welfare (equivalent variation)
Production
Employment
Real wage
Labor tax change (percentage points)
Permit price ($/metric ton)
Auction All GF Firms OBA Hist OBA VA
+ Ltax
+ Ltax
+ Ltax
+ L Tax
-0.050
-0.154
-0.135
-0.046
-0.45
-0.70
-0.48
-0.40
0.12
-0.31
-0.14
0.02
0.47
-1.09
-0.47
0.10
-1.82
1.05
0.35
-0.66
59.12
57.87
80.11
62.05
Other results are consistent with Fischer and Fox (2004): Grandfathering is the
most costly policy in welfare terms in these kinds of second-best situations. OBA raises
the marginal cost (permit price) of emissions abatement, since fewer reductions arise
from output substitution, and this effect is most striking under historical OBA. However,
output substitution through trade is less cost-effective when one cares about net global
emissions reductions, not just domestic ones. We find an important result that arises
from imperfect coverage of domestic emissions: while Fischer and Fox (2004) found that
historical OBA was more effective at combating international leakage, here it also tends
to induce more domestic leakage. Furthermore, when the covered sectors are more
limited limited, the implicit allocation subsidy is less well targeted to the sectors that
need it: the trade-sensitive, emissions intensive ones.
We explore the underlying reasons through the next set of tables.
Allocations
Table 2 reports the allocations by sector according to historical emissions shares
and value-added shares. By historical emissions, electricity and refineries get the largest
shares, whereas by value added, the pulp and paper and chemicals industries get the most.
8
Initial Draft – Please Do Not Quote Without Permission
Table 2: Initial Allocation of Permits by Sector (thousands of metric tons of carbon)
Carbon Permits
Historical
Value-Added
Allocation
Allocation
621,064
166,386
126,269
5,675
15,336
12,667
10,299
61,543
99,509
337,254
5,976
33,612
14,877
72,613
13,074
216,654
423,237
423,237
1,329,641
1,329,641
Sector
Electricity
Petroleum and coal products (refined)
Natural gas
Iron and steel industry
Chemical industry
Non-ferrous metals
Non-metallic minerals
Paper-pulp-print
Climate Change Credit Corporation
Total
Prices
Table 3 shows the equilibrium prices for the covered sector products according to
allocation regime. Since some of these sectors, particularly electricity, represent major
inputs to other sectors, their price changes help explain some of the impacts on the
uncovered sectors as well.
Table 3: Allocation and Price Changes for Covered Sectors
Percentage Change in Price
Auction
GF
Hist. OBA VA OBA
Sector
Electricity
Petroleum and coal products (refined)
Natural gas
Iron and steel industry
Chemical industry
Non-ferrous metals
Non-metallic minerals
Paper-pulp-print
11.63
2.96
-1.10
0.79
1.12
0.98
0.75
0.20
11.56
3.09
-0.93
1.10
1.36
1.26
1.02
0.47
-1.25
-2.41
-0.50
0.04
0.05
0.01
0.08
0.00
8.75
2.50
-0.67
-1.99
-1.98
-1.39
-2.55
-3.42
In most cases, the price increases are highest under grandfathering. With
historical OBA, prices fall for the highest emitters, electricity and refining, since the
allocation subsidy more than compensates for the cost increases. With value-added
OBA, prices drop for the other covered sectors.
Production
Table 4 gives the percentage change in the value of production among the covered
sectors, the transportation sectors, and the rest. Compared to grandfathering, an auction
induces smaller reductions in output, due to the revenue-recycling effect that boosts the
9
Initial Draft – Please Do Not Quote Without Permission
real wage, labor supply, and demand. In fact, production expands among the uncovered
sectors with auctioning, which is the source of domestic leakage.
The two OBA scenarios cause smaller reductions in covered sector output,
particularly value-added OBA. Notably, historical OBA causes more contraction in the
transportation sectors due to the substantially higher permit price. Value-added OBA
causes more contraction among the uncovered sectors—even more than grandfathering—
since not only are their electricity and transportation input costs high, but their labor costs
are as well, facing more competition from the covered sectors.
Table 4: Percentage Change in Production
Production
Auction
Total
Covered
Transportation
Uncovered
GF
-0.45
-3.33
-4.27
0.08
Hist. OBA VA OBA
-0.70
-3.49
-4.45
-0.18
-0.48
-1.23
-5.03
-0.18
-0.40
-0.19
-4.80
-0.24
Emissions
Table 5 shows the distribution of carbon reductions across both the covered and
uncovered sectors in the economy. We see that some of the incidence of the emissions
permit policy is passed on to other sectors in the form of higher electricity and
transportation prices. In our scenarios, transportation prices always rise; however,
electricity prices are affected by the allocation mechanism. In scenarios where those
prices rise, output and thereby emissions contract in the uncovered sectors, as well as the
covered ones. In other words, the impact of the cost increase (such as compared to
imports) is more important than the incentive to substitute away from covered to
uncovered products. However, with historical OBA, which disproportionately rebates to
electricity, many sectors actually increase output—and thereby emissions.
Table 5: Change in Emissions by Sector and Baseline Emissions
(Thousands of Tons of Carbon)
Auction
Sector
Electricity
Petroleum and coal products (refined)
Natural gas
Iron and steel industry
Chemical industry
Non-ferrous metals
Non-metallic minerals
Hist. VA OBA
OBA
Percentage Change
-22.9
-22.8
-20.9
-21.7
-17.9
-17.8
-14.7
-18.0
-5.2
-5.4
-5.7
-4.9
-9.3
-9.2
-7.6
-7.5
-11.7
-11.6
-9.7
-7.0
-10.7
-10.7
-8.1
-7.7
-16.1
-15.9
-16.2
-12.7
Baseline
621,064
126,269
15,336
10,299
99,509
5,976
14,877
10
GF
Initial Draft – Please Do Not Quote Without Permission
Paper-pulp-print
Other transport
Water transport
Air transport
Crude oil
Coal
Agriculture
Transport equipment
Trade margins
Communication
Other machinery
Other mining
Food products
Wood and wood-products
Construction
Textiles-wearing apparel-leather
Other manufacturing
Commercial and public services
Final Demand
Total
13,074
168,972
9,881
119,663
4,982
100
12,635
4,495
17,764
426
3,934
45
12,474
2,794
2,872
3,185
3,624
31,851
260,543
1,566,642
-10.7
-11.6
-12.3
-16.5
-10.8
-25.7
-1.1
-3.2
-3.8
-3.9
-3.4
-5.5
-1.5
-3.7
-2.2
-2.8
-3.5
-3.5
-6.0
-15.5
-10.7
-11.6
-12.3
-16.4
-10.8
-25.6
-1.3
-3.2
-3.9
-4.1
-3.4
-5.5
-1.7
-3.7
-2.1
-3.0
-3.5
-3.7
-6.2
-15.4
-9.2
-12.7
-13.5
-18.9
-8.3
-25.0
0.9
0.8
0.4
0.3
0.6
0.3
1.9
0.4
0.9
1.2
0.6
0.5
-9.3
-14.9
-7.6
-11.9
-12.6
-17.7
-11.4
-25.2
-1.6
-2.7
-2.9
-3.2
-3.1
-2.9
-0.8
-3.1
-1.6
-2.1
-3.2
-2.8
-6.8
-14.8
Trade
Table 6 shows the change in net exports by sector, which is an indicator of
leakage (though carbon leakage also depends on the emissions rates of the foreign
competitors). The OBA scenarios are better at mitigating these impacts for the covered
sectors than the others; however, we see a striking difference between the two. Historical
OBA actually increases net exports of electricity and refined petroleum products, due to
the strong implicit subsidy for them, while value-added OBA boosts the other energyintensive sectors.
Table 6: Change in Net Exports
Sector
Electricity
Petroleum and coal products (refined)
Natural gas
Iron and steel industry
Chemical industry
Non-ferrous metals
Non-metallic minerals
Paper-pulp-print
Other Transport
Water transport
Air transport
Crude oil
Coal
Agriculture
Auction
-1,127
-2,435
398
-969
-10,508
-1,559
-1,261
-473
-5,401
-301
-10,247
12,701
2,832
659
GF
Hist OBA VA OBA
-1,087
92
-947
-2,303
3,045
-2,638
428
171
50
-966
-164
1,363
-9,970
-1,408
10,706
-1,518
-194
1,038
-1,230
-224
2,938
-417
-56
5,563
-5,263
-7,202
-7,783
-295
-438
-500
-10,070
-12,320
-12,146
12,617
8,996
12,461
2,793
2,677
2,597
820
-71
-1,111
11
Initial Draft – Please Do Not Quote Without Permission
Transport equipment
Trade margins
Communications
Other machinery
Mining
Food products
Wood and wood-products
Construction
Textiles-wearing apparel-leather
Other manufacturing
Commercial and public services
Total
1,069
763
370
3,390
-108
428
205
83
169
1,793
6,912
-2,619
850
717
405
3,050
-100
602
219
67
379
1,583
6,453
-2,235
402
285
105
1,017
4
-93
-3
15
150
698
2,040
-2,475
-2,088
-685
-232
-2,685
-277
-1,151
-780
-36
-1,129
-2,492
-3,960
-3,922
Note that the allocation regime in the covered sectors also causes substantial
differences in the net exports of the uncovered sectors, due to changes in their input costs.
In particular, while most of these sectors gain exports in the other scenarios, they lose
significantly with value-added OBA—enough to increase the total net export loss for the
economy.
Carbon Leakage
Table 7 shows carbon leakage by trading partner, as a percentage of US
reductions. It shows clearly that OBA does a better job of minimizing leakage.
Somewhat surprisingly, the overall leakage difference between historical and value-added
OBA is small, though the distribution is quite different. For example, historical OBA
induces less leakage from Canada, presumably through electricity trade, while valueadded OBA causes less leakage from China, which trades more energy-intensive
manufactured goods.
Table 7: Carbon Leakage as a Percentage of US Reductions
Country
Canada
Europe
Japan
Other OECD
Former Soviet Union
Central European Assoc.
China (incl. HK & Taiwan)
India
Brazil
Other Asia
Mexico + OPEC
Rest of World
Total
Auction
GF
1.49
5.49
1.03
0.54
1.00
0.66
2.50
0.74
0.41
1.83
1.36
2.15
19.21
Hist OBA
1.46
5.42
1.02
0.54
1.00
0.65
2.47
0.74
0.41
1.81
1.35
2.13
19.00
12
0.65
4.94
0.96
0.53
0.87
0.59
2.15
0.68
0.35
1.64
0.84
1.70
15.90
VA OBA
1.01
4.93
0.85
0.48
0.81
0.58
1.85
0.62
0.33
1.43
0.97
1.92
15.79
Initial Draft – Please Do Not Quote Without Permission
Welfare
Table 8 shows the welfare impacts on the US and its trading partners. The world
fares best when the US uses auctioning with revenue recycling. While the US
experiences somewhat less of a welfare loss with value-added OBA, the rest of the world
experiences the least gain in that scenario.
Table 8: Change in Welfare (Equivalent Variation, Millions of 2001 $)
Auction
GF
Hist OBA
VA OBA
Country
Quantity % Quantity % Quantity % Quantity %
United States
-3,787 -0.05 -11,803 -0.15 -10,309 -0.13
-3,528 -0.05
Canada
-245 -0.06
-271 -0.06
-213 -0.05
-324 -0.07
Europe
5,648
0.10
5,636
0.10
5,602
0.10
4,494
0.08
Japan
1,533
0.06
1,660
0.06
1,393
0.05
907
0.04
Other OECD
-139 -0.05
-140 -0.05
-128 -0.05
-109 -0.04
Former Soviet Union
-639 -0.26
-588 -0.24
-829 -0.34
-1,005 -0.41
Central European Assoc.
228
0.08
187
0.06
246
0.08
352
0.12
China (incl. HK & Taiwan)
249
0.03
436
0.05
148
0.02
-430 -0.05
India
350
0.10
349
0.10
349
0.10
350
0.10
Brazil
174
0.05
159
0.05
189
0.06
239
0.07
Other Asia
377
0.05
536
0.07
278
0.04
-230 -0.03
Mexico + OPEC
-3,351 -0.36
-3,349 -0.36
-3,347 -0.36
-3,509 -0.37
Rest of World
-736 -0.07
-808 -0.07
-787 -0.07
-554 -0.05
Excluding US
3,449 0.025
3,805 0.028
2,899 0.021
180 0.001
World
-338 -0.002
-7,998 -0.038
-7,410 -0.035
-3,348 -0.016
Conclusion
Simple economic models have produced a clear preference for using tradable
emissions permits and auctioning the revenues to reduce labor taxes. However, this clear
preference has not translated into the policy sphere, in which emissions permits are nearly
always almost entirely given away, as in the US SO2 trading program, the Clear Skies
proposals, McCain Lieberman, and in the nascent EU carbon trading program.
Furthermore, the actual policy situation is rarely simple. Competition through
international trade, carbon leakage, and administrative issues likely hold more sway than
tax interaction problems. In the case of such multiple problems, the allocation of
emissions permits can have even more important effects. Specifically, grandfathering
can have even more costly effects in terms of welfare and competitiveness, though some
agents will receive a windfall. Output-based allocation may then emerge as a reasonable
option to combine gratis allocation with incentives that mitigate impacts on consumers
and trade. However, the mechanism for determining the sector-level distributions has
important consequences and tradeoffs. A value-added rule can result in higher overall
13
Initial Draft – Please Do Not Quote Without Permission
welfare and less leakage than with auctioned permits; however, its impacts on the
uncovered sectors can be more deleterious.
14
Initial Draft – Please Do Not Quote Without Permission
Bibliography
Bernard, A., C. Fischer, and M. Vielle. (2001) “Is There a Rationale for Rebating Environmental
Levies?” RFF Discussion Paper 01-31. Washington, DC: Resources for the Future.
Complainville, Christophe and Dominique van der Mensbrugghe. (April 1998) “Construction of
an Energy Database for GTAP V4: Concordance with IEA Energy Statistics,”
Manuscript. Paris: OECD Development Centre.
Fischer, Carolyn. (1999) “Rebating Environmental Policy Revenues: Output-Based Allocations
and Tradable Performance Standards,” Manuscript. Washington, DC: Resources for the
Future.
Fischer, Carolyn, and Alan K. Fox. (2004) “Output-Based Allocations of Emissions Permits:
Efficiency and Distributional Effects in a General Equilibrium Setting,” Discussion paper
04-37. Washington, DC Resources for the Future.
Fox, Alan K. (2002) “Incorporating Labor-leisure Choice into a Static General Equilibrium
Model,” Working paper. Washington, DC: U.S. International Trade Commission.
Fullerton, Don, and Gilbert Metcalf. 2001. "Environmental Controls, Scarcity Rents, and PreExisting Distortions," Journal of Public Economics, May.
Goulder, Lawrence H., ed. 2002. Environmental Policy Making in Economies with Prior Tax
Distortions, Edward Elgar.
Goulder, Lawrence H., Ian Parry, Roberton Williams III, and Dallas Burtraw. 1999. “The CostEffectiveness of Alternative Instruments for Environmental Protection in a Second-Best
Setting”, Journal of Public Economics 72(3):329-360, 1999.
Parry, Ian W.H. and Roberton C. Williams. 1999. Second-Best Evaluation of Eight Policy
Instruments to Reduce Carbon Emissions, Resource and Energy Economics, 21: 347-373.
Rutherford, Thomas F. and Sergey V. Paltsev (September 2000) “GTAPinGAMS and GTAPEG: Global Datasets for Economic Research and Illustrative Models,” Working paper.
Boulder, CO: Department of Economics, University of Colorado.