Technical Support Document (TSD)
for the final Transport Rule
Docket ID No. EPA-HQ-OAR-2009-0491

Assurance Penalty Level Analysis Final Rule TSD

U.S. Environmental Protection Agency
Office of Air and Radiation
June 2011

1

Deliberative - Do not cite or quote


-------
Assurance Penalty Level Analysis Final Rule TSD

This Technical Support Document (TSD) supports EPA's determination that the final
Transport Rule's assurance provision penalty requirement provides sufficient deterrence against
a state exceeding its assurance level. Section VII.E in the final Transport Rule preamble
discusses the assurance provisions, including the allowance surrender penalty analyzed in this
TSD. This TSD is organized as follows:

1.	Background

2.	Approach

3.	Results

1. Background

The final Transport Rule's limited interstate trading programs include assurance
provisions to ensure that the necessary reductions will occur within each covered state. The
assurance provisions limit emissions from covered units in a state to the state's emissions budget
plus variability limit, i.e., the state's assurance level.

As described in preamble section VI.D, EPA used a multi-factor analysis to determine
each state's emissions budget. Subsequently, as described in VI.F, EPA determined variability
limits for each state that reflect a percentage of the state's budget (e.g., 10%). This variability
limit is then added to the state budget to yield the state's assurance level. If emissions from
covered sources in a state in a compliance period exceed the state's assurance level, then EPA
applies additional criteria to determine which owners and operators of units in the state will be
subject to an allowance surrender penalty of two allowances per ton for their share of the
emissions over the assurance level.1 This penalty is in addition to the standard program
requirement that owners and operators of covered units hold one allowance for each ton emitted;
therefore, for any emissions identified by EPA as being over the state's assurance level, the
relevant owners and operators must submit a total of three allowances per ton - one of which is
for standard compliance for emitting under the program, and two of which are for the assurance
provision penalty.

As discussed in preamble section VI.F, EPA does not find reason to expect that emissions
from covered sources in any state will exceed that state's assurance level. The description and
tables below describe a sensitivity analysis EPA conducted to determine whether the two-for-one

1 The assurance provision allowance surrender penalty addressed in this TSD is distinct from the
penalties, discussed in preamble section VII.F, that apply to the trading program requirement to
hold allowances sufficient to cover emissions for each compliance period.

2

Deliberative - Do not cite or quote


-------
allowance surrender penalty provides a sufficient deterrent to keep emissions from covered
sources in each state from exceeding the assurance levels.

2. Approach

To determine if a penalty of two allowances for every ton of excess emissions would be
sufficient, EPA used the Integrated Planning Model (IPM) to assess a "TRPenaltyScenario"
whose results, along with the results of other IPM model runs for the Transport Rule, can be
found in the docket. More information on IPM can be found in the Documentation Supplement
for EPA Base Case v.4. lOFTransport - Updates for Final Transport Rule, which is also in the
docket. This penalty scenario offered covered sources in each state the choice to emit beyond the
state's assurance level and incur a fine for each excess ton worth twice the value of an allowance
(in addition to having to submit one allowance for emitting each ton, per standard compliance
procedures). In this analysis, the "state assurance level" is the state's emissions budget plus the
state's variability limit and corresponds to the "state emissions assurance level" in tables VI.F-1
and VI.F-2 of the preamble. The size of the penalty was calculated as twice the allowance price
for the relevant pollutant taken from the IPM analysis of the final rule's remedy
(TRLimitedTradingFinal), as shown in Table 1.

Table 1. Allowance prices (2007$) in the final remedy (TR Limited Trading Final) and
Penalty Costs in the "assurance penalty sensitivity" run (TR Penalty Scenario)



Emission Allowance
Prices ($/Ton) from
final remedy run
2012 2014

Emission Penalty Costs
($/Ton) in the assurance
penalty sensitivity run
2012 2014

SO2 Region 1 (TR)

971

1,127

1,942

2,254

S02 Region 2 (TR)

576

663

1,152

1,327

NOx Annual (TR)

497

577

994

1,153

NOx Ozone Season (TR)

1,321

1,532

2,642

3,064

It is important to consider that while the effective fine in this scenario (twice the value of
the relevant pollutant's allowance price) is a technically valid representation of the final rule's
penalty structure, it is an analytic understatement of the actual deterrence value of this penalty in
practice. The penalty in practice will have more of a deterrent effect than what this scenario
models it to have for two reasons. First, the penalty cost as modeled is fixed at twice the
allowance price in the final remedy scenario (TR Limited Trading Final) for every ton of
emissions in excess of a state's assurance level. In reality, excess emissions would increase the
allowance price (and therefore the cost of the penalty itself) since allowances would have to be

3

Deliberative - Do not cite or quote


-------
bought and surrendered for the penalty, raising allowance demand and thus making them more
valuable. The modeling imposed a fine and did not adjust the allowance pool to account for
penalty surrenders, and so it understates the cost of the penalty incurred. Second, the model has
perfect foresight of all future emitting behavior and thus does not take any "risk" into account
when determining whether excess emissions are "worth it" at the penalty cost modeled. In
reality, owners and operators of covered units will assign a risk premium to the nominal penalty
consequence because they do not have perfect foresight and cannot be sure of their precise
emissions until the compliance period is complete. Therefore, program participants can be
expected to act more "conservatively" than the modeling would suggest when determining
whether excess emissions are "worth it" at the penalty cost modeled, which suggests again that
this analysis understates the deterrence value of the penalty in practice.

The sensitivity analysis presented in this TSD was based off of the main remedy analysis
presented throughout the preamble and Regulatory Impact Analysis for this final rule. In
common with that main remedy analysis, this sensitivity analysis assumed preliminary variability
limits that were smaller than the variability limits finalized in this rule. Because the final rule's
variability limits are larger than those analyzed in this sensitivity, the results presented below
overstate each state's economic interest in violating its assurance levels. In other words, the
analysis examined whether or not states have an economic interest in surpassing an upper bound
in permissible emissions in 2012 and 2014 (budget plus assumed variability limit) that is less
than the actual upper bound (budget plus finalized variability limit) imposed on states in 2012
and 2014 under the final rule. It therefore follows that the state's actual economic interest in
surpassing the actual upper-bound would be less than the projected results presented below
analyzed for the modeled (lower than actual) upper-bound. This relationship further increases
EPA's confidence in the conclusions it draws from the results presented below.

3. Results

EPA compared the state-level emissions in 2012 and 2014 in this analysis to the state
assurance levels to determine whether the penalty level deterred excess emissions. Tables 2
through 5 show the state assurance levels and modeled emissions from covered sources in each
state. The modeled allowance prices in 2012 and 2014 for each pollutant are shown at the
bottom of each table. The penalty for exceeding the assurance level would be equal to twice the
allowance price.

In no case do the covered emissions in a state exceed that state's assurance level in 2012
or 2014. This result indicates that the penalty offers a sufficient deterrent to ensure emissions do
not exceed assurance levels in 2012 and 2014. Even though the modeling of this scenario
understated the actual value of the deterrent in practice, in no state did the covered sources find it
economic to exceed the states' assurance levels in 2012 and 2014.

4

Deliberative - Do not cite or quote


-------
In some states, the covered sources in a state are modeled to have collective emissions
that are exactly equal to the state's assurance level. These projections occur because the model
operates under perfect foresight and perfect information, and it therefore allows sources in the
modeling to emit up to the state's assurance level with full certainty that emissions will respect
that constraint to the last ton emitted, successfully avoiding additional emissions which are
shown to be uneconomic under the assurance penalty. In reality, the deterrence value of the
penalty would likely lead a state's covered sources to act conservatively by emitting below a
state's assurance level (rather than exactly up to it) to ensure that unexpected fluctuations in
emissions do not result in a penalty.

In some cases, notably SO2 Group 1 states, the covered emissions are projected to be
significantly lower than the states' assurance levels in 2012. This would occur if covered
sources decide to reduce their emissions beyond what is required so they can bank the excess
allowances. These banked allowances could then used to cover emissions in future years.
Specifically, the SO2 Group 1 state budgets were determined by the feasible emission reductions
at $500/ton SC^in 2012 and $2300/ton SC^in 2014 (see Significant Contribution and State
Emissions Budgets Final Rule TSD). Covered sources in these states may decide to reduce their
emissions further than required in 2012 and 2013 and bank the unused allowances for use in
2014 and later years. This pattern effectively smoothes their emission reductions over time to
minimize total compliance costs in those states. The modeled allowance prices in those states
also reflect this smoothing of emission reduction patterns. For example, as seen in Table 2, the
2012 and 2014 projected allowance prices for Group 1 S02 states are closer to each other than
the marginal cost thresholds used to formulate their budgets in those years ($500 per ton in 2012
and $2,300 per ton in 2014). This banking behavior to smooth emission reductions over time is
shown in this analysis to be entirely consistent with each state's assurance levels in both 2012
and 2014.

As noted above, EPA's modeling of this scenario projects no instance in which covered
sources would find it economic to exceed a state's assurance level in any of the programs in
2012 or 2014. This analysis also projects that the penalty provides sufficient deterrence for
virtually all states in these programs over the 2020-2030 timeframe as well. However, the
projections appear to suggest small exceedances in two states in 2020 and in three states in
2030.2 In most of these cases, the projected exceedances are marginal - on the order of two

2 As previously noted, these findings are based on lower variability limits than included in the
final rule. EPA conducted a separate sensitivity analysis on the remedy (with results presented in
Appendix F of the RIA) incorporating the final variability limits. This analysis shows a dramatic
reduction in the number of states projected to approach their assurance levels in the 2020 and
2030 projections. For example, while the original remedy analysis (on which this TSD's
sensitivity analysis is based) projected in 2020 that 11 states would approach their assurance
levels for S02, 2 states for annual NOX, and 1 state for ozone-season NOX, the revised remedy

5

Deliberative - Do not cite or quote


-------
hundred tons of pollutant. EPA does not believe that these longer-run results actually indicate a
likelihood of these exceedances occurring. It is important to note that this modeling assumes
perfect foresight and perfect information, even into the 2020-2030 timeframe, and that under
those assumptions, unit owners and operators would be willing to expend banked allowances in
those years even to the point of paying assurance penalties on them. In reality, operators of
covered units do not have perfect foresight or perfect information and will therefore act more
conservatively than "optimal" banking patterns from IPM modeling would indicate.

As a result, EPA expects sources will collectively continue to respect their states'
assurance levels in those instances by sustaining the Transport Rule emission reductions into the
2020-2030 timeframe and banking allowances further into the future to guard against
unanticipated developments that the model does not capture. Consequently, EPA does not
believe that these limited small instances of projected exceedances in 2020 or 2030 are likely to
occur in the actual operation of these programs. However, EPA will monitor the pattern of
compliance with these programs over the long-run and will be prepared to adjust the penalty
accordingly if evidence suggests that increased deterrence would be necessary at those later
stages to encourage states to respect their assurance levels. At this point, EPA believes that it is
best to be sure in the initial years that the assurance penalty is effective in keeping emissions
within variability limits, while encouraging trading to lower costs and increase flexibility and
avoiding actions that have a chilling effect on activities. While doing this, EPA believe it is
important to remain mindful that we do want assurance of meeting emission reductions over
time.

EPA believes these findings support a determination that the penalty requirement of
surrendering two additional allowances for each ton of excess emissions provides a sufficient
deterrent in the final Transport Rule such that EPA does not expect the covered sources in any
state to exceed the state's assurance levels under these programs.

sensitivity analysis (including the larger, finalized variability limits) saw only 5 states approach
assurance levels for annual S02, no state for annual NOX, and 1 state for ozone-season NOX.
With many fewer states even approaching their assurance levels in the long term with the final
rule's variability limits, states are even less likely to exceed assurance levels in the 2020-2030
timeframe than the findings presented in this TSD's sensitivity analysis based on lower
variability limits originally modeled.

6

Deliberative - Do not cite or quote


-------
Table 2. Annual S02 Group 1 State Assurance Levels and Emissions in 2012 and 2014
(TRPenaltyScenario)



2012

2014



State Assurance

Penalty Case

State Assurance

Penalty Case



Level

Emissions

Level

Emissions



(thousand tons)

(thousand tons)

(thousand tons)

(thousand tons)

Illinois

258

210

137

128

Indiana

314

241

177

177

Iowa

118

75

83

78

Kentucky

208

146

117

117

Maryland

33

27

31

30

Michigan

214

190

158

158

Missouri

228

182

183

177

New Jersey

7

6

7

7

New York

23

20

13

13

North









Carolina

151

117

63

63

Ohio

341

229

151

151

Pennsylvania

307

250

123

123

Tennessee

163

97

65

65

Virginia

78

67

39

39

West









Virginia

161

119

83

83

Wisconsin

87

77

44

44

Table 3. Annual SO2 Group 2 State Assurance Levels and Emissions in 2012 and 2014
(TR Penalty Scenario)		



2012

2014

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

Alabama

238

219

235

173

Georgia

174

159

105

93

Kansas

46

41

46

46

Minnesota

46

43

46

45

Nebraska

72

65

72

70

South
Carolina

97

85

97

97

Texas

268

244

268

266

7

Deliberative - Do not cite or quote


-------
Table 4. Annual NOx State Assurance Levels and Emissions in 2012 and 2014
(TRPenaltyScenario)



2012

2014

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

Alabama

80

74

79

68

Georgia

68

61

45

40

Illinois

53

48

53

49

Indiana

121

110

119

110

Iowa

42

37

41

38

Kansas

34

31

28

24

Kentucky

94

84

85

76

Maryland

18

16

18

17

Michigan

66

59

64

57

Minnesota

33

31

33

31

Missouri

58

52

54

49

Nebraska

29

26

29

27

New Jersey

8

7

8

8

New York

20

18

20

17

North
Carolina

56

48

46

42

Ohio

102

85

96

84

Pennsylvania

132

118

131

117

South
Carolina

36

33

36

36

Tennessee

39

33

21

20

Texas

147

133

147

137

Virginia

37

33

37

35

West
Virginia

65

56

60

53

Wisconsin

35

31

33

30

Table 5. Ozone Season NOx State Assurance Levels and Emissions in 2012 and 2014
(TRPenaltyScenario)*



2012

2014

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

State Assurance

Level
(thousand tons)

Penalty Case
Emissions
(thousand tons)

Alabama

35

32

35

30

Arkansas

18

15

18

17

Florida

31

28

31

29

8

Deliberative - Do not cite or quote


-------
Georgia

32

27

21

18

Illinois

23

21

23

21

Indiana

52

47

51

47

Iowa

18

16

18

16

Kansas

15

13

12

10

Kentucky

40

35

36

32

Louisiana

16

14

16

14

Maryland

8

7

8

7

Michigan

28

25

27

24

Mississippi

12

11

12

11

Missouri

25

22

23

21

New Jersey

4

3

4

4

New York

10

8

10

8

North
Carolina

24

21

20

18

Ohio

44

35

42

36

Oklahoma

24

21

24

21

Pennsylvania

57

50

57

50

South
Carolina

15

14

15

15

Tennessee

17

14

9

8

Texas

69

63

69

64

Virginia

17

14

17

15

West
Virginia

28

23

26

22

Wisconsin

16

13

15

13

*As discussed in section III of the Transport Rule preamble, the final rule does not include the states of Iowa,
Kansas, Michigan, Missouri, Oklahoma, or Wisconsin in the ozone season program. EPA issued a supplemental
proposal to include these six states in the Transport Rule ozone season program.

9

Deliberative - Do not cite or quote


-------