United States Air and Radiation EPA420-R-02-020
Environmental Protection August 2002
Agency
vxEPA Non-Conformance
Penalties for 2004 and
Later Model Year Emission
Standards for Heavy-Duty
Diesel Engines and
Heavy-Duty Diesel Vehicles:
Response to Comments
> Printed on Recycled Paper
-------
EPA420-R-02-020
August 2002
for
for
to
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
-------
RESPONSE TO COMMENTS DOCUMENT
Non-Conformance Penalties for 2004 and Later Model Year
Emission Standards for Heavy-Duty Diesel Engines and Heavy-Duty Diesel Vehicles
Table of Contents
INTRODUCTION 3
SECTION 1: GENERAL NCP ISSUES 4
Issue 1.1: Shape of Penalty Curves 4
Issue 1.2: Comparison to Rulemaking Costs 5
Issue 1.3: NCP Implementation Timing 6
Issue 1.4: Evaluation of NCPs After Implementation 7
Issue 1.5: "Sunsetting" of NCPs 7
Issue 1.6: Additional Penalty Factors to Consider 8
Issue 1.7: Interaction of NCPs with Consent Decrees 9
SECTION 2: ISSUES REGARDING THE PROPOSED ANALYSIS 11
Issue 2.1: General Support for Analysis 11
Issue 2.2: Warranty, Maintenance, and Repair 11
Issue 2.3: Fixed Costs 12
Issue 2.4: Other Costs 13
Issue 2.5: Statistical Methodology 13
Issue 2.6: Determination of Upper Limit 14
Issue 2.7: Sources of Data 20
Issue 2.8: Discount Rate 21
Issue 2.9: Estimated Market Share 23
SECTION 3: ADDITIONAL RELEVANT ISSUES 24
Issue 3.1: Canadian Sales 24
Issue 3.2: Alternative Use of Penalty Funds 25
Issue 3.3: Issues Regarding Collection of Data Prior to and After Proposal 25
-------
List of Commenters and Abbreviations
California Air Resources Board
Caterpillar Inc.
Cedar Valley Freightliner
Cummins Inc.
Detroit Diesel Corporation
Engine Manufacturers Association
National Automobile Dealers Association
New York State Department of Environmental
Conservation
Pony Pack Inc.
Puget Sound Clean Air Agency
San Joaquin Valley Air Pollution Control District
Seneca Economics and Environment, LLC
Volvo Powertrain Corporation
CARB
CAT
CVF
Cummins
DDC
EM A
NADA
NYDEP
PPI
PSCAA
SJVAPCD
SEE
VPC
-------
INTRODUCTION
The EPA proposed this regulation on January 16, 2002, at 67 FR 2159. The proposal
announced the opportunity for written public comment until March 18,2002. The proposal also
provided notice of a public hearing which was held on February 15, 2002 in the Washington, DC
area.
Complete transcripts of the public hearing and the full text of each comment letter may be
found in Docket No. A-2001-25. All data and information relied upon by EPA in developing the
regulation also may be found in Docket No. A-2001-25. This docket is available for public
inspection and copying between 8:00 a.m. and 5:30 p.m., Monday through Friday, excluding
government holidays, at Room M-1500, Waterside Mall, 401 M Street, S.W., Washington, D.C.
This document summarizes the written and oral comments submitted at the public hearing
on February 15, 2002, as well as any comments received during the public comment period, and
records EPA's responses to those comments. The reader should note that many of the most
significant comments are also addressed in the preamble for the final rule and the responses in this
document cross-reference the corresponding discussion in the preamble where appropriate. The
responses presented in this document are intended (1) to augment the responses to comments that
appear in the preamble to the final rule, or (2) to address comments not discussed in the preamble
to the final rule.
EPA received approximately 20 written comments. A copy of each comment letter received
is included in the rulemaking docket. A list of commenters and the EPA docket item number
assigned to their correspondence is also included in the docket. All of the comments have been
carefully considered, and where determined to be appropriate, changes have been made in the final
regulation.
-------
SECTION 1: GENERAL NCR ISSUES
Issue 1.1: Shape of Penalty Curves
COMMENT: CARB commented that EPA should change the shape of the NCR curve to increase
the penalties for engines emitting near the applicable standard. Cummins did not specifically
address the shape of the curve, but commented on the use of MC50 and MC90, which both are used
to determine the shape of the curve. Specifically, they commented that MC50 should be set equal to
COC50 and that MC90 should be set equal to COC90. They argued that "the use of the single cost
component to derive MC50 is troublesome and results in inaccuracies in the NCR calculation." (The
values that they recommended for MC50 were higher than the values proposed, and would have
made the NCR penalties higher for compliance levels near the standard.) Cummins also
commented that, "by constraining the F factor to 1.1 to 1.3, the Agency is artificially lowering the
NCPs."
RESPONSE: We believe that we should generally follow the existing regulations in this rulemaking
regarding the shape of the NCR curves, unless we have reason to believe that the resulting NCPs
would not appropriately implement the requirements of the CAA. For example, we would not follow
the regulatory penalty formula if we determined that the resulting penalties would not remove the
competitive disadvantage to the complying manufacturers. However, neither CARB nor Cummins
provided a basis to believe that the proposed shape of the curves would not properly implement the
requirements of the CAA.
MC50 is intended to represent the marginal cost of fully complying with the standards for the
last increment of emission reduction for the average engine. Thus, it is appropriate to use a single
cost component to calculate MC50. This last increment should involve generally the most expensive
controls, especially the most expensive calibration adjustment. However, in our analysis of
minimum values for MC50, we effectively considered all cost components. We compared these
calculated values to the minimums to ensure that our estimates were no lower than the average.
Cummins mistakenly commented that EPA has set MC50 equal to COC50 in past
rulemakings. MC50 was actually set equal to COC50 divided by the difference between the UL and
the applicable standard. In some cases, the difference between the UL and the applicable standard
was 1.0 g/bhp-hr so that MC50 coincidentally had the same numeric value (with different units). If we
had used that approach in this rule, we would have estimated MC50 for LHD, MHD, and Urban Bus
as COC50 divided by 2.0 g/bhp-hr (4.5 g/bhp-hr - 2.5 g/bhp-hr). Similarly, we would have estimated
MC50 for HHD as COC50 divided by 3.5 g/bhp-hr (6.0 g/bhp-hr - 2.5 g/bhp-hr). The comparison of
this approach with the approach that we are using is described in the Technical Support Document.
Cummins did not provide a rationale why the marginal cost of compliance (MC50) should be
identical to a total cost of compliance (COC50), without regard to the magnitude of the emission
reduction associated with a total cost. We believe that such an approach would be inappropriate.
We also believe that the approach used in previous rulemakings is a simple approximation, and
when information is available to more precisely determine the marginal cost of the last step in
compliance, it is appropriate to use such information. It is also worth pointing out that if we used the
approach from previous NCR rulemakings, the MC50 values would be significantly lower than the
values being finalized and the values recommended by Cummins. This would have the opposite
effect from the values recommended by Cummins.
-------
With respect to the F factor, the 1985 rulemaking that established the first NCR regulations
determined that the use of a constrained F factor to determine the value of X would generally result
in properly shaped curves, and that the resulting NCPs would meet the requirements of the CAA.1
Cummins commented that the constraint on the F factor would result in NCPs that are too low, but
provided no information to show that the resulting NCR curves would not meet the requirements of
the CAA. We continue to believe that the final NCR curves developed using an F value of 1.3 are
sufficiently steep to remove the competitive disadvantage for the complying manufacturers and
appropriately implement the requirements of the CAA.
Issue 1.2: Comparison to Rulemaking Costs
COMMENT: Caterpillar commented that we should set the NCR using the same cost estimates
used in the standard-setting rulemaking. In their comments they compared those rulemaking costs
to similar cost categories described in the draft Technical Support Document for this rulemaking.
They stated that we did not justify our "departure from the cost estimates used in the 2004 Final
Rule."
Cummins commented that the difference in the cost numbers "estimated in 1998 for the
2004 rule emissions standards" from the cost numbers estimated in 2002 for the NCR "reflects the
engine manufacturers (and EPA's) better understanding of the actual cost of the technology based
on the actual cost to develop and to capitalize their plants to manufacture the technology."
RESPONSE: The CAA directs EPA to set NCPs that remove the competitive disadvantage to the
complying manufacturer. We believe that this is best implemented by using the latest and most
accurate information that is available. As we have done in previous NCP rulemakings, we collected
new cost information to not only determine average costs, but also 90th percentile costs. Also,
rulemaking costs generally reflect longer term costs, rather than the manufacturing and operating
costs associated with engine produced during the first model year of the standards. For these
reasons, we are calculating the final NCPs in the same manner as was proposed.
We disagree with Caterpillar that we did not justify why the analysis in this NCP rulemaking
differs from the analyses for the model year 2004 standard-setting rulemakings. These differences
were discussed in detail in Chapter 3 of the Draft Technical Support Document for the proposal for
this rulemaking. As described in that document, there are three main reasons for the differences.
First, the goal of this analysis is to estimate manufacturer and operator costs during the first year of
the new standards rather than to project the long-term costs. While we did make some attempt to
project cost decreases with time, the only category in which we separately estimated short-term and
long-term costs in the FRM was hardware cost category. The 2000 FRM analysis did not include an
analysis of short-term fuel, warranty, or repair costs. Thus, Caterpillar's comparison of the various
estimates of costs in the different categories is not valid. Second, the baselines for calculation of
compliance costs differ significantly due to issues associated with the Consent Decrees. More
specifically, the assumed baseline engine used in the previous analyses is very different than the
upper limit engine that served as the baseline for heavy-heavy engines cost analysis. Finally, we
now have more detailed information about costs identified in the earlier analysis, as well as cost
1 The product of MC50 times F is the slope of the NCP curve for compliance values less than X.
Larger values of F would result in steeper slopes and higher penalties for compliance levels near the
standard.
-------
categories not previously included. During the previous rulemakings, we requested cost comments
from manufacturers, but received very little input regarding our cost estimates. However, we
received a great deal of input from manufacturers for this rulemaking.
Issue 1.3: NCP Implementation Timing
COMMENT: EMA and Caterpillar commented that EPA must set NCPs in advance to allow
manufacturers adequate lead time. EMA stated that EPA should establish NCPs in the rulemaking
that establishes the emission standards. Caterpillar argued that NCPs "are not useful if they are not
set with sufficient lead time to allow manufacturers to make rational marketing decisions."
Cummins commented that the timing of this rulemaking relative to the implementation date
of the new standards is no later than past NCP rulemakings, and earlier than most of them. They
also commented that even considering the timing relative to the consent decrees, this rulemaking
was consistent with past NCP rulemakings. They stated that "as long as the NCP rule is finished
before October 2002, NCP manufacturers will have adequate time to price their products."
RESPONSE: Historically, NCPs have never been established in the rulemaking that establishes the
emission standards. Indeed, this is essentially impossible to do, given that such rulemakings often
occur many years prior to the effective date of the new emission standards, at a point in time in
which it is typically difficult, or even impossible, to clearly judge whether the necessary criteria are in
place in order to establish NCPs. For example, in the 1999 review, we assessed the NCP eligibility
criteria due to comments raised during the rulemaking process. We found at the time that two of the
three necessary criteria could be met, but that a judgement regarding the third - the determination of
the likelihood of a technological laggard - could not be made at a date so distant from the
implementation of the standards (64 FR 58472, 58513, October 29, 1999).
These NCPs are being set with lead time relative to the implementation of the new
standards that is comparable to or longer than lead time associated with past NCP rules (see table
below). The NCPs in this rule are being established almost 18 months prior to the effective date of
the new standards, a period of lead time that we believe is adequate and consistent with historical
NCP rulemakings. Cummins' comments regarding appropriate lead time relative to the Consent
Decrees is not relevant, as this rulemaking addresses the 2004 model year regulatory requirements.
Timing of Past NCP Rulemakings
Model Years for Which NCPs Were Set
1987 and 1988
1991
1994
1996 and 1998
Date NCPs Were Established
December 31, 1985
Novembers, 1990
December 28, 1993
February 23, 1996
Federal Register Citation
50 FR 53466
55 FR 46622
58 FR 68540
61 FR 6944
-------
Issue 1.4: Evaluation of NCPs After Implementation
COMMENT: The NYDEC commented that the NCR upper limit should be continuously reevaluated
and reduced as appropriate over time.
RESPONSE: As is discussed in Issue 1.5 (below), the nature of the annual adjustment factor in the
NCR formula greatly diminishes the practical application of NCPs for many years after initial
applicability. Of course, one way to reduce the size of the NCR in the "out years" would be to move
the compliance level closer to the standard. This would have same effect as reducing the upper
limit. Thus whether the use of NCPs is partially inhibited by the increase in the NCR penalty rate
overtime, or those using NCPs tighten up their compliance levels to reduce the penalty per engine,
the effect being sought by the commenter (i.e., lower emitting engines in use) will be achieved.
Thus, EPA expects there will be no need to reduce the UL in a future rulemaking.
Issue 1.5: "Sunsetting" of NCPs
COMMENT: The SJVAPCD commented that, although they support the use of an annual
adjustment factor, they recommend making the NCPs available for only a limited amount of time in
order to provide an added incentive for manufacturers to produce complying engines in a timely
manner. They do not suggest what an appropriate time period would be, however.
RESPONSE: This issue is broader than the scope of this NPRM, but it was in fact considered in the
NPRM and FRM which established the generic NCP rule (see 50 FR 9204 and 50 FR 35374 and
supporting material in public docket EN -85-02.) . The NCP equation which was developed in that
rulemaking incorporates an annual adjustment using two factors. The first is a straightforward
adjustment for inflation effects on the dollar value of the NCP relative to the initial year. The second
factor in the NCP formula links the annual rate of NCP increase to the rate of NCP use, and
increases the real inflation-adjusted penalty with time to the extent that the penalties are used. The
equation for calculating time and usage effects is shown below and is described in detail in the
regulations at 40 CFR 86-1 1 1 3-87.
where: AAF = annual adjustment factor
i = year under consideration since NCP availability
I = percentage increase in overall CPI for previous year relative year NCPs first
offered for this rule
A = usage adjustment factor A = 0.10 for first year of adjustment and 0.08 thereafter
frac = fraction of engines using NCPs in this subclass for previous year
Referring to the NCP equation, these effects are cumulative since the NCP for any given year n is
calculated using the running product of the AAF term (see 86.1 113-87). These equations taken
together show that:
1) No adjustments occur in the first year.
2) If NCPs are not used in the second or subsequent years the value of the NCP increases
with the rate of inflation.
3) If NCPs are used the NCP for the subsequent year increases by a factor dependent on
the fraction of engines in a given subclass using NCPs ( see fracM above).
-------
4) Since NCPs should be most valuable when new standards are first being met, and less
important in later years after the NCR is implemented, the rate of this adjustment increases with
each year (i). This affect is moderated somewhat by the factor( A,) which is set at 0.10 for the
second year and 0.08 for subsequent years.
For example, if there were zero inflation and NCPs were used on 10 percent of the engines
sold for 5 years, the penalty rate would increase as shown below:
Year of NCR Use
1
2
3
4
5
Penalty Rate Increase Over First Year
n/a
12.3%
24.6%
39.82%
58.7%
EPA believes that the construction of NCP penalty formula provides adequate means that
will practically limit their use for a large number of years after initial applicability.
Issue 1.6: Additional Penalty Factors to Consider
COMMENT: Cummins commented that EPA should add "an additional amount in the range of
several thousand dollars" to the cost of the NCP to address buyer perception issues. DDC
commented that they share the concern that the cost-based NCP will not fully equalize the
purchasing decision between compliant and non-compliant engines, and that EPA should ensure
that all of the factors are considered and quantified. DDC also commented that EPA include in the
penalty a cost associated with the environmental disbenefit.
Cummins argued that there is an "exaggerated fear of buying unproven technology" and that
"simply setting NCPs equal to the actual costs to develop, manufacture and operate compliant
technology may not provide adequate disincentives to purchase existing (higher emitting)
technology engines." They suggested that one way of measuring the additional premium that
should be added to the straight cost calculation of NCP would be to look at the margin lost to the
lost-sale manufacturer, since "the choice to purchase existing (higher emitting) engines represents a
lost sale to the engine manufacturer that was prepared to sell an environmentally-more-beneficial
engine."
SJVAPCD strongly supported the view that the NCP should go beyond removing the
competitive disadvantage from a complying manufacturer. They suggest that the cost of an NCP
should "significantly exceed the competitive benefit gained by building a non-conforming engine."
RESPONSE: EPA sought comment on this issue and included a lengthy discussion on the matter in
the NPRM. However, since we believed that the proposed NCPs would be sufficient to remove the
competitive disadvantage for the complying manufacturers, no specific proposal was included for an
additional factor. The comments that we received that supported the inclusion of an additional
factor were conclusory and did not provide any information that would lead us to believe that
-------
finalizing NCPs based only on the estimated costs of compliance would be insufficient to remove the
competitive disadvantage for the complying manufacturer. Therefore, since the commenters
provided factual basis to believe that effects of purchaser perception need to be accounted for
separately, we still believe that no additional factor is necessary.
We also believe that developing an additional factor to address potential purchaser
reluctance would be problematic. Incorporating Cummins idea of including profits from lost sales is
one consideration, but since it does not directly reflect a cost of compliance, it would be difficult to
implement. Evaluating this option would require price elasticity, price, and cost information not
available to EPA at this time and even so still may not capture the relatively intangible nature of
consumer behavior in this situation where decision may not be based solely on price. DDC's idea of
incorporating a cost element related to the environmental disbenefit of higher emitting engines
reflects a cost to society but not necessarily a cost to the purchaser.
With respect to the SJVAPCD comment, we believe that the existing NCP formulas that use
"worst-case" costs as well as average costs will frequently result in there being some margin
between the penalty and the economic gain that might otherwise be obtained from selling a non-
conforming engine. That is why engine manufacturers have generally not used NCPs extensively in
the past.
The points raised by the commenters present some interesting issues for future
consideration. However, we believe the NCPs being finalized will appropriately implement the
requirements of the CAA without adding an additional factor.
Issue 1.7: Interaction of NCPs with Consent Decrees
COMMENT: Cummins commented that the 1998 NCPs multiplied by a factor of 1.5 do not reflect
the cost of compliance to achieve the 2004 standards. They stated that it is imperative that the
Agency promulgate NCPs for 2004 that accurately reflect the cost of compliance to meet the 2004
standards.
Caterpillar commented that EPA's NPRM is not timely, given that EPA claim's this rule will
impact the NCPs for Consent Decree companies in October, 2002. Caterpillar stated that EPA's
proposal cannot apply to engines set to be introduced in seven months. Caterpillar commented that
the NCP final rule must clearly state that the rule does not apply to Consent Decree engines.
Caterpillar commented that it was improper to consider the Consent Decrees in this
rulemaking. They added that, assuming that it was proper to consider the Consent Decrees, then
EPA should consider the requirements that will take effect in October of 2002. They stated that
these upcoming requirements will require manufacturers comply with FTP and EURO III levels of
2.5 g/bhp-hr for NOx + NMHC.
RESPONSE: As was stated in the proposal, this rulemaking establishes NCP's for the 2004 model
year NMHC+NOx emission standards for heavy-duty diesel engines. We agree with Cummins to
the extent that they support NCPs that accurately reflect the actual costs of compliance with the
2004 standards. However, Cummins comments regarding the penalty provisions of the Consent
Decrees are not relevant to this rulemaking. The analyses associated with this rule were performed,
and decisions were made independently from the penalty provisions of the Consent Decrees.
Caterpillar's comment regarding the timing between this rulemaking and the Consent
-------
Decrees are not relevant to this rulemaking. This rule establishes NCPs under the regulations for
the model year 2004 NMHC+NOx standard. Caterpillar's comments that these NCPs cannot apply
to Consent Decree engines and that EPA should state this final rule does not apply to the Consent
Decree engines are not relevant to this rulemaking. The only issue before the Agency and being
decided in this rulemaking is the applicability of NCPs, under EPA regulations, for model year 2004
and later engines. Issues of Consent Decree interpretation or implementation are not before the
Agency and are not relevant to this rulemaking.
Caterpillar's comment that EPA should consider the requirements that will take effect in
October of 2002, instead of those already in effect is addressed in section 2.6.
10
-------
SECTION 2: ISSUES REGARDING THE PROPOSED ANALYSIS
Issue 2.1: General Support for Analysis
We received general comments from CARB, Cummins and DDC supporting our analysis (or
the rulemaking in general). We also received more specific comments on our analysis. These
more specific comments are discussed below.
Issue 2.2: Warranty, Maintenance, and Repair
COMMENT: DDC provided detailed recommendations to modify our analysis of warranty,
maintenance and repair costs. First, they commented that their historical information suggests that
the per-mile post-warranty repair rate should be 35 percent less than the warranty repair rate, rather
than 50 percent less, as we used in our draft analysis. They also commented that our post-warranty
repair cost is too low because it does not include the engine manufacturers' markup for replacement
parts. They argued that this markup will result in a cost (dollars per repair) to the operator for post-
warranty repairs that is significantly higher than the cost to the engine manufacturer for warranty
repairs.
With respect to demurrage costs, DDC agreed with our proposed approach of estimating the
costs based on rental prices. However, they suggested that we drop the cost of insurance, since
insurance for rental vehicles would likely be included in the owners normal insurance policy. They
also commented that the total downtime for each repair would normally be 1.5 to 2 days when
logistical factors such as vehicle pickup and delivery are included. Finally, they commented that
EPA could alternatively estimate the demurrage cost as two days of lost revenue plus meals and
lodging for the driver.
With respect to scheduled maintenance, DDC commented that a typical EGR engine should
experience three more oil changes than a current engine, and will require EGR maintenance at
rebuild.
In addition to these public comments on maintenance issues, we also received updated
information from several manufacturers relating to maintenance. This information was also
considered in our final analysis.
RESPONSE: Based on the public information provided by DDC and confidential information
provided by other manufacturers, we continue to believe that the best estimate for the per-mile post-
warranty repair rate is 50 percent less than the warranty repair rate. While DDC projects that per-
mile repair rates will only drop by 35 percent after the warranty periods, at least one other
manufacturer projects that per-mile repair rates will only drop more than 50 percent. Therefore, we
are basing our final estimates of post-warranty repair rates on a 50-percent decrease. However, we
do believe that our estimated cost per repair in the draft analysis for heavy-heavy duty engines
($700 per repair) was slightly high. Based on updated information from manufacturers, we now
believe that the repair cost for heavy-heavy duty engines is more likely to be $500 per repair. As
described in our analysis, using a lower cost per repair will have the effect of increasing the number
of repairs. We agree with DDC that we should markup the cost of parts for post-warranty repairs.
We agree with DDC's comments that we should drop the insurance costs for our estimate of
11
-------
demurrage costs. However, we do not agree that we should increase the time allotted for each
repair. We still believe that most repairs will be completed within one day. While a few repairs may
take more than one day, many others may be completed in less than one day, or even during other
scheduled maintenance. Thus, our estimated demurrage costs are equal to the approximate cost of
renting a vehicle for a single day. We continue to believe that a rental based cost estimate is more
reliable than a lost revenue estimate, which would be highly dependent on the type of freight being
hauled.
We have included all of the updated information regarding scheduled maintenance into our
final costs estimates.
Issue 2.3: Fixed Costs
COMMENT: DDC provided detailed recommendations to modify our analysis of fixed costs. They
believe that our proposed estimates of fixed costs were amortized over too many engines. First,
they stated that manufacturers will attempt to recover the fixed costs associated with the 2004
standards before the new 2007 standards begin to be phased-in. They stated that "it is not correct
to assume that the planned EGR system for 2004 will in fact be useable in 2007 and beyond." They
argued that manufacturers will take advantage of the averaging provisions for model years 2007-
2009 to produce a single technology product rather than produce both very low NOx engines and
the 2004-era engines. They suggested that we use a three-year amortization period instead of the
proposed five-year period.
DDC also commented that we should base our cost analysis on modeled market volume
rather than current sales volumes. They argued that the compliance costs associated with the 2004
standards will affect the market and must be accounted for.
RESPONSE: We continue to believe that a five-year amortization period is appropriate for this
analysis. In our recent final rule setting new emission standards for 2007 and later heavy duty
diesel engines, we projected that manufacturers would rely on their 2004 technologies to help
achieve the more stringent 2007 standards. While it is theoretically possible, as DDC suggested,
that a manufacturer could comply on average with the 2007 phase-in requirements using a fully
optimized NOx adsorber on an engine without EGR, we do not believe that is likely to be feasible
when the standards are fully phased-in. Thus, we fully expect that manufacturers will be using their
2004 technologies in 2010. Given this, we believe that a five-year amortization period is appropriate
for this analysis.
With respect to the annual sales volumes, we believe that the 2000 sales volumes represent
a good estimate of what sales will be in 2004. There are reasons to believe that sales in 2004 may
be higher, but there are also reasons to believe that sales in 2004 may be lower. As shown in figure
2.3-1, since 1990, there has been a long-term trend of increasing annual sales volumes. However,
much of this increase was the result of the high economic growth rates observed in the 1990s. At
this point, we cannot necessarily project that growth rates between now and 2004 will be that high.
Moreover, some manufacturers have expressed concerns that purchaser perceptions may slow
sales as the new technology is introduced into the market.
12
-------
Figure 2.3-1
Total Sales
Thousands C
S./(
350
300
250
200
150
100
50
1£
Canada Factory Sales of HH Trucks
/\
X\ ^{ ^
/ a
\^
88 1990 1992 1994 1996 1998 2000 2002
Issue 2.4: Other Costs
COMMENT: Several manufacturers provided updated cost data in their CBI comments. DDC also
commented in support of the use of a multi-year fuel cost for heavy-heavy duty engines, given their
expected lifetimes.
RESPONSE: We incorporated the updated manufacturer information into our analysis. In general,
this updated information did not substantially change our analysis. The specific cases where we
received updated information are identified in Chapters of the Technical Support Document.
We agree with DDC that a multi-year average fuel price is appropriate for this analysis.
Therefore, we are using a five-year average fuel price in the final analysis. We believe that a
five-year average best approximates how purchasers will make purchase decisions, considering the
economic significance of changes in fuel consumption rates. This issue is discussed in more detail
in Chapter 3 of the Technical Support Document.
Issue 2.5: Statistical Methodology
COMMENT: Caterpillar commented that we incorrectly combined average costs for each category.
They implied that our method only worked if the costs occur in "fixed proportions", and argued that
we should have summed total costs for each manufacturer and then sales-weighted these total
costs, rather than summing the sales-weighted cost in each cost category. Cummins supported our
use of the sales-weighted mean of the costs forCOC50 instead of using median costs.
RESPONSE: It is not possible to estimate average costs in the manner suggested by Caterpillar
because not all manufacturers provided cost data for each subcategory. Some of these
manufacturers acknowledged that, while they were not able to estimate the cost for a category such
as oil changes, they did believe that there would be some additional cost. More importantly,
Caterpillar is incorrect in claiming that the two methods would produce different results. It is
mathematically equivalent to sales-weight total costs for engine and to sum sales-weighted
averages of subcategories of costs. This is true without regard to whether the costs occur in "fixed
13
-------
proportions." An example of this is shown in the following table. Both methods result in an
estimated average cost of $982.53.
Manufacturer 1
Manufacturer 2
Manufacturer 3
Sales
1000
2000
1500
Sales
Weighted
Avg.
Cost Category 1
$100
$150
$200
$136.50
Cost Category 2
$400
$300
$300
$331.75
Cost Category 3
$500
$500
$800
$514.28
Total
$1000
$950
$1300
$982.53
Issue 2.6: Determination of Upper Limit
COMMENT: DDC and Cummins supported our proposed upper limits. Caterpillar commented that
EPA should not base the NCPs for heavy-heavy duty engines on an upper limit of 6.0 g/bhp-hr.
NYDEC also opposed the 6.0 g/bhp-hr upper limit for heavy-heavy duty engines, because of their
concerns about "excess emissions in the field". They supported setting the upper limit at 5.3 g/bhp-
hr. We did not receive comments opposing the upper limits for the other service classes.
Caterpillar stated that it was improper for EPA to consider the Consent Decrees in setting
the upper limits. Caterpillar stated that EPA cited no technical analysis which justifies the use of a
6.0 g/bhp-hr NMHC+NOx Upper Limit. They also argued that the UL should be based on the
capabilities of the technological laggards in 2004. They argued that this approach effectively counts
the combined costs of complying with the 1998 NOx standard and the 2004 NMHC+NOx standard.
Caterpillar states there is "no statutory or regulatory justification for an approach that utilizes the
costs of bringing an engine that does not comply with the present emission standards into
compliance with the next emission standards." Caterpillar also argues an Upper Limit of 6.0 g/bhp-
hr is not appropriate because it counts the costs of "eliminating purported defeat devices even
though the defeat device prohibition is not new." They further argued that the compliance costs of
the Consent Decrees were "intended as punishment" and should not be considered to be costs of
compliance with the 2004 emission standards.
Caterpillar stated that basing the upper limit on Consent Decree engines would put non-
Consent Decree companies and new entrants to the market place at a disadvantage. They also
argued that it will create a disincentive to investment in emission reduction because it would allow"
a manufacturer to stay in the market where that manufacturer's engines would not have complied
with the previous emission standard or UL (not just the current emission standard and UL)."
Caterpillar stated that even if the Consent Decree emission levels were relevant to this rulemaking,
EPA selected the wrong levels. Caterpillar stated that the Consent Decrees specify the 2004 FTP
standards must be met by October 1, 2002, as well as a Euro-3 emission limit equal to the 2004
FTP standards, all of which must be met without the use of defeat devices.
Caterpillar stated that EPA attempted to justify an UL of 6.0 because the Agency did not
collect cost data for HHDDE's other than from a 6.0 g/bhp-hr NMHC+NOx baseline, but Caterpillar
suggests that a "flaw" in EPA's approach "is not a basis for arbitrarily choosing a 6.0 g/bhp-hr UL".
14
-------
Finally, Caterpillar stated that EPA improperly used non-existent off-cycle emissions requirements to
set a UL. Caterpillar states that the NPRM implicitly concludes that a compliant engine would
comply with Euro-Ill emission level, but such emission requirements do not apply until 2007, and
therefore EPA has included costs of compliance for 2004 which are really costs of compliance for
2007.
RESPONSE: We have decided to promulgate, as proposed, Upper Limit values of 4.5/bhp-hr
NMHC+NOx for light-heavy, medium-heavy, and urban bus engines, and an Upper Limit value of
6.0 g/bhp-hr NMHC+NOx for heavy-heavy duty diesel engines. These upper limit values will
generally allow technologically lagging manufacturers to continue producing their current products
until they can complete the development of their new engine models.
We understand the concerns raised by the NYDEC regarding the potential for in-use
emissions that are above the standard. However, NCPs by their very nature allow production of
engines which do not meet the otherwise applicable emission standards. We do not agree with
NYDEC that the emissions allowed by this UL would be "excess emissions" in the sense of
emissions attributable to defeat devices. This action does not weaken in any way the existing
prohibition of defeat devices. A manufacturer whose engine has high emissions on, for example,
the EURO III would need to certify to a comparable FTP level. Thus, the manufacturer would be
required to pay an appropriate penalty that is consistent with the engines actual performance. The
issue is whether the emissions that would be allowed to occur by the establishment of an UL for
heavy-heavy duty engine of 6.0 g/bhp-hr is allowed and appropriate under the CAA. We believe
that the CAA allows us to set an upper limit that allows continued production of current engines, and
thus allows us to set the UL at 6.0 g/bhp-hr. Moreover, we believe that the penalties that would
apply for compliance levels between 4.5 g/bhp-hr and 6.0 g/bhp-hr are large enough to provide a
strong disincentive to produce engines that have emissions any higher than technologically
necessary.
We disagree with Caterpillar's general comments that it was improper for EPA to consider
the Consent Decrees in setting the upper limit. As discussed in the preamble to the proposal and
for this final rule, the selection of an UL of 6.0 g/bhp-hr reasonably should allow existing HHDDE's in
the marketplace which are manufactured by technological laggards to continue to be sold while the
manufacturer works on the technologies necessary to bring such engines into compliance with the
model year 2004 standards, without also requiring a concurrent design change to produce engines
to a UL below the existing levels. The consent decrees have created a unique situation under which
engines which would not otherwise be certified by EPA are currently allowed to be manufactured
and sold. The vast majority of new HHDDEs do not comply with the existing emission requirements,
and emit at NOx levels well in excess of the current emission standard. The selection of an UL of
6.0g/bhp-hr reflects that reality, in order to meet the goal described above. Caterpillar's comments
suggest that EPA should ignore the existence of the consent decrees in considering the
establishment of the Upper Limit. We disagree with this comment. Ignoring the Consent Decree
would likely lead to setting an UL of 4.5 g/bhp-hr as with the other service classes. This would
interfere with meeting the goal noted above, and be inconsistent with the other reasons noted below.
Caterpillar states that EPA cited no technical data nor any statements from engine
manufacturers that suggests a technical laggard could not comply with an UL of 4.5g/bhp-hr
NMHC+NOx for HHDDEs. We disagree with the assumption inherent in this statement.
Caterpillar's statement implies that EPA must perform a technical analysis which demonstrates that
the level below the emission performance of existing engines could be achieved by technological
laggards. Neither the CAA nor EPA's regulations require that EPA demonstrate that the Upper Limit
15
-------
value is the lowest emission level achievable by the technological laggards in the given time frame.
CAA Section 206(g)(2) requires that no certificates of conformity will be granted " if the degree by
which the manufacturer fails to meet any standard promulgated under section 202(a) with respect to
such class or category exceeds the percentage determined under regulations promulgated by the
Administrator to be practicable." The statute requires that the Upper Limit value be the value
"determined... to be practicable". In past rulemakings EPA has interpreted its obligation to
determine a "practicable" emission level to not require that the most stringent level achievable by
technological laggards. For example, our regulations call for the UL to be set at the prior emission
standard (where one is applicable), and do not call for EPA to determine whether the prior emission
standard is the lowest level achievable by the technological laggards. That regulatory provision was
based on setting the UL at the level of existing control, to ensure all heavy duty engines could meet
it, to avoid the need for a separate design phase, and because of concerns over the ability to readily
determine a level below the prior standard that would be practicable for all manufacturers. See 50
FR 35376-35377 (August 30, 1985) and 67 FR 2164-2165 (January 16, 2002).
With respect to the likelihood of technological laggards developing, we have placed into the
docket for this rulemaking information from two heavy-heavy duty diesel engine manufacturers
which leads EPA to conclude there is likely to be at least one technological laggard in model year
2004.2 As specified in EPA's existing regulations, one of the criteria for establishing an NCP is EPA
must find "that there is likely to be a technological laggard" (see 40 CFR 86.1103-87(a)(2)).
In addition, EPA did provide data at the time of the proposal which indicates an Upper Limit
of 6.0 g/bhp-hr NMHC+NOX would likely be achievable by all manufacturers in model year 2004.
We included a memorandum to the docket for this rule which summarizes the certification data from
model year 2001 heavy-duty diesel engines, and this memorandum was discussed in the proposal
(See 67 FR 2165). This data shows that for the heavy-heavy service class, all engine families for
Model Year 2001 had Euro-3 emission performance at or below 6.0 g/bhp-hr NMHC+NOx.3 This
data shows there are a number of engine families with emission levels greater than 4.5 g/bhp-hr
NMHC+NOx. If EPA were to establish an Upper Limit of 4.5 g/bhp-hr NMHC+NOx, or even 5.3
g/bhp-hr NMHC+NOx (a level for which we also requested comment), many of these engine families
would not meet this value and would require additional development and perhaps the application of
new technologies. It is reasonable to establish the UL so as to avoid requiring such a separate
design cycle.
Caterpillar suggested that EPA's selection of a 6.0g/bhp-hr UL effectively combines the
costs of complying with the 1998 NOx standard and the 2004 NMHC+NOx standard, and that this
implicitly combines the costs of removing alleged defeat devices. Caterpillar states there is "no
See April 5, 2002 press release from Detroit Diesel Corp. in which they state "As
previously announced, DDC will initially make approximately 95% of its current ratings available in
October 2002." EPA believes it reasonable to project this company may need NCPs for model year
2004 in order to offer a full product range. See also EPA Memorandum "Summary of Meeting
between US EPA and Cummins Inc. on April 17, 2002 regarding 2004 On-highway Heavy-duty
Diesel Engine Nonconformance Penalty Rule", in which this company suggests some part of their
product line of heavy-heavy engines may not be ready by model year 2004. A copy of both of these
documents is available in the docket for this rule.
3 EPA Memorandum "Summary of Model Year 2001 Heavy-duty Diesel Engine HC and NOx
Certification Data", copy available in the docket for this rulemaking.
16
-------
statutory or regulatory justification for an approach that utilizes the costs of bringing an engine that
does not comply with the present emission standards into compliance with the next emission
standards." They also argued that our costs included the compliance costs of the Consent
Decrees. We disagree with Caterpillar's comments. We did not combine the costs of separately
complying with the 1998 and 2004 standards, nor did we calculate Consent Decree costs. Rather,
we have estimated the cost of compliance for model year 2004 relative to the UL of 6.0 g/bhp-hr.
We recognize that our estimate of these compliance costs will be similar to the costs of separately
complying with the 1998 and 2004 standards. However, the purpose of our calculation is separate
and distinct - to assess as accurately as possible the cost of compliance for model year 2004
compared to an engine at the UL. EPA believes setting an UL at 6.0 g/bhp-hr is most appropriate
for the reasons noted above and in the preamble. An UL of 6.0 g/bhp-hr best implements the
policies behind an UL. EPA believes it is more appropriate to address issues of enforcement of past
noncompliance with the defeat device prohibition through an enforcement mechanism. In selecting
an UL in this NCP rulemaking, we intend to do no more than pick a level that best implements the
policy behind setting an UL, as expressed in prior NCP rulemaking.
We disagree with Caterpillar's comments that what EPA has done in this rule is not justified
under the statute or EPA's regulations. To the contrary, CAA Section 206(g)(2) directs EPA to
establish an Upper Limit through a rulemaking which is "determined... to be practicable". The
statute provides a large degree of discretion to EPA in determining a practicable UL. Nothing in the
statute prohibit EPA from selecting the UL values promulgated in this final rule. As noted earlier,
we believe that the UL values being established are appropriate under this provision of the CAA.
Caterpillar also argues an Upper Limit of 6.0 g/bhp-hr is not appropriate because it counts
the costs of "eliminating purported defeat devices even though the defeat device prohibition is not
new." EPA agrees that the defeat device prohibition is not new. However, this does not negate the
reality that model year 2001 engines currently being produced by HHDDE manufacturers have high
off-cycle emissions, and these same engines are the engines which are the basis for the UL of
6.0g/bhp-hrforthe reasons noted above. The 2004 model year standards, including the defeat
device prohibition, are clearly a new more stringent standard. This is true whether compared to the
1998 standard with the defeat device prohibition, or to the Consent Decree limits. The existence of
such a new standard meets one of the regulatory criteria that must be met before NCPs are allowed.
However, this new standard does not dictate the level of the UL, which is chosen for other reasons
We disagree with Caterpillar's comments that a 6.0 g/bhp-hr UL would disadvantage non-
Consent Decree companies and new entrants to the market place. In fact, an Upper Limit set at 6.0
provides non-Consent Decree companies and new entrants to the market the option of producing
engines that emit up to 6.0 g/bhp-hr NOx rather than 4.5 g/bhp-hr, provided that they pay the
appropriate penalties. . Therefore, a non-Consent Decree company or new entrant to the market
who is a technological laggard and whose engines currently emits in the 4.5-6.0 range will benefit by
the selection of a 6.0 Upper Limit, since the UL at 6.0 would allow the company to continue to sell
today's product while they invest the resources to comply with the new standard. Implicit in
Caterpillar's comments is the assumption that the NCPs are higher because of the choice of 6.0
g/bhp-hr for the UL. However, for values below 3.4 g/bhp-hr (the X value for HHDE), the penalties
are defined by the marginal cost parameters F and MC50. Since F and MC50 are independent of the
UL, the penalties would be identical in this range for any UL value considered. As described below,
we did not estimate the cost of compliance for an UL of 4.5 g/bhp-hr. However, depending on what
techniques the manufacturers would otherwise have used to meet a 4.5 g/bhp-hr level, it is also not
clear that the penalties between 3.4 g/bhp-hr and 4.5 g/bhp-hr would have been significantly lower if
the UL had been set at 4.5 g/bhp-hr and, in fact, there are technical scenarios in which it could have
17
-------
been higher. However, Caterpillar has provided no data which suggests that non-Consent Decree
companies or new entrants intend to employ NCPs instead of complying with the standards. In fact,
we received no comments from non-Consent Decree companies or new entrants which would
indicate they are concerned that an UL of 6.0 puts them at a disadvantage. At least one such
company who competes in the heavy-heavy service class, DaimlerChrysler (manufacturer of
Mercedes-Benz), was directly notified by EPA of the publication of the proposal, and we received no
comments from this company.4 In addition, the statute requires the EPA establish NCPs which
"shall remove any competitive disadvantage to manufacturers whose engines or vehicles achieve
the required degree of emission reduction" (see CAA Section 206(g)(3)(E)). The statute requires we
remove any disadvantage which the complying manufacturer may encounter. While we do not
believe that NCPs place any non-Consent Decree company or new entrant at a competitive
disadvantage, the statute does not require that we establish an NCP which removes any
disadvantage to the non-complying manufacturer. As articulated in the preamble for this final rule
and in previous NCP rules (see for example 50 FR 35374), NCPs provide an alternative to a
manufacturer who is already in the market so that a manufacturer need not exit the market if they
are unable to comply with a new emission standard. New entrants to the market are by definition
not already in the market, and NCPs have not previously been promulgated to provide a mechanism
by which an engine or vehicle manufacturer could gain entry to the U.S. market by selling non-
complying products.
We disagree with Caterpillar's comment that a 6.0 UL for HHDDE's create's a disincentive to
invest in emission reductions. First, NCP's are specifically designed to encourage compliance with
the applicable emission standard, in particular for engines near the upper limit value. The per
engine penalty can be much greater than the average cost of compliance, and for engines at the UL,
the per engine penalty is intended to represent the 90th percentile costs. Nevertheless, even if the
UL of 6.0 g/bhp-hr resulted in lower penalties than a 4.5 g/bhp-hr value, there would still be
significant incentive to reduce emissions. Selling products in a competitive market where your costs
(including penalties) are at the high end of the range provides a strong incentive to apply
technologies to reduce the penalty being paid, and thus lower emissions.
We disagree with Caterpillar's comments that EPA selected the wrong emission level from
the Consent Decrees. Caterpillar's comments suggest that EPA should set the Upper Limit for the
heavy-heavy engines at the October 1, 2002 limits set in the Consent Decrees. However, the
October 1, 2002 levels from the Consent Decrees are the 2004 model year emission standards (2.5
g/bhp-hr NMHC+NOx). Setting the UL for the 2004 standards at 2.5 g/bhp-hr NMHC+NOx would be
inconsistent with the intent of §206(g) of the CAA. With the UL equal to the standard, there would
be no way for a manufacturer to employ the NCP. Such an approach would be inconsistent with the
policy behind setting an UL at a prior, higher level, to provide time for technological laggards to
come into compliance.
We disagree with Caterpillar's comments that EPA's selection of the UL was arbitrary
because of limitations in the data collected. While it is true that for the HHDDE's, all of the
confidential cost data we received included the cost of compliance for model year 2001 engines to
comply with the 2004 FTP standards, and all of the MY2001 engines were in fact operating at
emission levels on the order of 6.0 g/bhp-hr NOx+NMHC, that is not the primary reason why EPA
See E-mail message from William Charmley, US EPA to "Interested Parties", including Mr.
Jim Chen (icchen@hhlaw.com), attorney for Mercedes Benz. Copy available in the docket for this
rule.
18
-------
selected an UL of 6.0. As discussed in the proposal, we considered and requested comment on
alternative UL values of 4.5 g/bhp-hrand 5.3 g/bhp-hr NMHC+NOx (see 67 FR 2165). The primary
reason for the selection of an UL value of 6.0 is this value will allow technological laggards to
continue to produce existing engines (through the payment of a penalty), and allow such laggards to
focus their research and development efforts on complying with the 2004 emission standards. We
have concluded that this UL value is most consistent with the objective of the CAA's NCR provision,
as well as our previous NCR rulemakings. The reality of the existing market-place is that the vast
majority of HHDDE's are emitting at emission levels on the order of 6.0 g/bhp-hr NMHC+NOx.
Since the role of an NCR is to address the problems associated with a transition from a prior
emission requirement to a new more stringent requirement, it is appropriate to consider the effects
of the defacto HHDDE emission limits established by the Consent Decrees in establishing the UL for
the 2004 model year NMHC+NOx standard. The existence of the Consent Decrees October 1,
2002 emission requirements does not remove the technical issues laggards face, and we have in
fact received information which leads us to conclude a technological laggard is likely to exist for the
model year 2004 heavy-heavy service class, and such engines are currently operating at emission
levels on the order of 6.0 g/bhp-hr NMHC+NOx.
It is correct that we were not able to collect from engine manufacturers true cost data for a
complying heavy-heavy engine which operates at a 4.5 g/bhp-hr level to comply with the 2.5 g/bhp-
hr standard. All of the companies who provided data to us currently make products which emit at
off-cycle emission rates on the order of 6.0 g/bhp-hr. Those companies do not know what the costs
of compliance to go from a level of 4.5 to 2.5 would be, since they do not make such products. EPA
did request cost data from heavy-heavy companies who do make engines without high-off cycle
emissions, but those companies declined to provide EPA with cost estimates. However, even if we
had accurate estimates of the compliance costs associated with an UL of 4.5 for heavy-heavy duty
engines, an UL of 6.0 would still be more appropriate, for the reasons noted above.
We disagree with Caterpillar's comments that we have improperly included costs of
compliance with model year 2007 supplemental emission standards in establishing NCPs for the
model year 2004 standards. In 2004, engine manufacturers must comply with the FTP emission
standards, as well as the prohibition of defeat devices. In 2007, manufacturers must comply with
the FTP emission standards, which will include supplemental standards referred to as the
supplemental steady state standard and the not-to-exceed standard, as well as the prohibition of
defeat devices (See 40 CFR 86.004-11 and 40 CFR 86.007-11). Caterpillar states that the
proposal "implicitly concludes" that engines must comply with the 2007 supplemental standards in
2004. There is nothing in the record which supports this statement. In fact, this NCR is based on
compliance with the 2004 standards and the defeat device prohibition, not the 2007 standards. In
our request for cost data from engine manufacturers, EPA was clear that we were requesting cost
information for engines which companies are designing to comply with the 2004 standards. Our
request for information contained the following statements (See EPA Memorandum "Collection of
Compliance Cost Estimates for the Purpose of Establishing NCPs for the 2004 Heavy-Duty Diesel
NMHC+NOx Emission Standard", copy available in the docket for this rule, docket item ll-B-02):
• ".. we are collecting information in preparation fora notice of proposed rulemaking
regarding the establishment of non-conformance penalties (NCPs) for the model
year 2004 heavy-duty diesel engine NMHC+NQx standards."
• ".. we would like to gather company specific cost information regarding your cost of
compliance with the 2004 standards."
19
-------
"For the 2004 compliant engines, this should include a description of the
technological and calibration changes necessary to meet the emission standards."
• "The 2004 engine fixed costs should only include those costs attributable to meeting
the 2004 emission standards."
There is no evidence which supports Caterpillar's comments on this issue. We did not
modify the manufacturer costs (which were requested for compliance with the model year 2004
standards) in anyway to reflect compliance with the 2007 standards. We expect that manufacturers
would include in their estimates for MY 2004 compliant engines that such engines would not include
defeat devices, since an engine equipped with a defeat device would not comply with the 2004
requirements. Manufacturers at their option may have considered EPA's recent guidance regarding
defeat devices (such as Advisory Circular 24-3). Any such consideration would not be inconsistent
with our request that manufacturers cost estimates should be based on compliance with the model
year 2004 standards, as this guidance discusses defeat device prohibition. The technological
approaches to comply with the defeat device prohibition in model year 2004 would in any case be
similar to those used to comply with the model year 2007 supplemental standards for 2.5 g/bhp-hr
NOx+NMHC engines. This overlap in technology, however, does not invalidate the information
received from manufacturers concerning the model year 2004 standards. No adjustment was made
to the manufacturer data, nor was any adjustment necessary to address this overlap. Therefore, we
continue to believe that it is appropriate to rely on this manufacturer information.
Issue 2.7: Sources of Data
COMMENT: Caterpillar commented that EPA based the proposed NCP on incomplete information.
They stated that EPA's request for cost data from a number of engine manufacturers prior to the
proposal did not address all of the relevant issues. Specifically, Caterpillar commented that EPA's
request for information did not define critical terms like the term Upper Limit, and did not ask the
manufacturers to state the assumptions they used as to the regulatory status of the baseline engine.
Thus, they argue that the data collected would not be comparable. They also comment that it is
unclear whether the data collected by EPA reflects the costs of compliance with accelerated 2004
emission standards. Caterpillar also commented that EPA's request for cost data prior to the
proposal did not include all of the information EPA needed to complete the rule. Specifically,
Caterpillar points to EPA's statement in the proposal regarding the proposed Upper Limit value of
6.0 g/bhp-hr NMHC+NOx for the heavy-heavy service class for which EPA noted it "does not believe
it could readily develop the cost figures for such a development phase."
RESPONSE: We disagree with Caterpillar's comments that EPA's request for cost data prior to the
proposal did not address all of the relevant issues. First, it was not necessary for the collection of
the cost data prior to the proposal to define the term Upper Limit, or to inform manufacturers at that
time what Upper Limit values EPA was considering for the proposal, since EPA was not requesting
comments at that time, but rather we were requesting data. EPA's data request did specify both the
"baseline" engine and the "control" engine, for the purposes of estimating the costs of compliance.
As documented in an EPA memorandum to the docket, EPA's request for cost data clearly stated
that the baseline engine was "a model year 2002 engine", and the control engine was a "2004
20
-------
compliant engine".5 Before the proposal, we discussed with each manufacturer their submissions
to confirm that we properly understood them. This included discussions of the baseline engines.
We did, in fact, receive cost information from some manufacturers in response to our request for a
control engine that we determined were costs of compliance with the consent decree's October 1,
2002 requirements. We rejected and did not use these data in our analysis, and in every case we
requested and received from the engine manufacturer costs of compliance for the model year 2004
standards.
We disagree with Caterpillar's assertions that our data request prior to the proposal did not
request all of the data EPA needed for our proposal, and that our request was flawed. EPA's data
request prior to the proposal was not intended to request comment on all aspects an NCP rule.
That is the purpose of the public notice and comment process. Caterpillar's specific comment is in
reference to EPA's statement in the proposal that "EPA does not believe it could readily develop the
cost figures for such a development phase" (67 FR 2165). This statement was regarding the costs
of a development phase from an UL level of 6.0 g/bhp-hr NMHC+NOx to an UL level of 4.5 g/bhp-hr
NMHC+NOx for the heavy-heavy service class. This statement does not reflect a flaw in the data
EPA requested from engine manufacturers. Our request for data from manufacturers was not for
them to speculate on costs for a development program which they did not undertake. Rather, our
request from manufactures was for them to provide us with data that reflected the actual
development programs they were performing (i.e., making a model year 2001 engine comply with
the 2004 emission standards). For most heavy-heavy service class engine manufacturers, they
have not needed to undertake the development of a HHDDE which complies with an emission level
of 4.5 starting with an engine that complied with an emission level of approximately 6.0. That is, the
HHDDE manufacturers who are subject to the Consent Decrees manufacturer products which have
high off-cycle emissions, and these products are not required to comply with EPA's existing
regulations (i.e., the 1998 requirements). Most CD manufacturers will go directly from today's
product to a product which complies with the 2004 standard, by-passing engines which comply with
the 1998 regulations.
Again, EPA's request prior to the proposal was for the cost estimate for the development
programs that the manufacturer was actually undertaking keyed to costs for the 2004 model year,
not for a theoretical development program, or a prior model year. In addition, EPA took comment on
all of these issues, and received no additional or different data other than the data discussed above
and in the Technical Support Document. We also received no indications that manufacturers
misinterpreted our request for data.
Issue 2.8: Discount Rate
COMMENT: Two of the manufacturers commented that EPA should finalize the NCPs based on a
seven percent discount rate. One of the manufacturers commented that EPA should use a three
percent discount rate. Caterpillar cited a study by Ibbotson Associates to support using a real
discount rate of seven percent or higher. DDC stated that they use seven percent for their internal
analyses. Cummins cited EPA's Science Advisory Board and current interest rates for U.S.
Treasury Bills as support for using a three percent discount rate. Seneca Economics and
5 See EPA Memorandum "Collection of Compliance Cost Estimates for the Purpose of
Establishing NCPs for the 2004 Heavy-Duty Diesel NMHC+NOx Emission Standard.", copy available in
the docket for this rule, Docket Item ll-B-02.
21
-------
Environment (SEE) submitted comments critical of EPA's use of a seven percent rate for both
compounding and discounting and suggested that EPA should use different rates for past and future
cash flows relative to the base year.
RESPONSE: The purpose of a discount rate in the context of an NCP analysis is to reflect the
impact of manufacturer and user business decision making criteria with regard to meeting the CAA
principle that NCPs remove the competitive disadvantage for the complying manufacturer. The
rates used in the NCP should reflect the time value of money used by manufacturers and users in
business decisions and transactions. EPA has historically used a seven percent rate in rulemakings
in a societal costs context. The commenters supported the seven percent value as well as values
above and below it. Unfortunately, no user such as truck or bus fleet or owner commented on this
issue. The comments from SEE on the basic principles which should be followed in establishing
compounding and discount rates address the basic reason why EPA raised the issue in the NPRM.
That reason was that we anticipated that manufacturer costs and user costs may need to be
handled differently. However, SEE provided no data to assist in the analysis. The manufacturers
generally supported EPA's value, but it is in fact higher than the after-tax Treasury bill rate
suggested by SEE and is perhaps more in line with the time value of money/market risk premium for
future costs suggested by SEE.
For purposes of this rule, EPA is using of a seven percent rate. Based on present
information, it appears to be adequately representative of industry practice. However, EPA will likely
continue to include in future analyses sensitivity on other rates and to examine whether the rates for
future rules should be assessed in a manner similar to that suggested by SEE.
To put this issue into context, the effect of using different discount rates is shown in the tables
below. If a smaller discount rate had been used for both pre-production and operating costs, the NPV of
the fixed costs would have been lower, but the NPV of the operating costs would have been higher. The
net effect of a smaller discount rate would generally be penalties that are higher. While there are
differences, the differences are not large.
NCP Calculation Parameters with 7% Discount Rate
COC50
COC90
MC50
F
UL
X
Light-Heavy
$1,240
$2,710
$2,000
per g/bhp-hr
1.3
4.5
3.0
Medium-Heavy
$2,740
$4,930
$1,400
per g/bhp-hr
1.3
4.5
4.0
Heavy-Heavy
$6,810
$12,210
$5,600
per g/bhp-hr
1.3
6.0
3.4
Urban Bus
$3,930
$6,660
$3,800
per g/bhp-hr
1.3
4.5
3.3
22
-------
NCP Calculation Parameters with 3% Discount Rate
Parameter
COC50
COC90
MC50
F
UL
Light Heavy-Duty
Diesel Engines
$1,200
$2,770
$2,000
per g/bhp-hr
1.3
4.5
Medium Heavy-
Duty Diesel
Engines
$2,830
$5,340
$1,630
per g/bhp-hr
1.3
4.5
Heavy Heavy-Duty
Diesel Engines
$7,290
$13,720
$6,510
per g/bhp-hr
1.3
6.0
Urban Bus
Engines
$4,090
$7,230
$4,420
per g/bhp-hr
1.3
4.5
Issue 2.9: Estimated Market Share
COMMENT: Caterpillar stated that we should use the market share from the 2001 Ward's
Communication, Factory Sales to sales-weight our costs. Caterpillar provided this market share
data for 1999 and 2001 in their written comments.
RESPONSE: We agree that we should use the most recent sales data, and have updated our
analysis. Since the proposal we have received several estimates of sales for 2001, and have
updated the sales figures in final analysis for each service class for which we received new data.
"Data Received from Heavy Duty Diesel Engine Companies regarding 2001 Medium-Heavy and
Heavy-Heavy Diesel Engine Sales", William Charmley, EPA Memorandum, copy available in the docket
for this rulemaking.
23
-------
SECTION 3: ADDITIONAL RELEVANT ISSUES
Issue 3.1: Canadian Sales
COMMENT: Volvo and DDC commented that EPA should address the potential for operators to
circumvent U.S. requirements by purchasing their vehicles in Canada. DDC also recommended that
EPA "clarify the meaning of 40CFR86.1113-87(g)(1)(i) which, in relevant part, states that
non-conformance penalties must be paid "for all nonconforming engines or vehicles produced by a
manufacturer.... and distributed into commerce"." They stated that it is unclear whether the phrase
"distributed into commerce" is intended to mean only U.S. directed production or whether it includes
other engines that receive the NCP certification label, such as products delivered to Canada and
Mexico.
RESPONSE: The EPA Administrator issues a certificate of conformity for those vehicles or engines
determined to be in compliance with the Act and its implementing regulations. For those engines
covered by certificate of conformity, 40 CFR §86.095-35 (a) states:
The manufacturer of any motor vehicle (or motor vehicle engine) subject to the applicable
emission standards ( and family emission limits, as appropriate) of this subpart, shall, at the
time of manufacture, affix a permanent legible label, of the type and in the manner described
below, containing the information hereinafter provided, to all production models of such
vehicles (or engines) available for sale to the public and covered by a Certificate of
Conformity under §86.091-30(a).
The emissions label identifies the engine and provides additional specific information relative to
emission levels, date of manufacturer, model year, engine family, power, etc., and includes a
prominent statement that the engine conforms to U.S. EPA regulations for the applicable model
year.
Nonconformance penalties allow a manufacturer to introduce into commerce heavy-duty
engines(HDE) or heavy-duty vehicles (HDV), including light-duty trucks(LDT), which fail to conform
with applicable emission standards upon payment of a penalty. Accordingly, there are additional
regulatory requirements that apply to these engines and vehicles. Specific labeling requirements at
40 CFR §86.095-35 (h) (1) state:
Light-duty trucks and heavy-duty vehicles and engines for which nonconformance penalties
are to be paid in accordance with § 86.1113-87(b) shall have the following information
printed on the label required in paragraph (a) of this section, ..."The manufacturer of this
engine/vehicle will pay a nonconformance penalty to be allowed to introduce it into
commerce at an emission level higher than the applicable emission standard. The
compliance level (or new emission standard) for this engine /vehicle is XXX."
When labeling an engine as specified in 40 CFR §86.095-35(a) and (h), a manufacturer clearly
states that the engine conforms to U.S. EPA regulations and that the nonconformance penalty will
be paid for any engine on which the NCP label is applied. Labeling an engine as such without
payment of the penalty would be inappropriate and would misrepresent the status of that vehicle or
engine. The NCP payment is the basis for allowing the higher than applicable emission standard for
specific engine/vehicle. Without the NCP payment, the emission standard for such an engine is the
stated applicable emission standard and not the compliance level that would be applicable to an
24
-------
engine/vehicle under the NCR provisions.
In regard to the question of whether the phrase "distributed into commerce" is intended to
mean only U.S. directed production or whether it includes other engines that receive the NCR
certification label, such as products delivered to Canada and Mexico, it is clear that any engine
which bears the U.S. EPA certification label is available for introduction into commerce. It will
assumed to be distributed into commerce and should pay the appropriate penalty to the U.S.
Environmental Protection Agency in accordance with the NCP requirements. For those engines that
do not meet U.S. EPA emission requirements or that are intended solely for export, the engine
manufacturer must label the engines as such in accordance with the requirements for an export
exemption as stated in 40 CFR § 85.1709. The export label on these engines /vehicles would not
state that the engine conforms to U.S. EPA regulations and also would not state that a penalty had
been paid in order to allow its introduction into commerce. Provided engines/vehicles are properly
labeled, there will be little or no potential for operators to circumvent U.S. requirements by
purchasing their vehicles in Canada or Mexico. Vehicle purchased in Canada or Mexico for use in
the U.S. are subject to EPA regulations on imported vehicles.
Issue 3.2: Alternative Use of Penalty Funds
COMMENT: The SJVAPCD commented that revenues generated by the NCPs should be allocated
to ozone non-attainment areas to be used for projects that would compensate for the emission
reductions lost as a result of the use of NCPs. The PSCAA comments that the penalty money
should be made available for regional PM and toxics reduction projects. Similarly, the NYDEC
suggested that the NCP revenues should be directed "to the extent permitted by law" to diesel
retrofit projects to reduce NOx from on- or off-road heavy-duty engines.
Pony Pack Inc. (PPI) suggested that in lieu of paying NCPs manufacturers should be
provided the option of installing an auxiliary power unit (APU), in which case, PPI argues, "the
penalties assessed would actually go toward solving the problem." (PPI is a manufacturer of APUs).
PPI states that their proposal would "reduce emissions, which was supposed to be EPA's original
intent."
RESPONSE: Section 206(g) of the Clean Air Act, 42 U.S.C. §7525(g), authorizes EPA to establish
nonconformance penalties, but it does not authorize EPA to retain and use any penalty monies paid
by a manufacturer. Absent such authority to retain and use penalty monies received, the
Miscellaneous Receipts Acts, 31 U.S.C. §3302(b), requires that such monies be deposited in the
General Revenue Fund of the U.S. Treasury. Funds deposited in the General Revenue Fund may
then be appropriated by Congress.
Issue 3.3: Issues Regarding Collection of Data Prior to and After Proposal
COMMENT: Caterpillar commented that for the collection of cost data prior to the proposal EPA
"did not hold any industry workshops or meetings with the industry to discus its survey or the data
that would be used in the NCP rule." Caterpillar also commented that "EPA gave manufacturers
only two weeks to complete surveys. Given the significant potential effect of the NPRM .... this
plainly was not sufficient to gather the information necessary to promulgate NCPs."
In comments submitted by Caterpillar on April 8, 2002 and again on April 30, 2002
25
-------
Caterpillar commented that they were concerned that EPA requested updated information from
engine manufacturers after the close of the comment period. Caterpillar commented that if there
were limitations in the data used for the rule, EPA should have extended the comment period and
publically requested the necessary data. Caterpillar commented that they are concerned that the
final NCP rule will be based upon information from Caterpillar's competitors for which EPA has
provided no opportunity for comment. Caterpillar also commented that they are concerned with the
time frame for the request of additional information. Caterpillar commented that EPA provided less
than 20 days (April 8, 2002 letter from Caterpillar) or less than 10 days (April 30, 2002 letter from
Caterpillar) to respond on a matter that will have "a profound impact on the industry, the economy,
and air quality."
RESPONSE: We do not agree with Caterpillar's comment that it was necessary for EPA to hold an
industry workshop prior to the proposal for this rulemaking. Prior to the proposal EPA collected cost
data from a number of engine manufacturers to help in preparing the proposal for this rule. This
data collection request was provided in writing to a number of engine manufacturers, including
Caterpillar, as detailed in an EPA memorandum to this docket ("Collection of Compliance Cost
Estimates for the Purpose of Establishing NCPs for the 2004 Heavy-Duty Diesel NMHC+NOx
Emission Standard.", docket item ll-B-02). This data collection request was provided in a consistent
format to all manufacturers from whom EPA requested the information. We believe this collection
format was sufficient to provide EPA with adequate cost information to provide the basis of a
proposal. Caterpillar has provided no reason to explain how holding a workshop would result in
more or different data, especially given the confidential nature of the data submitted.
Caterpillar commented that EPA only provided two-weeks to respond to our pre-proposal
request for cost information from engine manufacturers. We believe this amount of time was
sufficient for the collection of cost information which was already in the manufacturers possession.
EPA was not requesting detailed analysis or comment on aspects of a proposal, we were simply
requesting information which manufacturers already had. To the extent a manufacturer needed
additional time they could have requested more time from EPA. In fact, a number of manufacturers
did request additional time, and EPA still made use of their data. Caterpillar provided no indication
to EPA that their cost data was not accurate, so we did not ask them if they needed additional time.
We disagree with Caterpillar's comment that EPA should have extended the comment
period because we requested updated information from some manufacturers after the close of the
comment period. After the close of the comment period, EPA contacted all of the engine
manufacturers who had previously supplied EPA with confidential cost information who had not
indicated in their public comments whether or not the data they previously supplied continued to be
accurate. We contacted Caterpillar (and other engine manufacturers) in order to confirm whether or
not the cost data they had provided in the late summer/early fall of 2001 continued to be an accurate
representation of their cost estimates for compliance with the model year 2004 standards. As
indicated in Caterpillar's letter to EPA of April 30, 2002, EPA also requested confidential internal
company documents from a number of engine manufacturers who had submitted confidential cost
data. The purpose of this request was to provide an additional verification that the data
manufacturers supplied to EPA was consistent with internal company documents which companies
use for briefings to senior management. Caterpillar's comments imply that EPA may not verify and
confirm the accuracy of cost data relevant to the rule after the close of the public comment period.
relevant to the rule. EPA believes it is appropriate to verify and confirm the accuracy of the cost
data to ensure that NCPs accurately reflect the competitive disadvantage to complying
manufacturers.
26
-------
Caterpillar commented they are concerned they do not have an opportunity to comment on
any updated information EPA receives from their competitors in response to EPA's request for
information after the close of the comment period. However, the purpose of EPA's request to
certain manufacturers was to confirm the validity and accuracy of the confidential cost data used for
the proposal. Caterpillar already had an opportunity to comment on this data during the public
comment period, and to provide its own data or information.
We also disagree with Caterpillar's comment that we did not provide them sufficient time to
respond to EPA's request for information. Caterpillar's comments state that EPA did not provide
enough time for Caterpillar "to respond on a matter that will have profound impact on the industry,
the economy, and air quality." However, we did provide a 60 day public comment period which
allowed Caterpillar and other interested parties an opportunity to comment on this rulemaking. In
EPA's requests for updated data and documents, we were not asking Caterpillar to respond on
specific issues, we were simply requesting Caterpillar to either provide data or to provide internal
company documents already in Caterpillar's possession. In response to each of EPA's request for
information, Caterpillar was able to provide the data, and they did not indicate that the information
they provide was inaccurate or limited by the time frame of the request.
27
------- |