SAVINGS FROM THE  APPLICATION OF TRADING AND
 AVERAGING TO HEAVY  DUTY  ENGINE REGULATION
               Prepared  for:

    U.S.  Environmental Protection Agency
             August  25,  1986
          SOBOTKA & COMPANY, INC.
            2501  M  Street,  N.W.
                 Suite  550
          Washington, D.C.   20037
               (202)887-0290

-------
                             TABLE  OF  CONTENTS
  I.  INTRODUCTION
     A.  Heavy Duty Engine Regulation and Emissions Averaging                  1
     B.  Issues Discussed  In  this
        Report:  Trading and  Averaging of NOx and PM	1
     C.  Methods  Used  to Evaluate  Savings from Averaging and Trading   .   .   2
     D.  Assumptions and Limitations of the Study  	   3
     E.  Summary  of the  Results	6


 II.  CALCULATION OF COSTS

     A.  Summary  of Types  of  Costs	8
     B.  Fuel  Effects  of Meeting NOx Standards	8
     C.  Trap  Effects	11


III.  SCENARIOS CONSIDERED

     A.  Regulatory Scenarios  	  15


 IV.  HOW AVERAGING AND  TRADING  GENERATE SAVINGS

     A.  An Example in Which  Averaging" Would Yield Cost Savings    ...  19
     B.  The Limits to Savings	20
     C.  Prorating to  Ensure  Emissions Do Not Rise	20
     D.  Prices of Credits and Their
        Relationship  to the  Allocation of Savings 	  22
     E.  Illustration  of Savings Relationships 	  23

-------
                        TABLE OF CONTENTS (continued)
 V. RESULTS

    A. Cost Comparisons by Scenario	27
    B. Comparisons of Emissions Levels    	  33
    C. Costs per Ton of Emissions Reductions	33
    D. Savings per Vehicle	36
VI. SENSITIVITY ANALYSES

    A.  Savings Due to Averaging and Trading In
       Comparison to a Baseline with Emissions Equal  to the Standard .    .   41
    B.  Reclasslflcatlon of Light-Heavy-Duty Engines  	   49
    C.  Sensitivity of Results to Assumed Functional Relationships    .    .   57
    D.  Banking	59
    APPENDICES

    A.  Optimization
    B.  Background Material  from ERC
    C.  Input Data for Computation of Savings  from Regulatory Flexibility
    D.  Detailed Results

-------
I.   INTRODUCTION
A.   Heavy-Duty Engine Regulation and Emissions Averaging

     On March 15, 1985 EPA promulgated  regulations  which significantly tighten
NOx emissions standards for heavy-duty  engines  (HDEs)  and  emissions of partic-
ulate matter (PM) for heavy-duty diesel engines  (HDDEs).  The  NOx  standard for
the 1988-90 model year HDEs is 6.0 g/BHP-hr, with a more stringent  standard of
5.0 g/BHP-hr effective for 1991 and later model  year engines.  Model year 1988-
90 HDDEs will be required  to meet a PM standard of 0.60 g/BHP-hr.  In 1991, all
HDDEs except  urban  bus engines must  meet a  0.25  g/BHP-hr  standard.   Urban
buses, which  are excluded  from averaging,  are  not  included  in  this  study.

     Standards for NOx and  PM  for 1988 are  intended to be met by each individual
engine family.  NOx and PM  limits  for 1991, on  the other  hand, may be  met by
groups of engines:  some  engine  families  may have  emissions in excess  of the
standards so long as the  group meets the standards on average.  Engines will be
considered to be  part of  the same  group,  or   "averageable  set,"  if  they are
built by the same firm and fall within the  same  subclass.   The four subclasses
envisioned are  heavy-duty  gasoline  engines  (HDGE),  light-heavy-duty  diesel
engines (LHDDE), medium-heavy-duty diesel  engines (MHDDE),  and heavy-heavy-duty
diesel engines  (HHDDE).  This added  flexibility,  termed  "emissions  averaging,"
is intended to  allow  cost  reductions for the manufacturers  with  no compromise
in meeting the air quality goals of the regulations.
B.   Issues Discussed in this Report:   Trading and Averaging of NOx and PM

     The purpose of  this report  is  to quantify  the  implications  of  carrying
this concept of averaging further.   It shows the annual  savings  that  could be
realized if additional  flexibility  is  permitted  in meeting  the   same  overall
emissions targets.   The industry as a  whole  could  be  allowed  to meet the  stan-
dards "on  average"  by permitting  the  inter-firm trading  of   "credits"  fo

-------
                                      -2-

emlssions below the  standards.  This system, termed "emissions trading," extends
the intrafirm benefits of averaging by enlarging the averageable sets.  Trading
increases the total   savings  while at  the  same time offering  increased  equity
across firms (since  not all firms would be able  to  take advantage of the benefits
of averaging alone to the same degree).

     This report also explores the cost-saving  potential  of  allowing averaging
or trading  across broad  sets of  engine types, instead  of  within narrow  sub-
classes.  Comparisons are  made among  three subclass  assumptions.   The  first
assumption is that HDE emissions averaging/trading  is  "restricted" to the  four
subclasses described above, with  averaging  or trading allowed within each class
but not between them.  The  second  assumption, referred to as "partially restric-
ted" averaging  or  trading,  involves   averaging/trading  among  the three  MODE
subclasses, but not  between  the two  HDE classes  (MODE  and HDGE).   The third
assumption, "unrestricted"   averaging   and  trading,  permits  averaging/trading
among the  three  HODE  subclasses  and  between   the  two HDE  classes   (HDDE  and
HDGE).
C.   Methods Used to Evaluate Savings from Averaging and Trading

     Cost savings under flexible regulations—averaging  or  trading—are  calcu-
lated by estimating the total costs of compliance with regulations traditionally
formulated on a  command-and-control  (every engine family must  pass)  basis  and
comparing these  costs  to  the costs  of  meeting  the same emissions goal  in  the
most efficient way allowed by more  flexible rules.  The most  efficient  way to
meet a  given  emissions goal  is found  by applying the economic  principle  that
the greatest efficiency in production is reached  only if the marginal  costs of
production are the same for all  production units.   (In  this  case, the "product"
is emissions reduction, and the  "production units" are  heavy-duty engines.)   In
practice, achieving  equal  marginal   costs  of  pollution control  among  engine
classes means that  some  engines  should emit more than average,  while other
engines balance these  excess emissions  by emitting less than  average.   Engine
types that  have  emissions below the  standard  accrue credits.   The   number  o

-------
                                      -3-

credlts that are thus created is proportional  to the number of grams per brake-
horsepower hour (g/BHP-hr) below the  standard.  Where averaging  or  trading  is
permitted across engine classes/subclasses, to ensure that total  (and to aid  in
ensuring localized) emissions do not  exceed the  levels  emitted  if each  vehicle
met the standards individually, account must be taken of differing characteris-
tics of each engine class or subclass.  Hence, in such cases, emissions  are not
necessarily averaged  or  traded on  a  one-for-one  basis;  they  are  "prorated"
according to sales, miles traveled  per year, years  of useful  life, and/or power
output per mile.   (An  example  of of  the prorationing concept is  presented  in
Section IV.)

     Marginal costs of emissions reductions are  calculated  based  on functional
relationships (cost/emission curves) relating NOx emissions levels, PM emissions
levels, and  percentage   increases  in  fuel consumption.   The  functions  were
developed from estimates of the technical relationships made by a  specialist  in
HDE emissions.^./  Individual functional  relationships were  developed for each
of eight engine type and usage  classifications (see Appendix B for a discussion
of the classification  system),  including  both  gasoline and diesel  types.   An
optimizing algorithm is used to find  the combination of emissions  levels across
different types of vehicles that equalizes marginal  costs.  (See Appendix A  for
an explanation of the optimization  program.)

     The distribution  of the  savings  from trading  for  individual  firms  is
estimated in  this  study  on  the basis   of  the  savings  realized   by  different
engine type  and  usage classes, and individual  firms' projected  shares  in the
sales of those classes.  That is, a  firm with a large share of the  market  for
those engine  types  that benefit  substantially  will  gain  relatively  more.
0.   Assumptions and Limitations of the Study

     Costs of meeting  the  target  levels  of  only  the NOx  and PM  regulations
effective in 1991  are estimated  in  this  study.  Target levels  are set at  80
     JV  See Section II.B. and Appendix B.  Data supplied by C. Weaver  of ER

-------
                                      -4-

percent of the standards to take into account production variation in emissions
or the normal  in-use deterioration of emissions performance over time, and thus
to include a. cushion to  allow for  engine-to-engine  variability.   Given  this
assumption regarding maintenance of  design cushions, the  "effective standards"
become 4.2 g/BHP-hr for  NOx and 0.22  g/BHP-hr PM.  Gasoline engines are included
in NOx averaging and trading only; diesel engines are  included  in both  NOx and
PM averaging and trading.  Trucks to  be sold in California,  which has separate
standards, were not incorporated in this study.

     The analyses in this  report  consider  emissions from an  engine only up to
the first  rebuilding.   Under  some  circumstances,  allowing  averaging  between
engines with  different   likelihoods  of  being   rebuilt could have  unintended
effects on emissions. For  instance,  an emissions trade may allow  a one g/BHP-hr
increase in NOx emissions for a LDDE-DI to  be offset by a  one g/BHP-hr decrease
in a LHDGE if each  were  assumed to travel the same number  of miles before being
rebuilt and exert the same number of  BHP-hrs  per mile.  This trade  would not
result in  any  emissions increase  if neither engine is rebuilt.   If, however,
the LHDDE-DI j_s  rebuilt  and the  LHDGE  is  not,  the  excess  emissions   of the
LHDDE-DI after the  rebuild will not  be offset  and  overall  emissions will rise.

     This study is  intended  to  be indicative  of  the  kinds and  general  magni-
tudes of percentage savings that would be  seen with more  flexible regulations,
but estimates of total emissions control costs,  cost  savings, or the distribution
of those savings over the  various  firms  cannot be made with  a  great degree of
precision.  The  main reason  for  this limitation  is  that   technological  and
market forecasting   for  points  in time as  far  away  as 1991  is  necessarily an
inexact undertaking.  If  sales in the  industry are much greater,  of  course,
total costs and total savings will  also be  greater.   Also, because the analysis
is driven by assumptions about  market shares  held by different types of vehicles
and different firms, the results will be sensitive to unforeseeable shifts (and
past experiences have shown such shifts to  be quite common).

     The recognition that  exact  predictions  of savings  are  impossible  led to
the decision to  limit  the  number  of  factors  considered  to those  considere

-------
                                      -5-

most Important.  Left out of  the  analysis are examinations of  how  differences
in defining useful  lives of the vehicles and some of the less  important catego-
ries of  emissions  control  costs   (including  research,  testing,  and  hardware
changes to meet NOx standards) affect the results.

     The analysis is limited  to examining the maximum savings  possible  due  to
the added flexibility of  emissions  averaging  and trading.   The  degree  to which
the firms would take advantage of the program to approach  the  maximum savings
is not considered, though it  is  expected that the  use of  the  system  would  be
extensive: firms have  an  economic  incentive  to  trade  credits  to  lower  their
costs.

     The study also  considers  cost savings for  only one future  year, meaning
that issues of spreading  emissions over  time  in  the  most  efficient  manner
(banking) are  not  addressed,  except in a  sensitivity analysis  in  Chapter  VI.
Banking would increase  cost savings potential.

     During the course  of this study, every effort was made  to use cost, sales,
technology, market  share, and  other information which was completely consistent
with EPA's Regulatory Impact Analysis  (RIA) for the 1991 HDE  NOx  and  Particulate
Emissions Standards  (see  Appendix C).  'However,  in  several  areas this  was  not
possible because the data  or  other information  needed  for  this  study  was  not
addressed directly  in  the RIA.  Thus,  to  meet the   needs of  this  study,  some
information from other  sources was  used and in  some  cases  additional  analyses
were necessary.  This in no way detracts from the  results of this  study,  however,
since the point here is to  evaluate the incremental  cost savings of different
averaging and trading programs over a given base  case.  As long  as the  input  is
consistent for the  base case and the averaging and trading cases, the percentage
cost savings  estimated  are accurate.   It  is also   reasonable  to  assume  that
percentage savings  found here reflect those which would  occur if this  analysis
had been incorporated directly into EPA's  RIA.

     Also, it  should be noted that  due  to  the  different  input parameter  values
used here, the baseline costs  for the 1991 standards are somewhat  higher  tha

-------
                                      -6-

those found in EPA's RIA.   However,  EPA still supports the  analysis  presented
in its RIA as the best  estimate of the costs of compliance.
E.   Summary of the Results

     A summary of the results, presented  in  Exhibit  I,  shows  that  averaging of
NOx emissions and of  PM  across  all  HDEs  (the least  restrictive  averaging  sce-
nario) could save over $191  million  per year, or  19 percent  of  the industry's
emissions control  costs  in  the  absence  of  averaging.   When the  industry is
divided into two subclasses, HDGEs and HDDEs, emissions  averaging  within those
groups saves $158 million, almost 16 percent of  industry  control  costs.   These
savings may be compared to the savings of $123 million  (12%  of baseline  costs)
that result  from averaging  emissions  under  the  current  averaging  regulations,
which divides engines  into  four  subclasses  (HDGE,  LHDDE,  MHDDE,  and  HHDDE).
(Estimates of cost  savings  were  not  previously  calculated  in the EPA  FRM of
March 15, 1985.)

     Trading can save  an additional  $107 million,  or  13 percent  of industry-
wide control costs,  in  the  least  restrictive   scenario,  over  and above  the
savings from averaging.  A substantial  portion  of the savings would  go  to the
firms with  lower  gains  under averaging  without  trading.    Trading  not  only
results in  a near-doubling of the cost  savings  under averaging  alone  but  also
widens the  set  of  firms  participating in the  gains from  increased regulatory
flexibility.  Trading thus helps  to  keep differential effects on  the firms  to a
minimum.

     Trading saves   less  compared  to  averaging  i.f  more  restrictive  subclass
assumptions are made.  If  HDEs are divided  into  two  subclasses,  trading  can
save an  additional   $9 million compared  to  averaging;  if four  subclasses  are
assumed, trading would save $8 million.   Detailed descriptions of the scenarios
and the savings are included in Section  III  and  Appendix D.

-------
                                      -7-

                                   Exhibit I
                            EMISSIONS CONTROL COSTS
                         (millions of dollars per year)
Baseline - No Averaging or Trading

Four Subclass Averaging *
Two Subclass Averaging **
One Class Averaging ***
Control Costs

  $ 1,009.6

  $   886.5
      851.4
      818.1
                                                            Savings
Four Subclass Averaging and Trading *   $   878.7
Two Subclass Averaging and Trading **       842.4
One Class Averaging and Trading ***         710.9
vs. Baseline:
$ 123.1    12.2%
  158.2    15.7
  191.5    19.0

vs. Averaging:
$   7.8     0.9%
    9.0     1.1
  107.2    13.1
  *  Restricted
 **  Partially restricted
***  Unrestricted

-------
                                      -8-

II.  CALCULATION OF COSTS

A.   Summary of Types of Costs

     The costs  Imposed  by the regulations  that  are considered  in  this  report
fall into the following categories:   1) costs due to fuel  consumption increases
resulting from  NOx-reduction techniques,  2)  costs due to  fuel  consumption in-
creases caused  by  PM-reducing trap  oxidizers  ("traps"),  and  3)   initial  and
maintenance costs of traps.   These are  not  the only costs imposed by the regula-
tions, but they are by  far  the most  important,  and are covered  in more detail
below.


B.   Fuel Effects of Meeting NOx  Standards

     Using conventional control  technology,  reducing  NOx  emissions from heavy-
duty truck engines* is expected to result in increases in fuel consumption.  The
percentage increase will  vary with the  NOx emissions  level,  increasing sharply
as emissions per BHP-hr  approach  4.0 grams.  The  size of the  increase  at any
given emissions level  also varies  considerably across different types of engines
and depends heavily on the  control  technique used.  Exhibit  II-l  shows point
estimates of the tradeoff between NOx and fuel consumption increases at various
NOx levels (targets, not standards) as  provided  by ERC Inc. (EPA's RIA analysis
was not adequately detailed to be used  here)V.   The  estimates are shown for 9
types of engines types,  which are described in  Appendix  B.   For computational
convenience, and to obtain estimates  of the consumption increases at intermedi-
ate NOx levels,  smooth hyperbolic functions were  fit to  the  point estimates.
These functions are shown in Exhibit  II-2.

     Translating these percentage  fuel  consumption increases in dollar  values
(assumed to be imposed on  the manufacturers by market forces—though the average
     ]_/  Estimates for gasoline engines assume that 3-way catalysts, which have
not yet  been  shown  to  be  suitable  for  heavy-duty  use,  will  not be  used'

-------
                                      -9-
                                  Exhibit II-l

               POINT ESTIMATES OF NOx/FUEL CONSUMPTION TRADEOFFS
ENGINE*/USAGE TYPE

1. Light-Heavy
   Indirect Injection

2. Light-Heavy
   Direct Injection

3. Standard
   Medium-Heavy

4. Premium
   Medium-Heavy

5. Line-Haul

6. Vocational
   Heavy-Heavy

7. Light-Heavy
   Gasoline

8. Medium-Heavy
   Gasoline
 NOx TARGET (g/BHP-hr)
 2.5   3.0   3.5   4.0   4.5   5.0   6.0   8.0
          FUEL CONSUMPTION INCREASES
 15%    8%          2%          0%    0%    0%


             12%          6%          1%    0%


             16%          7%          3%    0%


             12%          6%          1%    0%

              8%          4%        0.5%    0%


             10%          5%          1%    0%


6.5%    5%        2.5%          1%    0%    0%


6.5%    5%        2.5%          1%    0%    0%
*  All  are diesel  engines, except 7 and 8.

-------
            NOx/FUEL CONSUMPTION  FUNCTIONS
z
z
o
u
o:
o
u
                                                                  o
                                                                  x
                                                                -< m
                                                                -o i—
                                                                m
                                                                PD <-> m
                                                                oo o x
                                                                o z =r
                                                                i— co -••
                                                                i—i cz cr
                                                                o 3 -••
                                                                  -O c-r
                                                                  o •-«
                                                                  "Z. I
                                                                    ro
                                                                CD 70
                        TARGET G.'BHF-HR OF IJOx

-------
                                      -11-

cost increase for all  competing engines may be passed through)  was accomplished
by multiplying them by  EPA estimates  of  the discounted cost  of fuel  for the
engines per percent increase in fuel  consumption.^/  Fuel  costs were discounted
to allow for  the fact that  truck  purchasers are  less  affected by  costs  that
will not be incurred for several years.   These  values are  shown in  Appendix C.
C.   Trap Effects

     Two cost impacts of using traps to control  PM emissions may be identified.
First, the traps themselves  are  expensive to buy and to  replace if  and  when
they wear out.   Costs for initial and  replacement  trap  oxidizer  systems  were
provided by EPA£/ for different engine/use types.

     Second, the use of the traps increases fuel  consumption  by roughly 0.5 to
1 percent.2.7  This impact is translated into dollars in the  same  way that  NOx-
related fuel consumption increases are treated.

     Under the baseline  scenario,  all  diesel  engines would  need  traps.  Under
averaging, however, only some engines  would need  to  be fitted  with traps since
those with traps  will  be averaged in  with  those without traps.   To calculate
the total  costs  of  the regulations,  then,  it  is  necessary  to  estimate  the
fraction of engines needing traps.  This depends on the efficiency of the trap,
the PM  standard,  and the engines'  PM emissions  levels  without  traps  (called
"engine-out" emissions).
     2.7  These fuel consumption functions were not used directly; EPA estimates
were used:  Regulatory  Impact  Analysis,  Oxides of* Nitrogen  Pollutant Specific
Study and Summary  amir Analysis of Comments, Control of  Air  Pollution from New'
Motor Vehicles and New Motor Vehicle Engines:  Gaseous  Emission  Regulations for
1987 and  Later  Model  Year Light-Duty  Vehicles,  and for 1988  and  Later  Model
Year Light-Duty Trucks and Heavy-Duty Engine s; P a rt i cu1 ateTmTssion Regulations
7or~lg8B  and  Later ModTPYear Heavy^Duty  Diesel  Engines.  USEPA Uffice of Air
and Radiation, Office of Mobile Sources, March 1985.
     £/  See Appendix C.
     2/  See Appendix C.

-------
                                      -12-

     Traps have been assumed to be 80 percent efficient, trapping 80 percent of
engine-out PM emissions.^/  Engine-out PM  emissions  are expected to vary  with
NOx emissions-and engine  types  in a manner  similar to the relationship  of NOx
emissions to fuel  consumption increases:   Exhibits II-3 and II-4 show the point
estimates provided  by  ERC  and  the  curves developed  by Sobotka  & Co.,  Inc.
(SCI) for various  NOx levels.

     Using this information, and  given a  NOx and PM standard,  it  is straight-
forward to calculate the percentage of traps needed for a single type of engine
to meet the standard individually.  For example, if a NOx standard is such that
engine-out PM  is  0.5  g/BHP-hr, and  the PM standard is  0.22,  then  (0.5-0.22)/
(80% * 0.5), or 70% of engines will need traps.£/  Not  all  engine types need to
meet the PM standard exactly, though, since the averageable set includes various
engines made  by  a  firm.   Firms  are expected  to  lower their  total   costs  by
placing the traps  where they are most cost-effective:  on engines where  the tons
of PM  removed by the  traps are  large compared  to  the  costs imposed  by the
traps.  A  simple  algorithm  is  used in the  analysis  to "remove" traps  on  more
and more engines,  starting with those on which the traps are least cost-effective
until the  firm's  average  PM emissions just meet  the  standard.   Appendix  A
provides detail on this procedure.

     Total regulatory costs for the  baseline  scenario  are  found by  summing NOx
costs and  adding  the  costs  imposed  by the traps  to  get the  costs  per truck.
These costs are then multiplied  by the  projected sales of the type of engine/use
class considered and by the share that the manufacturer is  projected to have of
that engine/use class by 1991 (from EPA; see Appendix C).
     }_l  Regulatory Imapact Analysis,  March 1985,  op.  cit.
     £/  The RIA  (p.  2-65)  shows  that  70% would  need  traps with  engine-out
emissions of 0.42  to  0.54 g/BHP-hr and  a target  of  0.25;  this  is  consistent
with the text.

-------
                                      -13-


                                  Exhiblt II-3

                        POINT ESTIMATES OF ENGINE-OUT PM

                    EMISSIONS ASSOCIATED WITH NOx EMISSIONS
ENGINE*/USAGE TYPE

1. Light-Heavy
   Indirect Injection

2. Light-Heavy
   Direct Injection

3. Standard
   Medium-Heavy

4. Premium
   Medium-Heavy

5. Line-Haul

6. Vocational
   Heavy-Heavy
                            NOx TARGET (g/BHP-hr)
                            2.5   3.0   3.5   4.0   4.5   5.0   6.0
                                          8.0
ENGINE-OUT PM EMISSIONS (g/BHP-hr)
0.60  0.52
0.46
0.45
            0.65


            0.75


            0.58

            0.45


            0.54
      0.50


      0.60


      0.44

      0.37


      0.40
      0.34  0.30


      0.44  0.40


      0.32  0.28

      0.28  0.25


      0.30  0.27
*  All  are diesel  engines.

-------
             NOx/PM TRADEOFF FUNCTIONS
o:
I

a.
tn
a:
o
                    TARGET G/BHP-HR OF NOx

-------
                                      -15-

III. SCENARIOS CONSIDERED


A.   Regulatory Scenarios

     Costs are computed and compared under seven scenarios:  a baseline scenario
In which each  Individual engine  must  meet the standards; three  scenarios  that
allow averaging; and  three  more that allow  trading  in  addition  to  averaging.
(Trading may  be  thought of  as averaging  across  manufacturers.  Because  more
than one  manufacturer is  involved, the  system  of  accounting   for  emissions
increases and  decreases  is more  complicated,  but  the  concept   is  identical.
This concept is described in Section  IV'.) The  scenarios  differ  in terms  of the
restrictions or boundaries  of  their  averageable  sets, i.e.,  in  terms of  the
groups of engines which must meet the  standards on average.

     The differences among the scenarios  are  illustrated  in  Exhibit  III  with a
brief explanation of  each.  The  exhibit  shows  an  industry  comprised of  two
firms, A and B, each  of which markets  eight  families  of HDEs: two  HDGEs  (Al,
A2, Bl, and  B2); two  LHDDEs (A3, A4,  B3,  and B4); two MHDDEs (AS, A6, B5,  and
B6); and two HHDDEs  (A7, A8,  B7, and B8).  The engine families distributed  in
each box  represent  averageable  sets:  in a given  scenario,  emissions  from  one
engine family may be averaged  with  emissions  from any  engine  family  in the  same
box, but not with emissions from engines in other boxes.

-------
                                      -16-


                                  Exhlblt III
                              REGULATORY SCENARIOS
Baseline Scenario
Firm: A
No Averaging or Trading
Subclass:
HDGE LHDOE MHDDE HHDDE
Engine Al
A2
Bl
B2
A3
A4
-B3
B4
A5
A6
B5
B6
A7
A8
B7
B8
In the baseline scenario no averaging or trading  is permitted, and emissions from
each individual engine family are required to comply with standards.
Four Subclass Averaging
Firm: A

      B
Subclass:
HDGE LHDDE MHDDE HHDDE
Engine "Kl
A2
Bl
B2
A3
A4
B3
B4
A5
A6
B5
B6
A7
A8
B7
B8
In this scenario, which closely  resembles  current regulations, the averageable
sets are large enough to  include different engine  families  built by the  same
manufacturer and  falling  into the  same heavy-duty subclass.   For  instance,  A3
and A4, which  are  both  LHDDEs built  by Firm A, are  shown  in the same  box  to
indicate that  their  emissions  may  be averaged together.   Emissions from  A3,
however, could not be averaged in with  emissions  from B3 (a LHDDE built by the
other firm), or  from  A5 (an engine built  by  the  same firm but  in a different
class).
Two Subclass Averaging
Firm: A
Subclass:
HDGE LHDDE
Engine Al
A2
Bl
B2
A3
A4
B3
B4
MHDDE
Ab
A6
B5
B6
HHDDE
A7
A8
B7
B8
This scenario widens the  averageable set to  allow emissions  from  any  diesels
produced by a given  firm to be  averaged in with  emissions  from any other  of
that firm's diesels.   For example,  engine  A3 now  falls  in the  same box,  or
averageable set, with HHDDE A8.

-------
                                      -17-
                            Exhibit III (continued)
One Class Averaging
Firm:
Subclass:
HDGE
A
B
Engine Al
A2
Bl
B2
LHDDE
A3
A4
B3
B4
MHDDE
A5
A6
B5
B6
HHDDE
A7
A8
B7
B8
The final averaging scenario widens  the  averageable set still further to include
all of a firm's  engines,  whether they are diesel  or gasoline  fueled,  small  or
large.  (An important  exception  is  that  gasoline engines,  which do  not  emit
significant amounts of PM,  are  not  included in PM averaging.)  In the  exhibit,
all eight
other.
of A's engine families are in the same  box  and  all  of B's are in the
Four Subclass Averaging and Trading
Firm: A
Subclass:
HDGE LHDDE MHDDE HHDDE
Engine Al
A2
Bl
B2
A3
A4
B3
B4
A5
A6
B5
B6
A7
A8
B7
B8
This scenario permits averaging of emissions across firms, which is referred to
as trading.  The  same  subclass restrictions apply in  this scenario  as  in the
first averaging scenario, however:  gasoline fueled  engines  are in a  separate
averageable set from LHDDEs,  which in  turn  are separate  from MHDDEs,  and so
forth.  In the exhibit, emissions  from engines A3  and  A4 may now  be averaged
with emissions from B3, but  not with Al, Bl, or A8.
Two Subclass Averaging and Trading
Firm: A
Subclass:
HDGE LHDDE
Engine Al
A2
Bl
B2
A3
A4
B3
B4
MHDDE
A5
A6
B5
B6
HHDDE
A7
A8
B7
B8
In this scenario, the restriction on  averaging  different  heavy-duty  subclasses
of diesels is relaxed; all HDDEs produced by any  firm  are included  in  the  same
averageable set.  The exhibit  shows  all  engine families  in  one of  two  boxes:
all gasoline  engines  are  in  one  box,   and  all  diesels  are  in  the  other.

-------
                                      -18-
                            Exhibit III (continued)
One Class Averaging and Trading
Firm:
Subclass:
HDGE
A
B
Engine Al
A2
Bl
B2
LHDDE
A3
A4
B3
B4
MHDDE
A5
A6
B5
B6
HHDDE
A7
A8
B7
B8
The least restrictive  scenario  includes all HDEs  in  the same averageable  set
(though, as  with  one  class  averaging,  gasoline  engines are  not included  in
particulate  matter averaging).

-------
                                      -19-

IV.  HOW AVERAGING AND TRADING GENERATE SAVINGS


A.   An Example in Which Averaging Would Yield Cost Savings

     Allowing a given emissions target to be met with  averaging  or trading  can
reduce costs because it allows emissions control  activities that are relatively
inexpensive to be  substituted for those  that  are relatively  expensive.  As  a
simple example, consider a manufacturer whose total sales are one HHDDE and  one
standard MHDDE.  The  total  costs  of  meeting  the same  NOx  standards may  be
$3,000 for the HHDDE and $1,000 for the MHDDE, for a  total  of $4,000.  The fact
that the total costs  are higher  for the  larger  truck does not  by itself mean
that emissions controls are necessarily more costly per kilogram of NOx removed,
since $3000  spent  cutting emissions in  a large,  intensively  used truck  might
easily reduce total emissions  by  just  as much as  $3000  spent  on three smaller
trucks.  The important consideration in determining  if savings are possible is
not total  costs  but  the change in total  costs occurring  for  given  changes in
emissions.

     In other words, suppose that  removing an additional  ton of NOx by tighten-
ing the controls on  the  HHDDE would  raise the total  costs  of  control  for this
truck to $3010,  and  that increasing  emissions from the MHDDE  by one ton  would
reduce costs  for that  truck   to  $980.   The  same total  emissions  would take
place, but with one ton of emissions  "transferred" from the HHDDE to the MHDDE.
By reallocating  the  emissions  controls  in  this  way,  the manufacturer  could
ensure that  the  same  emissions  would   occur,  but at  a total  cost  of  $3990.

     In this example, we would  say that the marginal  costs of emission control
were $10 per additional  ton for the' HHDDE and  $20 per additional ton  for  the
MHDDE, since these are  the  dollar changes in total costs that  result  from  one
ton changes  in emissions.   It  is  this  difference between the marginal  costs of
emissions control that allow  savings to be generated:  the  marginal savings  per
ton of emissions transferred  from one  truck  to the other  is  equal to the dif-
ference between  their  marginal costs:   $10 per  ton transferred  from the  MHDD

-------
                                      -20-

to the HHDDE.   In terms of per engine costs, the costs for the HHDDE have risen
by $10 and the costs for the MHDDE have fallen by $20.
B.   The Limits to Savings

     It pays to  keep transferring emissions  from one  engine to the  other  so
long as the difference  between  the marginal  costs  remains.   As  emissions  are
reduced more and more  from  the  HHDDE, however, the marginal  cost  for reducing
emissions from it will tend to  rise.  The  opposite  will happen to the marginal
cost of emissions  reductions  for the MHDDE as the  controls are eased  on  it.
The gap  between  their marginal  costs  will  therefore  narrow, and  eventually
disappear.  At that  point, the transferring  should  stop,  since  all  of  the
potential gains will  have been  squeezed out,  and  the maximum savings  will have
been realized.  The  process of  transferring  to  save  control  costs  is  called
"averaging" because  of  the requirement that  per-vehicle  emissions  standards
still be met  "on  average"  after the  transferring.   If  a  one ton  increase  is
always balanced by a  one ton  decrease,  then  average emissions for the  set  of
vehicles covered--the "averageable set"--are  constant.  The  same  is  true with
respect to trading, i.e., the  tons balance out—only in this case, the average-
able set is the overall  industry.
C.   Prorating to Ensure Emissions Do Not Rise

     In an averaging  or trading program, emissions  from one engine  class are
allowed to increase  so  long  as emissions from  another class are  reduced.  To
ensure that total emissions  do  not  rise as a result,  the  program  must provide
for an appropriate system for testing whether the decreased  emissions  from one
class are  sufficient  to  cancel  out the  increased  emissions  from the  other.

     This might not be a simple matter.   In the example above, emissions changes
are described in one  ton increments.  However, regulations are generally written
not in terms of tons of  emissions over  the  life of  a  vehicle, but  in  terms  o

-------
                                      -21-
grams per brake  horsepower-hour  (g/BHP-hr).   Lowering the emissions  of Engine
Class B by 0.2 g/BHP-hr is not necessarily enough to compensate for an increase
in Engine Class  A's  emissions  of the same 0.2  g/BHP-hr.  The  two  changes  will
cancel out if the same number  of horsepower  hours  will  be used by trucks using
Engine Class B as will be used by  trucks  using Engine Class A.  This circumstance
is unlikely,  and  if  Class  A is  installed  in more  trucks,  or generates  more
horsepower per mile  in use,  or travels  more  miles  per year,  or typically has a
longer useful life, then  a given  reduction in emissions by  Engine  Class B  will
not make up  for the increase from Class A.   A  greater number of tons of emissions
will  result  over the  life of the engines  than  if  each class had been held exactly
to the standard.

     A partially  restricted  or  an  unrestricted averaging  or  trading  program
could essentially ensure  that  emissions  would  not  increase by  prorating  the
reductions and increases  in certification  levels to take sales, power, use pat-
terns, and length of useful lives into  account.  The  reduction  of 0.2 g/BHP-hr
for Class B  in the example above  should be multiplied by  a  factor proportional
to the total  number  of horsepower-hours expected to  be  expended by the engines
of Class B  before  comparing  the  reductions  to  similarly-prorated  increases  in
emissions of  engines  in  Class A.   The  following example illustrates  how  this
would work.
                                   Class A
Sales per year:                       1,000
BHP-hrs/mile:                         0.25
Miles per year:                      20,000
Years of useful  life:                    9
Total BHP-hrs:                  45,000,000
Total metric tons of
emissions from a change of
0.2 g/BHP-hr:                           90
Total metric tons of
emissions from a change of
1.0 g/BHP-hr:                          450
  Class B
      500
     0.20
   15,000
        6
9,000,000
       18
       90

-------
                                      -22-

Because each year the  Class  A fleet  would  generate five times  the  horsepower
hours of the Class  B  fleet, a change  in Class A's certification level  would have
to be  balanced  by  an opposite change in Class  B's  level  that was  five  times
greater.  A 0.2 g/BHP-hr emissions increase  for  Class  A would be  fully  offset
by a  1.0  g/BHP-hr  decrease   in emissions for Class  B;  then a 90  ton  increase
would be offset by  a  90 ton decrease.

     The regulatory scenarios discussed  in  this  paper  are assumed to  employ  a
prorating system,  identical  to  that described  above,  to   ensure that  total
emissions are the same with and without  averaging or trading.  In the operation
of the method for calculating cost savings,  all  changes  in emissions  levels are
converted into tons over the  life of  a year's sales fleet before being compared,
or weighted by the sales fleet's  total horsepower hours  before being averaged.
Traded credits are  measured and priced in terms  of tons  of total  emissions over
the life of the sales fleet.
D.   Prices of Credits and Their Relationship to the Allocation of Savings

     The example given  above  is simple partly  because the whole  "averageable
set" of engines  is  assumed to  be  built by  the same  manufacturer.   Thus,  the
fact that the total  emissions  control costs  rise for  one  of the engines  is  not
important; the manufacturer is more than compensated  by the lower  costs  of  the
other.  If the averageable  set  cuts across boundaries between  firms,  however,
as it does in  "trading" proposals,  new  issues are  raised.   The  same principles
of savings and  their  relationship  to  differences  in  marginal  costs apply  if
engines are built by  different  companies,  but the  problem of  compensating  the
manufacturer of the  engine whose  total control  costs are  increased  arises.

     This problem is taken care  of  by  awarding  credits (prorated, as described
above) for each  g/BHP-hr  below  the standard to a manufacturer  whose  engines
emit less than  the  maximum allowed.   These  credits  may  be sold  to firms  in
complementary positions—those  increasing  emissions and  reducing  cpsts.   The
purchasing firms would be required  to show that they  had  obtained  a credit  fo

-------
                                      -23-
each unit of emissions their products would emit over the allowed emissions for
each vehicle.

     The establishment of provisions for  awarding  the  credits and for allowing
them to  substitute  for emissions  reductions  creates  a small  but  potentially
efficient market for trading of  credits.   Market  forces will tend to force the
prices of the  credits  toward the marginal  cost  of generating  them  (and  their
marginal value  in   substituting  for  emissions  reductions).   Assuming  trades
occur when  it  is in the participants'  economic  self-interest  to do  so,  i.e.,
they trade on the basis of their marginal  costs and each manufacturer makes use
of the  market  to its  greatest  extent, then  all  manufacturers'  products  will
(after all the trades  are  completed)  show the same marginal  cost of emissions
reductions.

     Market participants will  benefit  more  from  the   market   for credits the
greater the  difference between  their  marginal   costs  and  the  industry-wide
marginal cost.   Those  with  costs  that are much  higher than  average  realize
large advantages both  because  the  average savings they realize per  kilogram
traded are large and  because they  will find  it  useful to buy  many  kilograms'
worth of credits.   Those with  costs much  lower than average will gain  just as
much--generating and  selling a  larger  number  of credits  at  a  large  average
profit.  Those  with marginal  costs  close to  the  industry's  marginal  costs
before trading will  gain  little by trading,  but  they will  not  lose  anything
either.  This is illustrated in the following  section.
E.   Illustration of Sav-ings Relationships

     Exhibit IV illustrates the  relationship  between costs  relative  to  ipdus-
try-wide marginal  costs and  the  cost savings available through  trading.   Each
of the three panels  shows  a marginal  cost curve representative of industry-wide
marginal costs of emissions reductions (per ton), in addition to a marginal cost
curve for one of  three different  firms:  A,  B, and C.  Before  trading, each  of
the three  firms  must  meet  the  same emissions  standard,  and are assumed  t

-------
                                   -24-

                                Exhibit IV
              TRADING: SAVINGS AS RELATED TO MARGINAL COSTS
$/ton
  MCa -
MC,
                MCa.
                                                   Marginal Cost - firm a
                 Marginal Cose - industry
                                                 Emissions Reduction - Firm A
 $/ton
                                                   Marginal  Cost -  firm  b

                                                     Marginal  Cost  -  industry
                                                  Emissions Reduction - Firm B
 $/ton
   MC  -
                                       Marginal Cost - industry
                                                                   MC,
             Marginal Cost - firm c -**'   . MCC
                                        ER,,
                                                  Emissions Reduction - Firm C

-------
                                      -25-

achieve that standard by setting emissions reductions at ERa,  ER^,  and ERC.   At
these levels of  emissions  reductions, the  firms'  marginal  costs  of  emissions
reductions (MCa»  MC-b, and   MCC)  vary widely with respect to the industry average
(MCi): B's cost for removing additional tons is just slightly above  the average,
but A's is much higher and  C's is much lower.

     If trading of emissions reduction credits is allowed, C finds it worthwhile
to increase its emissions  reduction,  generating  credits that may  be  sold to A
and (to a lesser degree) to B.  These credit purchases  allow  A and  B  to  reduce
their emissions reductions, saving enough  in  costs to  pay  for the  credits  and
leave a net saving.

     The savings  of  the various  firms  after trading are  illustrated  under  the
assumption that the  credit  price (per  ton  of  emissions  reduction)  reaches
equilibrium at the industry  average  marginal   cost before trading.   Each  firm
will  set  its  emissions reduction levels  so  that their  marginal   costs  under
trading equal  the price of credits.  (Emissions  reduction levels are  shown  as
ERa', ERb', and ERC'; marginal costs  are  MCa'  etc.)  Excluding its  payments  for
the credits, A's costs  fall by  the integral  of its marginal  costs  from ERa  to
ERa'—the area of the trapezoid  ERaMCaMCa'ERa'.   Payments  for  the credits equal
the price  (MCi)  times  the  quantity   (ERa  minus  ERa'), which leaves  A's  net
savings as the area of the  shaded triangle.

     The shaded triangle  representing B's  savings is  much  smaller--the tri-
angle's base (the number of credits  purchased)  and height (the net savings  on
the marginal ton  of emissions  reductions) are  both  proportionately  smaller than
for A's.  Thus, the net savings  fall   rapidly  as  a firm's marginal   cost  before
trading approaches the industry-wide  marginal  cost.

     C's situation is the  reverse of that  of  A's, but it leads to equivalent
net savings.  C's costs actually  rise  as it increases  its  emissions  reduction
levels, but the credits it  generates  in the process may be sold at  a price that
yields an  overall  gain.   This  gain   is  shown as  the  shaded triangle  in the
third panel of the diagram:  the triangle's dimensions  are  again equal  to the

-------
                                      -26-

number of  units  of  credits  traded  (ERC*  -  ERC)  and  the difference  between
the firm's marginal  cost  before and after trading  (MCC - MCC').   Net  savings
resulting fro/n trading are, as shown by the diagrams, independent  of  whether a
firm has  relatively  high  or  relatively low  marginal  emissions  control  costs.
The savings are,  however,  much greater for firms whose  costs  differ sharply in
either direction  from the  industry as a whole.

-------
                                      -27-


V.   RESULTS


     The first part of  this  section presents the  emissions  control costs  for

individual  firms and the industry for each of the regulatory  scenarios.   First,

the control costs  and  savings of  NOx  and PM regulations are  compared  for  all

scenarios.   This is followed  by  a  presentation  of the  differences  between  the

scenarios in  emissions  levels,  marginal   costs  of  emissions control,  and per-

engine costs and savings.


     As discussed  in  Section  III,  the costs  are  estimated  by  combining  the

costs of fuel  consumption  increases due  to NOx  control  and traps,   and  the

hardware and maintenance costs of traps.   The total costs are higher than would

be calculated under EPA's assumptions, due to differences in  the treatment of

fuel efficiency  gains   over  time.   The   savings  from  averaging and  trading,

though, are consistent  with  EPA assumptions.
A.   Cost Comparisons by Scenario


     Exhibit V-l shows that substantial  savings,  both  in absolute terms and as
a percent of total  emissions  control  costs, would be  realized  by the industry

under the One Class  Averaging  scenario.^./  The industry would save almost $192
million per year, 19 percent of  baseline control  costs, if emissions averaging
     }j  The dollar figures  for  the savings from averaging  and  trading listed
for each firm do not,  it should be remembered, refer only to manufacturing cost
savings to the  firms.   Most of  the  savings are  fuel  savings that  accrue  not
directly to the  firms  but  to  the purchasers  of the engines.   Thus,  the term
"savings" is really a  short-hand way to say "savings directly to the firms plus
increases in the value  of  the firms' products."  In large measure, the increased
product values resulting  from the fuel  savings  will  be  captured by the firms—
just as a  firm  with a  less efficient product is  forced  to  reduce its price to
stay competitive, a firm  with  a  more efficient product  is able to increase  its
price without losing sales.

-------
                                                  Exhibit V-l

                                                  COST SAVINGS
                                        ONE CLASS - AVERAGING AND TRADING
                                               (dollars In millions)
Firm

1 Bluebird
2 Chrysler *
3 Ford
4 CumraLns
5 Caterpillar
6 Daimler-Benz
7 KHD
8 General Motors
9 Navistar
10 Hlno Motors
11 Deere
12 Mack
13 Onan
14 Perkins
15 Renault
16 Saab
17 Isuzu
18 Iveco
19 Volvo
20 White
Baseline
100Z traps
No Averaging Averaging


$0 0
3
41
382
139
8
1
201
84
T
0
91
0
0
12
1
12
5
a
a
90
06
97
44
46
31
96
94
49
90
53
41
90
97
08
33
56
27
11

SO
3
33
329
124
7
1
131
59
3
0
79
0
0
11
1
9
5
7
7

0
9
9
5
1
6
1
4
6
2
8
8
3
8
4
0
9
2
3
1
Dollar and Percentage Savings from
, Trading Averaging
Trading (vs Averaging) (vs No Averaging)

(SO
-0
2
302
114
6
1
113
56
3
0
73
0
0
10
0
9
4
6
6

0)
8
0
4
0
9
1
2
1
0
7
2
2
7
4
9
3
a
7
4
(S)
$0 0
4 7
32 0
27 1
10 1
0 7
0 1
18 2
3 5
0 3
0 1
6 6
0 1
0 1
1 1
0 1
0.6
0 4
0 7
0 7
U)
128
121
94
8
8
9
8
13
5
8
10
8
47
10
9
10
6
7
9
9
a
6
2
2
1
6
5
a
a
i
0
3
4
0
3
0
3
9
2
6
(S)
$0 0
0 0
7 1
53 4
15 3
0 8
0 2
70 6
25 4
0 3
0 1
11 8
0 1
0 1
1 5
0 1
2 4
0 4
0 9
1 0
(X)
0 0
0 0
17 4
14 0
11 0
9 /
12 3
34 9
29 9
7 4
9 9
12 8
29 1
9 9
11 8
9 9
19 5
7 2
11 2
12 /
                                                                                                                                   I
                                                                                                                                  ro
                                                                                                                                  CO
        INDUSTRY TOTAL
                               $1,009 6
                                            $818 1
                                                      S/10 9   $107.1
                                                                          13 1   $191
                                                                                            19 0
• The negative sign for Chrysler Indicates that for this firm the value of traded credits would
  exceed the cost of generating them enough to more than offset the costs oi compliance with
  the regulations

-------
                                      -29-

were allowed.^./  Still  further  savings—over  $107  million—would accompany the
introduction of trading  (the  One Class  Averaging  and  Trading  scenario),  espe-
cially aiding-those firms with  less  ability to  take  advantage  of the averaging
program.  Moreover, trading "evens out"  the distribution of savings that occurs
with averaging to  more  closely represent the  original  distribution  of  costs
incurred.

     The gains from averaging would  be  spread quite  unequally  in both absolute
and percentage terms.  Among domestic firms, General  Motors, Navistar, and Ford
would reap the highest percentage savings.  Bluebird and Chrysler, on the other
hand, would  save  almost nothing through averaging.  However,  no company loses
through averaging.

     This variation in savings  relates directly to the market segments in which
the different  firms  participate.   Those  with  broad  market  coverage—General
Motors is the prime  example,  with LHDDEs,  MHDDEs,  HHDDEs,  plus HDGEs—are able
to take full advantage  of an  averaging  scheme.  Those whose  product lines are
much narrower—with engines that differ  little in emissions characteristics—are
less fortunate: the marginal  costs  for  emissions  reductions in their different
engines are  similar  or  identical even  without  averaging,   so  the potential  to
lower costs by averaging is very limited.

     Incremental  gains  from trading  in  addition to averaging  are also distrib-
uted unequally:  some  firms   show  substantial  gains  while  others  are  barely
better off.  (Every  firm is at  least no  worse  off  under  trading since  it can
choose not  to  trade  credits  at all, leaving it  in  exactly the  same position
as with averaging  without trading.)  As  mentioned  above,  for many  firms the
distribution of large  and  small  gains  from trading  counterbalances in  some
important ways the distribution for" averaging  by  itself:   a  firm with  one  of
the largest  gains  under  averaging,  Navistar,  gains only   slightly  more  under
     */  About half of these savings  result from the fact that, under averaging,
manufacturers are  not  forced to  over-control  for  PM,  installing  expensive
traps that  generate  emissions reductions in  excess of those  required  to mee
the emission standards.

-------
                                      -30-
trading.  And Chrysler, one of the  firms  with  the  smallest average gains  with-
out trading, gains significantly  under trading.

     Again, the  reason  for this  pattern  lies in  the  breadth  of the  firms'
product lines: those firms with  product  lines  that come close  to matching the
industry's sales mix are able  to  bring the marginal  emissions control  costs for
all of their  products close to  the industry-wide average  without  trading with
other firms,  and  consequently are  faced  with very  little  scope  for  further
savings once  they are allowed to  trade with other  firms.   Conversely,  a firm
like Chrysler with  only one type  of engine--and one whose  emissions  charac-
teristics differ  sharply  from those of the  rest  of  the  industry—finds  great
opportunities to save from trading.

     As shown in  Exhibit  V-2, when averaging  is permitted within  two industry
subclasses (partially restricted  averaging) most firms'  emissions control  costs
are the same as with one-class averaging.  Ford and  General Motors, as the only
manufacturers of  both HDDEs  and HDGEs  are the  only ones which save  less  by
averaging under this subclass  assumption.

     For most firms,  averaging  and  trading  within  two  subclasses results  in
significantly lower savings than  those resulting from trading within one class.
Navistar and  Onan are the only  manufacturers with  larger  savings  with  trading
(as compared  to  only averaging)  in this  scenario  than  in  the previous  one.
However, in  absolute  terms the  savings  are slight;  percentage  increases are
partly due to the relative magnitudes of control costs under  averaging compared
to trading in both subclass scenarios.

     Exhibit V-3 shows that if (restricted) averaging, limited to four industry
subclasses, were  allowed,  firms manufacturing  more  than one  type of MODE  would
save less  than  in the  two  subclass scenario.   The  two  firms  saving the most
with industry-wide averaging,  Navistar and  General  Motors, gain the least with
four subclasses.  As expected, there is  no  change  in the savings  realized  by
firms with only HDGE or only one type of HDDE.

-------
                                                                Exhibit V-2

                                                                COST SAVINGS
                                                   TWO SUBCLASSES - AVERAGING AND TRADING
                                                           (dollars In millions)
                                                                            Dollar and Percentage Savings from
Engine Type*

C
C
C D
D
D
D
D
C D
D
D
D
D
D
D
D
D
D
D
D
D

1
2
3
4
5
6
7
8
9
10
11
Firm

Bluebird
Chrysler
Ford
Cummins
Caterpillar
Daimler-Benz
KHD
General Motors
Navistar
Hlno Motors
Deere
12 .Mack
13
U
IS
16
17
18
1'J
20
Onan
Perkins
Renault
Saab
Isuzu
Iveco
Volvo
White
Baseline
No Averaging Averaging

$0
3
41
383
139
8
1
202
84
3
0
91
0
0.
13
1
12
5
8
8

0
9
1
.0
4
5
3
0
9
5
9
5
4
9
0
1
3
6
3
1

SO
3
38
329
124
7
1
159
59
3.
0
79
0
0
11
1
9
5
/
7

0
9
8
5
1
6
1
9
6
2
8
8
3
8
4
0
9
2
3
1
Trading
Trading (vs Averaging)

SO 0
3 9
36.0
329.4
123 9
7 6
1 1
159 2
54 8
3 2
0 8
79 7
0 1
0 8
11 4
1 0
9 9
5 1
7 3
7 0
(S)
$0 0
0 0
2 7
0 2
0 3
0 0
0 0
0.6
4 7
0 0
0 0
0 0
0 2
0 0
0 0
0 0
-0 0
0 1
0 0
0.0
(Z)
0 0
0.0
7 1
0 0
0 2
0 5
0 1
0 4
8 0
1 0
0 5
0 0
71 5
0 5
0 3
0 5

1 0
0 3
0 4
Averaging
(v* No Averaging)
(S)
$0 0
SO 0
$2 3
$53 4
S15 3
SO 8
$0 2
$42 1
$25 4
SO 3
SO 1
$11 8
$0 1
$0 1
$1 5
SO 1
$2 4
$0 4
SO 9
SI 0
(X)
-0 0
0 0
5 6
14 0
11 0
9 7
12 3
20 8
29 9
7 4
9 9
12 8
29 1
9 9
11 8
9 9
19 5
7 2
11 2
12 7
                              TOTAL
                                            $1.009 6
$851 4    $842 4
                                                                               $9  0
                                                                                         11    Slbd  2
                                                                                                          15  7
• C = HDCE. D = HDDE

-------
                                                Exhibit V-3




                                                COST SAVINGS


                                  FOUR SUBCLASSES - AVERAGING AND TRADING


                                           (dollars In millions)
                                                                            Dollar and Percentage Savings from
Engine Type*

C
G
C




C















L
L M H
M H
M H
M H
L M H
L M
M
H
H
L
M
L H
N
L H
H
M H
L H

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Firm

Bluebird
Chrysler
Ford
Cummin*
Caterpillar
Daimler-Benz
KHD
General Motors
Navistar
Hlno Motors
Deere
Mack
Onan
Perkins
Renault
Saab
Isuzu
Iveco
Volvo
White
Baseline
No Averaging Averaging

SO
3
41
383
139
8
1
202
84.
3
0
91
0
0
13
1
12
5
8
a

0
9
1
0
4
5
3
0
,9
i
9
5
4
9
0
1
3
6
3
1

0 0
3 9
38 8
333 7
124 5
7 6
1 2
177 8
71 0
3 2
0 8
79 9
0 3
0 8
11 6
1.0
10.7
5 2
7 4
7 3
Trading Averaging
Trading (vs Averaging) (vs No Averaging)

0 0
3 9
36 5
330 7
124 4
7 6
1 2
176 5
69 9
3 2
0 8
79 9
0 3
0 8
11 6
1.0
10 7
5 2
7 4
7 1
(S)
$0 0
0 0
2 2
3 0
0 0
0 0
0 0
1 3
1 1
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 1
(Z)
0 0
0 0
5 7
0 9
0 0
0 0
0 0
0 7
1 6
0 0
0 0
0 0
5 0
0 0
0 1
0 0
0 0
0 1
0 0
2 0
(S)
$0 0
0 0
2 3
49 2
IS 0
0.8
U 2
24 2
13 9
0 3
0 1
11 6
0 1
0 1
1 3
0 1
1 /
0 4
0 9
0 9
(Z)
- 0 0
0 0
5 6
12 9
10 8
9 7
12 1
12 0
16 4
7 4
9 9
12 7
29 1
9 9
10 4
9 9
13 5
7 2
11 1
10 6
                                                                                                                                          I
                                                                                                                                         u>
                                                                                                                                         ro
                                                                                                                                          i
                     INDUSTRY TOTAL-
                                             I.009 6
$886 S    $878 7
$7 8
                                                                                        0 9   $123  1
                                                                                                         12 2
* G - HDGE, L = LHDOE, M = MHDDE, H = HHDDE

-------
                                      -33-
     General  Motors  and  Cummins,  the manufacturers  with  the most  diversity
in product line,  gain  the most  savings  from  trading  under the  four  subclass
scenario as opposed  to two subclasses.   White also saves more under averaging
compared to the baseline; all  other firms'  savings decline.

     The percentage  of revenues  saved from trading within  the entire  industry
plotted as a function of the  percentage  saved from averaging  are presented in
Exhibit V-4 (one class case) for the individual manufacturers.  Here the inverse
relationship between the  gains  from  averaging  and trading  for  some  firms is
readily apparent: Navistar, General Motors, and Isuzu all experience significant
cost savings with averaging,  but  little  additional savings when trading is also
permitted.  Ford, Chrysler, and  Bluebird,  however, save  little  or  nothing by
only averaging,  but  significant  savings   accrue  when  trading  is  allowed.
B.   Comparisons of Emissions Levels

     This section presents the changes in the patterns  of  emissions that would
result from  changes  to  more restrictive  regulatory  scenarios.    Exhibit  V-5
shows that as the  subclasses become more  restrictive, the  range  in emissions
levels across engine classes decreases.  For  example,  there  is  a difference of
only 0.09 between the minimum and maximum NOx levels when emissions  are averaged
in four subclasses, and  a 1.73  difference when  averaging  within one  class is
permitted.

     In all  scenarios, the  range in  both  NOx and  PM  levels are smaller  when
emissions are averaged  .than  when  emission  credits are  averaged   as  well  as
traded.
C.   Costs per Ton of Emissions Reductions

     In the course  of  solving for  the  most  efficient  emissions  levels,  the
optimizing model   used  for  this  analysis  generates  the  costs  of  additiona

-------
                                       Exhlhl
UJ
D

U
u.
O
C?
z
a:
u
O
or
L-
no
                 TKAuiNG  ANu  A'vtiRACiNG
                              % OF REVENUES SAVED, ONE CLASS
      0%
                 D NAVISTAR
 DGM

a isuzu
       a CUMMINS
        P MACK
                                             D ONAN
        DKHD
          CATERPILLAR
       n-WHCTE.VOLVO
       D RENAULT

         B DEERE, PERKINS, SAAB, D.BENZ
          BHINO,  IVE<
                                                                                       CJ
                                                                                       *.
                                                                                       I
                                                             D FORD
                                                         DDLU
IBIRD
                            SAVINGS FROM TRADING, 95 OF REVENUES

                           	  REGRESSION: R2=.64

-------
                                      -35-
One Class
Engine Class:
        LHD6E-
    LHDDE-IDI-
     LHDDE-DI
        MHDGE-
     MHDOE-NA
     MHDDE-TC
     HHDDE-LH
    HHDDE-NLH
                                  Exhibit V-5
                                EMISSIONS LEVELS *
                                 (grams/BHP-hr)
                      NOx
                      Averaging
             Trading
                2,
                3,
                5,
     25
     67
     06
                2.21
                4.80
                5.17
                4.90
                4.97
           PM
           Averaging

                  na
               0.431
               0.384
                  na
               0.108
               0.251
               0.246
               0.085
Trading

     na
  0.479
  0.419
     na
    ,106
    ,375
    ,204
                             0,
                             0.
                             0.
                             0.072
Two Subclasses
Engine Class:
        LHOGE
    LHODE-IDI
     LHDDE-OI
        MHDGE
     MHDOE-NA
     MHDDE-TC
     HHDDE-LH
    HHDDE-NLH
NOx
Averaging
4.21
3.50
4.20
4.19
4.41
4.13
4.20
4.16

Trading
4.22
3.37
4.43
4.19
4.36
4.55
4.21
4.20
PM
Averaging
na
0.443
0.377
na
0.114
0.238
0.246
0.087

Trading
na
0.493
0.487
na
0.115
0.425
0.174
0.086
Four Subclasses
Engine Class:
     LHDDE-DI
        MHDGE
     MHDDE-NA
     MHDDE-TC
     HHDDE-LH
    HHDDE-NLH
NOx
Averaging

     4.21
     4.19
     4.23
     4.19
     4.21
     4.19
     4.23
     4.14
Trading

   4.22-
   3.75-
   5.54
   4.19-
   4.23
    ,18
    ,23
4.
4,
   4.14
PM
Averaging
na
0.219
0.224
na
0.122
0.273
0.286
0.087

Trading
na
0.166
0.383
na
0.119
0.274
0.286
0.087
*  Figures for trading  scenarios  are actual emissions  levels;  those for aver-
aging scenarios are averages across  all manufacturers,  weighted by sales.  The
regulations specify that  no engines  may  exceed 6.0  g/BHP-hr  of NOx  or 0.60
g/BHP-hr for PM.   This  constraint  was  not  binding  in the  scenarios examined
(i.e., even under  unconstrained trading and  averaging no engine would excee
these limits).

-------
                                      -36-

emlsslons reductions  for  individual  firms and  for the  industry  as a  whole.
Exhibit V-6 presents  these marginal  costs per ton  of  PM  and  NOx (emitted over
the life of a'year's fleet)  for the regulatory scenarios.

     This data may be compared to the costs  of emissions reductions  from other
sources for cost-effectiveness analyses.   It  is  apparent that  reductions in PM
are more costly than NOx emissions reductions.  For all industry subclasses the
average PM control costs are  greater than those for NOx.
D.   Savings per Vehicle

     The savings allowed  by  industry-wide  averaging and trading  varies  widely
across individual   firms.   Exhibits V-7,  V-8,  and  V-9  present the  per-engine
costs and savings under the three  regulatory scenarios.  The  per  engine  values
are, in  a  sense,  normalized,  and  thus are  more  representative of  relative
equity achieved across  firms  from averaging and  from trading than  the  values
indicated in Exhibits V-l, V-2,  and V-3.

-------
                               -37-

                          Exhibit V-6
       MARGINAL CONTROL COSTS OF  EMISSIONS  CONTROL
                      (dollars  per  ton)
                         ONE CLASS
                                                    TWO CLASSES
                         SOX
                                  PM
                                                CAS
                                                NOX
                                                        NOX
DIESEL
   ?M
AVERAGING
Finn
3laeblrd
Chrysler
Ford
Cummins
Cat erp lilac
Daimler-Benz
KHD
General Hotors
Savlstar
Si.no Motors
Oeere
Mack
Onan
Perkins
Renault
Saab
Isuzu
Iveco
Volvo
White

31.167
1.186
1.270
4.110
4,245
4,443
4,333
1,880
3.3&9
4,716
4.436
4.109
1.594
4.436
4,363
,436
.074
.753
,275
,401



S18.584
6.026
5.935
5.739
5,990
7.427
13.059
5.547
5.744
6.027
13.583
5,744
5.795
5.744
6.002
5.521
5.358
5.769
SI. 167
1.186
1.169




1.177














34,399
4.110
4.245
4.443
4.333
3,653
3,349
4,716
4,436
4.109
1.594
4,436
4.363
4.436
4.074
4,753
4.275
4.401


S12.S41
6.026
5.935
5.739
5.990
6.306
13.059
5,547
5.744
6.027
13.683
5,744
5.795
5,744
6,002
5.521
5.358
5.769
TRADING
                      S2.321   $6.853
                                              SI.174  $4.009   $6.093
                                    FOUR CLASSES
GAS LIGHT
SOX SOX
AVERAGING
Firm
Bluebird
Chrysler
Ford
Cummins
Caterpillar
Daimler-Benz
KHD
General Motors
SavLstar
Blno Motors
Deere
Mack
Onan
Perkins
Renault
Saab
Isuzu
Iveeo
Volvo
White
SI. 167
1.186
1.169 S4
4



1.177 1
1



1

1

2


4

,399
,398



.581
.581



.594

,594

.696


.406
LIGHT
PM

S12.
12.



13,
13,



13.

13.

13.


12.

641
641



682
682



683

683

268


629
MEDI'JV
sex


S4
4
4
4
4
4
4
4


4
4
4
4
4
4
4


.436
,647
.463
.789
.615
.521
.715
.436


.436
.436
.436
.469
,753
.436
.435
MEDIUM
PM


SS
5
5
5
5
5
5
5


5
5
5
5
5
5
5


, 744
.595
.725
.750
.617
,684
.547
.744


.744
.744
,744
.720
,521
,744
,744
HEAVY
NOX


S4
34
S4
34
34



$4






$4



,109
,109
,139
.109
,109



.109






.109

HEAVY
PM


36,027
$6.027
$6,027
$6.027
S6.027



$6.027






$6,027

TRADING
               $1.174  $2.494  $13.337   $4.557   $5,659  $4,109  $6.027

-------
           Exhibit V-7
ONE CLASS - AVERAGING AND TRADING

   PER ENGINE COSTS AND SAVINGS
Sales
13
34.016
139.009
93.817
35,259
3.023
297
329,138
79,647
1.191
331
17,441
794
331
5.265
397
7,345
1.885
2,408
3,753
Firm
1 BlOeblrd
2 Chrysler
3 Ford
4 Cummins
5 Caterpillar
6 Daimler-Benz
7 KHD
8 General Motors
9 Navistar
10 Hlno Motors
11 Deere
12 Mack
13 Onan
14 Perkins
15 Renault
16 Saab
17 Isuzu
18 Iveco
19 Volvo
20 White
I.USLS ---- 	 -
Basel Ine
No Averaging Averaging
S22B
115
295
4,082
3,955
2,799
4,398
614
1,066
2,931
2,710
5,248
518
2,710
2,464
2,711
1.679
2.<>5u
3.436
2,160
$228
115
244
3.512
3.520
2,527
3.856
399
748
2,714
2.441
4.574
367
2,441
2,174
2.442
1,352
2,737
3,050
1,886
- - - - savings -------- .-.-
Savings from Savings from
Trading Averaging
Trading (vs Averaging) (vs No Averaging)
($64)
(25)
14
3,223
3,234
2.285
3,526
344
704
2,496
2.198
4,195
193
2.198
1.973
2,199
1.267
2,520
2,770
1,704
$292
140
230
289
286
242
330
55
44
219
243
379
174
243
201
243
85
^17
280
182
' ($0)
0
51
570
435
272
542
214
318
217
269
674
151
269
290
269
327
212
386
275
                                                                                     I
                                                                                    OJ
                                                                                    CO
                                                                                     I

-------
           Exhibit V-8
TWO SUBCLASSES -  AVERAGING AND TRADING
     PER ENGINE COSTS AND SAVINGS
Sales
13
3d, 016
139.009
93,817
35,259
3,023
298
329,138
79,647
1,191
331
17.441
794
331
5,265
397
7,345
1.885
2.408
3,753
Firm
1 Bluebird
2 Chrysler
3 Ford
4 Cummins
5 Caterpillar
6 Daimler-Benz
7 KHD
b General Motors
9 Navistar
10 Hlno Motors
11 Deere
12 Mack
13 Onan
14 Perkins
15 Renault
16 Saab
17 Isuzu
IB Iveco
19 Volvo
20 White
Baseline
No Averaging Averaging
$223
115
295
4.082
3,955
2,799
4.398
614
1,066
2.931
2,710
5.248
518
2,710
2,464
2,711
1.679
2.950
3,436
2,160
$223
115
279
3,512
3.520
2,527
3,856
486
748
2,714
2,441
4,574
367
2,441
2.174
2.442
1,352
2.737
3,050
1.886
- - - - savings -------
Trading
Trading (vs Averaging) (vs
$223
115
259
3.511
3,513
2.514
3.854
484
688
2.687
2,429
4,572
104
2,429
2,168
2.430
1.354
2.709
3.042
1,878
$0
0
20
2
7
13
2
2
60
27
12
2
262
!2
7
12
(2)
28
8
8
Averaging
No Averaging)
$0
0
17
570
435
272
542
128
318
217
269
674
151
269
290
269
327
212
386
275
                                                                                          LJ

-------
              Exhibit V-9
FOUK SUBCLASSES -  AVERAGING AND TRADING


     PER ENGINE COSTS AND SAVINGS
Sale*
13
34.016
139.009
93,817
35,259
3.023
297
329.138
79.647
1.191
331
17.4*1
794
331
5.265
397
7.343
1,885
2,408
3,753
Firm
1 Bluebird
2 Chrysler
3 Ford
4 Cummins
5 Caterpillar
6 Daimler-Benz
7 KHD
8 General Motors
9 Navistar
10 Hlno Motors
11 Deere
12 Mack
1 3 Onan
14 Perkins
IS Renault
16 Saab
17 Isuzu
18 Iveco
19 Volvo
20 White
Cons -----------
Basel Ine
No Averaging Averaging
$228
115
295
4.082
2.339
2.799
4,413
614
1.066
2.931
2,710
5.248
518
2.710
2.464
2.711
1.679
2.950
3.436
2.160
$228
115
279
3,557
2.088
2.528
3.877
540
892
2.714
2,441
4,580
367
2,441
2.208
2.442
1,451
2,737
3.054
1.932
- - - - bavlngs -------------
Trading Averaging
Trading (vs Averaging) (vs No Averaging)
$228
115
263
3.525
2,088
2,527
3,876
536
878
2.713
2.440
4.580
348
2.440
2.205
2.441
1,451
2,735
3.053
1,895
$0
0
16
32
0
1
0
4
14
1
1
0
IB
1
3
1
1
2
1
38
so-
0
17
525
251
271
536
74
174
217
269
668
151
2b9
256
269
227
212
383
228
                                                                                                     o
                                                                                                      I

-------
                                      -41-

VI.  SENSITIVITY ANALYSES
     This chapter contains the  results of  four analyses of  issues  related to
averaging and trading.   The first takes up the  question  of  the extent to which
the savings from averaging and  trading are attributable to  greater  efficiency
in attaining a given level of emissions reductions,  or  to the opportunity that
averaging/trading provides to meet the emissions  standards  exactly  instead of
over-controlling emissions in many  cases.  The second  examines  the potential
impact of a reclassification  of the  smallest heavy-duty engines  as  light-duty
trucks, and the  reduction in the  scope  of  averaging  and  trading  that  would
result from  this  action.  The  third  examines  the  degree  to  which  the  cost
savings results are sensitive to differences  in input cost  functions, i.e., in
assumptions about the relationships  of NOx control  levels  to fuel  consumption
and PM emissions.   The  final  analysis estimates the  potential  of emissions
credit banking to ease  the transition  from the 1991 PM standards to the more
stringent 1994 standards.
A.   Savings Due to Averaging and Trading in Comparison
     to a Baseline with Emissions Equal  to the Standard

     The averaging program (and  any  potential trading program)  is  intended to
increase the  efficiency of  emissions  control   by  allowing  real locations  of
control efforts to  engines  for  which  control is less costly,  while  ensuring
that overall emissions  are no  higher than they  would  be without  the  program.
For this report, the constraint  that emissions be no  higher  with an averaging/
trading program has been  interpreted to mean that they  must  be no  higher,  on
average, than  if  each  vehicle  met  the  target  exactly.  The  costs under  the
averaging scenarios are compared  to a baseline in which averaging is not permit-
ted, and substantial  savings  are shown (see Section  V-A).

     Total  emissions under the  no-averaging  or  trading  baseline modeled  were
actually lower than the standards  require (for technological  reasons discussed

-------
                                      -42-

below).  Thus, under the averaging program  we  forego some emissions reductions
that would have  been  forced upon manufacturers  by  anticipated trap efficiency
in combination- with traditional  "every engine must  pass" regulatory approach.
These foregone emissions  reductions  are not a matter of  concern  because they
are not  needed  to  meet  the  emissions  standards  established  by  EPA.   It  is
interesting to estimate,  however,  the  degree  to which  the  cost  savings from
averaging and trading  are  attributable  to  allowing  the  manufacturers  to avoid
over-controlling emissions  versus  the degree to  which  the savings  are due to
more efficient allocations of the same emissions reductions.  This estimate can
be made  by  comparing  the  costs  under  averaging  and trading  to  an artificial
baseline in which over-control  is avoided with no resort to averaging or trading
between different engine types.

     The reason that over-control of  PM emissions  can be expected if averaging
jjs_ not permitted is that:

     o  based on  the  analysis  in  the  RIA,  the  standards have been set  at  a
        level sufficiently stringent that  every engine would be forced to use a
        trap;

     o  traps are expected  to  be  so efficient that almost every engine using a
        trap will have emissions  well below the 0.22 g/BHP-hr level (standard
        with design cushion); and

     o  there are no  inherent  cost  advantages to designing traps that are less
        efficient.

In quantitative terms,  engine-out  PM emissions  are  typically in the  range  of
0.6 g/BHP-hr  (at  low  NOx  emissions levels).   The  design target level  unr'ap  a
standard of  0.25  g/BHP-hr will  be  about  0.22  g/BHP-hr.  This  level  probably
cannot be reached  without  a  trap,  and so  in the  absence of  averaging  every
engine would  need a trap.   With a trap, which  can  be expected to remove 80% of
the engine-out emissions,  emissions  would  be  only  about 0.12  g/BHP-hr,  well
below the target.   It  is  the  on/off, either/or  nature of the trap technology

-------
                                      -43-

that would force  manufacturers  to  over-control  their engines.   Manufacturers
could try  to  design  less  efficient  traps,  but  there  appears  to  be  little
economic incentive to do so.

     For purposes of comparison, however, we are free to postulate an  artificial
baseline in which each engine has  "part of a trap:"  the costs and effectiveness
of the trap are  reduced  in  proportion to the need  to  meet  the  target  exactly.
Costs under this  artificial  baseline  are  lower  than under  the  actual  baseline
because not every engine must pay  for a "whole"  trap, and emissions are exactly
at their target levels.  The  costs are  not  as low as under  averaging,  however,
because the distribution of traps  is not optimized across different engines with
different costs of  emissions  reduction.   The cost  advantages of  averaging can
therefore be split  into its  two  components:  its ability to  let  manufacturers
avoid over-control, and  its  ability  to let  manufacturers allocate traps  opti-
mally across engine types without  allowing emissions to rise.

     The results for the one, two, and four  subclass cases using the  artificial
baseline instead  of  the  actual  baseline are  presented in Exhibits  VI-1,  VI-2
and VI-3.  While  the  dollar  savings from trading over averaging are  the  same
as presented in the  body of  this  report, the  savings  from  averaging  over the
artificial baseline are  significantly  lower than  when  the  costs  for averaging
were compared  to  the actual  baseline.   The reason  that averaging  looks  less
attractive, of course,  is  that  the  artificial  baseline  has  the  savings  from
reducing the number of traps  already built into  it.

     One way to  see this comparison  is  to  examine  Exhibit VI-4, which  shows
four industry-wide  total  costs  for  each  subclass  assumption:  (1)  the  actual
no-averaging baseline,  (2) the  artificial  baseline, (3)  the  costs  with  aver-
aging, and  (4)  the costs with  tradTng.  In  the  body  of the report,  (3)  was
compared directly to  (1), and referred to as the savings from  averaging.   The
savings from trading, which  is the difference between (4) and (3), seemed small
by comparison.  As  the table makes  clear,  though,  the difference between (3)
and (1)  is really composed of  two components. The  first  component is  the  dif-
ference between (2) and  (1),  which  is  the savings attributable  to avoiding the

-------
                                                  Exhibit VI-1

                                        COST SAVINGS  ARTIFICIAL BASELINE
                                        ONE CLASS - AVbRAGINC AND TRADING
                                              (dollars In millions)
Firm
                          Artificial                          Dollar and Percentage Savings from
                               Baseline                           Trading            Averaging
                          No  Averaging   Averaging   Trading  (vs Averaging)    (v» No Averaging)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
11
IB
19
20

Bluebird
Chrysler
Ford
Cummins
Catcrpl 1 lar
Daimler-Benz
KHD
General Motors
Navistar
Hlno Motors
Deere
Mack
Onan
Perkins
Renault
Saab
IlllZII
Iveco
Volvo
Willie

SO
3
38
335
12)
7
1
1 78
71
3
0
81'
0
0
11
1
111
5
;
/

0
9
S
i
2
7
2
s
3
3
a
i
3
a
7
0
•>
2
4
J

$0
3
33
329
124
7
1
131
S9
3
0
79
0
0
11
1
9
S
7
/

0
9
9
5
1
6
1
4
6
2
a
8
]
8
4
0
•J
2
J
1

<$n
0
2
302
114
6
1
113
36
3
0
n
0
0
10
0
V
4
6
o

0)
a
0
4
0
9
1
2
1
0
7
2
2
7
4
9
)
a
7
t
($)
SO
4
32
27
10
0
0
18
3
0
0
6
0.
0
1
0
0
0
0
C)

0
7
0
1
1
7
1
2
i
3
1
6
1
1
1
1
6
4
7
7
(I
128
1 >.}
94
a
8
9
8
13
S
8
10
8
4/
10
9
10
b
7
9
•*
)
8
It
2
2
1
6
i
a
a
i
0
3
4
0
3
0
1
9
2
6
(S)
$0
0
S
6
1
0
0
47
11
0
0
0
U
0
0
0
1
0
1)
U

0
0
0
0
1
0
0
1
a
0
0
i
0
0
2
0
0
0
0
J
(I)
u a
U U
12 8
i a
0 9
0 4
0 b
26 4
Ib S
0 8
0 2
0 b
0 1
0 2
1 8
0 2
a a
u 8
0 4
. - "
        INDUSTRY TOTAL
                                  $1141  I)
                                             SB IB  1
                                                       >/io •)    $10;  i
                                                                           n  i
                                                                                   i/.' 'I
                                                                                              a ?

-------
                                   Exhibit VI-2

                         COST SAVINGS  ARTIFICIAL BASELINE
                       TUO SUBCLASSES - AVERAGING AND TRADING
                               (dollars In million!)


c
c
c




c
















D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
II
(I
Firm

1 Bluebird
2 Chrysler
3 Ford
4 Cumins
% Tat., (.III. r
6 Dalinlet Beni
7 KHD
B General Motors
9 Navistar
10 Hlno Motors
11 Deere
12 Mack
1 3 Onan
14 Perkins
15 Renault
16 Saab
17 Isuzu
18 Ivaco
19 Vulvu
10 White
Artificial
Basel Ine
No Averaging Averaging

SO
3
38
335
12%
/
1
178
71
1
U
BO
0
0
11
1
10
5
/
•

0
9
9
5
1
/
2
5
3
3
a
3
3
a
7
0
9
2
4
1

so
3
38
329
1.-4
/
1
159
59
3
0
79
0
0
11
1
9
5
7
'

0
9
8
5
1
b
1
9
6
2
a
8
3
a
4
0
9
2
3
1
Hollar and ?<=i.
Ti adlng
Trading (v» Averaging)

SO
3
36
329
l.'l
/
1
159
54
3
0
79
0
0
II
1
9
5
}
>

0
9
0
4
V
b
1
2
B
2
0
7
1
8
4
0
9
1
1
U
(S)
SO
0
2
0
II
II
U
0
4
0
II
0
u
0
0
0
-0
0
o
II

0
0
'
2
1
U
u
6
'
u
II
0
I
0
0
0
0
1
II
u

0
•>
1
3
5
1
II
'
*

$11
0
0
6
1
U
u
18
11
0
0
U
0
0
0
0
1
0
u
1)

u
0
1
0
1
u
u
6
a
0
0
">
.0
0
1
0
u
0
u
i
(*>
o a
U 0
0 4
1 8
u t
a 4
U b
10 4
lb 5
o a
U J
0 b
0 1
0 2
i a
0 2
8 8
0 8
II •>
1 M
                                                                                                                     -p.
                                                                                                                     01
                                                                                                                     I
TOTAI
                            $851  4
                                      •>H42 4
                                                 i'l 0

-------
                              Exhibit VI-3




                    COST SAVINGS. ARTIFICIAL BASELINE

                 FOUR SUBCLASSES - AVERAGING AND TRADING

                          (dollars In millions)


G
C
C




C
















L
L M
M
M
H
L H
L H
H
M

L
H
L H
H
L H
H
M
L H


1
2
3
H 4
H 5
H 6
H 7
H 8
9
10
11
H 12
13
14
15
16
17
ia
H 19
20
Firm

Bluebird
Chrysler
Ford
Cuomlns
Caterpillar
Daimler-Benz
KHD
General Motors
Navistar
Ulno Motori
Deere
Hack
Onan
Perkins
Renault
Saab
Isuzu
Iveco
Volvo
UliUe
Artificial
Basel Ine
No Averaging Averaging

$0
3
38
335
125
7
1
178
71
3
0
80
0
0
11
1.
10
i
7
7

0
9
9
5
2
7
2
5
3
3
8
3
3
a
7
0
9
2
4
3

0 0
3 9
38 8
333 7
124. i
7 6
1 2
177.8
71 0
3 2
0 8
79 9
0 3
0 8
11 6
1 0
10 7
5 2
7 *
7 3
Dollar and Percentage Savings from
Trading Averaging
Trading (vs Averaging) (vs No Averaging)

0 0
3 9
36 S
330 7
12* 4
7 6
1 2
176 S
69 9
3 2
0 8
79 9
0 3
0.8
11.6
1 0
10 7
5 2
7 *
7 1
(S)
$0 0
0 0
2 2
3 0
0 0
0 0
0 0
1 3
1 1
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 1
(X)
0 0
0 0
5 7
0 9
0 0
0 0
0 0
0 7
1 6
0 0
0 0
0 0
5 0
0 0
0.1
0 0
0.0
0 1
0 0
2 0
(S)
$0 0
0 0
0 1
i a
0 7
0 0
0 0
0 7
0 3
0 0
0 0
0.4
0 0
0 0
0 0
0 0
0 2
0.0
0 0
0 0
(X)
0 0
o o-
0 4
0 5
0 6
0 4
0 4
0 4
0 4
o a
0 2
0 5
0 1
0 2
0 2
0 2
2 1
0 8
0 4
0 3
                                                                                                                           -p.
                                                                                                                           Ol
                                                                                                                           I
INDUSTRY TOTAL
                         $891 0
$886 5    $878.7
$7 8
                                                                   0 9
                                                                           $4  5
                                                                                      0  5

-------
ONE CLASS
TWO SUBCLASSES
                                      -47-

                                  Exhibit VI-4
                       COMPARISON  OF  INDUSTRY TOTAL  COSTS
                             (dollars  in mill ions)
                  (1)
                  Baseline
                (2)
                Artificial
100% Traps
No Averaging
$ 1,009.6
1

1
Baseline
No Averaging
$ 891.0

(l)-(3) = $ 191.5
1 1
(3)
Averaging
$ 818.1
1 1
(3)-(4
1
(4)
Trading
$ 710.9
1
) = $ 107.2

                         (l)-(2)  = $  118.6      (2)-(3)  = $ 72.9
$ 1,009.6
$ 891.0
$ 851.4
$ 842.4
                                  (l)-(3)  = $  158.2
                                 	I    I
                                         (3)-(4)  = $ 9.0
                                       I
                         (l)-(2)  = $ 118.6      (2)-(3)  = $  39.6
FOUR SUBCLASSES
$ 1,009.6
  I	
$ 891.0
$ 886.5
$ S78.7
                                  (l)-(3)  = $ 123.1
                                 	I    I
                                         (3)-(4)  = $  7.8
                         (l)-(2)  = $ 118.6      (2)-(3)  = $  4.5

-------
                                      -48-

over-control  of PM.   The second component is the difference between (3) and  (2),
which is the  savings attributable to the efficiency of allowing  firms to  real-
locate emissions controls optimally across engine lines.  This second component,
while appreciable, is much closer to the magnitude of the  savings  from trading
(the difference between (4)  and  (3)).

-------
                                      -49-
B.   Reclassification of Light-Heavy-Duty Engines
     Manufacturers of some LHDEs--those for trucks in Class II-B--may choose to
reclassify their engines as light-duty truck (LOT) engines to take advantage of
test procedures that are less  technically  difficult or because of the availabil-
ity of off-the-shelf production technology.  This  reclassification could have a
significant impact on averaging and trading programs.

     The extent to which  this  reclassification  would take place  is  not known;
we have therefore re-estimated costs under averaging  and trading for two widely
differing cases:  one  in  which  all Class  II-B  trucks  drop  out  of  heavy-duty
averaging and trading, and one in which only half  drop out.
Estimation of Class II-B Sales and Shares

     Engines for Class  II-B  trucks  do not comprise  all  of an engine  class  in
this analysis.  Instead, as defined, they represent a subset of each one of the
three light-heavy sub-classes: LHDGE, LHDDE-IDI, and  LHDDE-OI.   The  balance  of
each of these three sub-classes  is  composed  of Class III and  Class  IV trucks.
Thus, to estimate  how many  heavy-duty trucks would remain  in  these  classes  if
100% of Class  II-B trucks  were  reclassified  as  light-duty  trucks, it  is  suf-
ficient to  use  the total  sales  and  sales  shares' for  Class  Ill's and  IV's.

     Based on figures published  in Automotive  News, we  estimated that  sales
of Ill's and IV's combined are in the range  of  30,000 per year,  with virtually
all of  those  sales by  GM.   The  distribution of these  sales  across the  three
light-heavy classes (LHDGE,  LHDDE-IDI,  and  LHDDE-DI)  is made  simpler by  the
fact that GM does not  produce LHDDE-DIs;  thus, all  30,000 should  be placed into
either LHDGE or  LHDDE-IDI.   We have assumed  that  this division  is the  same'as
the industry-wide division between gas and diesel  for this size class,  yielding
20,904 LHDGEs and 9,096  LHDDE-IDIs.

-------
                                      -50-
     Once the sales and  sales  shares  for  the 100% reclassification  case  have
been set, the .determination of  the sales and shares for the 50% case is simple:
sales and sales shares  for  the  50% case are unweighted averages of the sales and
shares for  the   "0%  reclassification"   (baseline)  and  100%  reclassification
cases.  A comparison  of the sales  and  shares  for each of  these  three  cases
are presented in  Exhibit VI-5.

     Results for  this  sensitivity  case are presented  in Exhibits  VI-6,  VI-7,
VI-8, and VI-9.  The exhibits  show  that,  for  the industry as  a  whole,  reclas-
sification would  have  only a moderate impact  on  the  savings  from  trading and
averaging in the  one-class case.  Savings  fall  by about 15% for  averaging and
for trading if half of II-Bs are reclassified, and by another  15% or so if the
other half are reclassified as  well.

     Under two-subclass averaging  assumptions, the changes caused by reclassifi-
cation are similar to  the  one-class case.   The  savings  from trading,  however,
are cut  dramatically  by the  reclassification  of half  of the  II-Bs  and  are
virtually eliminated if all II-Bs are reclassified.   Trading  produces  greater
savings when the  averageable set  includes  engine  lines  and engine technologies
that are very dissimilar; apparently, subclass  restrictions in  conjuction  with
the reclassification of II-Bs  would leave  very  little  cost-savings  from heavy-
duty truck trading.  (Of  course,  such reclassification might make  trading for
light-duty trucks  very  appealing;  an examination  of  this  issue is  beyond the
scope of this analysis.)  A summary table  of  industry-wide comparisons  across
scenarios is provided as Exhibit VI-10.

-------
                                                                  VI 5
                                      SfiiSHU SJHU.S RH MM AjanrwuH «•• iJiiir uwr unv
SALES SIMS (peuMt)
df«r- b
Ita^ ^i^* ftiiM*ii«i fiyr^i^- Rnd Qualm pillar
UU£
iHre-nn
UIIE-OI
HIDE
MIIEtM
HUE-IK
IIIIEIH
IIIIEIIH
211.668
132.811
44.278
169.112
29.818
51.114
77,405
U.7U1
0
0
0
0007
0
0
0
0
161
0
0
0
0
0
0
0
97
0
400
595
0
0
0
0
0
0
550
0
0
1 6
Ml
0
0
0
0
To 1
U.8
1J 7
13 7
ililJer-
ItaK
0
0
0
0
1 0
01
0 1
Calftldl
HI) KJUJII HfA&*r
0
0
0
0
0
02
02
02
/4 2
51 4
0
405
>> 5
a 5
11 4
II 4
U
456
0
0
17 7
0
0
IllIU
0
0
0
0
i 7
08
0
0
l^j-n* M ^ ^ Qkjf i Iliiclrui l^i "^'1* S^^> T **ft'
0
0
G
0
0
06
0
0
0
0
0
0
0
0
150
150
0
06
0
0
0
0
0
0
0
0
0
0
0
0 6
0
0
0
04
0
0
0
9 1
0
0
0
0
0
0
0
08
0
0
0
1 9
2 5
0
1 4
6 1
0
0
Ivtin
0
0
0
0
4 5
1 0
0
0
VUvo Mill*
0
0
0 '
0
0
1 4
06
06
0
0
2 5
0
0
52
0
0
755.JH1
        BlufcUd Uynltf   fad Orafca
                                                        Hi)
                                                                   llivld.ii
                                                                             Hun
                                                                                                  Oui Itifcuu HUM.IL
                                                                                                                              laui
                                                                                                                                            VUvo   Uiiic
UHE 1)6. a*>
MIIE-m 70.9h5
IJIIKOI 22.119
HUE 1W.11>
HIIEtM M.HId
HIIE-n: 51. 114
IIIIKIH 77.44)5
HllLMh Id. All
0
0
0
0007
0
0
0
0
14 6
0
0
0
0
0
0
0
89
0
400
595
0
0
0
0
0
0
550
0
0
1 6

M 1
0
0
0
0
Jh 1
If. 8
117
LI 7
0
0
0
0
1 0
52
0 1
0 1
0
0
0
n
0
02
1)2
(12
to 5
546
0
Ul 5
•» 5
21 5
II 4
11 4
0
4J 6
n
0
U 7

0
0
0
0
0
0
.! 7
On
0
0
I)
0
0
0
0
06
0
0
0
U
0
0
0
0
150
15 U
U
06
0
0
0
0
0
0
0
0
0
0
0
06
0
0
0
0 4
0
0
0
9 1
0
0
u
0
0
0
u
08
0
u
0
1 8
2 %
0
1 4
6 )
0
0
0
0
0
0
4 5
1 0
0
0
0
0
0
0
0
14
06
06
0
0
2 ^
0
0
52
0
0
5/5,972
                                                                                                                                                                     I
                                                                                                                                                                    01
1
Typis* ^M^t fri»«J4gj Oyrilcf Fixd QJIU^JD i
HUE
uiiK-on
HUE oi
HUE
HIIEtM
MIIE-V:
IIIIEIH
IIUEMH
20.414
0
2X818
51,304
77.405
0
0
0
oou;
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
545
0
0
0
0
0
0
0
0
0
1 61
Ml
C^ittr* QuuJer*
I'l 1 1  1
Ie8
11 7
13 7
0
0
0
U
1 0
52
01
0 1
1
HI)
0
0
I)
0
0
02
0 2
02
"feSl N
HDO
luuo
0
Ul •)
m 5
21 5
11 4
11 4
Mtxjr
0,
0
n
0
11 7
0
0
lluu
0
0
0
0
2 T
08
0
0
Ufa:
0
0
0
0
0
06
0
0
fet Ci*. fed*. h.~Ji S-b U.. N~,
0
0
0
0
0
0
150
150
0
0
n
0
0
0
0
0
0
0
0
0
06
0
0
0
n
u
n
u
9 1
0
0
0
0
0
0
0
08
0
0
0
0
0
0
1 4
6 1
0
0
0
0
0
0
4 5
i a
0
0
VUvo
0
0
0
0
0
1 4
06
06
Uau
0
0
0
0
0
5 2
0
0
UI1E
IIIIE-mi L
IIIIEW
Milt    Hilbiulk^yllfyr.uJbi: UH.II>
HIII.IM  Hilbmll>vy(lly lluueJ hbut: - li
MIIK1L  MdliiiilL^yrily blue] beu« - 1\
IIIU <*: • N
                  Built: - Ldimx L\k»U»
                               . IUJ

-------
                                           EahlLlt VI-6
                                           COST SAVINGS
                        ONE CLASS  IIAIF OF Cl ASS II-b KLiIl AiilKI tlJ AS LOT
                                       (dul lars In mil Hunt)
Finn
                           Baseline                            Dollar and Percentage  Saving*  IIOID
                             1001  traps                           Trading             Averaging
                           No Averaging   Averaging   Trading  
6
3
6

SO 01
2
28
322
124
/
1
119
51
3
0
79
0
0
11
1
9
b
7
6

D3
0
8
/
1
b
1
1
8
2
8
a
1
3
4
0
S
2
3
8

o
3
U
74
U
0
III
0
a
4
b
b

0)
H
b
3
2
1
1
0
8
0
7
;
1
7
i
9
9
8
8
I

SO
1
3d
20
7
0
U
14
1
0
u
J
0
0
0
0
u
0
0
0

0
8
b
S
9
b
1
1
0
2
1
1
1
1
9
1
6
3
S
b
(I
133
143
127
b
b
1
b
11
2
b
8
6
49
8
;
8
b
»
7
;
»
/
2
2
3
4
7
b
8
0
7
0
4
9
0
7
0
U
n
3
9
(S)
SO
0
3
iO
li
0
U
&9
17
0
0
11
0
0
1
0
1
0
0
0

0
u
b
0
3
8
>
1
4
3
1
8
1
1
4
1
/
4
9
9
(I)
1 1
U U
11 i
13 4
11 0
9 /
\J 3
a 4
2i 2
7 4
<* 9
1.' 8
29 3
9 9
1U 9
ID 0
n i
7 2
11 2
11 4
                                                                                                                                      ro
                                                                                                                                      i
        IHUUSTKY TOTAI
                                  $948 0
                                             $'83 1
                                                       $t>91 ->
                                                                 $92 0
                                                                           11 8   $164  7
                                                                                              17  4

-------
                                         Exhibit VI-7
                                        COST  SAVINGS
                         ONE CLASS  ALL CLASS II-B RLCI.AbSIHLD AS IDT
                                     (dollars In million*)
h Irm
                           Baseline                             Dollar and Percentage Savings  iioiu
                             10U1  irarn                            Trading            Averaging
                           No Averaging    Averaging   Tiadlng  (vs Averaging)    (vs Nu Averaging)
1 Bluebird
2 Cluy>lcr
3 toed .
4 Cuamlni
S C.. 1-rplllar
6 Daimler-Benz
7 KHD
8 General Motors
9 NavLstur
10 Hlno Motors
11 Deere
12 Ma<.k
13 On jr.
14 Perkins
15 Renault
16 Saab
17 I»uzu
IB Iveco
19 Volvo
20 Uhl la
SO 003

21
362
13V
b
1
Ibb
bi
3
0
VI

0
12
1
10
5
H
/
•
B
i
4
i
3
;
6
b
9
b
m
9
;
i
i
6
i
j
$0 003

_>j
Jib
124
7
1
10U
48
3
0
/a

0
11
i
9
5
7
u
A
8
U
1
6
1
4
8
2
8
U
*
a
4
0
1
2
3
•>
($0 00.')

-19
302
lilt
;
i
9*
46
3
0
lo

0
10
0
8
4
7
b
•
U
2
i
2
I
6
1
1
B
3
*
B
7
9
6
9
0
I
IS)
SO o
•
42 H
U 8
& 6
0 4
II 1
8 8
•i 7
0 2
0 0
3 b
•
0 0
0 7
0 1
0 &
0.3
0 4
U 4
m
179 t

1/9
4
4
&
4
8
b
i
i
4

i
6
b
b
i
S
u
•
8
4
6
7
6
1
6
1
9
4
•
9
0
9
8
1
2
u
(S)
$0 0
•
0 0
46 b
Ib 3
0 8
0 2
47 )
4 8
0 3
0 1
11 8
•
0 1
1 t
0 1
1 0
0 4
0 4
U /
(I)
1 1
•
U U
12 8
11 U
9 7
12 3
10 4
8 9
7 4
9 9
1.' B
•
9 9
9 9
9 1
9 b
7 2
11 2
•> -i
                                                                                                                                     I
                                                                                                                                     in
        INMISTHY  TlllAI
                                                       $074  7
                                                                 $rin
                                                                           III  6    SI U  4
                                                                                             u a
 All

-------
Lulllbll VI -H
COST SAVINGS
TWO SUBCLASStS HALF OF CLASS) II-B KLCLASblFlKI)
(dollars In iiillllons)

C
C
C




C















D
D
0
D
0
0
D
D
0
D
0
0
0
D
U
D
D
1)

1
2
1
4
5
6
7
8
9
10
11
12
11
14
15
16
11
IB
19
70
Firm

Blueblid
Chrysler
Ford
Cumnlns
Caterpl 1 lac
Dalmler-Bcuz
KIID
Ceaeral Motors
Navistar
Hlno Motors
be ere
Mack
Onan
Perkins
Renault
Saab
Isuzu
Iveco
Volvo
Whir.
Basel Ilia
100Z craps
No Averaging Averaging


$0 001
2
12
372
139
8
1
i;a
09
3
0
91
0
0
12
1
11
4
a
7
0
4
7
4
J
1
a
i
4
9
4
I
9
a
i
i
6
3
o

SO
i.
31
322
124
7
1
. J
41
3
0
79
0
0
11
1
9
4
7
f>

0
u
i
'
i
6
1
4
a
.»
8
8
1
8
4
0
4
2
1
a
AS LOT
Trading
Tiadlng (vs Averaging)

so
-'
11
122
124
7
1
I4B
51
1
0
79
0
0
11
1
9
i
7
li

0
0
9
'
0
6
1
I
6
2
a
8
0
a
4
0
b
1
i
a
(S)
SO
0
1
0
u
0
0
u
0
0
0
0
0
0
0
n
0
0
0
0

0
0
4
0
1
0
0
2
1
0
0
0
1
0
0
0
0
0
0
0
<»:
i
u
4
0
0
0
0
IJ
0
0
0
0
73
0
0
0
U
0
0
0

Av«rag Ing
(vs No Averaging)
1
1
0
4
G
1
1
0
•
*
7
i
0
1
4
2
2
0
a
0
2
(S)
SO
0
1
50
15
0
0
10
i;
0
0
11
0
0
1
0
1
0
0
0

0
0
I
0
1
8
2
4
4
1
1
a
i
i
4
1
'
4
9
9
<.,
0 U
o o
J 5
U 4
11 0
9 7
U> 1
I/ 0
25 2
7 4
» 7
12 8
29 6
9 7
10 9
10 0
14 2
/ 2
11 i
11 4
                                                                                                                              I
                                                                                                                             en
TOTAL
                              SMI*)
                                         r.si; a
                                                                u i
                                                                        I il  0
                                                                                   14  I)

-------
                                                      VI  9


                                             COST SAVINGS


                          TWO SUBCLASSES  ALL CLASS  II-B KECIASS1HEI) AS I Lit


                                         (dollars  In milllona)

c
c
C D
D
D
0
D
G D
D
D
D
D
0
U
D
D
D
D
D
D

1
2
3
4
S
6
7
8
9
10
11
12
11
14
li
16
17
18
19
20
E* 1 fat
9 i rio
Bluebird
Chryiler
Ford
Cunmliii
Caterpillar
Daimler- Ben*
KHD
General Molori
Navlitar
Hi no Hoion
Deere •
Mack
Oixn
Peikln*
Kenauli
Saab
Iluzu
Iveco
Volvo
Uhlte
B^ >e 11 ne
lUUt ir^pi

$0 003

21
162
119
B
1
Ibb
iJ
3
0
91

0
:2
1
10
b
a
/
•
8
b
4
b
1
7
6
5
9
b
•
9
7
1
1
6
1
^


$0 001

23
116
124
7
1
137
<.rf
3
0
79

0
11
1
9
S
7
b
•
B
0
1
b
1
b
8
2
8
8
•
B
4
0
1
2
3
5
Dollar *nJ P'l-cenrane Savings from
Hading Av«r»»ln»

$0 001

23
31b
124
/
I
HI
48
3
0
79

0
11
1
9
S
7
6
*
B
9
1
6
1
b
7
2
B
B
•
B
4
0
1
1
3
b

($)
S"

0
11
0
u
u
u
0
0
u
a

o
0
0
0
0
0
u

0
•
u
i
0
0
U
u
i
0
0
0
•
0
0
0
0
1
0
u
u>
u u
•
0 0
u o
0 0
0 U
0 0
U 0
U i
U 0
U 0
0 U
•
II U
u u
0 0
0 0
1 1
0 0
u u
(S)
SO

0
46
li
0
0
IB
4
0
0
11

0
1
0
1
0
0
u
u
•
0
i
]
8
2
1
8
3
1
B
•
I
i
1
0
4
9
;
0)
0 U
•
0 0
1.' B
11 0
9 ;
U J
11 6
B 9
7 b
9 7
\1 9
A
9 7
10 o
9 9
9 b
7 2
11 1
10 0
                                                                                                                                         I
                                                                                                                                         tn
                                                                                                                                         LTI
                  TOTAL
                                  SBob  1
                                              $7b4
                                                        5/ttl 8
                                                                    SO  1
                                                                              0 U   $10.' 1
                                                                                                11
* All en*} 1 lie &  i e< ,
                    »i I t led

-------
                                                        Exhibit VI-10



                                SUMMARY OF JNDUSTKY COSTS FOK LHUt KECLASSIHLATION iCLNAHIOS

                                                     (dul l«i* til nil II tuns)
                                                                    Duller *nJ Percentage Saving! I roin

                                 1001  tr*|>i                            Tiadlng            Averaging

                               No Averaging    Averaging    Tiadlng   (v» Averaging)     (v* Nu Averaging)



                                                                        LDT        $1.009 6     $bil 4    $812 t     $« 0      11   SI48 2     1 b  7



Half II-Bi rccUsiltled o LUT         948 0      bib 0     81 i 8      22      03    Til I)     U  (I


All II-Bi rcclanlflcd ^s LDT          80b 3      784 1     781 8      03     0 04    102 3     lib

-------
                                       -57-

 C.    Sensitivity of Results to Assumed Functional Relationships
      All  of the  results  of this report are driven by the underlying  functional
 relationships among  NOx  and PM  emissions  and  fuel  consumption.   While  the
 functions used  were developed  after  careful research  into  current and  likely
 future  technologies, there  is no guarantee that the  true relationships will  not
 turn  out  to be  somewhat  different.  There could be  unexpected  breakthroughs in
 emissions control  techniques,  especially given the  incentives  provided  by  the
 averaging and  trading  programs to  produce  engines  capable  of lower emissions
 than  those mandated  by the  standards.   On the  other hand, it  is  possible that
 projections for  some types of engines could turn  out to be overly optimistic.

      To investigate  the  sensitivity of the  results  of the analysis to moderate
 shifts  in the  functional  relationships,  the  potential gains from  averaging  and
 trading were  re-estimated  with altered  functions  for one  subclass.  A  scaling
 parameter was altered  in  the equations relating NOx to fuel consumption  and PM
 emissions for all MHDDEs, increasing the  fuel  consumption impact and the  engine-
 out levels of PM associated  with  a given NOx target  by  20%.   This change  was
 arbitrary in the  sense that  there  is no  more reason  to expect these function to
 be off  the true mark in one direction than the other.  The shift  in  the functions
 does  increase the difference between MHDDEs  and other classes, which could  be
 expected  to  make  averaging and trading look more  attractive.  Averaging  and
 trading could be expected to appear less attractive if the functions had been
.shifted in the opposite direction.

      The  results, Exhibit  VI-11,  show that  trading  would  indeed  provide more
 cost  savings  with the  shifted functions, relative  to the  savings  from  trading
 with  the   unsTiifted  functions and  relative  to the  savings  from averaging    In
 the two subclass case, trading would  save an estimated $14.2 million per year,
 58% more   than  the $9.0  million   in  savings from  trading estimated  with  the
 unshifted functions.   Averaging would also  save more if the shifted functions
 were  correct,  by a  margin  of  $173.1  million  per  year to  $158.2 million  per
 year.  This gain represents an increase  of  9.4%,  significantly lower than  for
 trading.

-------
                           txl.lLU  Vl-ll

                           COST SAVINGS

        TWO SUUCI-AiSLb  201 INCREASE  IN MIHUit COST JUNCTION

                       (dul lar»  In  mil I lon»)


c
c
C D
D
D
D
D
C D
D
D
D
D
D
D
D
D
D
D
D
D


1
2
3
4
i
6
7
B
9
10
11
12
13
14
1)
16
17
IB
19
20
Firm

Bluebird
Chrysler
Ford .
Cummins
Cai.. |>ll lar
Daimler-Benz
KHD
Ceneral Muloi*
Navistar
Hlno Minor*
Deere
Hack
Onan
Perkins
Renault
Saab
Isuiu
Ivcco
Volvo
White
Basel tne
100X traps
No Averaging Averaging


JO 00 J
3
41
387
IB;
22
1
262
lb/
b
2
91
0
2
36
3
2a
a
17
ji
9
1
1
9
1
a
a
j
9
6
b
4
b
b,
1
a
9
0
b


$0 003
3
38
333
169
21
1
21b
126
b
2
19
0
2
Jb
1
26
a
lb
20
9
7
4
7
2
b
4
4
6
b
a
3
b
0
0
1
b
7
4
Dollar and Percentage Savings from
Trading Averaging
Trading (vs Averaging) (v» No Averaging)


Su nui
3
36
333
169
20
1
214
121
b
2
79
U
2
31
2
2b
8
lb
19
9
0
2
2
4
6
a
4
4
4
7
1
4
S
8
4
i
b
6
<;>
SO
0
2
0
0
0
O
u
b
0
u
0
u
0
1
0
0
0
0
0

0
0
7
2
s
a
0
6
0
2
1
1
2
1
b
2
7
3
2
a
(i
\
0
7
0
0
i
0
0
4
'
4
0
66
4
4
6
2
*
1
J
)
1
0
U
1
1
a
0
j
0
•j
0
1
7
0
3
7
7
S
J
»
.*.
$0
0
2
b3
ia
0
0
4'
30
0
0
11
u
0
1
0
2
0
1
1

0
0
4
7
2
9
2
4
9
3
1
7
1
1
b
1
'
4
3
1
(It
U U
0 0
b 1
13 9
9 ;
4 1
11 ;
Id U
IV D
4 6
2 6
12 a
il 0
2 7
4 1
2 7
9 b
4 1
/ b
b U
                                                                                                                      I
                                                                                                                     LD
                                                                                                                     CO
                                                                                                                      I
TOTAL
                I.JB-' 8   SI.IO'J '
                                                 S>4  2
                                                            1  )   $1/1 1
                                                                              I J  b

-------
                                      -59-
D.   Banking
     In this section  some aspects of  emissions banking  are discussed, and  a
rough estimate of  the  potential  savings  possible using  a  bank  is  offered.
The Concept of an Emissions Bank

     Under emissions banking, a  firm  that  generates emissions credits  by  pro-
ducing engines that  more  than  meet  emissions  standards  is  permitted  to  hold
onto them--putting them  in a  "bank"--and  use them in  later  years to  offset
emissions that exceed the standards.   In effect,  an emissions  bank extends the
concept of averaging  and trading over time, enlarging the averagable set to  more
than one model year.

     Savings from banking  could be  large under certain  circumstances,  without
compromising air quality.   Concern  is sometimes  expressed,  however,  that the
use of  banking  could  lead  to  emissions increases, undesirable  variations  in
emissions over time, or to  disruptive  attempts to manipulate market  shares  by
radical year-to-year changes  in emissions  characteristics.  To  understand the
real effects  of  banking,  it  is important  to  lay  out  the circumstances under
which emissions banks would be attractive.
Circumstances in which Emissions
Banking Would be Attractive to Manufacturers
     The circumstances favoring the use of an  emissions bank are:

     1) if the banking rules allowed manufacturers  to dispense with the "design
        cushion"-- the gap between the  emissions standard  and the  self-imposed
        emissions target set so as  to  ensure  compliance in  use.  In  this  case,

-------
                                      -60-

        after the  first year  of  a  banking  regime,  banked  credits  might  take  the
        place of th'e  design  cushion,  allowing manufacturers to aim directly at
        the standard.  This  would  be less costly  to  manufacturers over  time,
        but would  also  result in greater emissions;

     2) if banking were allowed  as  a  remedy for  in-use  noncompliance;

     3) if the real  value to consumers  of vehicles with higher  emissions were
        expected to rise  over time.   Vehicles  with  higher emissions tend  to  use
        less fuel, and if  fuel  costs rise  appreciably  (or  if  interest  rates
        fall appreciably, which  increases the  present value of  fuel costs) then
        consumers  will  be more concerned with fuel  consumption and will pay  a
        higher premium  for more  efficient engines.   Manufacturers could gain by
        generating credits  now and using them in  the  future, to  allow them to
        increase emissions  and  fuel  efficiency  when efficiency  is more  highly
        valued;

     4) if there  were  a  perceived marketing  advantage  in  allowing  an  entire
        model  year's  output  to exceed emissions  standards, even  at the cost of
        holding  an earlier year's  output below the  standards;

        and, probably most significantly,

     5) if  regulations are  expected  to be  tightened   in  the future.   Banking
        could then be used to make  the adjustment  to tighter  regulations  easier
        and less  costly.  Manufacturers  would  cut emissions  sooner   than  the
        regulations require,  and then ease  down to the  limits  just  after  the
        tighter  regulations   went  into effect.  T
-------
                                 -61-
1) Averaging,  and possibly trading, will  already be permitted before bank-
   ing goes into  effect.   These programs themselves are being structured to
   force, the retention of a design cushion.  A  banking  program could also
   be structured  in  this  way, by  limiting the  use  of banked  credits to
   adjusting the  PEL for an engine family, and  not  permitting  the  credits
   to be  used  to  make  up  for  the  problems   of  an  individual  engine.

2) Because  of  legal  provisions  and  enforcement  concerns,  banking as  a
   remedy for in-use violations  is  not  a  serious possibility;

3) While it is  possible that credit values will  rise over time, it is not
   likely that  manufacturers will be sure enough that they will  to make the
   investment  in  generating  credits  attractive.   Credit  values could be
   driven down  as  well  as  up, and so an  investment  in  credits would be
   quite risky  if undertaken  on  the  basis of expectations  regarding fuel
   costs or interest  rates.  Added  to the  fact that  no  interest  would
   accrue on  banked  credits, thus  decreasing  their  value  over  time,
   it would probably not  seem worthwhile  to  accumulate credits  for this
   reason.

4) The idea that  a  firm would try  to gain  sales  in  the  future  by spending
   banked credits is rendered  dubious by the fact that sales would probably
   be lost initially while the credits  were being generated.  In addition,
   the costs of this strategy would come  well  before the benefits,  meaning
   that the costs of the capital  invested in  the strategy would reduce the
   strategy's  net value.   Only  if  consumers behaved  quite  asymmetrically
   with regard  to increases and decreases in performance —valuing slightly
   improved performance  highly   while   being   relatively  indifferent  to
   slight deteriorations of performance—would  this  plan  seem  worthwhile.

5) This set of circumstances  is  the most  plausible,  since tightened emis-
   sions standards   have  been  promulgated.  Manufacturers   would  have  a
   strong incentive to generate  and bank  credits before the standards were
   tightened, and then use the credits in  the period  following the change

-------
                                      -62-

        in the regulations.  This shift in emissions would  not, however,  cause
        a worrisome  spike  (increase)  in  air  pollution.    Instead,  it  would
        contribute to a  smoother  reduction in emissions than  would  take  place
        without banking,  and  would  allow  society  to  enjoy  its  emissions
        reductions sooner than otherwise.

     In sum, the  only  likely scenario  for  the use of  emissions banks is  one
in which the  effects both  for  the manufacturers  and   society  as a  whole  are
beneficial.
Estimating the Savings Potential  of Banking

     It is not  possible  to estimate  the cost  savings  that would  be possible
from banking without  knowing  the rules  under  which banking could  take place,
and discovering the technological relationships among emissions  and  costs both
now and in the  future.   We have, nonetheless,  made an estimate  of  the savings
that could be  realized  under  idealized conditions  to   show  the potential  of
banking and demonstrate how a  more complete analysis would be conducted.  Costs
for banking with  one class averaging and  trading  have  been estimated.   Since
only PM credits would be banked and manufacturers  of gasoline engines would not
have PM credits, results for two  subclass averaging and  trading are  expected to
be about the same as the one class case.

      In this example, we have assumed no technological  progress; in  particular,
we do not assume that traps will  become more efficient  in  1994  than  we assume
they will  be in  1991.   Regulations  for  PM are to be tightened  from  a standard
of 0.25 g/BHP-hr (with a target  of 0.22) to a  standard  of 0.10 g/BHP-hr (and a
target of 0.088) in 1994.  Many  firms would  not be  able  to meet  this  stri-nent
standard, even with averaging, so we have assumed that a  non-conformance penalty
(NCP) would be charged to firms for each ton over  the PM target.^.X  The per-ton
     }_/  The charging of a non-conformance penalty is illustrative. No NCP devel-
opment work for the 1994 standards has taken place.

-------
                                      -63-

charge was assumed to be equal  to the Industry-wide marginal  cost of PM removal
calculated to  prevail  under  the  tight  1994  standards:   $13,234  per  ton.

     Allowing banking would permit a spreading-out of the emissions tightening.
For instance, under  banking there  would  be an  early (pre-1994)  reduction  of
emissions from 0.22  g/BHP-hr down to 0.154  g/BHP-hr  (the  average  of  0.22 and
0.088), a level  which  would  generate  credits.   There  would  then be  another
period of emissions  at  0.154 g/BHP-hr  after the regulations tightened,  until
the credits   were  used  up.   Using  one-class averaging  assumptions,  emissions
control costs under this  banking scheme  would  be lower by  about  $6.5 million
dollars per  year, or  0.7 percent of  total  emissions  control costs,  than  they
would be  without  banking (see  Exhibit  VI-12).   Banking  would  save more  under
one-class trading assumptions than under one-class averaging: $12.7 million per
year, or 1.6 percent of emissions control  costs.

     The total savings  from banking would depend  on  the number of years  over
which the inter-temporal  averaging  would  take  place: the  sooner  the  credits
started accumulating, and the longer they  were used after the regulatory change,
the more would be saved.   The  time period can be estimated  by  considering the
capital cost to the manufacturers  of generating  a credit and holding  it,  with
no interest  accruing, until  it  can be  used  to  reduce costs.   Considering the
high real cost of capital,  we  estimate that at most a year's worth of credits
would be banked, and  so the total savings from  banking  in these circumstances
would be relatively  small.

-------
                                                             Lxlilbll VI-12



                                                              COST SAVINGS

                                                   ONE CLASS  BANKING v» NO BANKING


                                                         (Jollars In ml 11 lout)
Firm

1 Bluebird
2 Chryiler
3 Ford
4 Cummin*
5 Caterpillar
6 Daimler-Benz
7 KHD
a General Motors
9 Navlsiar
10 Hlno Motors
11 Deere
12 Hack
13 Onan
1* Perkins
16 Renault
16 Saab
17 Isuzu
IB Iveco
19 Volvo
20 Ulilto
Averaging 	 	 	 	 i ran ing 	 	 	 	 	 ---
Dollar and Percentage Dollar and Percentage
Savings trum Savlng> from
Ulllioul Avciaglng Ullh B. liking Ulll.oiil Trading Ullli Banking
Hllh Banking Banking (va Ultliuul Banking) Wltli Banking B.nking (v* Ulcliout Banking)


SO 003
3
31
312
131
8
1
14)
70
i
U
Hi
0
U
12
1
10
4
7
7
9
8
1
7
1
2
b
0
4
U
3
4
9
1
U
8
«
«
5


$0 003
3
3)
3)3
131
a
i
149
7U
3
U
8i
U
0
12
1
11
i
/
/
9
a
4
9
1
2
2
1
4
9
3
4
9
3
0
1
b
b
;
(S)
SO 0
0 0
0 0
1 3
0 1
0 1
0 0
3 7
U i
a I
0 U
0 0
0 0
0 U
0 2
U 0
0 4
U 1
0 0
0 -'
(I)
U 0
0 0
0 0
0 4
0 1
1 2
0 0
2 b
0 4
.' 9
0 U
0 0
0 U
U 0
1 6
0 0
3 6
1 b
0 0
2 b


$0 001
-0
4
321
1J
-------
                                   Appendix A
                                  OPTIMIZATION
     This section describes the operation  of the  procedure  (model)  used  to  find
the lowest-cost set of emissions  levels under  regimes that allow  averaging  or
trading.  It is an adaptation  of  a  non-linear  programming  approach to optimiz-
ation.  It finds the  least-cost, optimal solution by  setting  all  derivatives  of
total emissions  control   cost   to  zero,   while   holding   emissions  constant.

     The computer program takes as inputs cost and technical data on the  engines
sold by firms in the industry (see Appendix B for  details) and  emissions  targets
set by the  regulations.   It provides as  outputs  the emissions levels   of  each
type of engine  that  meet  the  standards  at  the  lowest possible  cost.   Other
outputs are the costs per vehicle of emissions control, the total  costs of the
controls for each  class  of engine  and  across engines,  emissions  levels, the
costs per ton  of emissions reductions, and the  expenditures on emissions reduc-
tion credits made for each type of engine.

     Three conditions are  met   by the  minimum cost  (most efficient)  set  of
emissions levels.  First,  the  total emissions equal  the  total  emissions  per-
mitted.  This condition is necessary because it  is not permissible to  emit total
pollutant loadings at higher levels  than those  allowed by  standards, and it  is
not efficient   to  emit at   lower  levels   (since  emissions  reductions  are  not
without cost).  Second, the  application of traps  has  been done  so  that  no  trap
removes less PM  per  dollar than  it could  remove if  installed  on any  engine
without a trap.   Finally,  the  marginal  cost of  removing a ton  of NOx  is the
same for every engine within the averageable set. .
Finding Marginal  Emissions Control  Costs

     Finding a set of emissions levels that meets the three conditions set  out
above is a straightforward, but not trivial,  problem.   It  requires  estimates o

-------
                                      A-2

the marginal  cost  of  NOx, which  in turn  requires  the  estimation of  marginal
costs of PM removal-.

     The reason the NOx  estimate depends  on  the  PM marginal  cost estimate  is
that these two pollutants are  interrelated: decreases in NOx emissions generally
cause increases in  PM  emissions, which  must  be compensated  for in  some  way.
Thus, the cost of  NOx  removal  includes  not only the costs of  fuel consumption
increases resulting directly from NOx-control actions,  but an  indirect cost  as
well.  Removing a ton  of NOx  can cause  an  increase in  PM emissions of  on  the
order of a tenth of a  ton (for an engine without a  trap).  To remove  this added
PM, a slightly  larger  proportion of  engines  in the averageable set  must  use
traps.  To be efficient,  the  traps  must be added to the engines  for which the
cost per ton of  PM removed is  the  lowest of all  engines  that do not already
have traps.  The  cost  per ton  of PM  removed is  calculated  as  the  total  trap
cost divided by the reduction  in PM emissions per  BHP-hr due to the  trap times
the number of BHP-hrs in  the engine's life.  The inverse of  the  number of tons
removed  per dollar spent  on additional  traps is equal to the  number  of dollars
per ton  of PM removed—in other  words,  the marginal  cost  of  PM removal.   Thus,
the cost of removing a ton of  NOx must  include the  marginal  cost per ton of  PM
removal  times  the number of tons of PM added by the removal  of one ton of NOx.
Formally:

       MC (NOx) =  d(TC)/(NOx) =  $TC/frNOx + &PM/$NOx *  d(TC)/d(PM)
The functional   relationships  needed  for  the  marginal   cost  computations  are
derived by differentiating equations  based  on the  data  provided by  ERC,  Inc.
(see Appendix B) that relate  fuel  consumption  increases  and  PM emissions to NOx
emissions levels.
The Algorithm for Finding the Most Efficient Allocation

     An optimization algorithm is  used to  find  the  optimal,  i.e.,  lowest-cost,
most efficient NOx levels and trap use patterns for each scenario.  The algorith
                                                                             i

-------
                                      A-3

works by  starting  with each  engine  type within  each averageable  set exactly
meeting the  NOx  standards,  and with  100% trap  usage.   This  results  in the
correct total- emissions  of NOx  (though  not at  lowest  cost) and  PM emissions
that are lower than necessary.  The first step in finding the optimal allocation
of emissions is to  calculate the  cost-effectiveness  of traps on  each type of
engine--the dollars  in  trap-related  costs imposed  on an engine per ton  of PM
removed.  The program  "removes"  traps,  starting with the  least  cost-effective
ones, until total PM  emissions rise to the  permitted level.  The  cost  per PM
ton removed  for the  addition of a trap  to  the marginal engine  type (the type
with the last trap removed by the program)  equals the marginal cost of removing
a ton of PM (given the NOx emissions levels).

     The next step in  the  optimization  procedure is  to  calculate  the  marginal
costs of  removing  a  ton of  NOx from each  of the engine  types.   As described
above, the PM increases predicted to result from the  reduction  of NOx by a ton
are charged to  the cost of  the NOx reduction  at  the marginal cost of PM reduction
(calculated as  described above).   For  example, if  removing  a ton of  NOx from
a certain engine costs  $1,000 by itself (due to increased fuel consumption), and
increases PM emissions by 1/10 of a ton, and a ton of  PM costs $6,000 to remove,
then the marginal  cost of NOx removal  for that engine is $1,000 plus $6,000/10,
or $1,600 altogether.

     These marginal   costs  will  differ  for different engines,  since the  NOx
emissions levels   have  not yet been  set  at  their  optimal   levels.   Newton's
method is used  to  adjust  each NOx level until the marginal  cost of NOx emissions
is equalized across engine types  (and  for  types for  which  some  have traps and
some don't, the MC  for NOx ton removal is  set equal for  trapped  and  non-trap
versions) while the average NOx  emissions level  is  kept  at the  level  that will
allow the standard to be met.

     The process  is then begun again, but with the  NOx  levels changed  to equal
those calculated  to equalize  marginal  costs  of NOx  removal.  The most efficient
trap allocation scheme  is again  found,  and the  marginal  cost  per  ton  of  PM
removed is again  found.   This  time,  the  marginal  cost of  PM  removal  is  close

-------
                                      A-4

to its correct value because more refined NOx levels are used as inputs.  Again
the NOx  levels  are adjusted  so that  the marginal  costs of  NOx  removal  are
equal.  This process is repeated until no  further  change in  NOx levels  is  seen
on successive iterations.

     Once the iterative process  is complete, the NOx levels and trap percentages
have been set so  as  to meet the emissions limits within  each  averageable  set;
the traps are allocated across  each  averageable set so  that  no traps  are less
cost-effective that traps would be on any  engines without  traps; and the marginal
costs of removing  a ton of  NOx (including  the cost of compensating for increased
PM production by increasing the  trap percentage)  are equal across each  average-
able set.

     Total costs  for  each  averageable set  may then  be  found  by  adding  the
increased fuel consumption  costs for each engine  type,  given its  optimal  NOx
level, and adding  the costs of the traps  used.

     Cost allocations  within  an averageable set  (especially  across different
manufacturers in  cases  in  which trading   is permitted) may be  found by adding
to the direct costs for each engine type  the cost  of purchasing enough  credits
to bring that engine  type  into  compliance  (with  compliance defined  as either
emitting few enough tons to meet standards if the  standards  had  been expressed
on a  per-truck  basis, or  holding  enough  credits  to emit  sufficient  tons  of
pollutants to make up  the  difference).   The  number of tons'  worth  of  credits
purchased for each engine  is  equal  to   the  difference  between the  engine's
emissions per BHP-hr  and   the  target  for the  industry, times  the number  of
BHP-hrs expected to be  generated by  that engine over its  life.   The price  per
credit is assumed  to be equal to the marginal  cost of generating more  credits.
Because the  average  emissions  are  set  it the  target   for  the industry,  net
expenditures on  credits across the averageable  set  are  zero.

-------
                                   Appendix B
                          BACKGROUND MATERIAL FROM ERC
     This appendix  contains   a  detailed  description  provided  by  Energy  and
Resource Consultants,  Inc.  (ERC)  of the  engine  classification system  used in
this report.  It also provides a discussion of the development of the functions
relating emissions to each other,  and to fuel  consumption.

     Not all of the  engine  families cited as typical  of  particular subclasses
are still in production.  This  fact does  not  change the applicability  of the
functional  relationships of NOx to  fuel  consumption or PM emissions  since the
functions were   based  on  projected   technologies,  rather  than  exclusively  on
current practice.

-------
                    Energy and Resource Consultants, Inc.
BVBMOTJI
to: Barry Galef (SCI)
'ran: Christopher Weaver (SRC)
iubject: Classification of heavy-duty diesel engines
; have developed a classification scheme for heavy-duty engines for use in our
rock on the benefits of emissions averaging.  This scheme includes six classes
if heavy-duty diesel truck engines, 1 class of diesel bus engines, and 2
lasses of heavy-duty gasoline engines, totalling nine classes in all. • The
ilasses are described below.
                 ^
  1.  Light-Heavy Duty mi Engines;  These are a relatively new class,
     offered mainly in trucks of Classes 2b through 4 (as well as in light
     trucks) .  The two major examples are the GM 6.2 liter and the IH 6.9
     liter engine, although  Onan also produces an engine in this class.
     These engines resemble  passenger-car diesels in characteristics.

  2.  Light-Heavy Duty PI Engines;  These are just beginning to cone on the
     market — examples include the DI diesels used in the new Isuzu Class 3
     trucks and the IVECO 8060.  A number of engines of this class have been
     developed in Europe and Japan, and we can expect to see more of them
     over the next decade.
  3.  fit-andard Medium-Heavy Engine*.  These engines are used  in a variety of
     medium-heavy applications, mostly in Class 7 and  the bottom of Class 8.
     Typically naturally aspirated and with less power, less  durable, and
     less efficient than the remaining classes.  The naturally-aspirated
     version of the Caterpillar 3208 is in this class, as is  the IH 9.0 liter
     engine.

  4.  Premium Medium-Heavy Engines*  These engines are also used mostly in
     Class 7 through the bottom of Class 8.  They differ from the standard
     medium-heavy engines in incorporating more "heavy-truck* features in
     their design, including turbocharging and improved fuel  injection
     systems.  The IH 466 engine is an example of this class.

  5.  Non Line-Haul Heavy-Heavy Engines;  These engines are generally built
     on the same blocks as the line-haul engines, but  include changes in
     calibration and accessories (such as turbcchargers) to make them more
     fit for stop-and-go operation, typically in rough service such as
     dump-trucks and logging trucks.  Fuel-economy  is  less of an issue with
     these applications, and high power at a broad  range of RJW is more
     important.  Some manufacturers distinguish these  as  "vocational11
     engines.

  6.  Mn«wHjiui Truck Pngtn»«.  These engines are large, heavy, and highly
     fuel-efficient.  They are optimized for best performance in highway
     cruising, generally with low rated REH.  Turbcchargers and other
     accessories are also optimized for best performance at cruising speeds.

-------
                    Energy and Resource Consultants, Inc.
      Some manufacturers identify these as "economy*  engines.   Cumnins  refers
      to them as "Formula* engines.
   7.   BjaJUaiOfia:   These are heavy-duty truck engines adapted Cor  use in
      transit buses.  These adaptations include provision for rear-mounting,
      derating to reduce smoke, and possible changes in fuel-pump calibration
      to use Diesel  1 rather than  2.  There are probably differences in
      turbocharging as well.  Examples are the Cumnins NHHTC and the various
      CDA "coach" models.
   8.  Lighfc-Bgjwy r?uty (ilflflolinf* Enqi-neg;  These basically resent? le
      light-truck engines, incorporating few "heavy-duty11 features except size
      and power output.  They are not intended for long running at high power,
      and are found mostly in Class 2B through 4.  Many axe also used in light
      trucks.
   9.  Medium-Heavy Duty p*anifpg Engine;  These are true heavy-duty gasoline
      engines, normally incorporating features such as heavy-duty valves,
      hardened valve-seat inserts, governors, etc.  They are used mostly in
      trucks of Classes 5 through 7.  EPA projections (with which I agree)  are
      that these engines will probably almost die out in the next decade,
      having been replaced by diesels.

There may be some practical difficulties in distinguishing the different
diesel-engine classes from each other — especially in distinguishing between
Classes 5 and 6.  If the differences in control costs for these two classes
are sufficiently low, we might want to combine them.  There is also some
grading-together of Classes 4 and 6, but this should be easier to handle.

-------
                     Energy and Resource Consultants, Inc.

                     THFf TPT
       u flhlnq
    Indirect Injection
    Typically -128 to 150 horsepower
    Not rated for continuous full power
    Short lifetime — about 100,000 to 150,000 miles
TVPJe
    Class 2B through 4 light-heavy trucks
    Heavier trucks up to 2B,000 GVW in light service
    Detroit 6.2 1
    IH 6.9 1
    Indirect injection with prechairber optimised for low emissions
    Natural aspiration or turbocharginy (no aftercooling)
    Distributor injection pump with full electronic governor
    Closed-loop electronic injection timing control
    Ceramic monolith/additive system looks most promising.
    Hay require an ignitor as well.

-------
                 Energy and Resource Consultants, Inc.
     ahin
Direct Injection
Typically- 120 to 150 horsepower
High rated speed — 2600 to 3600 RIM
Weight range typically 600 to 1000 pounds
Not rated for continuous full power
Moderately short lifetime

Commercial trucks in Classes 2, 4, and 5.
Possibly Class 2B as well.
Light Buses such as school buses.
ISU2U 3.9 1
IVEOO 5.5 1
Cunmins 3.9 1 "B* engines
              TechnolooipH

Turbocharging with jacket-water after cool ing
High-pressure in-line injection pump
Electronic governor control
Closed-loop electronic injection timing control
Electronically-modulated BGR at low NDx levels
Little information on traps in this class.
Ceramic-acnolith/additive system or catalyzed trap upstream
 from turbochager look most promising.

-------
                 Energy and Resource Consultants, Inc.
Naturally aspirated
Typically 150 to 250 horsepower
Usually not rated for continuous full power
Moderate rated speeds — typically 2000 to 2600 RIM
Moderate lifetime — typically 250,000 miles
Weight typically 1000 to 1500 pounds
Medium-heavy trucks from 16,000 to 50,000 pounds GVW
Usually not in high-speed service (e.g. garbage trucks/
snail dump trucks, delivery trucks, school buses).
Detroit 8.2 1 MA
Caterpillar 3208 NA
IH 9.0 1
High-pressure in-line injection pump
Mechanical timing advance
Mechanical governor
Improved breathing — 4 valves/cylinder
Mechanically-modulated HER at low NDx levels
        Traa
Ceramic monolith/additive system looks most promising.

-------
                 Energy and Resource Consultants, Inc.

                               BMfST1>ES
Turbocharged
Typically 170 to 250 horsepower
Usually not rated for continuous full power
Moderate rated speeds — typically 2000 to 2600 MM
Moderate lifetime — typically 250,000 milea
Weight typically 1000 to 1500 pounds

Medium-heavy duty trucks from 16,000 to 60,000 pounds GVW.
Detroit 8.2 1 turbocharged
Caterpillar 3208 turbocharged
IH DT and OTI 466
Turbocharging with jacket-water aftercooling
High-pressure in-line injection punf> with electronic governor
Closed-loop electronic injection timing control
Electronically-modulated BGR at low NDx levels
Ceramic monolith/additive, Daimler-Benz trap/additive, or catalyzed
monolith upstream from turbocharger  lock most promising.
Daimler-Benz system was developed for this class.

-------
                     Energy and Resource Consultants, Inc.

         r.TMR-HAnr
    Low or very low rated speed — 1600 to 2000 RIM
    Power output typically 250 to 450 horsepower
    Low to moderate torque rise
    Optimized for best efficiency at highway speeds and near full
          load
    Designed and rated for continuous operation near full power
    weight typically 2000 to 3000 pounds
    Very long design lifetime — 250,000 to 400,000 miles before
      rebuild, and rebuildable indefinitely.
          P  -
    Line-haul trucking
    Cunmins "Fleet", "Formula", and big-cam NTC models
    Detroit 6V-92 and 8V-92 (1800 and 1950 REM) models
    Caterpillar 3306 and 3406 "Economy" models
    Most Mack models
Emissions—
    Turbocharging with cold-charge aftercooling
    Ultra-high pressure in-line pump or unit injectors
    Full electronic governing
    Closed-loop electronic timing control
Particulate Trars
    The ceramic monolith/burner system looks most promising,
    because of potential for long life.  The Daimler-Benz and
    monolith/additive systems are also possibilities.

-------
                     Energy and Resource Consultants, Inc


rrJLQg Kt HEAVY-HEAVY    * •VnTAfPTflMAL* f tPM T.TM^-HJUTT.
    Moderate rated speed — 2200 to 2400 REM
    Power output typically 250 to 550 horsepower
    High torque rise and good efficiency over a vide range of
      speeds.
    Designed and rated for continuous operation near full power
    Weight typically 2000 to 3000 pounds
    Very long design lifetime — 250,000 to 400,000 miles before
      rebuild, and rebuildable indefinitely.

                   s
    On/off road operation — logging trucks, dimp trucks, trash
      trucks, heavy farm trucks, other specialized applications.
    Hauling heavy loads (e.g. gravel)  in stop-and-go or short-haul
          service.
    damans Power-Torque, Twin-Turbo 475, and KT Engines
    Hack 2-Valve engines
    Detroit 8V-71, 6V92, and 8V-92 hlgh-RFM nodels
    Caterpillar 3306, 3406, and 3408 •Vocational* ratings
   go i
    Turbocharging with jacket-water aftercooling (some will have
      cold-charge cooling)
    Ultra-high pressure in-line punp or unit injectors
    Full electronic governor
    Closed-loop electronic tuning control
    Small amount of BGR at low NDx levels
    Monolith/burner, monelicVadditive, or Daimler-Benz systems
    look promising.

-------
                     Energy and Resource Consultants, Inc.


      7;  BDS
    Moderate horsepower  heavy-heavy engines specialized for use in
      transit buses.
    Transit buses
    Intercity buses
    Detroit 6V71  and 6V92 "Coach"  Series
    Cummins NUHIC
Bniaflionfl— Related Technologies Aasmned

    Turbocharging with jacket-water aftercooling
    Ultra-high pressure in-line punf> or unit injectors
    Full electronic governor calibrated for minimal particulate
      emissions (at the expense of performance)
    Closed-loop electronic injection timing control
    Monolith/additive system, possibly with additional igniters,
    is most promising.  Catalyzed monolith upstream from turbo and
    Daimler-Benz system are also possibilities.

-------
                     Energy and Resource Consultants, Inc.

                     PDW GA.qnT.TTJE
    Gasoline fueled
    Low to moderate horsepower
    Not rated for continuous high-power operation
    Lack of "heavy-duty* features such as sodium-filled valves,
      valv^seat inserts
    Low durability — typically 100,000 miles
    Basically similar to light-duty truck engines.

    Light-heavy trucks in classes 2B, 3, and 4.
    Class 5 trucks operating under light loads.
Tial
    Ford 240, 300, 360
    01 292, 307, 454
    All Chrysler UDGE
                  Technoloj
    Fuel injection
    Exhaust-gas recirculation
    (Possibly)  Three-way catalysts at very low NDx levels
    Not applicable

-------
                     Energy and Resource Consultants, Inc.

rTA.cs Q< MvnTrM-HPJWY nmv GASOLllE EICI1ES
niotlrvnilahina
    Gasoline fueled
    Typically moderate horsepower
    Bated for continuous or near-continuous full-power operation
    Incorporate "heavy-duty" design features such as valve-seat
      inserts and sodium-filled valves
        Ami
    Medium-heavy trucks under moderate to severe service
    A few heavy-heavy trucks in light to moderate service

    Ford 370 and 429
    CM 250, 366, 427 and 454
    Host International Harvester HUGE
                  Technolo iea
    Fuel injection
    Exhaust-gas recirculation
    Not applicable

-------
                     Energy and Resource Consultants, Inc
ESTIMATES OP COST FUNCTIONS BY ENGINE CLASS
Estimates of trap-oxidizer costs and efficiency the relationships between NDx
and engine-out particulars and NOx and fuel-econooy were developed separately
for each of seven classes of heavy-duty diesel engines.  The NDx/fuel-econony
relationship was also estimated for two classes of heavy-duty gasoline
engines.  This was necessary, since different classes of engines have
different speed and operating characteristics, and thus different emissions
patterns.  The feasibility of emissions-related technologies such as
turbocharging and aftercooling also varies between engine classes.  Tables
XX.1 through XX.9 list the salient characteristics of each engine class, sane
typical engine models within each class, and the emissions-related
technologies which were assumed in estimating each class1 cost functions.

Estimation of trap-oxidizer costs and efficiency relied primarily on earlier
work by one of the authors.  The rationale and assumptions involved have been
documented elsewhere (Weaver, 1984b).  The estimates of NOx/particulate and
NDx/fuel-econony tradeoff functions relied primarily on reports of
manufacturer's tests on development engines.  These reports are mostly
confidential, and so cannot be cited directly.  Analysis and application of
data from individual tests involves a great ^yl of engineering judgement, and
tradeoff curves shown are not, in general, those for any particular existing
engine.  Rather, they show the emissions and fuel-econoy levels which are
estimated to be achievable by engines of each class in model year 1990.

The estimates developed represent a major improvement over the capabilities of
present-day production engines.  Such estimates are necessarily speculative,
and (given the rapid progress that has been made in the last few years) they
may be overly pessimistic.  On the other hand, unexpected delays in
development or the failure of some technologies to fulfill their projected
potential could reveal these estimates as having been overly optimistic
instead.

-------
                                   Appendix C

       INPUT DATA 'FOR COMPUTATION OF SAVINGS FROM REGULATORY FLEXIBILITY


Definition of Engine/Use Classes Covered In This Appendix
Class

LHDGE
LHDDE-IDI
LHDDE-DI
MHOGE
MHDDE-NA
MHDOE-TC
HHDDE-LH
HHDDE-NLH
ERC Number   Definition
    8
    1
    2
    9
    3
    4
    5
    6
   Light-heavy duty gasoline engine
   Light-heavy duty diesel  engine—indirect injection
   Light-heavy duty diesel  engine—direct injection
   Medium-heavy duty gasoline engine
   Medium-heavy duty diesel  engine—standard/naturally aspirated
   Medium-heavy duty diesel  engine--premium/turbocharged
   Heavy-heavy duty diesel  engine—line haul
   Heavy-heavy duty diesel  engine—non-line haul/vocational
These classes are described in Appendix B.
Data Used in Costing and Optimizing Model

BHP-hrs/Truck:

LHDGE   LHDDE-IDI  LHDDE-DI   MHDGE   MHDDE-NA  MHDDE-TC  HHDDE-LH
78,540   86,46086,460   164,450  338,365   364,820   788,800

Source:  BHP-hrs/mile (below) times miles  per truck (page C-2).
                                                           HHDDE-NLH
                                                            788,800
BHP-hrs/Mile:

LHDGE   LHDDE-IDI  LHDDE-DI   MHDGE   MHDDE-NA  MHDDE-TC  HHDDE-LH   HHDDE-NLH
0.714
0.786
0.786
1.495
1.829
1.972
2.720
2.720
Source:  EEA^/ (Appendix B) provides BHP-hrs/mile  for  gas  and  diesel  trucks  by
MVMA weight class for  1987  & 1992.  Linear interpolation  was  used to estimate
1991.  To transform the data on MVMA classes  into  the  ERC  classes used in this
report, MHDDE-NAs were assumed  to be combinations of MVMA Class  6 and 7 (weighted
by projected sales of  diesels  in  these classes) and MHDDE-TCs were assumed  to
be half of  MVMA  Class  6 diesel, half of MVMA  Class 7  diesel, and  all  of MVMA
Class 8-1.  HHDDE-LH  and HHDDE-NLH  were  assumed to  be MVMA   Class 8-2.  ' 3oth
LHDDE-IDI and LHDDE-DI were assumed to have the  same BHP-hr/mile  as MVMA  Class
4 diesel.  MHDGE was  estimated  as  a sales-weighted  average  of the BHP-hr/mile of
MVMA Classes 6 and 7.   LHDGE  was  assumed to have the same BHP-hr/mile  as MVMA
Class 2B-4.
I/ Historical and Projected Emissions Conversion Fractor and Fuel Econo
for Heavy-Duty Trucks labZ-ZUUZ, Motor Vehicle Manufacturers Association of t
United States, Inc., prepared by Energy and Environmental Analysis, Inc. (EF
December 1983. /
%t
/LJ
                                                                                >/7

-------
                                      C-2
1991 Sales by Class:.

 LHDGE   LHDDE'-IDI  LHDDE-DI   MHDGE    MHDDE-NA   MHDDE-TC  HHDDE-LH  HHDDE-NLH
211,663   132,833    44,278   169,332   29,838    51,305    77,404    38,702

Source:  Bob  Johnson,  EPA,  Office  of Mobile  Sources,  Ann  Arbor,  Michigan,
         February 1986.


Useful Life Miles/Truck:

 LHDGE   LHDDE-IDI  LHDDE-DI   MHDGE   MHDDE-NA  MHDDE-TC  HHDDE-LH   HHDDE-NLH
11U.UUU   110,000   110,000   110,000   185,000  185,000   290,000     290,000

Source:  EPA,  Office  of Mobile  Sources,  Ann  Arbor, Michigan,  February 1986.


Cost per One Percent Increase in  Fuel Consumption:

LHDGE   LHDDE-IDI  LHDDE-DI   MHDGE   MHDDE-NA  MHDDE-TC  HHDDE-LH   HHDDE-NLH

 $0        $54       $54        $0       $259      $259      $705       $705

Source:  Regulatory Impact  Analysis, March 1985,  op. cit., pages 2-29 and 3-42.


Fuel Consumption Impacts  of Traps:

     A 1.25  percent  fuel   consumption  penalty for  traps  was  assumed  for  all
heavy duty diesel engines.   This represents the  midpoint of the 1 to 1.5 per-
cent range based  on the use of  trap-oxidizer systems  using  ceramic monolith
substrates and fuel burners  for regeneration; if the  same traps  were used with
an electric  regeneration,  it is  likely the penalty  would  be about  the same.
However, if the ceramic fiber trap  is used, the fuel  penalty  would be somewhat
less: 0.5 to 1.0 percent.

Source:  Regulatory Impact  Analysis, March 1985,  op. cit., page 3-86.


Discounted Costs per Trap:

LHDGE   LHDDE-IDI   LHDDE-DI   MHDGE '  MHDDE-NA   MHDDE-TC   HHDDE-LH   HHDDE-LH

 NA       $370        $370       NA        $448       $448       $574       $574

The ceramic  monolith/fuel  burner  trap  oxidizer  is  used  to  determine the trap
costs for the  particulate  standard, due to the greater  uncertainty associated
with the other designs.  A  10 percent discount rate  is assumed over the life of
the engine.  If another trap oxidizer design were used which was more expensive,
for example, cost savings results (savings  from trap  avoidance)  would be greater.,

Source:  Regulatory Impact  Analysis, March 1985,  op.cit.. page 3-82.

-------
                                      C-3
Trap Efficiency:
80% of engine-out PM emissions are removed.
Source:  Regulatory Impact Analysis,  March 1985, op.cit.. page 2-65.

1991 Projected sales shares by manufacturer:^/
             LHDGE LHDDE-IOI LHDDE-DI MHDGE MHDDE-NA MHDDE-TC HHDDE-LH HHDDE-NLH
Bluebird
Chrysler 16.1
Ford I/ 9.7
Cummins
Caterpillar
Daimler-Benz
KHD
GM 74.2 51.4
Navistar 45.6
Hi no
Deere
Mack
Onan 0.6
Perkins
Renault 0.4
Saab
Isuzu 1.9
Iveco
Volvo
White

40.0 59.5
55.0
36.1
1.0

40.5 36.5
17.7
2.7






2.5 1.4
4.5

2.5


1.6
16.8
5.2
0.2
21.5
26.9
0.8
0.6


0.6
9.1
0.8
6.3
1.0
3.4
5.2


59.1
13.7
0.1
0.2
11.4



15.0






0.6



59.1
13.7
0.1
0.2
11.4



15.0






0.6

     These are rough  projections  based  on  historical   shares  (from EPA  sales
fractions for engine families),  estimates of market  penetration by new engines,
and data on sales volumes of gasoline  powered trucks and imports.
     These share projections are not precise enough to be considered predictions
of market shares of  individual  firms,  e.g.,  Cummins, in the future.   They do,
however, allow the analysis  to  reveal,  for  instance, what is  likely  to happen
to a "Cummins-like"  firm--one with a  large  share concentrated in  the  heaviest
diesels but with some presence in  other segments.
     Totals may not add to 100 percent due to rounding.
Source: (see page C-4)
     }_/ (see page C-4)
     £/ (see page C-4)

-------
                                      C-5
Functional Forms and Parameters:

     Equations relating NOx  levels to  changes  in  fuel  consumption and  to  PM
emissions were derived  by  selecting parameter values  for  hyperbolic functions
that defined  curves  closely fitting the  point  estimates of  the relationships
made by ERC.  For the  relationship of  NOx emissions to  increases  in fuel  con-
sumption, the functional form used in the analysis is as shown:
% increase in fuel  consumption       a   +  b/(c  + NOx)
                                                                NOx
115.9
800
4
6.15
-427.87
7100
15
15
-369.12
6200
15
12.5
where NOx is emissions of NOx measured  in  grams/BHP-hr,  and  a,  b, c, and d are
parameters with values that  depend on engine class:

     LHDGE  LHDDE-IDI  LHDDE-DI   MHDGE  MHDDE-NA  MHDDE-TC  HHDDE-LH  HHDDE-NLH

a    -57.37   -11.6    -115.9   -57.37    2.41
b    500       24       800      500      15.3
c      6       -1.5       4        6      -2.5
d      3        1         6.15    3      -0.63

     Fuel consumption functions at the  level  of detail  shown  in  Exhibits II-l
and 1 1-2  were  not  addressed  in  EPA's  RIA,  so  the ERC  functions  were used.
However, EPA did make point  estimates of fuel consumption impacts  for the 5.0
g/BHP-hr standard, and these are lower than  those developed by ERC.  The primary
reason for the  differences lies in the area  of  how technological advances by 1991
are considered.  EPA's analysis  concluded that the next  five years  would bring
improvements to overcome most of  any fuel   consumption penalty  associated with
the 1991 standards.   ERC's analysis  also considered technological improvement,
but evaluated fuel consumption increases as foregone gains.

    While there are  differences between EPA's and  ERC's analyses,  they  do not
affect the results  of this  study, since a consistent  set of  input  values was
used for  all  scenarios  addressed here  to evaluate the incremental  benefits
of trading and  expanded  averaging.

    For the relationship of  NOx emissions  to  PM emissions, the  functional form
used in the analysis is  as  shown:
PM
                f/(g  + NOx)
where NOx  is  emissions  of  NOx  measured  in grams/BHP-hr,  and  e, f, and  g are
parameters with values that depend on engine class:
     LHDGE  LHDDE-IDI  LHDDE-DI  MHDGE  MHDDE-NA  MHDDE-TC  HHDDE-LH  HHDDE-NLH
e
f
9
              0.407
              0.134
             -1.8
 0.15
 0.85
-1.9
 0.29
 0.65
-2.1
 0.18
 0.60
-2.1
 0.16
 0.60
•1.5
 0.15
 0.67
-1.8
Source of point estimates  that  were  the basis of the parameter estimates: Appen
dix B and Sections II-B and INC.

-------
                                   Appendix D
                                DETAILED RESULTS
     This appendix presents detail  on costs and emissions by firm and by subclass
for each of the scenarios.

      Full data for each  firm  is presented separately  for  the  three scenarios
in which averaging but not  trading  is allowed.   (The data shown for the industry
is not the sum or  average of data for the  firms since it refers to the situation
under trading.) Under trading,  the industry is treated as  if it were combined
into a  single,  large  firm: all  data  presented for the  emissions,  trap usage,
and per-engine costs for the industry  as  a whole are  valid for each firm, and
all total cost  figures  presented  for  the industry become  valid for individual
firms when they are adjusted for the firms' shares in the sales of the industry.
For the scenarios  in which averaging is restricted to transactions among limited
sets of engine types,  subtotals are shown for each subclass.

     The first column identifies the engine type/use class  for  the  row of data
in the other columns.   These classes are defined in  Appendix B.

     The next two  columns  show the optimal  grams of  NOx'  under  averaging for
engines without and with traps,  respectively.  Emissions of  NOx are kept somewhat
.higher for engines without  traps to minimize PM emissions.  For engines equipped
with traps, the  engine-out emissions  of  PM  are less  important,  because the
traps eliminate most of thenr.

     The next column shows the  percentage  of  engines that  would be  fitted "ith
traps.  The traps are  placed   mostly  on  the larger  engines  to minimize the
trap cost per ton of PM removed by them.

     The next two columns show the costs of emissions controls per engine.  The
first, labeled "Uithout Credits,"  shows only  the  direct  costs  of the traps and

-------
                                      D-2
the fuel  consumption  Increases  at  the emissions  levels  found  to be  optimal
under averaging (or trading, In the  case of the industry-wide data); the second,
labeled "With Credits," includes the  costs  of "purchasing" credits for  NOx  or
PM emissions  in  excess  of  the target.  Under averaging,  the  credits are  not
really purchased,  since  all  transactions  are intra-firm.   These  results  are
presented as  though  intra-firm markets  for  credits  existed,  with  accounting
prices set at the marginal  costs of  reducing emissions.

     The next two  columns  show the costs per engine  of purchased  credits  for
NOx and PM,  respectively.   Again,  these  costs are  internal  accounting  costs
only under averaging,  and  the  total  expenditures  sum to  zero  for each  firm.

     The next two columns present total costs, in millions  of dollars,  for all
of each firm's  sales  in each  class.   The  first  of  the  two columns  shows  the
totals under averaging of NOx and PM without adjusting for  credit transactions,
while the next column takes  these  transactions into account.  The  total  costs
over all classes are shown  at the bottom  of  the column.  It is the same  whether
or not credit transactions  are included,  since the  transactions  within a  single
firm sum to zero.

     The final column  shows annual  sales  by  class and by  firm.

-------
DETAILED RESULTS: ONE CLASS AVERAGING AND TRADING
NUX U11»1UN» -
(g/BHP-hr)

FIRM ENGINE

TYPE •

No Trap
•
Trap
rn
TRAP
USAGE
(X)
COSTS PSK TKUCK -----------
Without
Credits
With
Credits
Credit Purchases.
For NOX
Foe PM
1UIAL. U»19 	
(millions)
Without
Credits
With
Credits
YEARLY
SALES
INDUSTRY "
LUDGE
LHDDE -
LHDDE -
HHDCE
MHDDE -
MHDDE -
HHDUE -
HHUDE -

IDI
DI

NA
TC
LH
NLH
2 25
3.67
5 06
2 21
5 0*
5 17
4 98
1 19
na
3.60
4 73
na
4 80
4 95
4 82
* 97
na
0 00
0 00
na
1 00
0 00
0 49
1 00
$407 2
169 0
189 1
857 2
2.337 0
810 8
2.403 3
3,912 6
($24.8)
193 1
517 8
(66 4)
2.644 3
2,199 7
3.884 7
4,816 6
($432 0)
(129 3)
211 0
(923 6)
571 3
1,000 3
1.568 4
1, 702 b
na
153 4
117 7
na
(264 1)
388 6
(86 9)
(79B B)
$86 2
22 5
8 4
145 2
69 7
41 6
186 0
151 4
($5 3)
25.6
22.9
(11.2)
78 9
112 9
300 7
186 4
211.668
132.833
44.278
169.332
29.838
51.304
77.405
38.703
         Industry Total
                                                                 $710 9    $710 9   755,361
       1  Bluebird

         HHDCE
                           4  20
                                             na    $236 2    $236 2      $00
                                                          na    $0 003    $0 003
                                                                                        13
                                                                                                                               O
                                                                                                                               to
       2  Chrysler

         LHDCE
                           4  20
                                             na    $114  7    $114  7
                                              $0 0
                                                                  $3 9
                                                                    $3 9    34.016
       3 Ford
        LHDCE
        LHDDE  - DI
        MHDCE

        Ford Total
      4 Cu
            nlns
        LHDDE - DI
        MHDDE - TC
        HHDDE - LH
        HHDDC - NLH

        Cumnlns Toi .1
4 OS
6 65
4 01
4 38
4 51
4 29
4 42
  na
6 07
  IIJL
4 04
4 26
4 09
4 14
  na    $129 6    $114 1
0 42     208 5     450 2
  na     273 7     234 4
U 00    $351 5    $557 2
U 00   1,519 6   2.4j? /
0 36   3.9J9 6   4,234 0
1 00   6,0*4 8   5,253 9
($15
241
(39

$63
43/
63
(210
4)
7
4)

5
7
3
0)

0


$142
460
231
(630
na
0
na

2
5
1
3)
$2
3
27
$33
$8
1
180
139
7
7
6
.9
6
1
3
4
$2
8
23
$33
$13
2
193
120
4
0
.6
9
6
0
7
2
20
17
100
139
24

45
22
.620
.711
.678
,010
.353
827
, 758
.879
                                                                                           $329  5     S329  5     93.817
     As  noted In ilie text,  nic dm presented for I he Industry Is not the weighted average of the data for the
     firms    This It because the dat* tor the Induitrv shows the situation under trading, and the data for the
     firms  shows the situation under aveiaglng

-------
DETAILED RESULTS. ONE CLASS AVERAGING AND  TRADING
FIRM ENGINE
NUJC HUSSIONS -
(B/BHP-hr)
TYPE • No Trap Trap
I'll
TRAP
USAGE
(X)
COSTS PER TKUbK -
' Without Ulth
Credits Credits
Credit Purchases.
For NOX For PH
(millions)
Ulchout Ulth
Credits Credits
YEARLY
SALES
5 Caterpillar
HHDDE -
HHDDE -
HHDDE -
HUDDE -
NA
TC
LH
NLH
4 45
4 45
4 22
* 35
4.30
4 20
4 02
4 06
1 00
0 00
0 44
1 00
$2.892 1
1,59? 3
4,308 8
6.347 5
$2.832 8
2.4«4 2
4,228 8
5.252 1
$147 5
379 5
(216 6)
(484 0)
($206 8)
467 4
136 6
(611 4)
Caterpillar Total
6 Daimler
HHDDE -
HHDDE -
HHDDE -
HHDDE -
-fienz
NA
TC
LH
NLH

4 40
4 36
4 13
4 24

4 26
4 11
3 92
3 94

1 00
0 60
1 00
1 00

$2.958 7
2.391 7
5.829 3
6.720 6

$2,844 6
2.442 9
4.213 7
5,239 9

$83 6
14 )
$31 2
13 7
45 6
33 b
$124 1

$0 8
6 4
(J 3
0 1
$7 6

$0 3
0 5
0 4
$1.2

$36 4
6.9
33 6
19.9
4 0
18 7
11 8
$30.5
21 0
44 8
27.8
$124 1

$0 8
6 5
0 2
0 1
$7 6

$0 2
0.6
0 3
$1-1

$13 5
16 0
11.8
25.5
19 8
27 6
17 3
10.783
8.600
10.584
5,292
35.259

284
2.6/3
44
22
3,023

99
132
66
298

157.032
68.335
68,641
10.899
11.031
8.800
4.400
                                                                                                                                                          o
                                                                                                                                                          -p.
         General  Hulors  Total
$131 4    S131 4   329.13H

-------
DETAILED RESULTS. ONE CLASS AVERAGING AND TRADING
nux uiiaaiuna ~
(g/BHP-hr)
FIRM ENGINE TYPE * No Trap Trap
9 Navistar
LHDDE - IDI 3.59 3.47
HHDDE - NA 5 01 4 62
HHDDE - TC 5 11 4 71
Navistar Total
10 HLao Hotors
HHDDE - NA 4 33 4 20
HHDDE - TC 4 24 3 99
Hino Motors Total
11 Deere
HHDDE - TC 4.36 4 11
12 Hack
HHDDE - LH 4 29 4 09
HHDDE - NLH 4 42 4 14
Hack Total
13 Onan
LHDDE - IDI 4.30 4 15
14 Peiklns
TRAP
USAGE Without With Credit Purchases Without
(Z) Credits Credits For NOX For PM Credits

0 24 $298 1 $306 6 ($184 4) $192.9 $18 0
1 00 2.508 8 2,499 3 478 6 (488 1) 13 3
1 00 2.043 9 2.010 2 624 0 (657 6) 28 3
$59 6

1.00 $3.046 8 $2.853 7 ($5 4) ($187 8) $2 4
0 IS 2.049 2 2.436.6 10 7 376 7 08
$3 2

0 64 $2.442 6 $2.442.9 ($0 0) $03 $08

0 30 $3,814 8 $4,234 0 $104 0 $315 2 $44 4
1 00 6.092 3 5,253 8 (207 9) (630 5) 35 4
$79 8

0 66 $367 6 $367 0 $0.0 ($0.7) $0 3

(millions)
With
Credits

$18 6
13 2
27 8
$59 6

$2 3
1 0
$3 2

$0 8

$49 2
30 5
$79 8

$0.3

YEARLY
SALES

60.529
5.292
13.826
79.647

794
397
1.191

331

11.627
5.814
17.441

794

                                                                                                                                                          o
                                                                                                                                                          in
        MHDDE - TC
                           4  36
                                   411     0 64  $2.44.' 0  $^.442 9
                                                                         SO U
                                                                                   $08
                                                                                             $0.8
                                                                                                       $0 8
                                                                                                                   331

-------
DETAILED RESULTS. ONE CLASS AVERAGING AND TRADING
                        NOx EMISSIONS -
                             <8/BHP-hr)

    FIRM ENGINE TYPE •  No Trap    Trip
                  PH   COSTS PER TRUCK -
                TRAP
               USAGE   Ulthout      With
                 (X)   Credits   Credits
                                   ..---_-..   TOTAL COSTS 	
                                                              (millions)
                                   Credit Purchases:   Ulthout      With
                                   For NOX    For PH   Credits   Credits
                                                                    YEARLY
                                                                     SALES
      IS Renault
         LHDDE - IDI
         MHDDE - TC

         Renault Total
3 30
4 39
3.25
4 1*
0 00    $272 8
0 66   2,416 5
  $70 8
2,442 4
($340 t)
   43 4
$138 6
 (17 5)
 $0 2
 11.3

$11 4
 $0 0
 11 4

$11 4
   595
 4.670

$5,266
      16 Saab

         HIIDOE - TC
                           4 36
                                   4 11
                0 64  $2,442 3  $2,442 9
                                      $0 U
                                        $0 5
                                                                  $1  0
                                                                            $1  0
                                                                                       397
      17 Isuzu
         LHDDE - IDI
         LIIDDE - DI
         MHDDE - NA
         MHDDE - TC

         Isuzu Total
      18 Iveco

         MHDDE - NA
         MHDDE - TC

         Iveco Total
      19 Volvo

         MHDDE -  TC
         HHDDE -  LH
         HHDDE -  NLH
3 35
4.39
4 50
4 52

4 32
4.23

4 43
4.21
4.33
3 30
4 06
4 35
4 28

4 19
3 98

4 18
4 00
4 04
0 00
0 00
1 00
0 80

1 00
0 01

0 29
1 00
1 00
$254
347
2,833
2,376

$3.058
1,910

$1.938
5.538
6.406
0
6
6
7

6
7

2
1
1
S97
556
2.823
2.431

$2.854
2.434

$^.441
4.231
5,255
b
1
2
7

5
a

i
9
5
(S^'JB
67
201
188

($17
45

$248
(b57
(548
3)
8
1
4

7)
3

5
8)
5)
$141
140
(211
(133

($186
478

$254
(648
(602
8
7
5)
3)

.4)
9

3
4)
2)
$0 7
0 4
1 2
7 7
$9 9
$4 1
1 0
$5 2
$3 3
2 5
1 5
                                                                            $0 3
                                                                             0 6
                                                                             1 2
                                                                             7 9

                                                                            $9 9
                                                                            $3.9
                                                                             1  3

                                                                            $5  2
                                                                            $4.2
                                                                             1  9
                                                                             1  2
                                                                             2,580
                                                                             1,107
                                                                               430
                                                                             3.228

                                                                             7,345
                                                                             1,356
                                                                               529

                                                                             1.885
                                                                             1,720
                                                                               459
                                                                               229
         Volvo Total
                                                                                             $7 J
                                                                                                       $7 3
                                                                                                                2.408

-------
DETAILED RESULTS: ONE CLASS AVERAGING AND TRADING
                        NO* EMISSIONS -
                             (g/BHP-hr)

    FIRM ENGINE TYPE •  No Trap    Trap
                             PM   COSTS PER TRUCK -
                           TRAP
                          USAGE   Without       With
                            (Z)   Credit!    Credit*
                                                	   TOTAL COSTS 	
                                                                           (millions)
                                                Credit  Purchases.   Without      Ulth     YEARLY
                                                For  NOX    For PM   Credits   Credits      SALES
      20 White
         LHDDE - DI
         MUDDE - TC

         White Total
          * 25
          4.37
             3.90
             *  13
0 00
0 72
 $390 6
2.510 8
 $553.8
2,442 7
$17 2
 (7 2)
$146 0
 (60 9)
$0 4
 6 6

$7.1
$0 6
 6 5

$7 1
1.107
2.646

3.753
       * ENGINE TYPES
         LIIDCE
         LIIDDE
         LHDDE
         MHDCE
         MHDDE
         MHDDE
         HHDDE
         HHDDE
IDI
DI

NA
TC
LH
NLH
                                                                                                                                                          O
                                                                                                                                                          I
Llght-Heavy-Duty Gasoline Engine
Llglii. Heavy-Duty Diesel Engine - Indirect Injection
Llghi-Heavy-Duty Diesel Engine - Direct Injection
MedluiD-Heavy-Duty Gasoline Engine
Medlum-Heavy-Duty Diesel Engine - Naturally Aspirated
Medlum-Heavy-Duty Diesel Engine - Turbo Charged
Heavy-Heavy-Duty Diesel Engine - Line Haul
Heavy-Heavy-Quty Diesel Engine - Non-Line Haul

-------