COST EFFECTIVENESS ESTIMATES FOR
MOBILE SOURCE EMISSION CONTROL
George Miller
10 November 1977
• '..-v Submitted to:
Environmental Protection Agency
Office of Mobile Source Air Pollution Control
Purchase Order CD-7-017T-A
REPORT NUMBER
VRI EPA-3
FR 77-1 (R)
Vector Research, Incorporated
Ann Arbor. Michigan
-------
EPA-420-R-77-102
CONTENTS
Page
1.0 INTRODUCTION 1
2.0 GENERAL APPROACH AND ASSUMPTIONS .... 3
3.0 DETAILS OF THE ANALYSIS 7
3.1 Effectiveness Computations 7
3.1.1 Case 1: Standards Only 8
3.1.2 Case 2: Standards Plus Certification 9
3.1.3 Case 3: Standards Plus Certification
with Parameter Adjustment 14
3.1.4 Cases 4 and 5: Standards Plus Selective
Enforcement Audit 14
3.1.5 Case 6: Standards Plus Inspection and
Maintenance (Diagnostic) 20
3.1.6 Cases 7 and 8: Standards Plus Inspection
and Maintenance (Testing) 21
3.1.7 Cases 9 and 10: Standards Plus Recall 25
3.1.8 Case 11: Standards Plus Certification with
Parameter Adjustment Plus Selective
Enforcement Audit 27
3.1.9 Case 12: Standards Plus Certification with
Parameter Adjustment Plus Inspection
and Maintenance (Testing) 29
3.1.10 Case 13: Standards Plus Certification with
Parameter Adjustment Plus Selective
Enforcement Audit Plus Recall 30
3.1.11 Cases 14 and 16: Standards Plus Certifi-
cation with Parameter Adjustment Plus
Selective Enforcement Audit Plus
Inspection and Maintenance (Testing)
Plus Recall 30
3.1.12 Case 15: Standards Plus Selective Enforcement
Audit Plus Inspection and
Maintenance (Testing) 31
3.1.13 Case 17: Standards Plus Selective Enforcement
Audit Plus Recall 32
3.1.14 Case 18: Standards Plus Selective Enforcement
Audit Plus Inspection and Maintenance
(Testing) Plus Recall 33
3.1.15 Case 19: Standards Plus Certification Plus
Inspection and Maintenance (Testing) ... 33
-------
i i
CONTENTS
(Concluded)
Page
3.2 Cost Estimates 34
3.2.1 Standards 34
3.2.2 Certification 36
3.2.3 Selective Enforcement Audit 36
3.2.4 Inspection and Maintenance 37
3.2.5 Recall 37
4.0 RESULTS 39
REFERENCES 65
-------
i i i
FIGURES
Title
Determination of Best Estimate
and Range of Values for Average
Emission Rates for Case 2
(Standards Plus Certification)
Alternative Deterioration Assump-
tions for Best Estimates and Low
Estimates of IM Effectiveness for
a Single Vehicle
Effect of Recall on Emissions of a
Recalled Vehicle
-------
iv
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Page
4
10
12
16
18
23
24
35
42
43
44
45
46
47
48
49
v
TABLES
Title
Compliance Assurance Programs Included in the
Cost Effectiveness Computations
Regression Analysis of In-Use Emissions as a
Function of Certification Emissions
Regression Analysis of In-Use Emission Rates
as a Function of Mileage for Case 2
Regression Analysis of In-Use Emission Rates
as a Function of Mileage for Case 3
Regression Analysis of In-Use Emissions as a
Function of Assembly Line Emissions
Regression Equations Used to Relate Short-
Test Results to FTP Levels in the IM Simulation
Short Test Cutpoints and Corresponding Percent
Savings from Application of IM Simulation
Summary of Costs Used in the Analysis
Results for Case 1 - Standards Only
Results for Case 2 - Standards Plus Certification
Results for Case 3 - Standards Plus Certification
with Parameter Adjustment
Results for Case 4 - Standards Plus Selective
Enforcement Audit (100% Rejection Rate)
Results for Case 5 - Standards Plus Selective
Enforcement Audit (75% Rejection Rate)
Results for Case 6 - Standards Plus Inspection
and Maintenance (Diagnostic)
Results for Case 7a - Standards Plus Idle
Inspection and Maintenance (30% Failure Rate)
Results for Case 7b - Standards Plus Loaded
Inspection and Maintenance (30% Failure Rate)
-------
17
18
19
20
21
22
23
24
25
26
27
28
Page
50
51
52
53
54
55
56
57
58
59
60
61
vi
TABLES
(Continued)
Title
Results for Case 8a - Standards Plus Idle
Inspection and Maintenance (50% Failure Rate)
Results for Case 8b - Standards Plus Loaded
Inspection and Maintenance (50% Failure Rate)
Results for Case 9 - Standards Plus Recall
(40% Response Rate)
Results for Case 10 - Standards Plus Recall
(70% Response Rate)
Results for Case 11 - Standards Plus Certifi-
cation with Parameter Adjustment Plus
Selective Enforcement Audit
Results for Case 12 - Standards Plus Certifi-
cation with Parameter Adjustment Plus
Inspection and Maintenance
Results for Case 13 - Standards Plus Certifi-
cation with Parameter Adjustment Plus
Selective Enforcement Audit Plus Recall
Results for Case 14 - Standards Plus Certifi-
cation with Parameter Adjustment Plus
Selective Enforcement Audit Plus Inspection
and Maintenance (30% Failure Rate) Plus Recall
Results for Case 15 - Standards Plus Selective
Enforcement Audit Plus Inspection and
Maintenance
Results for Case 16 - Standards Plus Certifi-
cation with Parameter Adjustment Plus
Selective Enforcement Audit Plus Inspection
and Maintenance (50% Failure Rate) Plus Recall
Results for Case 17 - Standards Plus Selective
Enforcement Audit Plus Recall
Results for Case 18 - Standards Plus Selective
Enforcement Audit Plus Inspection and
Maintenance Plus Recall
-------
vi i
TABLES
(Concluded)
Number Title Page
29 Results for Case 19 - Standards Plus Certifi-
cation Plus Inspection and Maintenance 62
30 Cost Effectiveness - Federal Government Dollars
Spent Per Ton of Pollutant Removed 63
31
Cost Effectiveness - Total Dollars Spent Per
Ton of Pollutant Removed
64
-------
viii
-------
1
1.0 INTRODUCTION
This document is a final report of an effort conducted by Vector
Research, Incorporated, (VRI) to develop cost and effectiveness estimates
for alternative mobile source compliance assurance programs for the
Environmental Protection Agency (EPA). The estimates are intended to be
used as input to a mobile source strategy paper being developed by EPA.
This report includes a general description of the approach and assumptions
used in developing the estimates (in section 2.0), a detailed description
of the procedures used to develop the estimates for each compliance
assurance program studied (in section 3.0), and a presentation of the
results of the analysis (in section 4.0).
-------
2
-------
3
2.0 GENERAL APPROACH AND ASSUMPTIONS
The cost effectiveness computations were performed for a set of
alternative compliance assurance programs (referred to as cases) selected
by EPA and summarized in table 1. Each of these programs assumes the
existence of emission standards, and may also include one or more of the
following compliance techniques:
(1) a certification process as currently performed by EPA;
(2) a certification process in which EPA may adjust certain vehicle
parameters within manufacturer specifications before testing
any vehicle;
(3) EPA's assembly line testing program (Selective Enforcement Audit);
(4) a state-operated inspection and maintenance program involving
diagnostic checks on emission-related equipment;
(5) a state-operated inspection and maintenance program involving
emissions testing; and
(6) a recall program.
For each case, effectiveness was computed in terms of reduction in 1975
FTP emissions from pre-1968 levels. Separate effectiveness computations
were performed for four pollutants (HC, CO, N0X, and evaporative HC). All
analysis was based on 1975 standards in 49 states, and considered low-altitude,
non-California light duty vehicles only. Effectiveness is expressed as tons
of pollutant over a 100,000-mile vehicle life, while costs are expressed as
lifetime costs in 1975 dollars. Both total and government costs are presented
for each case.
-------
TABLE ^: COMPLIANCE ASSURANCE PROGRAMS INCLUDED IN THE COST EFFECTIVENESS COMPUTATIONS
COMPLIANCE TECHNIQUES
CASE1
CERTIFICATION
SELECTIVE
INSPECTION AND
INSPECTION AND
STANDARDS
CERTIFICATION
WITH PARAMETER
ENFORCEMENT
MAINTENANCE
MAINTENANCE
RECALL"
ADJUSTMENT
AUDIT2
(DIAGNOSTIC)
(TESTING)3
1
2
X
X
X
3
X
X
4
X
X (100%)
5
X
X ( 75%)
6
X
X
7a
X
X (30%)
7b
X
X (30%)
8a
X
X (50%)
8b
X
X (50%)
9
X
X (40%)
10
X
X (70%)
11
X
X
X (100%)
12
X
X
X (30%)
X (70%)
13
X
X
X (100%)
14
X
X
X (100%)
X (30%)
X (70%)
15
X
X (100%)
X (30%)
X (70%)
16
X
X
X (100%)
X (50%)
17
X
X (100%)
X (70%)
18
X
X (100%)
X (30%)
X (70%)
19
X
X
X (30%)
*Cases 7a and 8a assume the use of an idle test for inspection; cases 7b and 8b assume the use of a loaded test.
All other cases involving inspection and maintenance (testing) assume the use of the idle test.
2Parenthetical numbers in this column refer to assumed fraction of test orders which result in rejection
of the lot.
Parenthetical numbers in this column refer to failure rate which the states are assumed to enforce.
Parenthetical numbers in this column refer to assumed response rate to a recall.
-------
5
Because of uncertainty regarding the validity of many of the assumptions
of this analysis, and scarcity of data for some of the compliance techniques,
a range of estimates has been developed for each case. A best estimate is
presented as the cost and effectiveness judged most likely to result from
implementation of the case. Low and high estimates are presented as "reason-
able" lower and upper limits on effectiveness, with corresponding costs.
Because these ranges of estimates represent uncertainty in estimating the
effectiveness and costs of the various compliance techniques, the range tends
to be smallest for cases involving techniques for which substantial data exist
(such as certification).
The primary data sources used in this analysis were the 1974 and 1975
Emission Factors Programs. Data from 1975 vehicles from these programs pro-
vided a population of in-use vehicles for which to estimate the effectiveness
of various combinations of compliance techniques. Other data (such as 1975
certification results, data from the Restorative Maintenance Program, 1975
assembly-line data, and miscellaneous sets of evaporative emission data and
cost data) were used where needed;1 all data sources and computational pro-
cedures are described in the following section.
^11 data used were current as of July 1977.
-------
6
-------
7
3.0 DETAILS OF THE ANALYSIS
This section describes the analytical procedures used in the cost
effectiveness calculations. Section 3.1 discusses the effectiveness
assessment methodology for each combination of compliance techniques
evaluated, while section 3.2 describes the derivation of the cost estimates
used in the analysis.
3.1 Effectiveness Computations
Effectiveness results were computed for each case in terms of millions
of tons of pollutant saved over a precontrol situation. Precontrol emission
rates were assumed to average 7.55 grams/mile for HC, 86.79 grams/mile
for CO, 3.40 grams/mile for NO , and 3.33 grams/mile for evaporative HC
A
during the life of a given vehicle. In all cases the standards assumed to
be in effect for exhaust emissions were 1.5 grams/mile of HC, 15 grams/mile
of CO, and 3.1 grams/mile of NO . In addition, in-use evaporative HC levels
A
(measured with the SHED technique) were adjusted to give a modified SHED
standard for 1975. To do this it was assumed that prototype vehicles actually
built for certification 1n 1975 met the evaporative standard exactly.
Cannister test data on twenty 1973 model year vehicles in Los Angeles was
examined to develop a relationship between certification and in-use
emissions.1 In particular, the ratio of average in-use emissions to the
average certification level was computed to equal 2.11. This relationship
xThe particular in-use cannister data used correlated fairly well with SHED
data, which was also available for these twenty vehicles.
-------
8
was used to predict the standard from the average in-use level of 1975
vehicles. The result was a predicted evaporative standard of 0.834 when
expressed in units of grams/mile.
Each of the following subsections presents the approach used to compute
the effectiveness for a single case or group of similar cases. These presenta-
tions generally describe the procedures followed to compute average emission
rates in grams per mile under the various strategies investigated. In all
cases, these rates were converted to effectiveness values using the following
relationship:
r^ 3 (uj - c.j) LPk,
where
r.j = reduction in emissions of pollutant i resulting from implementation
of a particular case (in tons) ,
u.j = uncontrolled average lifetime emission rate of pollutant i
(grams per mi le),
c.j = controlled average lifetime emission rate of pollutant i resulting
from implementation of the case (grams per mile),
L = average vehicle life (taken to be 100,000 miles),
P = vehicle population to which the benefit applies (taken to be
10 million vehicles), and
k = conversion factor to convert grams to tons.
3.1.1 Case 1: Standards Only
To estimate the reduction in emissions resulting from promulgation of
standards without any additional compliance techniques, it was assumed that
-------
9
manufacturers would respond by designing vehicles which would just meet the
standards as prototypes. In order to convert these design emission levels
to levels for vehicles in use, regression analysis was performed on data
for 570 in-use vehicles from model year 1975 and certification data for
corresponding vehicles to develop a relationship between design exhaust
emission levels within a given engine family, and 50,000-mile in-use levels
experienced by vehicles in that same family. The regression results are
given in Table 2. The in-use data were for low-altitude, non-California,
light duty vehicles tested in the FY 1974 Emission Factors (EF) Program.
Mileage regressions described in section 3.1.2, were used to project these
in-use levels to 50,000 mile levels for each vehicle. Because the stan-
dard for evaporative emissions was inferred from the 1975 data (as described
earlier), the corresponding in-use level is simply the level actually
realized in 1975 vehicles of 1.76 grams/mile. No deterioration of evapora-
tive emissions with mileage was assumed.
A range of values around this best estimate was developed by assuming
that, at worst, there would be no reduction in emissions from precontrol
levels and, at the other extreme, the reduction would be no better than
the most optimistic estimate of the reduction achieved in the presence of
certification (I.e., the high estimate for case 2).
3.1.2 Case 2: Standards Plus Certification
The effectiveness estimates for the case in which standards and certi-
fication are 1n effect were based on linear regressions of EF data on 1975
model year vehicles. These regressions predict mean exhaust emission rates
as a function of vehicle mileage for a population of vehicles subject to
-------
TABLE 2: REGRESSION ANALYSIS OF IN-USE EMISSIONS
AS A FUNCTION OF CERTIFICATION EMISSIONS
Relation: + fl.. (1 - 1. 2, 3)
where
y^j = in-use emission level of pollutant i by a given vehicle j at 50,000 miles;
= average certification emission level (including deterioration) for
pollutant i for vehicles in the same engine family as vehicle j; and
o
oij, ^ = regression coefficients.
Result8:
i
ai
a
i
R2
Value
Significance
Value
Significance
1 (HC)
0.37177
0.0268
2.1617
0.0000
0.00861
2 (CO)
1.3517
0.0310
24.535
0.0000
0.00816
3 (N0X)
0.69533
0.0001
1.1748
0.0043
0.02692
-------
n
the conditions of this.case. The data used for these regressions were for
low-altitude, non-California, light duty vehicles tested in the FY 1974
and FY 1975 EF Programs. To obtain the best possible fit to the data, two
separate sets of regressions were performed and combined into a single
emissions-versus-mileage regression for each pollutant.1 The first set of
regressions used data for the 533 vehicles in the data base for which the
measured idle CO level (ICC) was less than 1.5 percent; the second set
used the remaining 232 vehicles for which the ICO was greater than or equal
to 1.5 percent. The results of the regressions, and the composite estimate
of emissions for each pollutant, appear in table 3. The best estimate of
emissions was then taken as the value of the composite regression equation
at 50,000 miles.
To develop an estimate for average evaporative emission rates, 1978
certification data were examined to determine the average percent of the
standard at which vehicles were certified. The average evaporative emission
rate for vehicles in the field was then assumed to equal the product
s • f • fc, where
s 3 assumed 1975 standard for evaporative emissions (.83 grams per mile);
f » average fraction of 1978 standard experienced on 1978 certification
test results (0.30); and
i " ratio of in-use evaporative levels to certification levels (2.11,
as computed in the derivation of the assumed 1975 standard).
Upper and lower bounds on exhaust emissions for this case were compu-
ted from upper and lower 95 percent confidence limits on the composite
1See [Williams, 1976] for a description of and rationale for the procedures
used to combine these regressions.
-------
TABLE 3: REGRESSION ANALYSIS OF IN-USE EMISSION RATES
AS A FUNCTION OF MILEAGE FOR CASE 2
i^1
Relation: e^(m) = + n^li m + nli^ + n^ + (v>21 m + n2i^ = m + ni 0 = K 2, 3)
where
e.(m) = average in-use emissions of pollutant i by a vehicle at mileage m;
n. = number of vehicles in sample k (where k=l for idle CO less than 1.5 percent,
and k=2 otherwise);
ro
m = vehicle mileage, in tens of thousands;
yki* nki = re9re5si°n coefficients;
_")"])+ y2i. and
Mi n^ + n2
- = nlqli * n2n2i
ni n^ + n2
-------
TABLE 3: REGRESSION ANALYSIS OF IN-USE EMISSION RATES
AS A FUNCTION OF MILEAGE FOR CASE 2
(Concluded)
Regression Results:
k
nk
i
\\
nk1
R2
K
Value
Significance
Value
Significance
1
1 (HC)
0.19280
0.0000
0.89979
0.0000
0.04471
533
2 (CO)
1.7824
0.0027
12.872
0.0000
0.01685
(IC0<1.5)
3 (N0x)
0.17265
0.0002
2.2916
0.0
0.02625
2
1 (HC)
0.21144
0.0228
1.7917
0.0000
0.02234
232
2 (CO)
2.7543
0.1793
39.486
0.0000
0.00783
(IC0>1.5)
3 (N0X)
0.0072075
0.9249
2.4507
0.0000
0.00004
LO
Composite Results:
i
"i
ni
1
(HC)
0.1984
1.1674
2
(CO)
2.0740
19.9562
3
(N0x)
0.1230
2.3393
-------
14
emission-versus-mileage regression lines at 50,000 miles (see figure 1).
The upper limit on evaporative emissions was taken to be equal to the best
estimate (since the best estimate resulted in an in-use value well below
the standard), while the lower limit used the in-use evaporative rate
actually experienced in 1975.
3.1.3 Case 3: Standards Plus Certification with Parameter Adjustment
To represent the effect of EPA parameter adjustment on the effective-
ness of certification, the mileage regressions described for case 2 were
repeated on only those 1975 vehicles in the EF data base which were within
manufacturer specifications on idle CO, idle RPM, and timing.1 The re-
gression results are presented in table 4. The high and low estimates
again were based on 95 percent confidence intervals around each of these
regression lines at 50,000 miles. Under the assumption that parameter
adjustment would not affect evaporative emissions, the case 2 estimates
for evaporative effectiveness were used for case 3.
3.1.4 Cases 4 and 5: Standards Plus Selective Enforcement Audit
The effectiveness of Selective Enforcement Audit (SEA) was estimated
with the use of 1975 assembly-line data provided voluntarily by the manu-
facturers to EPA1s Mobile Source Enforcement Division. Using these data
LA vehicle was considered to be within manufacturer specifications if its
idle CO was less than 0.5 percent, and if it was within 150 RPM of the
manufacturer's specification for idle RPM and within 2 degrees of specifi-
cation for timing. There were 314 such vehicles in the data base. Note
that the requirement that idle CO be less than 0.5 percent eliminated the
need for the kind of composite regression used for case 2.
-------
Emission rate
for a given
FIGURE 1: DETERMINATION OF BEST ESTIMATE AND RANGE
OF VALUES FOR AVERAGE EMISSION RATES FOR
CASE 2 (STANDARDS PLUS CERTIFICATION)
-------
16
TABLE 4: REGRESSION ANALYSIS OF IN-USE EMISSION RATES
AS A FUNCTION OF MILEAGE FOR CASE 3
Relation: e.. (m) = vi.j m + (i = 1> 2, 3)
where
e. (m) = average in-use emissions of pollutant i by a vehicle
at mileage m;
m = vehicle mileage, in tens of thousands; and
u.j > n.j = regression coefficients
Results:
i
ni
R2
Value
Significance
Value
Significance
1 (HC)
0.21532
0.0000
0.76435
0.0000
0.06059
2 (CO)
1.1438
0.1510
12.071
0.0000
0.00660
3 (N0X)
0.23168
0.0010
2.2729
0.0000
0.03395
-------
17
and 1975 EF data for 282 vehicles, regression analyses (summarized in
table 5) were performed to estimate the relationships between assembly-
line emission levels (with deterioration factors applied) and levels of
vehicles in use (converted to 50,000 mile levels using the mileage regres-
sion described in section 3.1.2). The best estimate of effectiveness then
assumed that manufacturer design, production engineering, and quality
control would result in vehicles which just met the standards at the
assembly line. The regression relationships were used to convert these
assembly-line levels to levels for vehicles in use, and these in-use levels
were further reduced to account for the effect of rejections by SEA. This
effect was determined by assuming that an SEA failure would result in the
emissions of failing vehicles in a class being reduced to the average
emission levels of passing vehicles in the same class. The voluntarily
submitted data were then used to estimate the average saving per failing
class1 corresponding to this assumption, and these savings were converted
to in-use reductions using the regression relations given in table 5. To
apply these reductions to overall emissions, it was assumed that
(1) 20 test orders per year would be conducted;
(2) 50 percent of vehicles in a tested class would have been sold (and
not subject to the reductions2) by the time of the test order;
(3) 100 percent of classes would fail for the case 4 estimate, while
75 percent would fail for the case 5 estimate; and
1Failing classes were identified in the data by assuming a 40 percent AQL
for each pollutant.
2In cases Involving recall as well as SEA, 1t was assumed that these
vehicles would be recalled (see section 3.1.10).
-------
18
TABLE 5: REGRESSION ANALYSIS OF IN-USE
EMISSIONS AS A FUNCTION OF
ASSEMBLY LINE EMISSIONS
Re lation:
where
ij
ij
Zij
Yi,6i
Y . Z. . + 5 •
1 1J 1
-------
19
(4) the average sales of a falling class would equal the average for
all classes included in the voluntarily submitted data base.
To develop a lower bound on this estimate, it was assumed that SEA
would create no deterrent, so that the only benefit of SEA would be in the
form of a reduction in the emissions of vehicles in failed classes (i.e.,
precontrol levels would prevail, except in failed classes). An upper
bound on the estimate was developed by assuming that, in addition to the
high-estimate deterrent resulting from the existence of standards (see sec-
tion 3.1.1), a further deterrent from SEA would cause manufacturers to
reduce the emissions of vehicles, in classes which would otherwise fail to
the point at which just 40 percent of the vehicles in each class would
fail.1 In addition to this deterrent effectiveness, it was assumed that
random fluctuations in measured emissions would still result in SEA failures,
so effectiveness due to reduction in emissions of vehicles in failed
classes was also included in the high estimate.
Best, low, and high estimates of the effectiveness of SEA in control-
ling evaporative emissions (assuming the use of a hot soak SHED test) were
computed in a similar manner to the computation of exhaust emission effec-
tiveness. However, lack of assembly line data for evaporative emissions
required assuming that
(1) slippage of evaporative emission from assembly line to in-use
conditions is in the same proportion as the slippage from certifi-
cation to in-use conditions and
1This assumes 40 percent to be the AQL for SEA, and that manufacturers are
deterred to the point of reducing emissions to this level before test
orders are conducted.
-------
20
(2) the percentage reduction in evaporative emissions resulting from
failure of a class on an SEA test order is equal to the average
percentage reduction assumed for exhaust emissions.
3.1.5 Case 6: Standards Plus Inspection and Maintenance (Diagnostic)
The best estimate for case 6 assumes that a diagnostic inspection
consists of checks on idle limiter caps, idle RPM, timing, EGR, and hoses.
Restorative Maintenance (R.M) data were examined to estimate percent reduc-
tions from in-use levels resulting from bringing these items within manu-
facturer specifications. There were only 26 vehicles in the RM data base
which met the conditions of this analysis. Of these vehicles, six were
within manufacturer specifications on the above parameters. The average
emissions of these six vehicles were lower than the average for the entire
26 vehicles by 56 percent for HC, 69 percent for CO, and 38 percent for
NO . These percentages were applied to actual 1975 emission levels, and
A
the results were then adjusted for the fact that case 6 assumes no certi-
fication program exists. This adjustment was made by'adding the difference
between case 1 and case 2 emissions to each result.1 To combine these
first-year benefits with the effects of deterioration and multiple annual
inspections, yearly deterioration factors developed from a run of EPA's
inspection and maintenance (IM) simulation [EPA, 1977] were applied to the
first-year emissions predicted from the RM data to produce an estimate of
total emissions throughout a vehicle's life. In particular, the following
:This and other adjustments for the absence of a certification program
assume that the effect of certification on emissions is constant over
all mileages.
-------
21
relationship was used to predict the deterioration factor (DF) in a given
year:
DF for year n = (tons emitted in year n)/(miles driven in year n)
(tons emitted in year 1)/(miles driven in year 1)
where all values on the right-hand side of the equal sign were taken from
the simulation.
To develop a high effectiveness estimate, the above process was re-
peated under the assumption that the diagnosis included everything in the
RM diagnostic check. Because the results for the best and high estimates
were based on data for a very small set of vehicles, the low estimate
assumed that no benefit would accrue from the program. For all three
estimates, the diagnostic checks were assumed to have no direct effect on
evaporative emissions, but only the deterrent effect resulting from the
existence of standards.1
3.1.6 Cases 7 and 8; Standards Plus Inspection and Maintenance (Testing)
The best effectiveness estimates for cases 7a, 7b, 8a, and 8b were
developed using EPA's IM simulation [EPA, 1977]. (The version of the simu-
lation used was that available in July 1977.) Input to the simulation
consisted of 1975 EF data adjusted for the assumed absence of certifica-
tion. (These adjustments were made by adding the difference between
case 1 and case 2 emissions to the measured emissions of each vehicle in
Although a diagnostic inspection could produce an evaporative benefit,
data were not available to evaluate the magnitude of such a benefit.
-------
22
the data base.) Additional assumptions and procedures used in conjunction
with the simulation were that:
(1) the idle test was used for cases 7a and 3a, while a loaded test
was used in cases 7b and 8b (the loaded test used was the Federal
Short Cycle);
(2) the loaded test (but not the idle test) was used to measure NO
X
emissions;
(3) outpoints were set to force a first-year failure rate of 30
percent in cases 7a and 7b, and 50 percent in cases 8a and 8b;
and
(4) failure of an inspection would result in reduction of emission
levels to the FTP equivalent of the short test standards.
Equations used to relate short test results to FTP levels in the simula-
tion appear in table 6. The short test cutpoints used and the resultant
benefits are summarized in table 7.
Effectiveness of IM at controlling evaporative emissions assumed the
use of a hot soak SHED procedure in all cases. Because evaporative emis-
sions are assumed not to deteriorate, it was not necessary to use the IM
simulation to estimate effectiveness. Instead, the evaporative cutpoint
was assumed to be set at the same fraction above the standard as was the
cutpoint for exhaust HC. A sample of evaporative test results was then
approximated using 1978 certification data in order to predict a fraction
of vehicles which would fail, and therefore have their emissions lowered to
the cutpoint. The result was a 4 percent failure rate for cases 7a and 8a,
and a 6 percent rate for cases 7b and 8b. The resulting emission reduction
was applied to the case 1 (no certification) emission levels.
-------
23
TABLE 6: REGRESSION EQUATIONS USED TO RELATE
SHORT TEST RESULTS TO FTP LEVELS IN
THE IM SIMULATION
TEST
POLLUTANT
RELATION
HC
FTP
HC
3
.0035396 • ST HC + .87178
IDLE1
CO
FTP
CO
=
4.7936 • ST CO + 15.668
HC
FTP
HC
=
.84449 • ST HC + .53378
LOADED2
CO
FTP
CO
1.1067 • ST CO + 7.4225
NOx
FTP
3
LO
t—
o
o
CO
•
+
X
§
1—•
to
•
LD
cr>
CO
CO
•
2These Idle test equations were used with cases 7a, 8a, 12, 14,
15, 16, and 18.
2These loaded test equations were used with cases 7b and 8b.
-------
24
TABLE 7: SHORT TEST OUTPOINTS AND CORRESPONDING PERCENT
SAVINGS FROM APPLICATION OF IM SIMULATION
CASE
SHORT
TEST CUTPOINTS1
RESULT2 (PERCENT BENEFIT)
HC
CO
N0X
HC
CO
N0X
7a
478.0
4.8
-
29
32
-
7b
2.4
30.6
5.3
30
30
1
3a
348.0
3.5
-
38
40
8b
1.9
24.1
CM
38
39
4
12
55.0
0.5
-
41
30
-
14
41.2
0.1
-
42
35
-
15
"255.4
2.5
-
38
39
-
16
21.2
0.1
-
45
36
-
18
255.4
2.5
-
38
40
-
^dle test cutpoints are in units of parts per million for HC
and percent for CO. Loaded test cutpoints are all in units
of grams per mile.
2Results for cases 14, 16, and 18 also include benefits from
recall, which were computed in conjunction with the IM simulation
computations.
-------
25
The low estimate in these cases assumed that IM would have no deter-
rent effect (so that vehicles with uncontrolled emissions would be subject
to IM) and that deterioration between inspections would occur more rapidly
than in the best estimate case. (Figure Z illustrates the two alternative
assumptions regarding deterioration.) To approximate this effect, one-
half of first-year percentage reductions from [EPA, 1977] were applied to
uncontrolled emission levels to develop a HC and CO benefit in grams per
mile. Low NO benefits in the loaded-mode cases were taken to be zero.
A
Low evaporative benefits in all cases applied the percentage reduction in
the best estimate to uncontrolled levels.
High estimates did not use the simulation but assumed that average
emission levels met the FTP standards at 50,000 miles in the field.1
3.1.7 Cases 9 and 10: Standards Plus Recall
The best estimate for recall was developed by matching vehicles in
the EF data base with classes identified on EPA's current list of potential
recalls [MSED, 1977]. (Emissions in the EF data base were first adjusted
for the absence of a certification program, as described earlier.) All
vehicles so identified were assumed to be recalled at 25,000 miles.2 Those
vehicles responding to the recall were assumed to have their emissions for
pollutants causing the recall reduced to the average emissions of unaffect-
ed cars at 25,000 miles. (In case 9, 40 percent of all vehicles were
however, see section 4.0 for a discussion of situations 1n which these
(and other) high estimates were revised to maintain internal consistency
among the estimates.
2Note that this use of the EF data base (as well as other uses) assumes
that the data consist of a sales-weighted sample from the national
population.
-------
EMISSION RATE
FOR A GIVEN
POLLUTANT
(GRAMS PER MILE)
EMISSIONS
WITHOUT IM
THIS AREA EQUALS
ONE-HALF FIRST-YEAR BENEFIT
UNDER BEST-ESTIMATE
ASSUMPTION
FTP STANDARD^
V FTP EQUIVALENT OF
^— SHORT TEST
CUTPOINT
1
—1
1
,
1
1
1 1
FIGURE 2: ALTERNATIVE DETERIORATION ASSUMPTIONS FOR
BEST ESTIMATES AND LOW ESTIMATES OF IM
EFFECTIVENESS FOR A SINGLE VEHICLE
8
VEHICLE AGE
(YEARS)
-------
27
assumed to respond, while case 10 assumed a 70 percent response rate.)
Deterioration after a recall was assumed to occur at the same rate as be-
fore the recall. (See figure 3 for an illustration of the assumed
emission rates of a recalled vehicle.) Deterioration rates were developed
from mileage regressions with 1975 EF data (see section 3.1.2).
Low effectiveness estimates assumed no deterrent effect, and assumed
that recalls would reduce uncontrolled emission levels to the same level
to which they were reduced in the best estimate. The same vehicles were
assumed to be recalled as for the best estimate. High effectiveness
estimates assumed that standards plus the threat of recall would cause
average emission levels to be at the FTP standards at 50,000 miles in the
field.- As a result, no recall actions were assumed necessary.
3.1.8 Case 11: Standards Plus Certification with Parameter Adjustment
Plus Selective Enforcement Audit
The best and low estimates for this case began with emission rates
for certification with parameter adjustment (see section 3.1.3) and sub-
tracted an average benefit associated with SEA failures. This latter
benefit was developed from the manufacturer-submitted assembly-line data
describing the effectiveness of an SEA program in the presence of certifi-
cation, and was converted to an in-use benefit using the regression
relationships developed for cases 4 and 5 (see section 3.1.4). It was
assumed that, even in the presence of certification with paramenter adjust-
ment, EPA would still be able to identify 20 classes which would fail an
-------
EMISSION RATE
FOR A GIVEN EMISSIONS
FIGURE 3: EFFECT OF RECALL ON EMISSIONS OF A RECALLED VEHICLE
-------
29
SEA test order.1 The high estimates used the same approach but assumed
a further SEA deterrent which would cause manufacturers to reduce emissions
in potentially-failing classes to the point at which just 40 percent of
the vehicles in each such class would fail (see section 3.1.4).
3.1.9 Case 12: Standards Plus Certification with Parameter Adjustment Plus
Inspection and Maintenance (Testing)
The best estimate for exhaust emissions in case 12 used 1975 EF data
for vehicles within manufacturer specifications on idle CO, idle RPM, and
timing as input to the IM simulation. (See tables 6 and 7 for a summary
of the simulation run.) The low estimate applied half the first year
percent emission reductions from this simulation run to the low estimate for
certification with parameter adjustment (case 3, see section 3.1.3).2
For evaporative emissions, the best estimate was equal to the best
estimate for certification with parameter adjustment (case 3). Since the
results for this case were well below the derived evaporative standard, and
because of the assumption that evaporative emissions do not deteriorate
with mileage, it was assumed few or no IM failures would occur. The low
evaporative estimate applied the low IM benefit derived for case 7a to the
low estimate for evaporative emissions in case 3.
All high estimates assumed that emissions met FTP standards at 50,000
miles in the field.
iNote that this SEA benefit was not applied to particular vehicles in the
EF data base, but was computed separately and the result used to reduce
overall emissions in the field.
2See section 3.1.6 and figure 2 for an indication of the rationale of this
approximation.
-------
30
3.1.10 Case 13: Standards Plus Certification with Parameter Adjustment
Plus Selective Enforcement Audit Plus Recail
The best estimate in this case began with the 1975 EF data for vehicles
within manufacturer specifications on idle CO, idle RPM, and timing. These
emission rates were adjusted for an average in-use benefit associated with SEA
failures1 (see sections 3.1.4 and 3.1.8). Vehicles in this data base were
then matched with classes identified in [MSED, 1977], and the benefit of re-
call of these vehicles was computed (see section 3.1.7). The net emission
rates including this benefit provided the case 13 best estimate. Evaporative
emissions were assumed to equal the case 3 level minus a reduction due to SEA.
(Recall was assumed to have no direct effect on these emissions, since there
are no evaporative recalls in [MSED, 1977]. In the future, of course, more
stringent evaporative emission standards may result in some evaporative recalls,
Low estimates were developed by subtracting average effects due to SEA
failures (see section 3.1.4) and recall (see section 3.1.7) from the low
estimates of emission rates for certification with parameter adjustment (see
section 3.1.3). High estimates assumed that all vehicles meet FTP standards
on all pollutants at 50,000 miles in the field.
3.1.11 Cases 14 and 16: Standards Plus Certification with Parameter
Adjustment Plus Selective Enforcement Audit Plus Inspection
and Maintenance (Testing) Plus RecalT
The best estimate in these cases used a modified version of the IM
simulation which included a dynamic representation of the effects of a
because recall was also present as a compliance technique in this case,
95 percent of the vehicles sold before any SEA test order were assumed
to receive the SEA benefit, as a result of responding to a recall. This
also is true of cases 14, 16, 17 and 18. (The 95 percent response rate
reflects the fact that the vehicles are new at the time of the recall and
their owners are therefore likely to be located and to respond.)
-------
31
recall campaign. Recalls were assumed to occur after two years of the IM
program. The effects of recall were computed as described in section
3.1.7. Input to the simulation consisted of 1975 EF data for vehicles
within manufacturer specifications on idle CO, idle RPM, and timing.1
These input values were modified to account for the average effect of SEA
failures (see sections 3.1.4 and 3.1.8). Recalls were assumed to affect
those vehicles in this data base corresponding to the classes identified
in [MSED, 1977]. The best estimate for evaporative emissions assumed no
benefit for IM or recall, since emissions in the field were already pre-
dicted to be well below the standard and since no evaporative recalls
appear in [MSED, 1977],
Low estimates were approximated by applying half the reduction computed
for the first-year best estimate (in percent) to the low estimate for this
situation without IM and recall (represented in case 11). The low estimate
for evaporative emissions applied the low IM benefit derived for case 7a to
the low estimate for evaporative emissions In case 3. High estimates for
all pollutants assumed FTP standards were met in the field at 50,000 miles.
3.1.12 Case 15: Standards Plus Selective Enforcement Audit Plus Inspection
and Maintenance (Testing)
For the best estimate 1n this case the IM simulation was run using
as input the 1975 EF data adjusted for the absence of certification (by
adding the difference between average emissions 1n case 1 and case 2 to
1See tables 6 and 7 for further information concerning the simulation run.
-------
32
each, data item), and further adjusted for the average effects of SEA (see
sections 3.1.4 and 3.1.8).1 No NO benefit from IM was assumed. Evapora-
X
tive benefits used the reduction computed for case 7a (standards plus IM)
and applied it to the case 4 result (standards plus SEA).
The low estimate applied half the first year percentage emission reduc-
tions computed from the simulation to precontrol emission levels, adjusted
for the effect of SEA. The low evaporative estimate applied the percentage
reduction achieved in the best estimate to the case 4 low results. All
high estimates assumed that cars meet the FTP standards at 50,000 miles
in the field.
3.1.13 Case 17: Standards Plus Selective Enforcement Audit Plus Recall
The best estimate in this case again used 1975 EF data adjusted for
the absence of certification and for the direct benefit of SEA (see sections
3.1.4 and 3.1.8). The benefit of recall to those vehicles in this data base
which matched classes identified in [MSED, 1977] was then computed. The
net emission rates including this benefit provided the case 17 best esti-
mate. Evaporative emissions were assumed to equal the case 1 level minus
a reduction due to SEA. (Recall is assumed to have no direct effect on these
emissions, since there are no evaporative recalls on the MSED list.)
Low estimates were developed by subtracting average effects due to
SEA failures (see section 3.1.4) and recall (see section 3.1.7) from the
precontrol emission levels. High estimates assumed that all vehicles meet
FTP standards on all pollutants at 50,000 miles in the field.
^ee tables 6 and 7 for further information concerning the simulation run.
-------
33
3.1.14 Case 18; Standards Plus Selective Enforcement Audit Plus Inspection
And Maintenance (Testing) Plus Recall
For case 18, the best estimate again used 1975 EF data adjusted for
the absence of certification and for the average direct benefit of SEA.
These data served as input to the IM simulation as modified to include the
effects of recall.1 Recalls were again assumed to affect those vehicles
in this data base corresponding to the classes identified in [MSED, 1977].
No NO benefit from IM was assumed. Evaporative benefits used the reduction
computed for case 7a (standards plus IM) and applied it to the case 4
result (standards plus SEA).
The low estimate for HC and CO applied half the first year percentage re-
ductions computed from the simulation to precontrol emission levels, adjusted
for the effect of SEA. The low evaporative estimate applied the percentage
reduction achieved in the best estimate to the case 4 low evaporative
estimate. All high estimates assumed that cars meet the FTP standards at
50,000 miles in the field.
3.1.15 Case 19: Standards Plus Certification Plus Inspection and
Maintenance (TestlnqT
The best estimate for this case applied the percentage benefits of IM
predicted by [EPA, 1977] to the results of case 2 (standards plus certifi-
cation). The best estimate for evaporative emissions applied the percent
reduction computed for case 7a to the best case 2 estimate.
The low exhaust estimate applied half the first year percentage re-
ductions from [EPA, 1977] to the case 2 estimates. The low evaporative
xSee tables 6 and 7 for further information concerning the simulation run.
-------
34
estimate again applied the case 4 reduction to the low case 2 estimate.
All high estimates assumed that cars meet the FTP standards at 50,000
miles in the field.
3.2 Coat Estimates
Development of cost estimates for the 21 cases evaluated in this
study involved developing best, high, and low cost estimates for each
compliance technique separately. These costs were then added for the com-
bination of compliance techniques present in a given case. Note that the
high and low cost estimates for a given technique are not necessarily the
greatest and least costs possible for that technique, but rather repre-
sent the costs most likely to accrue given the high or low benefit accrues.
The following paragraphs describe the estimates used for government and
total costs for each compliance technique. A summary of these cost
estimates appears in table 8.
3.2.1 Standards
For all best and high effectiveness estimates, promulgation of emission
standards was assumed to result in the installation of emission controls.
Therefore, the cost associated with the existence of standards was assumed
to be the manufacturer design and production costs required to install
these controls. Based on recent EPA estimates, an average cost of $170
per vehicle was assumed to be incurred for this purpose [EPA, undated].
In cases in which the low effectiveness estimate assumed that no such
controls would be installed, this cost was not assessed. In- all cases,
government costs were assumed to equal zero.
-------
35
TABLE 8: SUMMARY OF COSTS USED IN THE ANALYSIS
(MILLIONS OF DOLLARS)
COMPLIANCE
TECHNIQUE
BEST ESTIMATE
GOV'T.
COSTS
OTHER
COSTS
LOW ESTIMATE
GOV'T.
COSTS
OTHER
COSTS
HIGH ESTIMATE
GOV'T.
COSTS
OTHER
COSTS
[Standards
1700
.1700 or O1
1700
Certification or
Certification with
l Parameter Adjustment
3.26
60
3.26
50
3.26
70
Selective Enforcement
Audit
.70
19.1
.70
16.6
.70
21.6
Inspection and
Maintenance (All
Versions)
400
400
400
24.41 or
f42.712
Recall,
.52
24.41 or
42.77 2
,52
24.41 or
•42.712
,52
Ssumed to be 1700 when in presence of certification or certification with parameter
<3justment, and 0 otherwise.
^sumed to be 24.41 for a response rate of 40 percent and 42.71 for a response rate of
0 percent.
-------
36.
3.2.2 Certification
The costs of certification (and certification with parameter adjustment)
included government and manufacturer costs associated with the operation of
the certification process. These costs were based on estimates provided by
[Auerbach, 1977]. Certification process cost to the government was estimated
at $3.26 million, while manufacturer costs were assumed to range between
$50 million and $70 million. The $50 million estimate was used with low
effectiveness estimates, the $70 million estimate was used with high estimates,
and their average was used as the best estimate of manufacturer costs.
3.2.3 Selective Enforcement Audit
Costs of SEA were based on [Train, 1976], which states that "The SEA
program is estimated to cost about $0.7 million per year for the government,
about $1.6 million per year for company testing, and about $15-20 million
per year for modifications to automobiles." For the latter range of estimates,
the $15 million figure was used with low effectiveness estimates, the $20
million figure was used with high effectiveness estimates, and their average
was used as a best estimate. It should be pointed out that these costs
are not completely consistent with the assumptions of this analysis, since
the $15-20 million estimate apparently is based on the assumption that
vehicle design changes will be implemented to eliminate all SEA test failures.
(The current analysis assumes that test failures actually occur.) However,
these costs were used here because they represent EPA's current best estimate
of the cost of SEA.
-------
37
3.2.4 Inspection and Maintenance
Costs for all versions of inspection and maintenance were based on
[Walsh, 1976]. These estimates indicate that the average cost of an
inspection is approximately $5 per car. Because an average car will be
subject to eight annual inspections during its 100,000-mile life, a total
lifetime cost of $40 per car was assumed. The estimates in [Walsh, 1976]
also indicate that vehicle repair costs after an IM failure will be
approximately equal to the fuel economy savings resulting from the repair.
Therefore, no repair costs or fuel economy savings were explicitly included
in these estimates. No federal government costs were assumed to be incurred
for IM.. Note that these estimates were used on the assumption that differences
in costs of idle mode testing, loaded mode testing, and parameter inspection
are negligible.
3.2.5 Recall
The costs of recall were assumed to include:
(1) government costs for operating a recall program one year1 (assumed
to equal $200,000),
(2) government testing costs (assumed to equal $1000 per car tested,
where 20 cars are assumed to be tested for each recall action
considered), and
xThis estimate assumes that the cost of one year's operation of the program
reasonably approximates that portion of program costs associated with
recalls of vehicles for a single model year.
-------
38
(3) manufacturer costs to fix recalled cars (assumed to equal $12
per car fixed).
Note that high effectiveness estimates assume that no recall actions are
found to be necessary. Therefore, the manufacturer cost for repairing
recalled cars is zero. However, this repair cost is still assessed as an
approximation of the additional cost to manufacturers of building the cars
in such a way that the recalls become unnecessary.
-------
4.0 RESULTS
Effectiveness and cost results for all cases evaluated appear in tables
9 through 29.1 In generating some of the low and high estimates appearing
in these tables, effectiveness values were sometimes adjusted to maintain
internal consistency among the assumptions used to generate the estimates.2
These adjustments were necessary because assumptions and procedures used to
compute preliminary values for some of the low and high estimates were found
to be inconsistent when the calculations were actually made. For example,
the assumption initially used to compute the high estimate for cases 7 and 8
(standards plus IM) was that vehicles would meet the FTP standards in the
field. However, the high estimate for NO emissions in case 1 (standards
A
alone) was an emission rate which was below the FTP standards for vehicles
in the field. Since it seemed unreasonable that an upper bound on the
effectiveness of standards plus IM would be less than an upper bound on the
effectiveness of standards alone, the case 7 and 8 high estimate for N0x was
adjusted to equal the case 1 high estimate.
The following points may assist in interpreting and estimating the
validity of the effectiveness values appearing in the tables:
(1) All effectiveness estimates were based on available data. Because
certification is the compliance technique which has been in effect
longest (and therefore has the most data accumulated on it), estimates
of the effectiveness of certification are probably the most accurate.
iThese tables are located at the end of section 4.0.
2Note that these adjustments were required for low and high estimates only.
None of the best estimates was changed. All adjusted values are marked
with asterisks in the tables.
-------
40
(2) Deterrent effects for cases which have never actually been
implemented are based on assumptions which are believed to be
reasonable, but for which supporting data were not always available.
This is especially true of effectiveness estimates associated with
the promulgation of standards in the absence of a certification
program.
(3) Because of the limited amount of SHED data available and the laqk
of a meaningful evaporative emission standard actually in existence
for 1975 vehicles, results for evaporative emissions are necessarily
based on limited data and several "reasonable" assumptions. These
results should, therefore, be interpreted carefully.
(4) The results for case 6 (standards plus diagnostic IM) are based on
results for a very small number of vehicles and, therefore, should
be interpreted with caution.
(5) The results for all cases involving IM testing are sensitive to
the specific combinations of HC, CO, and NO cutpoints used to
A
achieve the required first-year failure rates. Other cutpoints
could result in the same initial failure rate but a somewhat
different effectiveness estimate.
(6) Results for cases involving the loaded-mode IM test (cases 7b
and 8b) show little additional benefit over the corresponding
idle-mode cases (cases 7a and 8a, respectively). This is
apparently because all these cases assume IM is implemented in
the absence of a certification process, so the resulting in-use
emissions are quite high before application of IM. As a result,
the greater precision of the loaded-mode test is, in a sense,
"wasted" on vehicles which are likely to fail any reasonable test
used on them.
-------
41
(7) Note that certification with parameter adjustment (e.g., case 3)
appears to produce an increase in N0x emissions concomitant with
the reduction in HC and CO emissions realized over either case 1
or case 2.
Tables 30 and 311 present cost effectiveness estimates for all
cases in units of either government dollars or total dollars spent per ton
of pollutant removed. These results were computed for a given case by allo-
cating costs equally among the four pollutants and dividing the result by
the effectiveness values. Values in these tables can be misleading because
they result from a somewhat arbitrary allocation of costs among pollutants,2
and because they mask the actual magnitude of effectiveness that can be
achieved and costs required to implement a given case. Therefore, cases with
low cost effectiveness are not necessarily the most desirable cases in the
context of budget constraints and/or specific pollution reduction goals.3
Furthermore, these values cannot be used directly to estimate the marginal
cost effectiveness of adding a new compliance technique to an existing
compliance assurance program.
^hese tables are located at the end of section 4.0.
2Although any such allocation is likely to seem arbitrary, it might be
possible to allocate costs in proportion to the relative effectiveness
of a given case at reducing the emissions of each pollutant, or in pro-
portion to subjective estimates of the relative degree to which each
pollutant is a "target" of the compliance techniques used in a given case.
3For example, promulgation of standards alone (case 1) produces a reduction
in emissions at essentially no cost to the government. It therefore has
a cost effectiveness of zero and appears highly desirable from a cost
effectiveness standpoint. However, it is generally considered to be an
unacceptable alternative because it does not result in a sufficient
reduction in vehicle emissions.
-------
TABLE 9: RESULTS FOR CASE 1 -
STANDARDS ONLY
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
5.32
46.25
0.07
1.73
0
1700
LOU
0
0
0
0
0
0
HIGH
6.40
70.76
1.22
3.09
0
1700
-------
TABLE 10: RESULTS FOR CASE 2 - STANDARDS
PLUS CERTIFICATION
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVEO)
CC
(MILLIONS G
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
5.94
62.17
0.50
3.09
3.26
1763.26
LOW
5.48
53.58
-.23
1.73
3.26
1753.26
HIGH
6.40
70.76
1.22
3.09
3.26
1773.26
-------
TABLE 11: RESULTS FOR CASE 3 - STANDARDS PLUS
CERTIFICATION WITH PARAMETER ADJUSTMENT
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOVT.
TOTAL
BEST
6.29
75.97
-.03
3.09
3.26
1763.26
LOW
5.88
69.23
-.63
1.73
3.26
1753.26
HIGH
6.70
82.72
.56
3.09
3.26
1773.26
-------
TABLE 12: RESULTS FOR CASE 4 - STANDARDS PLUS
SELECTIVE ENFORCEMENT AUDIT (100%
REJECTION RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
COST
(MILLIONS OF DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
5.74
58.16
0.09
1.86
.70
1719.80
LOU
1.78
20.80
0.08
0.54
.70
1717.30
HIGH
6.49
76.09
1.24
3.13
.70
1722.30
-------
TABLE 13: RESULTS FOR CASE 5 - STANDARDS PLUS
SELECTIVE ENFORCEMENT AUDIT (75%
REJECTION RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
5.72
57.50
0.08
1.83
.70
1719.80
LOU
1.47
17.69
0.06
0.41
.70
1717.30
HIGH
6.48
75.71
1.24
3.13
.70
1722.30
-------
TABLE 14: RESULTS FOR CASE 6 - STANDARDS PLUS
INSPECTION AND MAINTENANCE (DIAGNOSTIC)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
COST
(MILLIONS OF DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
6.74
68.01
0.19
1.73
0
2100
LOW
0
0
0
0
0
400
HIGH
6.83
70.76
1.22*
3.09
0
2100
~Adjusted for consistency.
-------
TABLE 15: RESULTS FOR CASE 7a - STANDARDS PLUS
IDLE INSPECTION AND MAINTENANCE (30%
FAILURE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
6.19
62.03
0.07
1.80
0
2100
LOW
0.87
20.57
0
0.12
0
400
HIGH
6.66
79.04
1.22*
3.09*
0
2100
~Adjusted for consistency.
-------
TABLE 16: RESULTS FOR CASE 7b - STANDARDS PLUS LOADE
LOADED INSPECTION AND MAINTENANCE (30%
FAILURE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
COST
(MILLIONS OF DOLLARS)
H€
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
6.22
61.04
0.10
1.80
0
2100
LOU
0.87
20.57
0
0.12
0
400
HIGH
6,66
79.04
1.22*
3.09*
0
2100
~Adjusted for consistency.
-------
TABLE 17: RESULTS FOR CASE 8a - STANDARDS PLUS
IDLE INSPECTION AND MAINTENANCE (50%
FAILURE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CO
(MILLIONS 0
ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
6.46
65.97
0.07
1.80
0
2100
LOU
0.87
20.57
0
0.12
0
400
HIGH
6.66
79.04
1.22*
3.09*
0
2100
c_n
O
~Adjusted for consistency.
-------
TABLE 18: RESULTS FOR CASE 8b - STANDARDS PLUS
LOADED INSPECTION AND MAINTENANCE (50%
FAILURE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
6.46
65.48
0.21
1.80
0
2100
LOW
0.87
20.57
0
0.12
0
400
HIGH
6.66
79.04
1.22*
3.09*
0
2100
~Adjusted for consistency.
-------
TABLE: 19: RESULTS FOR CASE 9 - STANDARDS PLUS
RECALL (40% RESPONSE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CO
(MILLIONS C
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
5,40
47.45
0.13
1.73
.52
1724.93
LOU
l.H
11.67
0.06
0
.52
24.93
HIGH
6.66
79.04
1.22*
3.09*
.52
1724.93
*Adjusted for consistency.
-------
TABLE 20: RESULTS FOR CASE 10 - STANDARDS
PLUS RECALL (70% RESPONSE RATE)
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
cc
(MILLIONS G
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
5.45
49.15
0.19
1.73
.52
1743.23
LOU
1.61
14.16
0.10
0
.52
43.23
HIGH
6.66
79.04
1.22*
3.09*
.52
1743.23
CJ1
CO
*Adjusted for consistency
-------
TABLE 21: RESULTS FOR CASE 11 - STANDARDS PLUS
CERTIFICATION WITH PARAMETER ADJUSTMENT
PLUS SELECTIVE ENFORCEMENT AUDIT
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
6.32
77.51
-.03
3.09
3.96
1783.06
LOU
5.91
70.77
-.62
1.73
3.96
1770.56
HIGH
6.79
88.05
1.24*
3.13*
3.96
1795.56
~Adjusted for consistency.
-------
TABLE 22: RESULTS FOR CASE 12 - STANDARDS PLUS
CERTIFICATION WITH PARAMETER ADJUSTMENT
PLUS INSPECTION AND MAINTENANCE
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS 0
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
7.12
81.85
-.03
3.09
3.26
2163.26
LOU
6.05
70.28
-.63
1.80
3.26
2153.26
HIGH
7.12*
82.72*
,56*
3.09*
3.26
2173.26
cn
en
*Adjusted for consistency.
-------
TABLE 23: RESULTS FOR CASE 13 - STANDARDS PLUS
CERTIFICATION WITH PARAMETER ADJUSTMENT
PLUS SELECTIVE ENFORCEMENT AUDIT PLUS
RECALL
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
N0X
EVAP.
GOV'T.
TOTAL
BEST
6.53
78.03
0.38
3.09
4.48
1826.29
LOU
6.20
72.45
-0.14
1.73
4.48
1813.79
HIGH
6.79*
88.05*
1.24*
3. 13*
4.48
1838.79
cn
a*
~Adjusted for consistency.
-------
TABLE 24: RESULTS FOR CASE 14 - STANDARDS PLUS CERTIFICATION
WITH PARAMETER ADJUSTMENT PLUS SELECTIVE ENFORCEMENT
AUDIT PLUS INSPECTION AMD MAINTENANCE (30% FAILURE
RATE) PLUS RECALL
ESTIMATE
EFFECTIVENESS
(MILLIONS OF TONS OF POLLUTANT SAVED)
COST
(MILLIONS OF DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
7.16
83.83
0.38
3.09
4.48
2226.29
LOW
6.20*
72.45*
-0.14
1.80
4.48
2213.79
HIGH
7.16*
88.05*
1.24*
3.13*
4.48
2238.79
~Adjusted for consistency.
-------
TABLE 25: RESULTS FOR CASE 15 - STANDARDS PLUS
SELECTIVE ENFORCEMENT AUDIT PLUS
INSPECTION AND MAINTENANCE
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
6.72
72.75
0.09
1.89
0.70
2119.80
LOU
2.20
29.40
0.08
0.60
0.70
417.30
HIGH
6.72*
79.04
1.24*
3.09*
0.70
2122.30
CJ1
00
*Adjusted for consistency
-------
TABLE 26: RESULTS FOR CASE 16 - STANDARDS PLUS CERTIFICATION
WITH PARAMETER ADJUSTMENT PLUS SELECTIVE ENFORCEMENT
AUDIT PLUS INSPECTION AND MAINTENANCE (50% FAILURE
RATE) PLUS RECALL
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
7.22
84.01
0.38
3.09
4.48
2226.29
LOU
6.20
72.63
-0.14
1.80
4.48
2213.79
HIGH
7.22
88.05
1.24*
3.13*
4.48
2238.79
~Adjusted for consistency.
-------
TABLE 27: RESULTS FOR CASE 17 - STANDARDS PLUS
SELECTIVE ENFORCEMENT AUDIT PLUS
RECALL
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS 0
ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
5.90
62.31
0.23
1.86
1.22
1763.03
LOW
2.67
26.72
0.18
0.54
1.22
60.53
HIGH
6.66
79.04
1.24*
3.09*
1.22
1765.53
O
*Adjusted for consistency.
-------
TABLE 28: RESULTS FOR CASE 18 - STANDARDS PLUS
SELECTIVE ENFORCEMENT AUDIT PLUS
INSPECTION AND MAINTENANCE PLUS RECALL
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
COST
(MILLIONS OF DOLLARS)
HC
CO
NOx
EVAP.
GOV'T.
TOTAL
BEST
6.72
73.12
0.23
1.89
1.22
2163.03
LOU
2.67
29.40
0.18
0.60
1.22
460.53
HIGH
6.72*
79.04
1.24*
3.13*
1.22
2165.53
cr>
~Adjusted for consistency.
-------
TABLE 29: RESULTS FOR CASE 19 - STANDARDS PLUS
CERTIFICATION PLUS INSPECTION AND
MAINTENANCE
ESTIMATE
(MIL
EFFECTIVENESS
LIONS OF TONS OF POLLUTANT SAVED)
CC
(MILLIONS C
1ST
F DOLLARS)
HC
CO
NOx
EVAP.
GOVT.
TOTAL
BEST
6.53
77.06
0.50
3.11
3.26
2163.26
LOU
5.61
59.55
-.23
1.79
3.26
2153.26
HIGH
6.66
79.04
1.22*
3.11*
3.26
2173.26
ro
*Adjusted for consistency.
-------
63
TABLE 30: COST EFFECTIVENESS - FEDERAL GOVERNMENT
DOLLARS SPENT PER TON OF POLLUTANT
REMOVED (BEST ESTIMATES)
CASE
HC
CO
N°x
EVAP.
1
0
0
0
0
2
0.137
0.013
1.630
0.264
3
0.130
0.011
col
0.264
4
0.030
0.003
1.944
0.094
5
0.031
0.003
2.188
0.096
6
0
0
0
0
7a
0
0
0
0
7b
0
0
0
0
8a
0
0
0
0
8b
0
0
0
0
9
0.024
0.003
1.000
0.075
10
0.024
0.003
0.684
0.075
11
0.157
0.013
d
0.320
12
0.114
0.010
»l
0.264
13
0.172
0.014
2.947
0.362
14
0.156
0.013
0.339
0.362
15
0.026
0.002
1.944
0.093
16
0.155
0.013
2.947
0.362
17
0.052
0.005
1.326
0.164
18
0.045
0.004
1.326
0.161
19
0.125
0.011 1
1.630
0.262
XN0X effectiveness for this case is negative.
-------
64
TABLE 31: COST EFFECTIVENESS - TOTAL DOLLARS
SPENT PER TON OF POLLUTANT REMOVED
(BEST ESTIMATES)
CASE
HC
CO
NOx
EVAP.
1
79.89
9.19
6071.43
245.66
2
74.21
7.09
881.63
142.66
3
70.08
5.80
«i
142.66
4
74.90
7.39
4777.22
231.16
5
75.17
7.48
109.37
234.95
6
77.89
7.72
2763.16
303.47
7a
84.81
8.46
7500.00
291.67
7b
84.41
' 8.60
5250.00
291.67
8a
81.27
7.96
7500.00
291.67
8b
81.27
8.02
2500.00
291.67
9
79.86
9.09
3317.17
249.27
10
79.97
8.87
2293.72
251.91
11
70.53
5.75
*1
144.26
12
75.96
6.61
175.02
13
69.92
5.85
1201.51
147.76
14
77.73
6.64
1464.66
180.12
15
78.86
7.29
5888.33
280.40
16
77.09
6.63
1464.66
180.12
17
74.71
7.07
1916.34
236.97
18
80,47
7.40
2351.12
286.12
19
82.82
7.02
1081.63
173.90
xNO effectiveness for this case is negative.
X
-------
65
REFERENCES
[Auerbach, 1977]
Auerbach, Janet L., US Environmental Protection Agency, personal
communication, 28 March 1977, with attachments.
[EPA, 1977]
US Environmental Protection Agency, "Appendix N - Emission Reductions
Achievable through Inspection and Maintenance of Light Duty Vehicles,
Motorcycles, and Light and Heavy Duty Trucks," Federal Register,
42FR22177, 2 May 1977.
[EPA, Undated]
US Environmental Protection Agency, Economic Assessment of the
Proposed 1978 Light Duty Truck Emissions Standards.
[MSED, 1977]
Mobile Source Enforcement Division, US Environmental Protection Agency,
Current Investigations of Alleged Emissions - Related Problems3
February 1977.
[Train, 1976]
Train, Russell E., Administrator, US Environmental Protection Agency,
letter to Honorable James T. Lynn, Director, Office of Management and
Budget, May 1976.
[Walsh, 1976]
Walsh, Michael P., US Environmental Protection Agency, The Need For
and Benefits of Inspection and Maintenance, October 19, 1976.
[Williams, 1976]
Williams, Marcia, US Environmental Protection Agency, "Estimates of
Emission Deterioration for 1975 Model Year Vehicles," internal
memorandum, August 6, 1976.
------- |