EPA-AA-IMS-81-9
Technical Report
UPDATE ON THE COST-EFFECTIVENESS
OF INSPECTION AND MAINTENANCE
April 1981
Tom Darlington
NOTICE
Technical Reports do not necessarily represent final EPA decisions or
positions. They are intended to present technical analysis of issues using
data which are currently available. The purpose of the release of such
reports is to facilitate the exchange of technical information and to inform
the public of technical developments which may form the basis for a final EPA
decision, position or regulatory action.
Inspection and Maintenance Staff
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Ann Arbor, Michigan
-------
NOTICE - REVISION TO THIS REPORT. IS IN PROGRESS
THIS REPORT IS CURRENTLY BEING UP-DATED. RECENT INFORMATION BEING INCORPORATED
INTO THIS REVISION WILL INCREASE THE COST PER TON OF HYDROCARBON (HO
REDUCTION AND MINIMALLY DECREASE THE COST PER TON OF CARBON MONOXIDE (CO)
REDUCTION. THE VALUES BELOW ARE FOR A FIVE YEAR PERIOD 1983-1987 AND A
MINIMUM I/M PROGRAM, AS DESCRIBED IN THE CURRENT REPORT. COSTS HAVE BEEN
DIVIDED EQUALLY BETWEEN HC AND CO. NO COST HAS BEEN CHARGED FOR TRAVEL TIME
TO INSPECTION STATION OR REPAIR FACILITY.
COST-EFFECTIVENESS
I/M COST EFFECTIVENESS REPORT HI CO
TABLE I OF CURRENT REPORT
(EPA-AA-IMS-81-9) $581/TON S53/TON
FUTURE UP-DATED REPORT S900/TON $51/TON
-------
Table of Contents
Page
1.0 INTRODUCTION 4
L.I Purpose 4
1.2 Summary 4
1.3 Selection of Base Period for Costs and Emission Reductions 6
1.4 Features of the Typical I/M Program 7
2.0 ANALYSIS OF I/M COSTS BETWEEN
JANUARY 1. 1983 AND DECEMBER 31. 1987 8
2.1 Effects of Inflation 8
2.2 Inspection Costs 9
2.3 Repair Costs 11
2.3.1 Causes of Failures and Failure Rates 11
2.3.1.1 Pre-1981 Vehicles 11
2.3.1.2 1981 and Later Vehicles 11
2.3.2 Per Vehicle Repair Costs 12
2.3.2.1 Pre-1981 Vehicles 12
2.3.2.2 1981 and Later Vehicles 13
2.3.3 Effect of Emission Performance Warranty /;
on 1981 and Later Vehicle Repair Costs 14
2.3.4 Total Repair Costs 15
2.4 Fuel Economy Savings Due to I/M . 16
2.5 Summary of I/M Costs and Savings 16
3.0 EMISSION REDUCTIONS ATTRIBUTABLE TO I/M
BETWEEN JANUARY 1. 1983 AND DECEMBER 31. 1987 17
4.0 COST-EFFECTIVENESS FINAL RESULTS 19
5.0 COMPARISON OF I/M TO OTHER POLLUTION CONTROL STRATEGIES 20
-------
List of Appendices
Appendix 1 Equation Used in Estimating I/M Cost-Effectiveness 23
Appendix 2 Operating and Planned I/M Program Types and 24
Inspection Fees
Appendix 3 Model Year Make-Up of One Million Vehicle Fleet 25
From 1/1/83 to 1/1/88; Number of Vehicles Failed
in I/M Program-
Appendix 4 Notes on Repair Costs in Operating I/M Programs 26
Appendix 5 Analysis of Increase in Consumer Cost-Effectiveness 27
of I/M Due to the Emissions Control System Performance
Warranty Regulations
Appendix 6 MOBILE2 Inputs Used in Calculating Emission 28
Reductions Attributed to I/M
Appendix 7 Equation for Estimating Fleet Emission Reductions 29
Appendix 8 Methodology Issues and Sensitivity 30
-------
1.0 INTRODUCTION
1.1 Purpose
The coat-effectiveness of an air pollution control strategy is the measure of
that strategy's cost relative to its ability to remove a particular pollutant
from the atmosphere. It is commonly expressed as a ratio of cost to the
emission reduction achieved by the strategy, where costs are estimated in
dollars spent in complying with the strategy, and emission reductions are
measured in tons.
The cost-effectiveness of Inspection and Maintenance (I/M) as an air pollution
control strategy has been studied before by the EPA and other groups.
However, over the past year new information on repair costs and inspection
fees and improved methods of calculating emission reductions attributable to
I/M have become available. These changes underscore the need for a fresh look
at the cost-effectiveness of I/M.
1.2 Summary
I/M cost-effectiveness was modeled by having an example fleet of one million
vehicles (gasoline light-duty vehicles only) participate for five years in a
hypothetical I/M program starting in 1983. The design of the hypotheticil T/M
program was typical of programs now being implemented. Total co^its war--
determined by adding together repair costs and inspection costs for the five
year period, then subtracting fuel savings attributable to the I/M program for
the five year period. Inspection and repair costs were estimated using data
from currently operating and planned programs. Emission reductions were
obtained using MOBILE2 (EPA's model for predicting the emission behavior of a
fleet of vehicles with and without I/M) to estimate the masses of HC and CO
emissions that would be removed during the five year period by the
hypothetical I/M program.
Table 1 summarizes the cost-effectiveness of I/M. All figures are La 1981
dollars; costs and emission reductions have not been discounted. Because most
areas which are implementing I/M require reductions for both HC and CO, the
costs of I/M have been allocated equally to both pollutants.
-------
5
Table 1
I/M Cost-Effectiveness* over a
Five-Year Period (1983-1987) in 1981 Dollars
Pollutant Allocated I/M Coat Mass Removed by I/M Cost-Effectiveness **
HC $27.05 million46,500 tons (U.S.) $ 581/ton
CO $27.05 million 512,600 tons (U.S.) $ 53/ton
* Some areas which are implementing I/M only need a reduction in one pollutant
(HC or CO) to meet the National Ambient Air Quality Standards. In estimating
cost-effectiveness for these areas, all costs should be allocated to the one
pollutant only. The cost-effectiveness for that pollutant would then be
double that shown in this table.
** Cost-effectiveness values in general vary with program design and type.
The EPA has explored other program designs and types which exhibit better
cost-effectiveness values than the "conventional" I/M (I/M which uses HC and
CO cutpoints) program whose cost-effectiveness is shown in this table.
Basically, these other programs use only an idle CO cutpoint, and contain
other enhancements which yield HC and CO reductions which are equivalent or
superior to those in a conventional I/M program. Cost-effectiveness values
for these programs range from $386/per ton to $434/ton for HC, and $23/ton to
$26/ton for CO. Complete discussions of these programs are found in the EPA
report entitled "Low-Cost Approaches to Vehicle Emissions Inspection and
Maintenance" (EPA-AA-IMS-81-7).
-------
1.3 Selection of Base Period for Costa and Emission Reductions
A number of different base periods can be used to calculate the
cost-effectiveness of an Inspection and Maintenance program. First, costs and
benefits can be estimated over one particular year of a specific I/M program
such as the Portland, Oregon I/M program. However, the results of this type
of approach reflect only the current model year mix of vehicles participating
in the I/M program. Consequently, it does not account for changes that take
place in the fleet of vehicles tested by the I/M program over the life of the
I/M program, as new vehicles are added to the fleet and old vehicles are
retired. A second approach _to estimating I/M cost-effectiveness is to
determine costs and benefits over the entire life of one "average" vehicle.
However, the results of this approach are also quite restricted in applicabil-
ity. Cost-effectiveness of I/M under this approach would be applicable only
to the model year and technology type of a particular vehicle being studied
and only for an I/M program that was operating over the life of that vehicle.
A third way to estimate the cost-effectiveness of an I/M program is to
calculate costs and benefits for a "typical" fleet which participates in a
"typical" I/M program for a period of several years, taking into account
changes which occur in the fleet participating in the I/M program during that
period. This is the approach used to estimate I/M cost-effectiveness in this
report.*
* The specific equation used to estimate cost-effectiveness is presented in
Appendix 1.
-------
1.4 Features of the Typical I/M Program
The fleet chosen to represent a typical I/M fleet was a one million vehicle
(gasoline light-duty vehicles only) fleet which had a changing model year
distribution. The period chosen was the five year period from January of 1983
through December of 1987. This corresponds to the first five years of many
new I/M programs now in the planning and implementation stages. The test used
in the I/M program was the Idle Test, with a stringency rate of 20% for
pre-1981 model year vehicles, and cutpoints of 1.22 CO and 220 ppm HC for 1981
and later vehicles (these cutpoints are expected to yield a failure rate
of 5-10% for 1981 and later-vehicles). The model year coverage of the program
was 1968 and later. The program did not include a repair cost waiver.*
The decision of which base period to use to calculate I/M cost-effectiveness
is an important one. Equally important, however, is that every calculation
uses that period in a consistent manner. The use of the five year base period
base described in preceding paragraphs to determine I/M program costs and
benefits for a one million vehicle fleet is the subject matter of the rest of
this report.
* A repair cost waiver is an administrative provision whereby cars are excused
from compliance with the I/M requirement if the owner has spent at least a
specified amount of money attempting to make them comply. A repair cost
waiver was excluded from our analysis for two reasons. First, MOBILE2 (EPA's
latest model for predicting the emission behavior of a vehicle or fleet of
vehicles) is not yet capable of calculating emission reductions for an I/M
program that includes a repair cost waiver. Second, although many states are
planning to implement a program with a repair cost waiver, two states with
operating programs (Oregon and New Jersey) do not use a repair cost waiver.
We have based our estimates of I/M repair costs in part on repair costs from
these two programs. So although our estimates of "typical" emission
reductions might be slightly overstated from what they would have been if a
repair cost waiver had been included, our average repair cost may well be
slightly overstated also. We do not expect I/M cost-effectiveness to be
materially different with a properly administered repair cost waiver.
-------
2.0 ANALYSIS OF I/M COSTS BETWEEN JANUARY 1. 1983 AND DECEMBER 31. 1987
There are many different types of costs incurred by different groups of people
during the implementation and operation of an I/M program. However, adding
together all the costs incurred by different people would not adequately
reflect actual costs, as many items would be counted more than once in this
type of approach. For example, where a state selects a contractor to build
facilities and perform I/M tests for the state, it would be double-counting to
add the contractor's capital costs for facilities and equipment to the
inspection fees motorists must pay to obtain an I/M test. The contractor
borrows money to pay for -initial capital facilities, and provides for the
amortization of loans with a portion of the inspection fee paid by motorists.
The relevant costs for the purpose of estimating the cost-effectiveness of I/M
are ultimate costs to the motorist. This includes inspection costs (fees) and
repair costs, less the dollar value of fuel savings. This report assumes that
all costs incurred by the state for administering the program and by the
contractor or the private garage for performing the inspection, are reimbursed
through the payment of inspection fees. In practice, most states have
established their inspection fees in this manner.
2.1 Effects of Inflation
Inflation will act to raise the operating costs (for example, the labor cost
of inspectors) of the I/M program, and will also act to raise repair costs
through higher labor and parts costs. However, we elected not to include
inflation's effects on I/M costs (except where unavoidable), and instead to
report all I/M costs for the five year analysis period from 1/1/83 to 12/31/87
in 1981 dollars.* There were several reasons for the decision to eliminate
inflationary effects. First, cost-effectiveness values of many other types of
emission control strategies have also been calculated without inflation's
effects. Compensating I/M cost-effectiveness for inflation's effects would
result in an incorrect comparison between I/M and other control strategies.
Second, the inflation rate is difficult to predict for a period of time from
two to seven years in the future. Cost-effectiveness numbers which are based
on current expectations of the future inflation rate could be significantly
off-base even before most I/M programs get underway. Third, inflation will
also act to raise consumer incomes over the life of the I/M program,
consequently, the relative cost of I/M (to consumer incomes) should remain
approximately constant.
* Costs have not been discounted to 1981. To discount or not to discount
costs is a methodology issue with proponents on both sides. Appendix 8
contains a discussion of this issue.
-------
2.2 Inspection Coats
Inspection costs are fees that are charged to motorists for receiving I/M
tests. Current fees from operating and planned programs range, from three to
fifteen dollars, with the average of all programs which have decided on a fee
being about eight dollars. Fees for each program are listed in Appendix 2.
Average fees arranged by program type are listed below in Table 2.
Table 2
Average Inspection Fees Listed by
Program Type *
Program Type Fee Number of Programs
Centralized: Contractor-Run $9.62 6
Centralized: State-Run $5.83 3
Centralized: All $8.36 9
Decentralized $7.52 10
Overall Average - $7.90 19
* This is a list of fees from states which have determined an inspection fee.
Some states have not yet decided on a program type or fee. See Appendix 2 for
a complete list of program types and fees.
The figures in Table 2 display a fairly wide range of values. Actually, one
would not expect them to be nearly identical, because each I/M program is
being designed to meet the needs of that particular state. For example, some
I/M states have an existing centralized safety inspection program, most
notably New Jersey. The New Jersey I/M program was grafted onto the existing
safety program which already consisted of a network of inspection stations.
The New. Jersey inspection fee ($2.50) reflects the fact that new facilities
did not need to be built to conduct I/M.
Many other states planning decentralized programs currently have decentralized
safety inspection programs. The average decentralized inspection fee in Table
2 may be overstated because some of the fees used in calculating the average
include the safety as well as the emission inspection. Also, Nevada's
emission inspection, the fee of which was used in the decentralized fee
average, also includes some basic engine adjustments.
Some of the inspection fees, particularly in states with decentralized
programs, are limited by law to a certain figure. It is not clear whether or
not these figures will change during the life of the I/M program. Also, the
generally low inspection fees in decentralized programs are suspect because
they may not fully reimburse garage operators for their time and capital
-------
10
costs; the operators may increase repair bills to compensate for the loss
incurred in performing an emission inspection only, or may simply absorb the
loss to retain regular customers' other business.
The contractor-run centralized programs have perhaps the most realistic fees.
These fees are least likely to be distorted by hidden subsidies, and are
therefore probably the most realistic market value fees for an emission
inspection.
For the purposes of the cost-effectiveness calculation in this report, we have
selected $10.00 as the coat- of-an emission inspection. This figure is close
to the average of $9.62 for the fees from the six contractor-operated
programs. In some cases the contractor has agreed to accept a level fee over
the life of the I/M program, therefore, the fee has taken into account the
contractor's best estimate of the impact of inflation on operating costs. In
other cases the contractor and the state agree to an inspection fee that
increases over the life of the I/M program. Although it was decided that in
general the effects of inflation were not to be included in I/M costs, this
$10.00 figure was not adjusted to remove inflation because of the difficulty
in determining what fraction of each of the six fees is attributable to
inflation.*
Practically all centralized I/M programs do not charge a separate fee for
retests.** Instead, the initial fee is set high enough to cover the cost of
performing reinspections. Therefore, retest fees have been excluded from7 our
cost-effectiveness analysis.
In order to estimate the total five year inspection cost of all vehicles
participating in the I/M program, it was necessary to determine how many
vehicles in the one million vehicle fleet each year would be either too young
(less than one year) or too old (1967 model year and earlier) to be
inspected. This was accomplished with EPA estimates of the national
distribution of vehicles by model year. Estimates of the number of vehicles
that would need an inspection each year are presented in Appendix 3. We
estimated that 4,880,000 inspections would be performed on the one million
vehicle fleet over the five year period; this results in a total inspection
cost of $48.8 million dollars.
* Cost-effectiveness values are relatively sensitive to inspection fees. If a
$6.00 fee (this is the fee charged for an emission inspection in New York) had
been selected instead of the $10.00 fee, the cost-effectiveness results would
have been $34 per ton for CO, and $372 per ton for HC, as compared to $53 per
ton for CO and $581 per ton for HC with the $10.00 inspection fee (costs
allocated half to HC, half to CO).
** The one exception is the change-of-ownership program in Los Angeles, which
charges a $7 retest fee. The fee in this program also provides for reimburse-
ment to the state of a multi-million dollar debt incurred from extensive
testing performed during early development of the program.
-------
11
2.3 Repair Costs
We have stated that we will analyze the cost-effectiveness of I/M from the
perspective of a fleet of one million vehicles whose model year distribution
changes with time. One of our reasons for taking this approach is that there
are some important differences between pre-1981 and 1981 and later vehicles
which have a profound impact on total repair costs. First, the failure rate
on 1981 and later vehicles is expected to be lower than the failure rate on
pre-1981 vehicles. This difference has an impact on total repair cost for the
fleet, as only vehicles that are failed from an I/M program must obtain an I/M
repair. Second, the types of repairs on 1981 and later vehicles are expected
to be quite different than the types of repairs that are being performed on
pre-1981 vehicles in operating programs. These differences in types of
repairs may result in differences in the cost of repairs also. These
fundamental differences between pre-1981 and 1981 and later vehicles, which if
ignored would greatly distort the true picture of I/M cost-effectiveness, are
discussed in the following sections.
2.3.1 Causes of Failures and Failure Rates
2.3.1.1 Pre-1981 Vehicles
Most pre-1981 I/M test failures are caused by misadjustments in engine
parameters. These misadjustments can occur over a wide range of possible
settings. Therefore, I/M program officials can control the I/M failure'rate
for pre-1981 vehicles to any desired level by selecting CO and HC inspection
standards (or cutpoints) accordingly. Most states are designing their I/M
programs with about a 20% or higher failure rate for pre-1981 vehicles.
Therefore, we will use a 20% failure rate for pre-1981 vehicles for the
purpose of calculating I/M cost-effectiveness.
2.3.1.2 1981 and Later Vehicles
I/M program officials have less control over the failure rate for 1981 and
later vehicles than for Che older vehicles. This is due to two important
differences between them. First, recent EPA regulations* require auto
manufacturers to make idle mixture settings on 1981 cars much more resistant
to misad justment than was the case for older cars. Consequently,
misadjustments of the idle mixture, the factor contributing to the majority of
I/M test failures for pre-1981 vehicles, will be rare on 1981 and later
vehicles. Second, some fraction of the 1981 and later fleet will fail an I/M
test for reasons which do not apply at all to pre-1981 vehicles. Most 1981
and later vehicles will come equipped with closed-loop fuel metering systems
controlled- by an on-board computer that receives signals from various sensors
which measure the operating air/fuel ratio, temperature, speed, and power
level of the engine. The computer then adjusts the fuel metering system, via
electromechanical actuators, to achieve a desired air/fuel ratio. Failure of
a sensor, the computer, air actuator, or the electrical connections between
them can cause high-emissions on the I/M test.
* 44 F.R. 2960, commonly refered to as the Parameter Adjustment Regulations.
-------
12
There are some reasons for an I/M failure that will apply for both pre-1981
and 1981 and later vehicles, such as ignition system problems, vacuum leaks,
severe engine wear, and gross tampering with emission controls used on both
types of vehicles.
The overall result of the special character of I/M failures among 1981 and
later vehicles is that there will be a bimodal distribution of I/M test
results. Most vehicles will have very low I/M test scores, but some will have
very high scores. The failure rate will be affected to some extent by the
choice of an I/M test, since the more complicated tests (a loaded test or a
two-speed idle test which- requires cars to pass on both the idle and
high-speed idle portions) can detect more of the vehicles with failures in
their fuel metering systems.. The choice of inspection cutpoints will also
affect the failure rate to some extent. But for an I/M program which uses the
simple idle test, EPA expects that it would be impossible to adjust the
cutpoints so that the failure rate was less than 5 percent without sacrificing
most of the program's effectiveness, or more than 10 percent without failing
many cars that actually have low emissions.
For these reasons, we have chosen 7% as the failure rate for 1981 and later
vehicles for the purpose of estimating repair costs from 1981 and later
vehicles. This is within the range of 5-10% just mentioned. We believe it is
a reasonable estimate for any I/M program that uses the single idle test and
cutpoints in the range of 1.0-3.0% CO and 200-300 ppm HC.
2.3.2 Per Vehicle Repair Costs
2.3.2.1 Pre-1981 Vehicles
Extensive data on I/M repair costs for pre-1981 vehicles has been obtained
from operating programs in New Jersey, Oregon, California and Arizona. This
data is summarized in Table 3. Sources and notes on this data, including
discussions on the methods used in each state to collect the data, are
presented in Appendix 4.
The reader may be surprised to notice that the average I/M repair costs seems
to be substantially less than the price of a full tune-up, which many people
in non-I/M cities currently have performed on their cars on an annual basis.
This should not be interpreted as throwing doubt on the accuracy of I/M repair
costs in Table 3, since there is a sound explanation for the difference in
price. The reason for the difference in price between an I/M repair and a
tune-up is that an I/M repair is usually a single repair such as carburetor
adjustment, where a tune-up may involve several items such as spark plug and
air filter replacement, carburetor adjustment and adjustments of spark timing
and dwell. In addition, once an I/M program is underway, there is often
competition and price cutting by the repair industry.*
* Price competition may be an explanation for the reduction in average I/M
repair costs in New Jersey and California. A 1973-74 post card survey in New
Jersey found the average repair cost to be $36. The 1979 and 1980 New Jersey
surveys, although conducted in a different manner than the 1973-1974 survey,
found average I/M repair costs to be: $28 and $18.71, respectively. In
California, the 1979 average cost of repair is $32. For the latter two
quarters of 1980, the average cost of repair is less than $30. Sources for
this data are listed in Appendix 4.
-------
13
Table 3
Average Pre-1981 Vehicle I/M Repair Cost
from Operating Programs *
I/M Program Period Covered Average Repair Cost
New Jersey 1979 $28.00
1980 $18.71
Arizona 1979 30.00
1980 29.74
Portland, Oregon 1980 17.00
California 1979 32.00
1980 (July - September) 29.09
1980 (October - December) 28.82
* Sources for these costs are listed in Appendix 4.
The data in Table 3 indicate that the average cost of repair for pre-1981
vehicles is in the range of $17.00 (Portland) to about $30.00 (California).
We have selected $25.00 (1981 dollars) as the average cost of I/M repairs for
pre-1981 model year vehicles for use in our cost-effectiveness analysis.*
2.3.2.2 1981 and Later Vehicles
There is at present considerable uncertainty as to the average I/M repair cost
for 1981 and later vehicles. As noted above, very few of the 1981 and later
vehicles which fail I/M will do so because of misadjusted idle mixture
settings, the item most commonly corrected during I/M repairs on pre-1981
cars. The 1981 and later cars will share with the older cars repair of such
items as ignition problems, vacuum leaks, and tampering of some emission
control components they have in common. But some of the newer cars will also
require repairs to their closed-loop fuel metering systems. These repairs
will range from simply reconnecting a wire to a sensor or actuator to
replacing a sensor, an actuator, or the computer. The average repair cost
will very much depend on how often each type of repair is necessary. If most
repairs require a replacement sensor or actuator, the average repair cost
could be more.than on pre-1981 vehicles. It should not be expected that new
* If the highest typical (modal) value of $29 had been used as the average
pre-1981 I/M repair cost, our cost-effectiveness results would have been
$606/ton for HC and $55/ton for CO, compared to $581/ton for HC and $53/ton
for CO for an average pre-1981 cost of repair of $25.
-------
14
computers would be needed on many of the 7 percent of the vehicles which fail
I/M, because the auto manufacturers work hard to make them reliable. It would
be ideal if repair costs for 1981 and later vehicles would be taken from the
operating I/M programs, as it was for pre-1981 vehicles. Data on repair costs
for 1981 and later vehicles is unfortunately difficult to obtain because most
1981 vehicles in operating I/M programs will not need an inspection until
1982. However, a limited amount of repair cost data is available on vehicles
equipped with roughly the same types of emission control systems being used on
1981 and later vehicles. The data come from California's change-of-ownership
Vehicle Inspection Program in the Los Angeles area, where vehicles are tested
only when ownership is transferred from one person to another. As early as
1977, certain manufacturers started introducing vehicles with closed-loop fuel
control and three-way catalysts that are the forerunners of the type of
technology used on 1981 and later vehicles.* Thirty-nine of these vehicles,
with model years ranging from 1977-1981, were failed in the California program
in the period of July to December 1980. The average mandatory repair cost
(i.e., cost of repairs to meet inspection standard) of the vehicles was
$26.67, compared to $28.82 for all vehicles of the same model years.
Many of the cars in the 39 vehicle sample from California were Volvos with
closed-loop fuel injection systems. The repair costs for these vehicles may
be unrepresentative of the true cost of I/M repairs for 1981 and later
vehicles because they have adjustable (non-sealed) idle mixture screws, a
different type of computer and fuel metering approach, and fewer engine
sensors than most 1981 and later vehicles. For these reasons, we have
selected $30.00 as a conservative repair cost estimate for 1981 and later
model year vehicles in our cost-effectiveness analysis.
2.3.3 Effect of Emission Performance Warranty on 1981 and
Later Vehicle Repair Costs
The EFA recently promulgated Emission Control System Performance Warranty
Regulations (45 F.R. 34802, May 22, 1980, hereinafter referred to as the
Warranty) for 1981 and later model year vehicles. The regulations require
manufacturers to provide a warranty which entitles a vehicle owner to emission
related repairs at the vehicle manufacturer's expense if, among other
conditions, the vehicle fails an "approved" short test. The regulations are
designed to protect an individual from having to pay for an expensive repair
that may be due to the design or manufacture of the vehicle. Many I/M
programs are expected to implement approved short tests so that their citizens
may have the benefit of this coverage. Although the Warranty is expected to
reduce the cost to consumers of I/M-motivated repairs on 1981 and later model
year vehicles, it should not affect the overall costs of repair for these
vehicles. The repair cost not borne by the consumer who exercises his
warranty will be borne by the manufacturer and ultimately by either
stockholders and/or new car buyers throughout the U.S. Overall repair costs
* Auto manufacturers introduced this technology in California first (as far
back as 1977) because California's emission standards were more stringent than
federal standards.
-------
15
will not have been reduced, but rather will have been redistributed. The
degree of redistribution is presently unknown. In summary, the Emission
Performance Warranty will probably improve the cost-effectiveness of an I/M
program from the perspective of consumers who must participate in the local
I/M program and from the perspective of their local representatives, but leave
overall I/M cost-effectiveness from the perspective of the U.S. as a whole
unchanged. The administrative costs to the auto manufacturers and their
dealers might increase the total cost of I/M, but only very slightly. An
example of how the apparent cost-effectiveness of I/M for consumers (who must
participate in the I/M program) can be improved with Warranty coverage is
discussed in Appendix 5. For -the rest of the report, we have excluded the
effect of the Warranty on reducing repair costs for 1981 and later vehicles
from our analysis of the cost-effectiveness of I/M.
2.3.4 Total Repair Costs
We have selected $25.00 as our repair cost for pre-1981 vehicles, and $30.00
as our repair cost for 1981 and later vehicles. In estimating total repair
costs for the five year I/M program, it was necessary to first determine for
each year of the program how many of the one million vehicles were pre-1981
vehicles, and how many were 1981 and later vehicles (these figures are
provided in Appendix 3). This was accomplished by using EPA'3 estimates of
the national distribution of vehicles by model year which are consistent with
MOBILE2.* These estimates were combined with the I/M failure rates (20Z for
pre-1981, 7Z for 1981 and later) to estimate total repair costs. We estimated
that a total of 137,095 1981 and later vehicles would be failed during the
five years and need repairs, and 584,300 pre-1981 vehicles,would be failed and
need repairs. These figures yield a $18.7 million repair cost over the five
year I/M program.
We need to mention three factors which could have an impact on overall repair
costs, but have been omitted from our analysis of cost-effectivess because of
insufficient data. First, some portion of the repairs received by failed
vehicles would have been voluntarily purchased by vehicle owners eventually,
even in the absence of an I/M program. Many vehicles which receive a spark
plug replacement after failing an I/M test for high HC probably fall in this
category. These costs could justifiably be deducted from I/M total repair
costs. Second, it is possible that I/M, as a periodic preventive maintenance
function, increases vehicle engine life. Third, it is also possible that I/M
reduces overall maintenance costs for owners who begin to rely on I/M to tell
then when to tune their vehicles; these owners previously may have followed a.
tune-up schedule that was more frequent then actually needed. Inclusion of
these considerations in our cost-effectiveness analysis would have improved or
lowered our total repair cost numbers.
* MOBILE2 is EPA's latest model for predicting the emission behavior of a
vehicle or a fleet of vehicles over a period of time.
-------
16
2.4 Fuel Economy Savings Due to I/M
Fuel economy savings attributable to the "typical" I/M program were taken from
a technical report issued by the EPA entitled "Update on the Fuel Economy
Benefits of I/M Programs" (EPA-AA-IMS-81-10). This report presented the
average dollar savings (in less fuel consumed) for each vehicle participating
in a "typical" I/M program which was the same as the "typical" I/M program
used in this report. The Fuel Economy report showed that the average 1981
dollar savings per year per vehicle inspected (that is, averaged over the
entire fleet, not just those vehicles receiving repairs) in the I/M program
would be $2.74.* This figure,_ .when multiplied by the number of inspections
performed over the five year period (4,880,000; see Appendix 3), yields a
savings of 13.4 million dollars.
2.5 Summary of I/M Costs and Savings
A summary of the aggregate inspection cost, repair cost, and fuel savings for
the one million vehicle fleet which participated in a five year I/M program
from 1/1/83 to 12/31/87 is listed in Table 4.
Table 4
Summary of Total I/M Costs and Savings
Cost Item Cost (1981 Dollars)
Inspection $48.8 million
Repair $18.7 million
(Less) Fuel Savings ($13.4 million)
Total $54.1 million
* Fuel economy benefits in the basic I/M program come only from 1981 and later
vehicles. Therefore, fleet average fuel economy benefits improve with time as
more 1981 and later vehicles are added to the fleet.
-------
17
3.0 EMISSION REDUCTIONS ATTRIBUTABLE TO I/M
Emission reductions attributable to I/M are best calculated by MOBILE2, which
is EFA's latest model for predicting the emission behavior of a vehicle or
fleet of vehicles over a period of time. MOBILE2 can also predict the
behavior of a vehicle or fleet of vehicles participating in an annual I/M
program. The emission reductions attributable to I/M are calculated as the
differences in CO and HC emission rates between a fleet of vehicles not
participating in an I/M program, and a fleet of vehicles participating in an
I/M program.
MOBILE2 requires two different kinds of inputs in estimating non-I/M and I/M
vehicle emission rates. The first category of inputs are local transportation
inputs, for which we have used national average values. The second kind of
inputs are I/M-related, such as failure rate for pre-1981 vehicles and model
year coverage. A complete list of the inputs used in determining emission
reductions attributable to the "typical" I/M program (described in section
1.4) is found in Appendix 6.
The emission reductions attributable to the "typical" I/M program are
presented in Table 5. Reductions are calculated as the difference in emission
rates between the non-I/M and I/M scenarios. Reductions are calculated at the
midpoint of each year in order to most accurately estimate the average
reduction during that year.
Table 5
Reduction in Emission Rates
Attributable to I/M Programs
Midpoint of
Year
1983
1984
1985
1986
1987
1988
Light Duty Vehicle
Fleet Emission Rates
Without I/M
HC *
(g/mi)
3.57
3.25
2.96
2.71
2.51
2.42
CO
(g/mij^
36.34
33.69
31.39
29.56
28.10
27.47
Light Duty Vehicle
Fleet Emission Rates
With I/M
Emission Rate
Difference
HC *
(g/mi)
3.30
2.60
2.19
1.92
1.73
1.66
CO
(g/mi)
33.57
27.50
23.19
20.07
18.85
18.37
HC
(g/mi)
.27
.65
.80
.79
.78
.76
CO
(g/mi)
2.77
6.19
8.20
9.49
9.25
9.10
Total HC emissions, including evaporative emissions.
-------
18
The reader may notice that the emission rate difference for 1983 is substan-
tially lower than the emission rate differences for the remaining years. This
is due to the Eact that on July 1, 1983, only about half of the one million
vehicle fleet has been tested, since the program began on January, 1983. By
July 1, 1984, all of the fleet has been tested, consequently the emission rate
difference of the fleet between non-I/M and I/M scenarios at that date is
higher.
One half of the emission reductions in 1988 are included in our estimates of
cost-effectiveness, even though only 1983 through 1987 costs are included.
This is necessary to preserve the cause and effect relationship between costs
and emission reductions. Some of the emission reductions obtained in 1988 are
due to repairs performed in 1987, because vehicles inspected and repaired in
1987 have emission benefits that decline with time but continue through 1988.
The emission rate differences presented in Table 5 and an estimate of the
number of miles traveled by the fleet of one million vehicles in each year
were used to calculate the masses of HC and CO removed by the five year I/M
program. The equation used in this calculation is presented in Appendix 7.
Masses of HC and CO removed by the I/M program are presented in Table 6. The
"typical" I/M program operating for 5 years on a fleet of one million vehicles
is capable of reducing the amount of CO emitted during the five years by
512,600 tons and the amount of HC emitted by 46,500 tons.
Table 6
Pollutant Masses
Removed by the Typical I/M Program
Year CO (tons) HC (tons)
35,100 3,420
78,400 8,240
103,900 10,100
120,300 10,000
117,200 9,900
57.700 4.800
Totals 512,600 46,500
-------
19
4.0 COST-EFFECTIVENESS FINAL RESULTS
The typical I/M program incurred a total cost over the five year period from
January 1, 1983 to December 31, 1987 of 48.8 million 1981 dollars. Over the
same period, the I/M program was responsible for the reduction in 512,600 tons
of CO emissions, and 46,500 tons of EC emissions.
Table 7 summarizes the cost-effectiveness of Inspection/Maintenance based on
the above numbers. Because most areas which are implementing I/M require
reductions in both HC and CO in order to attain the National Ambient Air
Quality Standards for ozone aad CO, the costs of I/M have been allocated
equally to both pollutants.
Table 7
Cost-Effeetivenes s*
of I/M in 1981 Dollars
Pollutant Allocated I/M Cost Mass Removed by I/M Cost Effectiveness
HC
CO
27.05 million
27.05 million
46,500 tons (U.S.)
512,600 tons (U.S.)
581/ton
53/ton
* Some areas which are implementing I/M only need a reduction in one pollutant
(HC or CO) to meet the National Ambient Air Quality Standards. In estimating
cost-effectivness for these areas, all costs should be allocated to one
pollutant only. The cost-effectiveness for that pollutant would then be
double that shown in this table.
-------
20
5.0 COMPARISON OF THE COST-EFFECTIVENESS OF I/M
WITH OTHER POLLUTION CONTROL STRATEGIES
Table 8 lists the cost-effectiveness values for I/M along with those for a
number of other control measures. The I/M program has a cost-effectiveness
which compares favorably with control measures both recently and yet to be
implemented. For example, for HC, the cost-effectiveness of I/M of $581 per
ton is slightly higher than recently promulgated new motor vehicle emission
standards for cars and trucks, but lower than transit improvements and
reductions in emissions from automobile coating operations. For CO, the
cost-effectiveness of I/M -is -similar to new car standards and much more
cost-effective than transit improvements.
The reader should exercise caution in comparing cost-effectiveness values for
different control measures. The presence of air pollution control strategies
with greater or less cost-effectivness numbers than other strategies does not
imply that there is a cut-off cost-effectiveness above which no strategy is
implemented. There are three other issues which must be considered. First,
cost-effectiveness data should be evaluated in the context of overall clean
air goals: that urban areas that cannot demonstrate attainment of the
National Air Quality Standards must approach attainment as expeditously as
possible. Second, the size of the reductions available from a strategy must
be studied. Although a strategy may have low cost-effectiveness numbers, the
amount of emission reductions realizable by implementing that strategy may
only be a fraction of the emission reductions needed to demonstrate attainment
of the National Air Quality Standards. Lastly, there may be hidden
incompatibilities between different cost-effectiveness values which arise as a
result of different techniques used in estimating a strategy's
cost-effectiveness values. Therefore, a reader should usually view a
cost-effectiveness comparison as a rough or imprecise comparison.
-------
21
Table 8
Cost-Effectiveness of Control Measures
Measure
I/M
1981 Pass. Car Emission Stds1
1984 Gas Truck Stds2
Traffic Controls-*
Transit Improvements-^
Auto Coatings^
Fabric Coatings^
Bulk Plants5«°
Gas Stations^
Being Widely
Implemented?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Cost-Effectiveness
(dollars/ton)
HC CO
581 53
470 41
253 8
617 51
14599 1382
1205
40
net savings
327
Notes;
(1) Cost' of 1981 new passenger car emission standards compared to 1975
standards; Source: "Regulatory Analysis and Environmental Impact of Final
Emission Regulations for 1984 and later Model Year Heavy Duty Engines".
USEPA, December, 1979, pp. 159.
8500
(2) Ibid. pg. 6 Value shown is gasoline heavy duty trucks greater than
GVW.
(3) Source: 1979 SIP for Pima County, AZ.
(4) Although these measures have been or are being implemented in some areas,
most transportation control measures are still being studied.
(5) Source: "Phase I Air Quality and Economic Impacts for the N.Y.
Metro/Hartford Regional Study"; Contract 13-AQ-7718; GCA Corp, Bedford,
MA; October, 1980; prepared for Che National Commission on Air Quality.
Of the 13 stationary source control measures listed in the referenced
document, only the four measures shown in this paper account individually
for more than 1 percent of 1987 expected emission reductions.
(6) Net savings means cost of recovering emissions is less than the value of
the recovered product.
-------
22
APPENDICES
-------
23
Appendix 1.
Equation Uaed in Estimating I/M Coat-Effectiveness
CE - 1C + RC - FS
2EM
where CE is the cost-effectiveness of I/M in dollars spent per ton of
pollutant removed,
1C is the total five year inspection cost,
RC is the total five year repair cost,
FS_ is the total five year fuel savings, and
EM is the total amount of HC or CO emissions (in U.S. tons)
removed by the I/M program.
Note:
(1) This cost-effectiveness equation allocates half of the cost to HC, the
other half to CO.
(2) All figures in the report are presented in 1981 dollars.
-------
24
State
Connecticut
Massachusetts
Rhode Island
New Jersey
New York
Pennsylvania
Virginia
Washington D.C.
Delaware
Maryland
Georgia
Kentucky
North Carolina
Tennessee
Illinois
Indiana
Michigan
Ohio
Wisconsin
New Mexico
Texas
Missouri
Colorado
Utah
Arizona
Nevada
California
Oregon
Washington
Appendix 2.
Operating and Planned I/M Program
Types and Inspection Fees
Safety
Inspection
Type of Program (Y-Yes, N»No) Fee
Centralized, Contractor-Run
Decentralized
Decentralized *
Centralized, State-Run *
Decentralized
Decentralized
Decentralized
Centralized, State-Run
Centralized, State-Run
Centralized, Contractor-Run
Decentralized
Centralized
Decentralized
Centralized, State-Run
Centralized, Contractor-Run
Centralized, Contractor-Run
Decentralized
Centralized
Centralized, Contractor-Run
Centralized, State-Run
Undecided
Decentralized
Decentralized
Centralized, Contractor-Run
Centralized, Contractor-Run *
Decentralized
Undecided **
Centralized, State-Run *
Centralized, Contractor-Run
Fee Covers
N
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
N
N
N
N
N
N
N
N
N
Y
Y
Y
Y
N
N
N
N
N
$10.00
10.00
4.00
2.50
12.00
Undecided
3.50
Undecided
Undecided
9.00
3.00
5.00-10.00
3.65-10.00
Undecided
13.00
10.00
10.00
Undecided
Undecided
9.00-10.00
Undecided
3.50
Undecided
Undecided
5.75
12.00-17.00 +
Undecided
5.00
10.00
Emis. and Safety
Em is. and Safety
Emis. and Safety
Emis. and Safety
Emis. and Safety
Emissions Only
r
Emissions Only
* Operating program as of 1/1/81.
** California operates a change-of-ownership centralized, contractor-run I/M program in Los
Angeles. The fee for initial inspection is $11.00.
*** Some states have decided to have a centralized program, but have not yet decided whether
it will be run by the state or a contractor.
+ The Nevada emission inspection includes some basic engine adjustments.
-------
25
Appendix 3
Model Year Make-up of One Million Vehicle
Fleet from 1/1/83 to 1/1/88
1967 and
Year 1981 and Later 1980-1968 Earlier *
215,000 747,000 38,000
287,500 685,500 27,000
382,000 - '•-•• 597,000 21,000
490,000 492,000 18,000
584.000 400.000 16,000
Totals ** 1,958,500 2,921,500
* These vehicles are not tested in the hypothetical I/M program.
** The total number of inspections performed over the five year period is the
sum of 1981 and Later and 1980-1968 tested; or 4,880,000. The number of
vehicles failed in each category is presented below:
Number of Vehicles Failed
in I/M Program
Vehicle Category Number Tested Failure Rate Number Failed
1981 and Later 1,958,500 72 137,095
1980-1968 2,921,500 20Z 584,300
-------
26
Appendix 4
Notes on
Repair Costa in Operating I/M Programs
New Jersey - Repair costs in New Jersey were obtained by auditors who regular-
ly visit reinspection garages. Forty auditors randomly selected twenty job
receipts from garage records of vehicles applying for reinspection (for a
total of 800 records) and wrote down the vehicle's model year and make, the
reason it failed, and parts -and labor cost of repairs. This study was
conducted in late 1979. Additional details on the study are available from a
report entitled "The Cost of the New Jersey Motor Vehicle Exhaust Emission
Inspection and Maintenance Programs", State Contract #41410-400-212-396.
Arizona - Each vehicle receiving an I/M test in Arizona receives a form that
presents the test results of that vehicle. On the back of the form is a
section which is to be completed by the mechanic performing repairs for
vehicles failing the I/M test. This section includes information on the
repairs performed and their costs. Each month, 10% of the forms returned to
Arizona during.reinspection are randomly selected and analyzed to determine
average repair cost for that month. 1979 repair costs in Table 3 were
obtained from "A Survey of Operating Inspection/Maintenance Programs"; R.F.
Klausmeier, D.K. Kirk, 17 April 1980, EPA Contract 68-02-2538. 1980 repair
costs were obtained in a conversation with Fred lacobelli, Chief, Bureau of
Vehicular Inspection, Arizona Department of Health Services, 2/13/81.
Port land, Oregon - Repair cost data is obtained periodically by handing out
repair forms to vehicle owners whose cars fail the I/M test. Filling-out of
the form is voluntary; forms are collected by inspectors when the vehicle
returns for a retest. Repair costs in Table 3 came from Cost of Repair
Survey, May through July 1980, Department of Environmental Quality. This
repair cost survey consisted of 7832 total responses.
California - California's data collection system for repair costs is very
similar to Arizona's. Each vehicle receiving an I/M test is given a form
which presents the results of the test. Owners whose vehicles fail the I/M
test must have Che back of the form filled out by the person performing
repairs. This form is resubmitted to the inspector on successful completion
of the retest, and the repair data is automatically entered on a computer tape
file. Repair cost data is analyzed on a quarterly basis. 1979 data from
Table 2 was obtained from "A Survey of Operating Inspection/Maintenance
Programs"; R.F. Klausmeier, D.K. Kirk, 17 April 1980, EPA Contract
68-02-2538. 1980 data was obtained from Vehicle Inspection Program's Average
Codt of Mandatory Repairs tables, July through September 1980, and October
through December, 1980. These samples together consisted of approximately
68,000 responses.
-------
27
Appendix 5
Analysis of Increase in Consumer Cost-
Effectiveness of I/M due to the Emissions
Control System Performance Warranty Regulations
Since the Emission Control System Performance Warranty only applies to 1981
and later vehicles, and most 1981 and later vehicles will not need inspection
in I/M programs until 1982, very little is known about the portion of
I/M-motivated repair cost ~ that will be borne by manufacturers under the
warranty. But if 502 of the 1981 and later vehicle repair cost is borne by
the manufacturer, the cost-effectiveness of our "typical" I/M program for
consumers who must participate in the I/M program improves by about four
percent. See the table below.
Percent of 1981 and Later
Repair Cost Borne by
Manufacturer Pollutant Cost-Effectiveness *
OZ HC $581/ton
CO 53/ton
50Z HC $560/ton
CO 51/ton
75Z HC $548/ton
CO 50/ton
Tons are U.S. tons
-------
28
Appendix 6.
MOBILE2 Inputs Used in Calculating
Emission Reductions Attributed to I/M
I/M Input8
Program length: 1/1/83 through 12/31/87.
No mechanic training.
No waivers.
Vehicle classes covered by the I/M program are light duty vehicles.
Stringency rate of 20% for pre-1981 model year vehicles; default identifi-
cation rate of 502 for 1981 and later model year vehicles.
1968 and later model years are subject to the I/M program.
Transportation Inputs Value
Vehicle Average Speed --- _-____- ___ 19.6 mph
Ambient Soak Temp. 75°F
Percentage of Non-Catalyst
Equipped VMT accumulated
in Cold Start mode --------------------- 20.6%
Percentage of Catalyst
Equipped VMT accumulated
in Hot Start mode 27.3%
Percentage of Catalyst
Equipped VMT accumulated
in Cold Start mode 20.6%
-------
29
Appendix 7.
Equation for Estimating
Model Fleet Emission Reductions
M - VEHS x MILES x EFD x CF
Where M • mass of pollutant in U.S. tons
VEHS » Number of vehicles in model I/M program (one million)
MILES * Average number of miles traveled per vehicle (11,507.9)
EFD =» Difference in emission factors between vehicles not exposed to
the model I/M program and vehicles exposed to model I/M program.
There is a different EDF for each pollutant each year. The EDF
for a particular year is estimated at the midpoint of that year.
CF » Conversion factor which converts grams to tons (1 U.S. ton »
908,000 grams)
-------
30
Appendix 8
Methodology Issues and Sensitivity
This appendix contains discussions of other factors which could either lower
or raise our cost-effectiveness numbers. Included are discussions of future
costs discounted to present values, time costs of motorists obtaining
inspections and repairs, repair costs other than mandatory repair costs for
failed vehicles (i.e., anticipatory maintenance), and the effect of vehicle
operating conditions on I/M emission reductions.
Discounting - The method of discounting costs to present values is a common
practice which is used to determine the lump sum of money that could be
invested today (at current interest rates) that would pay for an I/M program
in the future. We omitted the practice of discounting in the interest of
simplicity; it would have the effect of improving or lowering our present I/M
cost-effectiveness numbers. However, the reader who wishes to determine
discounted I/M cost-effectiveness has all the information in the report
necessary to perform this task.
Motorists' Time and Travel Costs - The time and travel costs to the motorist
of obtaining an inspection and, if necessary, a repair, is a cost factor which
we omitted from our cost-effectiveness analysis because of the lack of
reliable data on these costs. Should this data become available, it can
easily be factored into our analysis. It would have the effect of raising- our
cost-effectiveness numbers.
Other Repairs - In figuring total I/M repair costs, we included only mandatory
I/M repair costs of vehicles failed in our typical I/M program. Some analysts
would argue that I/M raises repair costs of passing vehicles by some amount
also; as some people perform anticipatory maintenance on their vehicles prior
to obtaining an I/M inspection. We are aware of no data on the total cost of
I/M anticipatory maintenance, therefore, we have omitted this cost from our
analysis. Including anticipatory maintenance costs would have the effect of
raising our cost-effectiveness numbers.
The reader should recall that there is a group of failed vehicles on which
repairs were performed that would have been performed in the absence of an I/M
program. The vehicle which fails for excessive HC and obtains a spark plug
replacement is probably a member of that group. Although inclusion of this
consideration in our cost-effectiveness as analysis would have improved or
lowered our cost-effectiveness numbers, it was also excluded because of
insufficient data.
Vehicle Operating Conditions -Local vehicle operating conditions (average
speed temperature, etc.) have an effect on the quantity of emission reductions
obtained from an I/M program as predicated by MOBILE2 We have used standard
national average values for speed, temperature, and other vehicle operating
conditions (see Appendix 5).
------- |