EPA Report No.:
EPA-AA-CPSB-82-03
Technical Report
Cost Analysis of Proposed Changes to 40 CFR Part 600
to Improve Fuel Economy Labeling and the
Fuel Economy Data Base
Clifford D. Tyree
September 1982
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 in 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.
U. S. Environmental Protection Agency
Office of Mobile Sources
Certification Division
Certification Policy and Support Branch
2565 Plymouth Road
Ann Arbor, Michigan 48105
-------
-2-
I. Introduction
The following analysis has been completed to estimate the
economic impact on the automotive industry of the various rule
changes proposed in the Notice of Proposed Rulemaking (NPRM)
entitled 40 CFR Part 600 Fuel Economy of Motor Vehicles--
Revisions to Improve Fuel Economy Labeling and the Fuel Economy
Data Base. This analysis supersedes the January 1981 analysis
of the same title.
II. Organization of Analysis
The analyses in this paper are grouped in the same order
as each proposal discussed in Section II of the NPRM. The
estimated cost of the various proposals are summarzied in Table
I of this analysis.
III. Proposed Modifications
A. In-Use Adjustments for Label Values
This NPRM proposes constant shortfall ad3ustment
factors to be applied to model type fuel economy values to
arrive at the label values. This specific proposal will not
change the number of labels nor the number of tests required
under the current program. The impact of this proposal on the
current program will be in the calculation procedures only.
EPA computer personnel estimate that no more than eight hours
would be needed per manufacturer to make the required computer
program changes. An additional eight hours of programming time
per computer might be allocated for implementing the software
on computers located at several assembly sites. It was esti-
mated on the basis of data from the Motor Vehicle Manufacturers
Association that there are approximately 100 assembly sites.-L
Assuming the worst case that all manufacturers have computer
systems for generating label values, the total cost of this
proposal would be: the number of manufacturers times the esti-
mated reprogramming hours per manufacturer, plus the number of
hours required per computer for implementation, all multipled
by the cost per hour of computer programmer work. At an esti-
mated $30 per hour for programming (based on prevailing con-
tractor rates), the total cost would be $47,000;
1. Motor Vehicle Manufacturers Association of the United
States, Inc. Information Handbook; Facilities, 1980.
-------
-3-
((32 mfrs)(24 hrs per mfr) + (100 computers)(8 hrs per
computer)) $30 per hr.
= $47,040
B. Minimum Data Requirements for Labeling
The current program allows that any base level,2
where certification data satisfied the minimum data require-
ments, could be represented only by a low sales volume group of
vehicles. This proposal will require tnat each base level have
data generated from a vehicle which represented the highest
selling configuration in the base level. This should not
increase annual data requirements for initial labels over the
current system because each manufacturer must currently supply
data to meet the requirements for their preliminary corporate
average fuel economy (PCAFE). The current minimum aata
requirement for preliminary CAFE is that data be provided on
configurations that represent at least 90 percent or. the pro-
jected sales, in decreasing order. As a result, this proposed
change in data coverage represents a "pull ahead" in coverage
for initial labeling and not an increase in annual data
requirements. Because of this, it is estimated that this
change would not result in a testing cost increase to the manu-
facturers for initial labels. Any cost attributable to pulling
ahead testing is likely to be insignificant. The following
section covers costs attributed to the minimum labeling data
proposal as they apply to labeling subsequent to the initial
labels.
C. Label Updating
This NPRM proposes two types or label updating.
First, relabeling at approximately the middle of the production
year will be required and, second, label updating will be
required any time during the production year design changes
have been incorporated that increase any one of three specific
design parameters beyond their highest tested vaxue. In the
first case, the manufacturer must relabel if any decrease in
fuel economy results, and in the second case, the manufacturer
will only relabel if the fuel economy decreases by a specific
amount.
2. For a review of the technical terms associated with the
fuel economy, see: EPA Report NO. EPA-AA-CPSB-82-02, "Current
Fuel Economy Program," September 1982.
-------
-4-
1. Mid-Year Label Updating
There are three basic costs associated with the
proposal to update label values at mid-year. These costs are:
updating the sales projections, recalculating the label values,
and the cost associated with the minimum data requirements.
The costs associated with sales projection
updating cannot accurately be predicted as each manufacturer
has its own technique to forecast sales and we do not have
basic data on the resources expended in these forecasts. How-
ever, for this analysis we will present what we believe to be
reasonable estimates. If comments to this rulemaking indicate
our estimates are significantly understated, we will update
this analysis for the final rule.
For the seven major manufacturers, which account
for 90 percent of current production sales, we will assume one
person week will be expended in updating the sales projections.
For the remaining 25 manufacturers we will assume two person
days will be expended in updating the sales. We will also
assume the person assigned to this task will be at the same
technical level as the programmer in Section "A", above an
hourly cost of $30 per hour. The estimated cost for sales
updating is the number of manufacturers times the cost per
manufacturer or,
Cost of Sales Updating = 1(7 mfrs.)(40 hrs) +
(25 mfrs)(16 hrs) J $30/hr.
= $20,400.
The cost of recalculating the actual label values
would involve recalculating the label values and then updating
the computer at each manufacturing site. Since the actual fuel
economy data will already be in the data base as well as the
updated sales, our estimate is that it will take 2 hours of a
programmers' time per manufacturer to recalculate the label
values for their product line. In addition, assuming $3U per
hour programmer cost, 100 assembly sites and that it would take
four hours to update each on-site computer, the cost of recal-
culating would be:
Cost of Recalculating = I(32 mfrs)(2 hrs per mfr) +
(100 sites) (4 hrs per site)]
$30/hr.
= $13,920
Under the current system over 90 percent of the
labels are changed at least once prior to proposed mid-year
updating time. The current need for updating these labels
comes about because the values on each label depicting the
-------
-5-
range of comparable fuel economy from similiar vehicles change
and, therefore, the manufacturers are already relabeling to
incorporate this revised range at the proposed mid-year time
period. Because the majority of manufacturers are already
relabeling at the mid-year time period, albeit to change the
range values, including changed fuel economy values on tne
labels will not result in cost increase for the manufacturers.
Because this proposal will also require the same
data coverage (assuming B above is adopted) as is being pro-
posed for initial labels, there will be some additional testing
burden. An analysis was made of the 1980 data base to deter-
mine what impact this proposal would have on testing resources.
(The 1980 data base was selected because it is the latest model
year for which the final CAFE calculations are complete for tne
industry.) Since we do not have a way to know when any running
change will be put into production (or if it will be) it is
extremely difficult to determine what a particular manufac-
turer's product line will "look" like at mid-year and thus,
what if any additional tests would be required to have the pro-
posed minimum data coverage. The way we estimated the testing
impact was to look at the product line configuration at the
end-of-year and compare it to the product line coverage at the
time the labels were first calculated. This is worst case in
that all the testing requirements will be assumed to have
occurred prior to the mid-year relabeling time and further that
none of the other proposed changes in this rulemaking would
cause the manufacturers to minimize design changes after
initial labeling.
In the 1980 model year data base, 39 of the 860
base levels did not have the highest selling configuration
represented at the end of the year. This would indicate that
approximately 40 additional tests would be required to satisfy
the mid-year relabeling data requirements. In a review of the
testing trends of fuel economy data vehicles it was noted that
80 percent of these vehicles were reconfigured vehicles. Tnus,
32 of the above identified tests would most likely be performed
on reconfigured vehicles and 8 tests on new vehicles.
In a recent review of vehicle costs from data
submitted from several manufacturers the following costs were
derived:
A new vehicle $23,200/vehicle
A reconfigured vehicle $l,500/vehicle
Test Cost $l,450/test
Mileage accumulation $1.82/mile
Typical mileage at the first test
point for a new fuel economy data
car 4,000-miles
-------
-6-
Typical mileage accumulated between
tests on a reconfigured vehicle 500-miles
The resulting cost per test for new and reconfig-
ured vehicles is as follows:
New prototype vehicles cost per test =
(vehicle cost) + (mileage accumulation costs) +
(cost/test) = ($23,200/vehicie) +
($1.82/mile x 4,000 miles) + ($l,450/test) =
$31,930
Reconfigured vehicle cost per test =
(vehicle cost) + (mileage accumulation costs) +
(cost per test) = ($l,500/vehicie) +
($1.82/mile x 500 miles) + ($l,450/test) = $3,860
Using the above information the minimum aata
costs will be:
Vehicle Costs for Minimum Data Coverage =
(new vehicle) x (no. tests) +
(reconfigured vehicle) x (number of tests) =
($31,930) x (8 new tests) + ($3,860) x
(32 new tests) = $378,960
The total cost for mid-year relabeling would be:
sales updating cost plus label recalculation costs plus the
cost of any additional vehicle requirements or,
Mid-Year Relabeling Cost =
(sales updating) + (recalculation) +
(additional vehicles) = $20,400 + $13,920 +
$378,960 = $413,280
2. Relabeling Due to Design Changes
The second proposal for relabeling is to require
the manufacturers to recalculate any label that has nad base
levels within the model type affected by any one of three
specific design changes. Then, if the recalculated label
values have decreased by a specific amount the manufacturers
would have to relabel.
In reviewing the 1982 model year data base 115 of
the 850 running changes involved one of the three specific
design changes which could trigger a recalculation, depending
on whether the changes decreased fuel economy. This woula not
equate to 115 additional tests because the proposal will only
require additional data if inertia weight, road-load horse-
power, or axle ratio are increased beyond the range represented
-------
-7-
by test data by significant amounts. For example, if the
highest tested axle ratio was 3.00 a manufacturer would have to
submit data and recalculate if an axle of 3.30 was added ^10
percent above largest tested). Based on the recalculations,
the manufacturer would relabel only if the recalculation indi-
cated a decrease in fuel economy of 1.0 mpg or more for the
city value or, 2.0 mpg or more for the hignway value.
In a review of one manufacturer's 1982 running
change submissions, less than half of the applicable running
changes increased the values above the tested range. For this
analysis, half of the running changes are .assumed to trigger
the need for recalculations.
In order to prorate the recalculation costs
($20,400 for sales updating and $13,920 for recalculation)3,
the following assumptions were made: First, each running
change would affect one base level. This is likely a worst
case assumption in that several axie ratios could be added to
the same base level and only one of these (the largest numer-
ical) need be tested. The second assumption is that each base
level affects three label values. This is appropriate because
the domestic manufacturers frequently have five or more labels
affected by a single base level and each base level in a for-
eign manufacturer's product line usually affects one label and
sometimes two labels.
Using the above assumptions, tne 115 running
changes (r/c's) would change 173 labels.
(115 r/c's) x (0.5 of total r/c's triggering
recalculation) x (3 base level per r/c) = 173
These 173 labels represent 11 percent of all J.982
labels and would cost the industry approximately $3,800 to
recalculate.
Recalculation cost due to design changes =
[(sales updating) + (recalculation)] x
(fraction of labels affected)
= [($20,400) + ($13,920)] 0.11
= $3,775
3. See Section C.I for details on these data.
-------
The additional data costs to update tor design
changes will be the number of vehicles multiplied by vehicle
costs. Since all of the data will be running change data it is
appropriate to assume that all of the data will be generated
using reconfigured vehicles. The approximate cost will be:
Additional Data Costs = (reconfigured vehicles)
(no. of tests) = ($3,860 per vehicle)(117 x .5)
= $225,810
The total cost for updating label vaiues aue to
design changes made during the model year is thus estimated to
be:
Updating Due to Design Changes =
(recalculation) + (additional data costs) =
$3,775 + $225,810 = $229,585
D. Modifications to the Label Information
The current fuel economy program requires one value
to be depicted on the label, the city value. This proposal
will require two numbers (a city and highway value) be depicted
on the label. In addition, two standard formats are proposed,
one of which will be adopted in the final rule. For the pur-
pose of this analysis, both of these changes will require a new
label format and some programming changes. At an estimated 24
hours for art work and reprogramming, $30 per hour for art work
and programming, and 8 hours to implement the software changes
on each of 100 manufacturer computers, the estimated total cost
of this proposal would be:
[(32 mfr)(24 hours) + (100 computer)
(8 hours/computer)]($30/hr.)
= $47,040
This cost does not reflect the time savings that will
be realized by the manufacturers by having a stanaard label
format which will eliminate all advance EPA approval of the
formats.
E. Proposed Technical Amendments to Cut Cost and Improve
the Data Base
1. Elimination of the Preliminary CAFE Calculation
The proposed regulations would eliminate the
preliminary CAFE and require that the final CAFE simply include
the data from vehicle configurations with total production
volumes of at least 90 percent of the manufacturer's totai
-------
-9-
model year production. Currently, preliminary CAFE aata are
required from the vehicle configurations which, taken in order
of decreasing sales, represent 90 percent or more of tne sales
of each significant base level and at least one test in each
nonsignificant hase level.
This proposal would result in some decrease in
testing and in other administrative costs associated with the
preliminary CAFE. First, the paperwork ana personnel time
associated with the generation of the preliminary CAFE would be
eliminated. This time currently involves both the original
generation of the preliminary CAFE values and recalculations to
include base levels added to manufacturers' product lines after
the initial calculation of the preliminary CAFE. Second, the
requirement of 90 percent coverage across the product line,
instead of coverage of the top 90 percent of each significant
base level as required for the preliminary CAFE, would provide
manufacturers greater flexibility in the selection of test
vehicles. Third, test vehicles can be more precisely targeted
to represent actual sales for the calculation of the CAFE
rather than estimated sales for calculation of the preliminary
CAFE which is calculated early in the model year, reducing the
need for additional tests where sales projections were inaccu-
rate. Finally, whereas the bulk of the final CAFE data are
currently required to be included in the preliminary CAFE sub-
mitted early in the model year, eliminating the preliminary
CAFE allows testing to meet CAFE requirements to be spread out
more evenly over the year resulting in less need for overtime
and more efficient use of testing facilities.
If the preliminary CAFE were eliminated resources
expended in compiling the list of all of the venicles, data
vertification, data requirements, etc., would be significantly
reduced. Thus, a very conservative estimate for the industry
assumes a time savings of 12 40-hour person work weeks would
likely be realized by each of the seven largest manufacturers.
(7 mfrs)(12 weeks per mfr)
(40 hours per week)($30 per hour) =
$100,800
Additional smaller time and cost savings would be realized oy
the other manufacturers but is not accounted in this analysis.
The elimination of preliminary CAFE along with
relaxing the data requirements for the final CAFE calculations
should result in a reduction in the number of test venicies.
In 1981, there were over 200 tests which were conducted soley
for CAFE purposes. With the simplified and more tlexible test
requirements, each manufacturer will be able to more precisely
target the need for test vehicles. While it is not possible to
accurately determine the test savings, we believe it well
-------
within reason to expect a reduction in the number of test
vehicles from 200 to 190 or five percent. As indicated in
Section "C," 80 percent of the tests are performed in recon-
figured vehicles with the remainder requiring new vehicles.
New Vehicle Cost Savings = (vehicle costs) x
(no. of tests)
= ($31,930) x (2)
= ($63,860)
Reconfigured Vehicles Cost Savings = (cost per test) x
(no. of testsj
= ($3,860) x (8)
= $30,880
Total Cost savings = New Car + Reconfigured
= $63,860 + 30,880
= $94,740
Elimination of the preliminary CAFE would also
eliminate the exemption from submitting running change fuel
economy data where a manufacturer's preliminary CAFE level is
sufficiently above the applicable model year standards. About
17 percent, or 22, of the fuel economy test requirements are
annually eligible for exemption. Using the cost of reconfig-
ured vehicle the cost of discontinuing tne running change
exemption provision would be:
Cost = [(reconfigure vehicle) +
(mileage accum) + (cost of test)]
(No. of tests) =
[($1,500 per vehicle) + ($1.82 per mile)
(500 miles) + ($1,450 per test)] (22) = $84,920
Thus a conservative estimate of the net savings from this
proposal would be over $110,000 annually:
Additional Cost
Cost = of Running Change
Testing
Cost Savings
from Elimination
of the Preliminary
CAFE Calculation
Cost Savings
from Decreased
Testing Because
of More Precise
Targeting of
Test Vehicles
= $84,920 - $100,800 - $94,740
= - $110,620
-------
-11-
2. Fuel Economy Adjustments for High Mileage Test
Vehicles
Currently, the regulations allow a maximum mile-
age accumulation of 10,000 miles for fuel economy vehicles.
This proposal will continue to allow a maximum of 10,UOO
miles. However, under this proposal all fuel economy data
generated on a vehicle/engine system combination with over
6,200 miles will be adjusted to approximate the fuel economy at
4,000 miles. This proposal will not increase the number of
tests required since running beyond 6,200-mile is still an
option.
3. Drive Train Separation
This part of the proposal recommends that front-
and rear-wheel drive systems be separated into different base
levels. Since we are already separating these drive systems
under the general authority of the current regulations this
part of the proposal simply makes current practice explicit in
the regulations. No additional costs will be involved. In
addition to the drive system separation, above, lockup auto-
matic transmissions and all transmissions with overdrive
gearing will be separated into separate fuel economy base
levels. Since there are very few cases where this separation
would be involved, this proposal has no significant industry-
wide impact.
4. Interior Volume
Changes are being proposed to the measurement
methodology to account for (1) hatchback and station wagon
interior volume, (2) driver and front-seat leg room, and ^3)
luggage measurements if the two-seater vehicle classification.
These proposed changes will not affect the number of labels,
test vehicles, or the vehicle classificaton system and, as a
result, should not impose any additional cost on the industry.
IV. Summary
The annual cost of all of the proposals in this NPRM
package for all manufacturers is estimated to be $626,325.
The attached table summarizes the estimated cost of each
proposal. Some cost savings to manufacturers which are diffi-
cult to quantify, such as the possiblity of more efficient use
of test facilities because of the elimination of the PCAFE,
have not been calculated and included in the net cost of the
proposals in the NPRM.
-------
-12-
Cost Summary of the
Proposed Rulemaking
Regulation Issue Estimated Cost
1. In-Use Adjustment $ 47,040
2. Minimum Data Requirements Negligible
3. Relabeling
a. Mid-year $413,280
b. Design Changes $229,585
4. Label Format $ 47,040
5. Technical Amendments
a. Eliminate PCAFE -$110,620
b. High Mileage Adjustment -0-
c. Drive Train Separation -0-
d. Interior Volume -0-
Total $626,325
------- |