Cohenetal_RuralDelivery.txt

RURAL DELIVERY AND THE UNIVERSAL SERVICE OBLIGATION: A
Quantitative Investigation
Robert H. Cohen William W. Ferguson Spyros S. Xenakis
Postal Rate Commission Office of Technical Analysis and
Planning
July 31, 1992

1. Introduction

1.1. Scope and Purpose
It is widely believed that it costs more to provide rural areas
with postal service than urban areas. This belief is based
primarily on the perception of a cost differential between rural
and city delivery.1 This perception is one of the bases for the
argument that a universal service requirement is necessary to
assure the continuation of rural delivery or at least the level of
service currently accorded rural areas.2
The purpose of this paper is to analyze rural delivery costs and
compare them with city delivery costs. Using routine cost data
submitted in the course of postal rate proceedings, Section 2 of
this paper compares the cost of rural and urban delivery, Section 3
shows the relationship of rural delivery cost to population
density, Section 4 analyzes the "profitability" (viz., contribution
to overhead) of serving rural areas, and Section 5 presents some
concluding remarks and a brief summary. Before turning to Section
2, some background information on rural delivery and city delivery
is offered.
The views expressed in this paper are those of the authors and
do not necessarily represent the opinions of the Postal Rate
Commission.
1
According to the 1991 Comprehensive Statement on Postal
Operations (p. 66), USPS delivery costs represent 30 percent of
total costs; window service and mail processing, 32 percent;
transportation, 7 percent; and administrative, building occupancy,
and all other, 31 percent.
2
This paper focuses on the detailed comparison of delivery costs,
between rural and urban areas. It does not purport to show the
bearing, if any, this cost analysis might have on questions
regarding the complex issue of the postal monopoly, which
encompasses many issues beyond the scope of this paper.


1.2. Relation of Rural Routes to Demographically Designated
Rural Areas
The Postal Service has about three times as many city delivery
letter routes as rural routes. City delivery routes serve
geographic locations within the boundaries of a post office, while
rural routes generally serve areas falling outside these
boundaries.3 City and rural carriers are in separate labor unions,
and their compensation is determined separately based on different
factors.
The United States Census Bureau (1990 data) reports that
slightly under 25 percent of the 250 million people living in the
United States live in rural areas.4 The remainder live in urban
locations.5,6,7
In 1991, the United States postal system provided service to 102
million delivery points. City routes served 78.5 million and rural
routes (including contract routes) served 23.4 million delivery
points.8,9 A total of 95 million delivery points
3
United States cities and the areas served by their post offices
often expand to absorb surrounding areas served by rural routes. It
takes a great deal of time to administratively convert a rural
route to a city delivery route. Thus, some rural routes will serve
areas annexed by cities and their post offices.
4 Bureau of the Census press release, CB91-334, Dec. 18,
1991.
5 As defined for the 1980 census, urban areas include: (a)
places of 2,500 or more inhabitants incorporated as cities,
villages, boroughs (except Alaska and New York) and towns (except
in the New England states, New York and Wisconsin), but excludes
those persons living in the rural portions of extended cities; (b)
census designated places of 2,500 or more inhabitants; and (c)
other areas, incorporated or unincorporated, included in urbanized
areas. All non-urban areas are rural.
6
The Census Bureau reports that in 1980 there were 6,619
incorporated places with more than 2,500 inhabitants. In 1991, the
United States Postal Service provided city delivery to 6,625 post
offices, but not all were in incorporated areas.
7
The United Nations has estimated that 27.5 percent of the
population in the more developed nations would live in rural areas
in 1990. See World Population Trends and Policies, 1987 Monitoring
Report, The United Nations, Department of International Economic
and Social Affairs, Population Division, page 176.
8 Comprehensive Statement on Postal Operations, 1991, p. 49.
9
Highway contract routes (star routes) are similar to rural
routes. The difference is that the carrier is a contractor to the
Postal Service, not an employee. Contract routes serve 1.4 million
delivery points.
are residences; the remainder are businesses. Thus, rural (and
highway contract) routes serve slightly less than 25 percent of
total residential delivery points; this is the same as the
percentage of the population living in rural areas.


1.3. Description of Rural and City Delivery
All rural routes use vehicles to deliver to a box placed along
the roadside, and virtually all provide six-day-a-week delivery. A
rural route is defined in terms of the roads it traverses. Homes or
businesses not located on one of these roads must place a mail
receptacle along the route traveled. For this reason, boxes will
frequently be clustered where a rural route intersects roads not on
the route. In this sense, rural service is inferior to city
delivery where service is provided to (or in close proximity to)
each building served.
Most city delivery routes are "park-and-loop" routes. The
carrier on these routes uses a vehicle to drive to various points
along the route where the carrier dismounts and delivers to a
portion of the route on foot. Some city delivery routes, called
"curb line routes," use vehicles to provide curbside delivery to a
mail receptacle along the curb as is done by rural routes. A third
type of city delivery route is the "foot route."
City routes are further categorized as "business,"
"residential," and "mixed" (business and residential) routes.
Business routes (consisting of at least 70 percent business
deliveries), which account for less than one percent of all
possible city deliveries,10 are five-day-per-week routes. Mail on
all other city routes is delivered six days a week. Very few routes
service businesses or residences exclusively. For example, on
residential routes (which account for 94 percent of all possible
city deliveries), businesses account for five percent of possible
deliveries. The percentage of possible deliveries on rural routes
that are businesses is not known. Although the activities of city
and rural carriers are similar, some minor differences exist:
-rural carriers spend about three percent of their total time
providing retail services
10
"Possible delivery" is used to describe a household or business
address (including apartments and suites) to which mail might be
delivered by city carriers. "Box" is used to describe the
receptacles each family or business sets up on a rural route to
which mail might be delivered. The two terms stand for similar
concepts and for purposes of this paper are used
interchangeably.
(e.g., selling postage stamps, rating and receiving parcels,
etc.);
-city carriers provide no retail services;
-rural carriers transport mail to and from small post offices
along their route; and
-city carriers sweep more collection boxes along their routes
than do rural carriers.



2. Comparison of the Cost of City and Rural Delivery

2.1. Comparison of Carrier Time Required to Serve City and
Rural Areas
Table 1 shows that the average time per day per possible
delivery is 1.04 minutes for city delivery11 and 1.07 minutes for
rural delivery. (This small difference would be further reduced if
retail activities of rural carriers were not counted.) The virtual
equality of the average carrier time to serve urban and rural
customers is a major finding of this paper. Part of the explanation
for this finding can be inferred from table 1- businesses require
considerably more time per possible city delivery than do
residences.
The average possible delivery on business routes requires five
times as much carrier time as the average possible delivery on
residential routes. There are several reasons for this large
difference in carrier time:
-Businesses receive almost three times as many pieces per
possible delivery as do residences. The statistics shown in table 1
(row 2) reflect both in-office and out-of-office delivery time.12
In-office time is closely related to volume.
-In business areas, carriers travel by foot, and many deliver in
large office buildings to individual suites.
11
This analysis includes only city delivery letter routes. It
excludes parcel post and support routes which primarily serve
business districts in larger cities.
12
City and rural carriers typically spend between three and four
hours in the office preparing for delivery.
Table 1
Selected Statisticsa for City Delivery and Rural Delivery Routes
(1989)
City Item Residential Park & Foot Curb Loop
City Residential, Business All Businessb and Mixed Rural Park
& Total All Foot Curb Loop Total


1.
Possible Deliveries11.5 15.1 46.9 73.4 0.6 12.6 15.7 49.7
78.1 20.5 (millions)



2.
Minutes per Day per Possible Delivery 1.00 0.83 1.01 0.97
5.55 1.21 0.85 1.06 1.04 1.07 (in-office and

out-of-office)



3.
Seconds per Piece13.24 9.54 12.93 12.22 21.44 14.78 9.69
12.98 12.50 14.94

(in-office and out-of-office)


4.
Pieces per Day per Possible Delivery 4.53 5.21 4.70 4.78
15.52 4.89 5.28 4.92 5.01 4.30


a See Appendix A for source.
b
Reflects the fact that business routes receive service only 5
days per week.
-There are no curb delivery routes in business areas. Curb
routes are suitable only for residential areas. Table 1 (row 2)
shows that they require less time than either foot or park-and-loop
routes.
On the other hand, a direct comparison of rural to city
residential delivery reveals that rural carrier time per possible
delivery is only ten percent greater than city residential route
delivery time. (Seven percent higher if retail service costs are
eliminated from rural.) There are at least three major reasons for
rural route time per possible delivery being so close to the
corresponding time for city residential routes:
-Rural routes are the functional equivalent of curbline city
delivery, the most efficient form of city residential delivery.
Curbline routes, however, account for only 21 percent of city
residential possible deliveries.
-Rural routes have only 4.3 pieces per possible delivery, while
city residential routes have
4.8. Thus they incur less in-office costs.13
-As described earlier, rural customers who do not live along the
rural route must place a mail receptacle along the rural route.
Thus, rural mail boxes tend to cluster where roads (not on the
route) intersect with the carrier's route. A rural carrier can
serve a cluster of boxes much faster than if the individual boxes
were spread out along the route where the carrier would have to
slow down, stop, and accelerate for each one.
Thus far, we have examined carrier time per possible delivery or
per box. We now turn to the carrier time per piece delivered. Table
1 (row 3) shows that carrier time per piece delivered is 12.5
seconds per piece for all city and 12.2 seconds for residential
city. This contrasts with 14.9 seconds per piece for rural routes.
Thus, on a delivered piece basis, rural routes use 20 percent more
carrier time than do all city routes, and 22 percent more than city
residential routes. The major explanation for this is fewer average
pieces per possible delivery per day for rural routes (row 4).
13
Part of the reason for the difference in pieces per possible
delivery may be due to the fact that five percent of possible
residential deliveries are businesses, and it is thought, but not
known, that a lesser percentage of possible deliveries on rural
routes are businesses.


2.2. Labor Cost
The previous section compared city and rural delivery on the
basis of time. This section estimates postal labor delivery cost
that corresponds to units of time in order to convert time into
money in subsequent sections.
The Postal Service uses both full-time regular carriers and
casual employees (less than full-time or temporary) on its city and
rural routes. Casual employees are paid lower wages and have fewer
fringe benefits. Consequently, their cost to the Postal Service is
far less than for full-time employees.
Full-time rural carriers' compensation is slightly lower than
full-time city carriers. They incur less overtime and the rural
carrier work force has a higher proportion of casual employees. As
a result, rural carrier labor cost to the Postal Service in 1989
averaged $20.60 per work hour, or 34.3¢ per minute.14,15 In
contrast, in 1989 city carrier labor cost the Postal Service $24.49
per work hour, or 40.8¢ per minute.16
The difference in labor costs for rural and city carriers has
its roots in the development of the two crafts. In another postal
system there might be no differences in the compensation of city
and rural carriers or it might be much larger. In the United
States, rural wages are generally lower than urban wages.
For purposes of city and rural delivery cost analysis, we
present (1) a comparison using the actual labor costs of the two
crafts, and (2) a comparison using the average labor costs of all
Postal Service collective bargaining employees. This will allow
both an actual cost analysis and a resource comparison of city and
rural routes. The average labor cost per bargaining unit employee
in 1989 was $24.09 per work hour or 40.2¢ per minute.17 This is
very close to the average city carrier cost.
14
Highway contract carriers have compensation much lower than
rural carriers. Because they are not postal employees, we do not
include them in postal labor costs.
15
See Appendix B for derivation. Includes wages, premium payments
(e.g.,

17 See Footnote 15.


2.3. Delivery Vehicle Cost
Rural carriers furnish their own vehicles and provide all
maintenance, repairs, and fuel, for which they are paid an
allowance.18 In 1989, rural carriers received an average of 34
cents per mile as a motor vehicle allowance. The average length of
a rural route is 55 miles. The average annual cost per rural route
is shown in table 2.
Those city delivery carriers who make use of a vehicle are
furnished with one by the Postal Service which also provides all
maintenance, repairs, and fuel. City carriers drive an estimated 15
miles per day. Analyzing Postal Service accounts for depreciation,
fuel, and maintenance for city delivery carriers, we have estimated
the average city delivery vehicle cost per route.19
This is also shown in table 2.

Rural vehicle cost per box or possible delivery is twice the
average city carrier vehicle cost per possible delivery. Rural
vehicle cost per delivered piece is two and a half times city
carrier cost per delivered piece. It should be borne in mind that,
though the "city" column divides total city vehicle cost by
18
In 1989, the rural carrier vehicle allowance was 31 cents per
mile or a minimum of $12.40 per day, whichever was greater.
19 See Appendix C.
the total number of city routes (including foot routes), only 84
percent of possible city deliveries are made by city carriers using
vehicles.


2.4. Comparison of Direct Labor Plus Vehicle Cost to Serve City
and Rural Areas
Table 3 combines labor cost with vehicle cost for city and rural
carriers. It shows that when vehicle costs are added, the
difference in cost per box per day between city and rural carriers
depends heavily on which labor cost is used. Using actual labor
costs, the city cost per box per day is 7.5 percent higher. Using
the average bargaining labor cost, city delivery is 8 percent
lower. On a cost-per-piece basis, city costs are 8 percent lower
using actual labor costs and 21 percent lower using the average
bargaining labor costs.




3. Relation of Rural Delivery Cost to Population Density
A priori, population density should have an important effect on
rural delivery cost. We have no data available which directly
relate rural delivery cost to population density, but it seems very
likely that boxes per mile is highly correlated with population
density.
In order to examine the impact of density on cost, rural routes
have been divided into quintiles based on boxes per mile. Table 4
displays the relevant data. It can be seen that there is
wide variation between the quintiles. In the extreme, the
average number of boxes per route differs by a factor of two. The
average number of miles differs by a factor of 4.5. Average boxes
per mile differ by a factor of nine. Moreover, with the exception
of daily evaluated time and daily pieces delivered per box, the
average values of all variables change monotonically. Thus, they
are correlated with boxes per mile.
Table 4
Distribution of Rural Routes by Density (Boxes per Mile)
Selective Averagesa (1989)

a See Appendix A for source.
The first or least densely populated quintile stands out from
the remaining four. Its time per box is half again larger than the
mean for all rural routes and it is nearly two standard deviations
greater than the mean for all rural routes. The other four
quintiles are less than one standard deviation from the mean.
Moreover, the first quintile stands apart in that its seconds
per piece is also nearly two standard deviations greater than the
mean, while the other four are all less than one standard deviation
from the mean.20 Thus, the two measures of cost for the first
quintile are substantially greater than for the other four
quintiles. This will be seen clearly in Section 4, where the
profitability of rural delivery is calculated for each
quintile.
It would not be surprising to find that the percentage of boxes
which are businesses increases as population density increases. If
this is true, it could at least partially explain why the number of
pieces delivered per box is so much larger for the fourth and fifth
quintile than it is for the other three.
Finally, the variability (or elasticity) of time with respect to
volume for the five quintiles differs greatly:
All 1st 2nd 3rd 4th 5th Routes
.29 .37 .44 .53 .57 .44
Thus, for example, if the volume in the first quintile were to
double, total cost would increase by 29 percent. The variation
between quintiles can be explained by greater fixed costs in the
less densely populated quintiles than in the more populated
quintiles. The time required to drive the route is fixed, and
driving time represents a greater proportion of total cost in the
less densely populated quintiles.
Figure 1 displays the minutes per box per day as a function of
density. It can be seen that time per box drops off sharply at the
low end of the density spectrum and then it flattens.
20
The daily evaluated time per box for all routes (1.07 minutes)
has a standard deviation of 0.30 minutes. The daily evaluated time
per piece for all routes (14.94 seconds) has a standard deviation
of 4.64 seconds.
2.400000
2.200000
2.000000
1.800000
1.600000
1.400000
1.200000
1.000000
0.800000
0.600000 0 5 1015202530354045

In order to isolate the effect of density on time per box, the
elasticity of time per box with respect to density has been
calculated using the route evaluation factors (which are used to
determine rural carrier pay). Holding constant the pieces per rural
box and route miles, a one percent increase in density reduces
out-of-office time per box by 0.44 percent and total time per box
by 0.27 percent.21
21
The relation between time per box and boxes per mile (density)
is exponential, or linear in logarithms. The Pearson correlation
coefficient between the logarithmic values of the two variables is
-78.60 percent. The econometrically estimated constant elasticity
coefficient of a simple log- linear model with time per box as the
dependent variable and boxes per mile as the explanatory variable
is -27.45 percent with a t value of -266.22.


4. "Profitability" of Serving Rural Areas22
This section addresses the question: Does the United States
Postal Service find it remunerative to serve rural areas? A
revenue/cost model of rural delivery has been constructed to
determine the profitability23,24 of rural delivery by quintile. Its
simple structure is presented below:
Model for Calculating Profit (Loss) Per Box

22
"Profit" here means contribution to fixed overhead costs over
and above paying the fixed overhead of rural delivery.
23
The model presented here is valid only for an unsubsidized
postal system such as the Postal Service.
24
The simplifying assumption is made that there is no cost
difference between city and rural mail with respect to mail
processing, transportation, and retail service.
Line 1-the revenue per rural box per day is calculated by
multiplying average Postal Service revenue per piece (from the
Revenue, Pieces and Weight Report for 1989) times the average
number of pieces per box per day (from table 4). Line 2-rural
delivery cost per box per day consists of (1) labor cost,
(2) vehicle cost, and (3) indirect costs.25 Labor cost per box
is obtained by multiplying labor cost per minute (from Section 2.1)
by the average number of minutes per box (from table 4). Vehicle
cost per box per day are from table 2.26 Because there are two
labor costs per minute, one actual and the other theoretical, the
model is used separately for each labor cost and provides two
annual profit (loss) computations.
Line 3-here the model takes into account the cost of getting
mail to the point of delivery (e.g., processing, transportation,
administrative, retail, etc.). This is done by bifurcating the
average attributable cost27 per piece into delivery attributable
cost28 and nondelivery attributable cost.29 Rural carrier cost and
the nondelivery attributable cost are then subtracted to arrive at
a profit (loss) per box per day.
Observers familiar with Postal Rate Commission costing
procedures will recognize that rural delivery cost per box includes
the attributable and institutional costs associated with that
function. The Commission found in the most recent omnibus rate case
(Docket R90-1), that 39 percent of rural delivery costs are
attributable (i.e., vary with volume) and 61 percent are
25
Appendix D presents indirect costs in more detail and shows how
they are calculated.
26
Dr. John Haldi has brought to our attention that the daily
vehicle cost per box is not the same for all quintiles, as was
assumed in an early version of the paper, but decreases with
population density. The estimated daily vehicle cost per box in the
revenue/cost model now recognizes the density differences among the
five quintiles:

27
Attributable costs are the postal costs causally traceable to
mail. They consist predominantly of volume variable costs. The
remaining costs are called institutional costs, and they can be
treated as fixed. In the United States, the Rate Commission
determines the attributable cost for each class and subclass of
mail. Its most recent analysis is contained in the Docket R90-1
Opinion and Recommended Decision.
28
These are the attributable (direct and indirect) costs arising
from the delivery function. We have estimated them for purposes of
this analysis.
29 These are total attributable costs minus delivery
attributable costs.
institutional (i.e., nonvariable).30 Because such a large
percentage of the costs are fixed, the profitability calculation is
sensitive to both the average revenue per piece and the average
pieces per box. If the postal system operated at a lower scale,
revenue per piece would need to be higher to maintain the same
revenue cost balance. It is not clear, without further analysis,
what would happen to profitability if the system operated at a
lower scale.31
The average revenue per piece for the entire Postal Service is
used to estimate the average revenue per box. We have no
independent estimate of the average revenue for mail delivered by
rural carriers. If mail delivered by rural carriers has a different
composition compared to the system as a whole, profitability
conclusions would vary. In that case, the nondelivery attributable
costs would vary in the same direction as revenue, but not by
enough to offset the revenue change.
Table 5 provides both the input and output for the profitability
calculations for all rural routes in 1989. It can be seen that by
serving all rural routes and using actual labor costs, the Postal
Service realized an average profit of 10.8 cents per box per day,
or a total annual profit for all rural routes of $669 million. To
put this figure in perspective, the total accrued expenditures for
the Postal Service in 1989 were $39 billion.32
Table 5 shows that, using the average labor cost for all
bargaining employees, the profit drops to $283 million. The profit
from serving all rural routes is obviously highly sensitive to the
labor cost of rural carriers. However, using either labor cost
figure, rural delivery was profitable for the Postal Service.
If, in 1989, the Postal Service's overall surplus had been much
larger, the profit from rural delivery would no doubt have been
greater. Conversely, if the year had been one in which the Service
had a sizeable deficit, rural delivery would have been less
profitable. Fortunately for purposes of this analysis, the Service
essentially broke even in 1989 and so the profit from
30
In Docket R90-1, the variability or elasticity of evaluated
rural route costs with respect to volume was estimated at 44
percent by the Postal Service.
31
It does seem clear that the ratio of mail delivered to rural
boxes to mail delivered to city addresses is an important factor in
determining profitability.
32 The Postal Service surplus for 1989 was under $100
million.
rural delivery need not be interpreted based on the Postal
Service's overall financial results.
Table 6 displays the profitability of all five quintiles of
rural routes (based on population density) using both actual
compensation and average bargaining compensation. It can be seen
that the profit per box differs substantially from the first (or
least densely populated) quintile of rural routes to the fifth (or
most densely populated) quintile. Using actual labor cost only, the
first quintile was unprofitable, while, using average bargaining
labor cost, the first two quintiles were unprofitable.
Because the most densely populated quintiles of routes serve
more boxes, their total profit is disproportionate to their per-
box profit. Using either labor cost, the third, fourth, and fifth
quintiles were profitable.

Given the assumptions discussed above, rural delivery is
remunerative and it is unlikely that it would be abandoned if the
universal service requirement were eliminated. Some observers,
however, might expect the Postal Service to either drop or reduce
the level of service to the boxes in the first or second quintiles
of rural routes. The first quintile comprises only 2.5 percent of
all addresses served by rural and city carriers combined. The
second serves 3.9 percent.



5. Concluding Remarks and Summary

5.1. Concluding Remarks
While the boxes served by the quintile of routes serving the
least densely populated areas are unprofitable, we believe that it
is unlikely that the Postal Service would discontinue service to
them (or try to decrease their level of service) if the universal
service requirement were eliminated.
-The total loss on those boxes is small relative to total costs
of the Postal Service.
-Because these routes are scattered all over the country, boxes
on these routes are not easily identifiable without consulting an
extensive list. Consequently, it would be costly for firms to
separate mail addressed to these boxes from their remaining
mail.
-The transaction costs involved in putting pieces addressed to
boxes in the first quintile of rural routes in the hands of another
delivery firm, which would serve these addresses, would also be
high.
-If these addresses were dropped from the delivery network, it
would likely reduce the volume of mail sent by these addresses to
the remaining portions of the delivery network. Thus, profitable
volume would be lost.
Perhaps the above four points are simply underlying reasons for
the truism that for common carriers serving the general population,
larger service networks (be they mail, package, overnight, or
telephone) are more valuable to customers and providers than
smaller service networks. It is no accident that, within the United
States, United Parcel Service provides ubiquitous service for
parcels.33 Federal Express and other overnight carriers do the same
for overnight delivery. Moreover, the major long distance telephone
carriers also provide ubiquitous service. Quite possibly, all of
these common carriers find that sparsely settled portions of the
country are unprofitable to serve. That these organizations provide
universal service suggests that rural areas would receive postal
service even absent a universal service requirement.


5.2. Summary
-In the United States postal system, there is no real difference
in the carrier time required to serve city and rural addresses.
-The average city delivery cost reflects the higher cost of
serving businesses compared to residences.
-Rural delivery cost reflects a lower level of service than city
delivery.
33
UPS does not provide parcel service at the ordinary rate to the
Alaska bush where there are no roads and service must be provided
by air.
-Because fewer pieces are delivered per box on rural routes than
per possible delivery on city routes, the per piece delivery cost
is higher for rural routes than for city routes.
-The cost of delivery per box for the least densely populated
quintile of rural routes is much higher than the average for all
rural routes.
-The revenue from mail delivered to rural areas as a whole
exceeds the cost of handling and delivering that mail.
-There is a loss on serving the least densely populated quintile
of rural routes.
-It is likely that if the universal service requirement were
eliminated, even the most sparsely populated rural areas would
receive service.
Appendix A. Major Data Sources


A.1. City Delivery Carrier Data
The 1989 city delivery carrier data used in this paper are based
on information from several Postal Service data systems. Total city
delivery carrier work hoursa come from payroll hours accounting
systems data made available in the most recent rate proceeding.
These work hours are apportioned among city delivery carrier route
types on the basis of cost allocations from the In-Office Cost
System,b an ongoing work sampling system that is used to allocate
costs for certain labor crafts among different activities and rate
categories for ratemaking purposes.
Information concerning the total number of possible deliveries
is taken from the 1989 City Delivery National Totals Report.
Carrier Cost System (CCS) data are used to allocate total possible
deliveries among different route types and to determine the average
pieces per possible delivery for different delivery and route types
on letter routes. The CCS is used in rate cases to determine
attributable costs and associated distribution keys for certain
city carrier activities. CCS data are collected throughout the year
from over 500 thousand sampled stops on city delivery letter
routes.


A.2. Rural Delivery Carrier Data
Most rural routes are evaluated routes. Evaluated routes are
those routes for which the rural carrier's annual salary is
calculated using a set of standard time allowances. Time standards
are applied to workload elements (e.g., mileage, delivery boxes,
quantity of mail by shape, etc.) to calculate the total evaluated
time required to serve a rural route, and thus the salary of the
carrier serving that route.
To measure the workload elements needed to calculate the
evaluated time, a National Mail Count for most rural routes is
conducted periodically in accordance with the labor agreement
between the United States Postal Service and the union of rural
carriers. The statistics on rural routes presented in this paper
are based on the 1989 National Mail Count data.c
a
Docket R90-1, USPS-LR-F-342.
b Docket R90-1, USPS-T-13, W/S 7.0.7, p. 1.
c
The Postal Service used a sample of the 1989 National Mail Count
data to measure the elasticity (variability) of rural carrier costs
with respect to mail volume. Docket R90-1, USPS-T-13, Appendix
F.
The 1989 National Mail Count was conducted for 24 delivery days
from September 5 to October 2, 1989, and included 44,775 rural
routes out of a total of 46,197. Data for a few of the counted
rural routes appeared to be internally inconsistent. Counted routes
that had one or more of the following properties were deleted:d

-A difference between weekly actual and evaluated time greater
than 1,000 minutes.
This edit resulted in 931 deletions. Data from the remaining
43,844 rural routes were used to calculate the statistics presented
in this paper.
d The Postal Service performed a similar edit on the sample of
routes selected for measuring the elasticity of rural carrier
costs. Docket R90-1, USPS-T-13, Appendix F, p. F-8.


a
PRC Library Reference 5, Docket R90-1. b USPS City Delivery
Statistics National Totals, FY 1989.
c
L.1/L.2. d Six (6) delivery days per week times 52 weeks less 10
holidays.
e
L.3/L.4. f L.1/L.6. g L.7/L.4. h L.8/5.01. The denominator
(5.01) is the average number of pieces per
day per possible delivery for city delivery routes from table
1.

a
Costs for rural carriers estimated as a percentage of rural
carrier direct labor costs to total labor costs.
b
46,197 rural routes from the USPS FY 1989 Rural Delivery
Statistics National Report times 442.79 (the average number of
boxes served per route) from the 1989 Rural Carrier National Mail
Count.
c
L.4/L.5.
d Six (6) delivery days per week times 52 weeks less 10
holidays.
e
L.6/L.7.