BEFORE THE
POSTAL RATE COMMISSION
WASHINGTON, D.C. 20268-0001
POSTAL RATE AND FEE CHANGES, 1997
Docket No. R97-1
DIRECT TESTIMONY
OF
SANDER GLICK
ON BEHALF OF
MAGAZINE PUBLISHERS OF AMERICA
I
2
3
I.
4
Performance
5
PPC
6
consulting
7
rapidly since our inception
8
500, a compilation
9
joining the firm, I have performed
10
Autobiographical
My name
Sketch
is Sander
Corporation
provides
A. Glick.
(PPC), a consulting
management,
services
and governmental
I am a Senior
information
Analyst
at Project
firm based in Sterling,
technology,
and
to private and public sector clients.
Virginia,
environmental
The firm has grown
in 1991; last year we were number 272 on the Inc.
of the fastest growing
private companies
economic
in America.
and cost analysis
Since
for both private
clients.
II
12
I attended
the Maxwell
University,
School
13
Syracuse
where I received
14
in 1994, and Carleton
15
Physics in 1993. While at Syracuse
16
Center for Technology
17
administering
18
companies
19
Research
College,
Science
where
I received
University,
F’ublic Affairs
I was a graduate
Foundation-funded
at
degree
a BA. magna cum laude, in
Policy and assisted
the costs and benefits
and Development
and
a Master of Public Administration
and Information
a National
regarding
of Citizenship
assistant
in the
in developing
and
survey of more than 500
of working
with Federally-funded
laboratories.
20
21
Following
my formal education,
22
the end of 1996, I was promoted
23
assisted
24
life-cycle
cost
25
collecting
data to implement
26
compliance
27
estimators.
the Department
cost estimates
to Senior Analyst.
of Energy
of cleaning
I joined PPC in 1994 as an Analyst.
by developing
up nuclear
weapon
the analysis.
Since joining
methods
production
I
PPC, I have
for estimating
sites
I have also developed
and reviewed cost estimates
prepared
At
and
the
then
regulatory
by other cost
I
II.
Purpose
of Testimony
and Summary
Conclusions
2
In this testimony,
3
I review the Postal Service’s
method
for determining
4
rural carrier salaries and the Postal Service’s
5
I find that the Postal Service’s costing methodology
violates the long established
6
principle
must be consistent
7
way the cost is incurred
8
results in an anomalous
9
delivered
that the distribution
rural carrier costing methodology,
of a cost to subclass
and the attribution
methodology.
result: the cost distributed
with the
This inconsistency
to a subclass
of mail per flat
is about 15 percent higher than the amount the rural carrier is actually
10
paid to deliver a flat while the cost distributed
II
than the amount
12
anomaly
and to make rural carrier cost distribution
and attribution
13
propose
an improvement
proposed
14
distributing
the rural carrier
per letter is about 15 percent lower
is paid to deliver
to the Postal
Service’s
a letter.
To correct
this
consistent,
methodology
I
for
rural carrier costs to subclass.
15
16
Ill.
Rural Carrier
Salaries
17
Unlike city carriers who are paid on an hourly basis, the Postal Service
18
19
pays rural carriers on evaluated
20
comprise more than 90 percent of salary costs for rural carriers) based upon the
21
amount
22
example,
23
regardless
of work they perform
routes (salaries for carriers on evaluated
(e.g., the number
of letters they deliver).
routes
For
a rural carrier is paid for .0791 minutes for every letter he delivers,
of how long it actually takes him to deliver the letter.
24
25
Table 1 shows the evaluation
factor, or minutes allowed per unit of work
minutes
delivered),
26
(e.g.,
allowed
27
measures
28
the average FY 1996 rural carrier salary of $21.07 per hour (I-R-H-212,
29
Line 63, Column E), would be paid for performing
(USPS-T-5,
per
letter
for all rural
carrier
workload
WP-B, W/S 10.1 .l) and the amount a carrier, being paid
2
one unit of each task.
W/S-I,
Table 1. Evaluation
3
4
For example,
Factors
for Each Task
a carrier being paid the average rural carrier salary would
Because the
5
be paid five cents to deliver a flat and 2.8 cents to deliver a letter.
6
“average”
7
five cents is the volume variable (or marginal) rural carrier cost for delivering
8
flat.
carrier is paid five cents to deliver a flat (regardless
of the volume),
one
9
10
For 1996,
the Postal
Service
based
11
evaluations
[the National Mail Count] conducted
I2
of United
States
13
evaluations,
Postal
Service
“rural
carrier
on route
in the fall of 1995.” (Response
to MPAIUSPS-Tl7-10).
the Postal Service counted the workload
3
salaries
On these
of individual
route
rural carriers
I
for each route evaluation
item shown in Table 1. To determine
2
hours for which each rural carrier would be paid, the Postal Service multiplied
3
the count for each route evaluation
4
then summed hours across all route evaluation
5
individual
6
“evaluation”
item by its respectke
the number of
evaluation
factor and
items. The FY ‘I 996 salary of an
rural carrier was based upon the number of hours calculated
conducted
from the
in the Fall of 1995.
7
Because
8
9
carriers are paid based upon workload,
hours, a perfect costing method would distribute
rather than actual work
the amount a carrier is paid to
IO
perform a unit of workload
II
five cents, the amount a carrier is paid to deliver a flat, to subclass for every flat
12
delivered).
13
and therefore
14
costing system, at a minimum, should ensure an equal markup on the amount a
15
carrier is paid to perform a unit of workload
16
if the Postal Service distributes
17
flat for each flat delivered,
18
more than the cost for delivering
19
20
21
22
IV.
for each unit of workload
(e.g., distribute
Because the National Mail Count (NMC) is only performed
does not perfectly
Rural Carrier Costing
costing methodology.
24
rural carrier costs (the attribution
25
volume-variable
mail volumes,, an appropriate
for each route evaluation
item (e.g.,
15 percent more than the cost for delivering
a
15 percent
a letter for each letter delivered).’
there are two steps to the Postal Service’s
First, Witness
Baron determined
- Determining
the volume variability
step). Then, Witness Alexandrovich
costs to subclass (the distribution
A. Attribution
in the Fall
Methodology
23
Witness
reflect annual
the Postal Service should also distribute
As for all cost segments,
26
27
28
29
performed
of
distributed
step).
Volume Variable Cost
Baron first divided accrued costs into those for evaluated
30
and those for other routes based upon payroll data (See Ta.ble 2).
31
defined
sixteen of the route evaluation
items, shown
4
in Table
routes
He then
1, as variable
because
“the time required
delivered
on the route.”
required for completion
for completion
The remaining
is unaffected
varies
7
8
by route volume.” (USPS-T.-17 at 66-69).
9
Cost by Route Type
cost ($OOOs)
82,801,424
273,010
--
Individually
for evaluated
and other
average units of each route evaluation
routes,
--___I
he then
item performed
the
per week per route from
the NMC and multiplied
II
evaluation
12
For example, the average weekly activity level estimated for the letters delivered
13
item equals 5,713 letters per week per route.
14
evaluation
I5
per week per route for delivering
item to determine
by the evaluation
determined
10
16
17
this figure
with volume
items were fixed bescause “the time
Table 2. PI 1996 Rural Carrier Accrued
Route Type
Evaluated
Other
proportionately
the “average
weekly
factor
for each route
minutes for the given item.
The product of this level and the
factor of 0.0791 minutes per letter equals an estimated
Finally, Witness
letters in FY 1996.”
7 at 70).
Baron divided the sum of the average minutes per week
18
per route for all variable route evaluation
19
per route for all route evaluation
20
variability
21
74).
for evaluated
(USPS-T-l
452 minutes
items by the average minutes per week
items, fixed and variable, to obtain the volume
routes and other routes (See Table 3). (USPS-T-17
22
Table 3. Volume
23
24
Route Type
Evaluated
Other
Variability
of Rural Carrier Costs
Variability (%)
49.04
49.87
25
’ My improvements focus on the “Letters Delivered’ and “Flats Delivered’
costs account for approximately 80 percent of all rural carrier costs.
5
costs bewuSe these
at
I
6. Distribution
2
3
Witness
- Distributing
Alexandrovich’s
Cost to Subclass
first step in distributing
4
carrier costs was to disaggregate
5
evaluation
6
variable
7
route (excluding
8
results of this process for evaluated routes.
item.
volume-variable
To do this, he essentially
route evaluation
vehicle
loading
volume-variable
to average
and markups
rural
rural carrier costs by route
apportioned
items in proportion
volume-variable
time).’
cost to
minutes per week per
Table
4 provides
the
9
10
II
I2
13
14
Table 4. Base Year 1996 Evaluated Route Costs by Variable
Evaluation Item
Witness Alexandrovich
route evaluation
then developed
distribution
keys for each volume-
15
variable
16
Delivered”
17
based upon volumes from the rural Carrier Cost System (CCS).
For example,
18
the distribution
item was the
19
volume of flats delivered
20
using these distribution
’ The derivation
item “cost pool” shown
Route
and “Letters
Delivered”
key for the “Flats
of average
cost pools,
Delivered”
by rural carriers.
in Table 4~. For the “Flats
these
distribution
route evaluation
It is important
to note that, before
keys, the Postal Service reclassified
minutes
per week
per route is described
6
in Section
keys were
MA.
some letters from
I
rural CCS as flats primarily to account for the fact that the definition
2
the NMC is different than the standard postal definition
V.
of a flat in
of a flat.3
Data Anomaly
Table 1 shows that the average
rural carrier would have been paid 2.8
7
cents for each letter delivered
8
Year.
9
2.4 cents (about 15 percent lower than the amount the rural carrier is paid) and
10
and 5.0 cents for each flat delivered
In contrast, the Base Year 1996 cost distributed
the cost distributed
in the Base
to subcla:ss per letter was
per flat was 5.7 cents (about 15 percent higher).4
II
I2
The reason for this anomaly
is that Witness
Alexandrovich,
step, used NMC data to disaggregate
consistent
13
with the attribution
14
variable
I5
pools, but used volumes from the CCS to distribute
16
past cases, the Postal Service argued that shape data from the NMC is more
17
reasonable
costs to the “Letters
Delivered”
rural carrier volume
and “Flats Delivered”
rural carrier cost
these costs to subclass.
than that from CCS:
18
19
20
21
22
“The primary source of the discrepancy appears to be small flats
which accidentally are recorded as letters. The discrepancy
results from a definition of ‘letlen’ and ‘flats’ that is unique to
rural routes. The shape of rural letters is defined as 5’ in hei:ght
23
24
2s
26
27
28
29
30
31
32
33
34
35
36
37
38
or less. Anything with a greater height is a flat. By the standard
Postal definition (in the Domestic Mail Manual), a letter can
have a height of up to 6 118’. These pieces of mail are shaiped
like letten. but in fact are greater than 5” in height. They would
be considered letters except by experts in Rural Carder mail
shape definitions.... The National Mail Count is the basis for the
carrier’s salary.... Therefore, they [cartiers] would have an
incentive to insure that none of their flats get misclassified as
letters....
The 2858R surveys [CCS in this case], on the other
hand, do not appear to carriers as potentially beneficial or
harmful to them.... [For this test, data collectors] are experts in
distinguishing the details of the different subclasses, so there is
no reason to believe they are making any mistakes in this alaa.
The shape of mail, on the other hand, is different for rural routes
than for city routes. The shape is not the main focus of this test,
and furthermore, is inconsistent with the shape definition for city
’ For more detail on the mail shaDe adiustment. please refer to Section
V of this testimony
or Docket NO.
R90-1. USPS-T-13.
Appendix
F.
’
’ Cost distributed
per route evaluation
item is equal to the rural carrier cost for a route evaluation item
divided by the number of units (e.g., mail volume) for that route evaluatiOn item.
In
routes. Therefore, it seems reasonable to conclude that some
pieces... are being recorded as letters instead of flats.’ (Docket
No. R90-1. USPS-T-13, Appendix F, Page F-26 - F-28).
I
1
3
4
5
For this reason, the Postal Service in Docket No. R90-1, and all dockets
6
since5, implemented
7
letters (as a percentage
8
equal to letters (as a percentage
9
the mail shape adjustment
10
in two inconsistencies
a procedure
called the mail shape
adjustment
to adjust
of letters and flats) in the 2858R (or rural CCS) to be
of letters and flats) in the NMC.
does not fully correct the problem.
In this case,
This can be seen
which remain after the mail shape adjustment.
II
First, based upon NMC volumes, Witness Alexandrovich
12
13
make up 52 percent
14
Tables 10.1.1 and 10.2.1).
15
indicate that letters make up 59 percent
16
Second, as described
17
than the volume variable cost of rural carrier flat delivery - the amount a carrier
18
is paid to deliver a flat - while the cost distributed
19
volume variable cost of rural carrier letter delivery - the amount a carrier is paid
20
to deliver
a letter.
21
attribution
and distribution
22
addresses
these anomalies,
23
Service’s mail shape adjustment
24
25
26
21
VI.
Proposed
Witness
of rural carrier letter/flat
,found that letters
mail volume (USPS-T-5,
W/P B,
CCS volumes, even after the mail shape adjustment,
of rural carrier letter/flat
earlier in this section, the cost distributed
I propose
mail volume.
per flat is higher
per letter is lower than the
that the Postal Rate Commission
are consistent
by making
Section VI proposes
ensure
an adjustment
a modification
that properly addresses
that
that fully
to the Postal
the problem.
Methodology
Alexandrovich’s
workpapers
28
route in an average week, “Letters Delivered”
29
the sum of “Letters Delivered”
indicate
that, on an average
account for about 51.7 percent of
plus “Flats Delivered”
mail (See Table 5)’
5 The Postal Service’s proposed mail shape adjustment is described in LRH.193
‘USPS-T-5, W/P B. Tables 10.1.1 and 10.2.1
8
rural
Table 5. Average
I
2
Volume
Letters Delivered
Flats Delivered
Total Flats/Letters
Letter (%)
3
4
Volume
Delivered
Evaluated Routes
571,336
535.884
1,107.220
51.60%
I propose
Per Week Per Route
Other Routes
316,918
286,336
605,254
52.69%
a mail shape adjustment
that recodes
a sufficient
amount
5
letters such that the ratio of FY 1996 letters to letters and flats combined
6
CCS be equal
7
adjustment
8
MPAXl,
9
flats.
to 51.7 percent.
Performing
will result in the anomaly
this mail shape adjustment
any other
described
recodes
above.
mail shape
in Exhibit
1 out of every 4.01340
letters as
This solves the anomaly and ensures that the markup (in this case, mark
down) on flats is equal to the markup on letters.
II
resulting
12
Delivered”
VII.
from
As derived
10
13
14
15
16
letter/flat
of
Base Year 1996 distribution
Exhibit MPAX2
keys for the “Flats Delivered”
shows the
and “Letters
cost pools.
Conclusions
There is an inconsistency
between
In Docket No. R90-1, Witness
volume data from the NMC and the
17
rural CCS.
Barker found that this was primarily
18
due to the fact that rural flats are defined differently
19
in the Domestic Mail Manual.
20
cost to classes with a high proportion
21
high proportion
This inconsistency
than the standard definition
results in attributing
too much
of flats and too little cost to classes with a
of letters.
22
The Postal Service’s mail shape adjustment
23
24
the problem.
25
adjustment
26
MPA 3-l.
27
distribution
28
2.
does not colnpletely
For this reason, I propose that the Commission
described
Adopting
in Section VI of my testimony
address
us,e the mail shape
and illustrated
in Exhibit
this adjustment will result in the Base Year rural carrier cost
for “Letters Delivered”
and “Flats Delivered”
9
shown in Exhibit MPA 3-
Exhibit MPA 3-3 estimates the difference
I
by
subclass
between
the
2
costs
3
methodology
4
of the TYAR
5
Service’s
6
rural carrier costs by $22.7 million and total Periodical
7
cost piggybacks,
8
Table 6 disaggregates
and the MPA proposed
cost difference,
roll forward
model.
USPS
in Test Year After Rates (TYAR)
proposed
methodology.
rural
by $27.2 million.
should
methodology
rerun
decreases
the Postal
Periodical
costs, taking into account
Using TYAR volumes from Exhibit USPS3OF,
the cost reduction within the Periodical
class by subclass.
9
Table 6. TYAR Periodicals
10
II
costing
To obtain a precise estimate
the Rate Commission
The proposed
carrier
Cost Reduction
Subclass
In-County
Cost Reduction
$2,389
Regular
Nonprofit
Classroom
$18,937
$5,726
$126
12
IO
by Subclass
ExhlbH MPA 3-l. MPA Proposed Mail Shape Adjustment Summary (Votuml!r tn 000s)
Postal Setvlcc Attrlbutlon Step Percentages
LETTERS
51.70% [l]
48.30% [2]
FLATS
FY ,996 RCCS Data
LETTERS
FLATS
22.2w467
10.044259
[3]
[4]
68.66% 151
31.14% [6]
16.674142
15,5im64
m
[S]
51.70% [9]
48.30% (101
If the 1966 Rural CCS data had the same percentages of letters and flats as
in the Nalimal Mail Count, there would have to be the
following distribution:
LETTERS
FLATS
This would require an adjustment of
1 out of every
[I] USPS-T-5. W/P
121USPS-T-5 W/p
[3] USPS Response
[4] USPS Response
5.533.325
4.013404
6. W/S 10.1.1 and 10.2.1. Column 2
6, W/S 10.1.1 and 10.2.1. Column 2
to MPANSPS-T52CD,
TaMe 2
to MPANSPS-T52CD.
Table 2
Fl=W(P1+l41)
M=W(P1+[41)
m=~v(131+~41)
l~l=1W[3l+l41)
Pl=[11
[101=121
[11]=[3]-r7l
[12]=[3v111]
Papa l
,&en
pieces
would have to be reclassifed
IllI
as flats. (121
0.0%
8.2%
t5.19i
0.9%
5.9%
3, .a%
2,413
t.3a
221)
2.047
4.513
2.79ws3
2.905.2,3
I15.593
468.31*
6.292.322
CERTIFICATE
OF SERVICE
I hereby certify that I have this day served the foregoing
participants
of record in this proceeding
Practice.
December
30, 1997
in accordance
document
upon all
with section 12 of the Rules of