BEFORE THE
POSTAL RATE COMMISSION
WASHINGTON, D. C. 20268-0001
POSTAL RATE AND FEE CHANGES,
2000
:
.-
DIRECT TESTIMONY OF
MARK F. RAMAGE
ON BEHALF OF
UNITED STATES POSTAL SERVICE
Docket No. R2000-1
.--.
USPS-T-2
CONTENTS
AUTOBIOGRAPHICAL
SKETCH.. ..................................................................................
AND SCOPE OF TESTIMONY
..............................................................
.2
I.
PURPOSE
II.
Sample Design ........................................................................................................
.3
Ill.
Data Collection
........................................................................................................
.4
IV.
Data Validation
and Editing .....................................................................................
.5
V.
Estimation .................................................................................................................
TABLES
Cv’s for Mailprocessing Costs
C\rs for City
Citv Carrier In Office Costs
CV’s for Subervisors
Supervisors and Special Delivery Costs
Table 1
Table 2
Table 3
ASSOCIATED
LIBRARY REFERENCES
USPS-LR-I-12/R2000-1
USPS-LR-I-13/R2000-1
USPS-LR-I-14/R2000-1
USPS-LR-I-15/R2000-1
IOCS Statistisical and Computer Documentation (source
code and data on CD-ROM)
IOCS-CODES Computer System Documenation
Handbook F-45, In-Office Cost System, Field Operating
Instructions
IOCS-CODES Computer System Source Code (on
CD-ROM)
WORKPAPERS
Workpaper
A
Base Year 1998 LIOCATT for City Carriers (A-l and A-2)
3
6
2
DIRECT TESTIMONY
OF
MARK F. RAMAGE
AUTOBIOGRAPHICAL
I am a Mathematical
SKETCH
7
My name is Mark F. Ramage.
Statistician
in Cost Systems,
a
Finance.
9
present position, I am the program manager for the In-Office Cost System.
I have been employed in this capacity by the Postal Service since 1997. In my
10
held positions as Sr. Mathematical
Statistician and as Sr. Operations
II
for the Postal Service between 1983 and 1987. In those positions,
12
issues for the Carrier Cost Systems and for the In-Office Cost System.
I have also
Research analyst
I worked on statistical
13
-
'4
From 1987 to 1997, I was a Statistician for the Office of the Consumer
15
Postal Rate Commission.
16
before the Postal Rate Commission.
17
Mathematical
19
I received a Bachelor of Science Degree in Mathematics
20
Degree in Statistics in 1980, both from the University of Maryland.
21
additional postgraduate
22
Washington
Statistician
Advocate of the
In that capacity, I testified on a number of statistical issues
From 1980 to 1983, I was employed
as a
with the Bureau of the Census.
work in mathematical
in 1978 and a Master of Arts
I have completed
statistics and survey sampling at George
University.
23
_
24
I am a member of the American Statistical Association
25
Society.
and the Washington
Statistical
3
1
2
I.
PURPOSE
AND SCOPE OF TESTIMONY
3
The Postal Service’s
4
perfectly with, or specifically
6
the Postal Service must rely on various statistical systems and special studies to
6
provide estimates
7
estimates of revenues, volumes and attributable
a
The In-Office Cost System (IOCS) is a work sampling system designed to produce cost
9
estimates for various employee
revenue and cost accounting
systems do not generally coincide
identify, individual categories
of mail or service.
of accrued costs for certain postal operating functions,
Therefore,
and to provide
costs for the various categories
of mail.
activities in the office.
10
11
The purpose of my testimony
is to describe IOCS.
12
design, data collection methodology,
13
the types of estimates
My testimony describes
the editing and processing
the sample
of IOCS sample data,
produced from the IOCS, and the reliability of major estimates.
14
15
II.
Sample Design
16
The In-Office Cost System uses a probability sample of employee activity to develop
17
estimates of proportions
ia
for certain functions,
19
mail categories.
of employee work time spent on various office functions,
the proportions
of time spent handling and/or processing
and
specific
20
21
The IOCS is an ongoing system with a sample selected for each week.
22
4-5 weeks prior to each sample week, sample employees
23
current payroll files for the IOCS sample offices.
Approximately
are selected from the most
Employees are sampled independently
I
4
1
within Cost Ascertainment
2
Mailhandlers,
3
Selected employees
4
week for observation.
(1) Clerks, (2)
Group (CAG) for each of five employee crafts:
(3) City Carriers, (4) Special Delivery Messengers,
and (5) Supervisors.
are then randomly assigned an instant in time during the sample
5
6
The IOCS sample design documentation
7
contained
6
USPS-LR-I-12/R2000-1.
9
sample design, Section IV describes the sample selection programs, and Appendix H
lo
(CD-ROM)
along with sample selection programs are
in the IOCS Statistical and Computer System Documentation,
Section II of the IOCS System Documentation
describes the
contains the sample selection programs.’
11
-
12
Ill.
Data Collection
13
The IOCS data collection instrument consists of a data entry program residing on a
14
laptop computer.
15
On-Site Data Entry System (IOCS-CODES).
16
collector locates the sample employee and enters data describing the employee’s
17
observed activity directly into CODES.
ia
collector through the appropriate
19
the entered data. Sample data from various laptops at one site are transferred
20
base unit computer (a desk top PC). Then, at the end of each week, the data are
21
transmitted
22
processing.
This hardware and software is referred to as the IOCS Computerized
At the scheduled
sample time, the data
The data entry software guides the data
questions and performs basic consistency
checks on
from the base units to a mainframe computer for further editing and
to a
5
1
2
An overview of field IOCS test administration
3
Documentation
4
procedures
5
(USPS-LR-14/R2000-1).
6
code are contained
7
respectively.
(USPS-LR-I-12/R2000-1).
are contained
is contained
in section V of the IOCS
Detailed descriptions
in the IOCS Field Operating
IOCS-CODES
of IOCS data collection
Instructions,
Handbook
computer system documentation
in USPS-LR-I-13/R2000-1
F-45
and source
and USPS-LR-I-15/R2000-1,
a
9
IV.
Data Validation
and Editing
10
After the IOCS data are transmitted
to the mainframe computer,
11
SAS programs perform data validation, editing, and automated
12
Included in this process is the assignment
13
programs.
a series of COBOL and
error correction.
of activity codes used in subsequent
costing
14
15
Data validation and editing procedures
16
procedures)
17
and Computer
Documentation,
18
USPS-LR-I-12
describes the validation and editing programs, Appendix
19
flowchart
20
edit rules, Appendix
21
describes the encirclement
rules used to assign special service activity codes.
’ Thmughout this testimony. appendices
USPS-LR-I-12/R2000-1.
A through I refer to the appendices
are contained
documenting
(and the programs associated
in Section VI and several appendices
USPS-LR-I-12/R2000-1.
the assignment
D documents
of the IOCS Statistical
Specifically,
of activity codes, Appendix
the within county periodicals
with these
Section VI of
B contains a
C describes
detailed
edits, and Appendix
of the IOCS Statistical
E
All
and Computer Documentation,
6
1
validation and editing programs are contained on the CD-ROM described
in
2
Appendix
3
V.
4
The IOCS sample data are used to produce estimates of costs by function for each craft
5
group.
6
sample design and accrued costs for CAG and craft group.
7
be associated
a
is then converted
H.
Estimation
Cost weight factors are assigned to each sample observation,
reflecting both the
Hence, a dollar amount can
with each record in the IOCS data file. The cost weighted
into a final SAS data set which is used for producing
IOCS data file
cost estimates.
9
-~
10
Several craft reports are produced from the final IOCS data file that provide CRA
11
spreadsheet
12
of this report is included as Workpaper
inputs.
One such report is the City Carrier LIOCATT report.
A hard-copy
A to my testimony.
13
14
As with any sample based estimate, selection of a different sample (according
15
same sample design), could result in slightly different estimates.
16
variation one could expect due to sampling alone is quantified
17
variation (CV).
18
1 presents cost estimates, their estimated Cv’s, and 95 percent confidence
19
Cost Segment 3.1 Mail Processing.
20
Carriers, and for Supervisors
CV’s can be used to produce confidence
to the
The amount of
by the coefficient of
intervals for estimates.
Table
intervals for
Tables 2 and 3 provide similar estimates for City
and Special Delivery Messengers.
21
-
22
The development
23
CV estimation,
of the cost weighting factors, production
and descriptions
of the final SAS data files,
of various craft reports are provided in Section VII of
7
1
USPS-LR-I-12/R2000-1.
2
CV estimation,
The source code for weighting,
and craft reports is contained
production
of final data files,
on the CD-ROM, described
in Appendix
H.
.
a
Table 1
BY98 MODS-based estimated mean distributed
costs (000’s) and CVs, All Dffices
Cost Segment 3.1 Mail Processing - Clerks and Mailhandlers
Subclass
Cost Est.
std
Deviation
95% Lower
Limit
95% Lower
Limit
CV
First
4.972,053
1.226.433
161,924
37,759
22.659
15,014
5,570
2.600
5,016.465
1.255.860
192.841
42,855
0.46%
1.22%
3.06%
6.89%
Priority
Express
Mailgrams
645,444
94,151
166
8,336
3,160
103
629,106
87.919
0
661.762
100,383
367
1.29%
3.38%
62.13%
Periodicals
Within County
Regular
Non Profit
Classroom
13,126
566,261
94,885
5,177
1,607
9,775
4,161
953
9,976
547,102
86,730
3,310
16.276
585,420
103,040
7,044
12.24%
1.73%
4.38%
18.40%
94,386
303,067
2,111,573
46,614
427,658
4,711
7.790
19,437
3,365
9,659
85.152
287,799
2,073.476
40.019
408,726
103,820
318,335
2,149.670
53,209
446,590
4.99%
2.57%
0.92%
7.22%
2.26%
Standard tB)
Parcels -Zone Rate
Bound Printed Matter
Special Rate
Library Rate
246,615
128,518
82,890
11,860
7,622
6,288
5,084
1,579
231.877
116,194
72.925
8,765
261,753
140,842
92,855
14,955
3.09%
4.69%
6.13%
13.32%
USPS
Free for Blind/Handicapped
International
129,896
14,924
261,969
4,653
1,351
6,037
120,365
11,100
250,137
139,407
18,748
273,801
3.74%
13.07%
2.30%
36,966
31,865
1,759
554
3,655
122
485
2,867
41.241
2,107
1,372
368
248
N/A
N/A
324
N/A
2,231
41,095
35,731
2,460
1,041
N/A
N/A
1,120
N/A
45,614
5.70%
6.19%
20.92%
4483%
N/A
N/A
68.87%
N/A
5.41%
Letters
Presort
Private
Presort
and Parcels
Letters and Parcels
Mailing Cards
cards
Single Piece Rate
Regular Carrier Route
Regular Other
Non Profit Canter Route
Non Profit Other
Registry
Certified
Insurance
coo
Money Orders
Stamped Envelopes
Special Handling
P.O. Box
Other Special Services
11,817.062
4,927,641
I .197,006
171,007
32,662
32,837
27,939
1,038
67
N/A
N/A
0
N/A
36,868
9
Table 2
BY98 LIOCAlT estimated mean distributed
costs (000’s) and CVs, All Offices
Cost Segment 6.1 Mail Processing - City Carriers Direct Labor Inputs
Subclass
El&Q@
Letters
Presort
Private
Presort
and Parcels
Letters and Parcels
Mailing Cards
Cards
Cost Est. z
95% Lower
Limit
Std
Deviation
95% Lower
Limit
CV
1.062.576
469,772
54,433
10.878
9,443
6.361
2.221
1,480
1,044,067
457,305
50,080
15,978
37,597
2,601
0
1,655
516
0
33,961
1,589
0
41,233
3,613
0
4.93%
19.65%
0.00%
Periodicals
Within County
Regular
Non Profit
Classroom
8.069
133.157
29,824
381
913
3,467
1,595
165
6,280
126,362
26,697
58
9,859
139,952
32,951
703
11.32%
2.60%
5.35%
43.22%
Standard (A)
Single Piece Rate
Regular Carrier Route
Regular Other
Non Profit Carrier Route
Non Profit Other
13,348
284,702
525.706
20,000
106,645
1,133
4,999
6,600
1,332
3,044
11,128
274.904
512,769
17,390
100,680
15,568
294,500
22,610
112,610
8.49%
1.76%
1.26%
6.66%
2.85%
6,927
6.895
1;868
634
785
743
362
333
5,369
5,439
1,159
0
8,466
8,352
2,576
1,287
11.33%
10.77%
19.36%
52.60%
USPS
Free for Blind/Handicapped
International
10.549
983
10.955
1,077
275
948
8,439
443
9,097
12,659
1,523
12,814
10.20%
28.02%
8.65%
Registry
Certified
hursnce
coo
Money Orders
Stamped Envelopes
;~;~w;~dling
1,683
27,400
1.075
728
0
0
334
1,677
296
251
0
0
0
150
565
1.026
24,112
495
237
0
0
0
131
3,200
2,336
30,687
1.655
1,220
0
0
0
719
5,415
19.66%
6.12%
27.52%
34.40%
0.00%
0.00%
0.00%
35.28%
13.12%
Priority
Express
Mailgrams
Standard (6)
Parcels -Zone Rate
Bound Printed Matter
Special Rate
Library Rate
Other Special Services
Total
0
425
4,308
1.081,085
482,239
56,787
21,770
538.642
0.89%
1.35%
4.06%
7.84%
2.842,120
2 Standard deviations are estimated from a bootstrapped FORTRAN approximation to Carrier LIOCATT cost estimates. Confidence
interYak and Cv’s are developed by applying the standard deviation estimates to the actual BY99 CFSAinput. See
USPS-LR-I-12!R2000-1 for the programs used.
_
’
10
Table 3
BY98 estimated mean distributed
costs and CVs, All Offices
Supervisors
and Special Delivery Messengers
Subclass
Cost Est.
Mail Processing
Central Mail Mark-up
Window Service
Administrative end Support
City Delivery Carriers
Special Delivery Messengers
Rural Delivety Carriers
Vehicle Service
Employees/Labor
Higher Level Supervisors
Gen Supv.: Mail Processing
Gen Supv.: Collection/Delivery
Supervisor Training
Quality Control/Revenue Prot.
Supv. Of Mixed Clk/Mailhandler
Supv. Of One or More Crafts
Other
Total
So&al Deliver
Street Time
Other
Total
Activities
873.518
61,796
144,315
29,363
541,761
6,443
22,288
34,661
713
245,253
746
414
49,275
34.908
97,040
581,643
788,311
Std Deviation
95% Lower
Limit
95% Lower
Limit
cv
8.371
2.708
6,943
1,804
6.780
836
4,974
1,559
250
5,127
259
222
2,116
1,693
3,467
10.183
10,155
857,111
56,486
130,708
25,826
526,473
4,805
12,539
31,606
223
235,204
238
0
45,127
31,589
90,245
561,684
768,408
889,924
0.96%
67,104
157,923
32,899
555,049
8,080
32,038
37,716
1,203
255,301
1,254
849
53,422
38,226
103,835
601,603
808,Jl4
4.38%
4.81%
6.15%
1.25%
12.97%
22.32%
4.50%
35.07%
2.09%
34.76%
53.61%
4.29%
4.85%
3.57%
1.75%
1.29%
723
723
58.183
10,117
61,016
12,951
1.21%
6.27%
3,512.447
Messenoers
59,600
11,534
71,133
-