BEFORE THE
POSTAL RATE COMMISSION
WASHINGTON, D.C. 20268-0001
POSTAL RATE AND FEE CHANGES, 2006
)
Postal Rate Commission
Submitted 12/21/2006 4:18 pm
Filing ID: 55483
Accepted 12/21/2006
Docket No. R2006-1
INITIAL BRIEF
OF
VALPAK DIRECT MARKETING SYSTEMS, INC., AND
VALPAK DEALERS’ ASSOCIATION, INC.
William J. Olson
John S. Miles
Jeremiah L. Morgan
WILLIAM J. OLSON, P.C.
8180 Greensboro Drive, Suite 1070
McLean, Virginia 22102-3860
(703) 356-5070
Counsel for:
Valpak Direct Marketing Systems, Inc. and
Valpak Dealers’ Association, Inc.
December 21, 2006
TABLE OF CONTENTS
Page
TABLE OF AUTHORITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
STATEMENT OF THE CASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
ARGUMENT
I. STANDARD REGULAR AND ECR COST COVERAGES PROPOSED
BY THE POSTAL SERVICE ARE PROFOUNDLY OUT OF BALANCE
AND NEED SUBSTANTIAL ADJUSTMENT BY THE COMMISSION . . . . . I-1
A. A Lower Cost Coverage Should Be Applied to the ECR Subclass . . . . . . . . I-4
1. Proper application of the noncost criteria supports a cost coverage
near the systemwide average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-4
a.
b.
c.
d.
e.
f.
Fairness (criterion 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-4
Value of Service (criterion 2) . . . . . . . . . . . . . . . . . . . . . . . . . . I-5
Effect on Mailers (criterion 4) . . . . . . . . . . . . . . . . . . . . . . . . . . I-7
Availability of Alternatives (criterion 5) . . . . . . . . . . . . . . . . . . . I-8
Degree of Preparation (criterion 6) . . . . . . . . . . . . . . . . . . . . . . I-10
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-11
2. Continued imposition of an excessive cost coverage on ECR produces
Nonprofit ECR rates that do not reflect proper recognition of the
favored treatment due nonprofit mail . . . . . . . . . . . . . . . . . . . . . . . I-11
B. ECR’s Cost Coverage Should Be Set Below that of Regular . . . . . . . . . . I-12
1. ECR’s cost coverage should be near the systemwide average and below
that of Regular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-12
2. Undue deference towards existing rate relationships between
Regular and ECR undermines realization of the purpose behind the
establishment of separate subclasses . . . . . . . . . . . . . . . . . . . . . . . I-13
C. Postal Accountability and Enhancement Act — Transition Rule . . . . . . . . I-17
II. THE COMMISSION SHOULD CONTINUE TO TREAT MAIL
PROCESSING COSTS AS FULLY ATTRIBUTABLE . . . . . . . . . . . . . . . II-1
A. Witness Bozzo’s Analysis Does Not Withstand Scrutiny . . . . . . . . . . . . . II-2
ii
1. Witness Bozzo fails to address economies of scale . . . . . . . . . . . . . . II-2
2. Setup and takedown costs do not vary uniformly with mail volume,
but they nevertheless can be volume variable . . . . . . . . . . . . . . . . . II-5
3. Witness Bozzo would increase institutional costs by $2.4 billion . . . . . II-8
B. Many of Witness Bozzo’s Non-volume Variable Costs Are Attributable . . II-8
1. Non-volume variable costs are incremental costs . . . . . . . . . . . . . . . II-9
2. Witness Bozzo’s rebuttal testimony lacks merit . . . . . . . . . . . . . . . II-10
C. Witness Bozzo Failed to Address the Commission’s Concerns . . . . . . . . II-15
1. Witness Bozzo did not address the Commission’s analysis
regarding setup and takedown costs . . . . . . . . . . . . . . . . . . . . . . . II-15
2. Witness Bozzo did not address the impact of his analysis on
worksharing discounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II-16
III.
WITNESS HALDI (VP-T-2) IDENTIFIES SPECIFIC PROBLEMS
WITH CURRENT COSTING SYSTEMS AFFECTING STANDARD
MAIL, WHILE WITNESS MITCHELL (VP-T-3) ASKS THAT THE
COMMISSION ADDRESS ALL SUCH “FAIRNESS IN COSTING”
PROBLEMS IN A SUBSEQUENT DOCKET . . . . . . . . . . . . . . . . . . III-1
A. The Problem in Costing of Saturation Letters and Flats Created by
Capacity Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-3
B. Dr. Haldi’s Testimony in Docket No. R2005-1 . . . . . . . . . . . . . . . . . III-4
C. Dr. Haldi’s Testimony in Docket No. R2006-1 . . . . . . . . . . . . . . . . . III-5
1. The Existence of Capacity Constraints on Motorized Routes . . . . . . . III-6
2. The Effect of Uneven Distribution of Saturation Mail . . . . . . . . . . . III-7
3. The Effect of Permanently Bumping Saturation Letters to DPS
Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-7
4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-9
iii
D. Witness Mitchell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-10
IV.
WITNESS MITCHELL’S TRADITIONAL APPROACH TO
STANDARD MAIL RATE DESIGN SHOULD BE UTILIZED
AND HIS PROPOSED RATES FOR STANDARD MAIL
SHOULD BE RECOMMENDED . . . . . . . . . . . . . . . . . . . . . . . . . . IV-1
A. Witness Mitchell’s Standard Mail Rate Design Is Superior to
that of Witness Kiefer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-1
B. Witness Crowder’s Proposal to Burden ECR Letters with Unrecouped DAL
Costs, Which Are for the Exclusive Benefit of Saturation Flats, Must Be
Rejected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-6
C. Witness Mitchell’s Rebuttal Testimony Demonstrates Why the
Commission Should Reject Pitney Bowes Witness Panzar’s
Application of ECP to the Rate Design for Standard Letters and Flats . . . IV-9
CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-14
iv
TABLE OF AUTHORITIES
Page
Statutes
39 U.S.C. § 3622 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
39 U.S.C. § 3622(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II-8
39 U.S.C. § 3626(a)(6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-2
39 U.S.C. § 3623 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Postal Reorganization Act of 1970 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-2
Postal Accountability and Enhancement Act, H.R. 6407 . . . . . . . . . . . . . . . I-18, I-19, II-2
Regulations
39 CFR § 3001.20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Cases
National Easter Seal Society v. United States Postal Service,
656 F.2d 754 (D.C. Cir. 1981 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-5
Postal Rate Commission Opinion and Recommended Decision
Docket No. R84-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-8
Docket No. R87-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-8
Docket No. R90-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-12, I-13
Docket No. R94-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-12
Docket No. MC95-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-3, I-13, I-14, I-17, IV-10
Docket No. R97-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-14
v
Docket No. R2000-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-4, I-14, I-15, II-16, III-2
Docket No. R2001-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-15, I-16, II-1
Docket No. R2005-1 . . . . . . . . . . . . . . . . . . . . . . . I-7, I-15, I-17, II-1, II-15, III-1, III-5
BEFORE THE
POSTAL RATE COMMISSION
WASHINGTON, D.C. 20268-0001
POSTAL RATE AND FEE CHANGES, 2006
)
Docket No. R2006-1
INITIAL BRIEF
OF
VALPAK DIRECT MARKETING SYSTEMS, INC., AND
VALPAK DEALERS’ ASSOCIATION, INC.
STATEMENT OF THE CASE
On May 3, 2006, the United States Postal Service filed a request, pursuant to the Postal
Reorganization Act (39 U.S.C. §§ 3622 and 3623), for a recommended decision by the Postal
Rate Commission on certain rates and fees, including proposals relating to Standard Mail
rates, as well as certain changes to the Domestic Mail Classification Schedule (“DMCS”).
On May 5, 2006, the Commission issued a Notice of Filing of the Postal Service’s
request (Order No. 1464).
In accordance with Order No. 1279 and Rule 20 of the Commission’s Rules of Practice
and Procedure (39 CFR § 3001.20), Valpak Direct Marketing Systems, Inc. and Valpak
Dealers’ Association, Inc. each filed a notice of intervention on May 4, 2006. These
intervenors have proceeded jointly in this proceeding, and are referred to collectively as
“Valpak” or “VP.”
2
Postal Service Request
The Postal Service Request for a Recommended Decision initiating this proceeding
requested rate and fee changes affecting all classes of mail, and asserted that without those
changes the Postal Service would incur a revenue deficiency of $5.874 billion in the requested
test year (FY 2008). According to the Postal Service’s request, the requested rates would
generate a revenue deficit of approximately $787,000 in the test year. Request, p. 2.
The Postal Service’s case-in-chief presented a systemwide average rate increase of 8.5
percent, with Standard Enhanced Carrier Route (“ECR”) rates receiving above-average
increases of 8.9 percent, and Standard Regular rates also receiving above-average increase of
9.6 percent. Direct testimony of Postal Service witness O’Hara, USPS-T-31, pp. 27, 29 (as
revised August 25, 2006).
Discovery of the Postal Service’s Case-in-Chief
Valpak conducted written cross-examination of 23 Postal Service witnesses with respect
to their identified direct testimony, and oral cross-examination of nine Postal Service witnesses
(with the pages of that oral cross-examination set out below).
Bradley V. Pafford
USPS-T-3
Thomas W. Harahush
USPS-T-4
Thomas E. Thress
USPS-T-7
Mico Milanovic
USPS-T-9
Eliane Van-Ty-Smith
USPS-T-11
A. Thomas Bozzo
USPS-T-12
Marc A. Smith
USPS-T-13
Tr. 6/1341-1358
Tr. 10/2685-2709
3
Michael D. Bradley
USPS-T-14
Tr. 13/3830-3861
John P. Kelley
USPS-T-15
Dion E. Pifer
USPS-T-18
Michael W. Miller
USPS-T-20
James W. Page
USPS-T-23
Samuel T. Cutting
USPS-T-26
Daniel Talmo
USPS-T-27
L. Paul Loetscher
USPS-T-28
John P. Kelley
USPS-T-30
Tr. 12/3516-3522
Donald J. O’Hara
USPS-T-31
Tr. 17/5245-5273
Altaf H. Taufique
USPS-T-32
James M. Kiefer
USPS-T-36
Drew Mitchum
USPS-T-40
Marc D. McCrery
USPS-T-42
Tr. 11/3281-3302
Joyce K. Coombs
USPS-T-44
Tr. 13/3739-3773
Scott J. Davis
USPS-T-47
Tr. 9/2233-2244
Tr. 5/1128-1178
Commission Notices of Inquiry Nos. 2 and 3
On July 21, 2006, the Commission issued Notice of Inquiry No. 2, which elicited
comments on the Postal Service’s abandonment of the presort tree in the development of
Standard Mail rates, instead substituting “a multitude of free-standing rate categories, each
with its own benchmark rate.” Id., pp. 2-3. The Commission also sought comments on the
proper application of Efficient Component Pricing (“ECP”) to Standard Mail ratesetting.
4
On July 26, 2006, the Commission issued Notice of Inquiry No. 3, which elicited
comments on the Postal Service’s proposal to “de-link” the design of rates for single piece and
workshared mail within the First-Class letters subclass, and invited mailers to comment on
proper application of the Efficient Component Pricing concept to design of postal rates.
Valpak filed its Comments in response to NOI Nos. 2 and 3 on August 17, 2006.
Valpak’s response emphasized the primacy of costs in the statutory rate-setting process, with a
discussion of the distinction between worksharing-related costs (which are properly the focus
of ECP), and other costs such as shape and weight (to which ECP should not apply). Valpak
also promoted the use of incurred costs rather than avoided costs in ratesetting, and supported
retention of the presort tree in the development and presentation of Standard Mail rates.
Valpak Direct Testimony
Valpak sponsored the direct testimony of witnesses Robert W. Mitchell (VP-T-1 and
VP-T-3) and Dr. John Haldi (VP-T-2) on issues related to Standard Enhanced Carrier Route
Mail, filed on September 6, 2006. During discovery, interrogatories, and requests for
production of documents were propounded to witnesses Mitchell and Haldi by the following
parties: Advo, Inc. (ADVO/VP-T1-14, 15-19); Newspaper Association of America
(NAA/VP-T1-1-38); and the Postal Service (USPS/VP-T1-1-15, 16-27, 28-31, 32-33, 34-36;
USPS/VP-T2-1-12; and USPS/VP-T3-1-2). The responses of witnesses Mitchell and Haldi to
112 of these interrogatories were designated as written cross-examination in the transcript (Tr.
22/8067-68, Tr. 23/8587-8622, Tr. 25/8807-8948).
Additionally, two interrogatories and requests for production of documents were
propounded to Valpak by Advo, Inc. (ADVO/VP-1-2), but were not designated for the record.
5
On October 27, 2006, oral cross-examination was conducted of witness Mitchell on his
direct testimony VP-T-1 (Tr. 23/8798-9039).
On October 31, 2006, oral cross-examination was conducted of witness Haldi on his
direct testimony VP-T-2 (Tr. 25/8623-31).
There was no oral cross-examination of witness Mitchell on his direct testimony
VP-T-3.
Intervenor Direct Testimony
Valpak submitted a total of 35 interrogatories and requests for production of documents
to Pitney Bowes Inc. witness John C. Panzar (PB-T-1) and Saturation Mailers Coalition
witness Pete Gorman (SMC-T-1). All 35 of the responses to the Valpak interrogatories were
designated as written cross-examination. These responses appear in the record as follows:
VP/PB-T1-1-29, Tr. 26/9175-9209 and 9213-39; and
VP/SMC-T1-1-6, Tr. 28/9536-43.
Rebuttal Testimony
The following pieces of rebuttal testimony addressed the testimonies of witnesses
Mitchell and Haldi. Cross-examination was conducted as reflected on the pages noted.
A. Thomas Bozzo
USPS-RT-12
Tr. 36/12500-12530
James M. Kiefer
USPS-RT-11
Tr. 33/11222-11237
Maura Robinson
USPS-RT-10
Tr. 32/10767-10805
Marc D. McCrery
USPS-RT-14
Antoinette Crowder
SMC-RT-1
Tr. 35/11787-11836
Stuart W. Elliott
MPA et al.-RT-2
Tr. 36/12372-12381
6
Roger C. Prescott
MOAA-RT-1
Tr. 32/10959-10972
J. Gregory Sidak
NAA-RT-1
Tr. 32/10891-10913
Allan T. Ingraham
NAA-RT-2
Tr. 35/11871-11894
Valpak sponsored the rebuttal testimony of witness Mitchell (VP-RT-1). Witness
Mitchell’s rebuttal testimony critiqued Pitney Bowes witness Panzar’s analysis of the
appropriate relationship between costs and rates for letters and flats within the same subclass.
(In addition to rebuttal by witness Mitchell, witness Panzar’s testimony (PB-T-1) was rebutted
by Postal Service witnesses Maura Robinson (USPS-RT-10) and James M. Kiefer (USPS-RT11, pp. 15-19).) Witness Mitchell was cross-examined on his rebuttal testimony on December
4, 2006 (Tr. 35/11901-42).
I-1
I.
STANDARD REGULAR AND ECR COST COVERAGES PROPOSED BY THE
POSTAL SERVICE ARE PROFOUNDLY OUT OF BALANCE AND NEED
SUBSTANTIAL ADJUSTMENT BY THE COMMISSION.
Postal Service witness Donald J. O’Hara proposed a cost coverage of 214 percent for
Standard Mail ECR and Nonprofit ECR combined. USPS-T-31, revised August 25, 2006, p.
30, l. 23. This coverage implies an above-average ECR “rate increase of 8.9 percent.” Id., p.
29, ll. 3-4. Simultaneously, however, witness O’Hara acknowledged that most of the statutory
noncost criteria “might appear to indicate a cost coverage lower than that actually proposed for
ECR.” Id., p. 30, ll. 13-14. Nevertheless, he relies on two principal arguments to defend his
proposal of a high ECR coverage.
First, witness O’Hara noted that “[a]s mailer preparation increases, it reduces volumevariable costs but has no effect on institutional costs. Rates for otherwise similar mail (shape,
delivery, priority, type of transportation, etc.) can and should reflect to some degree the cost
differences due to different degrees of mailer preparation. At the same time, it is important to
retain a reasonable contribution per piece for ECR.”1 Id., ll. 16-20. Second, witness O’Hara
observed that “at proposed rates, contribution per piece is about the same in ECR and Standard
1
Valpak notes that small preparation changes at a point in time would change
volume-variable costs but not institutional costs; however, preparation changes can be large
and occur over periods of time, making witness O’Hara’s observation inapplicable. Also,
witness O’Hara presents no analysis of relative changes in mail preparation among subclasses,
and he makes no adjustments for changes in the systemwide average cost coverage, as the
Commission’s markup index is designed to do.
I-2
Regular (both round to 10.0 cents) despite the different cost coverages (214% and 177%2).”
Id., ll. 21-23.3
In response, witness Mitchell explained that “no justification exists for continuing the
historic cost coverage of Regular and ECR, case after case, much as though they are
worksharing categories within the same subclass.” VP-T-1, p. 93, ll. 9-11. Yet that would be
the effect if witness O’Hara’s arguments were accepted. Thus, witness O’Hara selects and
justifies coverages as if the establishment of distinct Standard Regular and ECR subclasses
never occurred. Witness Mitchell explained how witness O’Hara’s approach is misguided.
Once created, the cost coverage of a subclass should be set by the Commission through
independent application of the statutory non-cost criteria, not through an effort to preserve
relative levels of contribution that had once been implicit.
Based on witness Mitchell’s analysis of the criteria of the Postal Reorganization Act of
1970 (“the Act”), he recommends a cost coverage for Standard Regular of 180.2 percent and
for Standard ECR of 177.0 percent. At these coverages, the average rate increase for Regular
is 17.56 percent and for ECR is -8.47 percent. (The contribution from both Regular and ECR
2
With final small adjustments based on changes recognized in costs and revenues,
witness O’Hara’s final proposed Standard Mail Regular now rounds to 176 percent. See
explanation provided in spreadsheet attached to Response of United States Postal Service
Witness O’Hara to Questions Posed by Valpak at August 30, 2006 Hearings, filed September
1, 2006.
3
In oral cross-examination, however, witness O’Hara acknowledged that the
equality in contribution per piece had not been a goal of the Postal Service, but was instead just
“happenstance.” Tr. 17/5271, ll. 9-14.
I-3
combined under these proposal is $16.9 million greater than under the Postal Service’s
proposal.) VP-T-1, p. 97, ll. 7-10.
Mail Order Association of America (“MOAA”) witness Roger C. Prescott also
presented direct testimony which independently supports the testimony of witness Mitchell in
many particulars and urges a reduced coverage for ECR. See Direct Testimony of Roger C.
Prescott, MOAA-T-1.4
In Docket No. MC95-1, the Commission established separate Standard Regular and
ECR subclasses as a pricing improvement, to bring about increases in economic efficiency.
However, this commendable purpose will never be achieved so long as the respective
subclasses are locked into a relationship established for rate categories within former thirdclass bulk rate regular. Both fairness and the pursuit of economic efficiency require that the
two markups be adjusted, so that mailers can have the opportunity to respond to improved
price signals. Changes in the markups were implied by the Commission’s analysis when the
two subclasses were created, as was clear at the time. See VP-T-1, pp. 30-34. Unless the
substantial demand differences are recognized in adjusted markups, no efficiency
improvements will be realized, and the time and attention in Docket No. MC95-1 required to
create the subclasses will have come to naught.
As reviewed below, the record demonstrates that an objective application of the
statutory noncost criteria to ECR (and Nonprofit ECR combined) indicates a markup for ECR
4
Further, Advo, Inc. (“Advo”) and Saturation Mailers Coalition (“SMC”)
witness Antoinette Crowder endorses witness Mitchell’s observations on the excessive
coverage of ECR in her rebuttal testimony (SMC-RT-1, p. 2, l. 30, Tr. 35/11720).
I-4
that is below the markup for Regular and at or below the systemwide average (178.4 percent
under Commission costing at Postal Service proposed rates, USPS-LR-L-114.xls, sheet 2).
(Note: For brevity, the discussion below references Standard Mail ECR (or ECR) as including
both commercial ECR and nonprofit ECR, and Standard Mail Regular (or Regular) as
including both commercial Regular and nonprofit Regular.)
A.
A Lower Cost Coverage Should Be Applied to the ECR Subclass.
1.
Proper application of the noncost criteria supports a cost coverage near the
systemwide average.
a.
Fairness (criterion 1)
Only at the end of his analysis of the statutory noncost criteria does witness O’Hara
focus on the first criterion, fairness. “After considering all the criteria, I conclude that the rate
level proposed for ECR represents an appropriate balance among them and satisfies the
fairness and equity criterion.” USPS-T-31, p. 31, ll. 1-3. By contrast, witness Mitchell
begins his analysis with a discussion of the criterion-one fairness issue. See VP-T-1, pp. 6772.
Witness Mitchell reviews the Commission’s position, articulated in Docket No. R20001, that equal implicit markups are assumed as the baseline under noncost criterion 1. See VPT-1, p. 70, ll. 1-4, citing Op. and Rec. Dec., ¶ 5533. Had witness O’Hara acknowledged such
a starting point, it is simply inconceivable that the Postal Service could justify the very high
cost coverage which it has selected for ECR or the relatively low cost coverage for Regular.
For, as will be shown, not one of the noncost criteria — individually or in combination —
I-5
supports the imposition of a coverage on Regular that is below the coverage on ECR or a
coverage on ECR that is above the systemwide average.
In fact, the fairness criterion calls for a reduction in ECR coverage and an increase for
Regular. The effort to reach a reasonable coverage for ECR now has taken more than a
decade, achieving only minimal results, and prolonging this process any further would be
profoundly unfair to ECR mailers.
b.
Value of Service (criterion 2)
Witness O’Hara made the following observations about Regular and ECR:
In common with Standard Regular, the intrinsic value of service
for Standard ECR is relatively low (criterion 2), since it lacks
access to the collection system, receives ground transportation,
and its delivery may be deferred. Where the Postal Service is
able to accommodate mailer requests for delivery within specific
and sometimes relatively tight time frames, this again often
requires mailer preparation, coordination, and planning. For
the high-density and saturation rate categories, this may also be
facilitated by the regularity with which many of these mailings
are deposited. [USPS-T-31 (revised), p. 29, ll. 5-12 (emphasis
added).]
However, as witness Mitchell explained, although listing a product’s characteristics, as witness
O’Hara has done, may highlight its features and facilitate thinking about the value mailers
might place on it, the ultimate — and only relevant — consideration is whether the value is
actually there. The test for this is the decisions customers make in the market, which are
reflected in the elasticity measure. Analysis of the price elasticity does not support the Postal
Service proposed coverage for Regular or ECR. This is acknowledged by witness O’Hara,
who stated:
I-6
The [own-]price-elasticity of commercial Enhanced Carrier
Route is about -1.1, much higher than those of Standard
Regular, First-Class Mail Letters and Cards, or Periodicals,
indicating a rather low economic value of service. The
elasticity for Nonprofit ECR is about -0.3, but since commercial
ECR makes up about 83 percent of total ECR volume, the
commercial elasticity provides a good indicator of economic
value of service for the subclass as a whole. [USPS-T-31
(revised), p. 29, ll. 12-18 (emphasis added).]
Also, Postal Service witness Thomas E. Thress (USPS-T-7) testified that:
The own-price elasticity of Standard ECR mail is calculated to
be equal to -1.079 (t-statistic of -6.159). This is among the
highest own-price elasticities estimated in my testimony,
reflecting the competitiveness of the advertising market and the
extent to which relatively close substitutes exist for Standard ECR
mail. [USPS-T-7, p. 120, ll. 16-19 (emphasis added).]
ECR’s high (in absolute value) own-price elasticity demonstrates an extremely low
value of service. Such a low value of service should be recognized with a below-average
markup. Similarly, the low elasticity of Regular should be recognized with a markup that is
above that of ECR. The principal reason for creating the Regular and the ECR subclasses was
to facilitate recognition of their market characteristics, and the characteristics that received
primary attention were the disparate elasticities. The value available to be drawn upon in the
case of ECR is considerably lower than the value of Regular, or, for that matter, of any other
subclass except Express Mail.5 To the extent that recognition is given to this low reservoir of
value, a below-average cost coverage is indicated.
Thus, it is unreasonable under any circumstances that the Postal Service should propose
the highest cost coverage — combined with an above-average rate increase — on ECR, a
5
USPS-T-7, p. 9.
I-7
product with a much higher price elasticity than Standard Regular, First-Class Mail letters and
cards, or Periodicals.
c.
Effect on Mailers (criterion 4)
Witness O’Hara testified: “At 8.9 percent, the rate increase for ECR is very close to
the overall average, and the impact on ECR mailers should be acceptable (criterion 4).”
USPS-T-31 (revised), p. 29, ll. 19-20. No explanation of why this increase should be
“acceptable” was offered. Valpak submits that the proposed increase is not acceptable.
In the first place, the attention that should be given to this factor in this docket is not
great, either in regard to a rate increase for Regular or a rate decrease for Regular. As witness
Mitchell demonstrated, the ECR coverage in this case needs to move now to an appropriate
level, so that the coverage thus established can serve as a suitable base for any future regime of
price caps under now recently-passed legislation. For the same reason, the coverage on
Regular needs to be increased. VP-T-1, p. 80. Second, in the course of Docket No. R2005-1,
the last omnibus rate case, the parties that might be impacted by rate adjustments, small or
large, plus or minus, chose to postpone the inevitable “true-up” of rates and thereby should not
be allowed to block the necessary rate adjustments as being “too much too fast.” The
Commission recognized this development in its Opinion and Recommended Decision:
Cost-based rates have been the touchstone of postal ratemaking
for 35 years, and the Commission has significant concerns about
deviating from that policy, even for a limited time. [Docket No
R2005-1, Op. & Rec. Dec., p. i.]
Facing an additional, imminent rate case most participants are
willing to accede to the Postal Service’s preference, and defer the
complex cost analyses necessary to apply efficient component
pricing principles until that case. [Id., p. ii.]
I-8
... small equal increases now, to be followed by a proceeding to
“true-up” rates after a thorough examination of postal costs, is
consistent with sound public policy. The Commission’s
preference is to develop rates that accurately reward mailers’
worksharing. It is concerned that the delay in recognizing the
impact of recent innovations and improvements in postal
operations, coupled with the passage of time, will probably
result in unusually disproportionate increases and decreases in
different rates in the next case. The Postal Service and mailers
seem prepared for that possibility as they too recognize that
proper cost-based rates foster efficiency and promote a healthy
postal system. [Id., p. ii, emphasis added.]
Rate shock arguments are often raised in rate proceedings.
They are likely to be raised in the next proceeding as well, in
which case the Commission will assess their merits based on the
record developed in that proceeding. Parties should be aware
that the Commission will seek to obtain economically efficient
cost-based rates and appropriate allocations of institutional
burdens. [Id., ¶ 5032 (emphasis added).]
Substantial changes in coverage have been made in the past, when in order. In Docket
No. R87-1, the Postal Service proposed a cost coverage for Express Mail of 210 percent, down
from a 239 percent coverage established in Docket No. R84-1. This proposed cost coverage
would have resulted in an 8.7 percent decrease in rates. The Commission reduced the
coverage even further, to 169 percent (an increase in rates of 2 percent under PRC costing).
Thus, the net effect was a lowering of the cost coverage for Express Mail of 70 percentage
points in one case. Docket No. R87-1, Op. & Rec. Dec., ¶ 6037 and Appendix G, Schedule
3, p. 1.
d.
Availability of Alternatives (criterion 5)
Witness O’Hara identified the availability of alternatives for users of ECR mail as
relatively high. He added: “Due to its geographic concentration, both alternate delivery firms
I-9
and newspaper inserts may provide alternative ways of delivering the same advertising
messages.” USPS-T-31 (revised), p. 30, ll. 3-6.
Witness Mitchell presented several reasons why the presence of alternative carriers
should not be used as a basis for an unusually high cost coverage for ECR or Regular. See
VP-T-1, p. 88.
1.
Considerable importance attaches to recognizing the value of the mail service,
and the value of ECR is low.
2.
Rates for ECR, particularly those of the saturation categories, are well above
stand-alone costs.
3.
The alternative provided by private carriers is not as attractive as is often
presumed. The mailbox rule reduces their attractiveness, not postal rates.
4.
Although private delivery options may be available legally to significant
portions of ECR mail, many ECR mailers of these portions do not view private
delivery as an attractive option, partly because of the mailbox rule, and
therefore view themselves as constrained to use the Postal Service.
5.
It is not fair to ECR mailers who cannot legally use private services to elevate
their rates because they are grouped with mailers who can legally use such
services.
6.
It is neither competitive nor fair to elevate rates purposefully to levels that allow
competition to exist, even though handicapped by the mailbox rule, and then to
use the alternatives provided by that competition as a basis for keeping the rates
elevated.
7.
The proportion of Nonprofit mail that legally can be carried privately is much
smaller than the proportion of Commercial mail, and rates must be fair to the
Nonprofit mailers as well.
8.
As between Regular and ECR, there is no reason for a policy preference
favoring one more than the other due to any limitation in their alternatives.
Both are used primarily, if not exclusively, by businesses, and both have a
growing list of alternatives.
I-10
On balance, arguments relating to the availability of alternatives and fairly competing with
providers of alternatives can support a markup at or near the systemwide average, but do not
support the exceedingly high markup on ECR that exists under the Postal Service proposal.
e.
Degree of Preparation (criterion 6)
Witness O’Hara observed that ECR mail has a very high degree of preparation by
mailers, as even the basic rate category must be line-of-travel sequenced, and the high-density
and saturation categories are walk-sequenced. USPS-T-31 (revised), p. 30, ll. 7-9. ECR
mailers engage in preparation activities to a greater degree than any other subclass. ECR mail
is presorted, barcoded, prepared in line-of-travel or walk sequence, and more extensively
dropshipped. In addition, ECR mailers’ container usage, acceptance processes, and postage
payment procedures tend to reduce postal costs. Also, ECR mailings’ density, which may be
thought of as the number of pieces per carrier route, as well as their bulk nature, reduces
Postal Service costs. As a result, the Postal Service’s costs for ECR mail are notably low —
especially for saturation mail.
As noted above, witness O’Hara apparently views this high degree of preparation as
constituting the soundest justification for the Postal Service’s imposition of an above-average
rate increase and extremely high markup. By contrast, the logic behind criterion 6 is the
realization that, when setting postal rates, it is more efficient to recognize and encourage a
greater degree of preparation through appropriate pricing. When costs are low, a natural
outcome of the competitive process is low rates. This allows economic benefits from the low
costs to be realized. The use of low costs to justify excessive markups runs directly counter to
the proper application of this criterion — particularly given ECR’s low value of service. This
I-11
factor cannot logically be used as a basis to help support a markup on ECR of anywhere near
the current level, or even barely above the systemwide average.
f.
Conclusion
For the reasons discussed above, objective application of the applicable noncost criteria
to Regular and ECR indicate that the markup on ECR should be below that on Regular and at
or below the systemwide average. Again, the only apparent explanation for the Postal
Service’s commitment to overburdening ECR is the established pattern of treating Standard
Regular and ECR as constituents of the same subclass, alongside a general reluctance to
increase the institutional cost burden on any subclass of mail, here Standard Regular mail.
2.
Continued imposition of an excessive cost coverage on ECR produces
Nonprofit ECR rates that do not reflect proper recognition of the favored
treatment due nonprofit mail.
By statute, the rates for all Nonprofit mail in the Standard class are dependent on the
cost coverages selected for Regular and ECR. When the rates for Commercial ECR are
elevated unduly, as results from a high cost coverage, then rates for the associated Nonprofit
category also are elevated unduly. This raises the question of whether it is fair and appropriate
for a category like Nonprofit ECR to have an extremely high cost coverage, particularly
relative to the cost coverage of Nonprofit Regular.
The Postal Service proposed Nonprofit Regular and Nonprofit ECR rates tend to
discriminate against local nonprofits. Yet nothing in the nature of local charities (tending to
use the higher-density Nonprofit ECR rates) justifies imposing higher markups on their mail
when contrasted to national mailers (tending to use lower-density Nonprofit Regular rates).
Moreover, although it is the clear intent of Congress that both types of nonprofit mail receive
I-12
special consideration in the ratemaking process — that is the rationale behind P.L. 106-384
(Reduced Rate Mail Modification Provisions), as well as prior statutes predating the
establishment of the Postal Service — nothing in that (or prior) legislation suggests preferences
are due Nonprofit Regular vis-a-vis Nonprofit ECR.
B.
ECR’s Cost Coverage Should Be Set Below that of Regular.
In Docket No. MC95-1, Regular and ECR were created as separate subclasses to
deaverage then third class bulk-rate regular. As distinct subclasses, Regular and ECR (each
now combined by law with a corresponding nonprofit category) warrant an independent
application of the ratesetting guidance in the Act. Such an analysis favors reducing the ECR
cost coverage below that of Regular, and placing both subclasses near the systemwide average.
1.
ECR’s cost coverage should be near the systemwide average and below that
of Regular.
The Commission long held that the average cost coverage for what is now Standard
Mail, once designated third-class, should be somewhat below the coverage on First-Class and,
generally, below the systemwide average. See VP-T-1, pp. 94-95. In its Opinion and
Recommended Decision in Docket No. R90-1, it said: “Over time we have consistently found
that First-Class should bear a markup at, or only slightly above, systemwide average.
Similarly, we have consistently found that third-class bulk regular, another subclass which is
largely subject to the statutory monopoly, should also bear an approximately average markup.”
Docket No. R90-1, Op. & Rec. Dec., ¶ 4022. The Commission’s Opinion and Recommended
Decision in Docket No. R94-1 stated: “As in past dockets, the Commission considers the
I-13
higher demand elasticity of third-class mail as an important reason for maintaining a lower
third-class markup relative to First-Class letters.” Docket No. R94-1, Op. & Rec. Dec.,
¶ 4027.
Of course, since Docket No. R90-1, third-class bulk regular has been divided into
Standard Regular and ECR. The cited high elasticity (in absolute value) is a characteristic of
ECR only, not Regular. Accordingly, application of the Commission’s findings indicates that
the cost coverage on Regular should be near the coverage on First-Class and that the cost
coverage on ECR should be below the coverage on Regular.
Given ECR’s dramatically higher (absolute) elasticity, the cost coverage on ECR should
be somewhat below the coverage on Regular. For these reasons, witness Mitchell’s proposed
Standard ECR and Regular coverages should be adopted by the Commission.
2.
Undue deference towards existing rate relationships between Regular and
ECR undermines realization of the purpose behind the establishment of
separate subclasses.
No basis exists for perpetuating rate relationships or implicit cost coverages that existed
prior to the establishment of the Standard Regular and ECR subclasses. Particularly given the
possibility that future rate levels will reflect price caps applied to rate relationships established
in this docket, it is time to move decisively toward appropriate cost coverages and rate
relationships.
In prior cases, the Postal Service acknowledged the excessive cost coverage imposed on
ECR, but the rate relationship between carrier route (which became ECR) and the remainder of
bulk rate regular (which became Standard Regular) essentially has remained in place. For
example, during Docket No. MC95-1, when ECR was created, witness Joseph D. Moeller
I-14
said: “This filing is a first step in moving toward a more market-based structure. USPS-T-1.
If not for the desire to avoid major rate relationship changes, I would propose a lower cost
coverage for Enhanced Carrier Route.” Docket No. MC95-1, USPS-T-18, p. 7, ll. 4-7.
In Docket No. R97-1, witness O’Hara testified: “[t]he Postal Service is proposing a
cost coverage of 228 percent for the Enhanced Carrier Route (ECR) subclass, which results in
a 3.2 percent average rate increase. This is somewhat below the system-wide average
increase, reflecting a desire to lower the very high cost coverage of this subclass.” Docket
No. R97-1, USPS-T-30, p. 34. Witness O’Hara further observed in an interrogatory response
that, “were it not for the effects of rate increases on mailers and the other factors discussed ...
a lower cost coverage for Standard A ECR would have been appropriate, which would have
meant higher rate increases (and coverages) for other subclasses (such as Standard A
Regular).”6
In Docket No. R2000-1, witness Virginia J. Mayes identified a proposed cost coverage
of 208.8 percent for ECR “which results in a 4.9 percent average rate increase. This is
somewhat below the system average increase [of 6.4 percent], reflecting a desire to lower the
very high cost coverage of this subclass.” Docket No. R2000-1, USPS-T-32, p. 38, ll. 4-8.
Witness Mayes acknowledged that ECR “was established as a subclass with the intent of more
directly reflecting the unique cost and market characteristics of this mail” (id., Response of
witness Mayes to NAA/USPS-T32-21, Tr. 11/4324), but did not apply this principle to
Regular as well. While she referred specifically to mitigating the exposure of mailers to
6
Docket No. R97-1, Response to DMA/USPS-T30-2, Tr. 2/131.
I-15
“increases in their costs,” she did not address the need to take steps in the direction of
recognizing each subclass’ “unique cost and market characteristics.” Id., Response of witness
Mayes to NAA/USPS-T32-23, Tr. 11/4327. Both the costs of the mail involved and the need
to take steps toward improved rate positions should be recognized in this docket.
Witness Mayes also agreed with the concern that “a lower coverage for ECR would
have made it more difficult to design rates so that the Automation 5-digit rate in Standard
Regular was below the ECR basic rate.” Docket No. R2000-1, Response of witness Mayes to
NAA/USPS-T32-27, Tr. 11/4332. The Commission since has made it clear that the coverage
for ECR should not be biased or constrained by such a concern.7 The Commission has also
consistently held that an independent application of the factors in the Act is warranted for a
subclass.
Docket No. R2000-1 was the last docket in which an extensive record based on a
genuine adversarial process was developed. However, the Postal Service’s initial proposal in
Docket No. R2001-1 was conventional, in that a settlement was not anticipated at the time the
case was prepared and filed. There, witness Moeller identified the proposed ECR/NECR cost
coverage as 217.8 percent, reflecting “a 6.2 percent average rate increase for ECR, and a 6.5
percent increase for NECR. These are somewhat below the system average increase,
reflecting a desire to lower the very high cost coverage of this subclass.” Docket No. R20011, USPS-T-28, p. 36, ll. 15-19. Witness Moeller explained further:
7
“The Commission finds persuasive witness Mitchell’s arguments for decoupling
the ECR basic and Standard Regular 5-digit automation rates, and expanding the letter/flat
differential at the basic level to better reflect cost differences.” Docket No. R2005-1, Op. &
Rec. Dec., ¶ 6075.
I-16
Although the percentage rate increase proposed for this subclass
is below the system average in this case, many of the factors
considered above indicate a cost coverage lower than that actually
proposed. However, a lower markup would mean shifting more
of the burden of covering institutional costs to other subclasses.
[Docket No. R2001-1, USPS-T-28, p. 38, ll. 11-14.]
In response to an interrogatory, witness Moeller stated: “The ‘desire’ to lower the cost
coverage for ECR is based on examination of the pricing criteria, and comparison of the ECR
coverage to the coverages for other subclasses.” Docket No. R2001-1, Response of witness
Moeller to NAA/USPS-T28-10, Tr. 9/2544.
Yet, notwithstanding repeated assessments of Postal Service witnesses, the cost
coverages proposed for ECR continue to reflect the treatment of carrier route mail within third
class bulk rate regular. No objective application of the noncost criteria has yet been applied by
the Postal Service to ECR in this docket.
It is ironic that Postal Service witnesses repeatedly acknowledge the low value of ECR,
yet seek high contributions through high coverages. At the same time, they point to the
difficulty of “making up” excessive ECR contributions from other subclasses, which have
lower elasticities. The reality is, however, that Standard Regular can withstand rate increases
with relatively little volume loss (and attendant value loss) while ECR cannot. This was shown
during the cross-examination of witness Thress (USPS-T-7). Specifically, it was shown that a
rate increase of 9.6 percent in Regular would result in a volume loss of only 2.19 percent,
while a somewhat smaller rate increase of 8.1 percent in ECR would result in a volume loss of
6.45 percent. VP-XE-1 (Thress), Tr. 6/1357. The approach of the Postal Service in each
omnibus rate case for a decade has been to prolong pricing benefits for Regular mailers at the
I-17
expense of ECR mailers.8 In every recent case, the Postal Service has proposed a high cost
coverage, with the result that the volume of ECR is much lower than it should be. It has been
more than a decade since Docket No. MC95-1 established distinct subclasses of Standard
Regular and Standard ECR. It is time to establish reasonable and appropriate coverage levels
for both subclasses.
Cost coverage issues, particularly for ECR, were a matter of concern in Docket No.
R2005-1. While in the end the Commission recommended the across-the-board increase
proposed by the Postal Service, it indicated that “[t]hese issues are not unimportant” and said:
“Furthermore, the Postal Service’s representations that it will, in the relatively near future, file
a general omnibus rate request allay the Commission’s concerns that the issues raised by
Valpak and alluded to by others will be unheard for an extended period.” Docket No. R20051, Op. & Rec. Dec., ¶ 5079. That omnibus case is now here and Valpak believes the record
supports appropriate consideration and change at this time.
C.
Postal Accountability and Enhancement Act — Transition Rule
Witness Mitchell proposed alternative cost coverages, the choice of which is dependent
on whether the Commission determined that the Postal Service would be able to file another
rate case (following the current docket) under the current ratemaking rules. VP-T-1, pp. 94-
8
It has now reached the point where the relatively low contribution from
Standard Regular (see Initial Brief of Valpak, Docket No. MC2005-3, pp. 23-24, 27) is used to
justify negotiated service agreements (“NSAs”) that give substantial discounts from First-Class
rates to encourage migration from Standard Regular. If Standard Regular had a more
appropriate markup, the “value” in such NSAs would decline markedly.
I-18
97. What was not clear at the time witness Mitchell completed his testimony was whether
postal reform legislation would be enacted in the 109th Congress and whether a “price cap”
would be imposed on future rate increases.
Postal Service witness Maura Robinson said that witness Mitchell was asking the
Commission to “speculate” and “pre-judge.” USPS-RT-10, p. 22, ll. 8-9, Tr. 32/10721. Of
course, as evident from witness Mitchell’s testimony, he was not asking the Commission to
speculate whether postal reform would be enacted in the 109th Congress, especially in light of
the fact that the Commission’s Recommended Decision would not be issued until well after the
109th Congress adjourned. He was suggesting that passage of postal reform could affect
whether the Commission had only one or two rate cases to adjust ECR coverage before a price
cap went into effect.
At this point, it is known that the 109th Congress did pass the Postal Accountability and
Enhancement Act, H.R. 6407 (House of Representatives action took place on December 8,
2006, while Senate action took place on December 9, 2006), and President Bush signed it into
law on December 20, 2006. Further, it is known that this bill implements a price cap regime
for future rate increases. P.L. 109-___ (H.R. 6407), § 201. However, this bill does contain a
“Transition Rule” whereby the Postal Service would be able to initiate a rate request to be
considered according to the current ratemaking rules (i.e., not under a price cap) for a period
of one year after enactment. The Transition Rule reads as follows:
(f) TRANSITION RULE . — For the 1-year period beginning on the
date of enactment of this section, rates and classes for marketdominant products shall remain subject to modification in
accordance with the provisions of this chapter and section 407, as
such provisions were last in effect before the date of enactment of
I-19
this section. Proceedings initiated to consider a request for a
recommended decision filed by the Postal Service during that 1year period shall be completed in accordance with subchapter II
of chapter 36 of this title and implementing regulations, as in
effect before the date of enactment of this section. [P.L. 109-___
(H.R. 6407), § 201 (emphasis added).]
If it appears clear that the Postal Service will file such a rate case, the step to a rational
relationship between coverages for Standard ECR and Standard Regular could be made into
two steps. However, if it is not known that another rate case under current law will be filed,
the Commission should use the certain opportunity presented by the instant rate case to set
relative rates of Regular and ECR by adjusting relative coverages to conform with the noncost
criteria of the Act.
II-1
II.
THE COMMISSION SHOULD CONTINUE TO TREAT MAIL PROCESSING
COSTS AS FULLY ATTRIBUTABLE.
Since Docket No. R97-1, the Postal Service has endeavored to reduce the volume
variability of mail processing labor costs from the nearly 100 percent standard originally set by
the Commission in Docket No. R71-1. Notwithstanding these repeated efforts by the Postal
Service, the Commission never has deviated from this established methodology.1
In this docket, Postal Service witness A. Thomas Bozzo (USPS-T-12) again presented
the Postal Service’s analysis of the volume variability of Cost Segment 3 mail processing costs.
His testimony was essentially an update of his analysis presented (but not litigated) in Docket
No. R2005-1.2 This testimony again has generated significant opposition. Office of the
Consumer Advocate (“OCA”) witness Mark J. Roberts (OCA-T-1) explains various
econometric problems associated with Dr. Bozzo’s analysis, and United Parcel Service
(“UPS”) witness Kevin Neels (UPS-T-1) discusses data quality issues, as well as other
problems with Dr. Bozzo’s model. Dr. Haldi’s analysis (VP-T-2, pp. 10-55), discussed
below, exposes certain larger conceptual problems arising from Dr. Bozzo’s analysis.
1
The Commission has recommended settlement rates which incorporated the
Postal Service’s assumptions in Docket Nos. R2001-1 and R2005-1, but upon the express
understanding of the settling parties that the Opinion and Recommended Decisions in those
dockets had no precedential value as to matters not fully litigated. See, e.g., Docket No.
R2005-1, Stipulation and Agreement, ¶ 16 (July 22, 2005). In noting that the settlement did
not bind signatories to any ratemaking or costing principle or interpretation of law, the
Commission underscored that this “assures that the results of the settlement will not have
consequences beyond their immediate application.” Docket No. R2005-1, Op. & Rec. Dec.,
¶ 5089.
2
Docket No. R2005-1, USPS-T-12.
II-2
So long as all mail processing costs in Cost Segment 3 are treated as 100 percent
volume variable, and thereby 100 percent attributable, the issue of how to treat non-volume
variable costs simply does not exist. Under Dr. Bozzo’s proposal to reduce the volume
variability of mail processing costs, however, proper treatment of non-volume variable mail
processing costs — i.e., whether institutional, or product specific intrinsic — becomes a $2.4
billion issue which then would need to be resolved. In this docket, this complex question of
great significance was addressed only by Dr. Haldi. Based on the record in this docket, the
Commission should retain its established practice, as the Postal Service has not justified a
divergence from established practice.3
A.
Witness Bozzo’s Analysis Does Not Withstand Scrutiny.
1.
Witness Bozzo fails to address economies of scale.
Dr. Haldi observed in Section IV of his testimony (VP-T-2, pp. 17-31) that Dr.
Bozzo’s analysis of mail processing cost volume variability is deficient, in that it fails to
address economies of scale. Dr. Haldi explained in his testimony that:
“economies of scale” pertains to a declining envelope curve
(often referred to in economic textbooks as the long-run average
3
If the Postal Service wishes to continue its studies of this issue, it will have
other opportunities to present further findings and updates in the future. The recently passed
Postal Accountability and Enhancement Act (P.L. 109-___, H.R. 6407) contains several
provisions which no doubt will allow a fresh look at cost attribution issues. For example,
Section 204 requires an annual report from the Postal Service to the Commission that “shall
analyze costs, revenues, rates, and quality of service, using such methodologies as the
Commission shall by regulation prescribe....” Section 702 requires that the Commission make
a report containing “an estimate of the costs of the Postal Service attributable to the obligation
to provide universal service under current law....”
II-3
cost curve — or “LRAC”), which, in turn, is defined by the cost
curves of facilities of different sizes. If LRAC is not declining,
the long-run variability should be 100 percent (or greater). In
order to accept variability less than one, witness Bozzo should
have demonstrated that we are operating in a range of scale
economies. He has not done so. [VP-T-2, p. 27, ll. 3-10.]
Dr. Bozzo’s analysis focused on the volume variability of labor costs in individual
Management Operating Data System (“MODS”) cost pools, and in almost every cost pool
analyzed he found volume variability to be significantly less than 100 percent. If taken at face
value, this finding would indicate that an increase (or decrease) in volume of, say, 10 percent,
would result in costs increasing (or decreasing) by less than 10 percent. Under certain
circumstances, a finding that costs change less than proportionately with volume might be
interpreted to indicate not that costs are less than 100 percent volume variable, but rather the
existence of economies of scale in mail processing operations. (In fairness to Dr. Bozzo, he
explicitly disclaims having studied economies of scale, or how mail processing costs change as
capacity and plant size vary in response to changes in volume. Response of witness Bozzo to
VP/USPS-T12-4, Tr. 10/2656-57.)
The model used by Dr. Bozzo isolates and focuses on individual MODS cost pools. An
implicit assumption underlying his model is that the Postal Service’s mail processing
production function consists of cost pools that are completely independent of each other. As
Dr. Haldi pointed out, cost pools are interdependent in those facilities where some volume
frequently — or even routinely — gets directed to manual (e.g., because total volume exceeds
automated capacity). At such facilities, marginal cost may be rising sharply (because any and
all additional volume is being processed manually), yet in Dr. Bozzo’s analysis, both the
II-4
automated and the manual cost pools would have volume variability less than 100 percent
(because of setup costs), and lead to the incorrect inference that costs would increase less than
proportionately for any additional volume. VP-T-2, p. 23. This consideration is highly
pertinent, since 34.2 percent of total Postal Service mail processing costs in FY 2005
reflected manual processing. Response of witness Haldi to USPS/VP-T2-5, Tr. 23/8592. A
model incorporating the flawed assumption that every cost pool is independent makes it highly
unlikely that Dr. Bozzo ever would — or could — identify increasing marginal costs in any
facility, regardless of the reality. A more holistic approach would be necessary in order for
such a finding ever to emerge. An important limitation of Dr. Bozzo’s study of volume
variability in isolated MODS cost pools is his failure to show how facilities’ mail processing
costs respond to increases in volume, both within existing capacity limits and as capacity
changes in response to a permanent change in volume.
In response, Dr. Bozzo asserts that “Dr. Haldi’s criticism of the Postal Service models
as failing to provide information on ‘economies of scale’ is beside the point of the analysis”
(USPS-RT-5, p. 2, ll. 8-12, Tr. 36/12394). Dr. Bozzo’s study was limited to determining the
volume variability of Postal Service mail processing costs for individual cost pools, not how
mail processing costs vary with changes in capacity, or how costs vary within an entire facility
as volume changes. When cost pools for the same shape are interdependent, however, as they
so often are (e.g., an overflow of flats in the Automated Flats Sorter Machine (“AFSM”) 100
cost pool is shunted to the FSM 1000 or manual flats cost pool for processing), Dr. Bozzo’s
approach is not capable of depicting the correct production function relationship. The way in
which total mail processing costs change as volume and capacity change would seem to be
II-5
highly pertinent to attribution considerations, and an empirical investigation of longer-run
variability would be especially germane to assessing volume variability of mail processing
costs for attribution purposes. On this basis alone, the analysis presented by Dr. Bozzo in this
docket should not be utilized for the purpose of reducing attribution of costs to the classes and
subclasses of mail.
Dr. Bozzo further asserts that “the labor demand analysis bears on economies of
‘density’ rather than economies of scale.” USPS-RT-5, p. 4, ll. 1-2, Tr. 36/12396. He
suggests by analogy that Postal Service operations can be likened to the circumstances of an
airline which “can lower its average costs by filling otherwise-empty seats on both a 150-seat
jet and on a 400-seat jumbo.” USPS-RT-2, p. 4, ll. 15-17, Tr. 36/12396. However, he fails
to explain how an economic concept developed to reflect the economics of an airline system
properly applies to the volume variability of mail processing costs.4 Nor does the assumption
that Postal Service facilities always enjoy near universal excess processing capacity appear to
reflect reality.
2.
Setup and takedown costs do not vary uniformly with mail volume, but they
nevertheless can be volume variable.
The Postal Service processes mail by shape. Hence, all setup and takedown costs are
incurred with respect to specific shapes of mail; e.g., letters, flats and parcels. Within each
4
Just as aircraft sometimes have empty seats, it is undisputed that, during some
days of each month postal facilities have some unutilized capacity — i.e., economies of
density. However, the two industries also exhibit important differences. For example, the
Postal Service does not offer different rates for non-peak times during the day, the week, the
month, or the year, whereas airlines have time-of-day and time-of-week fares, and they change
their fare structure frequently.
II-6
shape, because of differing service standards, the Postal Service generally will process classes
separately and with different service standards, such as First-Class and Standard, at least until
they are ready to be merged for final delivery by carriers.
According to Postal Service witness Marc D. McCrery (USPS-T-42), when a facility
does not have sufficient volume to justify separate outgoing sortations of Standard and FirstClass Mail, the two may be merged in a single sortation. Response of witness McCrery to
VP/USPS-T42-7-10, Tr. 11/3106-3113. As Dr. Haldi points out, at some point, in response to
an increase in volume, the facility will incur the additional cost to set up and take down the
machine twice, instead of only once. Similarly, if originating volume decreased at such a
facility, at some point, the facility could be expected to merge the sorting of its First-Class and
Standard Mail, thereby eliminating the cost of one setup and one takedown operation. Thus,
within a low-volume range, setup and takedown costs are seen to vary with — and are caused
by — changes in volume (i.e., they are volume variable) although, as volume changes, the
variation resembles a discontinuous step function, rather than a continuous uniform function.
VP-T-2, p. 29. Since different subclasses such as First-Class and Standard are merged
routinely when the volume is low, and separated as volume increases, the additional cost of
separate setup costs for Standard Mail properly could, and perhaps should, be viewed as
volume variable costs, and therefore attributed.
Dr. Haldi makes a somewhat similar observation about facilities with multiple
machines and large volumes of mail. As volume increases, at some point, a marginal increase
in volume can result in the facility running identical sortation schemes on two (or more)
II-7
machines in order to meet service standards, while accommodating existing volume.5 When a
marginal increase in volume causes the same sort scheme to be run on an additional machine,
then for mail processed on the extra sort scheme(s), setup and takedown costs per unit of mail
volume, rather than decreasing, will increase. The situation is symmetrical with regard to
decreasing volume. When the same sort scheme is being run on two (or more) machines, at
some point, a marginal decrease in volume will enable the sort scheme on one machine to be
eliminated. At that point, a marginal decrease in volume will result in elimination of the entire
setup and takedown cost for one machine. Thus, in this high-volume range, setup and
takedown costs associated with all additional, identical sort schemes, also are seen to vary
with — and are caused by — changes in volume (i.e., they are volume variable), although as
volume changes, the change in cost again is a discontinuous step function, rather than a
continuous uniform function. VP-T-2, p. 30.
Situations such as those described above may arise only on limited occasions. But they
illustrate that, under certain conditions, setup and takedown costs are caused by changes in
volume, hence are volume variable, and should be attributed, regardless of whether economies
of density are deemed to exist.6 Under such circumstances, attribution becomes an empirical
matter which needs to be decided on the basis of the facts and pertinent data. As the
5
To borrow from Dr. Bozzo’s airline analogy, this situation has certain
similarities to an airline that has to add a second aircraft to a route in order to accommodate all
passengers who want to fly between two cities.
6
Such situations also illustrate that (i) economies of density and economies of
scale are quite different concepts, and (ii) they do not necessarily behave in the same manner.
II-8
necessary data have not been presented on the record in this docket, attribution should not
be changed.
3.
Witness Bozzo would increase institutional costs by $2.4 billion.
Using the volume variabilities of the MODS cost pools explicitly analyzed by Dr.
Bozzo, Postal Service witness Eliane Van-Ty-Smith (USPS-T-11) computed non-volume
variable costs of some $720.7 million. She extrapolated Dr. Bozzo’s analysis of volume
variability in the MODS costs pools which he analyzed to mail processing costs incurred
within all plants and BMCs, including those that do not have MODS, as well as post office
stations and branches. With this extrapolation, witness Van-Ty-Smith estimated total nonvolume variable mail processing cost to be $2.38 billion. USPS-T-11, Attachment Table 1,
pp. 32-33. As Dr. Haldi observes, it appears that the Postal Service would have the
Commission add virtually the entire amount to institutional costs and distribute it to all classes
and subclasses according to the non-cost factors contained in 39 U.S.C. § 3622(b). VP-T-2,
pp. 34-35. As explained above, it cannot just be assumed that none of these costs should be
attributed without further empirical investigation of the facts concerning their nature and their
testing on the record.
B.
Many of Witness Bozzo’s Non-volume Variable Costs Are Attributable.
Setup and takedown costs provide the basis for most of Dr. Bozzo’s findings that
volume variability of mail processing cost is less than 100 percent. Closer analysis indicates
that a substantial portion of those costs Dr. Bozzo finds to be non-volume variable deserve to
be classified as specific fixed intrinsic costs, and therefore should continue to be attributed.
II-9
1.
Non-volume variable costs are incremental costs.
Since mail processing is organized by shape, for the 11 MODS cost pools in which mail
is sorted, any cost deemed to be non-volume variable is clearly an incremental cost of the
sortation services provided to each respective shape involved, as Dr. Haldi observes. VP-T2, pp. 43-46. Thus, any non-volume variable costs in the BCS/DBCS cost pool are clearly
part of the incremental cost of letter sortation services provided to letter products. Similarly,
any non-volume variable costs in the AFSM 100 cost pool are part of the incremental cost of
flat sortation services provided to flat products. Finally, any non-volume variable costs in
parcel sorting cost pools are an incremental cost of parcel products. Postal Service witness
Dion E. Pifer (USPS-T-18) concurs that the setup and takedown costs of each sort scheme are
incremental to that particular sort scheme, but notes that “this does not imply that the nonvolume variable costs ... are incremental to any Postal Service product.” Response of witness
Pifer to VP/USPS-T18-4-6, Tr. 9/2223-25.
Within individual cost pools, such as the letter or flat sorting cost pools, some sort
schemes, such as an outgoing primary or secondary sort, may be performed exclusively, or
almost exclusively, on only one class of mail (e.g., First-Class Mail). Further, although most
classes of mail receive an incoming primary sort, because of different service standards, FirstClass and Standard Mail usually are kept separate during incoming primary sortation, and are
merged only when sorted to carriers or to carrier sequence. Whenever a particular sort scheme
is used for separate classes of mail, both the setup and takedown time are incremental to that
sort scheme and to the single mail product, or class of mail, sorted. Many of the sortation
schemes which the Postal Service runs every day are for a single class of mail.
II-10
Within MODS cost pools, many identifiable sortation operations are confined
exclusively to a single class or subclass of mail. At other times, when not confined solely to a
single class, the volume of mail from some other class is incidental to the sort scheme being
run. When any mail processing operations are exclusively, or primarily, for the benefit of
one class or subclass, it would seem clear that any non-volume variable costs arising from
those operations (i) pertain solely to the subclass in question, (ii) are an incremental cost of the
subclass which is served exclusively, or primarily, by the operation in question (e.g., sorting
or cancellation), (iii) are product-specific to the product in question,7 and (iv) thus should be
attributed as product specific intrinsic costs.
2.
Witness Bozzo’s rebuttal testimony lacks merit.
In his rebuttal testimony, Dr. Bozzo asserted that:
!
most setup and takedown costs are not volume-variable (i.e., “attributable” as
marginal costs);
!
a significant portion of setup and takedown costs are for multiple classes and
cannot be assigned to any class of mail as incremental cost;
!
even those setup and takedown costs that may be class-specific are not, in
general, the incremental costs of any subclass; and
!
the Postal Service incremental cost model includes “inframarginal” variable
costs, so the possibility that setup costs may represent costs that may vary with
volume in some respects, but not “on the margin,” is already incorporated in the
incremental cost estimates. [USPS-RT-5, p. 6, ll. 5-16, Tr. 36/12398.]
In response to Dr. Bozzo’s first point, it is clear that, whenever the same sort scheme
is being run on multiple machines because the volume is too great for one machine to
7
Witness Pifer (USPS-T-18) concurs; see response to VP/USPS-T1810(a), Tr. 9/2229.
II-11
accommodate all existing volume within the available operating window, the setup and
takedown costs on the additional machines are caused by the additional volume, and should be
considered volume variable, even though the variation in cost represents a slight discontinuity.8
As noted above, the extent to which identical sort schemes are run on multiple machines is an
empirical matter that needs to, and can, be settled only by relevant data, which have not been
submitted in this docket.
In response to Dr. Bozzo’s second point, setup costs incurred by multiple classes of
mail cannot properly be attributed to any class or subclass. He asserts that a “significant
portion” of the setup costs cannot be attributed for this reason, but he makes no effort to
quantify the portion of his non-volume variable costs that might reasonably be classified as
common costs. The bald assertion that all class-specific costs cannot be attributed lacks any
foundation and is completely conclusory.
In response to Dr. Bozzo’s third point, he argues that even when setup and takedown
costs are incurred for a single class of mail, they cannot be attributed to any mail in that class
if (i) it contains multiple subclasses, and (ii) two or more of those subclasses are being sorted
8
The Postal Service has a substantial number of costs that, technically speaking,
are discontinuous and do not vary at the margin, when “the margin” is interpreted to mean a
single piece — or just a very small quantity — of mail. Supervisor cost is an example of one
such “discontinuous” cost. In practice, however, these discontinuities are appropriately
ignored, and supervisor costs are treated as a continuous variable. See Dr. Haldi’s response to
USPS/VP-T2-19, Tr. 23/8612, for further discussion. Neither Dr. Bozzo, nor any other
Postal Service witness, explains what distinguishes the tens of thousands of discontinuities
associated with takedown and setup costs from all the other small discontinuities in the Postal
Service’s production cost function that routinely are treated as continuous, volume variable
and, hence, attributable. Without a solid, logical distinction, it would not be illogical for Dr.
Bozzo’s analysis to open the door to treating every “discontinuous” cost, such as supervisory
costs, as non-volume variable.
II-12
simultaneously on the same scheme. In his view, attribution is a two-step process requiring,
first, that virtually all attributed costs be volume variable and, second, that volume variable
costs change continuously with and reflect each marginal change in the volume of mail in
each subclass. Stated otherwise, even though setup and takedown costs may be incurred
solely for a single class of mail and, even though it is acknowledged that those costs are
caused by and are incremental solely to that class of mail, he rejects any attribution of those
costs on grounds that they are invariant with respect to the volume of mail in each subclass
being processed simultaneously on the sort scheme.9
Dr. Bozzo’s absolutist position concerning attribution of Cost Segment 3 setup and
takedown costs stands in sharp contrast to the attribution of Cost Segment 8 volume variable
service vehicle costs for local transportation (Vehicle Service Drivers). In that cost segment:
the volume variable costs of Vehicle Service Driver labor are
distributed to classes and subclasses of mail in the same
proportions as cubic feet of total (local and non-local) mail,
obtained from Revenue, Pieces, and Weight (RPW) statistics
(adjusted to include cubic feet for Mailgrams). [USPS-LR-L-1,
p. 8-3.]
9
Magazine Publishers of America, Inc. et al. (“MPA”) witness Stuart W. Elliot
(MPA et al.-RT-2) sought to rebut Dr. Haldi’s testimony, and in doing so he has demonstrated
the dangers associated with the type of purist position advocated by Dr. Bozzo. Dr. Elliot is
actually more rigorously consistent with Dr. Bozzo’s principles when he says that he would
demand total purity within the mail processing stream before he would allow the setup and
takedown mail processing costs to be attributed. On cross-examination, he stated that the
presence of even an incidental (defined as under 0.1 percent) percentage of Periodicals letters
within the First-Class letter processing stream would bar the attribution of any of these setup
and takedown letter processing costs, and presumably would cause them all to be converted to
institutional costs. Tr. 36/12377, ll. 14-22.
II-13
Using cubic feet of total (local and non-local) mail obtained from RPW as the distribution key
means, for any particular subclass, that the Segment 8 costs which are attributed to it do not —
and will not — change even if the subclass in question were to use significantly less local
transportation; e.g., because the subclass in question experienced a substantial increase in the
volume of mail entered at DDUs. By the same token, Segment 8 costs attributed to a
particular subclass will not change even if usage of local transportation were to undergo a
significant increase; e.g., because the Postal Service decided to backhaul a substantial volume
of mail in that subclass from DDUs to plants. Thus, attribution of costs within Segment 8 is
seen to be a fixed apportionment, regardless of usage, based on each subclass’s share of total
RPW cubic feet.
Because the Segment 8 costs attributed to each subclass are totally invariant with
respect to (i) actual usage of local transportation by each class and subclass, as well as
(ii) changes in the usage of local transportation by each class and subclass, no causal linkage
exists with respect to the amount apportioned to each subclass. Although it is volume
variable costs in Segment 8 that are being apportioned all the various classes, from the
perspective of any given class (or subclass), the end result is at least as arbitrary, if not more
so, than an apportionment of costs that are specific fixed and acknowledged to be clearly
incremental to a single class of mail. Despite the lack of causal linkage and the total
arbitrariness involved, the Postal Service nevertheless attributes these Segment 8 costs, and has
not proposed any change in costing methodology for that cost segment.
The attribution of Segment 8 costs thus is seen to be at total variance with Dr. Bozzo’s
stated criteria that, in order for costs to be attributed, they must change continuously with
II-14
and reflect each marginal change in the volume of mail in each subclass. The Postal
Service should not be allowed to have it both ways. If the current invariant method of
apportioning Segment 8 volume variable costs is acceptable, then there is no reason not to
attribute costs that clearly are specific and incremental to a single class of mail on a similar
basis. Conversely, if the criteria required by Dr. Bozzo’s two-step attribution method are held
to be the only valid way to attribute costs, then the attribution of Segment 8 costs to individual
classes and subclasses must be changed to incorporate a causal link to actual usage of local
transportation by each subclass of mail.
Dr. Bozzo’s fourth observation asserts that “the possibility that setup costs may
represent costs that may vary with volume in some respects, but not ‘on the margin,’ is already
incorporated in the incremental cost estimates.” USPS-RT-5, p. 6, Tr. 36/12398. This
statement fails to acknowledge that the Postal Service’s method for estimating incremental cost
is totally incapable of identifying any inframarginal costs that are specific fixed intrinsic costs
and that, when properly classified as such, should be attributed. In that respect, the Postal
Service’s incremental cost model is woefully deficient.
Finally, if adopted, Dr. Bozzo’s favored approach presents the Commission with the
curious situation that, as more postal products are offered, the less likely it is that any mail
processing operation would meet the economic test of being associated with a single
product, and, therefore, fewer and fewer mail processing costs will be attributed to any class
of mail. There should be no connection between the number of postal products and the level of
attribution of mail processing costs but, with Dr. Bozzo’s approach, there is a direct
correlation.
II-15
C.
Witness Bozzo Failed to Address the Commission’s Concerns.
As noted above, the Commission never has accepted the Postal Service’s proposed
methodology incorporating volume variability of less than 100 percent for mail processing
costs, notwithstanding repeated efforts by the Postal Service to modify this practice. One
consistent difficulty the Postal Service has yet to overcome is the repeated failure to address
prior concerns raised by the Commission.
1.
Witness Bozzo did not address the Commission’s analysis regarding setup
and takedown costs.
For example, in this docket, Dr. Bozzo failed to address explicitly an important
concern raised by the Commission in Docket No. R2005-1 regarding his methodology.
Specifically, the Commission observed:
1.
The setup/takedown time is always a part of each mail
processing run, so many runs would require many
setup/takedown times.
2.
The setup/takedown time is only present at the beginning
(setup time) and the end (takedown time) of a long series
of continuous nonstop mail processing runs, so many runs
would require only one setup/takedown time.
In Scenario 1, if the setup/takedown time would always be
present with each run, then it would be a part of the costs for
each run. Thus, if volume were to double and require a second
run, associated costs would double: variability is equal to one. In
Scenario 2, if the setup/takedown time costs would only be
present at the beginning and end of a long series of continuous
nonstop mail processing runs, then these costs would appear to be
very small. Thus, if volume were to double, the costs associated
with its processing would not appear to be significantly different:
variability is not significantly different from one. [Docket No.
R2005-1, Op. & Rec. Dec., App. I, p. 52.]
II-16
These concerns are quite similar to the points raised in Dr. Haldi’s testimony and
discussed above. And, as noted above, Dr. Bozzo not only has failed to refute the points
raised by Dr. Haldi; neither he (nor any other Postal Service witness) has addressed the
Commission’s prior analysis. Instead, he simply has reiterated prior Postal Service analyses
without re-considering the challenged foundations that underlie his model.
2.
Witness Bozzo did not address the impact of his analysis on worksharing
discounts.
Lastly, in neither his direct testimony nor his rebuttal testimony does Dr. Bozzo
address on the record the Commission’s concerns about the impact that his proposed attribution
methodology would have on worksharing discounts. For example, the Commission observed
in Docket No. R2000-1 that:
Because the variabilities estimated by the Postal Service are
dramatically lower [than 100 percent], using them to estimate the
costs avoided by worksharing would dramatically shrink the
estimated costs avoided. This would require an equally dramatic
reduction of the discounts offered for worksharing if they were
to accurately reflect the underlying cost savings using them to
estimate the costs avoided by worksharing would dramatically
shrink the estimated costs avoided. This would require an
equally dramatic reduction of the discounts offered for
worksharing if they were to accurately reflect the underlying cost
savings.... [Docket No. R2000-1, Op. & Rec. Dec., Appendix
F, p. 37 (emphasis added).]
Clearly, in this docket, the Postal Service again has failed to give the Commission an
adequate record basis on which to change its long-standing practice in the treatment of mail
processing costs.
III-1
III.
WITNESS HALDI (VP-T-2) IDENTIFIES SPECIFIC PROBLEMS WITH
CURRENT COSTING SYSTEMS AFFECTING STANDARD MAIL, WHILE
WITNESS MITCHELL (VP-T-3) ASKS THAT THE COMMISSION ADDRESS
ALL SUCH “FAIRNESS IN COSTING” PROBLEMS IN A SUBSEQUENT
DOCKET.
Cost-based rates always have been the centerpiece of postal ratemaking.1 Therefore, it
is axiomatic that good ratesetting must begin with accurate costs. The Postal Service expends
substantial resources in developing its cost estimates, but those efforts can and do fall short at
times. Notably, costing problems can occur when the assumptions underlying a given costing
approach fail to reflect actual operational conditions.
Underlying the costing methodology used by the Postal Service are a number of critical
assumptions that rarely are stated explicitly, and that are far from obvious. One such
assumption is that “Postal Service costing methods do not presuppose persistent processing
capacity constraints.”2 Nor do Postal Service costing methods presuppose persistent delivery
capacity constraints. Yet, as discussed below, information brought to light in this docket
shows that in fact such constraints do exist.
Another important assumption underlying the Postal Service’s costing methodology is
that cost is additively separable across cost components, such as mail processing cost pools.3
This assumption, known as “separability,” requires that cost pools be independent of each
1
See, e.g., “Cost-based rates have been the touchstone of postal ratemaking for
35 years....” Docket No. R2005-1, Op. & Rec. Dec., p. i.
2
3
The Postal Service acknowledged this in Docket No. R2005-1, Tr. 8D/5169.
See VP-T-2, p. 61, citing the Data Quality Study, Technical Report #1:
Economic Analysis, April 16, 1999, by A.T. Kearney Inc.
III-2
other in order for the Postal Service’s costing methods to validly reflect individual product
costs in all circumstances. As Dr. Haldi’s testimony states (VP-T-2, pp. 61-61, citing the Data
Quality Study), “the Postal Service’s costing methodology prohibits the level of volume (or the
level of capital) in one cost pool from affecting the level of cost in another cost pool.”
Therefore, as Dr. Haldi points out, “when the assumptions do not fit the facts, either because
they are too simplifying or just plain wrong, the conclusions do not follow and they should not
be relied on.” (Id., p. 63)
The existence of these critical assumptions means that when costs are generated using
an inappropriate methodology, they should not be taken at face value in ratemaking despite any
surface appearance of “correctness.”4 It is submitted that the costing problem associated with
saturation letters and flats discussed below is one of those cases. At such times, the
Commission either has adjusted Postal Service estimated costs, or has urged the improvement
of costing approaches for the next case, or both.5
4
As one example of a systematic problem with current Postal Service costing, see
the discussion, infra, at Section II discussing the error in Dr. Bozzo’s flawed assumption that
each cost pool operates totally independently of, and can be analyzed separately from, other
cost pools.
5
See, e.g., Docket No. R2000-1, Op. & Rec. Dec., ¶¶ 5604-05 (adopting an
intervenor’s estimate of the savings from reducing bundle breakage in periodicals and rejecting
the Postal Service’s estimate), and ¶ 5831 (urging the Postal Service to provide more analysis
of the relationship between cube and Parcel Post rates).
III-3
A.
The Problem in Costing of Saturation Letters and Flats Created by Capacity
Constraints
The basic problem is that the Postal Service’s third-bundle, or “extra” bundle, delivery
method creates a peculiar, previously overlooked problem for proper costing of ECR saturation
letters and flats (including both addressed flats, and unaddressed flats accompanied by DALs).
Stated simply, there is a limit to the number of extra bundles that city carriers can take directly
to the street. When that limit is reached, the Postal Service must make an operational decision
to take to the street either flats or letters, but not both. Whichever product is taken directly to
the street has the distinct benefit of zero in-office time for casing, and virtually no in-office
time in handling.6 Letters can be DPS’d or cased at considerably less cost than flats can be
cased.7 Consequently, it can be expected that flats are the products routinely taken to the
street, not letters. Therefore, whenever such a conflict occurs, the Postal Service costing
systems reflect higher incurred cost to handle the relatively low-cost letters, and lower incurred
cost to handle the relatively high-cost flats.
6
Postal Service witness Joyce K. Coombs (USPS-T-44) explained that when a
saturation mailing is taken directly to the street, the in-office cost of handling that mailing is
essentially zero. Tr. 13/3741, ll. 14-18. Obviously then, the lowest (zero) cost way to handle
saturation letters, if there were no saturation flats, would be to take them directly to the street.
7
Although automated sequencing of letters is now routine, automated sequencing
of flats is not yet a possibility. With respect to manual casing of letters and flats, the
traditional standard has been 8 and 18 per minute for letters and flats, respectively. In Docket
No. R90-1, Postal Service witness Shipe testified that saturation letters and flats could be cased
at rates of 47.4 pieces per minute for letters and 31.5 pieces per minute for flats. USPS-T-10,
p. 19, ll. 4-11. The test results reflected bound or enveloped flats. Roughly half of all
saturation flats consist of covers, or wraps. These are considered to be much more timeconsuming to case than bound catalogs, and carriers virtually never case covers or wraps. As
a result, in the Postal Service’s costing methodology, the most difficult-to-handle pieces wind
up with the lowest recorded cost of all saturation mail.
III-4
In addition to forcing letters to incur mail processing cost for DPSing (as opposed to
zero in-office cost), denial of treatment as extra bundles and sequenced mail also results in a 43
percent increase in on-street delivery cost, as discussed in more detail in Section C.3, infra.
If rates are predicated on flawed costs, mailers will be charged rates that are unfair, and
be given wrong pricing signals. If pricing signals are sufficiently incorrect, and mailers
respond rationally to those pricing signals, they could wind up converting letter mailings to
flats in order to enjoy the lower in-office handling costs, while increasing total postal costs.
B.
Dr. Haldi’s Testimony in Docket No. R2005-1
In Docket No. R2005-1, Dr. Haldi explained the costing problem in VP-T-2, Section
IV, pages 25-57. Naturally, his testimony was based on such information as was available at
that time, which came largely from Postal Service delivery witness Jeffrey W. Lewis (USPS-T30). Witness Lewis had testified on oral cross-examination that conflicts only occur on
walking portions of routes, as motorized carriers could work from many bundles
simultaneously, but perhaps less than 10 or 12. Docket No. R2005-1, Tr. 11/5997, ll. 21-25.
Further, in response to Dr. Haldi’s direct testimony, witness Lewis filed rebuttal
testimony (USPS-RT-2), which explained a brief study he conducted to try to disprove Dr.
Haldi’s theory about conflicts. That study assumed that conflicts could only occur on walking
portions of routes and he claimed that the number of occasions when capacity constraints
occurred were minimal, in the range of 10 percent. Advo, Inc. witness Antoinette Crowder
seemed to claim that conflicts could virtually never occur. ADVO-RT-1, p. 30-40. Lastly,
Postal Service witness Michael D. Bradley (USPS-RT-3), without mentioning (much less
III-5
attempting to justify) the several critical assumptions which underlie the Postal Service’s
costing model, provided the customary mathematical “proof” that the existing costing
methodology generates valid marginal costs at all times and under all conditions. Based on the
Docket No. R2005-1 testimony of witness Bradley “that current cost system is designed to
reflect current operating procedures”8 (USPS-RT-3, p. 18, Tr. 11/6055) and the testimony of
witness Lewis “suggesting the limited nature of a capacity constraint” (USPS-RT-2, p. 8, Tr.
11/5945.), the Commission rejected Dr. Haldi’s proposed alternative costing. Docket No.
R2005-1, Op. & Rec. Dec., p. 142, ¶ 6092.
C.
Dr. Haldi’s Testimony in Docket No. R2006-1
In this docket, the Postal Service provided new and different record information
which differs significantly from that testified to in Docket No. R2005-1. Dr. Haldi’s direct
testimony in this docket brings this new information to the Commission’s attention. See VP-T2, Section VIII, pp. 56-73. Interestingly, there was no rebuttal testimony from the Postal
Service in this docket seeking to minimize delivery capacity constraints, or to defend the
ability of Postal Service costing to develop accurate marginal costs in the face of constrained
capacity. Only Advo, Inc. (“Advo”) and Saturation Mailers Coalition (“SMC”) witness
8
The Postal Service’s In-Office Cost System (“IOCS”) cost methodology simply
observes what is being done (at some instant of time), not why anything is done in the manner
observed. This work-sampling methodology assumes that all activities observed in every cost
pool are completely independent of every other cost pool — which often is not the case.
Since the cost methodology is incapable of explaining why anything is the way IOCS records
it, all such inquiries typically necessitate a special study in order to glean any insights.
III-6
Antoinette Crowder weighed in as she had in the last rate case, seeking to deny the existence of
the entire problem.9 SMC-RT-1, pp. 33-40.
1.
The Existence of Capacity Constraints on Motorized Routes
In Docket No. R2005-1, witness Lewis (USPS-RT-2) took the position that capacity
constraints could never occur on motorized routes. When asked on cross-examination, he
testified that there were virtually no constraints operational on motorized routes: “It’s a
housekeeping thing, that’s all.” Docket No. R2005-1, Tr. 11/5997, ll. 21-25. If capacity
constraints ever occurred, witness Lewis argued, their existence was limited to the minority of
delivery points on the 46 percent of walking portions of routes (i.e., those that are subject to
the third-bundle contractual limitation). See, e.g., USPS-RT-2, p. 7, ll. 21-23.
In this docket, witness Coombs (USPS-T-44) helped expand and correct the record,
explaining the real constraints imposed by the basic design of the standard delivery vehicle,
known as the Long Life Vehicle (“LLV”). She described the physical limitations of the LLV.
Namely, it has room for only three “trays,” on the shelf that is within arm’s reach of the
9
Advo/SMC witness Crowder appears not to understand what a capacity
constraint is. Dr. Haldi used witness Coombs’ definition of a conflict as occurring whenever
more than one saturation mailing must be delivered on the same day, thereby forcing the Postal
Service to incur costs that would not arise if it had only one saturation mailing for delivery.
See VP-T-2, p. 66, ll. 6-9. Witness Crowder creatively redefines conflicts to the point where
she can find virtually none. For example, whenever two saturation flat mailings must be
delivered on the same day, she appears to find that no conflict exists if they can be collated,
thereby closing her eyes to the fact that collation forces the Postal Service to incur an
additional cost that substantially exceeds the near-zero cost of handling only one mailing.
SMC-RT-1, p. 36, ll. 9-11. The important point, which she prefers to ignore, is that
whenever a capacity constraint exists and is reached, at that point marginal cost increases,
perhaps substantially. Pretending that the marginal cost of additional mail is equal to the
marginal cost of existing mail does not make the problem go away, and it can result in wrong
price signals.
III-7
carrier for holding mail being delivered.10 Tr. 13/3751, ll. 13-15 and Tr. 13/3753, ll. 13-14.
Importantly, safety considerations limit the placement of additional trays of mail within reach
of the driver in the delivery area of the LLV. Tr. 13/3761, l. 1 – 3762, l. 2. Since one tray
always is used for DPS’d letters, and the second tray always is used for flats and letters sorted
in a carrier’s vertical flats case, only the third tray is available for extra bundles. The
preceding constitutes new information, and it means that the problem of capacity constraints
could occur on any and all routes — walking or motorized. This is major change in the record
from the last docket.
2.
The Effect of Uneven Distribution of Saturation Mail
Another important factor explicitly acknowledged by the Postal Service in this docket is
that “ECR saturation mailings are not uniformly distributed over all Zip codes.” It stands to
reason that high-volume, high-income areas receive a much higher number of saturation
mailings than do low-volume, low-income areas. Response to VP/USPS-T44-34(c)-(d), Tr.
13/3731. Conflicts are thus likely to occur often in certain areas, and virtually never in other
areas.
3.
The Effect of Permanently Bumping Saturation Letters to DPS Processing
Another important consideration is the degree to which the Postal Service has increased
DPSing of letter mail, particularly including saturation letter mail. Along these lines, in this
docket, the Postal Service proposes to eliminate the DDU entry discount for ECR letters (but
10
According to witness Coombs, the third shelf can hold a bundle consisting of
two flat mailings that have been collated in the office. Also, carriers are said to sometimes
place one extra bundle behind another extra bundle on the same shelf, and work from the two
bundles simultaneously.
III-8
not for flats). Dr. Haldi explained that the effects of this change are twofold: (1) it will
eliminate nearly any conflict that might arise from having a mailing of saturation flats and a
mailing of saturation letters to be delivered on the same day; and (2) it will result in saturation
letters consistently receiving higher handling costs, since the costs of two-pass DPSing can
never get as low as the zero in-office cost of extra-bundle treatment. This result will follow
even on those delivery routes without any conflict on account of multiple saturation mailings.
In a sense, the Postal Service has decided to bump ECR saturation letters permanently to DPS
equipment, so that it can reserve available extra-bundle capacity for the most expensive-tohandle saturation flats. Therefore, the 10 percent conflict estimate by witness Lewis in Docket
No. R2005-1 is now seen as quite irrelevant. Docket No. R2005-1, USPS-RT-2, p. 8, l. 5,
Tr. 11/5945. Postal operations increasingly treat low-cost saturation letters in a way that
causes them to incur greater mail processing and in-office costs than saturation flats, which
continue to be taken to the street as third bundles.
Additionally, when taken to the street as “extra bundles,” saturation mail is classified
as “sequenced mail,” and is one of several cost pools subjected to econometric analysis by
witness Michael D. Bradley (USPS-T-14) in his study of street time variability for the purpose
of distributing volume variable costs to each class and subclass. Tr. 13/3840, ll. 15-22. Using
witness Bradley’s results, for Standard Mail letters and flats Postal Service witness John P.
Kelley (USPS-T-30) developed the following street costs below the subclass level, which he
asserts to be “reasonable” (Tr. 13/3853, l. 19 – 3854, l. 12):
Regular letters: 1.81 cents
Sequenced letters: 1.22 cents
III-9
Notably, witness Kelly’s street costs for sequenced letters were 1.22 cents, and for regular
letters (virtually all of which are DPS’d) were 1.81 cents, fully 48 percent higher.
Consequently, it is seen that denying “extra bundle” treatment to saturation letters substantially
increases on-street carrier costs, as well as mail processing and in-office costs.11
4.
Conclusion
The problem of capacity constraints is now understood to exist on both walking and
motorized routes, is far more likely to occur on routes where saturation mail is concentrated,
and occurs wherever the Postal Service decides to DPS saturation letters and while taking
saturation flats directly to the street as extra bundles. In view of this new information, Dr.
Haldi again points out that when persistent capacity constraints are present, the Postal
Service’s costing systems generate the perverse result of identifying higher costs for the Postal
Service’s single product which is the most flexible and can be handled at the lowest cost, while
indicating that one of the least flexible and most difficult flat product to handle mail has a
lower cost.
The fundamental problem is that the Postal Service’s costing system does not — and
cannot — reflect costs affected by capacity constraints. The reason is easy to understand.
Once a capacity constraint is reached, another operating procedure then must be used to handle
some of the mail; i.e., costs then are incurred in one cost pool because of the situation in
11
Also, one has to ask, in the strongest possible way, how an analysis could yield
the result that the street costs, per letter, of DPS’d letters in trays could possibly be higher than
the cost of a sequenced letter, when the sequenced letter is handled individually, involves
reaching into a different tray to retrieve it, and must be merged with the other letters and flats
for the stop. Even a rudimentary understanding of carrier operations would seem to suggest
that something is seriously awry.
III-10
another cost pool. This in turn means that the operating procedures, or cost pools, have
become interdependent. This interdependence violates the separability assumption,
discussed previously, which requires that cost pools be independent of each other, and which
underlies the entire costing methodology.
When capacity is not constrained, the marginal cost of an additional saturation mailing
is less than and not the same as the marginal cost of a saturation mailing when capacity is
constrained.12 Since the Postal Service’s costing systems are incapable of estimating the
higher marginal costs of saturation mailings facing capacity constraints (and since the Postal
Service itself seems unwilling, or unable, to recognize these marginal costs in its own cost
estimates), unless the Commission recognizes them, rates based on this inequity will be
perpetuated. Accordingly, the Commission should take special note in this docket to guard
against any undue elevation in the rates for saturation letters based on over-stated saturation
letter costs and under-stated saturation flat costs.
D.
Witness Mitchell
In this docket, witness Mitchell submitted testimony (VP-T-3) related to the topic
discussed above. It introduces what he calls a “fairness in costing” issue, which arises when
capacity limitations cause one category of mail to be given priority over another category.
12
An effective capacity constraint — i.e., one that is binding — thus “means that
the marginal cost is most definitely not equal to the average volume variable cost.” VP-T-2, p.
68, ll. 3-4.
III-11
The situation is that the number of extra bundles that carriers can handle efficiently is
limited. If in such situations a choice has to be made whether to take out a bundle of flats or a
bundle of letters, but not both, it is in all likelihood the letters that will be handled in an
alternative way, probably at higher cost than extra-bundle treatment, given the way the Postal
Service costing systems develop costs. In other words: (1) if the flats are taken out as an extra
bundle, flats costs will be low, which would normally lead to flat rates that are low; and (2) if
the letters are taken out instead, letters costs will be low, which would normally lead to letter
rates that are low. Why would it be fair to letter mailers for flats to be taken out as the extra
bundle, or, why would it be fair to flat mailers for letters to be taken out as the extra bundle?
Because this problem will become more acute when the Flats Sequencing System
(“FSS”) system is fully deployed and carriers working from three bundles (DPS’d letters;
FSS’d flats; and residual flats and letters sorted in a vertical flats case) will be the norm,
witness Mitchell suggests that, after this docket is complete, the Commission inquire into this
important question of fairness in costing. VP-T-3, p. 12, ll. 7-10. This is a sound
recommendation which should be addressed at an appropriate time.13
13
Advo/SMC witness Crowder (SMC-RT-1) purports to rebut witness Mitchell,
but all she really does is formulate and then dismiss a strawman of her own. Witness Crowder
implies that witness Mitchell argues that a decision to give priority to flats as the extra bundle,
which would give lower costs (and likely rates) to flats and higher costs (and likely rates) to
letters, creates or implies a cross-subsidy from letters to flats, of the kind that economists are
prone to abhor. See SMC-RT-1, pp. 41-44, Tr. 35/11759–62. Short of that, however, all
witness Mitchell does is suggest that a hypothetical mailer of letters, who is paying higher rates
because of the priority given to flats, might feel that he is cross subsidizing flats. Witness
Crowder’s rebuttal testimony offers no information or analysis of any value to the Commission
on this point.
IV-1
IV.
WITNESS MITCHELL’S TRADITIONAL APPROACH TO STANDARD MAIL
RATE DESIGN SHOULD BE UTILIZED AND HIS PROPOSED RATES FOR
STANDARD MAIL SHOULD BE RECOMMENDED.
A.
Witness Mitchell’s Standard Mail Rate Design Is Superior to that of Witness
Kiefer.
In the past, the Commission typically has had only one complete presentation of
Standard Mail rates to consider — that provided by the Postal Service. In this docket,
however, the Commission has before it two alternative rate designs that have been fully tested
on the record — those of Postal Service witness Kiefer (USPS-T-36) and Valpak witness
Mitchell (VP-T-1).1 For reasons set out below, the traditional approach to Standard Mail rate
design, as employed by witness Mitchell, should be used again in this docket, as should the
rates he developed by use of that traditional approach.
The Postal Service rate design proposal for Standard Mail is developed and presented in
direct testimony by witness Kiefer (USPS-T-36), who also presented rebuttal testimony (USPSRT-11). Although witness Kiefer’s workpapers (contained in library reference USPS-LR-L36) show costs for all categories — and therefore passthroughs exist, at least implicitly, for all
categories — his unprecedented approach to rate design does not begin with costs and then
decide on an articulated basis the extent to which rates should reflect those costs. He also
abandons use of a presort tree which displays costs and rate relationships, and which has been
used in Standard Mail rate design since Docket No. R90-1. Instead, he asserts that the
1
At the very end of the case, Advo/SMC witness Antoinette Crowder (SMC-RT1) provided workpapers as part of her rebuttal testimony which purport to present an
alternative rate design, but did not provide a complete justification of her rates, except for
certain features. See n. 7 infra, for further discussion.
IV-2
established approach used in all recent omnibus rates cases that begins with costs is too
formulaic, and he claims to have developed an alternative method that, he says, helps him
“achieve” a Postal Service goal of “having more finely disaggregated and flexible rate
structures.”2 USPS-T-36, p. 12, ll. 25-26.
Witness Kiefer describes his new method as “iterating” among alternative rate levels
and rate relationships, using passthroughs as “levers,” until he finds new rates that result in the
desired overall revenue and that he views as “reasonable.” See Response to NAA/USPS-T369, Tr. 5/915, and Tr. 5/1132, l. 13 – 1134, l. 7. His basic rate design criteria appear to be
subjective assessments of various percentage rate increases, rather than a focus on changes in
costs and the deliberate selection of passthroughs. He develops rates for the nonprofit
subclasses in the same manner, with the added constraint that rates must accord with the 60percent rule.3 39 U.S.C. § 3626(a)(6).
2
The meaning of a rate structure being “flexible” is never made clear. It is
obvious that it has nothing to do with flexing, as in a fiberglass fishing pole, and it is difficult
to see how one rate being 80 percent of another rate is more (or less) flexible than being 60
percent of the other rate. Flexibility seems to be a euphemism for unfettered Postal Service
discretion to set rates without respect to costs and unconstrained by Commission-established
principles. A “flexible” structure makes it easier for rate relationships to oscillate over time,
according to changing, subjective Postal Service preferences. If flexibility removes the
discipline of acknowledging costs and deciding on transparent bases what the relationship of
costs to rates should be, it would not appear to be either fair or advisable.
3
As witness Mitchell explained, Nonprofit rates can be calculated directly,
although witness Kiefer may not have taken the time to examine the formulas involved. VP-T1, p. 114, n. 43.
IV-3
The result of witness Kiefer’s alterations is not something approaching an across-theboard rate increase. Indeed, in Commercial Regular, his rates range from a 19.2 percent
decrease (for Customized MarketMail (“CMM”)) to a 215.6 percent increase (for mixed
ADC/BMC Not Flat-Machinable (“NFM”) parcels entered at the DBMC). See sheet 27 in
USPS-LR-L-36, WP-STDREG.xls. The range of rate changes for Nonprofit Regular is
similar. The range within both ECR and Nonprofit ECR is narrower. It is clear that a large
rate increase or rate decline is considered acceptable to the Postal Service when it is in line
with Postal Service preferences, but not at any other time. As a result, occasions exist where a
misalignment between rates and costs is apparently acceptable. In most cases, the Postal
Service’s large percentage increases and declines are applicable to categories with small
volumes. In terms of criterion 1 (fairness), no explanation is provided for the Postal Service’s
apparent guiding principle that it is fair to give a mailer a large increase when the overall
volume of the mailer’s category is small, but it is not fair to impose significant price changes
on rate categories with large volumes. See VP-T-1, p. 123, n. 49. This aspect of the Postal
Service’s pricing strategy would appear to reflect “political” considerations, not statutory
criteria, which permit no such distinctions.
Witness Mitchell explains why witness Kiefer’s new subjective, iterative rate design
method should be rejected and why rates should be cost based, and developed in a transparent
manner, preferably in a presort tree.4 See VP-T-1, pp. 103-109. Witness Mitchell then
4
The Commission has expressed its belief in the importance of developing rates
in a transparent manner and the utility of presort trees for maintaining that transparency. See
Notice of Inquiry (“NOI”) No. 2, p. 2. In its comments in response to NOI Nos. 2 and 3 filed
on August 17, 2006, Valpak discussed the importance of transparency (p. 9) and the
IV-4
develops a complete schedule of rates for all four Standard Mail subclasses, honoring costs and
accepted ratesetting principles. See VP-T-1, pp. 98-197. In his workpapers, VP-LR-1,
witness Mitchell presents a complete, user-friendly spreadsheet that allows the user to test
alternative costs, passthroughs, etc., simply by changing a few inputs.5 See VP-T-1, p. 99, ll.
2-6. (Such workpapers differ markedly from those of witness Kiefer. See library reference
USPS-LR-L-36.) Unlike witness Kiefer, witness Mitchell discusses each cost, each
passthrough, and each important rate relationship, along with the history of these and previous
Commission opinions on each. Unlike witness Kiefer, he provides full and transparent reasons
for his recommendations. Aside from certain broader issues, discussed below, no party sought
to rebut specific aspects of witness Mitchell’s rate design on the record.
One key difference between witness Mitchell’s rates and those proposed by the Postal
Service, particularly in the Regular category, is the treatment of letters and flats. The Postal
Service proposes a letter-flat passthrough in the neighborhood of 58 percent. Consequently,
the per-piece contribution for flats under Postal Service rates is much lower than the per-piece
contribution of letters. Witness Mitchell explains why this outcome should not be accepted,
and proposes a corresponding passthrough of 95 percent. See VP-T-1, pp. 114-23, 156-61,
and 178-86. On rebuttal, witness Kiefer says: “Indeed, I am sympathetic to the view ... that
Standard Mail flats should bear a greater share of the Standard Mail institutional cost burdens”
helpfulness of continued use of presort trees (pp. 10-12).
5
Witness Kiefer actually appears to criticize witness Mitchell’s user-friendly
spreadsheets because they produce alternate rates by “only changing a few numbers” rather
than by some elaborate process. USPS-RT-11, p. 24, ll. 1-3, Tr. 33/11135(b).
IV-5
and believes that “relative letter-flat prices should adjust gradually.” USPS-RT-11, p. 20, ll.
21-23 and p. 21, ll. 4-5. Valpak believes that a substantial step should be taken in this case,
along the lines proposed by witness Mitchell. As noted above, the Postal Service clearly is
willing to propose large increases in other areas of Standard Mail, and it should be willing to
do the same here.
Another key difference associated with witness Mitchell’s rate design is his treatment of
rates within the Nonprofit subclasses. See VP-T-1, pp. 110-14, 161-62, and 188. Under the
Postal Service proposal for automation pieces at the mixed AADC and mixed ADC levels, for
example, flats have a negative contribution of 9.573 cents, while letters have a positive
contribution of 6.497 cents. See Response to USPS/VP-T1-22, Tr. 24/8926-27. Witness
Mitchell proposes the same discounts and rate differences for Nonprofit pieces as for
corresponding Commercial pieces, which avoids discrimination against the Nonprofits,
consistent with observations of the U.S. Court of Appeals for the D.C. Circuit in National
Easter Seal Society v. United States Postal Service, 656 F.2d 754 (D.C. Cir. 1981). See VPT-1, pp. 111-12.
Witness Mitchell opposed one classification change proposed by the Postal Service —
to eliminate automation basic ECR letters from the ECR subclass. See VP-T-1, pp. 123-27.
The fact is that these letters are in accord with the definitions of ECR and the logic that led to
its creation by the Commission in Docket No. MC95-1, and they qualify to be in the ECR
subclass just as much as other letters and flats in the subclass. On rebuttal, Postal Service
operations witness Marc D. McCrery testified concerning the handling of basic automation
letters, and argued for their elimination. See USPS-RT-14, pp. 3-6. His testimony misses the
IV-6
point. The presence of these letters in ECR was justified in Docket No. MC95-1 and is
justified now on economic grounds, not only on specific operational grounds.6 The Postal
Service did not justify this major classification change based on economic evidence. Witness
Mitchell has based the rates in his proposal on current Postal Service costs associated with
handling this mail as presented in accordance with Postal Service operational requirements. At
this point, there is certainly no basis for eliminating wholesale this established category of what
is bona fide ECR mail.
A related issue is whether rates for ECR non-automation basic letters should be
elevated artificially in order to keep them higher than the rates for Regular 5-digit automation
letters. Witness Mitchell convincingly explains why ECR rates should not be artificially
elevated. See VP-T-1, p. 120, l. 7 – p. 122, l. 7. The rates for these letters also should be
based on their costs, consistent with the recognition of the letter-flat difference, and the Postal
Service proposal should be rejected.
B.
Witness Crowder’s Proposal to Burden ECR Letters with Unrecouped DAL
Costs, Which Are for the Exclusive Benefit of Saturation Flats, Must Be
Rejected.
Advo/SMC witness Crowder, testifying on rebuttal (SMC-RT-1), develops, at the
coverage proposed by the Postal Service, what appears to be a set of alternative rates based on
6
One of witness McCrery’s arguments concerning ECR basic letters is that trays
might be mistagged, resulting in higher costs. USPS-RT-14, p. 5, ll. 2-5. This observation,
however, merely states the obvious: since Postal Service regulations specify tagging protocols
that are designed to assist optimum Postal Service processing, then costs can be presumed to
increase if trays are mistagged. However, that kind of cause/effect relationship is certainly not
limited to ECR non-automation basic letters.
IV-7
alternative costs for Commercial ECR (rejecting the costs of the Postal Service),7 and points to
general disagreement with witness Mitchell’s rates (and the Postal Service’s rates).
Unfortunately, Advo chose to file its alternative rates not when intervenors were required to
present their rebuttal to the Postal Service case, as well as their cases-in-chief (September 6,
2006), but two and one-half months later, at the end of the docket (November 20, 2006) in
rebuttal testimony when witness Crowder’s cost adjustments and proposed rates could not be
tested on the record through written discovery or rebuttal testimony. Therefore, witness
Crowder’s testimony has been subject only to oral cross-examination (December 4, 2006)
conducted shortly after the rebuttal testimony was filed, when only limited time was available.8
Witness Crowder’s testimony states that neither Advo, nor any other party, opposes the
Postal Service’s proposed surcharge on DALs. At the same time, her testimony also
acknowledges that in test year the proposed 1.5 cent surcharge for each DAL will recover only
7
Witness Crowder’s rates differ from those proposed by witness Mitchell due
largely to: (1) a different treatment of letters and flats, a matter discussed in another section of
this brief; (2) the selection of a lower pound rate than that proposed by the Postal Service and
adopted by witness Mitchell; and (3) different assumptions relating to the diversion of DAL
flats to on-piece-addressed flats. Witness Mitchell explained why the Postal Service procedure
for handling the DAL issue is reasonable in this case. See Response to NAA/VP-T1-38, Tr.
25/8881. He also explained how to recognize the DAL surcharge in rates, without
disadvantaging saturation flats with on-piece addresses. See VP-T-1, pp. 175-76.
8
Witness Crowder seeks to avoid the charge that she was filing direct testimony
as rebuttal testimony by stating that Advo still supported the Postal Service proposal (SMCRT-1, p. 3, l. 4, Tr. 35/11721), but adds “[i]f the Commission decides it must make changes
to the Postal Service-proposed ECR rates, then it should make them in this direction.” Id., p.
6, ll. 8-9, Tr. 35/11724. Witness Crowder’s alternative rates should have been filed initially,
and came too late for meaningful record evaluation.
IV-8
about one-third of the cost of handling all DALs that remain in the system. If half of all preexisting DALs remain in the system9 and pay the surcharge, those DALs will cause the
Postal Service to incur $67 million of expenses that will not be recouped by the surcharge. Tr.
35/11791, ll. 1-3.
When asked on cross-examination how the rate design in her testimony treated the
unrecouped expense of DALs that accompany saturation flats, she responded, “[t]hat is just
something that all ECR covers. It’s like an overhead.” Tr. 35/11791, ll. 11-12. Because the
DAL surcharge covers only a small fraction of the costs caused by DALs, witness Crowder
went on to explain that “DAL flat mailers are not paying as much of an institutional cost
contribution as non-DAL members [sic].... A DAL flat ... is not making as much contribution
as an on-piece address flat, and the difference in that contribution is made up by everybody
else. That’s what I’m trying to explain.” 10 Tr. 35/11791, ll. 20-22 and Tr. 35/11793, ll. 1-4;
emphasis added. To avoid any misunderstanding, she subsequently said:
Effectively what I am recommending is that the saturation DAL
flat rate pay a slightly less per piece contribution than on-piece
address saturation flat mailers. The difference in that
contribution would have to be made up by everybody else.
Everybody else. All ECR. [Tr. 35/11794, ll. 6-12 (emphasis
added).]
The essence of witness Crowder’s proposal is that saturation flat mailers who opt to
continue using DALs should have the extra costs which they impose on the system shared by
9
Half of all existing DALs is conservatively estimated to be around 2.2 billion.
The Postal Service has yet to submit any reliable data on the volume of DALs.
10
Having different pieces within the same rate category pay different contributions
is at complete variance with witness Crowder’s total embrace of ECP.
IV-9
all other ECR mailers, including ECR letters.11 Valpak respectfully disagrees. DALs have
never benefitted letter mailers. In fact, as explained in Section III, the preferential treatment
given saturation flats already causes the costs that are recorded for saturation letters to be
higher than they would be if letters and flats were treated in the same manner. At the very
least, all costs of DALs should be fully reflected in the letter flat differential, and any
unrecouped DAL expense perhaps should be shared only by all saturation flat mailers. There
really is no reason why the costs of DALs, which are exclusively for the benefit of saturation
flats, should be shared by any other rate category.
C.
Witness Mitchell’s Rebuttal Testimony Demonstrates Why the Commission
Should Reject Pitney Bowes Witness Panzar’s Application of ECP to the
Rate Design for Standard Letters and Flats.
In his rebuttal testimony VP-RT-1, witness Mitchell explains why the Commission
should reject Pitney Bowes Inc. (“Pitney Bowes”) witness John C. Panzar’s (PB-T-1)
suggestion that Efficient Component Pricing (“ECP”) principles should be applied in the
development of rate differentials between letters and flats. Valpak considers this to be an
important issue.
Witness Panzar’s reasoning is that if categories (parcels, in the application he
addresses) have different costs and different elasticities, they should probably be in different
subclasses, where notions of economic efficiency would be applied in developing their rates.
Witness Mitchell explains that if the elasticities are the same, or are not known to be different,
or if a decision is made not to treat them as though they were different, and the categories are
11
This differs from the Postal Service’s prior procedure of counting DALs in the
carrier costing systems as letters and charging an unknown amount of DAL costs to letters.
IV-10
housed in the same subclass, there is no reason to give zero weight to the importance of
economic efficiency, as it also can be recognized within subclasses. To make clear the
importance and effect of instead applying ECP, witness Mitchell explains how the simple
decision to house them in the same subclass and adopt ECP can cause a large step decrease
in the rate difference between letters and flats, and that such a discontinuity seems strange and
difficult to justify. VP-RT-1, p. 5, l. 18 – p. 6, l. 8.
As the Commission has done in the past,12 letters and flats should continue to be
distinguished for rate purposes. Witness Mitchell explains that they are obviously different on
their face, that they have different processing streams, different costs, and are used by different
mailers.13 VP-T-1, pp. 114-119. Even Advo witness Crowder, who seeks to blur the
distinction between letters and flats and who introduces varying other disaggregations in ECR
that might warrant special treatment, discusses letter mailers competing with flats mailers
(SMC-RT-1, p. 7, l. 4, Tr. 35/11725), and the possibility of some mailers making multiple
folds in thin flats to make them into letters (Tr. 35/11817, ll. 9-14). This may mean that she
believes some advertisers at the margin might move from one type of product to the other, but
even she nowhere suggests that flats mailers (such as Advo) are on the verge of changing to a
letter format, and she never suggests that letter mailers are on the verge of changing to flat
12
13
See, e.g., Docket No. MC95-1, Op. & Rec. Dec., ¶5592.
Witness Crowder misses much of the essence of letters and flats as postal
products when she mischaracterizes witness Mitchell’s argument that letters and flats are
separate products as being based on “cost and demand differences.” SMC-RT-1, p. 20, l. 8,
Tr. 35/11738.
IV-11
format (although this perverse response to postal pricing could be the result if letter-flat rates
are homogenized, as she would seem to prefer).
Notions of economic efficiency are general and should be used as a reference point for
all rate considerations — indeed, the distance of rates from costs and the associated losses in
consumer surplus are important and clear. This does not mean that other factors are wholly
unimportant. It does mean, however, that one needs a good reason (or reasons) to withdraw
from economic efficiency and focus exclusive attention on ECP. The primary justification
usually given relates to worksharing and the importance of technical efficiency, i.e., getting the
lowest-cost entity to do the work. This was pointed out strongly by Newspaper Association of
America (“NAA”) witness J. Gregory Sidak (NAA-RT-1).14 In addition, witness Mitchell, in
his response to ADVO/VP-T1-10 (Tr. 25/8824-835, especially 8832-34), points out a range of
other factors that might be associated with attaching guiding importance to ECP.
It is true that ECP is closely associated with technical efficiency and with the banner of
achieving lowest combined cost. Both of these sound good, and arguments against them per se
are unusual. But the issues need to be understood clearly.
14
In response to a pointed interrogatory on the application of ECP, NAA witness
Sidak demurred from disagreeing with parts of Pitney Bowes witness Panzar’s statement that
ECP should be applied to all cost differences by qualifying his statement with an assumption
that the value must be equal, an assumption with unclear meaning at this point, and by
declining to answer the second part of the question, which was whether worksharing arguments
relating to lowest combined costs apply to all cost differences. Response to ADVO/NAA-T14(b), Tr. 26/9132. Witness Crowder took this response to mean that witness Sidak agrees
broadly with witness Panzar. See SMC-RT-1, p. 6, n. 1, Tr. 35/11724. One should be very
careful using a limited and careful response in this way, especially when it is clearly
inconsistent with a series of other statements of the witness.
IV-12
The meaning of technical efficiency (and lowest combined cost) is clear when the task
at hand is understood and invariant. For example, it makes little difference regarding the
nature of the overall task whether the Postal Service, or a mailer, sorts the mail or transports
the mail. In either case, the same final result is accomplished. The mail is sorted, the mail is
transported, and the mail is delivered. The nature of neither the mail nor the service provided
changes. It is easy, then, to talk about whether the cost of doing the task one way is different
from the cost of doing the task another way, i.e., it is technically more efficient to accomplish
delivery of a given letter for 8 cents than it is to accomplish delivery of the same letter for 10
cents. But when the nature of the task changes, as in switching from a letter to a flat,
arguments relating to technical efficiency become unclear and indeterminate. For example,
how can one compare the technical efficiency of delivering a letter for 10 cents with the
technical efficiency of delivering a flat for 13 cents? It cannot be done. There is no basis for
saying that one is more efficient than the other. Choosing a letter instead of a flat is not a
matter of technical efficiency or of lowest combined cost.
When the issue is simply one of who does the work, it is easy to argue that, say, an
increase in a discount causes some quantity of work to shift to a lower-cost provider, and this
is clearly good. But when the issue is the rate difference between a letter and a flat, an
increase in the difference results in an increase in the quantity of letters and a reduction in the
quantity of flats. This means that more value is being received from the consumption of letter
services and less value is being received from the consumption of flat services. The
desirability of such a change cannot be measured by looking at the change in the rate and the
change in the cost to the Postal Service. Different product mixes are involved.
IV-13
Changes in relative rates and associated changes in relative volumes occur routinely in
markets. The firms involved produce different quantities of the products, incur different costs,
obtain different revenues, realize different profits, and evaluate their market position. There is
a strong basis for looking at the economic efficiency of the resulting outcome, but there is no
basis for looking at cost differences and revenue differences when prices are changed, and
reaching negative conclusions if the cost difference and the rate difference are not the same.
The conclusion is that ECP applies in a meaningful way to worksharing, but it does not
apply to letters and flats. Witness Sidak agrees with this. Response to ADVO/NAA-T1-4(b),
Tr. 26/9132. Witness Mitchell has explained it further in a postal context. In fact, although
witness Crowder advocates applying ECP to letters and flats to achieve better rates for flats,
she accurately describes ECP as involving primarily issues of technical efficiency and the
minimization of “end-to-end distribution costs.”15 SMC-RT-1, p. 22, l. 27, Tr. 35/11740.
15
Witness Crowder also notes that gains in dynamic efficiency can occur under
ECP. This generally refers to competition resulting in something being done at a lower cost.
However, the notion of dynamic efficiency loses much, if not all, of its meaning when the
issue involves changing from letters to flats.
IV-14
CONCLUSION
The cost coverages for Standard Mail proposed by the Postal Service are profoundly
out of balance, and should be adjusted by the Commission in accordance with witness
Mitchell’s proposals, substantially reducing the coverage for ECR to below the coverage for
Regular. The Commission should reject the Postal Service’s latest effort to treat substantial
mail processing costs as institutional costs. Postal Service costing systems do not accurately
reflect the relative costs of ECR saturation letters and flats, and that fact needs to be taken into
consideration when setting rates. General problems associated with Fairness in Costing needs
to be addressed in a subsequent docket. The Commission should reject the novel “iterative”
approach of Postal Service witness Kiefer, and recommend rates for Standard Mail set in
accordance with the traditional approach, as reflected in Witness Mitchell’s proposed rates in
VP-T-1.
Respectfully submitted,
_______________________
William J. Olson
John S. Miles
Jeremiah L. Morgan
WILLIAM J. OLSON, P.C.
8180 Greensboro Drive, Suite 1070
McLean, Virginia 22102-3860
(703) 356-5070
Counsel for:
Valpak Direct Marketing Systems, Inc., and
Valpak Dealers’ Association, Inc.
December 21, 2006