Before the

Postal Regulatory Commission

Modification of Analytic Principles UsedDocket No. RM2010-1

In Periodic Reporting (Proposal Twenty)

COMMENTS OF THE PUBLIC REPRESENTATIVE

IN RESPONSE TO ORDER NO. 311

(October 28, 2009)

I. INTRODUCTION

On October 6, 2009, the Postal Service (“USPS”) filed a Petition for a Proposed Methodology change in Analytical Principles with the Postal Regulatory Commission (“Commission” or “PRC”). A Commission Public Notice set the Postal Service Petition for public comment by October 28, 2009, with replies due upon determination of their necessity by the Commission.

Proposal Twenty updates the currently used density factors resulting from a special study undertaken in year 2001.[1] These density factors are used to calculate cubic feet occupied by various mail products in order to distribute among them the volume variable vehicle service driver costs (Cost Segment 8) and purchased transportation costs (Cost Segment 14).

The Postal Service purports that while performing the 2009 density special study,it has primarily adhered to themethodology used in the previous 2001 study[2]albeit with a few exceptionsdetailed below -

  1. Full and nearly full containers were selected for sampling in the 2009 special study. In the prior study, the data collectors manually added to partially filled trays before sampling.
  1. Recording piece density,for parcel shaped mail only, instead of load density used in the prior study. All other mail continues to be sampled using load density.
  1. The 2009 special study provides densities for products not separately stated in the previous study including Standard and First Class Parcel densities and eliminates obsolete products.

II. Discussion

Density factors play a vital role in the development of the distribution keys used to attribute certain volume variable costs to products. It is important that the Postal Service updates these factors on a regular basis to ensure that each product is being allocated its fair share of these costs. Therefore the 2009 transportation density special study is a welcome update to the prior study which included densities for eliminated products and no densities for the new products.

In its petition, the Postal Service maintains that the results from the 2009 update produced “relatively minor changes in the density factors”.Upon comparing FY08 original costs for Cost Segment 8 (Vehicle Service Drivers) and Cost Segment 14 (Purchased Transportation) with the simulated costs using the new density factors[3] calculated in the 2009 study, there is no change to cost per piece for total First Class Mail, Standard Mail or Periodicals in the Market Dominant Products group.

However, the cost per piece for Package Services changed significantly resulting in a decrease of 2.4 cents per piece of Cost Segment 8 and Cost Segment 14 costs attributed to the class as a whole.It appears that the higher volume in other subclassesmay have mitigated the net changes arising from using the new density factors, resulting in no change to the cost per piece for the subclass.

Changes to cost per piece were most noticeable primarily for parcel shaped mail for the Market Dominant Products as shown in the table[4] below –

FY2008 CS14/CS8 Original Cost vs. Simulation (new density factors) / 1.601
LINE NO. / CLASS, SUBCLASS, OR SPECIAL SERVICE / CRA CLASS / Net Changes / FY08 Volume / Change in Cost per Piece
UNITS / $(000) / (000) / $
1 / MARKET DOMINANT PRODUCTS
2 / FIRST-CLASS MAIL:
7 / SINGLE PIECE FLATS / 16 / -1,619 / 2,607,157 / $ (0.001)
8 / PRESORT FLATS / 17 / -848 / 772,584 / $ (0.001)
9 / SINGLE PIECE PARCELS / 19 / 18,212 / 595,014 / $ 0.031
10 / PRESORT PARCELS / 20 / 445 / 10,507 / $ 0.042
12 / STANDARD MAIL
18 / NOT FLAT-MACHINABLE AND PARCELS / 27 / 13,590 / 733,729 / $ 0.019
24 / PACKAGE SERVICES
25 / SINGLE PIECE PARCEL POST / 41 / -1,970 / 89,536 / $ (0.022)
26 / BOUND PRINTED MATTER FLATS / 42 / -5,627 / 289,623 / $ (0.019)
27 / BOUND PRINTED MATTER PARCELS / 43 / -6,699 / 308,561 / $ (0.022)
28 / MEDIA AND LIBRARY MAIL / 44 / -5,747 / 158,505 / $ (0.036)
29 / TOTAL PACKAGE SERVICES / -20,043 / 846,225 / $ (0.024)
30 / USPS MAIL / 125 / -6,209 / 823,685 / $ (0.008)
31 / FREE MAIL -BLIND HANDICAPPED & SERVICEMEN / 130 / -1,133 / 71,975 / $ (0.016)

The change in the measurement of parcel shaped mail from load density to piece densitycan be assumed,at least partially, to have caused the reduction in the cost per piece of attributed transportation costs. Due to the change in the methodology of calculating density for parcel shaped mail, weight per sampled piece,and not weight per sampled container as in the prior study, is afactor in its measurement. This places great importance on the sample size for parcel shaped pieces.

For example, if the weight of parcel shaped mail decreases due to the recession, along with a similar weight reduction among the other subclasses, an expected outcome of economic downtimes, and the sample pieces selected for measurement of parcel shaped mail did not capture the overall weight reduction,all non-parcel shaped mail measured by load density would be attributed a higher than their share of transportation costs when compared to transportation costs attributed to parcel shaped mail.

Therefore the sampling size becomes an important factor in piece density measurements. The coefficient of variation (CV) is a statistical measure, useful in calculating losses in sampling reliability. The CV for a single variable describes the dispersion of the variable in a way that does not depend on the variable's measurement unit. The rationale is that the higher the CV, the greater the dispersion in the variable.

It is not known and should be explored further, if the higher coefficient of variation in the 2009 Transportation density study when compared to its predecessor 2001 density study, in parcel shaped mail in the Market Dominant products group is caused by the change in methodology in the sampling of parcel shaped mail, a population that would include packages.

Excerpt from Table 2.1[5]
2009 Density Study / 2001 Density Study
Est. 95% Confidence Interval / Est. 95% Confidence Interval
Mail Category / Est. Coefficient of Variation / Mail Category / Est. Coefficient of Variation
Market Dominant
Package Services
PSVC Bound Printed Matter / 10.00% / PKG Bound Printed Matter / 2.88%
PSVC Media and Library Mail / 7.60% / PKG Media and Library Mail / 3.00%
PSVC Parcel Post / 5.10% / PKG Parcel Post Machin / 2.53%
PSVC Parcel Post NMO / 11.70% / PKG Parcel Post Non-Machin / 3.16%

Upon comparing the coefficient of variations provided in Table 2.1 Density Estimates by Mail Category from the 2001 study to the 2009 study, it appears that there were other products other than package services that had a high coefficient of variation but there was no method of determining if the same product had a similarly high CV in the previous study due to the difficulty in matching the breakdown of the categoriesin the two studies.

The last ACD for FY 2008 highlighted that package services had a volume of .42 percent[6] of total mail volume and some products in Package Services did not cover their costs. Notable among them are Media and Library Mail with an 87.9% cost coverage and Single-Piece Parcel Post with a 91.8% cost coverage[7]. The overall cost coverage for package services was 101.0% per the FY 2008 ACD[8].

Transportation costs are the third largest cost pool after mail processing and delivery. The attribution of less transportation costs to a product with a slim mark up would be beneficial in aiding its compliance with 39 U.S.C. §3622(c)(2), which requires each class to cover its attributable direct and indirect costs. Therefore, it is crucial that the sampling size and methodology used to calculate the density factors in the 2009 density study and the net cubic volume obtained from TRACS provide reliable estimates of cubic footage to ensure that each class is attributed only its share of volume variable transportation costs.

There has been anoverall reduction in sampling sites, sampling sizes and days of sampling used in the 2009 density study when compared to its predecessor study. Data was collected for six weekdays at each facility in the 2001 study[9], Monday through Saturday compared to two consecutive days in the 2009 study[10]. The reduction in days sampled may not have captured the heavier mail related to advertising on some days of the week.

There has also been a decrease in the number of sites selected for sampling,fourteen[11]randomly selected sites used in the 2009 study compared to forty eight sites[12]used in the 2001 study.

An additional area of concern is that the data for the density studies is collected over a limited period of time over the summer months. Data was collected between July and August 2009 and April 23, 2001 and May 12, 2001. It would be beneficial to know the weight to volume ratio difference between sampling in the summer months compared to other monthsto understandthe potential seasonal variations.

III. Conclusion

The 2009 Transportation density study is a welcome update and an improvement to its predecessor study with the provision of updated density factors and the elimination of subjectivity in only selecting full or partially full trays for sampling.

The Public Representative supports the proposal but urges the Commission to take into consideration areas of concern in the reduction in the sampling sites, sampling sizes and days of sampling that questions the reliability of the data provided by it.

Respectfully submitted,

Cassie D’Souza

Public Representative

901 New York Avenue NW Suite 200

Washington DC 20268-0001

202-789-6815

[1] Docket No. R2001-1

[2] Docket No. R2001-1, USPS-LR-K-33

[3] Docket No. RM2010-1, FY 2008 CS14/CS8 Original Cost vs. Simulation (new density factors)

[4] Docket No. RM2010-1

[5] Docket No. RM2010-1

[6] Annual Compliance Determination, Fiscal Year 2008, Page 70

[7] Annual Compliance Determination, Fiscal Year 2008, Page 72

[8] Same as footnote 7

[9]Docket No. R2001-1, USPS-LR-K-33

[10] Docket No. RM2010-1, 2009 Transportation Density Study Update

[11] Docket No. RM2010-1, 2009 Transportation Density Study Update, Table 3.1

[12] Docket No. R2001-1, USPS-LR-K-33, Table 3.2