Chapter2.html

[2001] In accordance with established practice in the Commission's rate proceedings, the Postal Service estimates its future revenue needs in this case by identifying a recently concluded fiscal year as a base period, adjusting its reported historical result through intermediate periods, and rolling results foward by incorporating the effects of numerous expense and volume factors to arrive at cost and revenue estimates in a selected test year. This process is described in the testimonies of witnesses Tayman, USPS-T-9, and Kashani, USPS-T-14.

[2002] As noted in the Introduction, the Postal Service used Fiscal Year 1998 as the base year from which to project costs and revenues. USPS-T-9 at 11. The Service selected Fiscal Year 2001 as the test year for ratemaking purposes in this proceeding.¹

[2003] The Commission's determination to use the more recent Fiscal Year 1999 reported results for projecting test year costs requires new bases for arriving at those estimates. Generally, those bases consist of: (1) the "basic update" of the rollforward program sponsored by witness Patelunas; (2) recognition of other changes in estimated test year costs and revenues, many of which were provided in response to Order No. 1294; and (3) miscellaneous adjustments and error corrections to assure accuracy and maintain consistency with other Commission findings in this case.

[2004] Recognizing that the Commission might conclude it should incorporate the updated information produced in response to Order No. 1294 in its analyses and recommendations, the Postal Service argues on brief that those materials "must be treated as a unified package" because they "do constitute a reasonably balanced update, in light of the circumstances." Postal Service Brief at I-17. (emphasis in original) The Commission has done so by considering the Service's suggested revisions on the merits in their entirety. The Commission has altered the Service's new cost and revenue items only where consistency with established methodologies, or the correction of errors, so requires.

[2006] Additional "breakthrough productivity" cost reductions. Additional cost reductions associated with the Postal Service's "breakthrough productivity" initiatives, quantified by witness Patelunas in USPS-ST-44, have been recognized.

[2007] Revised "Other Programs" costs and revenues. Adjustments have been made to reflect changes in "other programs" costs (such as advertising costs) and revenues (such as estimated revenue for e-Business programs in the test year), also documented by witness Patelunas in USPS-ST-44.

[2008] Revised personnel cost level change factors. Costs in appropriate segments have been increased to reflect updated wage rate information, including the Postal Service's substitution of the unreduced Employment Cost Index (ECI) percentage value for ECI minus one percent.

[2009] Recognition of the Field Reserve offset to cost reduction programs. In response to Presiding Officer's Information Request No. 14, witness Patelunas states that the update filed in response to Order No. 1294 should have incorporated a $200 million Field Reserve offset to certain operations cost reductions, but it was inadvertently omitted. Tr. 46D/21593. He further states that, had the Field Reserve been properly incorporated, the amount would have reduced total program savings, and would have applied primarily to savings in mail processing and window service clerks and mailhandlers. Ibid.

[2010] Rather than distributing the Field Reserve amount as proportional offsets to savings associated with the various cost reduction programs, it has been recognized as a special-purpose component of the contingency provision. See subsection D.

[2011] Hybrid billing determinants. In response to a request in Presiding Officer's Information Request No. 16 for revenue estimates by subclass and service that reflect FY 1999 billing determinants in the manner the Service deems appropriate, witness Mayes presented hybrid billing determinants for a year consisting of FY 1999 Quarter 3 through FY 2000 Quarter 2. Tr. 46D/21408-16. Subsequently, she also submitted a corrected and revised version of the hybrid billing determinants. id. at 21416-21. The Commission has used these billing determinants in their corrected form to calculate test year revenues and final adjustments.

[2012] Other updated cost level change factors. The Commission has used other updated indices to calculate test year cost estimates, including more recent CPI values to calculate cost-of-living adjustments and the non-personnel cost level change factors provided in witness Patelunas' Library Reference LR-I-421.

[2013] Revised RPYL amount. The amount of the provision for recovery of prior years' losses has been recalculated and increased to incorporate the estimated amount of the net loss in FY 2000. In the Postal Service's original filing, witness Tayman's computation includes an estimated net income of $66 million in FY 2000. USPS-T-9 at 48, Table 53. The materials filed in response to Order No. 1294 include an estimated loss of $325 million. Exhibit USPS-ST-44E. This change effectively increases the annual recovery amount by $43.6 million.

[2014] The Postal Service is required by the Commission's Rules to accompany any general request for a change in rates with two sets of forecasts of postal volumes. These forecasts must be based upon econometric fits of economic demand functions for postal services by class and subclass. The rule setting out these requirements reads as follows:

(i)An econometric demand study relating postal volumes to their economic and noneconomic determinants including postal rates, discounts and fees, personal income, business conditions, competitive and complementary postal services, competitive and complementary nonpostal activities, population, trend, seasonal patterns and other factors.

(ii)The actual or estimated volume of mail at the prefiled rates for each postal quarter beginning with the first quarter of the most recent complete fiscal year and ending one year beyond the last quarter of the future fiscal year.

(iii)The estimated volume of mail assuming the effectiveness of the suggested rates for each postal quarter beginning with the quarter in which the rates are assumed to become effective and ending one year beyond the last quarter of the future fiscal year.

[2015] To comply with this rule the Postal Service sponsors the testimony of witnesses Tolley, Musgrave and Thress. Witnesses Tolley and Musgrave present two sets of forecasts through the end of the government fiscal year (GFY) 2002:

[2016] "Before-rates" forecasts based on the assumption that Postal Service rates and rules remain unchanged.

[2017] "After-rates" forecasts based on the assumption that all of the rate increases proposed by the Postal Service are implemented as of October 1, 2000.

[2018] The relationship between the "before-rates" and "after-rates" forecasts is principally determined by a set of estimated price elasticities and by the differences in the price information for the two forecasts.

[2019] The forecasts are basically made at the level of mail subclasses and special services. At this level the econometric estimate of price and other elasticities may be applied in a fairly direct manner to yield forecasts of volumes by postal quarters. For First-Class and Standard A mail, volume forecasts by subclass are not sufficient because the mailstreams in these subclasses are composed of mail receiving discounts for different kinds of worksharing, including presorting and prebarcoding for automated processing. Witness Tolley's forecasts for First-Class and Standard A mail further divide these subclasses into various major worksharing categories using share models.

[2020] The forecasts are derived from econometric models fit to time series by USPS witnesses Thress and Musgrave. The models for classes and subclasses, and the manner in which they are applied to produce forecasts, have evolved slowly since similar models and forecasts were presented by witness Tolley in Docket No. R80-1. The share model for worksharing categories first appeared in approximately its present form in testimony for the most recent general reclassification case, Docket No. MC96-1.

[2021] The Commission's Rules have also evolved. Prior to R94-1 the Commission constructed Lotus 1-2-3 worksheets for its own use based upon the testimony of Postal Service witness Tolley and supporting library references. This was an obvious and risky duplication of effort since the Commission's worksheets had to be produced over a very short period of time and on the basis of largely written testimony that was not always perfectly explicit in describing the Service's calculations. Following Docket No. R90-1 the Commission addressed the problem by amending our rules. Beginning with Docket No. R94-1, the Postal Service has been required to supply the Commission with a usable electronic version of its volumes forecasting models and methods. The language of the rules describes in detail the minimum capabilities the Commission expects to find in the Service's submissions:

(j)(6)(iii) Subject to paragraph (a)(2) of this section, there shall be furnished in every formal request a computer implementation of the methodology employed to forecast volumes and revenues for each class and subclass of mail and postal service.

(iv) The computer implementation described in paragraph (j)(6)(iii) of this section shall be able to compute forecasts of volumes and revenues compatible with those referred to in paragraphs (j)(2), (j)(3) and (j)(5) of this section for:

[2022] Since R94-1, the Postal Service has supplied the Commission with Lotus 1-2-3 worksheets that are similar in function to worksheets that the Commission constructed for itself in Docket Nos. R87-1 and R90-1. The worksheets are well-designed and largely self-documented with descriptors and notes that have made it relatively easy to check them for correctness and to modify them to meet the Commission's requirements. Witness Thress is the author of the worksheets.

[2023] Although the Commission has considered from time to time using alternative forecasting methods proposed by OCA and other parties, the established forecasting methodology is the methodology that has been developed over the years by the Postal Service. The worksheets must do more than simply document the testimony of Service volumes witnesses. The Commission actually has to be able to use the worksheets during the course of a rate proceeding to modify the Service's forecasts, particularly the "after-rates" forecasts. The capabilities specified in the rules anticipate the Commission's needs and reflect past experience in making necessary changes to the Postal Service's own forecasts. Overall, the purpose of the Commission's rules is to ensure that the Postal Service's forecasting methodology is always available in a rate proceeding, even if the Commission should decide not to use it.

[2024] The Commission has rarely recommended postal rates and fees that exactly match those proposed by the Service in its filing. Adjusting the Service's volume forecasts to reflect the Commission' s recommended rates is a necessary and predictable aspect of the regulatory process, since the Commission must propose rates that allow the Postal Service to just recover costs plus a reasonable contingency.

[2025] Adjustments to the forecasts for other reasons are less regular and predictable, as the Commission follows the practice of making such adjustments only when there is a demonstrated need.

[2026] In Docket No. R84-1 the Commission moved the base year forward to improve the accuracy of its forecasts of volumes, revenues and costs in the test year. Advancing the base year minimized the impact of a set of net trends with a severe downwards bias that had been inserted by witness Tolley. In Docket No. R90-1 the base year was advanced and several other changes were made to deal with an unexpected change in economic conditions. During the 10-month course of the R90-1 proceedings the economy entered a recession. This recession was not anticipated in the Data Resources Inc. (DRI) forecasts of "economic determinants" such as income and the non-postal price indexes used in the Postal Service's initial filing. The Commission substituted a later DRI forecast of economic conditions so that the test year volume forecasts would include the effects of the recession. The Commission also corrected judgmental net trends superimposed on the forecasts by witness Tolley and added billing determinants for several proposed new categories of mail to correct the fixed-weight price indices. In principle and in practice the Postal Service's econometric models and worksheet implementations are supposed to allow the Commission to make these kinds of changes within a coordinated framework that is consistent with the econometric models upon which the forecasts must be based.

[2027] In supplemental testimony filed at approximately the mid-point of the present proceeding witness Thress disclosed that the Postal Service's econometric model and worksheets could not be used with then-current (June 2000) DRI forecasts of economic conditions to give valid forecasts of postal volumes during the test year. The reason given was that the Service's models and worksheets made no provision for a wide-ranging restatement and revision made by the U.S Department of Commerce in its historical income and consumption data. This revision rebased the DRI income and consumption series used in the Service's econometric models from 1992 dollars to 1996 dollars. Since the June 1999 DRI forecasts used in the Service's filing predated the rebasing, and any of the more recent DRI forecasts would postdate the rebasing, it was witness Thress' opinion that the Service's econometric model would have to be refit with the revised data before it could be used with a more recent DRI forecast to validly update the test year volume forecasts USPS-ST-46 at pp. 6-7. Witness Thress repeated this opinion in hearings.

(I)t is my judgement that if you are going to plug in new forecast data based on a new DRI forecast, which is based on restated Commerce Department data, that it would also be necessary for you to also plug in new elasticities which are estimated using consistent data.

[2028] The Commission accepts witness Thress' professional judgment on this point but finds much to criticize in the Postal Service's response to circumstances that has left us to depend on DRI economic forecasts that, at the time of the Postal Service's filing, were already almost 8 months old, and now are about 18 months old. The Commerce Department began to publish revised and rebased income and consumption data beginning in October 1999. USPS-ST-46 at p. 6. By November 1999 the Commerce Department's revisions of the historical data were still incomplete and extended back only 2-4 years according to witness Thress. Tr. 35/16851. The Postal Service could have re-estimated its demand equations to make them compatible with the revised and rebased Commerce Department series soon after November 1999, but, possibly, not soon enough to have used them for its filing. Apparently witness Thress was still using the unrevised income and consumption series to refit the demand model as late as November 1999. Id. at 16858.

[2029] The Postal Service filed its rate request in mid-January 2000. However, the forecasts for the filing were prepared in November and December 1999 from base year postal volumes that were supplied to witness Thress in October 1999. Id. at 16848. The proposed rates for the "after-rates" forecast were given to witness Thress around December 2, 1999. Id. at 16849. In November 1999 DRI published a new trendlong forecast that projected income and consumption for the revised and rebased Commerce Department series. At that point it should have occurred to the Service's volumes witnesses that they were using an econometric model and forecasting worksheets that would not comply with the Commission's rule 3001-45 (j) (6) (iv) quoted above.

[2030] The decision to use the June 1999 rather than the November 1999 DRI economic forecast was made in November 1999. Id. at 16850. Witness Thress, who made the decision, was aware of the more recent DRI forecast but chose to use the older one because it was consistent with the econometric model which had been fit to the old income and consumption series. Id. at 16851. Witness Thress explained the reasons for the decision at some length in his oral testimony. Id. at 16850-51. In November 1999, refitting the model with the revised data would have been problematic for several reasons. First, the revised data was historically incomplete although rebased income and consumption series would have been available for the entire period of the Postal Service's sample. Second, witness Thress believed that it would be necessary to reexamine the specification of the demand model and not just mechanically refit the old equations. Id. at 16852-53. Third, the June 1999 and November 1999 DRI forecasts are quite different. In June 1999 DRI thought the economy would level off and perhaps enter a recession within a year. By November 1999 it had become apparent that the economy was still growing strongly so the projected growth rates of the income and consumption series had all been substantially revised upwards by DRI. Id. at 16853. Using the November 1999 forecast would have been expected to produce a revenue surplus in the test year at the Service's proposed rates and requested contingency since all of the income elasticities in the demand models are positive numbers.

[2031] Later DRI forecasts continue to exhibit much higher trends for the economy's main consumption and income aggregates than the June 1999 forecast used by the Postal Service. This fact is evident from the projected growth rates for GDP for the June 1999 and May 2000 DRI forecasts shown in Table 1 of USPS-LR-I-447.

[2032] The use of the old June 1999 DRI forecast in the Service's January 2000 filing would have been avoidable if the November 1999 refit of the demand equations had used the best data available at the time, which were the Commerce Department's partially revised and rebased consumption and income data. Using the most current revised data for econometric time series estimation is an obvious best practice. Witness Thress testimony shows that he tries to follow this practice at least with respect to the Postal Service's volume data.

As a general rule, we try to update the equations on a quarterly basis probably. Generally speaking, whenever we get a new quarter, whenever there is a new quarter of volume data, . . .. And I think as a general rule, time permitting, I at least like to have my equations estimated using all the data I have, so that as we get a new quarter of data, if time permits, we will try to reestimate all the equations.

[2033] The DRI forecasts are econometric forecasts largely derived from equations fit econometrically to the Commerce Department's consumption and income time series. DRI's practice, according to witness Thress, is to re-estimate monthly.

My understanding is that DRI would again update their model every time they came out with a new forecast. So, in the case of DRI, I believe they update their model monthly, . . .. But I think they try each month certainly to incorporate whatever new information they have got versus what they had the month before.

[2034] By November 1999 DRI was following the best practice of using the partially revised and rebased Commerce Department data but the Postal Service was not.

[2035] Using the most current forecast of economic variables within the Postal Service's volumes forecasting worksheets would also be "considered better practice" "all other things being equal" according to witness Thress. "In theory", using the most current economic forecast tends to reduce errors in the Postal Service's volumes forecast. Id. at 16854. Rule 3001-45 (j) (6) (iv) is intended to enable the Commission to follow the better practice of using the most current information, including recommended rates, base year volumes, net trends, billing determinants, implementation date and DRI forecast, when it applies the Postal Service's econometric model to forecast test year volumes and revenues. For this reason the Postal Service's econometric models and forecasting worksheets need to have a useful life that is at least sufficient for the 10-month span of a postal rate case. In this regard the forecast worksheets are no different from the worksheets that the Postal Service and Commission use to predict test year costs.

[2036] The Postal Service's econometric models and worksheets for forecasting volumes were not usable as required by the Commission's rules when filed in January 2000. The worksheets could, of course, still be used mechanically Id. at. 16861. But they could not be used with the DRI economic forecasts of November 1999 (or later) to yield valid forecasts of postal volumes in the test year without plugging in new elasticities that would have to be obtained by re-estimating the demand equations. Id. at. 16862. This puts the worksheets out of compliance with rule 3001-45 (j) (6) (iv).

[2037] It is now clear that the DRI June 1999 forecast was unduly pessimistic. The difference in growth rate for Gross Domestic Product (GDP) between the DRI June 1999 baseline forecast and a comparable DRI forecast made in May 2000 is described in a Service Library Reference filed in August 2000:

In June of 1999, DRI forecasted real GDP to grow 4.1 percent from 1999 to 2001, with 2.1 percent growth from 2000 to 2001. In DRI's baseline May, 2000 forecast (DRI's Trendlong0500 forecast), however, real GDP is projected to grow 8.2 percent from 1999 to 2001 and 3.1 percent from 2000 to 2001.

[2038] In June 2000 the Postal Service re-estimated its econometric model and revised the test year volumes forecasts for its own use. The revised forecasts and forecasting worksheets are all included in USPS-LR-I-447. There were produced by the Service in August 2000 at the request of the Commission when it became apparent from the oral testimony of USPS witness Thress that the Service's econometric model had already been refit using the Commerce Department's revised and rebased series Tr. 35/16854-57 and 16864.

[2039] Tables 2 and 3 in USPS-LR-I-447 show how the Postal Service's filed forecasts and the forecasts from the refitted econometric model correspond. The tables compare filed forecasts and revised forecasts of GFY volumes as follows:

[2040] The filed and the revised forecasts differ very little in their predictions of volumes by class in the test year. This similarity in the forecasts is cited by the Service as one of the "compelling reasons to conclude that an attempt to update the volume forecast used in this proceeding in not warranted under current circumstances." USPS-LR-I-447 at p. 3. This accords with the recommendation of witness Thress in his Supplemental Testimony USPS-ST-46 and in hearings.

[2041] The Commission has used the filed volumes worksheets and the June 1999 DRI baseline economic forecast in its forecasts of test year volumes for the rates recommended by the Commission. The Commission notes that these forecasts correspond, not to a single current DRI baseline forecast, but to an average of DRI forecasts that are heavily weighted towards pessimistic and recession scenarios. Therefore, the test year volume forecasts include a significant built-in margin for downside error.

[2042] In the early history of the Commission, the volume and revenue forecasts provided by the Postal Service on the occasion of a general rate case were almost entirely judgmental. Now, and for many years past, the Commission's rules encourage the use of economic models, historical data, and econometric methods, and discourage reliance upon ad hoc methods and unsupported judgment in the preparation of forecasts. In every general rate proceeding since R80-1, the Postal Service and the Commission have relied upon the econometric research of witness Tolley. Since R90-1, the Service and the Commission have relied upon similar research for Priority and Express Mail conducted by witness Musgrave. The demand models used by witnesses Thress and Musgrave in this proceeding are recognizable variants of earlier models developed by witness Tolley. The models offered by witness Musgrave are not materially different from similar models for Priority and Express Mail that the Commission relied upon in R94-1. Witness Musgrave's models and estimation methods are still very much as described by the Commission in the R94-1 Opinion and Recommended Decision.

[2043] On the other hand, witness Thress conducted a thorough and effective revision of witness Tolley's models and econometric practice prior to the R97-1 proceeding. Many of these revisions corrected weaknesses and defects noted by the Commission in R94-1. In other respects, witness Thress' revisions appeared to be the result of a wide-ranging and open econometric reexploration of the underlying economic theory, the identification of suitable variables and the selection of appropriate estimation techniques for the Postal Service's volumes models. It was exactly the kind of econometric research that the Commission's Rules of Practice and Procedure for statistical evidence are intended to encourage. Witness Thress has continued to develop the Postal Service's econometric model and has eliminated several ad hoc estimation methods for cross-price and cross-volume elasticities that the Commission noted in its PRC Op R97-1, Appendix H at pp. 26-27.

[2044] Although there are many differences in detail in the equations fit for the R97-1 proceeding and the current proceeding there is nothing fundamentally new or different about the Postal Service's econometric models for forecasting volumes by class, subclass and worksharing category. Postal Service witness Thress also still uses the worksharing share model developed for the recent general reclassification proceeding Docket No. MC95-1. These models were all described and critiqued at some length in Appendix H to the Recommended Decision in Docket No. R97-1.

[2045] Witness Tolley (but not witness Musgrave) believes that it is still sometimes necessary to incorporate a term for recent unexplained trends in the volume forecasts. Therefore, he augments the model forecasts for some classes of mail with a net trend intended to represent a continuation of recent volume growth that cannot be attributed to movements in population, postal rates, income and other economic variables. The source of the net trends employed by witness Tolley is a forecast error analysis program described in the Technical Appendix to his direct testimony USPS-T-6 at A-28-A-34. The estimate that is most often selected is described as a "five-year mechanical net trend 1994q4 to 1999q4." It is the average annual trend unexplained by the demand model over the last five years of the sample. Witness Tolley uses his personal judgment to decide which mail categories will have net trends included in their forecasts and which will not. Thus, the Postal Service's volume forecasts should be viewed as dependent upon both the econometric studies performed by Postal Service witnesses Thress and Musgrave, and upon the personal judgments of witness Tolley with respect to the net trends.

[2046] The Commission has always regarded witness Tolley's error analysis program as an ad hoc method for estimating net trends that is being used in place of accepted econometric methodology. The accepted econometric methodology is to define a variable to represent a recent trend, include the variable in the specification of the demand equation, and estimate a coefficient for the variable along with the other parameters of the demand equation. The accepted econometric methodology has several advantages over witness Tolley's ad hoc procedure.

[2047] First, if a recent trend is really important then omitting a net trend variable from the demand models leaves estimates with a missing variable bias. Second, the statistical properties of all of the estimated parameters of the demand equation will be improved when an explanatory variable is added to capture an important recent net trend. Third, the estimated trend coefficient has all of the desirable properties of a generalized least squares estimate, whereas the statistical properties of witness Tolley's ad hoc estimates are unknown and may be undesirable. Fourth, the estimated net trend coefficient will have an associated "t-value" describing the accuracy of the estimate, whereas the accuracy of witness Tolley's net trends is a mystery. Fifth, the econometric methodology provides the appropriate setting for exploring refinements to the definition of the net trend variable itself. For example, witness Tolley's choice of a four or five year period for calculating net trends in his forecast error analysis is arbitrary and could easily be refined by witness Thress in the econometric research.

[2048] Witness Tolley's use of net trends to alter the forecasts has been sparing. For most subclasses, including all of the larger ones except Parcel Post, the net trend used in the forecast is one. Witness Tolley defends the use of the mechanical net trends for Parcel Post as an appropriate method for dealing with the delayed effects of a recent UPS strike on Parcel Post volumes. However, an inspection of witness Thress' Parcel Post equation shows that a dummy variable has already been included for this purpose.

[2049] The time that has elapsed since the filing of the current postal rate case has provided the Commission with the usual opportunity to compare the before-rates forecasts made by witnesses Tolley and Musgrave with four postal quarters of actual volumes. In Appendix I, Table I-1, the revised "before-rates" forecasts from witness Tolley's testimony are compared directly to the volumes shown in the Postal Service's quarterly reports of Revenue Pieces and Weight By Classes of Mail and Special Services. These reports have been submitted periodically during the current proceeding.

[2050] The before-rates forecasts continue to exhibit characteristics and patterns that the Commission has come to expect from similar comparisons with observed volumes in earlier proceedings. The comparison reveals again that an excellent overall performance masks large-but-offsetting forecast errors among the individual categories. Through the four postal quarters of PFY 2000, aggregate volume is predicted with considerable accuracy for total First-Class, total Periodicals and total Standard(A) mail. Typically, the percentage errors for all of First-Class and Standard A Mail lie within a range of several percent. However, the errors for most subclasses are larger in magnitude. The errors tend to be even larger for the smaller subclasses and worksharing categories of mail. On the whole the errors exhibit a pattern that could be explained by a fair amount of sampling error in the RPW statistics. Sampling errors would affect the RPW statistics for the smaller mail categories more severely that the larger or aggregated categories.

[2051] As in most earlier proceedings a comparison of predicted to observed overall volumes of mail does not support the hypothesis that the forecasts submitted by the Postal Service will systematically understate volumes during the test year. That is, the forecasts submitted by the Postal Service do not appear to have any overall bias through the four quarters of PFY 2000. However, as noted above, the Service's forecasts are based upon an overly pessimistic May 1999 DRI economic forecast.

[2052] Most of the differences between forecast and observed volumes do not appear to be entirely random from quarter to quarter. First, there is an evident tendency for differences to persist from quarter to quarter. This tendency for differences to persist may be explained in part by properties of the forecasting methodology used by Postal Service witnesses. This methodology forecasts off a base year rather than off the mean of the sample. Errors in the base year's RPW statistics are incorporated in the forecasts for the postal quarters that follow. For example, if the RPW volumes for Presorted Post Cards were high by 15 percent in the base year, this would be carried into the forecasts as a tendency for the quarterly forecasts of volumes to exceed actual volumes by 15 percent. Persistent differences may also be caused by incorrect net trends.

[2053] The second evident nonrandom pattern is to be seen in the errors for different worksharing categories for the same subclass. For example, in the errors for single-piece and worksharing First-Class letters and in the errors for regular presort and automation presort Standard Mail (A). The errors for these worksharing categories are very often offsetting. This means that witness Thress' share model is not doing a particularly accurate job of dividing the subclass volumes predicted by his econometric equations.

[2054] A side-by-side comparison of Postal Service and Commission after-rates volume forecasts in the Test Year is shown in Appendix I, Table I-2. The forecasts in column from left to right are: Column 1, the USPS forecast as filed using the USPS proposed rates and June 1999 DRI economic forecast, Column 2, the USPS forecast as filed with Priority and Express mail revisions, Column 3, the Commission's forecast with the recommended rates and June 1999 DRI economic forecast.

[2055] In addition to the update-related changes summarized in subsection A., the Commission has made other adjustments and corrections in costs and revenues where they have been found to be justified. A summary of the major non-update related changes in the revenue requirement is provided at the conclusion of this subsection, following a discussion of two adjustments proposed by participants.

[2056] DMA witness Buc proposes a reduction of approximately $93 million in supervisor costs in this case to correct what he characterizes as a flaw in the Postal Service's rollforward model. In Docket No. R97-1, witness Buc testified that supervisors' costs should be reduced in proportion with decreases in their managed employees' work hours resulting from cost reduction programs. The Commission agreed with the proposal and recognized a net decrease in supervisory costs of approximately $100 million, finding that the Postal Service had not effectively rebutted its factual premises with record evidence. PRC Op. R97-1, para. 2154.

[2057] Inasmuch as the Service made no such proportional adjustment to supervisory costs in this case, witness Buc reiterates his proposal on the same grounds. Tr. 22/9547-49. In response, the Postal Service sponsors the rebuttal testimony of witness Patelunas, USPS-RT-4. According to witness Patelunas, witness Buc's proposed adjustment is improper because it is inconsistent with operational realities, which effectively limit the opportunity to reduce supervisor costs in direct proportion to craft workhour savings in implementing cost reduction programs. Tr. 38/17142-43. Witness Patelunas also testifies that actual supervisor costs for FY 1998 were very close to both the Postal Service's and the Commission's original, unadjusted estimate in the R97-1 case, while the proportionally adjusted estimate shows a greater variance. This, he suggests, demonstrates that the argument underlying the adjustment is invalid. Id. at 17144. In summary, witness Patelunas testifies that supervisor cost savings opportunities should be reviewed in the context of the functions, obligations and environment of supervision, not "merely mechanistically piggybacked on direct labor costs." Id. at 17145.

[2058] Notwithstanding the testimony in rebuttal to the proposed adjustment, the Commission will retain this approach to estimating supervisor costs, resulting in a cost reduction of approximately $97 million. Witness Patelunas' testimony discloses operational limitations that might prevent supervisory costs from decreasing proportionately with craft workhours in implementing cost reduction programs, and the Commission welcomes more detailed presentations on this subject in future proceedings. However, Postal Service Library Reference USPS LR-I-1, "Summary Description of USPS Development of Costs by Segments and Components, Fiscal Year 1998," contains the following statements:

It is recognized that a change in employee workhours, caused by a change in mail volume, may not be accompanied immediately by a corresponding change in firstline supervisory workhours. However, for any substantial or prolonged change in the level of non supervisory employee effort for a given work activity, there will be an accompanying change in firstline supervisory requirements.

USPS-LR-I-1 at 2-2, §§ 2.1.1, 2.1.3. (emphasis added.) These statements, which are unchanged from corresponding provisions in USPS-LR-H-1 in Docket No. R97-1, lend support to the factual premise of witness Buc's proposed adjustment-namely, that in the long run developments that decrease employee work effort should also result in decreases in supervisory work effort, and that both will be reflected in volume-variable cost changes. There may be exceptions to this linkage, as witness Patelunas testifies, but in the absence of more detailed evidence, the Commission concludes that the proportional relationship on which the adjustment relies remains valid.

[2059] Regarding witness Patelunas' argument based on actual versus estimated supervisor cost results for FY 1998, the Commission does not view the reported variances as probative evidence of the invalidity of witness Buc's proposed adjustment. Many factors can influence the amount of cost actually incurred in this category in a fiscal year, and the effect of operational cost reduction programs has not been isolated and retrospectively analyzed and reported by the Postal Service on the record. Such analysis would be useful in establishing the effectiveness of the Service's cost reduction efforts, and the Commission recommends it for use in future proceedings.

[2060] Witness Haldi, testifying on behalf of the Alliance of Nonprofit Mailers, proposes that the Commission reduce the unit cost of Periodicals on the basis of a judgmental assessment that the Postal Service has under-invested in flats processing equipment that would have enhanced processing efficiency and thereby restrained cost increases. He testifies that the Service's spending on capital investment has been grossly inadequate, and has led to severe shortages of mechanized and automated sorting capacity for periodical and non-letter mail and of facility space for sorting such mail. Tr. 22/9625-42. According to witness Haldi, no obstacle or countervailing consideration has prevented the Postal Service from making adequate capital commitments that would have prevented these developments. Id. at 9643-48. On the ground that, "[t]he Postal Reorganization Act entitles the Postal Service only to those revenues needed to cover costs under `honest, economical and efficient management[,]'" Id. at 9622 (footnote omitted), witness Haldi proposes that the Commission redress the Service's inaction by disallowing 1.2 cents per piece for all Regular Rate, Nonprofit and Classroom periodicals mail in the test year, for a total of approximately $94 million. Id. at 9650-54.

[2061] In response to this proposal, the Postal Service filed rebuttal testimony sponsored by witnesses Dowling and Strasser. Witness Dowling, the Postal Service's Vice President of Engineering, testifies that the Service has made continuous progress since the early 1990's toward its ultimate objective of bringing flats automation along as far as letter automation has progressed. According to witness Dowling, the Service has pursued this objective through acquiring and upgrading where possible successive generations of flats processing equipment, as well as small parcel and bundle sorters. He also states that the Service is currently developing several designs for a flats bundle collator. Tr. 46A/20476-80. Further, while witness Dowling agrees that technological advancements in processing flats have lagged behind letter automation, he states that it has not been due to a lack of commitment. He testifies that the Service and its supplier have aggressively pursued development of new flats processing technologies, but observes that not all research activities lead to viable improvements. Id. at 20480-81.

[2062] Witness Strasser opposes ANM's proposed productivity adjustment on several grounds. He testifies that witness Haldi's adjustment is based on faulty premises concerning the appropriate level of capital investment by the Postal Service and a skewed selection of FY 1993 as a base period. Id. at 20201-03. Further, he states the proposed adjustment fails to account for numerous cost savings programs that inure to the benefit of Periodicals between the base year and the test year. Id. at 20203.

[2063] Witness Strasser observes that there are many, sometimes highly complex reasons for changes in unit costs, and states his opinion that it would not be prudent to disallow increased costs that may be due to changes in the makeup of mail or other causes. More generally, he states that such an adjustment would be unjustified in light of witness Dowling's demonstration that the Service has pursued flats automation opportunities in a responsible way. Ibid.

[2064] On brief, the Postal Service argues that the proposed disallowance of costs should be rejected because it is flawed factually, analytically, and legally. The Service contends that, contrary to witness Haldi's claims, the record reflects a consistent pattern of prudent investment, including research and development in pursuit of efficiency gains, that is fully supported by postal management and the Board of Governors. Postal Service Brief at II-12 through II-13. Analytically, the Service argues that witness Strasser exposed witness Haldi's failure to account for new and expanded cost savings programs for Periodicals between the base and test years, and his flawed reliance on total factor productivity and net investment rates in other industries. Id. at II-13.

[2065] Finally, the Service presents extensive legal argument to support its assertion that the proposed disallowance relies on a defective interpretation of the Reorganization Act. According to the Service, witness Haldi's testimony unjustifiably superimposes the "honest, economical, and efficient management" (or "HEEM") standard in § 3621 upon the § 3622 ratemaking criteria the Commission implements in rate proceedings. Id. at II-14-II-22. Further, the Service argues that disallowing future estimated costs on the basis of alleged past management failures would both violate the § 3621 breakeven requirement and involve the Commission in reducing the revenue requirement for disciplinary purposes, which the Court of Appeals found to be in excess of its authority in Newsweek v. United States Postal Service.² Id. at II-22-II-24. Finally, the Service argues that the breakeven requirement in § 3621 makes disallowance of actual costs because of HEEM concerns while still providing sufficient revenues a practical impossibility. In any event, the Service asserts, it has met its obligations under the HEEM standard with respect to Periodicals mail by giving special recognition to cost reduction programs in this proceeding. Id. at II-16-II-26.

[2066] The Commission declines to disallow any portion of the estimated costs of Periodicals mail on the basis of witness Haldi's proposal. While the Commission holds opinions on the extent of its role vis-à-vis the revenue requirement that diverge from those argued by the Postal Service-as the following subsection will elaborate-the Commission believes the facts must control the outcome of all such controversies. In view of the Postal Service's detailed testimony explaining its ongoing efforts to improve the efficiency of flats processing, the Commission finds insufficient justification for concluding that the Postal Service has incurred expenses in contravention of the honest, efficient, and economical standard of § 3621 with respect to Periodicals mail or flats generally.

[2067] Proportional reduction in supervisor costs. In addition to the discussion of this proposed adjustment in subsection C.1., Appendix D, Schedule D-3, displays the calculation of this reduction.

[2068] Increase in First-Class additional-ounce revenue. This revenue adjustment is discussed in § V.B.1.e., and its calculation is presented in Library Reference PRC-LR-3.

[2069] Miscellaneous adjustments in flats processing costs. Some of these reductions were proposed by participants as decreases in Periodicals costs, and these are discussed in § V.D.2.c. The cost reduction associated with reduced flats bundle breakage is a component of the Postal Service's "breakthrough productivity" initiatives, and is discussed in Appendix D.

[2070] Adjustments resulting from application of PRC attribution methodology. Applying the Commission's cost attribution methodology as discussed in various portions of Chapter III produces changes in the revenue requirement. The detailed application of these changes is presented in Library Reference PRC-LR-4.

[2071] Further updates. The Commission has recalculated projections of cost-of-living allowance additions to employee compensation by substituting the actual CPI-W indices for June through September, 2000, for the estimated January, 2001 and July, 2001 indices. This necessitates a recalculation of the estimated CPI-W indices for October, 2000 through September, 2001. These updates, described in Appendix D, result in increases in estimated COLA for FY 2001, displayed in Table D-2.

[2072] Miscellaneous corrections of USPS volumes, revenues and costs. These various corrections are documented in Appendix D and Library Reference PRC-LR-3.

[2073] Contingency provision. The Commission's recommendations regarding the amount of the provision for contingencies are discussed in subsection D.

[2074] Having arrived at an aggregate estimate of the costs likely to be incurred by the Postal Service in the test year, it is now appropriate to turn to a separate revenue item, the provision for contingencies. In identifying potential categories of "total estimated costs" of the Postal Service recoverable through postal rates and fees, such as operating expenses and various financial accounts, § 3621 explicitly includes "a reasonable provision for contingencies" without further description or explanation.

[2075] While the component of the Postal Service's revenue requirement included in particular rate requests as a contingency provision has been a focus of participants' presentations in past omnibus rate proceedings, the Service's proposed 2.5 percent contingency allowance has provoked extensive criticism and controversy in this case. A number of mail users and the Office of the Consumer Advocate have challenged the proposed amount as excessive, and offered testimony proposing recognition of lesser amounts in the revenue requirement. In addition to sponsoring rebuttal testimony responsive to these challenges, the Service has presented extensive legal argument on brief concerning such fundamental issues as the scope of the Commission's authority in this area and the respective roles of the Governors and the Commission in ratemaking under the Reorganization Act. The unusual prominence of this controversy requires particular attention in this opinion.

[2076] Postal Service Request. In Docket No. R97-1, the Postal Service's request provided for a 1 percent contingency, which had a dollar value of $605.5 million at the level of accrued costs estimated by the Service in that case. Witness Tayman characterized 1 percent as a "smaller" contingency, and testified that the amount represented the Postal Service's desire to keep rate increases as low as possible and below the level of growth in general inflation. He also cited the Service's recent financial success, the favorable economic climate at that time, and postal management's concern about the effect of the contingency on rate levels in support of a smaller contingency. However, he left the door open for a return to a larger contingency in the future, if necessary, because of changed circumstances. PRC Op. R97-1, para. 2024, citing USPS-T-9 at 38.

[2077] In this case, witness Tayman testifies in support of a larger proportional contingency of 2.5 percent. The dollar value of this proportion is $1.68 billion on a test year after-rates basis. USPS-T-9 at 43.

[2078] Witness Tayman presents several rationales for what he characterizes as a "mid-range" contingency of 2.5 percent, between the 1- and 2-percent amounts included in Docket Nos. R97-1 and R94-1 and the 3.5 percent contingencies included in earlier cases. He cites the following considerations: (1) recent financial performance less favorable than in the mid-1990s; (2) volume growth below historical norms and projections accompanied by delivery network growth; (3) the challenge of achieving a 1.5 percent workyear reduction in the test year; (4) apparently significant new pressures on salary and benefit cost levels; (5) an acceleration in health benefit cost increases and the prospect of significantly more costly labor contracts; (6) the increasingly competitive environment in which the Postal Service operates, including electronic alternatives and U.S. operations of foreign postal administrations; and (7) the Postal Service's inability to implement rates recommended in this proceeding until January, in the second quarter of the test year. Id. at 43-44.

[2079] Witness Tayman also provides historical variance analyses similar to those included in past Postal Service filings, which produce hypothetical test year variances ranging from -2.2 percent (or -$1.5 billion) to 2.3 percent (or a positive $1.5 billion). However, he states that these analyses are provided for informational purposes only, and opines that variance analyses should not be the basis for determining the need for a contingency or its size. His belief is based, in part, on recognition that "[t]he Postal Service's financial performance is under much greater pressure and is subject to substantially greater risks than it was at the time of the last two omnibus rate cases." Id. at 45.

[2080] Testimonies of Participants Opposing Postal Service Contingency Proposal. Several participants sponsored testimony opposing the Postal Service's proposal of a 2.5 percent contingency. These parties include the Association of American Publishers; the Coalition of Religious Press Associations; a consortium consisting of Direct Marketing Association, Advo, Inc., Alliance of Independent Store Owners and Professionals, Alliance of Nonprofit Mailers, Amazon.Com, Inc., American Business Media, American Library Association, Association for Postal Commerce, Association of Priority Mail Users, Inc., Coalition of Religious Press Associations, Dow Jones & Company, Inc., Florida Gift Fruit Shippers Association, Magazine Publishers of America, Major Mailers Association, McGraw-Hill Companies, Inc., Parcel Shippers Association, and Time Warner Inc;³ the Office of the Consumer Advocate; Parcel Shippers Association; and Val-Pak Direct Marketing Systems, Inc. and Carol Wright Promotions, Inc.

[2081] Testimony of witness Buc. In testimony sponsored jointly by 16 mail users and associations, Lawrence G. Buc opposes the Postal Service's contingency provision as being "neither reasoned nor reasonable." Tr. 22/9531. Following a review of the Commission's criteria and actions regarding the contingency provision in past proceedings, witness Buc asserts that witness Tayman's presentation in this case provides little support for the Service's proposed contingency provision. He criticizes Witness Tayman's dismissal of the Service's variance analysis as a basis for determining an appropriate contingency level, and asserts that five of the seven considerations he cites do not provide support because they concern financial challenges that cannot be considered "unforeseen and unforeseeable events." Id. at 9540-44.

[2082] Witness Buc also testifies that neither the current financial condition of the Postal Service nor general economic conditions provide support for the 2.5 percent contingency provision. According to witness Buc, the financial condition of the Service as measured by its equity position is far superior in this case compared with its status in the last two cases, and it is currently ahead of its cumulative target for equity restoration. Thus, the Service is better situated to withstand adverse unforeseen events than it was in cases in which contingency requests of 1 and 2 percent were approved. Id. at 9544-45. Further, he testifies that projections of relevant measures of test year inflation in the form of the consumer price index (CPI-W), the Employment Cost Index (ECI), and the Producer Price Index (WPI) do not indicate the need for a higher contingency provision in this case than in the previous two. Id. at 9545-46.

[2083] Given the results of the variance analysis presented by witness Tayman, the financial condition of the Postal Service, the state of the economy, and the guidance provided by the Commission's decisions over the past 25 years, witness Buc submits that a reasoned and reasonable contingency in this case should be no larger than those incorporated in either of the previous two omnibus rate cases. In comparing R94-1 to R2000-1, he testifies that the variance analysis and the Service's financial condition indicate that a much smaller contingency is needed in this case than the 2 percent provision adopted in R94-1. In comparing R97-1 to R2000-1, he testifies that the same indicia suggest a smaller provision, while the general state of the economy could support the same 1 percent contingency in this case as adopted in R2000-1. Consequently, witness Buc concludes that a reasoned and reasonable contingency is 1 percent in this proceeding. Id. at 9547.

[2084] Testimonies of OCA witnesses Burns and Rosenberg. The Office of the Consumer Advocate sponsored the testimonies of Robert Burns and Edwin Rosenberg, both of whom are affiliated with the National Regulatory Research Institute (NRRI), the research and public service organization for the National Association of Regulatory Utility Commissioners (NARUC), of which the Commission has been a federal member since 1971. Both witnesses oppose the Postal Service's 2.5 percent contingency provision, but offer somewhat different perspectives in doing so.

[2085] Testimony of witness Burns. Witness Burns addresses the purpose of the statutory provision for contingencies, the standards applied by the Commission in reviewing particular provisions in rate proceedings, and the adequacy of the reasons given by witness Tayman in support of the 2.5 percent contingency in the Postal Service's request. According to witness Burns, the purpose of the contingency provision is twofold: to provide a cushion against potential expenses caused by unforeseeable events, and to compensate for forecasting errors. Id. at 9710. Witness Burns observes that contingency reserves are used for the same purpose in the insurance industry, and are subject to a requirement that the provision be clearly related to future, uncontrollable events, rather than serving as a device to smooth out irregularities or volatility in earnings. Id. at 9710-11. This requirement is important, he testifies, because without it contingency reserves tend to become larger than necessary, and managers of the enterprise make less effort to limit cost increases within their control. Id. at 9712-14.

[2086] Witness Burns testifies that the Commission's approach to reviewing the Service's contingency request in past cases has been consistent with these precepts, by focusing on prevailing national economic conditions, requiring that postal management's subjective perception of risks be reasonably articulated, and requiring supporting substantial evidence, such as the results of variance analyses. According to witness Burns, the Commission's prior decisions suggest that proper assessment of a proposed contingency provision rests on: (1) careful evaluation of postal management's explanation of its subjective judgment on the proposed level; (2) objective review of potential forecasting errors; and (3) consideration of external factors such as key national economic indicators and economic stability. Id. at 9714.

[2087] The balance of witness Burns' testimony addresses the first mode of assessment. He asserts that Postal Service witness Tayman has not articulated a reasonable basis for substantiating management's subjective judgment in favor of a 2.5 percent contingency provision. Regarding witness Tayman's reference to unfavorable recent financial performance, witness Burns notes that the 1 percent contingency approved in R97-1 allowed the Service to finish Fiscal Year 1999 with net revenue of $363 million, notwithstanding substantial spending on Y2K computer system remediation in that period and the surrounding two fiscal years. Id. at 9716-17. Regarding volume growth below historical norms and the Service's planned 1.5 percent workyear reduction, witness Burns testifies that this is an area within postal management's scrutiny and control. Id. at 9717. Witness Burns similarly asserts that the "new pressures" on salary and benefit levels, health benefit cost increases, and labor contract costs are either within postal management's influence or subject to estimation from available indices and forecasts; he also claims witness Tayman's reference to "other uncertainties" affords no support in the form of substantial evidence.

[2088] With respect to the "increasingly competitive environment" in which the Postal Service purportedly operates, witness Burns notes that the Service's volume forecasts take such potential diversions of mail into account, and claims that witness Tayman's general reference to the Internet making inroads into mail volume is insufficient to influence the rational choice of a contingency level. Id. at 9718-19. Finally, on the subject of possible legislative change, witness Burns claims that the potential for an outcome beneficial to the Postal Service-in the form of increased ratemaking flexibility-appears to be at least as favorable as the potential for a negative outcome. Id. at 9719.

[2089] On the basis of his review of the risk factors presented by witness Tayman, witness Burns concludes that the Postal Service has failed to articulate a rational connection between potential uncontrollable risks in the test year and the proposed $1.69 billion contingency provision. He therefore concurs with witness Rosenberg's recommendation that the level of the contingency be kept at 1 percent. Ibid.

[2090] Testimony of witness Rosenberg. Witness Rosenberg opposes increasing the contingency provision to 2.5 percent as neither necessary for the continued successful operation of the Postal Service nor in the public interest. His opinion is based on his application of public policy considerations, regulatory principles, and evaluation of witness Tayman's testimony and exhibits.

[2091] Witness Rosenberg establishes a context for his recommendations by identifying postal ratemaking under the Reorganization Act as a variant of cost-of-service ratemaking that avoids some of its shortcomings. Particularly, he observes that the Postal Service is allowed to base its rate requests on its best estimates of costs in a prospective test year, rather than being tied to historical costs; that in addition to relying on projections of future costs a contingency allowance is available to provide a margin of safety against unforeseen and uncontrollable circumstances; and that if revenue and expense estimates nevertheless prove to be off target, the resulting operating deficit can be recovered in the future through use of the prior years' loss provision. Thus, he states, the Postal Service has three different levels of protection for breaking even financially while providing good service to consumers at reasonable rates. Id. at 9807-09.

[2092] Witness Rosenberg characterizes the function of the contingency provision as a form of insurance against unforeseen, unexpected, and uncontrollable adverse fluctuations in revenues or expenses. Because prospective cost estimates cannot account for all possible fluctuations, the contingency provision serves as a cushion against occurrences that could not reasonably be forecasted or foreseen. Additionally, the contingency serves implicitly as a means for lengthening the time between postal rate increases. The essential question, he submits, is: What is the optimum size of the contingency? Id. at 9810.

[2093] According to witness Rosenberg, a disciplined analysis of the question would consider: (1) the magnitude and types of uncertainties requiring a contingency, with particular attention to the state of the economy; (2) the Postal Service's historical experience with contingency provisions of various magnitudes; and (3) the short-run and long-run effects of either too large or too small contingency provisions on the Postal Service and its managers, and on the Service's customers. Based on his analyses of these considerations, witness Rosenberg testifies that a 2.5 percent contingency is not necessary at this time, and that a 1 percent contingency provision should be recommended. Id. at 9810-11.

[2094] Addressing the first consideration, witness Rosenberg presents statistics to support his conclusion that conditions in the national economy are relatively stable: the United States is currently enjoying the longest economic expansion in more than half a century, and is doing so in a climate of relatively low inflation. According to witness Rosenberg, these conditions should allow the Postal Service to meet its responsibilities with a minimum contingency provision. Id. at 9811-15.

[2095] Regarding the second consideration, he testifies that the Postal Service has been able to achieve a positive net income over the two most recent rate cycles with contingency provisions less than the requested 2.5 percent. He notes that the Service has generated a cumulative net income of $5.58 billion during the FY 1995-2000 period, while the contingency provision was set at 2 percent as the result of Docket No. R94-1 and 1 percent in Docket No. R97-1. On this basis, he concludes that nothing in the Service's recent operating history indicates a need for a 2.5 percent contingency provision. Id. at 9815.

[2096] Witness Rosenberg also asserts that adopting a 2.5 percent provision in this case would run counter to a downward trend he identifies in the contingency provision over time. He observes that the Postal Service has had nearly 30 years of experience operating under the Reorganization Act in a more businesslike manner, and that its forecasting ability is improving. Witness Rosenberg presents a chronological table showing, he testifies, coincident downward trends in CPI increases and the percentage value of the contingency provision in successive rate cases since 1976. Id. at 9816-19.

[2097] Witness Rosenberg also observes that the requested increase in the contingency provision from 1 percent to 2.5 percent requires the production of additional revenue in excess of 27 percent of the total revenue requirement deficiency identified by witness Tayman. An increase of this magnitude, he believes, requires well-reasoned justification. However, he cites the inadequacies in witness Tayman's rationale addressed by his colleague witness Burns and finds no support in the variance analysis presented by witness Tayman. He notes that the proposed 2.5 percent allowance lies outside the range of that variance analysis, and that its four scenarios produce results that both on average and in total involve revenue deficiencies of less than 1 percent. Id. at 9820-24. Witness Rosenberg also testifies that it would be useful for the Postal Service to develop some other analytical approach for this purpose, and provides examples from the electric utility, telephone, and natural gas industries. Id. at 9824-26.

[2098] On the third consideration he addresses-the short-term and long-term effects of smaller and larger contingency provisions on the Postal Service, its management, and its customers-witness Rosenberg advances several reasons for concluding that larger provisions are not preferable. If the cushion provided by the contingency allowance is too thick, he states, postal management will have a diminished incentive to manage economically and efficiently because the goal of breakeven can be achieved without having to make tough decisions in the face of higher cost levels or other adverse circumstances. Id. at 9826-27. According to witness Rosenberg, restoration of equity would not justify erring on the side of too large a contingency provision; this function should be restricted to the prior years' losses allowance, and a shorter amortization period should be requested if the Postal Service wishes to accelerate its rate of equity recovery. Id. at 9830-32. He also observes that a smaller contingency provision can be achieved by shortening the Postal Service's rate cycle, which will increase forecasting accuracy by shortening the horizon of estimation. Id. at 9828-30.

[2099] Finally, witness Rosenberg testifies that the contingency provision should not be increased to 2.5 percent in light of impact considerations. He notes that the required additional $1 billion will be extracted from the pockets of Postal Service customers, who will suffer a consequent lost opportunity cost. He denies that the arguable indirect benefit to customers of providing longer rate stability furthers the purpose the contingency provision is intended to serve. Id. at 9827-28. He also observes that increasing rates by an additional $1 billion to fund a larger contingency provision may be counterproductive, because it would degrade the Postal Service's position in what witness Tayman characterizes as an "`increasingly competitive environment,'" and at worst may produce a "`vicious cycle'" in which rising postal rates create more headroom for competitors, which would result in lower postal revenues and pressure for further rate increases. Id. at 9832-33, citing USPS-T-9 at 44.

[2100] Following the Postal Service's filing of information in response to Order No. 1294, OCA submitted additional testimony by witness Rosenberg in which he reconsiders his recommendation of a 1 percent contingency provision in light of the new presentation. According to witness Rosenberg, the Service's updated expense estimates may overstate the revenue requirement and the claimed revenue deficiency; in particular, he identifies the "Field Reserve" exclusion of $200 million from target cost reductions, the shift from ECI minus one to ECI for estimating growth in labor costs, and witness Patelunas' failure to recognize actual results for FY 2000 as possible bases for reducing the estimated net loss of $325.5 million. Tr. 41/18304-08.

[2101] Witness Rosenberg notes witness Patelunas' confirmation that the revised cost level estimates, based on more recent DRI forecasts identified in a table included in the rebuttal testimony of OCA witness Thompson, are likely to be more accurate than those contained in the original filing. On this basis, witness Rosenberg testifies that the more recent forecasts support a less generous contingency provision. However, he testifies that use of the May 2000 DRI indices for fuel prices, rather than the lower indices issued by DRI in July 2000, would tend to overstate these test year expenses. Id. at 18308-10.

[2102] Using the test year costs presented by witness Patelunas, which he believes may be overstated, witness Rosenberg estimates that a 1 percent contingency provision would produce a revenue surplus of $739.4 million. He states that the surplus will be even greater if costs are lower than estimated, or if the final results for FY 2000 are better than the Postal Service's predicted deficit. For these reasons, he maintains his previous recommendation of a 1 percent contingency provision. Id. at 18312.

[2103] Testimony of witness Haldi. Witness Haldi, testifying on behalf of Val-Pak/Carol Wright, opposes the Postal Service's contingency proposal as inadequately supported and excessive. Inasmuch as the Postal Service has recently added significantly to its forecasting capabilities, and the provision for recovery of prior years' losses is available in the test year, he submits that the Service should not need such a large contingency to insure against errors in forecasting. Tr. 32/15784-85. Indeed, because the prior years' losses recovery mechanism serves as a retrospective contingency allowance, he states that it is not necessary to be overly conservative about protecting against any shortfall during the test year via a large prospective contingency. Id. at 15787. He also testifies that surpluses should not be intentionally created by inflating the contingency provision in order to fund capital improvement programs, as the Postal Service has ample borrowing authority. Id. at 15785-87.

[2104] Witness Haldi further testifies that the Service's proposed contingency provision is not only unnecessarily excessive, it is also counterproductive to the sound management of the Postal Service. He observes that the fixed costs of the Service's delivery network are large, and that spreading those fixed costs to keep rates affordable to all users requires large volumes of mail. He also notes the recent expressions of concern by GAO and others about the prospect of major declines in future volume because of electronic diversion. In light of these considerations, he testifies that the Postal Service needs to keep rate increases to an absolute minimum to preserve necessary volume.

[2105] Witness Haldi notes that, contrary to this recommendation, the proposed 2.5 percent contingency accounts for almost half the aggregate rate increase in this case, which exceeds the rate of inflation. While superficially this appears to provide the Service additional cash as a cushion against inflation, witness Haldi submits that the higher rates it requires will stimulate competition based both on innovations in information technology and in more conventional media. Therefore, he asserts that the Service's proposed contingency is counterproductive, and should be reduced to no more than 20 to 25 percent of the projected test year shortfall without any contingency, i.e., to between $400 and $500 million in this case, with any actual additional shortfall to be made up through prior years' loss recovery. Id. at 15787-90.

[2106] Testimony of witness Zimmerman. In a similar vein, Parcel Shippers Association witness Zimmerman questions the Postal Service's willingness to assess a disproportionately large contingency allowance that will drive rate increases above the rate of inflation and have "the predictable effect of killing volume, thereby spreading non-variable institutional costs over a smaller base, with resultant adverse revenue effects. . . . " Tr. 29/14130. He remarks on the Postal Service's profitable operation following the R94-1 and R97-1 proceedings, in which it requested, and the Commission recommended, overall rate increases that were less than the rate of inflation in the general economy. Id. at 14129. If the requested overall 6.4 percent increase were reduced by 2.7 percent, the amount by which the assumed rate of inflation exceeds the CPI-W estimate, he observes that it would have the effect of reducing the increase by $1.5 billion. Accordingly, witness Zimmerman asserts that the contingency should be reduced by at least $1 billion, aligning the overall increase much more with the expected inflation rate in the general economy. Id. at 14130-31.

[2107] Testimony of witness Stapert. Dr. John C. Stapert, testifying on behalf of the eight-member Coalition of Religious Press Associations (CRPA), also testifies in opposition to the Postal Service's proposed contingency provision. He notes that the Coalition endorses the testimony of witness Morrow on the proper amount of contingency allowance that Periodicals should bear,⁴ but offers additional observations regarding aspects of the Service's proposal that he finds "particularly peculiar." Tr. 30/14445. He states that witness Tayman's explanation of the need for a $1.7 billion contingency allowance is unconvincing in light of the Service's cumulative earnings since the last two rate proceedings, and of its current revenue position. While the contingency purportedly is intended to guard against many unnamed uncertainties, witness Stapert cites witness Tayman's statement that it is reflected as a test year expense and cash requirement, and thus "[i]t is a certainty that USPS will spend every dime it can get, which hardly promotes efficiency." Id. at 14446.

[2108] Testimonies of DMA witnesses Buc and Bernheimer. Following the Postal Service's submission of supplemental testimony and other information in response to Order No. 1294, the DMA-led consortium filed the supplemental testimonies of witnesses Buc and Bernheimer, as provided for in P. O. Ruling R2000-1/71. Both testimonies address the appropriate level of contingency provision in light of the updated information provided by the Postal Service.

[2109] Supplemental Testimony of witness Buc. Witness Buc revises his earlier recommendation of a 1 percent contingency provision to one-quarter of 1 percent. The revision of his earlier contingency proposal is based on his review of the testimony of Postal Service witness Patelunas, USPS-ST-44. Tr. 38/17185.

[2110] According to witness Buc, there are four reasons for reducing the contingency to a fraction of 1 percent. First, he claims that witness Patelunas' use of the full value of the Employment Cost Index (ECI)-rather than ECI minus 1, used in the Service's initial presentation-as a basis for estimating the aggregate percentage pay increase in upcoming wage settlements justifies a reduction in the contingency. By increasing the basis for the wage settlement and including these increases in the various cost segments affected, he submits, the Postal Service has correspondingly reduced the risk to which it is exposed in the form of additional labor costs. Using the Postal Service's Library Reference LR-I-421, he calculates the appropriate reduction to equal $246.6 million. Id. at 17187-88.

[2111] The second basis cited by witness Buc is witness Patelunas' failure to include in his revised estimates for Postal Service cost reduction programs the full amount of cost reductions to which the Postmaster General has publicly committed. Witness Buc notes that witness Patelunas' revised estimates include an additional $544 million in cost reductions, $456 million less than the one billion dollar "breakthrough productivity" target publicly announced by the Postmaster General, and that Patelunas conceded on oral cross-examination that the savings could exceed his estimate. Id. at 17188. Additionally, he observes that the cost reductions used by witness Patelunas for his test year after-rates cost forecast are $200 million less than the cost reductions contained in the draft budget for Fiscal Year 2001. At a minimum, witness Buc submits, the contingency should be reduced by the $200 million of cost reductions that appear in the draft budget but not in the Postal Service's response to Order No. 1294. Id. at 17189.

[2112] The timing of the Postal Service's revision of its cost estimates provides the third basis cited by witness Buc for reducing the contingency. With their filing less than three months before the start of the test year, witness Buc submits that the shortened forecasting horizon reduces the risk of outcomes lying outside the range of predicted increases and decreases. Ibid.

[2113] Finally, witness Buc asserts that the very outcome of the Postal Service's re-estimation of test year after-rates costs justifies a lower contingency. He characterizes that process as "an experiment to determine the sensitivity of the deficiency with respect to changes in inflation rates." Ibid. Notwithstanding the substantial increases in key inflation indices shown in Exhibit USPS-ST-44AB, and the substitution of ECI for ECI minus one for estimating wage settlement increases, he notes that the effect on Postal Service net income is to increase the deficiency from $21.8 million in the initial filing to $275.3 million. Even with the inclusion of $200 million as a Field Reserve offset to cost reductions, he calculates that the increase in the revenue deficiency represents only 0.38 percent of the original estimated total of test year after-rates costs. Id. at 17189-90.

[2114] Supplemental Testimony of witness Bernheimer. Approaching postal finance from the "top down," rather than the Service's own "bottoms up" approach of rolling segmented costs forward from a base year for ratemaking purposes, witness Bernheimer presents an analysis of aggregate growth in Postal Service costs to support his conclusion that any contingency allowance "is unjustifiable, unnecessary, and uncalled for." Tr. 46A/20420. Inasmuch as detailed data are available for the first 11 accounting periods of FY 2000, and the Service has selected FY 2001 as its test year, witness Bernheimer states that there is sufficient information to make a very accurate estimate for the year immediately preceding the test year. Using an expense growth rate of 4.5 percent, notwithstanding the Service's cost-cutting programs, witness Bernheimer projects total expenses for FY 2000 to equal $64.513.5 billion, which represents a 3.9 percent increase over expenses for FY 1999. Id. at 20421.

[2115] Witness Bernheimer observes that the Postal Service's revenue requirement of $69.644.9 billion incorporates a growth rate of 8.0 percent over the estimate he calculates for aggregate expenses in FY 2000. According to witness Bernheimer, a growth rate of this magnitude is "exaggerated and unreasonable." Ibid. He observes that it is inconsistent with the seven-year average of 4.3 percent in expense growth from FY 1993 through his estimated result for FY 2000, and asserts that it cannot be explained by inflationary increases in transportation or employment costs or declining productivity. Id. at 20421-23. He also claims that the 6.8-point difference derived by subtracting the percentage estimate of volume growth from the corresponding percentage for expense growth for FY 2001 is anomalously high in comparison with the seven-year average of about 1.3 points, and could only occur as the result of the "grossest possible mismanagement" of the Postal Service. Id. at 20423-24.

[2116] Witness Bernheimer submits that this result will not occur, and for that reason proposes elimination of the full amount of the proposed contingency provision from the revenue requirement. In practical effect, he observes that the reduced revenue requirement will still represent an amount approximately 5.3 percent higher than his projection of total expenses for FY 2000, the highest rate of increase in the past eight years. At the same time, he notes, this increase would be occurring against the background of a 1.2 percent volume growth projection, the lowest rate of increase during the same period. Even with an additional reduction of $600 million, he notes that the resulting revenue requirement of $67.3 billion would incorporate an increase from estimated FY 2000 costs that matches the average of 4.3 percent for the past eight years. Id. at 20424-25.

[2117] Witness Bernheimer concludes with projections of Postal Service financial results assuming the level of expense he projects for FY 2000 and introduction of rates based on a revenue requirement that does not include a contingency provision. Assuming an average rate increase of 4.6 percent at a point one-third of the way through FY 200l, he projects a loss of $166 million for FY 2000, a profit of $331 million in FY 2001, and a loss of $309 million in FY 2002. Id. at 20426. For the test year, with an average expense increase of 4.3 percent and average unit volume growth of 2 percent, he projects that breakeven could be achieved with an average rate increase of 2.6 percent. Id. at 20427.

[2118] Supplemental testimony of witness Siwek. AAP witness Stephen E. Siwek also addresses the contingency provision in supplemental testimony. Observing that the inflation projections and other data submitted by the Postal Service in response to Order No. 1294 are based on information one year closer to the forecasted test year, witness Siwek expects the accuracy of the Service's forecasts to have improved. He also discounts witness Patelunas' statement that accomplishing the cost reductions related to breakthrough productivity programs will be challenging and involve a higher degree of risk, noting that witness Patelunas has no personal knowledge of risks associated with these programs. Siwek asserts that the overall reduction in risk resulting from updating all cost projections overshadows any greater risk associated solely with the cost reduction programs. Accordingly, should FY 1999 cost data be used in this case, he submits that the contingency must be reduced. Tr. 38/17096-100.

[2119] Rebuttal testimonies of witnesses Strasser and Zarnowitz. On rebuttal, the Postal Service defends its proposed contingency provision in the testimonies of witnesses Strasser and Zarnowitz. Witness Strasser responds to the parties' counter-proposals generally; witness Zarnowitz offers testimony on the future financial perils against which the contingency provision is intended to insure.

[2120] Testimony of witness Strasser. Witness Richard J. Strasser, Jr., Acting Chief Financial Officer and Executive Vice President of the Postal Service, devotes the majority of his rebuttal testimony to defending the proposed 2.5 percent contingency provision against the counterproposals summarized above. In general, he testifies that the proposed provision is a sound product of the Service's judgmental assessment of a variety of factors, including its expected financial condition, historical experience, the potential for unknown future adversities, and the financial, operational, and ratemaking policies established by the Board of Governors. By contrast, witness Strasser asserts, intervenors who argue that a contingency must be justified on the basis of variance analyses and other empirical information would in effect substitute their own judgments to establish a lesser allowance. Tr. 46A/20182-83.

[2121] He also states that the proposed 2.5 percent allowance falls well within a range of reasonableness established in prior proceedings, in which provisions from 1.0 percent to 5.0 percent have been recommended. Id. at 20183-84. Witness Strasser explains the proposed increase from the 2.0 and 1.0 provisions included in the Service's requests in Docket Nos. R94-1 and R97-1, respectively, as a return to a "more normal, but still modest level" justified by changed circumstances: management challenges in Fiscal Year 2000, accelerating inflation, and greater uncertainty regarding labor costs because of contract expirations during the test year. Id. at 20184-86.

[2122] Witness Strasser challenges each of the arguments on which witness Buc bases his proposal of a 1 percent contingency provision. He denies that the Service's proposal lacks an appropriate framework based on quantitative measures, claiming that the Commission's criteria recognize a combination of subjective and objective judgment without necessarily relying on quantitative methods. Id. at 20186-88. Witness Strasser also denies that the Postal Service's improved equity position, which witness Buc says could be further improved by better management of its real estate holdings, provides any basis for lowering the contingency. He asserts that there is no connection between the status of equity restoration and the Postal Service's vulnerability to unknown future adversities or shortfalls, and that witness Buc's claim regarding gains from improved real estate management is speculative. Id. at 20188-89. Witness Strasser also challenges witness Buc's testimony that the state of the economy supports a lower contingency provision, citing indications of increased inflation in DRI indices provided in Exhibit USPS-ST-44AB, witness Tayman's testimony regarding increasing competitive pressures, and the rebuttal testimony of witness Zarnowitz. Id. at 20189-91.

[2123] Witness Strasser criticizes the testimony of OCA witness Burns, denying his assertion that the Service has failed to articulate a reasonable basis for its subjective judgment regarding choice of a contingency provision and challenging witness Burns' analogy to the insurance industry. Unlike insurance industry reserves, witness Strasser testifies, the Postal Service's contingency allowance is designed to protect against both "known unknowns"-such as volume erosion due to the Internet or future legislation-and totally unknown adverse events. Furthermore, he states, the contingency has an important policy dimension, in that "it represents the level of risk that postal management is prepared to accept in directing the Postal Service's operations and finances." Id. at 20192-94.

[2124] Witness Strasser also challenges OCA witness Rosenberg's analysis in support of a 1 percent contingency provision, stating that the Commission has not established firm, objective guidelines conforming to his and witness Burns' "formalistic prescriptions for justifying the contingency." Id. at 20194. Furthermore, witness Strasser claims that an alternative analysis grouping the data used by witness Rosenberg more rationally supports the Service's proposed 2.5 percent contingency, rather than the lesser amount proposed by OCA. Ibid.

[2125] According to witness Strasser, witness Rosenberg's contention that the current economy is operating in a climate of relatively low inflation is based on a defective analysis purporting to show that inflation has both trended lower and become less erratic in recent years. First, witness Strasser notes that it relies totally on historical inflation data, rather than the forecasted data he asserts are more relevant to the test year. Second, witness Rosenberg's analysis focuses on the Consumer Price Index, not the Employment Cost Index witness Strasser claims is more relevant to the mix of services and goods used by the Postal Service. Third, witness Strasser argues that witness Rosenberg's grouping of inflation and contingency data into five-year intervals is arbitrary and masks the true relationships between inflation and contingency amounts. Id. at 20195.

[2126] Witness Strasser presents a table which differs from witness Rosenberg's by grouping data by rate case test year, and by the years feeding into each test year; by including all inflation data in this Docket, rather than data through the end of 1999; and by displaying the respective ECI index for each period, in addition to CPI-W. Id. at 20196, Table 1. According to witness Strasser, the increases in inflation indices since the Docket No. R97-1 test year support, rather than refute, the proposed 2.5 percent contingency, and taken alone could support an even higher contingency. Id. at 20196-97. Witness Strasser also cites the recent surge in fuel prices reflected in Exhibit USPS-RT-1A as support for a higher contingency, as the Postal Service has no mechanism for imposing a price surcharge rapidly, unlike private competitors. Id. at 20197.

[2127] Witness Strasser also disputes witness Rosenberg's assertion that the Service's recent string of positive net incomes supports a low contingency, stating a concern about the declining trend in the Service's net incomes that has developed despite recent financial successes and favorable economic conditions. He notes that the Service's response to Order No. 1294 estimates a net loss for FY 2000 of $325 million, and-notwithstanding cost decreases expected from breakthrough productivity initiatives and increased revenues due to revenue generation initiatives-a test year after-rates deficiency of $475 million. Witness Strasser also states that a favorable economy has not translated into strong volume and revenue growth for the Service, and may create perils in the form of additional pressure on postal wages, higher costs of borrowing, and the increased possibility of an economic slowdown. Id. at 20198-99. Given this high level of uncertainty, he asserts that it would be unreasonable for the contingency provision to be any lower than 2.5 percent. Id. at 20197-98.

[2128] According to witness Strasser, witness Rosenberg's advice that the Service should not use the contingency provision to restore equity is misplaced as a policy matter, inasmuch as the Service's equity is currently negative to the extent of almost $3 billion. More immediately, witness Strasser testifies that it is possible that most if not all of the contingency allowance will be consumed in FY 2001. He explains that rates will likely not be implemented until after the high-volume and -revenue Fall and Holiday mailing seasons, and that much of the remainder will be eliminated by the additional $651.5 in estimated test year costs identified in response to Order No. 1294. Other erosions could occur if volume growth continues to slow, breakthrough productivity cost savings are not realized, there is a shortfall in new revenue generation initiatives, or adverse legislation is passed. Id. at 20199.

[2129] Finally, witness Strasser denies witness Rosenberg's claim that the "safety nets" available to the Postal Service in the form of borrowing authority, the mechanism for recovery of prior years' losses, management's ability to control expenses, and the ability to request new rates on the basis of prospective revenue and expense estimates, should reduce the need for a contingency, as all of these factors are considered in arriving at a contingency, and none is intended to protect against incurring a loss as a result of unknown adverse event or errors in estimation. As an example, witness Strasser cites postal management's limited ability to control workhours in the face of the increase in the number of new delivery points. Id. at 20200.

[2130] Testimony of witness Zarnowitz. Dr. Victor Zarnowitz presents testimony to rebut statements by witnesses Buc, Burns, Rosenberg, and Stapert to the effect that economic conditions will continue to be stable, and that inflation will continue to be relatively low and predictable over the projected rate cycle in this proceeding. In general, he testifies that, although the U.S. economy has benefited from benevolent economic conditions since the mid-1990s, there has been a gradual increase in imbalances and risks that has accelerated in the past and current year, resulting in a much higher level of uncertainty about the direction of the economy. Tr. 41/18190.

[2131] Witness Zarnowitz testifies that there are signs of a slowdown in current U.S. economic activity and leading indicators. First, he states that the comparatively sluggish growth of the early 1990s reduces the claim that a new pattern of noninflationary growth and noncyclical prosperity is firmly entrenched, and he underscores the continued relevance of the business cycle. Second, while there is no sign yet of a slowdown, he finds new evidence of declines in the growth of consumption and employment in the second quarter of 2000, and increases in business investment and government expenditures are likely to prove temporary. Third, and most significantly, he finds warning signs of a slowdown in the Composite Index of Leading Economic Indicators (LEI), particularly in the financial sector, the high-plateau level of the U.S. Index of Lagging Indicators, and testimony of Federal Reserve Chairman Greenspan on July 20 stating an expectation of a leveling out of demand and a dampening of the "wealth effect" that has been driving consumer spending. Id. at 18190-94.

[2132] Witness Zarnowitz also cautions that the remarkable combination of low unemployment and low inflation of the recent past may not last indefinitely. He observes that recent declines in U.S. inflation made possible through lower import prices must be expected to decrease with the improvement of the economic climate abroad and the sharp increase in oil prices. In addition, he states that the containment of inflation through a coincidence of favorable "supply shocks" that have depressed prices cannot be comfortably projected into the future. Also, notwithstanding the low level of increase in "core" inflation as measured by the CPI, witness Zarnowitz anticipates that action by the Federal Reserve Board will have the effect of slowing the economy, and that a "soft landing" is by no means guaranteed. Id. at 18194-97.

[2133] Turning to trends in cost and productivity, witness Zarnowitz observes that despite variable gains in hourly wages from 1991 through 1998, the unusually sharp rise in the ECI in 1999-2000 supports concerns that the labor market may yet tighten sufficiently to force wage increases, leading to more price inflation or a squeeze on profits. He also notes that after variable rates of growth in most of the 1990s, productivity has stabilized and risen since 1997, although at rates that may look surprisingly moderate to the new technology enthusiasts. While profit variables have declined only mildly so far, witness Zarnowitz opines that intensified demand for wage and salary raises may squeeze profits sufficiently to produce a major slowdown. Id. at 18197-200.

[2134] According to witness Zarnowitz, the exceptionally strong but increasingly volatile equity markets are viewed as a bubble about to burst by some prominent finance scholars. Should market prices decline in order to revaluate stocks to reflect more realistic profit assumptions, he cites the expectation of one observer that the Federal Reserve Board does not have the power through interest rate changes to prevent the onset of recession. Even if the apparent overvaluation of equities is justified by a new economic paradigm in which computers and other high-productivity capital goods are substituted for labor, he notes that many seasoned observers predict a tighter monetary policy and higher interest rates to counter the "wealth effect" on consumption demand. Id. at 18200-02.

[2135] Witness Zarnowitz also identifies three other sources of uncertainty regarding the future of the economy. Notwithstanding the Federal Reserve Board's actions intended to keep inflation at moderate levels, he notes the sustained growth in gross domestic product, heavy investment by businesses in new equipment and software, and the continued widening of the trade deficit, which lead to uncertainties regarding further increases in interest rates. Id. at 18202-05. Witness Zarnowitz also testifies that after strong rates of growth in the 1995-98 period, some measures of the money supply have exhibited a low growth rate since that time, and this change could well contribute to slowing the pace of economic activity. Id. at 18205-06. He also sees potential perils in the low rate of personal saving, the great increase in private borrowing (including risky and expensive margin accounts), the huge level of private debt (at present more than 50 percent greater than GDP and increasing), and the high level of foreign borrowing associated with the burgeoning trade deficit. Id. at 18206-08.

[2136] In summary, witness Zarnowitz testifies that the gradual increase in the imbalances that tend to accompany economic booms has greatly accelerated during the past and current years, increasing the risks of a slowdown, higher inflation, higher interest rates, and possibly destabilization of the stock market. Hence he concludes that there is more uncertainty now about forecasts of the economy in the years ahead, including projections of the Postal Service. For this reason, he opines that the Service will generally need more protection of insurance against unexpected adverse events than it has in recent years. Id. at 18212-13.

[2137] Arguments of the parties. On brief, a number of participants argue for reduction of the 2.5 percent contingency provision included in the Postal Service's request. As part of its argument in favor of reducing the overall revenue requirement by at least $1.3 billion, the DMA-led Consortium argues that the Commission bears a legal responsibility to approve a contingency provision no larger than what it determines to be reasonable; that a reasonable provision must be based on substantial evidence and should cover only unforeseen expenses and forecasting errors; that the Postal Service has not justified its proposed 2.5 percent allowance; and that the Commission should approve a contingency provision no greater than 1 percent. Joint DMA et al. Brief at 2-20. If the Commission projects test year costs using update estimates, the Consortium argues that the contingency provision should not exceed 0.25 percent. Id. at 20-21.⁵ The Association of American Publishers,⁶ Coalition of Religious Press Associations,⁷ Greeting Card Association and Hallmark Cards, Inc.,⁸ the Commission's Office of Consumer Advocate,⁹ Parcel Shippers Association,¹⁰ and Val-Pak Direct Marketing Systems and Carol Wright Promotions¹¹ also advocate the use of a contingency provision significantly below the 2.5 percent incorporated in the Postal Service request.

[2138] In its brief, the Postal Service argues that its proposed contingency provision is supported by substantial record evidence-primarily the testimonies of witnesses Tayman, Strasser, and Zarnowitz-and is consistent with the levels of contingency provisions recommended in earlier Commission rate decisions. Postal Service Brief at II-2-II-8. The Service claims that participants opposing its contingency provision rely on "misapprehensions of fact or judgment" and ask the Commission to substitute their or the Commission's judgment for that of postal management. Id. at II-9-II-10.

[2139] The Postal Service's position finds one defender on brief. United Parcel Service argues that the proposed 2.5 percent contingency should be approved to prevent the possibility that services bearing a low cost coverage may fall below attributable cost or fail to contribute their fair share to institutional costs in the event of cost increases beyond those forecast. UPS Brief at 92.

[2140] In its reply brief, the Postal Service presents extensive and detailed arguments regarding the respective spheres of authority of the Board of Governors and of the Commission; the mechanism by which the level of the contingency provision is properly determined; and the soundness of participants' arguments as a possible basis for recommending a lesser amount than the contingency allowance included in the Service's request. Postal Service Reply Brief at II-1-II-39.

[2141] First, the Service argues that the Commission's authority to adjust the revenue requirement is severely limited, and that determination of a reasonable provision for contingencies lies outside that limited authority. The Postal Service acknowledges that the Commission has never shared the Service's views on the limits of PRC authority over the revenue requirement generally, and has consistently disagreed that the contingency provision is outside the purview of Commission review in rate proceedings. Nonetheless, the Service reviews some of the respective institutional declarations on the subject of authority over the revenue requirement, and cites judicial authority bearing on the division of institutional responsibilities. Id. at II-3-II-14.

[2142] Second, the Service argues that the contingency provision is properly established through "a basic policy choice subjectively determined by the Postal Service and its Board of Governors[,]" as "an important element of the Postal Service's financial policies." Id. at II-14-ll-15. Thus, the Service argues further, the contingency is not a cost estimate, a cumulation of cost estimates, or a variance around cost estimates, unlike other components of the revenue requirement that can be determined empirically. On this basis, the Service argues that the analytical approach taken by DMA and other proponents of a lower contingency provision does not allow a reasonable contingency to be developed, let alone quantified as equaling 1 percent or one quarter of 1 percent. Id. at II-14-II-20.

[2143] Finally, while not conceding as a legal matter that the Commission's views or precedents are an authoritative interpretation of what is "reasonable," the Service argues that the evidentiary bases on which its contingency provision rests are sound; that a new determination of reasonability must be made in the current circumstances of each case; and that there is no lawful precedent on which to base the requested reduction of the contingency provision. Id. at II-20-II-31. The Service also argues that neither its improved financial condition nor the state of the national economy justifies reducing the contingency; that the Service has not ignored historical variance analysis; that the contingency should not and may be reduced to encourage postal managers to control costs; and that an inadequate contingency in the face of increased uncertainty is likely to increase the need for future ratepayers to continue to subsidize current mailers through recovery of prior years' losses. Id. at II-38.

[2144] Bases of Commission recommendation. The Commission acknowledges the primacy of the Governors' authority to assure that rates and fees generate sufficient revenues to enable the Postal Service to perform its public mission. This authority is exercised both by the Governors' initiation of requests for rate changes under § 3622 and by their actions on the Commission's rate recommendations as provided in § 3625.

[2145] Between these two significant events, the Commission is charged with the responsibility of producing recommendations that are in accordance with the policies of Title 39 and the ratemaking factors prescribed in § 3622(b). Section 3624 commands that formal hearing procedures, including public participation in the making of an evidentiary record, precede and inform the Commission's recommendations.

[2146] In the first rate proceeding conducted under the procedures prescribed by the Reorganization Act, the Commission concluded-contrary to the urgings of the Postal Service at the time-that its statutory responsibilities require that the Service's aggregate cost and revenue estimates be subjected to the same scrutiny as its rate proposals. In analytical terms, the Commission found that its independent review of the revenue requirement is necessary to implement the legislative intent to introduce a system of checks and balances into the ratemaking process. PRC Op. R71-1 at I-268-I-269. More pragmatically, the Commission concluded that:

The purpose of our recommendation is to assist the Governors in performing their statutory functions under § 3625. Ultimately, the Governors may modify the Commission's recommended
decision . . .. But as this proceeding has plainly shown, if the Postal Service is not required to justify its estimates on the record, neither the Commission nor the Governors would be in a position to exercise their authority in a meaningful way.

[2147] The Commission has consistently conducted its rate proceedings in accordance with this general conclusion throughout its institutional history. Regarding the contingency provision particularly, the Commission held to a congruent view, capsuled in the opinion in Docket No. R87-1:

In prior dockets, we have concluded that the subjective element of the contingency determination entitles management's determination to a good measure of deference, but that it does not render that judgment unreviewable. As we noted in Docket No. R84-1, judgment implies opinion or assessment, and is not necessarily equated to management discretion. Because the statutory requirement that a contingency be supported by substantial evidence remains in effect, management must still provide such evidence, and the Commission must still review it.

[2148] In this proceeding, the Postal Service reminds the Commission of its continuing disagreement with this position, and "urge[s] that the Commission not listen to the voices that have influenced it in the past to challenge the Postal Service's revenue requirement, or other matters, on the basis of these fundamental disagreements over statutory authority." Postal Service Reply Brief at II-2-II-3. The Service offers its own interpretation of an extremely attenuated Commission responsibility vis-à-vis the revenue requirement, and cites judicial authority to support its views, including the view of one court that "the PRC must accede to the Board's estimates of the Service's revenue needs." Id. at II-8, quoting Time, Inc. v. U.S. Postal Service, 685 F.2d 760, 775 (2d Cir. 1982).

[2149] Once again, to the extent that the Postal Service is advancing the argument that the estimates of required revenue contained in its Request are immune from inquiry and appraisal on the record, or that the Commission's recommendations must approve them regardless of their record support, the Commission must respectfully agree to continue disagreeing in this area. The Supreme Court's decision in NAGCP IV,¹² also cited by the Postal Service, states:

Although the Postal Reorganization Act divides ratemaking responsibility between two agencies, the legislative history demonstrates "that ratemaking . . . authority [was] vested primarily in [the] Postal Rate Commission." [Citations omitted.] The structure of the Act supports this view. While the Postal Service has final responsibility for guaranteeing that total revenues equal total costs, the Rate Commission determines the proportion of the revenue that should be raised by each class of mail.

462 U.S. at 821. (Emphasis added.) (Footnote omitted.) The Commission's view of its and the Governors' statutory responsibilities regarding the revenue requirement is fully compatible with this declaration. The Governors, in consultation with postal management, decide the magnitude of required revenues to include in Requests, in accordance with § 3621. They also exercise discretion to act on the Commission's recommendations pursuant to § 3625; should they find, after resubmission of the Request, that the rates recommended in the decision on reconsideration will yield insufficient total revenues, they may modify the Commission's recommendations in accordance with the record and the policies of Chapter 36. In the intermediate process that the Commission is directed to conduct, revenue requirement matters are subject to the same substantive, on-the-record review as are other issues, and the Commission will make substantive, but not final, determinations and construct its recommendations accordingly.¹³

[2150] In this case, on brief, the Postal Service suggests that the degree of scrutiny applied to the contingency provision in past rate proceedings has been variable, with a perceived "lowering of the bar" when the Service's contingency proposal has been reduced from previous levels, as in Docket No. R97-1. Postal Service Brief at II-4-II-6. A "double standard" does not to operate in such circumstances. That appearance may result from the Commission's deference to the Governors' assessment of potential risks at that time, and the absence of participants' initiatives to explore the subject on the record. The applicable standard of "reasonability" is the same in all cases: whether, giving due deference to the Postal Service's judgment on the subject, the provision is rationally related to achievement of revenue sufficiency in the period under review.¹⁴ Varying circumstances in different rate proceedings will require different degrees of inquiry and review.

[2151] In this case, the Postal Service supports the 2.5 percent contingency allowance incorporated in its Request with the testimony of witness Tayman, as supplemented by the presentations of witnesses Strasser and Zarnowitz on rebuttal. Other participants have made voluminous presentations on the contingency provision, challenging the Service's proposal on a variety of grounds. In the Commission's view, it is appropriate to consider all material on the record that bears on the reasonability of the contingency provision.

[T]he purpose of the contingency provision set forth in 39 U.S.C. section 3621 is two-fold. First, it provides insurance against the possibility of misestimates of test year accrued revenues and expenses. As we have stated in the past, such variances are inherent in the forecasting process. Second, the provision is intended to protect against unforeseeable events, not capable of being prevented through honest, efficient and economical management, and which might have a significantly adverse impact on the financial position of the Service or upon its operations.

[2153] In presenting the Service's 2.5 percent contingency provision, witness Tayman acknowledges this two-fold purpose in his statement that, "[t]his amount is judged as reasonable against unforeseen events and forecasting errors, given the magnitude of the Postal Service's operations and expenses." USPS-T-9 at 43.

[2154] Witness Tayman reports the results of variance analyses contained in his Exhibit USPS 9-J "[I]n deference to the Commission's desire to evaluate forecast errors and their sources[.]" Id. at 44. However, he disavows them as the basis for determining the need for a contingency or its size, and further opines that relying on them for predictive value "would be both irresponsible and illogical." Id. at 45. "Regardless of what history shows," he asserts, "management must be allowed to assume its responsibility to determine the amount of contingency most appropriate for achieving its goals." Ibid.

[2155] This aspect of witness Tayman's testimony is troubling for several reasons. First, while the Commission has never enshrined variance analysis of historical costs and revenues as the definitive indicator of the appropriate magnitude of the contingency provision, it has recognized its value as an empirical measure of forecasting error and the relative magnitude of unforeseen events.¹⁵ The Commission adheres to its long-established opinion that, as a significant quantitative input to a blend of subjective and objective judgment, consideration of the results of variance analyses is an appropriate component of determining the reasonability of proposed contingency provisions.

[2156] Second, witness Tayman's assertion of the supervening importance of postal management's selection of a contingency amount "most appropriate for achieving its goals" is potentially problematical in view of the Commission's responsibilities. Whatever management goals might be served by the selection of a particular amount for a contingency provision, the Commission's review must be guided by the objective of providing reasonable assurance of revenue sufficiency for the Postal Service in accordance with § 3621.¹⁶ At the least, the Commission and the participants are entitled to a clear identification of the nexus between specific test year management goals and revenue sufficiency, and an explanation of why achieving those goals requires the specific contingency amount requested. While the Postal Service has explained why its operations goals are expected to generate a particular level of expenses, no such presentation has been made in this case regarding the provision for contingencies.

[2157] As is evident from the results witness Tayman reports, the 2.5 percent contingency allowance he defends lies outside the bounds of the variance analyses. If the Postal Service performed any other statistical or quantitative analyses bearing on the amount of the contingency provision, they were not provided by witness Tayman, witness Strasser, or the Postal Service institutionally.¹⁷ Consequently, the Postal Service's proposed contingency provision is not based on any empirical estimate of potential forecasting error.

[2158] Part of the Postal Service's defense of its contingency provision relies on the unknown and unknowable nature of future events with potentially negative impacts.¹⁸ However, the Commission cannot agree that the unknowability of future events puts choice of a contingency provision into the realm of purely subjective judgment. As long ago as Docket No. R77-1, the Commission stated its "view that over the long run the relative magnitude of unforeseen events (variances between estimates and actual results caused by uncontrollable events) will prospectively tend to display a certain degree of predictability, albeit not precise, with historical results." PRC Op. R77-1 at 32. (Footnote omitted.) For this reason, the Commission endorses the position of OCA witness Burns that a contingency provision must be reasonably related to a careful assessment of future uncontrollable events, and confirms the importance of historical analyses of the kind performed by witness Rosenberg to make such an assessment. In the years since R77-1, the Postal Service has become significantly more skillful at projecting its costs and revenues, and it should also have developed an extensive record of the frequencies and impacts of unknowable events such as natural disasters.¹⁹ Whether by variance analyses or some other reasoned, transparent technical exercise, the Commission maintains that examination of historical results is a valid and useful tool for gauging the probable magnitude of unforeseeable future events.

[2159] In addition to the results of the variance analyses, which suggest a comparatively low level of forecasting error based on historical results, the Commission's recognition of updated test year costs projected from actual FY 1999 data further reduces forecasting error. Moving the source of actual data one full year forward in time, while retaining FY 2001 as the forecast target, enhances the predictive accuracy of the forecasts by shortening their span.²⁰ Further, it justifies reevaluating the appropriate size of the contingency provision.²¹ All other things being equal, updating costs in this case should greatly reduce the need for the contingency provision's function of insuring against the possibility of misestimates of test year accrued costs.

[2160] Lacking any additional empirical information for guidance on an appropriate contingency provision, the Commission must evaluate the subjective claims of risk the Postal Service makes in support of its selection of an increased contingency provision. As in past cases, the Commission assesses these subjective claims by examining evidence bearing on the Postal Service's financial condition, the state of the national economy, and other relevant factors.²²

[2161] Beginning with the economic climate in which the Postal Service operates, the short-term outlook for the national economy does not appear to involve any significant risk of unforeseeable financial harm to the Service. OCA witness Rosenberg testifies that the United States continues to enjoy robust growth in the longest economic expansion in over half a century. Tr. 22/9815. Although witness Zarnowitz testifies that he detects what may be precursors of a potential reduction in economic activity, on oral cross-examination by OCA he confirmed that leading indices have occasionally declined during the last five years, but in each instance the index rebounded without turning negative. He agreed with a characterization of the various data collectively as a "mixed picture," and added: "We still are in an expansion that is relatively strong compared to the `60s and `80s." Tr. 41/18286.

[2162] The chief perils witness Zarnowitz and witness Strasser identify concern the potential for cost increases driven by inflation. The Postal Service, the Consortium and OCA argue that different indices of inflation, analytical timeframes, and interpretations should inform the choice of an appropriate contingency provision. However, the Commission's update of test year costs-which employs the more recent indices used by witness Patelunas in his rollforward exercise and CPI values for periods as recent as July-September of this year-render this disagreement moot. By using the most recent information available on the record, the Commission's test year forecasts minimize uncertainty concerning the impact of misestimates of economic activity on Postal Service costs.

[2163] Regarding the financial condition of the Postal Service, the evidence appears to be mixed. The Service enjoyed net revenue through FY 1999, and has continued to improve its equity position. However, it finished FY 2000 with a reported loss; the exact amount is not available on the record, as an audited result has yet to be publicly reported by the Postal Service.²³ Witnesses Tayman and Strasser also identify volume growth as an area of concern. However, in supplemental testimony explaining why the test year volume forecast initially filed by the Postal Service did not need to be updated along with cost estimates to account for actual FY 1999 data, witness Thress cited as one reason ". . . the fact that the initial forecast is performing quite well compared with the most recent actuals. . .." USPS-ST-46, at 1. This suggests a comparatively low level of uncertainty regarding potential forecast error in volume estimates for the test year.

[2164] The greatest potential source of uncertainty concerning the Postal Service's financial results in the test year appears to be ambitious cost reduction programs. Witness Tayman cites the "challenge" of achieving a 1.5 percent workyear reduction in the test year. USPS-T-9 at 44. In addition to other cost reduction programs incorporated in the revenue requirement, the Postal Service's updated costs for the test year reflect $744 million in "breakthrough productivity" cost savings. Without linking an amount to any specific program, the Postal Service's update also reflects a $200 million Field Reserve offset to cost savings, which "recognizes the difficulty in achieving the aggressive cost reductions in FY 2001, the first year of the Breakthrough Productivity Initiative." Tr. 46D/21595, fn. 2. This is a somewhat unusual, but appropriate, example of the kind of insurance against uncertainty the contingency provision provides, and the Commission believes it should be reflected in the amount of that item.

[2165] On balance, these considerations support a conclusion that a 2.5 percent contingency allowance is not necessary to assure revenue sufficiency in the test year, and thus is excessive. This conclusion is reinforced by the disproportional share of additional new revenues the proposed contingency provision constitutes.

[2166] As a table in the Brief of the OCA shows, the Postal Service's contingency provision represents 60 percent of aggregate revenue increase requested by the Postal Service in this case. While the Commission has directed its scrutiny primarily to the percentage value of the contingency provision in past proceedings,²⁴ the Service's proposed allowance represents a majority of the total requested revenue increase, an unprecedented proportion of revenue burden to distribute to the classes of service without attribution on the basis of cause. As Table 2-1 shows, since Docket No. R76-1 the proportion has rarely exceeded 30 percent.

Table 2-1

Comparing the Postal Service's R2000-1
Contingency Proposal with Prior Rate Cases
Case	Requested Increase in Revenue	Amount of Contingency	Contingency/ Revenue Increase

R76-1	$2.31 billion 1	$0.54 billion 2	23%
R77-1	$1.97 billion 3	$0.66 billion 4	34%
R80-1	$3.75 billion 5	$0.54 billion 6	14%
R84-1	$3.11 billion 7	$1.0 billion 8	32%
R87-1	$4.3 billion 9	$1.3 billion 10	30%
R90-1	$6.16 billion 11	$1.6 billion 11	25%
R94-1	$4.11 billion 13	$1.05 billion 12	26%
R97-1	$2.24 billion 15	$0.6 billion 14	27%
R2000-1 (2.5 %)	$2.8 billion 17	$1.68 billion 16	60%
R2000-1 (1 %)	$1.7815 billion 19	$0.6719 billion 20	38%
Average proportion of contingency to revenue increase for the eight omnibus rate cases is 26.4%.
Source: OCA Brief at 41, Table 1.

1 USPS Request, filed December 19, 1975. 3 USPS Request, filed July 13, 1977. 5 USPS Request, filed April 21, 1980. 7 USPS Request, filed November 10, 1983. 9 USPS Request, filed May 10, 1987. 11 USPS Request, filed March 6, 1990. 13 USPS Request, filed March 8, 1994. 15 USPS Request, filed July 10, 1997. 17 USPS Request, filed January 12, 2000, at 2.		2 PRC Op. R76-1 at 19. 4 PRC Op. R77-1 at 42. 6 PRC Op. R80-1, para 0138. 8 PRC Op. R84-1, Appendix A.9, 10. 10 PRC Op. R87-1, Appendix A. 12 PRC Op. R90-1, Appendix A. 14 PRC Op. R94-1, Appendix A. 16 PRC Op. R97-1, Appendix C. 18 Exh. USPS 9A. 20 Ibid.
19 The Postal Service's proposed contingency of $1.6798 billion (Exh. USPS-9A) is $1.0069 billion higher than witness Rosenberg's proposed contingency of $0.6719 billion (OCA-RT-2 at 4; Tr. 41/18303). If a commensurate $1.0069 billion reduction is made in the Postal Service's requested total increase in revenue, i.e., $2.7884 billion (Postal Service Request at 2), then the total increase in revenue would be approximately $1.7815 billion. 0.6719/1.7815 = 37.7%.

[2167] OCA's point is that the appropriate size of the contingency should be related to the size of the requested increase in revenues. The corollary is that large forecast error is more likely when projecting large changes (for example in a period of rapid inflation) than when projecting small changes (for example during stable economic times.) The Postal Service projects small, gradual increases in operating expenses that will result in a test year deficiency in operating revenues of approximately 2.6 percent assuming no rate increase. OCA contends that as economic conditions have been shown to be stable, OCA Brief at 43-51; and other causes of projection forecast error have been dramatically reduced, Id. at 53-55; there is no valid justification for a sharp increase in the size of the provision for contingencies.

[2168] Table 2-2 breaks out the uses that will be made of the revenues generated by the rate increases proposed by the Service and recommended by the Commission. The Request seeks $2,788 million in additional revenue, only 30 percent of which, $840 million, would be used to offset increases in test year operating expenses. Twice that amount, $1,680, is sought as a cushion against forecasting errors and unknown events.

Table 2-2
Allocation of Revenues from Rate Increase
	USPS Request $ Millions	PRC $ Millions

INCREASE IN REVENUE	2788 ¹	2508 ²
$ Allocated for Contingency	1680 ³	1012 ⁴
$ Allocated for PYL	268 ⁵	312 ⁶
Remainder to Offset Operations Expenses	840	1184

Request at 2.

Appendix C Total Revenues less TYBR revenues from Request at 2.

USPS-T-9 at 43.

⁴

Appendix C.

⁵

USPS-T-9 at 46.

⁶

Appendix C.

[2169] The voluminous materials provided by the Postal Service with its Request largely focus on the Service's operations and operating expenses. Initially, only two pages of fairly general statements provided by witness Tayman, USPS-T-9 at 43-44, are offered to support the $1,680 million additional revenues sought by the Service for contingencies. This discussion focuses primarily on potential causes of estimation error, and provides little justification for a larger cushion for unknown events. Even as supplemented by witnesses Patelunas and Strasser, there is no explanation of how much and why the contingency for unknown events is increased from the R97-1 level.

[2170] The magnitude of this amount is difficult to reconcile with the Postal Service's witnesses' expressions of concern regarding the increasingly competitive environment in which it operates. It is not the Commission's function to direct how the Postal Service should respond to competitive pressures. However, for ratemaking purposes it is difficult to interpret a perceived increase in the intensity of competition as a justification for increasing an item that will raise all rates in the aggregate.

[2171] The Postal Service has not justified a contingency provision of this magnitude on the basis of revenue need in the test year. Nor, in the Commission's opinion, would it be appropriate to defend a contingency allowance of this size on the ground that excess revenues will improve the equity position of the Postal Service, as witness Strasser suggests Tr. 46A/20199. This function is performed retrospectively by the provision for the recovery of prior years' losses, and is not a legitimate purpose of the contingency provision. Any additional revenue realized in advance of the recovery schedule through the contingency provision would constitute a test year profit, which has never been a legitimate objective of ratemaking under the Reorganization Act.

[2172] As noted above, participants have argued that the Commission should respond to the excessive contingency allowance included in the Postal Service's Request by recommending the 1 percent provision adopted in Docket No. R97-1, a fraction of 1 percent, or a zero contingency provision. Notwithstanding the reduction in potential forecasting error resulting from updating costs, the Commission finds that uncertainties surrounding the Postal Service's achievement of ambitious financial goals in the test year require a contingency provision appreciably greater than one quarter of 1 percent.

[2173] While their analytical approaches differ, several participants argue with merit that a 1 percent contingency should provide an adequate cushion to assure revenue sufficiency in the test year. OCA makes a particularly convincing argument, based on the testimonies of witness Rosenberg, that a 1 percent contingency should be sufficient during this period of economic stability and relatively low inflation, in light of 30 years' experience of ratemaking under the Reorganization Act and of the use of updated forecasts.²⁵

[2174] Nevertheless, the Commission recommends the incorporation of a 1.5 percent contingency allowance in the revenue requirement. In the Commission's opinion, this appropriately reflects the decrease in potential forecasting error resulting from use of updated costs, counterbalanced by consideration of the challenges to achievement of the Postal Service's financial goals in the test year.

[2175] In the Commission's view, a 1.5 percent contingency provision-which has a dollar value of $1,012 million on a test year after-rates basis-is near the outer boundary of reasonability in this case. As a gauge of this limit, the Commission notes that the total amount for all test year cost reduction programs incorporated in the revenue requirement is $1.1 billion.

[2176] As described in the previous section, II. C., the updated cost and revenue information provided by the Postal Service while this case was in progress indicate that operating expenses in the test year will be higher than initially projected. Additionally, at the rates the Commission recommends, volume losses will be reduced, and the retained volumes will add costs to the system. The Service also requests slightly more revenue for the recovery of prior years losses. These adjustments are reflected in Table 2-2. Nonetheless, 40 percent of the revenues generated by the rate increases recommended by the Commission will go to funding the provision for contingencies. This is the largest ever proportion of new revenues provided to fund a provision for contingencies.

Postal Service Request, Attachment C, Rule: 54(f)(2). Under the terms of § 54(f)(2), which prescribes use of a test year "beginning not more than 24 months subsequent to the filing date of the formal request[,]" the Service alternatively could have selected Fiscal Year 2002 at the test period.

⁴

Witness William A. Morrow testifies that unique factual circumstances in this case justify application of a zero contingency markup for Periodicals. Witness Morrow's proposal is addressed in the discussion of rate design for Periodicals, Chapter V. D.

¹³

Other authorities cited by DMA et al. in their analysis of the Commission's legal responsibility to review the contingency provision provide additional support for this position. DMA et al. Brief at 3-6.

¹⁴

In its opinion in Docket No. R84-1, the Commission stated: "In essence, a reasonable contingency provision should, then, better enable the Service to comport with the break-even requirement of § 3621 which mandates that, as nearly as possible, costs equal revenues plus appropriations." PRC Op. R84-1, para. 1017. (Footnote omitted.)

¹⁵

In the opinion in Docket No. R87-1, the Commission stated: "The Commission has never advocated that statistical analysis be the exclusive determinant of the proper contingency amount, nor that it should be accepted uncritically, in terms of its precision, or its ability to account for external factors. We maintain our view, however, forecasting errors have sources, and that much can be learned by systematically evaluating the behavior of those sources over time. We also adhere to our view expressed in Docket No. R77-1 that the relative magnitude of unforeseen events, including external events, over the long run will tend to display a degree of predictability, based upon historical results." PRC Op. R87-1, para. 2077. (Citations omitted.)

¹⁶

Witness Tayman's statement is a verbatim repetition of his testimony in Docket No. R97-1, suggesting that this position has become a rubric of postal management policy on the subject. Restatement of this position provides no substantive support or assistance to the Commission's analysis of the revenue needs of the Postal Service.

¹⁷

As the OCA observes on brief, the Postal Service objected to providing responses to interrogatories from both OCA and DMA et al. requesting analyses and other documents relating to its contingency proposal, claiming that responsive materials were predecisional and therefore protected by the deliberative process privilege. In light of this resistance, both OCA and the Consortium argue that the record is "utterly devoid of any evidence of a `systematic analysis' that may or may not have been made." OCA Brief at 66; DMA et al. Brief at 9. It is worth noting that any analyses that were actually relied on to arrive at the chosen 2.5 percent figure would no longer have been privileged, and therefore should have been provided.

²⁰

As OCA witness Rosenberg stated in his supplemental testimony, "the closer the Postal Service's estimates are to the forecasted period, the more accurate its forecasts are likely to be." Tr. 41/18308. On oral cross-examination by counsel for DMA, Postal Service witness Zarnowitz confirmed the superior accuracy of one-year forecasts over two-year forecasts, and stated that over a two-year span forecasts "decline within this period in accuracy, quarter-by-quarter." Id. at 18234.

²¹

"We agree with witness Quick that the size of the contingency should be reevaluated as the Test Year nears, if there has been a major, net reduction of forecasting error." PRC Op. R87-1, para. 2095.

²³

Witness Patelunas uses $325 million as an assumed loss result in his rollforward exercise, and the Commission's update preserves this assumption. To the extent the audited result for FY 2000 shows a smaller loss for the period, the Commission's test year revenue requirement will be higher than necessary. It is well to keep in mind that misestimation and unforeseen events may benefit the Postal Service.

²⁵

Witness Strasser presented a table purporting to show that inflationary trends support a 2.5 percent contingency provision. Tr. 46A/20196, Table 1. However, the table presents data only for fiscal years associated with the past four rate proceedings, and the relevance of the ECI data it presents to the choice of a contingency provision is questionable, as this index is already a component of the forecasts with which test year labor costs are estimated.