X. Batch Hot Cut Costing and Pricing (47 C.F.R. § 51.319(d)(2)(ii)(A)(4))

A. Framework for Analysis

In conjunction with the requirement for states to approve and implement a batch cut process, the FCC also directed state commissions to adopt Total Element Long Run Incremental Cost (TELRIC)-based rates for the approved batch cut activities. The FCC defines TELRIC rates as "forward-looking" and based on the use of the "most efficient telecommunications technology currently available and the lowest cost network configuration, given the existing location of the incumbent LEC's wire centers."¹⁴⁰ As noted by the FCC, "[t]hese rates shall reflect the efficiencies associated with batched migration of loops to a requesting telecommunications carrier's switch, either through a reduced per-line rate or through volume discounts as appropriate."¹⁴¹

The ILECs mailed testimony on January 7, 2004, presenting TELRIC pricing proposals for their BHC processes on January 28, 2004; other parties mailed reply testimony on this issue.

1. SBC Hot Cut Prices Position of SBC

SBC proposed a batch cut rate structure on a per line basis, with pricing variations according to time of day and/or other cost drivers. SBC's pricing proposal is based on its cost study reflecting non-recurring costs.¹⁴² SBC represents that its pricing proposal complies with FCC's TELRIC methodology set forth in its First Report and Order, CC Docket No. 96-98, FCC 96-325, adopted August 1, 1996, and also complies with the Commission's Consensus Costing Principles.¹⁴³ SBC cost witness Pearsons applied the Commission-approved shared and common cost allocation factor of 21% to its estimated TELRIC to derive SBC-proposed prices.¹⁴⁴ SBC developed flat, per line rates, reflecting the following factors:

· Process used (Enhanced Daily Process, Defined Batch Process, or Bulk Project offering)

· Type of hot cut requested (FDT, CHC or IDLC)

· Time of cut (normal business hours, expanded hours, or premium hours)¹⁴⁵

· SBC's proposed rates for the Enhanced Daily Process are based on the type of hot cut requested, as follows: Enhanced Daily Process - FDT Basic Option $14.70

· Enhanced Daily Process - CHC Basic Option $20.73

· Enhanced Daily Process - IDLC Basic Option $79.09

Since all cuts in the Enhanced Daily Process occur during normal business hours, no additional rate elements for out-of-hours conversions apply. Requests for hot cuts outside of normal business hours, reflected below under the "Expanded" option, are subject to minimum volume requirements as described in SBC's Batch Cut Proposal.

SBC's proposed rates for the Defined Batch Process are based on the type of hot cut requested and the time requested for the cut:

· Defined Batch Process - FDT Option

- Basic (M-F, 8:00 am - 5:00 pm) $10.61

- Expanded (M-F, 6:00 am - 8:00 am) $10.74

· Defined Batch Process - CHC Option

- Basic (M-F, 8:00 am - 5:00 pm) $12.70

- Expanded (M-F 6 AM - 8:00 am, 5:00 pm -
12:00 am, Sat 8:00 am - 5:00 pm) $12.75

· Defined Batch Process - IDLC Option

- Basic (M-F, 8:00 am - 5:00 pm) $77.35

SBC's proposed rates for the Bulk Project offering are based on the type of hot cut requested and the time requested for the cut. The Bulk Project Offering includes additional rates for out-of-hours and premium cut time options.

Bulk Project Offering - FDT Option

- Basic (M-F, 8:00 am - 5:00 pm) $10.58

- Expanded (M-F 6:00 am - 8:00 am,
5:00 pm - 12:00 am, Sat 8:00 am - 12:00 pm) $10.71




- Premium (M-F 12:00 am - 6:00 am,
Sat 12:00 am- 8:00 am)¹⁴⁶ $11.31

Bulk Project Offering - CHC Option

- Basic (M-F, 8:00 am - 5:00 pm) $12.67

- Expanded (M-F 6:00 am - 8:00 am,
5:00 pm - 12:00 am, Sat 8:00 am - 12:00 pm) $12.72




- Premium (M-F 12:00 am - 6:00 am,
Sat 12:00 am - 8:00 am)¹⁴⁷ $13.56

Bulk Project Offering - IDLC Option

- Basic (M-F, 8:00 am - 5:00 pm) $77.33

SBC's cost witnesses¹⁴⁸ defend the reasonableness of the activities, job titles, estimated times and probabilities of occurrences with respect to the local operations center (LOC), the central office (LFO-In) and the outside plant (LFO-Out) work groups in California included in the "Bill of Costs" Tab of the Cost Study in Attachment 1 to Pearsons' testimony. The "Bill of Costs" tab details each of the items associated with any CHC, FDT or IDLC hot cut ordered under the Enhanced Daily, Defined Batch and Bulk Project Processes. SBC claims that its proposed costs reflect a reduction from about 50% to 85% off the prices that CLECs currently pay for hot cuts.

The basic tasks in a hot cut, as previously discussed, involve administrative processes, running a jumper from the CFA to the frame, pre-testing the dial tone and phone number, performing the "lift and lay" and a final dial tone and phone number verification.

Two additional tasks are performed by the LOC when a line is migrated from an IDLC to a copper facility. The first task requires the LOC to contact the LFO-Out organization to confirm the appointment time on the order prior to the dispatch of the technician. The second task requires the LOC to receive a call from the technician when the cutover from IDLC to copper is ready to begin.

To compute TELRIC for each applicable activity required to provision the service/UNE, the job title performing the work and the labor rate associated with that job title was identified. The appropriate labor rate was multiplied by the time required to perform the activity, and then multiplied by the Work Group Occurrence Factor (WGOF)¹⁴⁹ and the Activity Occurrence Factor (AOF).¹⁵⁰

An experienced subject matter expert (SME) representing each workgroup identified the activities and the job titles required to provision each rate element identified within the non-recurring cost study. Each workgroup SME also provided activity times and percent occurrences associated with each activity based on their knowledge of the Batch Hot Cut Process to be implemented and how long it takes to perform similar activities today.

2. Response of MCI and AT&T to SBC Cost Study

AT&T and MCI each performed separate analyses of SBC's costs. AT&T and MCI claim that the SBC cost study fails to comply with the FCC's TELRIC rules, particularly Rule §51.505, Part (1) which requires the use of an efficient network configuration.

MCI argues that the batch hot cut cost study submitted by SBC does not assume a properly mechanized system, includes costs which are more appropriately recovered from SBC retail customers, and includes exaggerated work times and unnecessary work steps. MCI adjusted SBC's cost study for alleged errors relating to work steps, task times and costs associated with removing the retail customer's service from the SBC network, and applied the Commission-approved non-recurring fallout rate to generate TELRIC-compliant costs for a hot cut.¹⁵¹ The rates that MCI calculated, as presented in the testimony of witness Starkey are listed in the table below.¹⁵²



MCI recommends that the Commission adopt the rates in the table above for the following hot cut migration scenarios (voice only, UNE-L to UNE-L, line sharing, and line spitting) even though SBC has excluded the latter three from its batch hot cut proposal.¹⁵³ For migrations involving hot cuts of loops provisioned via an EEL, MCI recommends a different set of batch hot cut rates to reflect that such migrations could require additional work steps beyond those incorporated by SBC.¹⁵⁴ During the pre-wiring phase of a "cut-to-EEL" scenario, SBC would need to ready the interoffice DS0 circuit connecting the CLEC's distant collocation arrangement to the central office within which the cut will take place. MCI believes this is likely to take additional time beyond that required for the other hot cut scenarios.¹⁵⁵ Until more information is available, MCI witness Starkey assumed that these activities will require twice the amount of time SBC has identified for pre-wiring other types of migrations.¹⁵⁶ Thus, MCI recommends the following rates for a "cut-to-EEL" scenario.



AT&T likewise adjusted SBC's cost study, and recast SBC's cost and pricing tables, to reflect the adjustments of AT&T witness Turner. AT&T likewise claims that SBC's proposed rates are not TELRIC-compliant, and argues that SBC should be required to submit a TELRIC compliant model. AT&T witness Turner also presented an alternative pricing proposal that he claims still is not low enough to overcome impairment, but is closer to TELRIC compared to SBC's proposal. Turner's proposed adjustments to SBC's cost study are set forth in the matrix table in Attachment SET-1 to Turner's Reply Testimony (Ex. 115C). AT&T's recast TELRIC price tables were provided under seal, and are set forth in Ex. 115C (Attachment SET-6; Proprietary).

3. Disposition of Cost Adjustments for SBC TELRIC Prices

We review below the proposed MCI and AT&T adjustments to TELRIC, and determine to what extent, if any, the proposed adjustments are warranted. Based on our review of the proposed adjustments to SBC's TELRIC rates, we have computed revised rates for SBC's batch cut processes as set forth in Appendix 1 of this order. We adopt these prices on an interim basis to be applied to any initial offerings of hot cut options made available by SBC pursuant to further orders either of this Commission or the FCC. Parties' proposed adjustments, and our disposition thereof, are set forth below.

a) Incorporation of System Mechanization into Cost Structure

Rule 51 states that the TELRIC of an element should be measured based on the use of the most efficient telecommunications technology currently available and the lowest cost network configuration, given the existing location of the incumbent LEC's wire centers. MCI argues that FCC rules require that if technology currently exists to mechanize the process to perform hot cuts, TELRIC must be calculated as if that technology was being used today. MCI argues that such technology does exist and, as a result, SBC and Verizon's failure to account for it in their cost studies is inconsistent with the FCC's pricing rules. The TRO instructs that any batch hot cut pricing must comply with TELRIC, must reflect the efficiencies gained by performing hot cuts as batch rather than one at a time, and must reduce the price that CLECs pay.

MCI argues that the ILECs have increased mechanization for their retail services by continually improving the underlying technology and systems.¹⁵⁷ For example, a software matrix within IDLC equipment serves the role of the main distribution frame and allows SBC to "cut" a customer's IDLC loop without any manual intervention.¹⁵⁸ Thus, when SBC "turns up" a retail customer's service using IDLC, there is no need in most circumstances for SBC to dispatch a technician. With modern software platforms available from multiple IDLC vendors, a carrier can map any IDLC loop to nearly any port on its digital switch on a desktop computer.¹⁵⁹ MCI argues that these same systems, with certain modifications, can provide the same software-driven efficiency for loops being hot cut to another carrier's switch.¹⁶⁰ MCI claims that existing mechanized frame technology allows a carrier to cross connect all-copper pairs within a software-driven environment, via numerous methods (robotic, electronic matrix, etc.) without dispatching a technician to perform the function.¹⁶¹

SBC assumes IDLC penetration rates of approximately 50% in its cost studies.¹⁶² In those circumstances where copper would remain as the most efficient, least cost technology to be used in a forward looking network design, MCI proposes that SBC use existing technologies to allow a carrier to provision services without manual cross connect (so-called "lift & lay") activities. Automated distribution frame technology is already available from Telcordia and numerous IDLC vendors.¹⁶³ Moreover, Verizon uses automated frame technology and has stated that it would use this technology for purposes of accomplishing hot cuts for unbundled loops if required.¹⁶⁴

SBC argues that there is no commercially available mechanized frame that makes economic sense for SBC's network. SBC witness Mitchell testified that mechanized frames are limited to 5,000 lines or fewer, and have not passed the work standards that would be required to place them into widespread use. Telcordia (formerly Bellcore) has stated that the mechanized frame industry is "not ready for prime time." (Ex. 36 (Mitchell) at 7.) Telcordia concluded it would cost more to cut over loops with the mechanized frames that are available today than it costs to cut over loops manually.¹⁶⁵

Given the practical commercial constraints pointed out by SBC witness Mitchell, we agree that a mechanized frame technology is not commercially feasible today. Further efforts, however, to make such technology viable and cost-effective are important. Nonetheless, a forward-looking network for purposes of deriving TELRIC-based prices must be based on technology that exists today. Because such current technology does not include mechanized frames, we agree with SBC that TELRIC based prices applicable to current hot cut processes cannot incorporate such technology.

b) Charges for Removal of Customer Number and Switch Translations

SBC's initial BHC cost study included expenses associated with removal of the CLEC customers' telephone number and other switch-related instructions from SBC's switch.¹⁶⁶ MCI argued that these activities are not incremental to a request for a hot cut, but are part of the SBC's customer's cancellation of service and are recovered through SBC retail rates when SBC first acquires the customer.¹⁶⁷

We agree that these removal costs are not part of the hot cut process, and should be excluded from SBC's TELRIC prices to avoid double recovery. SBC witness Pearsons revised his calculations in response to MCI's concerns on this issue to exclude the costs of the SBC customer cancellation from BHC costs.¹⁶⁸ Accordingly, SBC's revisions appropriately resolve the dispute over this issue. We emphasize, however, that the ILECs are precluded from charging any of the disconnect charges identified by MCI and other CLECs in this proceeding, not just those associated with removal of a CLEC customer from UNE-P. This prohibition applies to Verizon as well as SBC.

c) Surcharges to Unbundle Loops Provisioned on IDLC

SBC's BHC process includes a provision for IDLC-provisioned loops. IDLC technology integrates the digital loop carrier system directly into a switch on a digital (DS1) basis. For IDLC-provisioned loops, the analog signal generated by the end-user's customer premises equipment is converted into a digital signal at a remote terminal. The digital signal is then multiplexed into a DS1 signal, and is transported (along with the other signals with which it has been multiplexed) to the central office over a high-speed digital feeder facility. At the central office, the feeder facility is terminated, and the IDLC traffic is routed as DS1-level signals to the digital line ports on the switch.¹⁶⁹

SBC argues that because voice traffic over IDLC-provisioned loops is delivered into the switch as a multiplexed, DS1-level signal, there is no technically feasible means of obtaining access to an individual IDLC-provisioned loop at the central office. Consequently, in order for an SBC customer served by an IDLC-provisioned loop to be cut over to another carrier's switch, the customer's service must be moved from the IDLC-provisioned loop to either a copper loop or a loop provisioned with UDLC technology (which, unlike IDLC, permits access to individual loops at the central office). To do this, an SBC must dispatch an LFO-Out technician to the field to move the end-user's service.¹⁷⁰

MCI contends that the LFO-Out work could be eliminated by unbundling the individual loops. MCI thus opposes the SBC surcharge equal to $91.50 per loop for unbundling IDLC loops so the customer may be migrated to the CLEC switch.¹⁷¹ MCI argues that technology exists that would allow CLECs to access loops, especially IDLC loops, in a more efficient, unbundled way via a number of options developed by Telcordia. MCI claims that such technology would avoid the need for manual dispatch to the central office or the remote terminal for purposes of a hot cut, but would instead most likely take place via software command with no manual intervention.¹⁷²

IDLC was implemented, in part, to reduce the number of dispatches required to provide combined services (i.e., retail, resale, UNE-P).¹⁷³ MCI argues that these same efficiencies could be captured for UNE loops, although SBC has declined to develop the required processes and systems.¹⁷⁴ Nonetheless, TELRIC rules require SBC to assume a network configuration employing the most efficient technology in the least-cost manner.

SBC declined to develop IDLC unbundling processes and system, at least in part, on the basis that it has deployed only a small amount of IDLC on a statewide basis.¹⁷⁵ However, the percentage of IDLC appears to be growing, as SBC witness De Luca noted in revised testimony, more than doubling the estimate of the amount of IDLC now in SBC loop plant statewide.¹⁷⁶ Moreover, in some wire centers, the concentration of IDLC facilities can be quite large, representing more than half of all residential customers.¹⁷⁷

MCI further claims that SBC California's sister company in Ohio performs IDLC unbundling through its "Facilities Modification Policy" and various 271-related agreements at no charge for UNE loops. If SBC agrees to perform IDLC unbundling in a different service territory, MCI argues that it should be willing to do so in California as well.¹⁷⁸

SBC, however, denies that SBC Ohio agreed to the contract provision to which MCI refers, citing pertinent excerpts from the Public Utilities Commission of Ohio recommended decision of the arbitration panel and from the Ohio interconnection agreement.¹⁷⁹ Upon review of these excerpts, we agree with SBC that the position SBC is taking in California with respect to IDLC is consistent with the position of SBC Ohio. Thus, MCI's argument is not convincing in claiming that actions in Ohio justify IDLC unbundling in California.

Moreover, MCI has not shown how IDLC loops could be unbundled using currently available technology, or how such unbundling would eliminate the LFO-Out work that is now required. AT&T joins MCI in arguing that the technology exists to enable IDLC-provisioned loops to be unbundled by software so that no manual work would be necessary. AT&T cites testimony Exhibit. 115 of Witness Turner. AT&T proposes that until SBC revises its TELRIC model to incorporate a more automated process for unbundling of IDLC loops, the CPUC should require the use of FDT process costs in lieu of IDLC costs. AT&T claims that these costs are more commensurate with the level of coordination that SBC has assumed for the IDLC process.

SBC refutes AT&T's claims in its Reply Brief (pp. 100-101). SBC cites to the reply testimony of witness Mitchell for an explanation of why IDLC unbundling is not feasible. As noted by SBC, the FCC has specifically rejected AT&T's claim that IDLC can be unbundled electronically through a software command, and declined to order that electronic loop provisioning be used in batch hot cut studies.¹⁸⁰ Accordingly, we find no basis to disallow SBC's cost for IDLC loops.

d) Internal Consistency of Estimates

SBC witness Pearsons served an initial cost study on January 7, 2004,¹⁸¹ with two subsequent revisions.¹⁸² The last revision resulted in a reduction in costs of 20%. Pearsons testified that the first revision was to correct the application of "premium" versus "average" labor rates for certain tasks carried on prior to the due date of the hot cut.¹⁸³ The second revision was to reflect concerns raised by MCI and AT&T's experts regarding some of the task times in Pearsons' initial study, ¹⁸⁴ but also to correct his task time estimates to reflect California-specific numbers rather than the 13-state wide numbers that Pearsons had used in his initial cost study filed on January 7, 2004.¹⁸⁵ Ms. Heki, SBC's LFO-IN expert, testified that the task times in Pearsons' cost study were higher than her expert estimates because Pearsons had used 13-state wide average input numbers.¹⁸⁶

Pearsons testified that he had removed the 13-state wide average numbers from his cost study during the second revision. MCI claims that discrepancies remain, however, between Pearsons' task time estimates and those of the workgroup experts who provided estimates. MCI claims that Pearsons' estimate of time for the LOC workgroup to perform a coordinated hot cut is 12 minutes, but his estimate for the exact same set of tasks is 14 minutes.¹⁸⁷

As verified by SBC in its reply brief, however, SBC witness Pearson did in fact use a 12-minute estimate for its LOC task times.¹⁸⁸ Accordingly, we conclude that SBC's LOC task time estimates are internally consistent.

e) Task Time for "External Issues"

AT&T witness Turner proposed to cut in half the cost study time for the LOC to "Resolve internal and external issues (pre-due date)," claiming that since hot cuts are done within the central office, there will only be internal issues (i.e., issues internal to the central office), and no external issues (i.e., issues relating to outside plant). (Ex. 115 (Turner), at 23.) SBC argues that Turner has misinterpreted the words "internal" and external" in the cost study. "Internal" refers to issues internal to the LOC, and "external" refers to issues outside the LOC (i.e., involving other SBC departments or the CLEC). With this explanation, SBC contends that Turner's reduction is invalid. We find SBC's explanation reasonable and accept its calculations on this point.

f) Application of Task Times on a Per-Order Basis Rather than Per-Loop Basis

Turner asserts that for six LOC tasks, SBC inappropriately attributed to each individual loop a task time that should apply only to each order (so that, for example, the charges would apply only once, rather than twenty times, to an order for twenty loops). (Ex. 115 (Turner), at 24.) SBC revised its cost study in response to this concern. Specifically, in its revision, SBC now assigns a time of 0.00 minutes to four of the six tasks under the "Additional Resource Driver" column on the Bill of Costs.¹⁸⁹ Thus, SBC now proposes to apply the time for those tasks on an order basis rather than on an individual loop basis, as Turner advocated. For the other two tasks to which Turner referred (see Id, Lines 9 and 17), SBC has reduced the time for loops after the first loop to one half of the time for the first loop (i.e., 0.50 minutes vs. 1.00 minutes). For those two tasks, the LOC Maintenance Administrator does spend time on each loop, but spends twice as much on the first loop as on the remaining loops in the same order.

We find SBC's revisions in response to Turner's criticisms to be conceptually reasonable. However, in comparing the times of those two activities with times of similar activities, we found inconsistencies. In those instances, we have revised SBC's calculations to apply its revised assumptions on a consistent basis. The applicable cost elements that we have adjusted in this manner are set forth in Appendix 2.

g) Travel Time to Unmanned Offices

AT&T Witness Turner claims that SBC overstated the time for LFO-In technicians to travel to unmanned offices; receive and review service orders; perform cross connects; and perform dial tone and ANI testing and lifts and lays. Turner claims that the LFO-In organization's time estimate for "Travel time to unmanned offices" should be reduced. (Ex. 115 (Turner), at 27-28.) Turner argues that SBC has provided no support for the labor times associated with travel in its cost studies.

SBC argues that the General Manager of SBC's Northern California LFO for Central Office Operations (Ex. 31 (SBC Panel), at 2) is better qualified to offer an opinion on travel time that is AT&T's witness who offers literally no support for his lower time estimate. In this instance, we conclude that SBC's subject matter expert is better positioned to provide an expert opinion concerning the travel time required to an unmanned SBC central office. We shall accept SBC's estimate of travel times.

h) Allocation of Task Time for Work Orders Unrelated to Hot Cuts

Turner contends that regardless of how long it takes the technician to travel to the CO, the batch hot cut study should reflect only a fraction of that time, because the technician, after arriving at the CO, may perform more that one task. While agreeing that a technician dispatched to an unmanned office may perform work in addition to hot cuts, SBC disputes AT&T's claim that the technician performs an average of four work orders when dispatched to an unmanned central office. (Ex. 115 (Turner), at 29.)

We accept as reasonable Turner's estimate that a technician performs an average of four work orders per dispatch. Turner relied on his past experience and the experience of peers that have actually managed these types of functions. We also agree with SBC, however, that a technician will not necessarily spend an equal percentage of time on each task. SBC argues that if a technician cuts over, for example, 12 loops, and performs three other discrete tasks at an unmanned CO (for a total of four work orders), the percentage of the technician's travel time that should properly be assigned to the CLECs that ordered the hot cuts would be 80%, not necessarily the 25% that Turner would use.¹⁹⁰

We agree that the weighting of work tasks other than hot cuts should take into account the relative percentage of time spent on each task. Yet, SBC assumed 100% of the technician's time applies to hot cuts in its calculation. No party, however, has provided evidence concerning the specific amount of time that a technician, on average, spends on hot cut tasks relative to other tasks at a given central office dispatch. Therefore, in the absence of specific evidence to indicate to what extent, if any, the time to cut over a loop is longer or shorter than the time required for a technician to perform other tasks, we shall adopt the AT&T assumption applying an equivalent duration to each task. Thus, based on Witness Turner's testimony that the technician performs an average of four work orders per central office, we shall assign 25% of the technician's time to hot cut activity. Accordingly, subject to true up pending receipt of more specific evidence we shall adjust SBC's TELRIC calculations to subtract 75% of technician time as being attributable to functions other than hot cuts.

i) Time to Receive and Review Service Orders

SBC's initial cost study allotted two minutes to "receive and review service order" for each loop in the order. In pre-filed testimony (at 29-30), Turner agreed that two minutes was appropriate for the first loop, but contended that only 0.50 minutes should be allotted for each subsequent loop. After consulting with its subject matter experts in this area, SBC reduced from two minutes to one minute the time for loops after the first loop on an order. In analyzing the times to receive and review service orders, we found this task to be similar to tasks in Lines 8 and 9. The task in Line 8 is to receive and input order onto cut log and includes a time of one minute for the initial loop and 0 minutes for additional loops.

The task in line 9 is to screen service order and verify facility assignments are correct and is given a time of one minute for the initial loop and 0.5 minutes for additional loops. In comparing the three similar tasks, we conclude that if the time to receive and review service orders for an initial loop on an order is two minutes, then the time to receive and review service orders for additional loops on an order should be 0.5 minutes.

j) Cross-Connect Times

AT&T contends that SBC inappropriately included IDF cross connect costs in its cost studies based on Turner's testimony that the cross-connect times include time to perform cross-connects on IDFs, and that IDFs are not forward looking. (Ex. 115 (Turner), at 30-34.) Turner believes that SBC assumed that 100% of hot cuts entail a cross connect on an IDF (as well as on the Main Distribution Frame) (see Id. at 31, lines 12-17). On that basis, Turner proposed to eliminate the IDF cross-connect time by cutting SBC California's cross-connect times in half (Id. at 34).

SBC denies that it assumed that every hot cut would entail a cross connect at an IDF. Because very few central offices in California have IDFs, SBC's cost study includes little or no time for cross connects on IDFs. (See Feb. 6, 2004 Tr. at 9479 (Pearsons).) SBC argues that because the few offices in California that do have IDFs have them for good reason, whatever minimal IDF-related costs may be included in the cost study are forward-looking. See id. at 9479-9480.¹⁹¹

We conclude that there is not enough evidence in the record to support either AT&T's or SBC's arguments regarding the cross connect times in SBC's cost study. We note that in a February 5, 2004 correction to the SBC cost study, Pearsons revised the times for cross connects from 8.71 minutes to six minutes for the initial driver and from 7.79 minutes to six minutes for the additional drivers. At this time, we accept the six minutes for the initial and additional drivers. However, in subsequent proceedings to produce finalized hot cut prices, we shall direct further review of this specific task time in order to ensure correct measurements of time for the SBC batch cut process.

k) Time to Perform Dial Tone and ANI Testing and Lift and Lay

Turner also argued that SBC time estimates to complete the lift and lay, ANI testing and dial tone check should be reduced to two minutes to reflect times he has seen in "various nonrecurring cost studies including those of SBC in other jurisdictions." (Ex. 115 (Turner), at 34.) Turner proposed to reduce from four minutes to two minutes the task time for this item (which appears at Line 5 of Ex. 88C), based on his belief as to how long this activity should take. (Ex. 115 (Turner), at 34.) SBC reduced its time estimate to three minutes based on further consideration by SBC subject matter experts. SBC argues that its subject matter experts' opinion as to how long this task takes should be given more weight than that of Turner.¹⁹²

We conclude that Turner's estimate of two minutes for completion of the three tasks at issue is more reasonable, and shall adopt it. SBC Panel Testimony states that the "lift and lay typically takes less than a minute."¹⁹³ AT&T witness Turner testified that the task of performing the dial tone and ANI testing are essentially electronic tests that should take no more than one minute.¹⁹⁴ Thus, taking these statements together, we find it reasonable to conclude that all three tasks could be completed within two minutes. SBC has not explained why an extra minute would be required. Accordingly, we shall reduce SBC's
three-minute estimate for these functions down to two minutes, for purposes of computing TELRIC prices.

l) Fallout Rate

AT&T contends that SBC's cost study incorporates high fallout rates. When a local service order does not flow through to service order generation without manual intervention by the LSC, the order is said to "fallout." Turner contends that SBC has failed to incorporate forward-looking flow through probabilities into the BHC Cost Study, and takes issue with the percentage figure cited by SBC. SBC explains that this percentage figure is for the LOC,¹⁹⁵ and does not pertain to the LSC or the ordering process at all, but to the LOC and provisioning.

The FCC defines flow-through as the "percentage of orders that an incumbent LEC processes electronically through its gateway and accepts into its back office systems without manual intervention (i.e., without additional human intervention once the order is submitted into the system)." Performance Measures Proposed Rulemaking, 13 FCC Rcd at 12,849, ¶ 71

MCI argues that the Commission has already determined that non-recurring costs for SBC should be set based on an expectation that orders will flow through on an automated basis, without the need for manual intervention, 96 percent of the time (i.e., a 4% fallout rate).¹⁹⁶ MCI proposes that the same assumption should be applied to the non-recurring costs and prices for SBC's batch hot cut processes.

SBC argues that the FCC has indicated that fallout is an attribute of the OSS ordering function, but does not refer to the OSS provisioning function. The FCC thus specifies that flow-through does not apply to provisioning:

Order Flow Through applies solely to the OSS ordering function, not the OSS provisioning function. In other words, Order Flow Through measures only how the competing carrier's order is transmitted to the incumbent's back office ordering systems, not how the incumbent ultimately completes that order. (Performance Measures Proposed Rulemaking, 13 FCC Rcd at 12,849, ¶ 71 (emphasis added).)

Turner extrapolates the concepts of flow-through and fallout into the provisioning of hot cuts, contending that fallout rates applicable in the LSC should also apply when the order is being provisioned at the LOC. Thus, all the sources for the fallout figures Turner relies on are OSS ordering functions. Since Turner relies exclusively on percentages that pertain to the LSC, the fallout percentage he advocates would not apply to the LOC.

Besides the fact that the fallout rate that applies in the LSC cannot appropriately be extended to the LOC, SBC argues that the LOC occurrence probability percentages in the cost study are appropriate and forward looking. They reflect internal mechanization efforts that are currently underway that will reduce manual intervention of pre-due date provisioning activities as well as due date activities required for a hot cut conversion. SBC provides an explanation in its reply brief as to why the specific LOC activities identified by Turner cannot be performed electronically 98% of the time.¹⁹⁷ We agree with SBC's explanation and conclude that AT&T and MCI's argument regarding high fallout rates is unfounded. We do not believe that acceptance of SBC's fallout rates conflicts with D.98-12-079, as claimed by MCI and AT&T. In that decision, the term "fallout rate" referred to the percentage of orders that are returned or fall out. SBC has determined, however, that a fallout rate of 100% applies to individual tasks, not to orders.

m) Removal of Overtime and Shift Differentials

In its cost testimony in Ex. 84C, SBC states that the labor rate used for base hours was calculated by removing overtime and shift differentials from the average labor rate. In our review of SBC's underlying TELRIC calculations, however, we found instances where it failed to remove overtime and shift differentials from the average labor rate. We have made the appropriate corrections to remove the effect of overtime and shift differentials as reflected in the TELRIC prices derived in Appendix 1.

n) Reduction of Time Estimates to Reflect Consolidation of Tasks

In our review of SBC's costs, we concluded that certain tasks identified as requiring separate time completion intervals could be consolidated into more compressed time intervals. Specifically, we conclude that the task (on Line 15 of the TELRIC table): "time to call the CLEC to advise cut complete-due date" is similar to the task on Line 13: "Receive call from CLEC and record start time and call LFO to begin cut- due date." We thus conclude that the "initial minutes" assigned to the line 15 task are not necessary. Instead, we include the Line 15 task within the time already allocated for the task on Line 14. "Receive call from LFO to advise cut complete and records stop time - due date." A similar adjustment is made for other entries in the costing calculation worksheet where those same tasks apply to other BHC process categories, as identified in Appendix 2.

4. Conclusion

Based upon our review of SBC's proposed rates, and consideration of the analysis done by AT&T and MCI, we conclude that the rates proposed by SBC for batch hot cuts are not TELRIC compliant and fail to substantially reduce existing hot cut rates to incorporate efficiencies of a batch process. While we do not accept all of the adjustments proposed by MCI and AT&T, we still conclude that certain adjustments to SBC's proposed costs are necessary in order to make it TELRIC-compliant.

Without the necessary adjustments, CLECs would bear a substantially increased economic burden because of the non-recurring charges they must pay for hot cuts compared to rates they currently pay to migrate customers to their network via UNE-P. These increased charges would cause a "sticker shock" that would make the migration far from seamless.¹⁹⁸

Although MCI has proposed a separate set of prices for a "cut-to-EEL' migration scenario, we do not adopt any prices for that migration scenario at this time. As discussed above in reference to hot cut migration scenarios, we have concluded that further development is needed before such a scenario could be implemented. Thus, it is premature to adopt prices for this scenario at this point.

B. Verizon's Proposed TELRIC Pricing for Hot Cut Processes

1. Position of Verizon

Verizon presented its TELRIC pricing proposal in the testimony of Ann Dean¹⁹⁹ covering three of Verizon's different hot cut options: (1) the "Basic" hot cut process utilizing the WPTS;²⁰⁰ (2) the large job or "Project" process and (3) the BHC process. In preparing its cost study, Verizon employed the same nonrecurring cost methodology as presented in its cost filing in R.93-04-003/
I.93-04-002. Verizon asserts that its costs are compliant with D.03-03-033, and thus proposes the following structure for the three hot cut options:




1) Hot Cut Ordering Charge: This charge is intended to recover the costs of processing and provisioning a hot cut order, broken down into "manual, semi-mechanized, and mechanized" categories. Ordering charges include activities relating to wiring and pre-wiring, coordination, and related management of the hot cut. Separate charges are assessed for the first versus additional units.




2) WPTS Coordination Expedite Charge: This charge applies when a CLEC requests service earlier than the next standard due date.




3) IDLC Surcharges: These charges cover costs due to substituting facilities before a cut can be made, primarily where the loop is provisioned using IDLC technology.

Verizon's proposed charges are summarized below:

Verizon argues that its cost study is consistent with TELRIC rules, the Commission's Consensus Costing Principles, and the Commission's nonrecurring rate structure. Verizon characterizes its study as a bottoms-up analysis that measures each cost arising from servicing individual CLEC requests for hot cuts. Verizon first identified applicable work activities and durations. For Regional CLEC Coordination Center (RCCC) and CO Frame organizations, Verizon determined work durations through surveys of Verizon employees based on a self-reported time and motion study. For Field Dispatch activities, Verizon used the sub-loop drive time study submitted in the UNE case
(R.03-04-003/I.93-04-002). For National Marketing Center (NMC) activities, Verizon developed costs using historical data from system generated reports. For Recent Change Memory Administration Center (RCMAC) activities, Verizon used approved work times from the N.Y. State PSC in C.98-C-1357.

Verizon then applied an "occurrence factor" to the work times to reflect the percentage of cases where the activity is currently required to determine average time required for the activity across all orders. Verizon also applied a "Forward-Looking Adjustment Factor" based on consultation with subject matter experts.

To convert the work times into a forward looking cost, Verizon multiplied the forward looking work time for each activity by the hourly labor cost per worker category, trended forward three years at a 4% annual escalation rate. Verizon finally applied a common overhead and gross revenue loading to the labor cost for each activity to reflect overheads, uncollectibles, and assessments. As a placeholder, Verizon utilized the common overhead loading factor that it proposed in its November 3, 2003 filing in R.93-04-003/I.93-04-002.

2. Response of MCI and AT&T to Verizon Cost Study

MCI argues that Verizon's batch hot cut cost analysis suffers from many of the same conceptual problems as does SBC's. MCI argues that Verizon's hot cut costs and prices do not incorporate the mechanization that is technically feasible and achievable through deployment of electronic unbundling of IDLC and automated main frame technology. Thus, MCI claims that Verizon's batch hot cut pricing proposal does not comply with the required TELRIC standard, producing exaggerated cost estimates for hot cuts, relative to the governing cost standard and, notably, prices higher than the UNE-P charges CLECs pay today in California to migrate customers.²⁰¹

MCI did not propose specific prices based on the Verizon hot cut cost model. MCI claims it was unable to do so given the complexity of Verizon's model and Verizon's failure to produce supporting data on a timely basis. MCI requests that the Commission instead adopt its conceptual adjustments proposed for Verizon and compel Verizon to produce a revised model recalculated consistent with the Commission's determinations. AT&T, however, was able to provide specific adjustments to Verizon's cost study and recalculated rates along the same conceptual lines advocated by MCI. MCI generally believes AT&T's analysis produces rate alternatives compliant with the TELRIC standard.²⁰²

MCI claims that Verizon's cost studies and rates suffer from both methodological and structural problems that systematically exaggerate rates relative to what the FCC's TELRIC methodology would produce. Verizon's cost studies rely upon estimates related to the time required to complete certain tasks undertaken by its technicians and other provisioning personnel. For each activity, Verizon's NRC model estimates the expected forward-looking time as a product of three components: activity durations, occurrence factor and a forward-looking adjustment factor (FLAF). These three components are estimated through three separate processes described by Verizon, yet little if any information validating Verizon's assumptions was available in time for MCI to analyze it.

AT&T likewise claims that Verizon's costs are not TELRIC-compliant. AT&T submitted its analysis through the testimony of Richard Walsh who testified that the appropriate cost for a hot cut in a forward-looking environment should be in the range of $5.²⁰³ AT&T argues that Verizon should be required to resubmit its cost study on a TELRIC-compliant basis. In the alternative, AT&T recommends that the Commission adopt the costs and rates proposed by AT&T witness Walsh, which are based on SBC's study, modified to reflect revisions that AT&T considers more forward looking.

3. Discussion

While we do not find that all of the adjustments to Verizon's batch cut process proposed by MCI and AT&T have been justified, parties' analysis of Verizon's hot cut costs was not sufficiently completed given the complexity of the analysis and the ambitious schedule under the nine-month proceeding. Therefore, we do not find the record sufficient at this point to adopt final prices for Verizon's hot cut processes. We conclude that before specific batch cut prices are approved for Verizon, further proceedings are necessary to compare the relative methodologies between SBC and Verizon, and to understand the basis for material disparities in cost treatment. We shall direct the ALJ to issue a procedural ruling to this effect. In the event that separate TELRIC-complaint prices cannot be approved for Verizon by the time that its batch cut processes are to be implemented, we shall direct that SBC's prices be used as a surrogate for Verizon.

We address below our disposition of specific issues that were raised by parties concerning Verizon's cost calculations.

a) Verizon Labor Rates

Verizon asserts that its labor rates are TELRIC-compliant. Verizon derived its labor rate by taking the actual 2002 wage rates for employees who perform hot cuts and dividing the 2002 wage expense by the number of productive hours. Verizon then added to the basic rate the cost of employee benefits, premium time, payroll taxes, and paid absences, as well as tools, motor vehicles, clerical support, and supervision of reporting personnel. Verizon then applied a 1.04% inflation factor, to trend the data through 2005, the assumed period during which the rates would be in effect.

AT&T presented the testimony of Robert Flappan who conducted an analysis of the hot cut labor rates proposed by Verizon. Flappan claims that Verizon's labor rates are not TELRIC compliant, but are based on embedded costs. Flappan summarizes Verizon's proposed fully loaded hot cut labor rates for six categories of operations in Table 1 of his testimony. In Tables 2 and 3, Flappan summarizes the comparison of his proposed TELRIC-adjusted labor rates versus those proposed by Verizon.

In Ex. 161C, Table 12, Flappan summarized AT&T's proposed adjustments to Verizon's labor rate to reflect a TELRIC rather than embedded cost basis.²⁰⁴ Flappan claims that the labor rate shown in the sixth column of his table represent efficient labor rates that would be achievable by a new entrant.

Flappan reduces Verizon's labor rates to exclude shared and common costs, which are not properly part of a TELRIC labor rate study. Flappan also claims that Verizon's costs are merely based on the historical experience of one firm, namely Verizon. Flappan proposes various adjustments that he claims are necessary to bring Verizon's labor rates into compliance with TELRIC.

Verizon disputes Flappan's claim that its labor rates are not TELRIC-compliant merely because they reflect recorded costs. Verizon asserts that it has presented the costs that it will actually incur going forward, and that its labor rates reflect the real costs of a large company with a unionized work force. Verizon disputes Flappan's claim that it has not reflected productivity gains. Verizon reflects productivity through its application of FLAF to the expected time needed to perform each hot cut activity. Adjusting both the labor rate and work time for productivity would result in double counting. Verizon criticizes Flappan's figures are being merely hypothetical, and based on Bureau of Labor Statistics data, resulting in rates as low as 30% below what Verizon will actually pay its workforce.

We agree with Verizon that TELRIC-based rates can be based on actual costs provided that those costs are otherwise forward-looking in nature. Generally, we find no basis to disallow costs that are Verizon-specific merely because they are higher than costs of other firms. We agree with AT&T, however, that shared and common costs should be excluded from base labor rates.

b) Verizon's Estimated Work Activity Durations

Verizon's cost studies identify the work steps believed to be required to accomplish the hot cut process encompassed by its various hot cut proposals.²⁰⁵ Verizon then estimates the time required by its technical personnel for each work step. The work steps were identified by subject matter experts, while activity durations were obtained from a number of sources including surveys, analyses in other states or reports generated by Verizon's internal systems.

MCI's witness provided an examination of the total forward-looking times for an individual initial two-wire hot cut contained in Verizon's cost study.²⁰⁶ The most time-intensive activities reflected in Verizon's cost study are associated with Verizon's CO Frame, NMC and RCCC organizations. While both the CO Frame and RCCC activities (representing the majority of work time captured in Verizon's cost study) take place in California, many of the time estimates and all occurrence factors come from sources other than California specific data.

Verizon did not present documentation specific to its survey studies supporting many of its duration estimates, until late in this proceeding, after the time when MCI's witness could have used this information in his analysis.²⁰⁷ The underlying databases allow the parties to view the sample sizes used by Verizon, i.e., how many responses were used to arrive at each time estimate. Likewise, the survey forms show whether the format or wording used might have biased the responses or prompted employees to overestimate or underestimate activity durations.²⁰⁸ The survey instructions and questionnaires used in California were virtually the same in substance as those used by Verizon in New York and Virginia. This survey process was reviewed by the FCC in the proceedings leading to its Virginia Arbitration Order.²⁰⁹

MCI argues that survey responses are subjective measures, and therefore, contain a potential source of bias, such as here, where respondents had several incentives to overestimate work times, and no serious incentives to underestimate them. Moreover, Verizon did not provide a proper statistical validation of the results. Specifically, Verizon did not show that its sample of responses in the study properly represents the "population," or durations of typical hot cut activities. In fact, for the majority of activities and sample sizes were less than what is considered to be sufficient for statistical analysis.²¹⁰

We agree that MCI has raised valid concerns regarding the reliability of Verizon's estimated work activity durations. We do not believe a sufficient basis exists for approval of the durations assumed by Verizon underlying its cost calculations. We direct that further work needs to be done to verify work durations before Verizon's costs can be finalized and approved.

¹⁴⁰ 47 C.F.R. § 51.505(b).

¹⁴¹ 47 C.F.R. § 51.319(d)(2)(ii)(A)(4).

¹⁴² The costs and rate structure that SBC proposes for its Batch Hot Cut Processes are based on the cost study sponsored in testimony of Scott Pearsons.

¹⁴³ D.95-12-016, Appendix C.

¹⁴⁴ The Ninth Circuit Court of Appeals reversed the Commission's findings on the 21% factor in AT&T Communications, Inc. v. Pacific Bell Tel. Co., 375 F.3d 894 (9^th Cir. 2004). SBC has filed a petition for rehearing on the Circuit Court opinion which is still pending. Accordingly, the 21% overhead factor is adopted herein subject to any final determinations made pursuant to disposition by the Ninth Circuit. Upon final determination of the appropriate overhead factor, we authorize a true up of applicable hot cut rates to reconcile the difference between the 21% versus the final adopted factor.

¹⁴⁵ Available request times vary among the three processes.

¹⁴⁶ Id.

¹⁴⁷ Id.

¹⁴⁸ SBC Joint Cost Witnesses are Domenic Cusolito, Dennis Deluca and Barbara Heki.

¹⁴⁹ The Workgroup Occurrence Factor is the percentage of time that a workgroup must be involved to provision the service. The WGOF may represent the fallout associated with a mechanized process, or the percent occurrence when two different workgroups share responsibility for a similar work activity.

¹⁵⁰ In OANAD, this occurrence was referred to as a task occurrence factor. It is the percentage of time that an activity must take place after the WGOF is considered.

¹⁵¹ Ex. 146 (Starkey 1/28 Reply), at 26.

¹⁵² Ex. 146 (Starkey 1/28 Reply), at 26. MCI provided, as Confidential Attachment 4 to Starkey's Jan. 28, 2004 batch hot cut pricing reply testimony, a copy of SBC's cost study including the proposed modifications of Starkey in support of the rates proposed in this table.

¹⁵³ Ex. 146 (Starkey 1/28 Reply), at 28.

¹⁵⁴ Ex. 146 (Starkey 1/28 Reply), at 28-29.

¹⁵⁵ Id.

¹⁵⁶ Id.

¹⁵⁷ Ex. 146 (Starkey 1/28 Reply), at 8 and Attachment 1.

¹⁵⁸ Ex. 146 (Starkey 1/28 Reply), at 9.

¹⁵⁹ Id.

¹⁶⁰ Id.

¹⁶¹ Ex. 146 (Starkey 1/28 Reply), at 11.

¹⁶² Ex. 146 (Starkey 1/28 Reply), at 11.

¹⁶³ Ex. 146 (Starkey 1/28 Reply), at 11 and Attachment 3.

¹⁶⁴ Ex. 146 (Starkey 1/28 Reply), at 12. Verizon made this statement when appearing before the State of New York, Public Service Commission, Proceeding on Motion of the Commission to Examine the Process, and Related Costs of Performing Loop Migrations on a More Streamlined (e.g., Bulk) Basis, Case No. 02-C-1425, Public Transcript (pp. 290-293), Testimony of Michael A. Nawrocki, On Behalf of Verizon New York, Inc.

¹⁶⁵ Tr. at 8963 (Mitchell).

¹⁶⁶ Ex. 146 (Starkey 1/28 Reply), at 13.

¹⁶⁷ Ex. 146 (Starkey 1/28 Reply), at 14.

¹⁶⁸ RT 60; Pearsons/9459.

¹⁶⁹ SBC Reply Brief at 97.

¹⁷⁰ Ex.113, SBC/DeLuca at 5-6.

¹⁷¹ Ex. 83 (Chapman 1/7/04 Direct and 2/5/04 revision), Pricing Schedule, p. 6.

¹⁷² Ex. 146 (Starkey 1/28 Reply), at 16.

¹⁷³ Id.

¹⁷⁴ Id.

¹⁷⁵ Ex. 146 (Starkey 1/28 Reply), at 17.

¹⁷⁶ Revised Direct Testimony of Dennis DeLuca was e-mailed to parties.

¹⁷⁷ Ex. 146 (Starkey 1/28 Reply), at 17.

¹⁷⁸ Id.

¹⁷⁹ SBC Reply Brief at 98-99.

¹⁸⁰ See TRO Sec. 487, n. 1517

¹⁸¹ Ex. 84C (Pearsons 1/7 Direct).

¹⁸² Ex. 86C (Pearsons 1/13 Revised Direct); Ex. 87C (Pearsons 2/5 Second Revised Direct).

¹⁸³ Tr. 2/6/04, (Pearsons), at 9457-9458.

¹⁸⁴ Tr. 2/6/04, (Pearsons), at 9458-9461.

¹⁸⁵ Tr. 2/6/04, (Pearsons), at 9467-9468.

¹⁸⁶ Tr. 2/3/04, (Heki), at 8811. Moreover, the Commission has rejected use of non-California-specific data in cost studies. See D.98-12-079, p. 28, rejecting GTE's non-recurring cost studies in part because the embedded data GTE used were not California specific.

¹⁸⁷ Tr. 2/3/04, (Cusolito), at 8778-8780.

¹⁸⁸ Ex. 88C, Bill of Costs, Lines 8-17; see also SBC Reply Brief, p. 105.

¹⁸⁹ See Ex. 88C, Lines 8, 10, 11 and 12.

¹⁹⁰ This assumes that the time to cut over a loop is equal to the time to perform each of the other tasks. There is no evidence in the record for any other assumption.

¹⁹¹ At the hearing, Pearsons testified that if there were any such costs, they would be "minuscule," but that he "would have to rely on Ms. Heki for that." (Id. at 9479.) MCI did question Heki (see Id., p. 9491 et seq.), but not on that subject.

¹⁹² Turner also claimed that SBC incorrectly included an Additional Resource Driver when the LFO-In closes out an order. Ex. 115 (Turner), at 35. SBC yielded to Turner on that point. (See Ex. 88C, entry for "Close Order in Frame Work Station - due date.")

¹⁹³ Ex. 31, SBC BHC Panel Testimony at 10-11.

¹⁹⁴ Ex. 115 (Turner) at 35:1-2.

¹⁹⁵ See Ex. 88C, line 10, Activity Occurrence Probability.

¹⁹⁶ See D.98-12-079 (Dec. 17, 1998), at 72-75. MCI asks that the Commission take administrative notice of its non-recurring cost order.

¹⁹⁷ SBC Reply Brief, pps 113-115.

¹⁹⁸ Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 11.

¹⁹⁹ Ex. 28C Testimony of Ann Dean and Ex. 30 C -Verizon's Proprietary Cost Study.

²⁰⁰ Verizon does not propose rates for Verizon's basic, non-WPTS coordinated conversion and hot cut coordinated conversion options which are being litigated in R.93-04-003/I.93-04-002.

²⁰¹ Ex. 146 (Starkey 1/28 Reply), at 3-12.

²⁰² Ex. 146 (Starkey 1/28 Reply), at 2-3.

²⁰³ Ex. 159 (Walsh 1/28 Reply) at 2:12-16.

²⁰⁴ Attachment RPF-2 "Verizon CA TRO Hot cut Labor rates Restate.xls" is attached to Flappan's Ex. 161C in support of his calculations developing proposed TELRIC labor rates for Verizon. The supporting worksheets in Excel workbook format are summarized on pp. 36-37 of his testimony.

²⁰⁵ Ex. 28 (Dean 1/7 Direct) at 11, 14 and 15.

²⁰⁶ Ex. 146 (Starkey 1/28 Reply), at 30-31.

²⁰⁷ Hearing Tr. (Harrelson), Feb. 27, 2004 at 10830-10831. Despite AT&T's February 2 record request for these survey forms, Verizon did not produce them until February 25, two days before the close of hearings.

²⁰⁸ Ex. 146 (Starkey 1/28 Reply), at 31-32.

²⁰⁹ Memorandum Opinion and Order, (Virginia Arbitration Order) CC Docket Nos. 00-218 and 00-251 (rel. August 29, 2003).

²¹⁰ Ex. 146 (Starkey 1/28 Reply), at 33.