At & T Communications of California, Inc. v. Pacific Bell Telephone Co.

 ORDER ON CROSS-MOTIONS FOR SUMMARY JUDGMENT WILKEN, District Judge. In this case, Plaintiffs AT & T Communications of California, Inc., MCI World-corn Network Services, Inc. and MCImetro Access Transmission Services LLC seek to overturn various decisions by the Public Utilities Commission of the State of California (CPUC) during proceedings that set network element rates for Pacific Bell Telephone Company (Pacific). Plaintiffs move for summary judgment (Docket No. 41), seeking to overturn certain aspects of the decisions. Defendants Pacific, the CPUC, and Commissioners of the CPUC in their official capacity (Loretta M. Lynch, Henry M. Duque, Richard A. Bilas, Carl W. Wood and Geoffrey F. Brown) oppose the motion. Pacific cross-moves for summary judgment (Docket No. 55), seeking to overturn other portions of the decisions. Plaintiffs, the CPUC and the individual Commissioners oppose the cross-motion. The matter was heard on May 3, 2002, and taken under submission pending the Supreme Court’s ruling in Verizon Communications, Inc. v. FCC, 535 U.S. 467 , 122 S.Ct. 1646 , 152 L.Ed.2d 701 (2002). Having considered all of the papers filed by the parties and oral argument on the motion, the Court DENIES Plaintiffs’ motion, and GRANTS Pacific’s motion in part and DENIES it in part. BACKGROUND I. Telecommunications Act of 1996 Local telephone service has, for most of its history, been treated as a natural monopoly. Until recently, States typically granted an exclusive franchise for local telephone service in each service area. However, recognizing that technological advances may have obviated the need for exclusive franchises, Congress enacted the Telecommunications Act of 1996, Pub.L. 104-104, 100 Stat. 56 (codified as scattered amendments to the Communications Act of 1934, 47 U.S.C. § 151 et seq.). The Telecommunications Act of 1996(Act) “fundamentally restructures local telephone markets” by imposing a variety of duties on local exchange carriers (LECs) that are designed to facilitate competition. See AT & T Corp. v. Iowa Utils. Bd., 525 U.S. 366, 371, 119 S.Ct. 721 , 142 L.Ed.2d 835 (1999) (Iowa Utils. Bd. I). Pacific is an “incumbent” LEC (ILEC). An ILEC is a. dominant LEC, as defined by 47 U.S.C. § 251 (h)(1), that was providing telephone service when the Telecommunications Act of 1996 became law. In short, an ILEC is a LEC that previously held a State-sanctioned monopoly on local telephone service in a particular service area. Plaintiffs are “competitive” LECs (CLECs) — LECs that are not ILECs. The Act provides that States may no longer enforce laws that impede competition, and ILECs are subject to a host of duties intended to facilitate market entry. Foremost among these duties is the ILEC’s obligation under 47 U.S.C. § 251 (c) to share its network with competitors. Under this provision, a requesting carrier (CLEC) can obtain access to an ILEC’s network in three ways: It can purchase local telephone services at wholesale rates for resale to end users; it can lease elements of the ILEC’s network “on an unbundled basis”; and it can interconnect its own facilities with the ILEC’s network. See Iowa Utils. Bd. I, 525 U.S. at 371-73 , 119 S.Ct. 721 . At issue here is the Act’s requirement that ILECs must lease unbundled network elements (UNEs) 1 to CLECs at rates that *1089 are “based on the cost of providing the element[s]” and “nondiscriminatory.” 47 U.S.C. § 251 (c)(3) & 252(d)(1). Specifically, § 251(c)(3) requires that an ILEC has the duty to provide, to any requesting telecommunications earner for the provision of a telecommunications service, nondiscriminatory access to network elements on an unbundled basis at any technically feasible point on rates, terms, and conditions that are just, reasonable, and nondiscriminatory in accordance with the terms and conditions of the agreement and the requirements of this section and section 252 of this title. An incumbent local exchange carrier shall provide such unbundled network elements in a manner that allows requesting earners to combine such elements in order to provide such telecommunications service. 47 U.S.C. § 251 (c)(3). Section 252(d)(1), which addresses pricing standards, provides: Determinations by a State commission of the just and reasonable rate for the interconnection of facilities and equipment for purposes of subsection (c)(2) of section 251 of this title, and the just and reasonable rate for network elements for purposes of subsection (c)(3) of such section— (A) shall be— (i) based on the cost (determined without reference to a rate-of-return or other rate-based proceeding) of providing the interconnection or network element (whichever is applicable), and (ii) nondiscriminatory, and (B) may include a reasonable profit. 47 U.S.C. § 252 (d)(1). II. The Federal Communication Commission’s Implementing Regulations The Federal Communications Commission (FCC) issued rules implementing the local competition provisions of the Act. See In re Implementation of the Local Competition Provisions in the Telecommunications Act of 1996, 11 F.C.C.R. 15499, 1996 WL 452885 (Aug. 8, 1996) (Local Competition Order). The FCC’s Local Competition Order adopted rules requiring that prices for UNEs be set under a cost methodology known as TELRIC (Total Element Long Run Incremental Cost). TELRIC measures the “forward-looking long run economic cost” of providing a network element because, according to the FCC, this “best replicates, to the extent possible, the conditions of a competitive market.” Local Competition Order ¶¶ 672, 679; see also id. ¶¶ 672-732; 47 C.F.R. §§ 51.503 , 51.505, 51.507. Under TELRIC, the forward-looking economic cost of an element is the sum of (1) the “total element long-run incremental cost” of the element (the forward-looking cost over the long run of the facilities and functions directly attributable to, or reasonably identifiable as incremental to, such element), and (2) a “reasonable allocation of forward-looking common costs” (economic costs incurred in providing a group of elements or services that cannot be attributed directly to individual elements or services). See 47 C.F.R. § 51.505 (b) & (c). In the Local Competition Order, the FCC defined the first category of costs, “directly attributable” costs, as costs that “are incurred as a direct result of providing the network elements, or can be avoid *1090 ed, in the long run, when the company ceases to provide them.” Local Competition Order ¶ 691. The FCC further explained: Directly attributable costs shall include costs such as certain administrative expenses, which have traditionally been viewed as common costs, if these costs vary with the provision of network elements. Retailing costs, such1 as marketing or consumer billing costs associated with retail services, are not attributable to the production of network elements that are offered to interconnecting carriers and must not be included in the forward-looking direct cost of an element. Id. The FCC also stated the importance of attributing costs directly to specific elements: “A properly conducted TELRIC methodology will attribute costs to specific elements to the greatest possible extent, which will reduce the common costs.” Id. ¶ 695 . The FCC in the Local Competition Order defined the second category of costs, “common costs,” as “costs that are incurred in connection with the production of multiple products or services, and remain unchanged as the relative proportion of those products or services varies (i.e., the salaries of corporate managers).” 2 Id. ¶ 676 . The FCC further explained: Certain common costs are incurred in the provision of network elements. As discussed above, some of these costs are common to only a subset of the elements or services provided by incumbent LECs. Such costs shall be allocated to that subset, and should then be allocated among the individual elements or services in that subset, to the greatest possible extent. Common costs also include costs incurred by the firm’s operations as a whole, that are common to all services and elements (e.g, salaries of executives involved in overseeing all activities of the business), although for the purpose of pricing interconnection and access to unbundled elements, which are intermediate products offered to competing carriers, the relevant common costs do not include billing, marketing, and other costs attributable to the provision of retail service. Given these common costs, setting the price of each discrete0 network element based solely on the forward-looking incremental costs directly attributable to the production of individual elements will not recover the total forward-looking costs of operating the wholesale network. Because forward-looking common costs are consistent with our forward-looking, economic cost paradigm, a reasonable measure of such costs shall be included in the prices for interconnection and access to network elements. Id. ¶ 694 (internal footnotes omitted). The FCC makes clear that in calculating common costs and allocating those costs to the direct costs of providing UNEs, all costs of retail services must be excluded, in order to calculate “the total forward-looking costs of operating the wholesale network.” Id. (emphasis added). In other words, the TELRIC methodology calculates the forward-looking cost to ILECs of providing UNEs, in a hypothetical competitive mar *1091 ket in which the ILEC is a wholesaler, leasing UNEs to CLECs. The ILEC’s retail operations (selling telephone services to consumers) are therefore irrelevant to the TELRIC pricing method, and must be excluded. As the FCC stated: “Retailing costs, such as marketing or consumer billing costs associated with retail services, are not attributable to the production of network elements that are offered to interconnecting carriers and must not be included in the forward-looking direct cost of an element.” Id. ¶ 691 ; see also id. ¶ 694 (“[F]or the purpose of pricing interconnection and access to unbundled elements, which are intermediate products offered to competing carriers, the relevant common costs do not include billing, marketing, and other costs attributable to the provision of retail service.”). Regarding the allocation of common costs, the FCC stated in the Local Competition Order: “[F]orward-looking common costs shall be allocated among' elements and services in a reasonable manner, consistent with the pro-competitive goals of the 1996 Act. One reasonable allocation method would be to allocate common costs using a fixed allocator, such as a percentage markup over the directly attributable forward-looking costs.” Id. ¶ 696 . The CPUC’s calculation of Pacific’s common costs, and its method of allocation of those costs to UNEs, are at issue here. Suits challenging the FCC’s implementing regulations were consolidated in the Eighth Circuit. The Eighth Circuit vacated the pricing rules on the ground that the Act authorized State utility commissions, not the FCC, to set rates. See Iowa Utils. Bd. v. FCC, 120 F.3d 753, 793-800 (8th Cir.1997). The court expressly declined to review the pricing rules on the merits. See id. at 800 . The court also vacated non-pricing rules that required incumbent carriers to combine unbundled network elements for competing carriers and prohibited incumbent carriers from separating already combined network elements before leasing them to competing carriers, known as combination rules. See id. at 813 , 819 & n. 39. The Supreme Court reversed the Eighth Circuit’s ruling that the FCC did not have jurisdiction to promulgate pricing rules, holding the' FCC had jurisdiction to “prescribe such rules and regulations as may be necessary in the public interest to carry out the provisions of this Act,” including the “jurisdiction to design a pricing methodology.” See Iowa Utilities Board I, 525 U.S. at 377, 385 , 119 S.Ct. 721 . The Supreme Court also reinstated the combination rule prohibiting ILECs from separating already combined network elements (Rule 315(b)), but did not address the combination rules requiring ILECs to combine network elements (Rules 315(c)-(f)). See id. at 393-94 , 119 S.Ct. 721 . On remand, the Eighth Circuit upheld the use of a forward-looking cost standard, but foreclosed the use of the TELRIC methodology because the Act required rates based on the actual, not hypothetical, cost of providing network elements. See Iowa Utils. Bd. v. FCC, 219 F.3d 744 (8th Cir.2000), aff 'd in part, rev’d in part and remanded sub nom. Verizon Communications, Inc. v. FCC, 535 U.S. 467 , 122 S.Ct. 1646 , 152 L.Ed.2d 701 (2002) (Iowa Utils. Bd. II). The Eighth Circuit also invalidated the combination rules that the Supreme Court did not address, Rules 315(c)-(f), holding that the Act required CLECS, not ILECS, to do any and all combining. See id. The Supreme Court granted petitions for certiorari to review the Eighth Circuit’s judgment and to determine the validity,of the FCC’s TELRIC pricing rules and combination rules found in Rules 315(c)-(f). In Verizon Communications, the Supreme Court upheld the FCC’s TELRIC pricing methodology against various chai- *1092 lenges by ILECs, and upheld the combination rules in Rules 315(c)-(f) as a valid exercise of the FCC’s discretion. See 122 S.Ct. 1646 . III. The CPUC Proceedings The proceedings at issue in this case involve the CPUC’s determination of rates that Pacific, as an ILEC, must charge CLECs such as Plaintiffs to lease UNEs. In determining these rates, the CPUC took several steps. First, the CPUC determined Pacific’s forward-looking costs that are directly attributable solely to the provision of UNEs (total element long-run incremental cost, or TELRIC). See Local Competition Order, 47 C.F.R. § 51.505 (b). Second, the CPUC determined Pacific’s forward-looking “common costs”: operating costs that are indivisible and cannot be attributed directly to a particular network element or aspect of Pacific’s operations, but rather are incurred generally as a result of Pacific’s firm-wide ’operations. See id. § 51.505(c). Third, the CPUC determined a common cost markup, which is applied to the costs directly attributable to providing UNEs in order reasonably to allocate the-common costs to the cost of providing UNEs. These determinations were made in a series of CPUC proceedings and resulting decisions, which are under review here. Prior to the adoption of the Act in 1996, the CPUC began hearings to determine how to open the network architecture of ILECs to competition. See The Open Network to Bottleneck Services and a Framework for Network Architecture Development of Dominant Carrier Networks, D.96-08-021, R.93-04-003, 67 CPUC 2d 221 (OANAD proceeding). The FCC then released its Local Competition Order establishing the TELRIC methodology for determining pricing in response to adoption of the Act. In December, 1996, the CPUC ordered Pacific to submit cost studies based on the TELRIC methodology. The CPUC decisions at issue here are all part of the OANAD proceeding that began in December, 1996. A. First Cost Decision In D.98-02-106 (First Cost Decision) 3 , the CPUC addressed, among other things, the amount of Pacific’s firm-wide forward-looking common costs. 4 The CPUC described this determination as follows: “In this phase, ... our task is to determine the retail portion of common costs that are likely to be incurred by Pacific in a forward-looking environment.” First Cost Decision at 63 n. 55. To determine Pacific’s forward-looking common costs, the retail portion must be excluded. Pacific had earlier filed cost studies claiming $2.2 billion in common costs, which had already been reduced to $1.2 billion in response to previous CPUC decisions applying the provisions of the Local Competition Order. Plaintiffs claimed that the $1.2 billion common cost amount was still too high because over $200 million of these costs were directly attributable to retail services and therefore were not common costs. Plaintiffs submitted a study purporting to show that over $200 million of Pacific’s claimed common costs must be excluded because they are retail-related *1093 and would not be incurred in the future if Pacific operated only as a wholesaler. See Administrative Record, Vol. 2 Ex. 2. The CPUC disagreed, finding that Plaintiffs’ study “assumes that in a forward-looking environment, Pacific will be able to avoid substantial amounts of overhead (such as general, administrative and executive expenses) related to its retail operations. There is as yet no empirical evidence to support such an assumption.” First Cost Decision at 63. However, the CPUC did find that approximately $68 million of Pacific’s claimed common costs were in fact directly retail-related, and must be excluded from Pacific’s forward-looking common costs. The CPUC explained that “some of these costs cannot truly be considered ‘common,’ because they have a clear retail component that, under the TELRIC methodology, may not be included in the determination of wholesale UNE costs.” Id. In other words, the CPUC excluded $68 million because it determined that, in a hypothetical forward-looking environment in which Pacific is solely a wholesaler, Pacific would not incur these retail-related common costs. The CPUC primarily excluded “expenses associated with retail sales and marketing activity, retail customer service activity, and amounts for general administrative expense that are derived by averaging the number of wholesale loops sold against the total number of access lines.” Id. at 63-64; see also id. App. A (unre-dacted) (reflecting a downward retail adjustment of approximately 8%, totaling $68 million, to Pacific’s total common costs, to arrive at wholesale portion of Pacific’s common costs). After further minor adjustments, the CPUC determined that Pacific’s firm-wide common costs totaled $996 million. Pacific argues that the common costs arrived at by the CPUC include only the common costs associated with Pacific’s regulated operations, and does not include costs associated with unregulated and Category III services (these will be referred to together as Category III services), which have their own overhead organizations whose costs are tracked separately. Pacific claims that the common cost figure excludes both the separately tracked Category III services costs and the shared overhead costs that are “direct billed” to the Category III services. B. Second Cost Decision In D.98-12-079 (Second Cost Decision) 5 , the CPUC determined Pacific’s total nonrecurring costs of providing UNEs at $375 million. 6 The CPUC also declined to establish recurring costs for one particular type of UNE, known as Operations Support Systems (OSS). OSS are the systems, databases, and personnel that ILECs use to order, provision, maintain, repair and bill customers for telephone service. See, e.g., Local Competition Order ¶ 505. The FCC’s Local Competition Order requires ILECs like Pacific to “unbundle” OSS, or permit CLECs to access and use their systems and personnel in a manner that is equivalent to the use the ILEC itself makes of them. Though Pacific had provided the CPUC with a cost study addressing the recurring costs of OSS, the CPUC concluded that the cost study was inadequate, and that “Pacific has not dem *1094 onstrated that these costs should be recovered from competitors.” See Second Cost Decision at 45-46 (“We find the cost studies presented by Pacific and GTEC do not demonstrate reasonable and supported expenses associated with implementing, developing and maintaining OSS systems and databases that should be directly assigned to competitors. Therefore, we reject both cost studies.”). The CPUC also stated that Pacific could seek to recover these costs in a separate proceeding before the CPUC, the Local Competition Proceeding. However, the CPUC expressly stated that it was “not predetermining that the costs presented here are eligible for recovery in that proceeding.” Id. at 46. C. OANAD Decision In D.99-11-050 (OANAD Decision) 7 , the CPUC adopted prices for Pacific to charge for UNEs, and completed the costing and pricing for Pacific’s UNEs. The parties challenge several of the rulings made in the OANAD Decision. 1. Allocation of Common Costs In the OANAD Decision, the CPUC allocated the previously determined common costs to the costs of UNEs by calculating a common cost “markup” to apply to the directly attributable UNE costs. Several parties offered various proposals to determine the common cost markup. The CPUC chose Pacific’s proposal of dividing Pacific’s forward-looking common costs by Pacific’s forward-looking direct costs of UNEs (also sometimes called “TELRIC costs” or “TELRICs”). The CPUC first used Pacific’s common costs of $996 million, determined in the First Cost Decision (with minor adjustments), as the numerator of the fraction used to determine the markup (expressed as a percentage). The CPUC then calculated the denominator, which consisted of the “total direct costs of the network elements” approved in the First Cost Decision ($4.814 billion) (with minor adjustments) plus the $875 million non-recurring cost figure approved in the Second Cost Decision, to arrive at a total of $5.189 billion in Pacific’s direct costs of UNEs (also called TEL-RICs). See OANAD Decision at 72. The common cost markup to be applied to Pacific’s direct UNE costs was then calculated by dividing the numerator ($996 million) by the denominator ($5.189 billion). This resulted in a markup of 19%. See id. The CPUC explained that Pacific’s proposal “is the most reasonable and should be adopted,” while “[ejach of the approaches suggested by other parties for computing a shared-and-common-cost markup suffers from significant infirmities.” Id. at 62-63. Plaintiffs’ approach, which the CPUC rejected and which Plaintiffs proffer again here, was to begin with the same numerator (Pacific’s common costs), but use a denominator that included not only the direct costs of UNEs but also the costs of Pacific’s other operations: retail services and Category III services (an extra $2.9 billion). Id. at 55-56. This approach would yield a common cost markup of about 13%. The CPUC determined that including Pacific’s costs of retail services and Category III services in the denominator of the markup fraction “would be both unfair and inconsistent with the TEL-RIC methodology,” because “it would be mixing apples and oranges to include retail costs in the denominator.” Id. at 64. The CPUC then quoted at length from an ar *1095 gument made by Pacific as support for its determination: [Plaintiffs have] ignored the fact that all of the shared and common costs that are retail-related have been removed from the shared and common costs identified in this phase. In D.98-02-106, the Commission explicitly addressed the issue of any retail-related dollars included in the shared and common expenses. In that decision, the Commission directed adjustments which resulted in the exclusion of any and all retail-related expenses from Pacific’s identified shared and common costs. Thus, the shared and common costs and the TELRICs adopted by the Commission exclude all retail-related costs. It is therefore entirely appropriate and proper to divide the non-retail shared and common costs by the non-retail TELRICs to obtain the non-retail minimum TELRIC markup for UNEs. Id. at 64 (quoting OANAD Decision Ex. 131-S at 5) (internal citations omitted). Plaintiffs, in response to a proposed decision by the CPUC adopting Pacific’s method, insisted as they do here that the denominator must include Pacific’s retail costs and Category III services costs as well as UNE costs, stating that the CPUC’s reasoning is “based on factual error because, as all parties including Pacific agree, no such thing as a ‘non-retail shared and common cost’ exists. Instead, the common cost number in the record of this proceeding is Pacific’s firmwide common cost.” Id. at 64 n. 63 (quoting OAN-AD Decision, AT & T/MCI Opening Comments at 16). The CPUC rejected this argument, stating: “We agree with Pacific that it is evident from an examination of D.98-02-106 that common costs not related to UNEs were removed from the common cost total adopted in that decision.” Id. The CPUC then quoted from the First Cost Decision (D.98-02-106), wherein it had held that “some of these costs [claimed by Pacific as common costs] cannot truly be considered ‘common,’ because they have a clear retail component that, under the TELRIC methodology, may not be included in the determination of wholesale UNE costs.” Id. (quoting First Cost Decision at 63). The CPUC also stated that “it would be unfair to include Category III services in the denominator, since these services have their own separate shared and common costs.” Id. at 65. The CPUC then quoted from a brief of Pacific’s, which stated that Pacific’s Chapter III businesses have their own overhead organizations, and to the extent they use Pacific’s overhead departments, the costs are directly billed to them and these billings are not reflected in Pacific’s common costs as previously determined. See id. Based on this, the CPUC concluded that because there are no Category III services costs in the determination of Pacific’s shared and common costs (the numerator), it would be improper to include any of these costs in the denominator. 2. Non-recurring Costs As stated above, the CPUC determined that the denominator of the common cost markup calculation consisted of Pacific’s costs of providing UNEs, which included recurring costs and $375 million in nonrecurring costs. See id. at 71 & n. 71. Pacific claims that the addition of the $375 million in non-recurring costs resulted in a double-counting of non-recurring costs. Pacific’s original cost study provided to the CPUC claimed a total of $4.83 billion in direct UNE costs, which included both recurring costs and a $583 million estimate of non-recurring costs. See, e.g., id. at 71 n. 71. In the First Cost Decision, the CPUC reduced this amount to $4.814 billion (after certain adjustments were made). That Decision stated that it was not addressing non-recurring costs. See First *1096 Cost Decision at 11-12. However, Pacific claims that the $583 million estimate of non-recurring costs was not removed by the CPUC in arriving at the $4,814 billion amount (primarily because it is arithmetically impossible to subtract $573 million from $4.83 billion and arrive at a result of $4,814 billion, even with the other minor adjustments made). In setting the final rates, the CPUC added the non-recurring costs determined in the Second Cost Decision, $375 million, to Pacific’s total direct costs of providing UNEs. Pacific claims the CPUC therefore double-counted Pacific’s non-recurring costs, by including the original estimate of $583 million and the later $375 million amount in the direct UNE cost denominator used to determine the common cost markup, instead of substituting $375 million for the earlier $583 million estimate. Pacific argues that this double-counting operated to inflate the denominator and therefore lower the common cost markup. 3. Risk Adder The OANAD Decision also addressed Pacific’s contention that a “risk adder” should be included as part of the cost of capital that is added to the price of Pacific’s UNEs in order to compensate Pacific for the risks associated with investing in a hypothetical, perfectly competitive market. The CPUC framed Pacific’s primary argument as follows: [Pacific] argues that the obligation to sell UNEs creates a risk for Pacific that it may not be able to recover its ‘sunk and irreversible’ investments in UNEs— i.e., that this investment may become stranded — if a CLEC purchasing UNEs suddenly decides it is time to switch customers served through those UNEs over to facilities owned by the CLEC. OANAD Decision at 19. In its Local Competition Order, the FCC stated that the currently authorized rate of return at the federal or state level is a reasonable starting point for TELRIC calculations, and incumbent LECs bear the burden of demonstrating with specificity that the business risks that they face in providing unbundled network elements and interconnection services would justify a different risk-adjusted cost of capital or depreciation rate. Local Competition Order ¶ 702. The CPUC rejected Pacific’s argument for a “risk adder” markup above the current cost of capital, for several reasons. See OANAD Decision at 37-53. First, the CPUC pointed out that Pacific did not propose that the full “risk adder” be taken on each UNE, likely because Pacific “realizes that it’s unlikely the Commission is going to go along with something that high.” Id. at 39-40. The CPUC also noted that Pacific’s witnesses “failed to offer any concrete proposal for discounting UNE prices when a CLEC agrees to purchase UNEs on a long-term basis,” even though such long-term contracts were offered as an alternative to the proposed “risk adder.” Id. at 40. Second, the CPUC found that “the investment risks Pacific will incur in the near future are more likely to be attributable to the provision of retail service than to the provision of UNEs.” Id. As the CPUC pointed out, “demand for UNEs is only one of the reasons that Pacific will be building new plant in the future, and thus is only one of many reasons why future plant might become stranded.” Id. Third, the CPUC pointed out that “regulatory requirements play at least as important a role as economic incentives in determining where and to what extent an ILEC will build facilities.” Id. at 41. Fourth, the CPUC determined that it would be inappropriate to “overprice UNEs by including a risk adder for risks *1097 that may never materialize,” and instead permitted Pacific “an opportunity to prove in the future that investment made solely to provide UNEs has become stranded because new entrants decided to switch from UNEs to their own facilities at the point when providing service through their own facilities became cost-justified.” Id. at 48. Last, the CPUC pointed out that Judge Susan Illston of this Court held that risk adders such as that proposed by Pacific “are inconsistent with the basic pricing standard contained in § 252(d)(1) of the Act.” Id. (citing AT & T Communications v. Pacific Bell, No. C 97-0080 SI, 1998 WL 246652 (May 11, 1998)). 4. OSS Recurring Costs In the OANAD Decision the CPUC again refused to set rates to compensate Pacific for recurring costs of providing OSS, because “[n]o OSS recurring costs were adopted” in the Second Cost Decision. See id. at 55 n. 54; Second Cost Decision at 45^6. The Second Cost Decision refused to determine recurring OSS costs because it rejected Pacific’s cost studies as failing to “demonstrate reason-, able and supported expenses associated with implementing, developing and maintaining OSS systems and databases.” Id. The Decision left open the possibility that Pacific could recover recurring OSS costs in a separate proceeding, if it believes that “these costs qualify for recovery as implementation costs in the Local Competition proceeding.” Second Cost Decision at 46. Pacific did not appeal the Second Cost Decision. Nor did Pacific raise the issue of its recurring OSS costs in the Local Competition Proceeding, as the CPUC suggested it might. Pacific claims that it failed to do so because the CPUC in two decisions after the Second Cost Decision suggested that it would not consider Pacific’s recurring OSS costs in the Local Competition Proceeding. See D.99-07-048 at 9; D.00-09-037 at 3 (Administrative Record, Yol. 1 Exs. 11 & 15). However, those two decisions simply restate what was said in the Second Cost Decision: Pacific “could reevaluate these costs to see if they qualified instead as infrastructure implementation costs, and if they did, submit them for consideration this Local Competition docket.” D.99-07-048 at 9. 8 5. Combination Requirements In the OANAD Decision, the CPUC first noted that the validity of the FCC’s rules in the Local Competition Order requiring ILECs to combine UNEs in ways that the ILEC may not use itself was not clear, because of pending legal challenges in the Eighth Circuit (now resolved by the Supreme Court, which upheld the validity of the combination rules). OANAD Decision at 141-44 (addressing FCC Rules 315(c)-(f), codified at 47 C.F.R. § 51.315 (c)-(0). The CPUC nevertheless concluded that it had authority under State law (Public Utilities, Code § 709.2(c)(1)) to require Pacific to create at cost-based rates new combinations of UNEs that it does not use itself, and ordered Pacific to provide such combinations upon request to all CLECs with which it has interconnection agreements. See id. at 144-45. The CPUC, acting, under authority of State law, ordered ILECs “to combine network elements in innovative ways (provided the requested combination is technically feasible, does not prejudice the rights of other CLECs, and results in adequate compen *1098 sation for the costs of providing the requested combination).” OANAD Decision at 145. This requirement, and its limitations, is identical to the FCC’s combination rules. The CPUC then held, in D.00-08-011 (Arbitration Decision), that Pacific is “required to combine [unbundled] network elements for AT & T,” 9 citing to the OAN-AD Decision. Arbitration Decision at 7. IV. Present Suit The central issue in this case is whether the CPUC properly determined, in accordance with the Act and the FCC’s implementing regulations, the rates at which Pacific must lease its UNEs to CLECs such as Plaintiffs, which rates are based on the CPUC’s determination of Pacific’s costs of providing UNEs. Plaintiffs contend that the CPUC misinterpreted and violated the requirements of the Act and the FCC’s implementing regulations when it set UNE rates for Pacific. More specifically, Plaintiffs claim that Pacific’s common costs were not calculated correctly, and that the markup for Pacific’s common costs should have been allocated to all of Pacific’s operations, not just to UNEs. Pacific and the CPUC Defendants disagree. Pacific cross-claims that the CPUC erred in several respects in setting rates. First, Pacific argues that a “risk adder” should have been included in Pacific’s costs to reflect the risk of stranded facilities. Second, Pacific claims that the CPUC erred by double-counting Pacific’s non-recurring costs, by adding $875 million to the denominator of the common cost markup instead of substituting that amount for the earlier $588 million estimate. Third, Pacific asserts that the CPUC unlawfully failed to set rates for recurring OSS costs. Fourth, Pacific suggests that the CPUC erred by ordering Pacific to provide CLECs with combinations of UNEs at cost-based prices, even where those combinations do not exist in Pacific’s network. Plaintiffs and the CPUC disagree. DISCUSSION I. Legal Standard for Summary Judgment Summary judgment is properly granted when no genuine and disputed issues of material fact remain, and when, viewing the evidence most favorably to the non-moving party, the movant is clearly entitled to prevail as a matter of law. Fed. R.Civ.P. 56; Celotex Corp. v. Catrett, 477 U.S. 317, 322-23, 106 S.Ct. 2548 , 91 L.Ed.2d 265 (1986); Eisenberg v. Ins. Co. of N. Am., 815 F.2d 1285, 1288-89 (9th Cir.1987). The moving party bears the burden of showing that there is no material factual dispute. Therefore, the Court must regard as true the opposing party’s evidence, if supported by affidavits or other eviden-tiary material. Celotex, 477 U.S. at 324 , 106 S.Ct. 2548 ; Eisenberg, 815 F.2d at 1289 . The Court must draw all reasonable inferences in favor of the party against whom summary judgment is sought. Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587 , 106 S.Ct. 1348 , 89 L.Ed.2d 538 (1986); Intel Corp. v. Hartford Accident & Indem. Co., 952 F.2d 1551, 1558 (9th Cir.1991). Material facts which would preclude entry of summary judgment are those which, under applicable substantive law, may affect the outcome of the case. The substantive law will identify which facts are material. Anderson v. Liberty Lobby, Inc., 477 *1099 U.S. 242, 248, 106 S.Ct. 2505 , 91 L.Ed.2d 202 (1986). II. Standard of Review of the CPUC’s Decisions Under § 252(e)(6) of the Act, the Court has jurisdiction to review a State commission’s decision to “determine whether the agreement or statement meets the requirements of section 251 of this title and this section.” The Court will consider de novo whether the CPUC’s decisions are in compliance with the Act and the FCC’s implementing regulations, US West Communs. v. MFS Intelenet, Inc., 193 F.3d 1112 , 1117 (9th Cir.1999), and will review factual determinations and all other issues under an arbitrary and capricious standard. Id. III. Plaintiffs’ Motion for Summary Judgment Plaintiffs move for summary judgment, seeking to overturn the CPUC’s determination of Pacific’s common costs and the determination of the common cost markup. A. The CPUC’s Determination of the Amount of Pacific’s Firm-wide Common Costs Plaintiffs argue that the CPUC improperly determined that Pacific’s firm-wide common costs were $996 million. 10 The CPUG agreed to reduce the previous common cost total by $68 million, which resulted in the $996 million total, in response to Plaintiffs’ request that over $200 million be removed from common costs. 11 Plaintiffs claim that the CPUC should have reduced the common cost total more than it did because Plaintiffs showed that over $200 million of the common costs were directly attributable to retail services or were “avoided retail costs” that must be excluded. Because this issue requires interpreting the FCC’s regulations, the Court reviews the CPUC’s decision on this matter de novo, while reviewing the CPUC’s underlying factual determinations under an arbitrary and capricious standard. The CPUC’s task was to calculate the common costs (costs that could not be attributed to a particular UNE or group of UNEs) that Pacific would experience in a forward-look *1100 ing, hypothetical environment in which Pacific operated a “wholesale network.” Local Competition Order ¶ 694. Another way to frame the issue is as follows: how much money would Pacific save in overhead and administrative expenses, such as the salaries of executive officers, if Pacific ceased to operate a retail business and became a wholesaler (leasing UNEs to CLECs)? Plaintiffs proposed that Pacific would save approximately $200 million in common costs such as overhead expenses. The CPUC disagreed, and instead reduced seven expense accounts (such as marketing, customer service, maintenance, support expenses, executive/planning, and general & administrative) by a certain percentage in order to reduce Pacific’s forward-looking common costs by the amount that the CPUC believed Pacific would save in a wholesale-only environment in which Pacific did not operate a retail business. See First Cost Decision App. A. Some expense accounts, such as marketing and customer service, were reduced by as much as 93% (because these primarily serve retail operations), while most were reduced by approximately 8%. See id. This resulted in a reduction of Pacific’s common costs by $68 million. The CPUC stated that Plaintiffs arrived at the over $200 million amount that they requested be removed by “multiplying the amount of overhead expenses claimed by Pacific for each of seven [expense] accounts by a percentage that they claim is retail-related.” Id. at 62. The Court finds that the CPUC did not err in calculating Pacific’s common costs. The CPUC has expertise in this area, and made an extremely fact-based and policy-based decision to reduce the expense accounts by percentages other than those proposed by Plaintiffs. Determining the forward-looking common costs of an ILEC is a necessarily hypothetical exercise that relies on speculation regarding the amount of an ILEC’s common costs that are attributable to retail services, and would not exist in a forward-looking, wholesale environment. Plaintiffs have failed to present a cogent argument to overturn the CPUC’s decision to reduce Pacific’s common costs by $68 million. 12 Plaintiffs also argue that the CPUC erred by failing to exclude from Pacific’s common costs “avoided retail costs” as described in FCC Rule 51.609. See 47 C.F.R. § 51.609 . Plaintiffs claim that Rule 51.505(d)(2) requires that the CPUC conduct an avoided retail cost analysis and exclude all avoided retail costs from common costs. 13 See 47 C.F.R. § 51.505 (d)(2). However, nothing in Rule 51.505(d)(2) requires that the CPUC conduct an avoided retail cost determination in calculating common costs. That Rule merely provides: The following factors shall not be considered in a calculation of the forward-looking economic cost of an element: (2) Retail Costs. Retail costs include the costs of marketing, billing, collection, and other costs associated with offering retail telecommunications services to *1101 subscribers who are not .telecommunications carriers, described in § 51.609. 47 C.F.R. § 51.505 (d)(2). Rule 51.609 is used to calculate actual retail costs avoided when an ILEC sells those services to another carrier for resale, and is not conducted in a hypothetical, forward-looking environment as is the case with common costs. Further, the Rule 51.609 standard as applied to resale rate proceedings has been held unlawful and vacated by the Eighth Circuit. See Iwoa Utils. Bd. II, 219 F.3d at 755-56 , cert. granted in part, denied in relevant part, 531 U.S. 1124 , 121 S.Ct. 878 , 879, 148 L.Ed.2d 788 (2001). Therefore, it would be unwise to hold that the avoided retail cost determination of Rule 51.609 must be applied here. The Court holds that the CPUC did not err in failing to conduct an avoided retail cost analysis in calculating Pacific’s common costs. The Court affirms the CPUC’s calculation of Pacific’s common costs, including its decision to reduce Pacific’s common costs by $68 million to exclude retail-related common costs. B. The CPUC’s Allocation of All of Pacific’s Common Costs to Unbundled Network Elements (UNEs) The FCC’s Local Competition Order states that “forward-looking common costs shall be allocated among elements and services in a reasonable manner, consistent with the pro-competitive goals of the 1996 Act. One reasonable allocation method would be to allocate common costs using a fixed allocator, such as a percentage markup over the directly attributable forward-looking costs.” Local Competition Order ¶ 696. Plaintiffs argue that the CPUC erred in allocating Pacific’s common costs by calculating the common cost markup in the way it did. The common cost markup was calculated by dividing Pacific’s common costs ($996 million) by the direct costs of UNEs ($5,189 billion). For purposes of this analysis, the actual amount of Pacific’s common costs and direct UNE costs is not important. The question Plaintiffs raise is whether the CPUC properly used Pacific’s direct UNE costs as the denominator of the common cost markup calculation, instead of including the direct costs of all of Pacific’s operations, including retail services and Category III services. 14 Plaintiffs argue that, in order reasonably and properly to allocate Pacific’s firm-wide common costs to the costs of UNEs, the common costs must be allocated to all of Pacific’s operations firm-wide. To do this, Plaintiffs claim, the CPUC should have included the direct costs of all of Pacific’s operations, including retail services and Category III services, as well as UNE costs, in the denominator. Plaintiffs argue that by including only direct UNE costs in the denominator, the CPUC effectively allocated all of Pacific’s firm-wide common costs exclusively to the cost of UNEs. Pacific and the CPUC assert that the CPUC correctly included only the direct costs of UNEs (that is, TELRICs) in the denominator of the common cost markup. They argue that because the numerator (Pacific’s common costs) excluded retail costs, it is “entirely appropriate and proper to -divide the non-retail shared and common costs by the non-retail TELRICs to obtain the non-retail minimum TELRIC markup for [UNEs].” 15 OANAD Decision at 64 (quoting Scholl Reply). Pacific and *1102 the CPUC assert that it would be “mixing apples and oranges to include retail costs in the denominator,” because the First Cost Decision excluded all directly attributable retail costs from the common cost numerator. Id. Pacific and the CPUC argue that Pacific’s firm-wide common costs reflect the costs that Pacific would incur in a hypothetical environment in which Pacific is solely a wholesaler, and conducts no retail activity. Because Pacific’s common costs are the hypothetical firm-wide common costs of Pacific as a wholesaler, they should only be divided by Pacific’s direct UNE costs, and not by its retail and Category III services (which would no longer exist in a wholesale-only environment). The Court concurs with the reasoning of Pacific and the CPUC. As discussed above, the TELRIC methodology calculates an ILEC’s costs in a hypothetical, forward-looking environment in which the ILEC is solely a wholesaler. See, e.g., Local Competition Order ¶ 694. The methodology embraced by the FCC excludes all retail costs from the calculation of common costs and the common cost markup, because retail costs should not be included in the prices that ILECs charge CLECs to lease UNEs. See id. ¶ 691, 694. The CPUC calculated Pacific’s common costs by removing all costs attributable to retail services, in order to arrive at the common-costs that Pacific would experience in a wholesale-only environment. In other words, Pacific’s firm-wide common costs are the common costs that Pacific would experience as a wholesale firm. In such a wholesale-only environment, it would make no sense to divide Pacific’s common costs by anything more than Pacific’s direct costs of UNEs. Pacific’s firm-wide, wholesale-only common costs are reasonably divided by only the direct costs of UNEs, and not to Pacific’s retail services or Category III services, which would not exist in a wholesale-only environment. The FCC’s Local Competition Order supports this conclusion, by stating that “[o]ne reasonable allocation method would be to allocate common costs using a fixed allocator, such as a percentage markup over the directly attributable forward-looking costs.” Local Competition Order ¶ 696. Pacific’s directly attributable forward-looking costs, in a hypothetical environment in which Pacific is solely a wholesaler, consist of only the direct costs of UNEs, and must exclude retail services and Category III services. The CPUC’s division of Pacific’s common costs only by the direct costs of UNEs was a reasonable method of calculating a common cost markup. 16 The CPUC’s determination of the common cost *1103 markup is affirmed. Plaintiffs’ motion for summary judgment is DENIED. IV. Pacific’s Cross-Motion for Summary Judgment Pacific cross-moves for summary judgment, arguing that the CPUC erred in several respects in the OANAD proceeding. A. Risk Adder Pacific asserts that the CPUC failed to include a risk adder as part of the cost of capital associated with a hypothetical, forward-looking network, to account for the risk of stranded facilities. The FCC’s Local Competition Order provides that “the currently authorized rate of return at the federal or state level is a reasonable starting point for TELRIC calculations, and incumbent LECs bear the burden of demonstrating with specificity that the business risks that they face in providing unbundled network elements and interconnection services would justify a different risk-adjusted cost of capital or depreciation rate.” Local Competition Order ¶ 702. The CPUC rejected Pacific’s argument for a risk adder, for several reasons: (1) Pacific faded to propose that the risk adder be taken on each UNE, and failed to offer a proposal for discounting UNE prices when a CLEC agrees to a long-term purchase, (2) Pacific’s investment risks are more likely to be attributable to providing retail services than to providing UNEs, (3) regulatory requirements play at least as important a role as economic incentives in determining where and to what extent an ILEC will build facilities, and (4) it would be inappropriate to overprice UNEs by including a risk adder for risks that may never materialize. See OANAD Decision at 39-43. Pacific argues that the CPUC rejected the risk adder in large part because Pacific failed to demonstrate that it would, in reality, experience increased borrowing costs associated with providing wholesale UNEs. Pacific claims that the CPUC instead should have assumed a hypothetical market. In essence, Pacific quarrels with the FCC’s rule that requires ILECs to demonstrate with specificity the risks that they face in providing UNEs in order to increase the cost of capital. However, the Supreme Court in Verizon upheld the FCC’s rule regarding risk-adjusted cost of capital, stating that “whether the FCC’s assumption about adequate risk adjustment was based on hypothetical or actual competition, it seems fair to say that the rate ... mentioned by the FCC ... is a ‘reasonable starting point’ for return on equity calculations based on the current lack of significant competition in local-exchange markets.” 122 S.Ct. at 1677-78 . The Supreme Court did not appear concerned about whether a proposed risk adder “was based on hypothetical or actual competition,” stating that rates were “to be adjusted upward if the incumbents demonstrate the need.” Id. The Court refused to invalidate the FCC’s rule, rejecting Pacific’s argument that “the TEL-RIC rules limit the depreciation and capital costs that ratesetting commissions may recognize.” Id. The Court found that “TELRIC itself prescribes no fixed percentage rate as risk-adjusted capital costs,” and “the FCC committed considerable discretion to state commissions on these matters.” 17 Id. *1104 As indicated above, the CPUC had several reasons for rejecting a risk adder, which were primarily based on the small likelihood of increased risks associated with providing UNEs to CLECs. The CPUC found that Pacific had not met its “burden of demonstrating with specificity that the business risks” that it faces justify a higher cost of capital, as required by the Local Competition Order. The CPUC properly rejected Pacific’s argument because it was not supported by the evidence submitted by Pacific. To the extent that Pacific is arguing that the Local Competition Order and the TELRIC methodology employed therein are faulty because they fail to utilize a hypothetical market, this would be an impermissible collateral attack on the TELRIC methodology, which was resolved against Pacific by the Supreme Court, and which a district court in any event may not entertain. 18 The CPUC’s decision not to include a risk adder is consistent with the FCC’s Local Competition Order, and is not arbitrary and capricious. Therefore, the Court affirms the CPUC’s decision to reject a risk adder. B. Double-Counting Non-recurring Costs Pacific argues that the denominator of the common cost markup was inflated, and the markup thereby deflated, by a double-counting of Pacific’s non-recurring costs. In the First Cost Decision, the CPUC determined that Pacific’s recurring costs of providing UNEs totaled $4,814 billion. In the Second Cost Decision, the CPUC determined that Pacific’s non-recurring costs of providing UNEs was $375 million. This amount was added to Pacific’s recurring UNE costs in the OANAD Decision, to arrive at the total direct costs of providing UNEs. However, Pacific had originally included an estimate of UNE non-recurring costs, of about $583 million, in its initial UNE cost studies that resulted in the recurring UNE cost total of $4.814 billion. Pacific claims that the $375 million amount should have been substituted for the $583 million estimate. It does not appear that this original estimate of non-recurring costs was ever removed, even though the First Cost Decision stated that it only addressed recurring costs. The CPUC’s primary argument that there was no double-counting is that it has satisfied itself that there was not. The CPUC claims that because the First Cost Decision only addressed recurring costs, the costs determined in that Decision could not have included any non-recurring costs, which were determined in a later proceeding. However, this circular argument fails in the face of simple arithmetic. Pacific’s initial cost studies reflected total UNE costs of $4.83 billion, and included an estimate of non-recurring costs of $583 million. See OANAD Decision at 71 n. 71. Considering these cost studies, the CPUC found that Pacific’s recurring costs total $4.814 billion. It is clear that the $583 million non-recurring cost amount was still included in the $4.814 billion result, because it is impossible to subtract $583 million from $4.83 billion and arrive at $4.814 billion, even considering the other minor adjustments made by the CPUC. Further, the CPUC adopted the $4.814 billion amount from Pacific’s revised cost *1105 studies after making adjustments which did not include removing the non-recurring cost estimate. See, e.g., First Cost Decision at 103, 106. In the face of this convincing evidence, the CPUC has provided no reasoning for its claim that it did not double-count other than its own assertion that it has satisfied itself that it did not. The CPUC failed to remove the original $583 million estimate of non-recurring costs when it added its $375 million non-recurring cost determination to Pacific’s costs of providing UNEs. This was arbitrary and capricious, given the evidence submitted by Pacific, and the CPUC’s failure to provide any support for its position that it did not double-count. The CPUC’s determination of Pacific’s direct cost of providing UNEs (the denominator of the common cost markup), and any decision which relies on this determination, must be vacated and remanded, so that the double-counting can be remedied. C. Rate for Recurring OSS Costs Pacific claims that the CPUC unlawfully failed to set a recurring rate for OSS costs in the Second Cost Decision. Pacific argues that the Second Cost Decision suggests that Pacific could recover recurring OSS costs in the Local Competition Proceeding, but Pacific was then barred from doing so at a later date. Pacific claims that the CPUC erred by failing to set OSS recurring rates, simply because it was unsatisfied with Pacific’s cost studies. However, the Court does not read the Second Cost Decision in the manner that Pacific does. First, the Second Cost Decision does not simply leave the issue of recurring OSS costs to another day. It explicitly rejects Pacific’s request for a determination of rates for OSS costs because the cost studies were insufficient: “We find the cost studies presented by Pacific and GTEC do not demonstrate reasonable and supported expenses associated with implementing, developing and maintaining OSS systems and databases that should be directly assigned to competitors. Therefore, we reject both cost studies.” Second Cost Decision at 46. This is a substantive decision by the CPUC that Pacific failed to support its request with adequate cost studies. The CPUC explained: “Pacific has not demonstrated that these costs should be recovered from competitors.” Id. at 45. Pacific failed to apply for rehearing of the Second Cost Decision, which prevents it from seeking relief in this Court. Second, the CPUC did not promise Pacific that it could seek recovery of OSS recurring costs in the Local Competition Proceeding, as Pacific contends. The CPUC did say that Pacific “may seek to recover any eligible costs” through the Local Competition Proceeding, but only if it “believe[s] these costs qualify for recovery as implementation costs.” Id. at 46. As Pacific has pointed out, implementation costs are non-recurring costs, which are clearly different from recurring costs. Therefore, Pacific was put on notice that it likely could not seek recurring OSS costs in the Local Competition Proceeding. The CPUC made this clearer, by stating that “we are not predetermining that the costs presented here are eligible for recovery in that proceeding.” Id. The CPUC also stressed that it had “discussed here our concern regarding the reasonableness of the costs,” and directed Pacific to review and document its OSS costs carefully before seeking recovery in another proceeding. 19 This provides more support for the CPUC’s argument that Pacific should have *1106 applied for rehearing of the Second Cost Decision if it wanted to challenge the CPUC’s failure to set a rate for recurring OSS costs. Because the CPUC rejected Pacific’s request to set a rate for OSS recurring costs in the Second Cost Decision, and Pacific never challenged that ruling, it cannot do so here. The CPUC’s decision not to set a rate for OSS recurring costs is affirmed. D. Requiring Pacific to Provide New Combinations of UNEs Pacific argues that the CPUC erred by ordering Pacific to provide CLECs with combinations of network elements at cost-based prices, even where those combinations do not exist in Pacific’s network. The Eighth Circuit has twice held that the FCC’s combination rules, which require ILECs to provide CLECs with new combinations of network elements at cost-based prices, violate the plain language of the Act, Iowa Utils. Bd. I, 120 F.3d at 813 (invalidating Rules 315(b)-(f)); Iowa Utils. Bd. II, 219 F.3d at 758-59 (reaffirming invalidation of Rules 315(c)-(f)), but the Supreme Court in Verizon reversed the Eighth Circuit, finding the combination rules valid. See Verizon, 122 S.Ct. at 1682-87 (reversing Eighth Circuit’s ruling in Iowa Utilities Board II that invalidated Rules 315(c)-(f)); see also AT & T Corp. v. Iowa Utils. Bd., 525 U.S. at 393 , 119 S.Ct. 721 (reversing Eighth Circuit’s ruling in Iowa Utilities Board I that invalidated Rule 315(b), but not addressing Rules 315(c)-(f)). The Supreme Court upheld the FCC’s combination rules as a reasonable exercise of the FCC’s discretion. Pacific concedes that the Supreme Court’s decision upholds the FCC’s rules regarding combination requirements. However, Pacific argues that the combination rule adopted by the CPUC in the OANAD Decision, and the contract language imposed by the CPUC to enforce that ruling in an agreement between Pacific and AT & T, is broader than the FCC’s rule approved by the Supreme Court, in that it requires Pacific to provide AT & T with “any combination of Network Elements requested by AT & T to serve its customers.” Pacific Supp. Briefing, Ex. C at 6 (Pacific-AT & T Interconnection Agreement Att. 6 § 3.1 (Aug. 14, 2000)). As the Supreme Court clarified, the FCC’s rule requires an ILEC to combine elements “when the requesting carrier is unable to combine, when it would not place the incumbent at a disadvantage in operating its own network, and when it would not place other competing carriers at a competitive disadvantage.” Verizon, 122 S.Ct. at 1687 (internal citations omitted). Rule 315(c) states that an ILEC need only combine elements if the combination is “(1) technically feasible; and (2) would not impair the ability of other carriers to obtain access to unbundled network elements or to interconnect with the incumbent LEC’s network.” 47 C.F.R. § 51.315 (c). Pacific’s argument that the CPUC’s combination requirements are broader than the FCC rule is not well-taken. In the OANAD Decision, the CPUC expressly included the limitations contained in the FCC’s combination rules, stating: We ... conclude that our unbundling authority under California law includes the power to order ILECs to combine network elements in innovative ways (provided the requested combination is technically feasible, does not prejudice the rights of other CLECs, and results in adequate compensation for the costs of providing the requested combination). OANAD Decision at 145. In its Arbitration Decision, the CPUC held that Pacific is “required to combine [unbundled] network elements for AT & T,” citing the OANAD Decision. Arbitration Decision at 7. The agreement between Pacific and AT *1107 & T which Pacific argues requires Pacific to provide AT & T with “any combination of Network Elements requested by AT & T to serve its customers” also states, in the Introduction section, that it “sets forth the unbundled Network Elements and Combinations of unbundled Network Elements (‘Combinations’) that PACIFIC agrees to offer to AT & T in accordance with its obligations under Section 251(c)(3) of the Act and Applicable Laws,” and specifically states that this' “includes, without limitation, the FCC’s First Report and Order [Local Competition Order].” Pacific Supp. Briefing, Ex. C at 1. In other words, this agreement expressly adopts the FCC’s combination rules. There is no evidence that the CPUC, in any decision before the Court, has required Pacific to combine elements in a manner that is broader than that required by the FCC’s combination rules, which have been approved by the Supreme Court in •Verizon. The combination requirement enunciated by the CPUC in the OANAD Decision, and applied in the Arbitration Decision, is consistent with the Act and the FCC’s combination rules. 20 The CPUC’s decision on this issue is affirmed. CONCLUSION For the foregoing reasons, Plaintiffs’ motion for summary judgment is DENIED. The CPUC’s determination of Pacific’s common costs and its calculation of Pacific’s common cost markup are affirmed. Pacific’s cross motion for summary judgment is GRANTED in part and DENIED in part. The CPUC’s calculation of Pacific’s total direct costs of UNEs is vacated and remanded to correct the double-counting of non-recurring costs. The relevant CPUC decisions that depend upon the incorrect calculation of Pacific’s total direct costs of UNEs are vacated and remanded as noted above. IT IS SO ORDERED. 1 . A “network element” is a facility or equipment used in the provision of telecomnaunica- *1089 tions service. See 47 U.S.C. § 153 (29). 2 . “Common costs” are sometimes called "shared common” costs or "firm-wide common” costs, and are referred to by the CPUC as "shared and common” costs. A separate category of costs described above, often referred to as "shared family” costs, are costs that are common to a certain group or subset of an ILEC’s operations, such as UNEs. These costs are described as a form of common costs in the first portion of the quoted material above. However, reference to "common costs” in the Court's Order refers only to the first concept (firm-wide common costs), and not to shared family costs, which can be directly attributed to a subset of elements or services such as UNEs. 3 . Interim Decision Adopting Cost Methodology, Evaluating the Hatfield Computer Model, and Deciding Other Issues Related to Cost Studies of Pacific Bell's System, Rulemaking on the Commission’s Own Motion to Govern Open Access to Bottleneck Services and Establish a Framework for Network Architecture Development of Dominant Carrier Networks, D.98-02-106 (Feb. 19, 1998) (Administrative Record, Vol 1 Ex. 6). 4 . The First Cost Decision also addressed the amount of Pacific’s "shared family” costs of network elements (i.e., costs common only to network elements, but not to firm-wide operations). 5 . Rulemaking on the Commission s Own Motion to Govern Open Access to Bottleneck Services and Establish a Framework for Network Architecture Development of Dominant Carrier Networks, D.98-12-079 (Dec. 17, 1998) (Administrative Record, Vol. 1 Ex. 9). 6 . Non-recurring costs were defined by the CPUC as “one-time expenses associated with initiating or disconnecting a service.” See Second Cost Decision at 12. 7 . Interim Decision Setting Final Prices for Network Elements Offered by Pacific Bell, Rulemaking on the Commission's Own Motion to Govern Open Access to Bottleneck Services aná Establish a Framework for Network Architecture Development of Dominant Carrier Network, D.99-11-050 (Nov. 18, 1999) (Administrative Record, Vol. 1 Ex. 12). 8 . The CPUC also explained: "We expressly warned [in the Second Cost Decision], however, that this opportunity to reevaluate the costs should not be misconstrued as a prejudgment by the Commission that the costs would be recoverable from customers.” D.99-07-048 at 9. 9 . Opinion, Application by AT & T Communications of California, Inc., et al. (U 5002 C) for Arbitration of an Interconnection Agreement with Pacific Bell Telephone Company (U1001 C) Pursuant to Section 252(b) of the Telecommunications Act of 1996, D.00-09-011 (Aug. 3, 2000) (Administrative Record, Vol. 1 Ex. 13). 10 . Pacific claims that this argument was not raised in the Complaint. However, the Complaint alleges: "In applying the FCC’s TEL-RIC pricing rules to Pacific Bell’s unbundled network elements, the PUC erroneously included in Pacific Bell’s prices a massively inflated amount of 'common costs.’ " Complaint ¶ 9. The Complaint also alleges that these common cost figures "reflected many costs that were directly attributable to Pacific Bell’s retail and other operations and thus were not common costs at all.” Id. ¶ 34 . The allegations in the Complaint are sufficient to put Pacific on notice of Plaintiffs’ argument regarding the calculation of the amount of common costs. 11 . Plaintiffs argue that the CPUC’s reason for declining to exclude the $200 million amount, that there "is as yet no empirical evidence to support” Plaintiffs' assumption regarding avoidable overhead costs, is improper because Pacific has the burden of proof of the "specific nature and magnitude” of the common costs it claims. However, the burden of proof cited by Plaintiffs only applies to arbi-trations, which this is not. Local Competition Order ¶ 695. Even if such a burden applied, Plaintiffs have not shown that Pacific did not meet its initial burden of providing support for its calculation of common costs. Plaintiffs are simply arguing that Pacific has failed adequately to respond to Plaintiffs’ challenge of Pacific’s costs. This is not a question of burden of proof, but of the substantive merits of each party’s argument. See also Verizon, 122 S.Ct. at 1678 ("TELRIC rate proceedings are surprisingly smooth-running affairs, with incumbents and competitors typically presenting two conflicting economic models supported by expert testimony, and state commissioners customarily assigning rates based on some predictions from one model and others from its counterpart.”). 12 . The CPUC relied in part on its factual determination that there "is as yet no empirical evidence to support” Plaintiffs’ assumption that "in a forward-looking environment, Pacific will be able to avoid substantial amounts of overhead (such as general, administrative and executive expenses) related to its retail operations.” First Cost Decision at 63. The Court reviews such factual determinations under an arbitrary and capricious standard. Plaintiffs have failed to show that the CPUC's finding on this issue was arbitrary and capricious. 13 . Avoided retail costs are costs "that an incumbent LEC would no longer incur if it were to cease retail operations.” Local Competition Order, ¶ 911. 14 . To the extent that this is an issue of interpretation of the FCC's rules and regulations,, the Court will review the CPUC’s decision de novo. The Local Competition Order requires that allocation of common costs be “reasonable.” A plainly unreasonable allocation method is therefore reversible. 15 . The CPUC supported its allocation method on the basis of its discretion to set rates for an ILEC’s UNEs, and its expertise. However, the Local Competition Order requires that the method of allocation be "reasonable.” General assertions of discretion and expertise will not justify a decision which is unreasonable. 16 . Pacific and the CPUC also argue that Category III services costs should not be included in the denominator for an additional reason. Pacific argued below, and the CPUC agreed, that Category III services should not be included because no Category III costs are included in Pacific’s common costs. Pacific asserts that its Category III services have their own overhead organizations, and to the extent they use Pacific's overhead departments, the costs are directly billed to them and such billings were removed by Pacific from the common costs figures. See id. at 65-66. Plaintiffs argue that there is no evi-dentiary basis for the CPUC's finding that Pacific's common costs do not include any Category III costs. However, the CPUC supports its finding with testimony from one of Pacific’s witnesses, Richard Scholl, who the CPUC quoted as follows: "[TJhere are no shared and common costs of Category III and non-regulated services in the shared and common costs identified in Pacific Bell’s TELRIC study.” Id. at 66 (quoting Scholl Reply). This is sufficient to support the CPUC’s finding. Given this determination that all Category III services costs were billed directly to Category III services and removed from the common costs, it is reasonable to exclude Category III services costs from the denominator for this reason as well. 17 . In its supplemental brief in light of the Supreme Court's Verizon decision, Pacific argues for the first time that the CPUC erred in adopting the ten percent cost of capital that it adopted prior to considering the risk adder, because that figure was taken from a 1989 CPUC decision. See OANAD Decision at 28-29. However, on appeal Pacific only raised the issue of the failure of the CPUC to include a “risk adder” on top of the adopted cost of capital. Pacific cannot raise this new issue of *1104 the amount of the adopted cost of capital (prior to consideration of a risk adder) at this late date. 18 . The Hobbs Act grants the courts of appeals exclusive jurisdiction over all reviewable final orders of the FCC. See 28 U.S.C. § 2342 (1); 47 U.S.C. § 402 (a); Wilson v. A.H. Belo Corp., 87 F.3d 393, 396-97 (9th Cir.1996). A complaint filed in district court that seeks relief that cannot be granted unless a final FCC order is enjoined, set aside, suspended or held to be invalid should be dismissed for lack of subject matter jurisdiction. See id. at 397 . 19 . Pacific chose not to raise the issue of recurring OSS costs in the Local Competition Proceeding. 20 . Further, the CPUC may impose requirements under State law that are not inconsistent with the Act. Because of the uncertainty regarding the validity of the FCC's combination rules at the time of the OANAD Decision, the CPUC based its decision to impose a combination requirement on its independent authority under State law. See 47 U.S.C. § 252 (e)(3), § 253(b); see also OANAD Decision at 144-45; Arbitration Decision at 7.