MCI WorldCom Communications, Inc. v. Bellsouth Telecommunications, Inc.

PRYOR, Circuit Judge: The key issue in this appeal is whether the Florida Public Service Commission complied with the Telecommunications Act of 1996 and corresponding federal regulations when it approved substantial parts of the pricing plan for the lease of telecommunications equipment urged by BellSouth Telecommunications, Inc. BellSouth and the Florida Commission appeal a declara *1166 tory judgment that invalidated part of the pricing plan approved by the Florida Commission. They argue that the district court erroneously held that the BellSouth Telecommunications Loop Model failed to adhere to the Telecommunications Act and federal regulations because the pricing plan used multiple “scenarios” instead of the single most efficient, lowest cost network configuration to calculate the rate for the lease of wire loops. MCI WorldCom Communications, Inc., and Florida Digital Network, Inc., cross-appeal and argue that the district court erroneously approved the inflation factor used in the BellSouth model. Florida Digital Network also argues that the district court erroneously approved the geographic cost-based deaver-aging model adopted by the Florida Commission. We conclude that the district court erred when it determined federal law forbids the use of multiple scenarios, and we remand this action to the district court to evaluate whether each scenario in the pricing model approved by the Florida Commission complies with federal law. We also conclude that the Florida Commission did not err when it approved the inflation factor and the geographic cost-based deaveraging model. We reverse and remand in part and affirm in part. I. BACKGROUND To explain the background of this- appeal, we address four matters. First, we describe the technology relevant to this appeal. Second, we provide an overview of the regulatory scheme. Third, we describe the pricing model adopted by the Florida Commission. Fourth, we outline the procedural history of this appeal. A. An Overview of the Relevant Technology A local telephone network consists of several elements, and three of these components are central to this appeal. The first element of a local telecommunication network is its wire loops, also known as local loops. Wire loops are the telephone wires that connect each residential customer to the network of the local carrier. Loops are made of either copper or fiber optic wire, and the capabilities and cost of the loop are dependent on its type. Although copper wire is less expensive than fiber optic wire for short loops, fiber optic is more cost-efficient for longer loops. Some services such as Digital Subscriber Line technology (DSL, a type of high-speed internet service), can be offered only over copper wire, notwithstanding its potentially higher cost. The second element of a local telecommunications network is its switches. Local loops connect to switches, which are computers that route calls on the network. When the wire loop is fiber optic, the switch-loop combination can be either “integrated” or “universal.” In an “integrated digital loop carrier,” the switch and wire loop operate as one unit because the wire loop is integrated directly into the switch. In a “universal digital loop carrier,” the local loop and the switch are independent. For universal digital loop carrier technology, the lessee of the loop may provide its own switch, but for integrated digital loop carrier technology, the lessee must use the switch-loop combination of the lessor because it is cost-prohibitive to decouple the wire loop from the switch. The third element of a local telecommunications network is its wire centers. Wire centers are where the switches are located. Wire centers act as a bridge between the wire loops and the central office of the carrier, which allows long distance calls to be placed. B. The Telecommunications Act Before the Telecommunications Act became law, most areas were served by a *1167 single local exchange carrier, now known as the “incumbent local exchange carrier.” See AT&T Corp. v. Iowa Utils. Bd., 525 U.S. 366, 371, 119 S.Ct. 721, 726, 142 L.Ed.2d 835 (1999). Over the years, the incumbent local carrier constructed hardware networks to deliver residential and commercial telephone service to the area. Id. Because they were without competition and were often compensated based on how much they spent (the “rate-of-return method”), incumbent local carriers had an incentive to construct networks that were inefficient. See Nat’l Rural Telecom Ass’n v. FCC, 988 F.2d 174, 178 (D.C.Cir.1993). The Telecommunications Act was enacted to “uproot[ ] the monopolies that traditional rate-based methods had perpetuated.” Verizon Commc’ns Inc. v. FCC, 535 U.S. 467, 488, 122 S.Ct. 1646, 1660, 152 L.Ed.2d 701 (2002). The Act preempted state laws that protected local monopolies, and it imposed on local carriers affirmative duties to facilitate market entry by new local carriers, known as “competitive local exchange carriers.” See AT&T Corp., 525 U.S. at 371, 119 S.Ct. at 726. Central to this appeal is the duty of an incumbent local carrier to provide access to its network to competitive local carriers. See 47 U.S.C. § 251. The Telecommunications Act requires incumbent carriers to make available to potential competitors their “unbundled network elements.” Id. § 251(c)(3). The Act encourages incumbent and competitive local carriers to negotiate access rates. Id. § 251(c)(1). In the event an agreement cannot be reached, any party may petition the state telecommunications commission to arbitrate any open issues. Id. § 252(b)(1). Once arbitration has been invoked, the state commission must adhere to federal law when it sets the rates. See id. § 252(e)(1); Verizon Cal. Inc. v. Peevey, 413 F.3d 1069, 1071 (9th Cir.2005) (citing AT&T Corp., 525 U.S. at 385, 119 S.Ct. at 733). Congress delegated to the Federal Communications Commission the authority to promulgate regulations that govern the setting of rates. See 47 U.S.C. § 251(d)(1). The methodology the FCC selected is called the Total Element Long-Run Incremental Cost method. 47 C.F.R. § 51.505. The TELRIC of an element is the “forward-looking cost over the long run” of an element, “taking as a given the incumbent LEC’s provision of other elements.” Id. § 51.505(b). The TELRIC must be measured “based on the use of the most efficient telecommunications technology currently available and the lowest cost network configuration.” Id. § 51.505(b)(1). The regulations also require geographically-deaveraged rates: “State commissions shall establish different rates for elements in at least three defined geographic areas within the state to reflect geographic cost differences.” Id. § 51.507(f). Congress delegated to each state commission the authority to approve the interconnection agreement, including the pricing of unbundled network elements. See 47 U.S.C. § 252(e). The Telecommunications Act also allowed incumbent local carriers to participate in the long-distance service market upon approval of the FCC. Id. § 271(d)(3). In a section 271 proceeding, the FCC permits the incumbent local carrier to enter the long-distance market only after the carrier implements the “competitive checklist,” one item of which is “[n]on-discriminatory access to network elements in accordance with the requirements of sections 251(c)(3) and 252(d)(1) of this title.” Id. § 271 (e)(2)(B)(íí). Section 271 proceedings are streamlined; the FCC must approve or deny the petition of the local carrier within 90 days of receiving it. Id. § 271(d)(3). *1168 C. The Pricing Model Adopted by the Florida Commission This appeal arises from the pricing model approved by the Florida Commission that sets the rates for the use of the network elements of BellSouth. The pricing model adopted by the Florida Commission is largely based on the BellSouth Telecommunications Loop Model. Three aspects of the model adopted by the Florida Commission are relevant to this appeal. First, the BellSouth model employs three “scenarios” that model the different types of wire loops instead of a unitary network comprised of all three types of wire loops. Second, the BellSouth model incorporates an “inflation factor” to account for its cost of capital. Third, the model approved by the Florida Commission creates three tiers for geographic cost-based deaveraging. The first aspect of the BellSouth model is its use of three scenarios to compute the TELRIC of each unbundled network element. The first scenario is “Copper Only”: all loops in this hypothetical network are required to be copper. Bell-South maintains that this scenario reflects the type of network that a competitive local carrier would require to provide DSL service. The second scenario is BST2000. The BST2000 scenario uses copper for wire loops up to 12,000 feet and uses fiber optic for loops of longer lengths. In the BST2000 scenario, all fiber optic switch-loop combinations are universal rather than integrated. This technology allows the competitive local carrier to lease the loop but to supply its own switch (i.e., the loop “stands alone”). BellSouth maintains that this scenario reflects the type of network that a competitive local carrier would require if it sought to provide its own switches and wished to lease only the wire loops of the incumbent local carrier. The third scenario is referred to as the “Combo” scenario. The Combo scenario uses only integrated digital loop carrier technology for its fiber optic switch-loop combinations. The competitive local carrier must lease both the wire loop and the switch under this scenario. For each scenario, the average cost per unit for the unbundled network element was calculated by dividing the total forward-looking cost for the wire loops in the scenario by the total number of wire loops in the scenario. BellSouth thus treats its wire loops as three separate unbundled network elements: copper, universal digital loop carrier, and integrated digital loop carrier. BellSouth charges the competitive local carrier based on which technology it requests. The second aspect of the pricing model approved by the Florida Commission that is relevant to this appeal is the use of an inflation factor. When calculating the TELRIC for each unbundled network element, the BellSouth model also adds an inflation factor, which BellSouth contends reflects “inflation that will affect the cost of equipment that BellSouth will purchase over a period of several years.” BellSouth maintains that the inflation factor is independent of the “cost of capital” factor that the BellSouth model also takes into account in its pricing model. The third relevant aspect of the pricing method approved by the Florida Commission is its methodology for geographic cost-based deaveraging. The wire centers were divided into three zones in the following manner. First, the wire centers were divided into five groups such that “the average rate in each zone is no more than 20% higher or 20% [lower] than the forward-looking cost of providing that element.” These five groups were then reduced to three by combining the two high *1169 est-priced zones and combining the two lowest-priced zones. D. The Background of This Appeal In December 1998, several competitive local exchange carriers petitioned the Florida Commission to hold a hearing to establish rates for the unbundled network elements owned by BellSouth in Florida. In October 2000, the Florida Commission held the hearing, and BellSouth introduced the BellSouth Telecommunications Loop Model. The competitive local carriers objected to the BellSouth model on several grounds relevant to this appeal. Their first objection was to the use of three scenarios instead of one to compute the TELRIC of each unbundled element. According to the Florida Commission, MCI advocated the use of a single scenario, the Combo scenario. The second objection challenged the use of an “inflation factor.” MCI argued that the use of the inflation factor amounted to “double counting” because the Bell-South model already accounted for inflation under its “cost of capital” factor. The third objection challenged the method used for geographic cost-based deaveraging. The BellSouth model proposed three zones based solely on geography. MCI proposed the “Sprint approach,” which constructed six rate zones in which the rate of each element in the zone was within 20% of the average rate for that zone. The Florida Commission concluded its proceedings on September 27, 2002. The Florida Commission determined that the three-scenario approach of the BellSouth model was consistent with federal law. The Florida Commission also approved the inflation factor used in the BellSouth model. The Florida Commission rejected both the BellSouth and Sprint approaches to geographic cost-based deaveraging and adopted its own methodology. The Florida Commission used a modification of the Sprint approach to generate five groups with 20% variance, but the Florida Commission then consolidated the two most expensive zones and the two least expensive zones to create three deaveraged geographic cost groups. The order stated that it reduced the number of groups to alleviate “administrative burden.” MCI filed suit against BellSouth and the Florida Commission in federal district court seeking declaratory and injunctive relief under section 252. See id. § 252(e)(6). Florida Digital Network was permitted to intervene. MCI and Florida Digital Network raised three arguments relevant to this appeal. First, MCI and Florida Digital Network argued that the BellSouth model violated federal law because the use of multiple scenarios to model the wire loops of the network failed to comply with TELRIC. Second, MCI and Florida Digital Network argued that the use of the “inflation factor” in the pricing plan violated federal law because it double-counted certain expenses. Third, Florida Digital Network argued that the method adopted by the Florida Commission to allocate geographic cost-based deaveraging zones was not supported by the record. The district court concluded that the BellSouth model conflicted with federal law. The district court found that the multiple-scenario approach of the Bell-South model was contrary to FCC regulations for two reasons. First, the BellSouth model failed to take “as a given the incumbent LEC’s provision of other elements,” 47 C.F.R. § 51.505(b), and instead “focused on a particular loop type or combination to the exclusion of others.” Second, the model was based on scenarios “where the particular [unbundled network element] occupies the entire network as opposed to what is likely to be requested and used.” See id. § 51.511(a). *1170 The district court affirmed other portions of the order of the Florida Commission. First, the district court upheld the use of the inflation factor by the BellSouth method. The district court concluded that the inflation factor neither resulted in double counting nor was contrary to federal regulations because it “reflects the growth costs of the hypothetical network during the rate period, which is typically three to four years.” Second, the district court upheld the geographic cost-based deaver-aging method adopted by the Florida Commission. The district court noted that federal law does not require the Florida Commission to use more than three zones and found that there was “sufficient record evidence to support the [Florida] Commission’s approach.” II. STANDARD OF REVIEW This Court reviews de novo questions of law. AT&T Commc’ns of the S. States, Inc. v. BellSouth Telecomms., Inc., 268 F.3d 1294, 1296 (11th Cir.2001). Federal courts generally “accord no deference to the state commission’s interpretations” of federal law. AT&T Commc’ns of Va., Inc. v. Bell Atl.-Va., Inc., 197 F.3d 663, 668 (4th Cir.1999); accord AT&T Commc’ns of Cal., Inc. v. Pac. Bell Tel. Co., 375 F.3d 894, 904 (9th Cir.2004) (“We also consider de novo whether the agreements comply with the Act and its implementing regulations.” (internal quotations omitted)); MCI Telecomms. Corp. v. Bell Atl.-Pa., 271 F.3d 491, 516 (3d Cir.2001). The factual findings of the state agency will not be disturbed unless they are arbitrary and capricious or not supported by substantial evidence. See Sw. Bell Tel. Co. v. Waller Creek Commc’ns, Inc., 221 F.3d 812, 816 (5th Cir.2000); accord GTE S., Inc. v. Morrison, 199 F.3d 733, 745 (4th Cir.1999). III. DISCUSSION This appeal presents three issues. First, BellSouth and the Florida Commission argue that the district court erroneously concluded that the use of multiple scenarios in the BellSouth model violates TELRIC. Second, MCI argues that the district court erroneously approved of the use of the inflation factor in the BellSouth model. Third, Florida Digital Network argues that the district court erroneously approved the geographic cost-based deav-eraging methodology adopted by the Florida Commission. We address each argument in turn. A. The District Court Erroneously Concluded That TELRIC Forbids Multiple Scenarios. BellSouth and the Florida Commission argue that the district court erroneously concluded that the use of multiple scenarios in the BellSouth model conflicts with TELRIC. BellSouth and the Florida Commission argue that the district court failed to credit FCC precedents, in section 271 proceedings, that found the use of multiple scenarios compliant with TEL-RIC. They alternatively argue that, even without deference to the FCC, the district court erroneously concluded that the Bell-South method violates TELRIC. Our review of the multiple scenarios is divided in two parts. We begin by addressing the level of deference we afford to statements made by the FCC regarding TELRIC in section 271 proceedings. Because we conclude section 271 proceedings provide little or no guidance as to the requirements of TELRIC, we then conduct our own review of whether TELRIC prohibits the use of multiple scenarios. We explain why the use of multiple scenarios is not forbidden by TELRIC so long as each scenario complies with TELRIC. *1171 1. Section 271 Proceedings Are Not Precedent for Section 252 Proceedings. Section 271 permits incumbent local carriers to enter the long-distance market upon approval of the FCC. 47 U.S.C. § 271(d)(1). The FCC may approve the application only if fourteen requirements are met. See id. § 271(c)(2)(i)-(xiv). One requirement is that the incumbent local carrier must offer “[n]ondiscriminatory access to network elements in accordance with the requirements of sections 251(c)(3) and 252(d)(1).” Id. § 271(c)(2)(h). The FCC must evaluate each of these requirements and render its decision within 90 days of the filing of the petition. Id. § 271(d)(3). BellSouth and the Florida Commission argue that this Court must defer to decisions of the FCC that found the use of multiple scenarios was consistent with TELRIC for purposes of the section 271(c)(2)(h) requirement. See In-Region, interLATA Servs. in Ga. & La. Order, 17 F.C.C.R. 9018, 9041-42 ¶¶ 38-42 (2002); see also In-Region, interLATA Servs. in Fla. & Tenn. Order, 17 F.C.C.R. 25,828, 25,840 ¶ 23 (2002); In-Region, interLATA Servs. in Ala., Ky., Miss., N.C. & S.C. Order, 17 F.C.C.R. 17,595, 17,621-25 ¶¶ 56-63 (2002) (hereinafter the “Five-State Order"). Further, BellSouth and the Florida Commission argue that the Florida and Tennessee Order endorsed the very pricing scheme challenged in this appeal. These arguments fail for at least three reasons. First, the section 271 decisions cited by BellSouth and the Florida Commission are far from a clear endorsement of the use of multiple scenarios by the FCC. In the Georgia and Louisiana Order, the FCC considered the application of BellSouth to enter the long-distance market in Georgia and Louisiana. See generally 17 F.C.C.R. 9018. The FCC discussed the multiple scenario approach of the BellSouth model and concluded that the parties opposing the application “ha[d] not presented evidence sufficient to show that the Louisiana Commission erred in its decision or to overcome the current evidence BellSouth has presented as to why the use of multiple scenarios is appropriate.” Id. at 9042 ¶42. The FCC elaborated, “[W]e have never held that an appropriate application of TELRIC precludes such an approach. Accordingly, we cannot conclude that the Louisiana Commission committed any clear error in adopting it.” Id. This statement of the FCC is not an approval of the multiple scenarios approach; the FCC stated instead that it had never disapproved of the use of multiple scenarios in a section 252 proceeding. See id. Closely related is the second reason the arguments of BellSouth and the Florida Commission fail: the standard of review applied by the FCC in a section 271 proceeding is highly deferential to the state communications commission. See, e.g., Five-State Order, 17 F.C.C.R. at 17,624 ¶ 61 (“We defer to the analyses of the state commissions, and we therefore reject WorldCom’s criticism of the multiple scenario approach.”). BellSouth has stressed the deferential standard of review applicable to section 271 proceedings in its filings before the FCC: “The Commission should place great weight on the state commissions’ determinations that BellSouth’s rates are TELRIC-compliant. As the Commission has explained, it does not engage in de novo review of rates in section 271 proceedings.” Brief for BellSouth at 29, Fla. & Tenn, Order, 17 F.C.C.R. 25,828 (No. 02-331) (internal quotations and citations omitted). This Court, in contrast, exercises de novo review of the decision of the state commission. See AT&T Commc’ns of the S. States, 268 F.3d at *1172 1296; accord Bell Atl-Va., Inc., 197 F.3d at 668 (holding that “we accord no deference to the state commission’s interpretations” of federal law); Bell Atl.-Pa. Serv., 271 F.3d at 516 (“[A] state utility commission’s interpretations of the Act are reviewed de novo ... because the state commissions are not federal agencies to which deference is due.”). If this Court were to defer to the FCC, which had, in turn, deferred to the state commission, it would render our de novo review of the state commission meaningless. Third, the FCC has itself disavowed the precedential value of its opinions in section 271 proceedings. See In-Region, Inter-LATA Servs. in Kan. & Okla. Order, 16 F.C.C.R. 6237, 6246-47 ¶ 19 (2001). In the Kansas and Oklahoma Order, the FCC explained that section 271 proceedings are streamlined and should not be delayed by questions best resolved in other fora: As the Commission stated in the SWBT Texas Order, despite the comprehensiveness of our local competition rules, there will inevitably be, in any section 271 proceeding, new and unresolved interpretive disputes about the precise content of an incumbent LEC’s obligations to its competitors — disputes that our rules have not yet addressed and that do not involve per se violations of self-executing requirements of the Act. The section 271 process simply could not function as Congress intended if we were generally required to resolve all such disputes as a precondition to granting a section 271 application. Congress designed section 271 proceedings as highly specialized, 90-day proceedings for examining the performance of a particular carrier in a particular State at a particular time. Such fast-track, narrowly focused adjudications are often inappropriate forums for the considered resolution of industry-wide local competition questions of general applicability. Second, such a requirement would undermine the congressional intent of section 271 to give the BOCs an incentive to open their local markets to competition. That incentive would largely vanish if a BOC’s opponents could effectively doom any section 271 application by raising a host of novel interpretive disputes in their comments and demanding that authorization be denied unless each one of those disputes is resolved in the BOC’s favor. Finally, simply as a matter of statutory construction, few of the substantive obligations contained in the local competition provisions of sections 251 and 252 are altogether self-executing; they rely for their content on the Commission’s rules. Id. (footnotes omitted) (citing In-Region, InterLATA Servs, in Tex. Order, 15 F.C.C.R. 18,354, 18,367 ¶27 (2000)). That the FCC occasionally reaches the same conclusion in a section 252 proceeding as it had in a section 271 proceeding does not undermine this clear language. See, e.g., Metro Teleconnect Cos., Inc. v. Verizon Md. Inc., 18 F.C.C.R. 9033, 9035 ¶2 (2003). We join our sister circuits and conclude that section 271 proceedings provide, at most, persuasive guidance when evaluating an appeal under section 252. See, e.g., MCImetro Access Transmission Servs., Inc. v. BellSouth Telecomms., Inc., 352 F.3d 872, 880 n. 6 (4th Cir.2003); AT&T Corp. v. FCC, 220 F.3d 607, 630-31 (D.C.Cir.2000). What matters instead is the text of the federal regulations regarding TELRIC. We next conduct a de novo interpretation of whether the TELRIC regulations allow the use of multiple scenarios. 2. TELRIC Permits the Use of Multiple Scenarios, So Long As Each Scenario Complies with TELRIC. BellSouth and the Florida Commission argue that, even without deference to the *1173 section 271 proceedings, this Court should find that the pricing model adopted by the Florida Commission complies with TELRIC, but MCI and Florida Digital Network argue, as the district court concluded, that the use of multiple scenarios violates TELRIC. MCI and Florida Digital Network argue, alternatively, that, even if multiple scenarios are permissible under TELRIC, the scenarios adopted by the Florida Commission violate TELRIC. To resolve this controversy, we first consider whether TELRIC permits the use of multiple scenarios, an issue of first impression for the federal courts. Because we conclude that it does, we then review the standard each scenario must meet to comply with TELRIC and remand the evaluation of each scenario, under that standard, to the district court. a. The Use of Multiple Scenarios Is Consistent ivith TELRIC. The TELRIC value of an unbundled network element is “the forward-looking cost over the long run of the total quantity of the facilities and functions that are directly attributable to, or reasonably identifiable as incremental to, such element, calculated taking as a given the incumbent LEC’s provision of other elements.” 47 C.F.R. § 51.505(b). This cost 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.” Id. § 51.505(b)(1). The “forward-looking economic cost per unit” is then determined by dividing the TELRIC for the network element by “the sum of the total number of units of the element that the incumbent LEC is likely to provide to requesting telecommunications carriers and the total number of units of the element that the incumbent LEC is likely to use in offering its own services.” Id. § 51.511(a). The TELRIC methodology requires that the per unit cost of an unbundled network element be calculated by finding the total cost for the element in a hypothetical most efficient network and dividing by the number of units that will be put into use by the incumbent or a competitive local carrier. BellSouth and the Florida Commission argue that no single scenario for wire loops can be “most efficient” because different services require different types of wire loops. To support this position, BellSouth and the Florida Commission argue that a competitive local carrier that offers DSL will require all wire loops to be copper; likewise, a competitive local carrier that intends to use its own switches must lease loops that use the universal digital loop carrier technology. In short, because different competitive local carriers will request different elements, there is no unitary “most efficient” network. We agree with BellSouth and the Florida Commission that TELRIC does not prohibit the use of multiple scenarios in a pricing model. The use of multiple scenarios classifies different types of wire loops as different network elements. BellSouth and the Florida Commission, for example, maintain, without dispute, that DSL service can be offered only via copper wire loops, so the Copper Only scenario represents a network element consisting of DSL-capable wire loops. The Combo scenario likewise represents the network element of wire loops and switches that employ the integrated digital loop carrier technology. The definition of “network element” in the Telecommunications Act supports an interpretation that depends on separate features, functions, and capabilities: The term “network element” means a facility or equipment used in the provi *1174 sion of a telecommunications service. Such term also includes features, functions, and capabilities that are provided by means of such facility or equipment, including subscriber numbers, databases, signaling systems, and information sufficient for billing and collection or used in the transmission, routing, or other provision of a telecommunications service. Id. § 153(29). Nothing in this definition requires that the network element be defined so broadly, as MCI and Florida Digital Networks argue, as to encompass all wire loops; rather, the focus of the definition on separate “features, functions, and capabilities” evidences the intent of the FCC to encourage narrowly-defined network elements. Id. This interpretation also comports with the parties’ understanding of the technology of telecommunications. The parties agree, for example, that DSL is a different “telecommunications service” than local telephone service, and DSL requires different equipment (ie., copper wire loops). Likewise, the parties agree that the equipment that provides integrated digital loop carrier service is different from universal digital loop carrier equipment. We conclude that TELRIC permits an incumbent local carrier to define its unbundled network elements narrowly to separate wire loops with different capabilities and characteristics into different network elements through the use of multiple scenarios. This conclusion is consistent with the result reached by the FCC in section 271 proceedings, see, e.g., Ga. & La. Order, 17 F.C.C.R. at 9041-42 ¶¶ 38-42, and by many state commissions, see, e.g., id. at 9041 ¶ 40; Five-State Order, 17 F.C.C.R. at 17,621 ¶ 56. The district court erred in concluding otherwise. b. Each Scenario Must Comply with TELRIC. Although TELRIC allows the use of multiple scenarios, that conclusion does not end our inquiry. MCI and Florida Digital Network argue that, even if multiple scenarios are permissible, the scenarios in the model approved by the Florida Commission nevertheless violate TELRIC. Although the district court did not address this issue, we may consider it to determine whether the judgment of the district court may be affirmed on this alternative ground. See Cochran v. U.S. Health Care Fin. Admin., 291 F.3d 775, 778 (11th Cir.2002) (“[W]e may affirm for any reason supported by the record.”). Because TELRIC permits a pricing model to use multiple scenarios when each scenario represents its own network element, it follows that each scenario must itself comply with the requirements of TELRIC. See 47 C.F.R. § 51.503(b) (“An incumbent LEC’s rates for each element it offers ... shall be established ... — (1) Pursuant to the forward-looking economic cost-based pricing methodology set forth in §§ 51.505 and 51.511[.]” (emphases added)). In other words, the incumbent local carrier must compute for each scenario “the forward-looking cost over the long run of the total quantity of the facilities and functions that are directly attributable to, or reasonably identifiable as incremental to, such element, calculated taking as a given the incumbent LEC’s provision of other elements.” Id. § 51.505(b). The design of each scenario must be “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.” Id. § 51.505(b)(1). Although “enormous flexibility is built into TELRIC,” AT&T Corp., 220 F.3d at 616, it is not possible to reconcile the *1175 “most efficient ... lowest cost network configuration” requirement of section 51.505 with a network that contains loops with costs that vastly outstrip their utility. 47 C.F.R. § 51.505(b)(1). A scenario, for example, that is included in a pricing model specifically for DSL providers, cannot be the “most efficient ... lowest cost network configuration,” Id., if it contains wire loops that are so long that DSL cannot be provided over those loops. The touchstone — as with any efficiency-based model — is whether a rational local carrier would likely use each loop modeled by the scenario. Other language in the FCC regulations supports this interpretation. See id. § 51.511(a). To determine the per-unit cost of a network element, the TELRIC for the element is divided by “the sum of the total number of units of the element that the incumbent LEC is likely to provide to requesting telecommunications carriers and the total number of units of the element that the incumbent LEC is likely to use in offering its own services.” Id. (emphasis added). A scenario that contains loops that an incumbent local carrier is neither “likely to provide” to competitive local carriers nor “likely to use” in offering its own service artificially inflates this average cost by including units for which there is no demand. That result would ran counter to the pro-competitive purpose of the Telecommunications Act. See AT&T Corp., 525 U.S. at 366, 119 S.Ct. at 724. In sum, so long as a rational local carrier would choose to substitute another technology or forgo the use of the wire loop altogether instead of paying the inflated cost for the wire loop, the scenario containing that wire loop must be considered inefficient and in violation of TEL-RIC. After thorough review, we cannot say whether MCI and Florida Digital Network can satisfy their burden of proving that the pricing plan adopted by the Florida Commission violates TELRIC. We cannot determine from the record on appeal whether each scenario satisfies the standard of efficiency defined by TELRIC, and the arguments of the parties are not sufficiently developed for us to conduct a meaningful analysis of that issue. The resolution of this issue is best left, in the first instance, to the district court where the parties may more fully develop their arguments. We remand this issue to the district court. B. The Inflation Factor Is Consistent with Federal Law. MCI and Florida Digital Network argue that the use of the “inflation factor” in the BellSouth model violates TELRIC. MCI and Florida Digital Network contend that including the inflation factor results in double counting because the BellSouth model already includes a factor to account for its cost of capital. MCI and Florida Digital Network argue that, by allowing double counting, the Florida Commission granted BellSouth more than a “normal” economic profit in violation of TELRIC. MCI and Florida Digital Network also contend that neither TELRIC nor the Telecommunications Act allows the additional inflation factor. BellSouth and the Florida Commission dispute that the inflation factor constitutes double counting. BellSouth and the Florida Commission contend that there are two types of inflation: “(1) inflation reflected in the increased cost of money over the period of years in which the rates will be in effect, and (2) inflation that will affect the cost of equipment that BellSouth will purchase over a period of several years.” BellSouth and the Florida Commission argue that TELRIC authorizes both forms of inflation because they both contribute to the cost of maintaining the hypothetical most efficient network. We agree with BellSouth and the Florida Commission. *1176 BellSouth and the Florida Commission correctly argue that the two types of inflation are independent of one another. The first type of inflation is “general inflation for which investors demand compensation through the cost of capital.” The second type of inflation reflects “specific inflation related to an investment or asset” (ie., inflation related to the acquisition of materials and services over the duration of the agreement). The two inflation rates, one general and one specific, need not be the same; as MCI concedes, the latter form of inflation may even be negative (ie., deflation) if the costs associated with the hypothetical network decrease due to improved technology. The argument of MCI and Florida Digital Network that TELRIC does not authorize recovery for inflation likewise fails. The TELRIC of an element is based on the forward-looking cost of the hypothetical most efficient, lowest cost network. 47 C.F.R. § 51.505(b). Inflation (or deflation) of the cost of materials and services for the hypothetical network is not listed as one of the “factors that may not be considered” in calculating the cost. Id. § 51.505(d). Because interconnection agreements span several years, it is necessary to account for changes in industry-specific costs over that period. Nothing in the Telecommunications Act or TELRIC bars the Florida Commission from including an inflation factor for costs associated with the hypothetical network in the pricing model, and the decision of the Florida Commission was not arbitrary and capricious. C. The Geographic Cosh-Based Deaver-aging Method Adopted by the Florida Commission Complies with Federal Law. Florida Digital Network argues that the district court erroneously approved the geographic cost-based deaver-aging method employed by the Florida Commission. Florida Digital Network contends that the method adopted by the Florida Commission lacks a basis in the record and fails to promote competition. We disagree. The regulations provide, “State commissions shall establish different rates for elements in at least three defined geographic areas within the state to reflect geographic cost differences.” Id. § 507(f). For states, such as Florida, that do not have “existing density-related zone pricing plans,” the “state commissions must create a minimum of three cost-related rate zones.” Id. This requirement recognizes that “deaveraged rates more closely reflect the actual costs of providing interconnection and unbundled elements.” Implementation of the Local Competition Provisions in the Telecommunications Act of 1996, 11 F.C.C.R. 15,499, 15,882 ¶764 (1996) (hereinafter “Local Competition Order”). It costs less per unit to provide local carrier service for an urban area than for a rural area, for example, so costs for these areas should not be averaged together but instead treated separately. See id. at 15,879-80 ¶ 760. Before the Florida Commission, Bell-South advocated for a three-zone method that was based on geography, but left wide variance in prices within each zone. MCI and Florida Digital Network advocated for the “Sprint approach,” which “group[ed] wire centers into zones based on the average cost of all UNE loops in the wire center, regardless of geographic or political subdivisions.” Each element in the rate zone would be no more than 20% above and no less than 20% below the average rate for all elements in the zone. The Sprint approach, as executed by MCI *1177 and Florida Digital Network, resulted in six pricing zones. The Florida Commission ultimately adopted neither approach. The Florida Commission rejected the BellSouth model because it was not cost-based, and it rejected the Sprint approach because it resulted in too many rate zones and was therefore “administratively burdensome.” Instead, the Florida Commission ran a “revised iteration of the Sprint approach” that produced five rate zones and then consolidated the two least expensive zones into a new zone and the two most expensive zones into another. This resulted in a three-zone model. The argument of Florida Digital Network that the decision of the Florida Commission to adopt its own methodology is not supported by the record fails because the initial step in the methodology adopted by the Florida Commission was presented as the Sprint approach. Florida Digital Network concedes that the manner in which the number of zones was reduced from five to three was a “variation” on the Sprint approach. The further requirement that “the cost data available in the proceeding implies that three zones is the most reasonable choice for BellSouth” is supported by the record. We conclude that the record supports the decision of the Florida Commission. Although Florida Digital Network argues that the methodology fails to promote competition, section 51.507 does little to cabin the discretion of a state commission when devising a geographic cost-based deaveraging method. See 47 C.F.R. § 51.507(f). The only requirements imposed by the section are (1) the method must be based on “geographic areas within the state,” (2) the method must “reflect geographic cost differences,” and (3) there must be “a minimum of three cost-related rate zones.” Id. The methodology adopted by the Florida Commission complies with each of these requirements. The methodology chosen by the Florida Commission is based on the Sprint approach, and Florida Digital Network concedes this satisfies the first and second requirements. The third requirement is satisfied because there are three cost-related zones. The argument advanced by Florida Digital Network ultimately turns on its dissatisfaction with the zones chosen by the Florida Commission, not with the failure of the zones to comport with federal law. Neither Congress nor the FCC required a specific degree of deaveraging (e.g., a maximum variance within each zone) even though either could have imposed such a restriction. Instead, the FCC chose to impose only a requirement that there be at least three zones. See Local Competition Order, 11 F.C.C.R. at 15,832 ¶765 (‘We conclude that three zones are presumptively sufficient to reflect geographic cost differences in setting rates for interconnection and unbundled elements.”). The Florida Commission complied with this requirement, and it was within its discretion to choose among the many methods that would do so. Because the decision of the Florida Commission to adopt this method of geographic cost-based deaveraging was not arbitrary and capricious, we affirm the decision of the district court upholding the deaveraging model. IV. CONCLUSION We reverse the decision of the district court that held TELRIC prohibits the use of multiple scenarios to model wire loops and remand to allow the district court to determine whether the scenarios approved by the Florida Commission are consistent with the interpretation of TELRIC provided in this opinion. We affirm the decision of the district court to uphold the use of *1178 the inflation factor and the geographic cost-based deaveraging method adopted by the Florida Commission. REVERSED and REMANDED in part, AFFIRMED in part.