Full opinion text
Justice Souter delivered the opinion of the Court. These cases arise under the Telecommunications Act of 1996. Each is about the power of the Federal Communications Commission to regulate a relationship between monopolistic companies providing local telephone service and companies entering local markets to compete with the incumbents. Under the Act, the new entrants are entitled, among other things, to lease elements of the local telephone networks from the incumbent monopolists. The issues are whether the FCC is authorized (1) to require state utility commissions to set the rates charged by the incumbents for leased elements on a forward-looking basis untied to the incumbents’ investment, and (2) to require incumbents to combine such elements at the entrants’ request when they lease them to the entrants. We uphold the FCC’s assumption and exercise of authority on both issues. I The 1982 consent decree settling the Government’s antitrust suit against the American Telephone and Telegraph Company (AT&T) divested AT&T of its local-exchange carriers, leaving AT&T as a long-distance and equipment company, and limiting the divested carriers to the provision of local telephone service. United States v. American Telephone & Telegraph Co., 552 F. Supp. 131 (DC 1982), aff’d sub nom. Maryland v. United States, 460 U. S. 1001 (1983). The decree did nothing, however, to increase competition in the persistently monopolistic local markets, which were thought to be the root of natural monopoly in the telecommunications industry. See S. Benjamin, D. Lichtman, & H. Shelanski, Telecommunications Law and Policy 682 (2001) (hereinafter Benjamin et al.); P. Huber, M. Kellogg, & J. Thorne, Federal Telecommunications Law §2.1.1, pp. 84-85 (2d ed. 1999) (hereinafter Huber et al.); W. Baumol & J. Sidak, Toward Competition in Local Telephony 7-10 (1994); S. Breyer, Regulation and Its Reform 291-292, 314 (1982). These markets were addressed by provisions of the Telecommunications Act of 1996 (1996 Act or Act), Pub L. 104-104, 110 Stat. 56, that were intended to eliminate the monopolies enjoyed by the inheritors of AT&T’s local franchises; this objective was considered both an end in itself and an important step toward the Act’s other goals of boosting competition in broader markets and revising the mandate to provide universal telephone service. See Benjamin et al. 716. Two sets of related provisions for opening local markets •concern us here. First, Congress required incumbent local-exchange carriers to share their own facilities and services on terms to be agreed upon with new entrants in their markets. 47 U. S. C. § 251(c) (1994 ed., Supp. V). Second, knowing that incumbents and prospective entrants would sometimes disagree on prices for facilities or services, Congress directed the FCC to prescribe methods for state commissions to use in setting rates that would subject both incumbents and entrants to the risks and incentives that a competitive market would produce. § 252(d). The particular method devised by the FCC for setting rates to be charged for interconnection and lease of network elements under the Act, § 252(d)(1), and regulations the FCC imposed to implement the statutory duty to share these elements, § 251(c)(3), are the subjects of this litigation, which must be understood against the background of ratemaking for public utilities in the United States and the structure of local exchanges made accessible by the Act. A Companies providing telephone service have traditionally been regulated as monopolistic public utilities. See J. Bon-bright, Principles of Public Utility Rates 3-5 (1st ed. 1961) (hereinafter Bonbright); I. Barnes, Economics of Public Utility Regulation 37-41 (1942) (hereinafter Barnes). At the dawn of modern utility regulation, in order to offset monopoly power and ensure affordable, stable public access to a utility’s goods or services, legislatures enacted rate schedules to fix the prices a utility could charge. See id., at 170-173; C. Phillips, Regulation of Public Utilities 111-112, and n. 5 (1984) (hereinafter Phillips). See, e. g., Smyth v. Ames, 169 U. S. 466, 470-476 (1898) (statement of case); Munn v. Illinois, 94 U. S. 113, 134 (1877). As this job became more complicated, legislatures established specialized administrative agencies, first local or state, then federal, to set and regulate rates. Barnes 173-175; Phillips 115-117. See, e. g., Minnesota Rate Cases, 230 U. S. 352, 433 (1913) (Interstate Commerce Commission); Shreveport Rate Cases, 234 U. S. 342, 354-355 (1914) (jurisdictional dispute between ICC and Texas Railroad Commission). See generally T. Mc-Craw, Prophets of Regulation 11-65 (1984). The familiar mandate in the enabling Acts was to see that rates be “just and reasonable” and not discriminatory. Barnes 289. See, e. g., Transportation Act of 1920, 49 U. S. C. § 1(5) (1934 ed.). All rates were subject to regulation this way: retail rates charged directly to the public and wholesale rates charged among businesses involved in providing the goods or services offered by the retail utility. Intrastate retail rates were regulated by the States or municipalities, with those at wholesale generally the responsibility of the National Government, since the transmission or transportation involved was characteristically interstate. See Phillips 143. Historically, the classic scheme of administrative rate-setting at the federal level called for rates to be set out by the regulated utility companies in proposed tariff schedules, on the model applied to railroad carriers under the Interstate Commerce Act of 1887,24 Stat. 379. After interested parties had had notice of the proposals and a chance to comment, the tariffs would be accepted by the controlling agency so long as they were “reasonable” (or “just and reasonable”) and not “unduly discriminatory.” Hale, Commissions, Rates, and Policies, 53 Harv. L. Rev. 1103, 1104-1105 (1940). See, e. g., Southern Pacific Co. v. ICC, 219 U. S. 433, 445 (1911). The States generally followed this same tariff-schedule model. Barnes 297-298. See, e. g., Smyth, supra, at 470-476. The way rates were regulated as between businesses (by the National Government) was in some respects, however, different from regulation of rates as between businesses and the public (at the state or local level). In wholesale markets, the party charging the rate and the party charged were often sophisticated businesses enjoying presumptively equal bargaining power, who could be expected to negotiate a “just and reasonable” rate as between the two of them. Accordingly, in the Federal Power Act of 1920,41 Stat. 1063, and again in the Natural Gas Act of 1938, 52 Stat. 821, Congress departed from the scheme of purely tariff-based regulation and acknowledged that contracts between commercial buyers and sellers could be used in ratesetting, 16 U. S. C. §824d(d) (Federal Power Act); 15 U. S. C. §717c(c) (Natural Gas Act). See United Gas Pipe Line Co. v. Mobile Gas Service Corp., 350 U. S. 332, 338-339 (1956). When commercial parties did avail themselves of rate agreements, the principal regulatory responsibility was not to relieve a contracting party of an unreasonable rate, FPC v. Sierra Pacific Power Co., 350 U. S. 348, 355 (1956) (“its improvident bargain”), but to protect against potential discrimination by favorable contract rates between allied businesses to the detriment of other wholesale customers. See ibid. Cf. New York v. United States, 331 U. S. 284, 296 (1947) (“The principal evil at which the Interstate Commerce Act was aimed was discrimination in its various manifestations”). This Court once summed up matters at the wholesale level this way: “[WJhile it may be that the Commission may not normally impose upon a public utility a rate which would produce less than a fair return, it does not follow that the public utility may not itself agree by contract to a rate affording less than a fair return or that, if it does so, it is entitled to be relieved of its improvident bargain. In such circumstances the sole concern of the Commission would seem to be whether the rate is so low as to adversely affect the public interest — as where it might impair the financial ability of the public utility to continue its service, cast upon other consumers an excessive burden, or be unduly discriminatory.” Sierra Pacific Power Co., supra, at 355 (citation omitted). See also United Gas Pipe Line Co., supra, at 345. Regulation of retail rates at the state and local levels was, on the other hand, focused more on the demand for “just and reasonable” rates to the public than on the perils of rate discrimination. See Barnes 298-299. Indeed, regulated local telephone markets evolved into arenas of state-sanctioned discrimination engineered by the public utility commissions themselves in the cause of “universal service.” Huber et al. 80-85. See also Vietor 167-185. In order to hold down charges for telephone service in rural markets with higher marginal costs due to lower population densities and lesser volumes of use, urban and business users were charged subsidizing premiums over the marginal costs of providing their own service. See Huber et al. 84. These cross subsidies between markets were not necessarily transfers between truly independent companies, however, thanks largely to the position attained by AT&T and its satellites. This was known as the “Bell system,” which by the mid-20th century had come to possess overwhelming monopoly power in all telephone markets nationwide, supplying local-exchange and long-distance services as well as equipment. Vietor 174-175. See also R. Garnet, Telephone Enterprise: Evolution of Bell System’s Horizontal Structure, 1876-1909, pp. 160-163 (1985) (Appendix A). The same pervasive market presence of Bell providers that made it simple to provide cross subsidies in aid of universal service, however, also frustrated conventional efforts to hold retail rates down. See Huber et al. 84-85. Before the Bell system’s predominance, regulators might have played competing carriers against one another to get lower rates for the public, see Cohen 47-50, but the strategy became virtually impossible once a single company had become the only provider in nearly every town and city across the country. This rejgulatory frustration led, in turn, to new thinking about just and reasonable retail rates and ultimately to these cases. The traditional regulatory notion of the “just and reasonable” rate was aimed at navigating the straits between gouging utility customers and confiscating utility property. FPC v. Hope Natural Gas Co., 320 U. S. 591, 603 (1944). See also Barnes 289-290; Bonbright 38. More than a century ago, reviewing courts charged with determining whether utility rates were sufficiently reasonable to avoid unconstitutional confiscation took as their touchstone the revenue that would be a “fair return” on certain utility property known as a “rate base.” The fair rate of return was usually set as the rate generated by similar investment property at the time of the rate proceeding, and in Smyth v. Ames, 169 U. S., at 546, the Court held that the rate base must be calculated as “the fair value of the property being used by [the utility] for the convenience of the public.” In pegging the rate base at “fair value,” the Smyth Court consciously rejected the primary alternative standard, of capital actually invested to provide the public service or good. Id., at 543-546. The Court made this choice in large part to prevent “excessive valuation or fictitious capitalization” from artificially inflating the rate base, id., at 544, lest “‘[t]he public ... be subjected to unreasonable rates in order simply that stockholders may earn dividends,’” id., at 545 (quoting Covington & Lexington Turnpike Road Co. v. Sandford, 164 U. S. 578, 596 (1896)). But Smyth proved to be a troublesome mandate, as Justice Brandéis, joined by Justice Holmes, famously observed 25 years later. Missouri ex rel. Southwestern Bell Telephone Co. v. Public Serv. Comm’n of Mo., 262 U. S. 276, 292 (1923) (dissenting opinion). The Smyth Court itself had described, without irony, the mind-numbing complexity of the required enquiry into fair value, as the alternative to historical investment: “[I]n order to ascertain [fair] value, original cost of construction, the amount expended in permanent improvements, the amount and market value of its bonds and stock, the present as compared with the original cost of construction, the probable earning capacity of the property under particular rates prescribed by statute, and the sum required to meet operating expenses, are all matters for consideration, and are to be given such weight as may be just and right in each case. We do not say that there may not be other matters to be regarded in estimating the value of the property.” 169 U. S., at 546-547. To the bewildered, Smyth simply threw up its hands, prescribing no one method for limiting use of these numbers but declaring all such facts to be “relevant.” Southwestern Bell Telephone Co., 262 U. S., at 294-298, and n. 6 (Brandeis, J., dissenting). What is more, the customary checks on calculations of value in other circumstances were hard to come by for a utility’s property; its costly facilities rarely changed hands and so were seldom tagged with a price a buyer would actually pay and a seller accept, id., at 292; West v. Chesapeake & Potomac Telephone Co. of Baltimore, 295 U. S. 662, 672 (1935). Neither could reviewing courts resort to a utility’s revenue as an index of fair value, since its revenues were necessarily determined by the rates subject to review, with the rate of return applied to the very property subject to valuation. Duquesne Light Co. v. Barasch, 488 U. S. 299, 309, n. 5 (1989); Hope Natural Gas Co., supra, at 601. Small wonder, then, that Justice Brandéis was able to demonstrate how basing rates on Smyth’s galactic notion of fair value could produce revenues grossly excessive or insufficient when gauged against the costs of capital. He gave the example (simplified) of a $1 million plant built with promised returns on the equity of $90,000 a year. Southwestern Bell Telephone Co., supra, at 304-306. If the value were to fall to $600,000 at the time of a rate proceeding, with the rate of return on similar investments then at 6 percent, Smyth would say a rate was not confiscatory if it returned at least $36,000, a shortfall of'$54,000 from the costs of capital. But if the value of the plant were to rise to $1,750,000 at the time of the rate proceeding, and the rate of return on comparable investments stood at 8 percent, then constitutionality under Smyth would require rates generating at least $140,000, $50,000 above capital costs. The upshot of Smyth, then, was the specter of utilities forced into bankruptcy by rates inadequate to pay off the costs of capital, even when a drop in value resulted from general economic decline, not imprudent investment; while in a robust economy, an investment no more prescient could claim what seemed a rapacious return on equity invested. Justice Brandéis accordingly advocated replacing “fair value” with a calculation of rate base on the cost of capital prudently invested in assets used for the provision of the public good or service, and although he did not live to enjoy success, his campaign against Smyth came to fruition in FPC v. Hope Natural Gas Co., 320 U. S. 591 (1944). In Hope Natural Gas, this Court disavowed the position that the Natural Gas Act and the Constitution required fair value as the sole measure of a rate base on which “just and reasonable” rates were to be calculated. Id., at 601-602. See also FPC v. Natural Gas Pipeline Co., 315 U. S. 575, 602-606 (1942) (Black, Douglas, and Murphy, JJ., concurring). In the matter under review, the Federal Power Commission had valued the rate base by using “actual legitimate cost” reflecting “sound depreciation and depletion practices,” and so had calculated a value roughly 25 percent below the figure generated by the natural-gas company’s fair-value methods using “estimated reproduction cost” and “trended original cost.” Hope Natural Gas, 320 U. S., at 596-598, and nn. 4-5. The Court upheld the Commission. “Rates which enable the company to operate successfully, to maintain its financial integrity, to attract capital, and to compensate its investors for the risks assumed certainly cannot be condemned as invalid, even though they might produce only a meager return on the so-called ‘fair value’ rate base.” Id., at 605. Although Hope Natural Gas did not repudiate everything said in Smyth, since fair value was still “the end product of the process of rate-making,” 320 U. S., at 601, federal and state commissions setting rates in the aftermath of Hope Natural Gas largely abandoned the old fair-value approach and turned to methods of calculating the rate base on the basis of “cost.” A. Kahn, Economics of Regulations: Principles and Institutions 40-41 (1988). “Cost” was neither self-evident nor immune to confusion, however; witness the invocation of “reproduction cost” as a popular method for calculating fair value under Smyth, see n. 5, supra, and the Federal Power Commission’s rejection of “trended original cost” (apparently, a straight-line derivation from the cost of capital originally invested) in favor of “actual legitimate cost,” Hope Natural Gas, supra, at 596. Still, over time, general agreement developed on a method that was primus inter pares, and it is essentially a modern gloss on that method that the incumbent carriers say the FCC should have used to set the rates at issue here. The method worked out is not a simple calculation of rate base as the original cost of “prudently invested” capital that Justice Brandéis assumed, presumably by reference to the utility’s balance sheet at the time of the rate proceeding. Southwestern Bell Telephone Co., 262 U. S., at 304-306. Rather, “cost” came to mean “cost of service,” that is, the cost of prudently invested capital used to provide the service. Bonbright 173; P. Garfield & W. Lovejoy, Public Utility Economics 56 (1964). This was calculated subject to deductions for accrued depreciation and allowances for working capital, see Phillips 282-283 (table 8-1) (“a typical electric utility rate base”), naturally leading utilities to minimize depreciation by using very slow depreciation rates (on the assumption of long useful lives), and to maximize working capital claimed as a distinct rate-base constituent. This formula, commonly called the prudent-investment rule, addressed the natural temptations on the utilities’ part to claim a return on outlays producing nothing of value to the public. It was meant, on the one hand, to discourage unnecessary investment and the “fictitious capitalization” feared in Smyth, 169 U. S., at 543-546, and so to protect ratepayers from supporting excessive capacity, or abandoned, destroyed, or phantom assets. Kahn, Tardiff, & Weisman, Telecommunications Act at three years: an economic evaluation of its implementation by the Federal Communications Commission, 11 Information Economics & Policy 319, 330, n. 27 (1999) (hereinafter Kahn, Telecommunications Act). At the same time, the prudent-investment rule was intended to give utilities an incentive to make smart investments deserving a “fair” return, and thus to mimic natural incentives in competitive markets (though without an eye to fostering the actual competition by which such markets are defined). In theory, then, the prudent-investment qualification gave the ratepayer an important protection by mitigating the tendency of a regulated market’s lack of competition to support monopolistic prices. But the mitigation was too little, the prudent-investment rule in practice often being no match for the capacity of utilities having all the relevant information to manipulate the rate base and renegotiate the rate of return every time a rate was set. The regulatory response in some markets was adoption of a rate-based method commonly called “price caps,” United States Telephone Assn. v. FCC, 188 F. 3d 521, 524 (CADC 1999), as, for example, by the FCC’s setting of maximum access charges paid to large local-exchange companies by interexchange carriers, In re Policy and Rules Concerning Rates for Dominant Carriers, 5 FCC Rcd. 6786, 6787, ¶ 1 (1990). The price-cap scheme starts with a rate generated by the conventional cost-of-service formula, which it takes as a benchmark to be decreased at an average of some 2-3 percent a year to reflect productivity growth, Kahn, Telecommunications Act 330-332, subject to an upward adjustment if necessary to reflect inflation or certain unavoidable “exogenous costs” on which the company is authorized to recover a return. 5 FCC Red., at 6787, ¶ 5. Although the price caps do not eliminate gamesmanship, since there are still battles to be fought over the productivity offset and allowable exogenous costs, United States Telephone Assn., supra, at 524, they do give companies an incentive “to improve productivity to the maximum extent possible,” by entitling those that outperform the productivity offset to keep resulting profits, 5 FCC Red., at 6787-6788, ¶¶7-9. Ultimately, the goal, as under the basic prudent-investment rule, is to encourage investment in more productive equipment. Before the passage of the 1996 Act, the price cap was, at the federal level, the final stage in a century of developing ratesetting methodology. What had changed throughout the era beginning with Smyth v. Ames was prevailing opinion on how to calculate the most useful rate base, with the disagreement between fair-value and cost advocates turning on whether invested capital was the key to the right balance between investors and ratepayers, and with the price-cap scheme simply being a rate-based offset to the utilities’ advantage of superior knowledge of the facts employed in cost-of-service fatemaking. What is remarkable about this evolution of just and reasonable ratesetting, however, is what did not change. The enduring feature of ratesetting from Smyth v. Ames to the institution of price caps was the idea that calculating a rate base and then allowing a fair rate of return on it was a sensible way to identify a range of rates that would be just and reasonable to investors and ratepayers. Equally enduring throughout the period was dissatisfaction with the successive rate-based variants. From the constancy of this dissatisfaction, one possible lesson was drawn by Congress in the 1996 Act, which was that regulation using the traditional rate-based methodologies gave monopolies too great an advantage and that the answer lay in moving away from the assumption common to all the rate-based methods, that the monopolistic structure within the discrete markets would endure. Under the local-competition provisions of the Act, Congress called for ratemaking different from any historical practice, to achieve the entirely new objective of uprooting the monopolies that traditional rate-based methods had perpetuated. H. R. Conf. Rep. No. 104-230, p. 113 (1996). A leading backer of the Act in the Senate put the new goal this way: “This is extraordinary in the sense of telling private industry that this is what they have to do in order to let the competitors come in and try to beat your economic brains out... . “It is kind of almost a jump-start.... I will do everything I have to let you into my business, because we used to be a bottleneck; we used to be a monopoly; we used to control everything. “Now, this legislation says you will not control much of anything. You will have to allow for nondiserimina-tory access on an unbundled basis to the network functions and services of the Bell operating companies network that is at least equal in type, quality, and price to the access [a] Bell operating company affords to itself.” 141 Cong. Rec. 15572 (1995) (remarks of Sen. Breaux (La.) on Pub. L. 104-104). For the first time, Congress passed a ratesetting statute with the aim not just to balance interests between sellers and buyers, but to reorganize markets by rendering regulated utilities’ monopolies vulnerable to interlopers, even if that meant swallowing the traditional federal reluctance to intrude into local telephone markets. The approach was deliberate, through a hybrid jurisdictional scheme with the FCC setting a basic, default methodology for use in setting rates when carriers fail to agree, but leaving it to state utility commissions to set the actual rates. While the Act is like its predecessors in tying the methodology to the objectives of “just and reasonable” and nondiscriminatory rates, 47 U. S. C. § 252(d)(1), it is radically unlike all previous statutes in providing that rates be set “without reference to a rate-of-return or other rate-based proceeding,” § 252(d)(l)(A)(i). The Act thus appears to be an explicit disavowal of the familiar public-utility model of rate regulation (whether in its fair-value or cost-of-service incarnations) presumably still being applied by many States for retail sales, see In re Implementation of Local Competition in Telecommunications Act of 1996,11 FCC Rcd. 15499, 15857, ¶704 (1996) (First Report and Order), in favor of novel ratesetting designed to give aspiring competitors every possible incentive to enter local retail telephone markets, short of confiscating the incumbents’ property. B The physical incarnation of such a market, a “local exchange,” is a network connecting terminals like telephones, faxes, and modems to other terminals within a geographical area like a city. From terminal network interface devices, feeder wires, collectively called the “local loop,” are run to local switches that aggregate traffic into common “trunks.” The local loop was traditionally, and is still largely, made of copper wire, though fiber-optic cable is also used, albeit to a far lesser extent than in long-haul markets. Just as the loop runs from terminals to local switches, the trunks run from the local switches to centralized, or tandem, switches, originally worked by hand but now by computer, which operate much like railway switches, directing traffic into other trunks. A signal is sent toward its destination terminal on these common ways so far as necessary, then routed back down another hierarchy of switches to the intended telephone or other equipment. A local exchange is thus a transportation network for communications signals, radiating like a root system from a “central office” (or several offices for larger areas) to individual telephones, faxes, and the like. It is easy to see why a company that owns a local exchange (what the Act calls an “incumbent local exchange carrier,” 47 U. S. C. § 251(h)) would have an almost insurmountable competitive advantage not only in routing calls within the exchange, but, through its control of this local market, in the markets for terminal equipment and long-distance calling as well. A newcomer could not compete with the incumbent carrier to provide local service without coming close to replicating the incumbent’s entire existing network, the most costly and difficult part of which would be laying down the “last mile” of feeder wire, the local loop, to the thousands (or millions) of terminal points in individual houses and businesses. The incumbent company could also control its local-loop plant so as to connect only with terminals it manufactured or selected, and could place conditions or fees (called “access charges”) on long-distance carriers seeking to connect with its network. In an unregulated world, another telecommunications carrier would be forced to comply with these conditions, or it could never reach the customers of a local exchange. II The 1996 Act both prohibits state and local regulation that impedes the provision of “telecommunications service,” § 253(a), and obligates incumbent carriers to allow competitors to enter their local markets, § 251(c). Section 251(c) addresses the practical difficulties of fostering local competition by recognizing three strategies that a potential competitor may pursue. First, a competitor entering the market (a “requesting” carrier, § 251(c)(2)) may decide to engage in pure facilities-based competition, that is, to build its own network to replace or supplement the network of the incumbent. If an entrant takes this course, the Act obligates the incumbent to “interconnect” the competitor’s facilities to its own network to whatever extent is necessary to allow the competitor’s facilities to operate. §§ 251(a) and (c)(2). At the other end of the spectrum, the statute permits an entrant to skip construction and instead simply to buy and resell “telecommunications service,” which the incumbent has a duty to sell at wholesale. §§ 251(b)(1) and (e)(4). Between these extremes, an entering competitor may choose to lease certain of an incumbent’s “network elements,” which the incumbent has a duty to provide “on an unbundled basis” at terms that are “just, reasonable, and nondiseriminatory.” § 251(c)(3). Since wholesale markets for companies engaged in resale, leasing, or interconnection of facilities cannot be created without addressing rates, Congress provided for rates to be set either by contracts between carriers or by state utility commission rate orders. §§ 252(a)-(b). Like other federal utility statutes that authorize contracts approved by a regulatory agency in setting rates between businesses, e. g., 16 U. S. C. § 824d(d) (Federal Power Act); 15 U. S. C. §717c(c) (Natural Gas Act), the Act permits incumbent and entering carriers to negotiate private rate agreements, 47 U. S. C. § 252(a); see also § 251(c)(1) (duty to negotiate in good faith). State utility commissions are required to accept any such agreement unless it discriminates against a carrier not a party to the contract, or is otherwise shown to be contrary .to the public interest. §§ 252(e)(1) and (e)(2)(A). Carriers, of course, might well not agree, in which case an entering carrier has a statutory option to request mediation by a state commission, § 252(a)(2). But the option comes with strings, for mediation subjects the parties to the duties specified in §251 and the pricing standards set forth in § 252(d), as interpreted by the FCC’s regulations, § 252(e)(2)(B). These regulations are at issue here. As to pricing, the Act provides that when incumbent and requesting carriers fail to agree, state commissions will set a “just and reasonable” and “nondiscriminatory” rate for interconnection or the lease of network elements based on “the cost of providing the ... network element,” which “may include a reasonable profit.” §252(d)(1). In setting these rates, the state commissions are, however, subject to that important limitation previously unknown to utility regulation: the rate must be “determined without reference to a rate-of-return or other rate-based proceeding.” Ibid. In AT&T Corp. v. Iowa Utilities Bd., 525 U. S. 366, 384-385 (1999), this Court upheld the FCC’s jurisdiction to impose a new methodology on the States when setting these rates. The attack today is on the legality and logic of the particular methodology the Commission chose. As the Act required, six months after its effective date the FCC implemented the local-competition provisions in its First Report and Order, whieh included as an appendix the new regulations at issue. Challenges to the order, mostly by incumbent local-exchange carriers and state commissions, were consolidated in the United States Court of Appeals for the Eighth Circuit. Iowa Utilities Bd. v. FCC, 120 F. 3d 753, 792 (1997), aff’d in part and rev’d in part, 525 U. S. 366, 397 (1999). See also California v. FCC, 124 F. 3d 934, 938 (1997), rev’d in part, 525 U. S. 366, 397 (1999) (challenges to In re Implementation of Local Competition Provisions in Telecommunications Act of 1996, 11 FCC Red. 19392 (1996) (Second Report and Order)). So far as it bears on where we are today, the initial decision by the Eighth Circuit held that the FCC had no authority to control the methodology of state commissions setting the rates incumbent local-exchange carriers could charge entrants for network elements, 47 CFR § 51.505(b)(1) (1997). Iowa Utilities Bd. v. FCC, supra, at 800. The Eighth Circuit also held that the FCC misconstrued the plain language of § 251(c)(3) in implementing a set of “combination” rules, 47 CFR §§51.315(b)-(f) (1997), the most important of which provided that “an incumbent LEC shall not separate requested network elements that the incumbent LEC currently combines,” § 51.315(b). 120 F. 3d, at 813. On the other hand, the Court of Appeals accepted the FCC’s view that the Act required no threshold ownership of facilities by a requesting carrier, First Report and Order ¶¶ 328-340, and upheld Rule 319, 47 CFR §51.319 (1997), which read “network elements” broadly, to require incumbent carriers to provide not only equipment but also services and functions, such as operations support systems (e. g., billing databases), § 51.319(f)(1), operator services and directory assistance, § 51.319(g), and vertical switching features like call-waiting and caller I. D., First Report and Order ¶ ¶ 263, 413. 120 F. 3d, at 808-810. This Court affirmed in part and in larger part reversed. AT&T Corp. v. Iowa Utilities Bd., 525 U. S., at 397. We reversed in upholding, the FCC’s jurisdiction to “design a pricing methodology” to bind state ratemaking commissions, id., at 385, as well as one of the FCC’s combination rules, Rule 315(b), barring incumbents from separating currently combined network elements when furnishing them to entrants that request them in a combined form, id., at 395. We also reversed in striking down Rule 319, holding that its provision for blanket access to network elements was inconsistent with the “necessary” and “impair” standards of 47 U.S.C. § 251(d)(2), 525 U. S., at 392. We affirmed the Eighth Circuit, however, in upholding the FCC’s broad definition of network elements to be provided, id., at 387, and the FCC’s understanding that the Act imposed no facilities-ownership requirement, id., at 392-393. The case then returned to the Eighth Circuit. Id., at 397. With the FCC’s general authority to establish a pricing methodology secure, the incumbent carriers’ primary challenge on remand went to the method that the Commission chose. There was also renewed controversy over the combination rules (Rules 315(c)-(f)) that the Eighth Circuit had struck down along with Rule 315(b), but upon which this Court expressed no opinion when it reversed the invalidation of that latter rule. 219 F. 3d 744, 748 (2000). As for the method to derive a “nondiscriminatory,” “just and reasonable rate for network elements,” the Act requires the FCC to decide how to value “the cost ... of providing the . . . network element [which] may include a reasonable profit,” although the FCC is (as already seen) forbidden to allow any “reference to a rate-of-return or other rate-based proceeding,” § 252(d)(1). Within the discretion left to it after eliminating any dependence on a “rate-of-return or other rate-based proceeding,” the Commission chose a way of treating “cost” as “forward-looking economic cost,” 47 CFR §51.505 (1997), something distinct from the kind of historically based cost generally relied upon in valuing a rate base after Hope Natural Gas. In Rule 505, the FCC defined the “forward-looking economic cost of an element [as] the sum of (1) the total element long-run incremental cost of the element [TELRIC]; [and] (2) a reasonable allocation of forward-looking common costs,” § 51.505(a), common costs being “costs incurred in providing a group of elements that “cannot be attributed directly to individual elements,” § 51.505(c)(1). Most important of all, the FCC decided that the TELRIC “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 incumbentfs] wire centers.” § 51.505(b)(1). “The TELRIC of an element has three components, the operating expenses, the depreciation cost, and the appropriate risk-adjusted cost of capital.” First Report and Order ¶ 703 (footnote omitted). See also 47 CFR §§ 51.505(b)(2)-(3) (1997). A concrete example may help. Assume that it would cost $1 a year to operate a most efficient loop element; that it would take $10 for interest payments on the capital a carrier would have to invest to build the lowest cost loop centered upon an incumbent carrier’s existing wire centers (say $100, at 10 percent per annum); and that $9 would be reasonable for depreciation on that loop (an 11-year useful life); then the annual TELRIC for the loop element would be $20. The Court of Appeals understood §252(d)(l)’s reference to “the cost ... of providing the . . . network element” to be ambiguous as between “forward-looking” and “historical” cost, so that a forward-looking ratesetting method would presumably be a reasonable implementation of the statute. But the Eighth Circuit thought the ambiguity afforded no leeway beyond that, and read the Act to require any forward-looking methodology to be “based on the incremental costs that an [incumbent] actually incurs or will incur in providing... the unbundled access to its specific network elements.” 219 F. 3d, at 751-753. Hence, the Eighth Circuit held that § 252(d)(1) foreclosed the use of the TELRIC methodology. In other words, the court read the Act as plainly requiring rates based on the “actual” not “hypothetical” “cost ... of providing the . . . network element,” and reasoned that TELRIC was clearly the latter. Id., at 750-751. The Eighth Circuit added, however, that if it were wrong and TELRIC were permitted, the claim that in prescribing TELRIC the FCC had effected an unconstitutional taking would not be “ripe” until “resulting rates have been determined and applied.” Id., at 753-754. The Court of Appeals also, and for the second time, invalidated Rules 315(c)-(f), 47 CFR §§51.315(c)-(f) (1997), the FCC’s so-called “additional combination” rules, apparently for the same reason it had rejected them before, when it struck down Rule 315(b), the main combination rule. 219 F. 3d, at 758-759. In brief, the rules require an incumbent carrier, upon request and compensation, to “perform the functions necessary to combine” network elements for an entrant, unless the combination is not “technically feasible.” Id., at 759. The Eighth Circuit read the language of § 251(c)(3), with its reference to “allowing] requesting carriers to combine ... elements,” as unambiguously requiring a requesting carrier, not a providing incumbent, to do any and all combining. Ibid. Before us, the incumbent local-exchange carriers claim error in the Eighth Circuit’s holding that a “forward-looking cost” methodology (as opposed to the use of “historical” cost) is consistent with § 252(d)(1), and its conclusion that the use of the TELRIC forward-looking cost methodology presents no “ripe” takings claim. The FCC and the entrants, on the other side, seek review of the Eighth Circuit’s invalidation of the TELRIC methodology and the additional combination rules. We granted certiorari, 531 U. S. 1124 (2001), and now affirm on the issues raised by the incumbents, and reverse on those raised by the FCC and the entrants. Ill A The incumbent carriers’ first attack charges the FCC with ignoring the plain meaning of the word “cost” as it occurs in the provision of § 252(d)(1) that “the just and reasonable rate for network elements ... shall be ... based on the cost (determined without reference to a rate-of-return or other rate-based proceeding) of providing the . . . network element . . . The incumbents do not argue that in theory the statute precludes any forward-looking methodology, but they do claim that the cost of providing a competitor with a network element in the future must be calculated using the incumbent’s past investment in the element and the means of providing it. They contend that “cost” in the statute refers to “historical” cost, which they define as “what was in fact paid” for a capital asset, as distinct from “value,” or “the price that would be paid on the open market.” Brief for Petitioners in No. 00-511, p. 19. They say that the technical meaning of “cost” is “past capital expenditure,” ibid., and they suggest an equation between “historical” and “embedded” costs, id., at 20, which the FCC defines as “the costs that the incumbent LEC incurred in the past and that are recorded in the incumbent LEC’s books of accounts,” 47 CFR § 51.505(d)(1) (1997). The argument boils down to the proposition that “the cost of providing the network element” can only mean, in plain language and in this particular technical context, the past cost to an incumbent of furnishing the specific network element actually, physically, to be provided. The incumbents have picked an uphill battle. At the most basic level of common usage, “cost” has no such clear implication. A merchant who is asked about “the cost of providing the goods” he sells may reasonably quote their current wholesale market price, not the cost of the particular items he happens to have on his shelves, which may have been bought at higher or lower prices. When the reference shifts from common speech into the technical realm, the incumbents still have to attack uphill. To begin with, even when we have dealt with historical costs as a ratesetting basis, the cases have never assumed a sense of “cost” as generous as the incumbents seem to claim. “Cost” as used in calculating the rate base under the traditional cost-of-service method did not stand for all past capital expenditures, but at most for those that were prudent, while prudent investment itself could be denied recovery when unexpected events rendered investment useless, Duquesne Light Co. v. Barasch, 488 U. S., at 312. And even when investment was wholly includable in the rate base, ratemakers often rejected the utilities’ “embedded costs,” their own book-value estimates, which typically were geared to maximize the rate base with high statements of past expenditures and working capital, combined with unduly low rates of depreciation. See, e. g., Hope Natural Gas, 320 U. S., at 597-598. It would also be a mistake to forget that “cost” was a term in value-based ratemaking and has figured in contemporary state and federal ratemaking untethered to historical valuation. What is equally important is that the incumbents’ plain-meaning argument ignores the statutory setting in which the mandate to use “cost” in valuing network elements occurs. First, the Act uses “cost” as an intermediate term in the calculation of “just and reasonable rates,” 47 U. S. C. § 252(d)(1), and it was the very point of Hope Natural Gas that regulatory bodies required to set rates expressed in these terms have ample discretion to choose methodology, 320 U. S., at 602. Second, it would have been passing strange to think Congress tied “cost” to historical cost without a more specific indication, when the very same sentence that requires “cost” pricing also prohibits any reference to a “rate-of-return or other rate-based proceeding,” § 252(d)(1), each of which has been identified with historical cost ever since Hope Natural Gas was decided. The fact is that without any better indication of meaning than the unadorned term, the word “cost” in § 252(d)(1), as in accounting generally, is “a chameleon,” Strickland v. Commissioner, Maine Dept. of Human Services, 96 F. 3d 542, 546 (CA1 1996), a “virtually meaningless” term, R. Estes, Dictionary of Accounting 32 (2d ed. 1985). As Justice Breyer put it in Iowa Utilities Bd., words like “cost” “give ratesetting commissions broad methodological leeway; they say little about the ‘method employed’ to determine a particular rate.” 525 U. S., at 423 (opinion concurring in part and dissenting in part). We accordingly reach the conclusion adopted by the Court of Appeals, that nothing in § 252(d)(1) plainly requires reference to historical investment when pegging rates to forward-looking “cost.” B The incumbents’ alternative argument is that even without a stern anchor in calculating “the cost... of providing the . . . network element,” the particular forward-looking methodology the FCC chose is neither consistent with the plain language of § 252(d)(1) nor within the zone of reasonable interpretation subject to deference under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, 843-845 (1984). This is so, they say, because TELRIC calculates the forward-looking cost by reference to a hypothetical, most efficient element at existing wire centers, not the actual network element being provided. 1 The short answer to the objection that TELRIC violates plain language is much the same as the answer to the previous plain-language argument, for what the incumbents call the “hypothetical” element is simply the element valued in terms of a piece of equipment an incumbent may not own. This claim, like the one just considered, is that plain language bars a definition of “cost” untethered to historical investment, and as explained already, the term “cost” is simply too protean to support the incumbents’ argument. 2 Similarly, the claim that TELRIC exceeds reasonable interpretative leeway is open to the objection already noted, that responsibility for “just and reasonable” rates leaves methodology largely subject to discretion. Permian Basin Area Rate Cases, 390 U. S. 747, 790 (1968) (“We must reiterate that the breadth and complexity of the Commission’s responsibilities demand that it be given every reasonable opportunity to formulate methods of regulation appropriate for the solution of its intensely practical difficulties”). See generally Chevron, supra, at 843-845, 866 (“When a challenge to an agency construction of a statutory provision, fairly conceptualized, really centers on the wisdom of the agency’s policy, rather than whether it is a reasonable choice within a gap left open by Congress, the challenge must fail”). The incumbents nevertheless field three arguments. They contend, first, that a method of calculating wholesale lease rates based on the costs of providing hypothetical, most efficient elements may simulate the competition envisioned by the Act but does not induce it. Second, they argue that even if rates based on hypothetical elements could induce competition in theory, TELRIC cannot do this, because it does not provide the depreciation and risk-adjusted capital costs that the theory compels. Finally, the incumbents say that even if these objections can be answered, TELRIC is needlessly, and hence unreasonably, complicated and impracticable. a The incumbents’ (and Justice Breyer’s) basic critique of TELRIC is that by setting rates for leased network elements on the assumption of perfect competition, TELRIC perversely creates incentives against competition in fact. See post, at 548-551. The incumbents say that in purporting to set incumbents’ wholesale prices at the level that would exist in a perfectly competitive market (in order to make retail prices similarly competitive), TELRIC sets rates so low that entrants will always lease and never build network elements. See post, at 549-550. And even if an entrant would otherwise consider building a network element more efficient than the best one then on the market (the one assumed in setting the TELRIC rate), it would likewise be deterred by the prospect that its lower cost in building and operating this new element would be immediately available to its competitors; under TELRIC, the incumbents assert, the lease rate for an incumbent’s existing element would instantly drop to match the marginal cost of the entrant’s new element once built. See ante, at 550; Brief for Respondents BellSouth et al. in Nos. 00-555, etc., pp. 28-29. According to the incumbents, the result will be, not competition, but a sort of parasitic free riding, leaving TELRIC incapable of stimulating the facilities-based competition intended by Congress. We think there are basically three answers to this no-stimulation claim of unreasonableness: (1) the TELRIC methodology does not assume that the relevant markets are perfectly competitive, and the scheme includes several features of inefficiency that undermine the plausibility of the incumbents’ no-stimulation argument; (2) comparison of TELRIC with alternatives proposed by the incumbents as more reasonable are plausibly answered by the FCC’s stated reasons to reject the alternatives; and (3) actual investment in competing facilities since the effective date of the Act simply belies the no-stimulation argument’s conclusion. (1) The basic assumption of the incumbents’ no-stimulation argument is contrary to fact. As we explained, the argument rests on the assumption that in a perfectly efficient market, no one who can lease at a TELRIC rate will ever build. But TELRIC does not assume a perfectly efficient wholesale market or one that is likely to resemble perfection in any foreseeable time. The incumbents thus make the same mistake we attributed in a different setting to the FCC itself. In Iowa Utilities Bd., we rejected the FCC’s neeessary-and-impair rule, 47 CFR §51.319 (1997), which required incumbents to lease any network element that might reduce, however slightly, an entrant’s marginal cost of providing a telecommunications service, as compared with providing the service using the entrant’s own equivalent element. 525 U. S., at 389-390. “In a world of perfect competition, in which all carriers are providing their service at marginal cost, the Commission’s total equating of increased cost (or decreased quality) with ‘necessity’ and ‘impairment’ might be reasonable, but it has not established the existence of such an ideal world.” Id., at 390. Not only that, but the FCC has of its own accord allowed for inefficiency in the TELRIC design in additional ways affecting the likelihood that TELRIC will squelch competition in facilities. First, the Commission has qualified any assumption of efficiency by requiring ratesetters to calculate cost on the basis of “the existing location of the incumbent[’s] wire centers.” 47 CFR § 51.505(b)(1) (1997). This means that certain network elements, principally local-loop elements, will not be priced at their most efficient cost and configuration to the extent, say, that a shorter loop could serve a local exchange if the incumbent’s wire centers were relocated for a snugger fit with the current geography of terminal locations. Second, TELRIC rates in practice will differ from the products of a perfectly competitive market owing to built-in lags in price adjustments. In a perfectly competitive market, retail prices drop instantly to the marginal cost of the most efficient company. See Mankiw 283-288,312-313. As the incumbents point out, this would deter market entry because a potential entrant would know that even if it could provide a retail service at a lower marginal cost, it would instantly lose that competitive edge once it entered the market and competitors adjusted to match its price. See Brief for Respondents BellSouth et al. in Nos. 00-555, etc., at 28-29. Wholesale TELRIC rates, however, are set by state commissions, usually by arbitrated agreements with 3- or 4-year terms, see Brief for Respondent Qwest Communications International, Inc., in Nos. 00-511, etc., p. 39; Reply Brief for Petitioners Worldcom, Inc., et al. 6; Reply Brief for Respondent Sprint Corp. 7, and n. 3; Reply Brief for Petitioner AT&T Corp. 11-12; and no one claims that a competitor could receive immediately on demand a TELRIC rate on a leased element at the marginal cost of the entrant who introduces a more efficient element. But even if a competitor could call for a new TELRIC rate proceeding immediately upon the introduction of a more efficient element by a competing entrant, the competitor would not necessarily know enough to make the call; the fact of the element’s greater efficiency would only become apparent when reflected in lower retail prices drawing demand away from existing competitors (including the incumbent), forcing them to look to lowering their own marginal costs. In practice, it would take some time for the innovating entrant to install the new equipment, to engage in marketing offering a lower retail price to attract business, and to steal away enough customer subscriptions (given the limited opportunity to capture untapped customers for local telephone service) for competitors to register the drop in demand. Finally, it bears reminding that the FCC prescribes measurement of the TELRIC “based on the use of the most efficient telecommunications technology currently available,” 47 CFR § 51.505(b)(1) (1997). Owing to that condition of current availability, the marginal cost of a most efficient element that an entrant alone has built and uses would not set a new pricing standard until it became available to competitors as an alternative to the incumbent’s corresponding element. As a reviewing Court we are, of course, in no position to assess the precise economic significance of these and other exceptions to the perfectly functioning market that the incumbents’ criticism assumes. Instead, it is enough to recognize that the incumbents’ assumption may well be incorrect. Inefficiencies built into the scheme may provide incentives and opportunities for competitors to build their own network elements, perhaps for reasons unrelated to pricing (such as the possibility of expansion into data-transmission markets by deploying “broadband” technologies, cf. post, at 552 (Breyer, J., concurring in part and dissenting in part), or the desirability of independence from an incumbent’s management and maintenance of network elements). In any event, the significance of the incumbents’ mistake of fact may be indicated best not by argument here, but by the evidence of actual investment in facilities-based competition since TELRIC went into effect, to be discussed at Part III-B-2a-(3), infra. (2) Perhaps sensing the futility of an unsupported theoretical attack, the incumbents make the complementary argument that the FCC’s choice of TELRIC, whatever might be said about it on its own terms, was unreasonable as a matter of law because other methods of determining cost would have done a better job of inducing competition. Having considered the proffered alternatives and the reasons the FCC gave for rejecting them, 47 CFR § 51.505(d) (1997); First Report and Order ¶¶ 630-711, we cannot say that the FCC acted unreasonably in picking TELRIC to promote the mandated competition. The incumbents present three principal alternatives for setting rates for network elements: embedded-cost methodologies, the efficient component pricing rule, and Ramsey pricing. The arguments that one or another of these methodologies is preferable to TELRIC share a basic claim: it was unreasonable for the FCC to choose a method of setting rates that fails to include, at least in theory, some additional costs beyond what would be most efficient in the long run, because lease rates that incorporate such costs will do a better job of inducing competition The theory is that once an entrant has its foot in the door, it will have a greater incentive to build and operate its own more efficient network element if the lease rates reflect something of the incumbents’ actual and inefficient marginal costs. And once the entrant develops the element at its lower marginal cost and the retail price drops accordingly, the incumbent will have no choice but to innovate itself by building the most efficient element or finding ways to reduce its marginal cost to retain its market share. The generic feature of the incumbents’ proposed alternatives, in other words, is that some degree of long-run inefficiency ought to be preserved through the lease rates, in order to give an entrant a more efficient alternative to leasing. Of course, we have already seen that TELRIC itself tolerates some degree of inefficient pricing in its existing wire-center configuration requirement and through the rate-making and development lags just described. This aside, however, there are at least two objections that generally undercut any desirability that such alternatives may seem to offer over TELRIC. The first objection turns on the fact that a lease rate that compensates the lessor for some degree of existing inefficiency (at least from the perspective of the long run) is simply a higher rate, and the difference between such a higher rate and the TELRIC rate could be the difference that keeps a potential competitor from entering the market. See n. 27, infra. Cf. First Report and Order ¶ 378 (“[I]n some areas, the most efficient means of providing competing service may be through the use of unbundled loops. In such cases, preventing access to unbundled loops would either discourage a potential competitor from entering the market in that area, thereby denying those consumers the benefits of competition, or cause the competitor to construct unnecessarily du-plicative facilities, thereby misalloeating societal resources”). If the TELRIC rate for bottleneck elements is $100 and for other elements (say, switches) is $10, an entering competitor that can provide its own, more efficient switch at what amounts to a $7 rate can enter the market for $107. If the lease rate for the bottleneck elements were higher (say, $110) to reflect some of the inefficiency of bottleneck elements that actually cost the incumbent $150, then the entrant with only $107 will be kept out. Is it better to risk keeping more potential entrants out, or to induce them to compete in less capital-intensive facilities with lessened incentives to build their own bottleneck facilities? It was not obviously unreasonable for the FCC to prefer the latter. The second general objection turns the incumbents’ attack on TELRIC against the incumbents’ own alternatives. If the problem with TELRIC is that an entrant will never build because at the instant it builds, other competitors can lease the analogous existing (but less efficient) element from an incumbent at a rate assuming the same most efficient marginal cost, then the same problem persists under the incumbents’ methods. For as soon as an entrant builds a more efficient element, the incumbent will be forced to price to match, and that rate will be available to all other competitors. The point, of course, is that things are not this simple. As we have said, under TELRIC, price adjustment is not instantaneous in rates for a leased element corresponding to an innovating entrant’s more efficient element; the same would presumably be true under the incumbents’ alternative methods, though they do not come out and say it. Once we get into the details of the specific alternative methods, other infirmities become evident that undermine the claim that the FCC could not reasonably have preferred TELRIC. As for an embedded-cost methodology, the problem with a method that relies in any part on historical cost, the cost the incumbents say they actually incur in leasing network elements, is that it will pass on to lessees the difference between most efficient cost and embedded cost. See First Report and Order ¶ 705. Any such cost difference is an inefficiency, whether caused by poor management resulting in higher operating costs or poor investment strategies that have inflated capital and depreciation. If leased elements were priced according to embedded costs, the incumbents could pass these inefficiencies to competitors in need of their wholesale elements, and to that extent defeat the competitive purpose of forcing efficient choices on all carriers whether incumbents or entrants. T