dissenting:
The must-carry provisions of the Cable Television Consumer Protection and Competition Act of 1992 (“1992 Act”) are ably set forth by the majority at pp. 3-6. In essence, they require cable operators to set aside just over a third of their channels for local broadcast stations. (The one-third-plus figure comprises both the commercial stations protected under § 4 of the Act and the noncommercial stations protected under § 5.) In addition, each privileged broadcast station has a right to its specific “channel position” on a cable system. 1992 Act at §§ 4(b)(6) and 5(g)(5).
In considering cable, Congress confronted a very real problem — one for which it has an easy remedy entirely consistent with the First Amendment. The problem is that cable systems control access “bottlenecks” to an important communications medium. In fact, while there are several thousand cable operators in the United States, only 53 communities are served by more than one operator. Debates, 138 Cong.Rec. S400, S436 (comments of Senator Gorton). There are well-developed regulatory responses to this sort of situation. The “bottleneck” holder may be ordered to serve all parties that meet neutral criteria for service. The Federal Communications Commission itself administers just such regulations in assuring that interstate long-distance telephone companies have access to local telephone networks. 47 U.S.C. § 202; and see, e.g., MTS and WATS Market Structure Phase III, 100 FCC 2d 860, 861 (1985) (adopting rules and policies relating to access). Indeed, courts have sometimes taken on the supervision of such compulsory access under the “essential facilities” doctrine of the antitrust laws. United States v. Terminal Railroad Ass’n, 224 U.S. 383, 32 S.Ct. 507, 56 L.Ed. 810 (1912). These solutions are, of course, imperfect; for example, they pose tricky pricing issues.1 But mandatory access rules of this sort give no special privilege to one set of access seekers over another.
In fact, in 1984 Congress adopted provision for neutral, compulsory access to cable in § 612 of the Cable Communications Policy Act of 1984 (“1984 Act”), 47 U.S.C. § 532, a solution somewhat refined in § 9 of the 1992 Act. This provision entitles independent programmers to lease access on cable systems, at reasonable prices to be set by the FCC. Although the channels available for lease under § 612 are limited to a specified proportion of each operator’s channels (depending on the total number that the operator possesses, 47 U.S.C. §§ 532(b)(l)(a)-(d)), *58Congress remains free to expand the fraction of channels available. The only programmers excluded from the benefits of § 612 are those affiliated with the cable operator, i.e., the very ones who would be the likely beneficiaries of operator discrimination and who are therefore in no need of the law’s intervention.
The must-carry provisions, by contrast, extend the privilege of access only to a special class of . unaffiliated programmers — local commercial and non-commercial television stations. Congress rested its decision to promote these stations in part, but quite explicitly, on a finding about their content — that they were “an important source of local news and public affairs programming and other local broadcast services critical to an informed electorate.” See 1992 Act, § 2(a)(ll). Moreover, because of FCC licensing requirements, every such local broadcaster is legally bound to “provide programming responsive to issues of concern to its community.” Report and Order, MM Docket No. 83-670, 98 FCC 2d 1076, 1091-92 (1984). See also National Association of Broadcasters (“NAB”) Corrected Memorandum at 24-25 (emphasizing force of mandate to carry local content). In other words, by virtue of provisions of law outside the 1992 Act, the term used by the Act to define the benefitted class automatically entails content requirements.
A number of parties challenge these mandates, among them the cable operators who must carry local broadcast stations in place of their own choice of programming, and, even more significantly, the unaffiliated cable programmers, such as Discovery Network, producer of the Discovery and Learning Channels, whose programming will be supplanted. I shall apply current First Amendment analysis to their claims, but before doing so I think it useful to imagine a few hypothetical cases outside the area of cable. If the answers to those hypotheticals are as clear as I think they are, then — unless there is something terribly special about cable, apart from the bottleneck issue addressed by § 612 of the 1984 Act — the must-carry provisions must fall.
1. The Washington Post develops and patents a special “paper-springer” that enables it to deliver papers at a tiny fraction of others’ costs. Assume that paper delivery costs represent a very large share of total costs, so that this technology is a source of overwhelming monopoly power. Congress decrees that The Post must license the paper-springer — but only to publishers of local, neighborhood papers (i.e., not citywide ones). This cuts out The Washington City Paper, The Washington Informer, and The Washington Times.
2. A state is concerned that large shopping centers represent vital gathering places of citizens and that center owners may not willingly allow the sort of leaflet-ting or soap-box oratory associated with a vibrant democracy. Accordingly it requires owners to allow such leafletting and oratory regardless of the leafletters’ or speakers’ message. See PruneYard Shopping Center v. Robbins, 447 U.S. 74 [100 S.Ct. 2035, 64 L.Ed.2d 741] (1980). But, finding that local residents are an especially important source of this vibrant debate, it limits the privilege to persons living within four miles of each shopping center. This excludes the proverbial “outside agitator”.
These scenarios are in fact less troublesome under the First Amendment than the must-carry rules; the parties selected to benefit from the bottleneck-breaking rules are defined in content-neutral terms. Here, by contrast, the beneficiaries are local stations that are required by government to include specific content that the government has deemed especially worthy.2
*59Standard of Review. The standard of review in First Amendment cases usually depends on whether the regulation is content-based, leading to “strict scrutiny”, Arkansas Writers’ Project, Inc. v. Ragland, 481 U.S. 221, 231, 107 S.Ct. 1722, 1729, 95 L.Ed.2d 209 (1987); Carey v. Brown, 447 U.S. 455, 461-62, 100 S.Ct. 2286, 2290-91, 65 L.Ed.2d 263 (1980), or content-neutral, leading to more relaxed scrutiny under the formula of United States v. O’Brien, 391 U.S. 367, 88 S.Ct. 1673, 20 L.Ed.2d 672 (1968), as modified in Ward v. Rock Against Racism, 491 U.S. 781, 109 S.Ct. 2746, 105 L.Ed.2d 661 (1989). Here the statute is content-based under controlling precedent. The must-carry provisions mandate speech: they require cablecasters to carry the speech of local broadcast stations. As the Supreme Court held in Riley v. Nat’l Fed. of the Blind of North Carolina, 487 U.S. 781, 108 S.Ct. 2667, 101 L.Ed.2d 669 (1988), “Mandating speech that a speaker would not otherwise make necessarily alters the content of the speech. We therefore consider the Act [mandating certain disclosures by charitable solicitors] as a content-based regulation of speech.” Id. at 795, 108 S.Ct. at 2677. This alteration of content is equally certain here; under the must-carry mandate, cablecasters must replace the programs they would have chosen, regardless of origin, with programs selected by local broadcasters.
The Riley Court did not, however, immediately proceed to apply strict scrutiny. Rather, addressing a claim that the speech was basically commercial, the Court rejected that view on the ground that the commercial elements were “inextricably intertwined” with fully protected non-commercial speech. Id. at 796, 108 S.Ct. at 2677. More generally, it said that the Court’s “lodestars in deciding what level of scrutiny to apply to a compelled statement must be the nature of the speech taken as a whole and the effect of the compelled statement thereon.” Id. While here we have no issue of commercial speech, I read the Riley analysis as a caution against assuming that all compulsions of speech, even fully protected speech, must receive strict scrutiny. We must look further.
Where a government regulation of conduct has an “incidental” burden on speech, the Court looks to whether the interest asserted by the government “is unrelated to the suppression of free expression”, Texas v. Johnson, 491 U.S. 397, 407, 109 S.Ct. 2533, 2541, 105 L.Ed.2d 342 (1989) (internal quotations omitted), or, as the Court expressed it later on the page, “unconnected to expression”, id. Application of a test for “incidental” burdens on speech seems highly artificial here. Given the finite number of cable channels,3 replacement of the cablecaster’s choice of programs with those of local broadcasters suppresses the alternative programs as completely as if Congress had ordered them shut down; there is nothing “incidental” about the burden. Further, one of the’ interests explicitly asserted by Congress was the benefit of “local news and public affairs programming”, 1992 Act, § 2(a)(ll); and indeed the class of chosen beneficiaries is defined by other legal provisions as persons bound to include some minimum level of local programming content. Moreover, the Court applies strict scrutiny to rules that control editorial discretion, Miami Herald Pub. Co. v. Tornillo, 418 U.S. 241, 94 S.Ct. 2831, 41 L.Ed.2d 730 (1974), which surely comprises the choice of programs. See Leathers v. Medlock, — U.S. —, —, 111 S.Ct. 1438, 1442, 113 L.Ed.2d 494 (1991) (noting that cable “is engaged in ‘speech’ under the First Amendment, and is, in much of its operation, part of the ‘press’ ”); see also City of Los Angeles v. Preferred Communications, Inc., 476 U.S. 488, 494, 106 S.Ct. 2034, 2037, 90 L.Ed.2d 480 (1986) (noting that cable operators exercise significant editorial discretion).
On the other hand, in PruneYard Shopping Center v. Robins, 447 U.S. 74, 100 S.Ct. *602035, 64 L.Ed.2d 741 (1980), the Court assessed mandated speech without any explicit reference to strict scrutiny. The case hardly establishes that speech mandates need not pass such scrutiny, however. ^ Free speech was on both sides of the balance in Prune-Yard, for the challenged state rule entitled anyone to engage in various expressive activities such as speaking and leafletting at shopping centers. The owner’s only First Amendment claim was its right to silence, and the Court found that interest of little weight in context: the owner had dedicated the property to commercial use open to the public; it could readily disassociate itself from the ideas expressed; and the rule made no inroad into editorial functions. Id. at 87-88, 100 S.Ct. at 2044. While the Court did not use the language of strict scrutiny, the rule would likely have survived such review, given the state’s purpose of assuring expressive opportunities on a completely nondiscriminatory basis, and the weakness of the shopping center’s competing speech interests. Thus PruneYard cannot reasonably be taken to allow speech mandates to escape strict scrutiny.
The Supreme Court has also developed rules governing the First Amendment evaluation of taxes on the media. While the must-carry regulations are not in form taxes, they are in substance indistinguishable from a tax-subsidy combination — a tax falling exclusively on cablecasters, in an amount large enough to fund a subsidy for local broadcasters to enable them to lease cable channels. The Court has insisted on a “compelling justification” for any use of the taxing power “to single out the press”. Leathers, — U.S. at —, 111 S.Ct. at 1443; see also Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U.S. 575, 585, 103 S.Ct. 1365, 1371-72, 75 L.Ed.2d 295 (1983). Where a tax not only singles out the press but “targets a small group of newspapers”, even by entirely content-neutral categories, the government must carry a “heavy burden” to sustain the distinction. Id. at 591, 592-93, 103 S.Ct. at 1375, 1375-76. See also Arkansas Writers’ Project, Inc. v. Ragland, 481 U.S. 221, 107 S.Ct. 1722, 95 L.Ed.2d 209 (1987) (imposing “heavy burden” of justification for tax that singles out a subset of the press for a special tax, based on broad content categories).
Leathers itself, of course, upheld a tax statute that distinguished between cable operators and broadcasters, and indeed specified that in the context of a generally applicable tax the distinction was not subject to heightened scrutiny. — U.S. at— —, 111 S.Ct. at 1444-47. Although the Court observed broadly that “differential taxation of speakers, even members of the press, does not implicate the First Amendment unless the tax is directed at, or presents the danger of suppressing, particular ideas”, id. at —, 111 S.Ct. at 1447, the observation was in the context of a “broad-based, content-neutral sales tax” that fell on virtually every good or service sold in Arkansas, id. The treatment of First Amendment values as irrelevant cannot be transported to the context of a burden imposed on one set of speakers for the direct and explicit advantage of a limited class of their competitors — a class whose programming must, as a matter of law, include content of a type specified by the government. Accordingly, I conclude that the proper test is strict scrutiny.
Application of the test. To survive strict scrutiny, regulation of speech must serve a “compelling” governmental purpose and its “means must be carefully tailored to achieve those ends.” Sable Communications v. FCC, 492 U.S. 115, 126, 109 S.Ct. 2829, 2836, 106 L.Ed.2d 93 (1989). First we must identify the government’s interests, and then assess the fit between each of those interests and the must-carry requirements.
The Senate Report articulated the congressional interests as follows:
(1) preserving the benefits of local television service, particularly over-the-air television service; (2) promoting the widespread dissemination of information from diverse sources; and (3) promoting fair competition in the video marketplace.
S.Rep. 102-92 at 58, U.S.Code Cong. & Admin.News 1992, at 1191. These purposes are reflected in §§ 2(6) — (15) of the Act. Because this statement of purposes involves some overlap, the analysis will be clearer if we reformulate them as follows: (A) preservation of *61open access to cable in order to assure diverse programming (see purposes (2) and (3) above); and (B) preservation of local broadcasting (see purposes (1) and (3) above). This second reason involved not only a desire to promote “local news and public affairs programming”, see § 2(a)(ll) of the Act, but .also a wish to assure that local broadcasters could continue to serve those persons who either did not, or could not, subscribe to cable. I address diversity first, then the preservation of local broadcasters.
Open Access for Diverse Programming
I assume that, at least at some level of abstraction, the interest in promoting diversity of views on cable is compelling. With regard to the broadcast spectrum the Court has characterized programming diversity as an “important governmental objective”, Metro Broadcasting, Inc. v. FCC, 497 U.S. 547, 566, 110 S.Ct. 2997, 3010, 111 L.Ed.2d 445 (1990), and a lengthy passage in Miami Herald Pub. Co. v. Tornillo, 418 U.S. at 247-54, 94 S.Ct. at 2834-38, though inconclusive, appears to regard the interest in correcting newspaper monopolies as extremely valuable. In Tomillo, however, the Court balked at the state’s remedy — a right-of-reply statute — on the ground that it would interfere with editorial discretion. The interference took two forms: first, the statute compelled “editors or publishers to publish that which reason tells them should not be published”, id. at 256, 94 S.Ct. at 2839 (internal quotations omitted), and second, it tended to dissuade editors from publishing those controversial statements that would trigger the right of reply, id. at 257, 94 S.Ct. at 2839. Framing the case in modern doctrine, the Court may be said to have found the solution too intrusive to satisfy the requirement of careful “tailoring”.
Here, too, the fit between the interest and the regulation is extremely weak. It is far from clear that giving local broadcasters an entitlement to be carried will increase program diversity at all. Because eablecasters now carry the vast majority of local stations (as explained further below), the must-carry rules may have little effect, but where they have any, it will be only to replace the mix chosen by eablecasters — whose livelihoods depend largely on satisfying audience demand — with a mix derived from congressional dictate.
Moreover, in considering “fit” one must look at the available alternatives. As the Court said recently in assessing the fit of a statute that burdened commercial speech, “[I]f there are numerous and obvious less-burdensome alternatives to the restriction on commercial speech, that is certainly a relevant consideration in determining whether the ‘fit’ between ends and means is reasonable.” City of Cincinnati v. Discovery Network, Inc., — U.S. —, — n. 13, 113 S.Ct. 1505, 1510 n. 13, 123 L.Ed.2d 99 (1993). Here, one obvious alternative is expansion of the access provisions of § 612 of the 1984 Act. Those access provisions are far less burdensome — in the critical sense of minimizing government interference in the choice of who will have access to cable. Under § 612, all programmers are eligible for leased access except the ones that don’t need it — the affiliates of a cable operator. 47 U.S.C. § 532(b)(1).
Even neutral remedies such as § 612 are subject to challenge under Tomillo, as they replace the cablecaster’s speech with that of others. But a number of distinguishing factors argue for upholding the leased access solution. First, a relatively large number of channels is available (ranging from the teens up to 40, 50 and more), so that the scope of the monopolization is greater than in Tomillo. Second, the channels are discrete, so that viewers will be less likely to attribute all speech on all channels to the cablecaster. And there is no chilling effect on the cable-caster’s programming of the remaining channels. Finally, the benefitted group precisely fits the legislative concern — all unaffiliated programmers.
The last is conspicuously missing from must-carry. The difference between the two is the difference between the state rule upheld in PmneYard Shopping Center and the hypothetical set forth at the start of this opinion — a PmneYard-type entitlement limited to a specific class of would-be speakers defined in terms of their location. Indeed, the must-carry requirements are worse than the PmneYard hypothetical, for admission to *62the privileged class requires, as a matter of law, the carrying of specific program content.
It is quite true, of course, that Congress found in 1992 that leased access under § 612 has not been as effective as predicted in 1984, pointing to cumbersome enforcement procedures and undue cablecaster control over prices and other conditions of access. See ELRep. 102-628 at 39-40. But the Report went on to “restate[ ]” the Committee’s “belief that access requirements establish a form of content-neutral structural regulation “which will foster the availability of a diversity of viewpoints to the listening audience’.” Id. at 40. More important, to the extent that Congress spotted defects in § 612, it set out to correct them. The 1992 Act amends § 612 by clarifying and broadening the FCC’s authority to set maximum reasonable rates and other reasonable terms and conditions for access. See 1992 Cable Act, § 9. Given § 612, and Congress’s authority to expand its scope, the must-carry rules do not provide a reasonable fit with the diversity rationale. Preservation of Local Broadcasting
The second interest that Congress has invoked is the preservation of local broadcasting. Section 2(a)(10) of the Act is explicit: “A primary objective and benefit of our Nation’s system of regulation of television broadcasting is the local origination of programming. There is a substantial governmental interest in ensuring its continuation.” In part this interest derived from the local content of such broadcasting — “an important source of local news and public affairs programming.” 1992 Cable Act § 2(a)(ll). But in addition the interest derived from an indirect concern for the households not served by cable — both those unwilling or unable to pay for the service and the 10% for whom such service is physically unavailable. Here Congress reasoned that if broadcast stations ever lost the advertising revenues attributable to the audiences they reach through cable service, they would be driven from business and thus be unable to provide service for those without cable. There was also special concern that local broadcast stations were exceptionally vulnerable as potential victims of cablecaster bottleneck control, because of their competition with cable for advertising revenue. See 1992 Act, § 2(a)(14)~ (16). I will first address the issue of local content, then the preservation of over-the-air service.
Local content. It seems extremely doubtful that forcing local affairs content on First Amendment speakers could ever qualify as a compelling interest (at least outside the special preserve of broadcasting itself, to which I return below). If government “may not select which issues are worth discussing or debating in public facilities”, Police Department of Chicago v. Mosley, 408 U.S. 92, 96, 92 S.Ct. 2286, 2290, 33 L.Ed.2d 212 (1972), it is hard to see why it may do so on private facilities.
Even if by some stretch an interest in local content were “compelling”, far less restrictive means are readily available. Congress is free to subsidize favored speech if the market is too weak to sustain the quantity that Congress seeks. See, e.g., Rust v. Sullivan, —U.S.—,—, 111 S.Ct. 1759, 1774, 114 L.Ed.2d 233 (1991); cf. Regan v. Taxation with Representation, 461 U.S. 540, 544-48, 103 S.Ct. 1997, 2000-02, 76 L.Ed.2d 129 (1983); compare Arkansas Writers’ Project (striking down content-based tax burden). Given those means, Congress cannot advance specific content by requiring a competing class of First Amendment speakers to carry the favored speech. “[T]he concept that government may restrict the speech of some elements of our society in order to enhance the relative voice of others is wholly foreign to the First Amendment”. Buckley v. Valeo, 424 U.S. 1, 48-49, 96 S.Ct. 612, 649, 46 L.Ed.2d 659 (1976).
Preserving over-the-air TV service. I assume that as an abstract matter government has a compelling interest in assuring access to TV for those unwilling or unable to subscribe to cable, and especially those beyond cable’s physical reach. The first difficulty with the argument here is that there is no evidence that this access is in jeopardy. The interest, as the Court said of a concern for breaches of the peace in Texas v. Johnson, 491 U.S. 397, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989), simply “is not implicated on this record”. Id. at 407, 109 S.Ct. at 2541.
*63The most plausible evidence would be of local TV stations going under and of licensees turning in their licenses. Nothing remotely like that is in sight — the facts show quite the opposite. Plaintiff National Cable Television Association asserts (without contradiction) that in the years since the D.C. Circuit Court of Appeals invalidated the FCC’s first effort at must-carry, Quincy Cable TV, Inc. v. FCC, 768 F.2d 1434 (D.C.Cir. 1985), the number of commercial broadcast stations has increased by 22% (from 919 to 1118), the number of educational broadcast stations by 15% (from 316 to 363), and the number of cities receiving broadcast television by 16% (from 514 to 594). NCTA’s Statement of Material Facts ¶¶ 12-14. Thus, not only are licenses not being turned in for want of profitable opportunities, but evidently entrepreneurs continue to seek the additional licenses that the FCC makes available. As a result, they are both deepening and spreading the coverage of over-the-air TV. Whatever risk there may be in the abstract has completely failed to materialize.
Defendants and defendant-intervenors urge in response that this record means little; since Quincy, the shadow of possible congressional intervention has fallen across the cable industry, causing operators to be less ruthless than they would be if must-carry were invalidated. Remove the shadow, they claim, and the operators will show their trae monopolist colors.
This argument is incurably flawed. If the constitutionally fatal aspect of the must-carry rules is the absence of evidence of any visible threat to the over-the-air TV industry, then the rules would become constitutional the minute the industry was visibly threatened. Accordingly, cablecasters would continue to operate under the same shadow of congressional intervention, and over-the-air TV would continue to thrive. Invalidation of must-carry for want of evidence would preserve the supposedly necessary threat.
Supporters of must-carry thus turn to evidence of (1) instances where cablecasters have actually dropped broadcast stations and (2) structural relationships in the industry, which, they say, indicate the threat is real. Indeed, in its findings, Congress noted a “marked shift in market share from broadcast television to cable television”, 1992 Act, § 2(a)(13), and also endorsed the structural argument, id. at §§ 2(a)(12), (14^15).
The congressional finding on actual effects is not disputed — nor is it supportive of must-carry. That cable has grown faster than broadcast does not in the slightest suggest that broadcast is in peril.
The record before Congress did, however, include evidence of some cable operators’ dropping broadcast channels. Specifically, the Senate Report cites evidence that, of 4303 cable systems disclosing data, 869 had dropped one or more local stations in a total of 1,820 instances.4 S.Rep. 102-92 at 43. Because only half of the more than 8000 cable systems disclosed data, the committee extrapolated to find that approximately 1700 cable systems have denied carriage in approximately 3600 instances. Id.
These facts do not support an inference that over-the-air TV is at risk. First, they indicate (by simple subtraction) that 80% of cable systems have never dropped a single local broadcaster. Second, it is impossible to know what to make of the 3600 estimated instances of dropping when we do not know either the percentage of total broadcasters that the figure represents or the character of the broadcast channels dropped. What pertinent data there is on the record states that the average cable system in 1990 carried just over eight local broadcast signals (7% more than were carried in 1985). Klein Affidavit at ¶ 13. Taking Congress’s figure of over 8000 cable systems, this would indicate that there are at least 64,000 local broadcast signals being carried on cable systems (clearly this figure counts single stations multiple times, just as does the raw figure on drops). Next to this figure, the 3600 instances of droppage5 do not support an inference of *64peril to the broadcast industry or even any segment of that industry. In fact, plaintiff NCTA submits further, uncontradicted evidence that in 1988, well after the FCC must-carry rules had been struck down, 98% of all the broadcast stations that would have qualified for mandatory carriage were still being carried despite the absence of such a requirement. Klein Affidavit at ¶ 12.
There is furthermore no information about the characteristics of the broadcasters actually dropped. Such distinctions are important. Plaintiff Time-Warner Entertainment (“TWE”) alleges without contradiction that the must-carry rules will require it to carry, in its Staten Island, New York line-up, not one but two stations from Bridgeport, Connecticut. TWE Memorandum in Support of Motion for Summary Judgment at 12. If many of the stations dropped have as remote a link to the dropping cableeaster’s market as do these Bridgeport stations, it would explain why the cable operators’ right to drop has coexisted with the broadcast industry’s continued growth. Thus we have nothing to connect the limited evidence of cable operators’ actual conduct with any inference that a threat to broadcast is imminent or serious.
The structural argument is that the competition between cable and broadcast for advertising revenue gives cable operators an incentive to drop broadcast stations. See 1992 Act, §§ 2(a)(12), (14-15). (Taken at its best, this theory would lend no support to § 5 of the Act (relating to noncommercial TV), as the educational stations do not advertise.) The evidence cited in the Senate Report consists of a projected 17% growth rate in cable advertising revenues. S.Rep. 102-92 at 44. Yet rapid growth in advertising seems easily attributable to the growth of cable, and perhaps some branching into advertising-supported programs. In itself, it shows neither a purpose to undermine broadcast TV nor the start of a campaign to do so. Broadcast television evidently still accounts for 92% of all television advertising revenues. See Hendricks Declaration at ¶ 22(d).6
Further, cable television appears to continue to be dependent on local broadcasters as critical suppliers of programming. Congress explicitly found that “broadcast programming that is carried remains the most popular programming on cable systems, and a substantial portion of the benefits for which consumers pay cable systems is derived from carriage of the signals of network affiliates, independent television stations, and public television stations.” 1992 Act, § 2(a)(19). In fact, local broadcaster programming accounts for approximately two-thirds of total cable viewing hours. S.Rep. 102-92 at 35. Moreover, the ratio of cable subscription fee revenues to advertising revenue is evidently 25:1. TWE Reply Memorandum in Support of Motions for Summary Judgment at 15. So long as local broadcast programs continue to be so popular, it appears that cable simply cannot afford to drop broadcast channels to any significant degree.
Finally, even if local broadcasters were in perceptible peril, it would not follow that Congress could secure their survival by must-carry provisions. Again, less intrusive alternatives are obvious. First, if evidence of any risk should appear, Congress could get broadcast TV into the homes of cable viewers (and thus preserve the stations’ advertising *65revenues) by expanding leased access under the neutral provisions of § 612 of the 1984 Act. Second, if broadcasters’ programs were not popular enough to enable them to pay the leasing fees set by the FCC, Congress could subsidize the difference to the extent necessary. Compare Rust v. Sullivan, — U.S. at —, 111 S.Ct. at 1774.
In reaching these conclusions, I do not question Congress’s power to engage in critical fact-finding. The degree of deference that we owe such findings is hotly disputed between the parties. Compare NCTA Reply Memorandum at 41-42 (citing Sable Communications of California v. FCC, 492 U.S. 115, 129, 109 S.Ct. 2829, 2837, 106 L.Ed.2d 93 (1989); Landmark Communications, Inc. v. Virginia, 435 U.S. 829, 843-44, 98 S.Ct. 1535, 1543-44, 56 L.Ed.2d 1 (1978), with DOJ Answer at 37 (citing Metro Broadcasting, Inc. v. FCC, 497 U.S. 547, 572, 110 S.Ct. 2997, 3012-13, 111 L.Ed.2d 445 (1990). My view does not turn on any resolution of that dispute, because the legislative findings simply do not support the inferences needed to sustain must-carry. The problems are that (1) there is no finding of any present or imminent harm; (2) the evidence of some dropping of broadcast channels in itself fails to show any widespread problem; (3) the proliferation of local broadcast stations since the end of the FCC’s must-carry rules undermines any inference of a problem; (4) the findings as to structure and incentives, taken together with the evidence of cable’s dependence on broadcasting, fail to raise the concern beyond the level of speculation; and (5), even if the hazard were perceptible, the record does not address the less intrusive alternatives. If findings as scantily connected to the conclusion as these can justify must-carry, then the door is open — even in the area of First Amendment rights — to exercise of the most naked interest-group preferences. Cf. Cass R. Sunstein, “Naked Preferences and the Constitution”, 84 Colum.L.Rev. 1689 (1984).
Thus I conclude that — unless the analysis of First Amendment issues in the special eontext of broadcast for some reason supervenes — the must-carry provisions violate the First Amendment.
Proposed Use of Broadcast Analysis
Some defendants urge the court to use the sort of diluted First Amendment analysis applied to broadcasting in Red Lion Broadcasting Co. v. FCC, 395 U.S. 367, 89 S.Ct. 1794, 23 L.Ed.2d 371 (1969). DOJ Answer at 26; NAB Corrected Memorandum, etc. at 46. Use of the broadcast standards would be completely inappropriate.
The weakened status of broadcasters under the First Amendment arose from the Supreme Court’s conclusion that discretionary government allocation of channels was necessary for the orderly use of the broadcast medium — or at any rate that such allocation was not materially more intrusive on First Amendment values than any visible alternative. In National Broadcasting Co. v. U.S., 319 U.S. 190, 63 S.Ct. 997, 87 L.Ed. 1344 (1943), the Court upheld the FCC’s so-called “Chain Broadcasting” regulations (aimed at preventing monopolization of radio licenses) against a variety of statutory challenges and a First Amendment claim. “[R]adio inherently is not available to all.” Id. at 226, 63 S.Ct. at 1014. “That is its unique characteristic, and that is why, unlike other modes of expression, it is subject to governmental regulation.” Id. Thus, “The right to free speech does not include ... the right to use the facilities of radio without a license.” Id. at 227, 63 S.Ct. at 1014.
Although the precise question was not before it, the Court appeared to assume that these factors logically entailed discretionary government allocation of licenses, so that such allocation presented no First Amendment problem so long as the choice was not openly based on the applicants’ “political, economic or social views, or upon any other capricious basis.” Id. at 226, 63 S.Ct. at 1014.7 This of course overlooked content-neutral bases for handling the problems, such as (1) the rule of first possession, which in fact preceded government allocation and *66tracked the doctrine of prior appropriation,8 (2) auctions, and (3) lotteries.9 These content-neutral alternatives evidently went by default, as those raising the First Amendment claims — the incumbent broadcasters- — • had no incentive to undermine the system under which they held their own quasi-monopoly licenses. NBC of course paved the way for Red Lion, which upheld, for the broadcast medium, the sort of government-managed right to respond that the Court later struck down for the print media in Tornillo. Compare Red Lion, 395 U.S. at 386-401, 89 S.Ct. at 1804-12, with Tornillo, 418 U.S. at 247-58, 94 S.Ct. at 2834-40. Here, of course, not even the defendants or defendant-intervenors suggest that discretionary allocation of channels is necessary or even appropriate to preserve access to cable. In the face of § 612, such a claim would be untenable.
Defendants and defendant-intervenors, however, seeking to deny cable operators full First Amendment protection, point to three special characteristics of cable that in their view justify government management of the channels: (1) the cable operators’ need for rights-of-way, along or under city streets, in order to lay their cable; (2) the cable industry’s prior enjoyment of special government privileges; and (3) the ability of cable operators to engage in “private censorship” against the speech of broadcasters. Neither alone nor in the aggregate do these justify subjecting cable to diminished First Amendment protection.
Rights-of-way. State, city and county governments have property interests in the streets, so that the placement of cable commonly requires government consent. There may also be instances where the federal government holds a proprietary interest essential to the laying of cable. These governments may seek to condition their consents on the operators’ waiver or abandonment of their First Amendment rights. If so, courts would presumably assess these efforts under the doctrine of unconstitutional conditions. Elrod, v. Burns, 427 U.S. 347, 357-60, 96 S.Ct. 2673, 2681-83, 49 L.Ed.2d 547 (1976); Perry v. Sindermann, 408 U.S. 593, 596-98, 92 S.Ct. 2694, 2697-98, 33 L.Ed.2d 570 (1972); Western Oil & GasAss’n v. Cory, 726 F.2d 1340 (9th Cir.1984) (state method of calculating charges for use of right-of-way invalidated under negative commerce clause), aff'd 471 U.S. 81, 105 S.Ct. 1859, 85 L.Ed.2d 61 (1985); Omega Satellite Products Co. v. City of Indianapolis, 694 F.2d 119, 128-29 (7th Cir.1982). We need not speculate on how such eases would work out. Here the federal government is not attaching conditions to the grant of a property interest; it is acting in a purely regulatory role. It cannot lean on proprietary powers that are out of the picture.
Prior benefits. Defendants argue that cable’s initial growth depended, perhaps essentially, on government assistance in securing the right to retransmit the programs of broadcast channels, free of charge. See 17 U.S.C. § 111, as amended (compulsory copyright access); see DOJ Memorandum in Support of Motion to Dismiss at 27-28. This history seems completely irrelevant. If the Union Pacific Railroad were to own a newspaper, its history as' a recipient of government largesse on a grand scale would surely not condemn it to second-class First Amendment protection. Of course current federal government benefits would provide an opportunity for attachment of conditions; but again, just what conditions might prove justifiable in the face of the unconstitutional conditions doctrine is not before us. With the adoption of § 6 of the 1992 Act, Congress gave broadcast stations an unequivocal right to withhold consent to retransmission.
Cablecasters’ power to censor. Finally, intervenor-defendant NAB argues that cable operators can use their position to censor broadcasters. NAB Corrected Memorandum at 46. This argument amounts to nothing more than a reiteration of the fact that cable operators control a bottleneck. The mere fact that such control might force some programmers to find different ways to reach the public (which all programmers have via *67§ 612, and which broadcasters uniquely have via the airwaves) does not give the government any right to force access on behalf of a preferred class of speakers. This follows a fortiori from Tornillo.
In short, the special character of cable is limited to one feature: bottleneck control of access to an important medium. A straightforward, speech-neutral solution to that problem exists and has already been employed by Congress in § 612 of the 1984 Act, as amended by § 9 of the 1992 Act. I see no constitutional obstacle to the expansion of that solution if Congress should deem it appropriate. Given the availability of that remedy (together with such subsidies as Congress might find suitable), I do not see how Congress can constitutionally deny cable operators the ordinary rights of any First Amendment speaker.10
Conclusion
The must-carry regulations in the 1992 Cable Act clearly burden the protected speech of cable operators, in favor of local broadcasters whose programming content is in material part specified by law. In requiring cable systems to carry a special group of competing speakers, Congress directly, not incidentally, restricts the cable operators’ exercise of editorial discretion. None of the interests advanced by Congress supports such a burden. The diversity rationale fits poorly with mandatory carriage of a specific group of programmers, being served "with neutrality by § 612’s provision for leased access. Although the interest in protecting over-the-air TV for non-cable-subscribers may be compelling in the abstract, the findings in the record neither indicate any real threat nor suggest any flaw in the less burdensome and obvious means for addressing any such threat. For these reasons, I respectfully dissent, and would declare the must-carry provisions to be unconstitutional abridgments of the First Amendment rights of cable operators and unaffiliated programmers.
. See, e.g„ William J. Baumol, "Deregulation and Residual Regulation of Local Telephone Service”, Ch. VII, American Enterprise Institute Studies in Telecommunication Deregulation (draft, March 3, 1993). (Forthcoming as William *58J. Baumol and J. Gregory Sidak, "Toward Competition in Local Telephony,” M.I.T. Press, 1993).
. Government line-drawing in First Amendment cases can be analyzed both in terms of the distinction between content-based and content-neutral rules, and in terms of underinclusiveness. An underinclusiveness claim, when the classification is not content-based, is extremely weak; typically the law's challenger is saying in essence that the law should fall because it burdens too little speech. See, e.g., Walsh v. Brady, 927 F.2d 1229, 1236 (D.C.Cir.1991). This seems counter-intuitive, and such claims rarely prevail. See, e.g., Austin v. Michigan Chamber of Commerce, 494 U.S. 652, 110 S.Ct. 1391, 108 L.Ed.2d 652 (1990). Here, however, the operators, and especially the independent cable programmers, do not say that the law fails because the burden is too narrow; they say it fails because the beneficiaries are so narrowly defined, and that the *59underinclusiveness thwarts rather than furthers the legitimate purpose of diversity. Even apart from the content-based character of the provision, then, the claim is stronger than the usual one of underinclusiveness.
. There is uncontroverted evidence that 2000 cable systems, serving one-third of all subscribers, have no excess channel capacity. NCTA Statement of Material Facts at V 6. It is further undisputed that there are many unaffiliated, non-broadcast programmers seeking access, but who will lose out because of the preferential treatment of broadcasters. Id. at ¶ 8; Klein Affidavit at ¶ 6.
. Clearly some stations were dropped on more than one occasion, indicating that being dropped must in many instances have been temporary. We have not been directed to any evidence of how many stations may have been permanently dropped.
. The record doesn't indicate the time frame within which these drops occurred. If it was a *64period of years, this even further weakens the inferential chain.
. The National Association of Broadcasters points to evidence before Congress that on some occasions cablecasters repositioned broadcast channels and placed less popular programs at the former broadcast position. NAB Corrected Memorandum, etc. at 22. This is urged as confirmation that cable operators will pursue advertising revenue at the expense of viewer satisfaction. At oral argument cable representatives noted in response that the data referred to periods immediately after the switch, when the new programs had little or no reputation with audiences. In any event, the reasoning of the broadcasters is hard to follow, and seems especially ironic coming from firms that depend exclusively on advertising revenues. Broadcasters might well increase their audiences by dropping advertisements, but no one expects them to do so; it is recognized that they accept detriments to audience satisfaction where justified by increases in net revenue. That cable operators may make some such calculations is no evidence of anti-competitive purpose. In fact, because cable operators' revenues are overwhelmingly due to subscription fees, see below, advertising seems likely to be less common on cable.
. I use the qualifying adverb "openly” advisedly. Careful research suggests, as anyone would expect, that often the political allegiance of applicants has been critical in license allocation. See, e.g., Lucas A. Powe, Jr., American Broadcasting and the First Amendment 68-84 (1987).
. See Thomas W. Hazlett, The Rationality of U.S. Regulation of the Broadcast Spectrum, 33 J.L. & Econ. 133, 147-52, 163 (1990).
. Cf. 47 U.S.C. § 309(i).
. Because of my conclusion on the First Amendment challenge to the must-cariy provisions, I do not reach the contention of TWE and Daniels that those provisions also represent an unconstitutional taking of cablecasters’ property in violation of the Fifth Amendment. I do not, however, regard the claim as frivolous. The creation of an entitlement in some parties to use the facilities of another, gratis, would seem on its face to implicate Loretto v. Teleprompter Manhattan CATVCorp., 458 U.S. 419, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982), where the Court struck down a statute entitling cable companies to place equipment in an owner’s building so that tenants could receive cable television. The NAB responds that Loretto is limited to "physical” occupations of "real property". See NAB Opposition at 33. But the insertion of local stations' programs into a cable operator's line-up presumably is not a metaphysical act, and presumably takes place on real property.
NAB also argues that cable operators have no "historically rooted expectation of compensation” for the right to use their cable services, see Loretto, 458 U.S. at 441, 102 S.Ct. at 3179, evidently because of (1) local governments' practices of imposing carriage requirements as a condition of granting franchises and right-of-way access, and (2) prior similar federal impositions, seemingly granted in exchange for privileged entitlements to retransmit broadcasters' signals. Both seem of limited relevance for the reasons discussed above in connection with the parallel Red Lion theories: the local governments are not involved here, and the federal government has removed the quid for which it now seeks a quo.