Opinion for the Court filed by Circuit Judge WILLIAMS.
Dissenting opinion filed by Circuit Judge SENTELLE.
STEPHEN F. WILLIAMS, Circuit Judge:Petitioner BellSouth Corporation challenges the constitutionality of Section 274 of the Telecommunications Act of 1996 (the “Act”), 47 U.S.C. § 274, and of the Federal Communications Commission’s order implementing that provision.1 Section 274 limits the ability of Bell operating companies (“BOCs”) to provide “electronic publishing,” a category that includes disseminating news articles, offering literary material, and providing services similar to the Lexis/Nexis and Westlaw databases. BellSouth says § 274 is an unconstitutional bill of attainder, stressing the fact that the subjects of its restrictions, the BOCs, are singled out by name. Bell-South also complains that § 274 impermissibly abridges its First Amendment rights of free expression. We reject both challenges.
* * *
The story behind the Telecommunications Act of 1996 has often been told, although electronic publishing restrictions have usually amounted to little more than a subplot. In 1982 a consent decree was entered in settlement of the government’s 1974 antitrust suit against AT&T. That decree, as modified by the district court, became known as the “Modification of Final Judgment,” or “MFJ.” See United States v. American Tel. & Tel. Co., 552 F.Supp. 131 (D.D.C.1982), aff'd sub nom. Maryland v. United States, 460 U.S. 1001, 103 S.Ct. 1240, 75 L.Ed.2d 472 (1983). The MFJ required AT&T to divest itself of its local exchange monopolies. Under the reorganization plan approved by the district court, the twenty BOCs eventually named in the 1996 Act were spun off from AT&T and grouped into seven regional Bell operating companies, or “RBOCs” (now five thanks to mergers), of which BellSouth is one.2
The MFJ initially prohibited the BOCs from providing “information services,” defined to include electronic publishing. The prohibition rested on two concerns commonly voiced about regulated monopolists operating in fields adjacent to their monopolies. First, to the extent that the monopolist’s good or service is an input for the adjacent industry, the monopolist may offer its own enterprise discriminatory advantages, in this case “favorable access to the local network.” 552 F.Supp. at 189. Second, the monopolist may use monopoly revenues to subsidize its associated enterprise. Id. In a “triennial review” process established by the decree, the Department of Justice moved to lift the information services restrictions, and no party to the decree opposed the motion. The district court ultimately did lift them. United States v. Western Electric Co., 767 F.Supp. 308 (D.D.C.1991), aff'd, 993 F.2d 1572 *61(D.C.Cir.1993). We will return later to the analysis supporting that result, which Bell-South says helps its constitutional case against § 274.
The 1996 Act rescinded the MFJ, see Pub.L. No. 104-104, § 601, 110 Stat. 143 (1996), and changed the entire telecommunications landscape. Several key provisions of the Act apply to incumbent local exchange carriers generally, such as 47 U.S.C. § 251, requiring them to offer nondiscriminatory access and interconnection to local competitors. Sections 271 through 276 of the Act, however, entitled “Special Provisions Concerning Bell Operating Companies,” are applicable to the BOCs and their affiliates alone.3 For example, § 271 establishes requirements that must be met before the BOCs can break into the long distance, or “interLATA,” market, see SBC Communications, Inc. v. FCC, 138 F.3d 410 (D.C.Cir.1998); § 273 bars the BOCs from manufacturing and selling telecommunications equipment until they have received authorization to enter the interLATA market; and § 275 prohibits BOCs (other than Ameriteeh) from providing alarm monitoring services for five years, see Alarm Industry Communications Committee v. FCC, 131 F.3d 1066, 1067 (D.C.Cir.1997). In general these provisions simply maintained, and in most cases loosened, various restrictions to which the BOCs were already subject under the MFJ. By contrast, the provision at issue here-§ 274-reimposed on the BOCs some of the information services restrictions that had been lifted in 1991. BellSouth challenges only that provision and the FCC order of implementation.4
Section 274 provides:
No Bell operating company or any affiliate may engage in the provision of electronic publishing that is disseminated by means of such Bell operating company’s or any of its affiliates’ basic telephone service, except that nothing in this section shall prohibit a separated affiliate or electronic publishing joint venture operated in accordance with this section from engaging in the provision of electronic publishing.
47 U.S.C. § 274(a). Section 274’s restrictions expire on February 8, 2000, four years from the date of the Act’s passage. § 274(g)(2).
As is evident from its text, § 274 provides two pathways for BOCs wishing to enter electronic publishing: the “separated affiliate” route and the “joint venture” route. The statute defines a separated affiliate as “a corporation under common ownership or control with a Bell operating company that does not own or control a Bell operating company and is not owned or controlled by a Bell operating company.” 47 U.S.C. § 274(i)(9). An “electronic publishing joint venture” is a “joint venture owned by a Bell operating company or affiliate that engages in the provision of electronic publishing which is disseminated by means of such Bell operating company’s or any of its affiliates’ basic telephone service.” 47 U.S.C. § 274(i)(5). Section 274 imposes several structural requirements on both separated affiliates and electronic publishing joint ven*62tures. See generally 47 U.S.C. § 274(b). For example, each such entity must maintain books, records, and accounts separately from the BOC with which it is affiliated, § 274(b)(1), may have “no officers, directors, and employees in common” with a BOC, § 274(b)(5)(A), and may “own no property in common,” § 274(b)(5)(B).
The Act defines “electronic publishing” broadly as
the dissemination, provision, publication, or sale to an unaffiliated entity or person, of any one or. more of the following: news (including sports); entertainment (other than interactive games); business, financial, legal, consumer, or credit materials; editorials, columns, or features; advertising; photos or images; archival or research material; legal notices or public records; scientific, educational, instructional, technical, professional, trade, or other literary materials; or other like or similar information.
47 U.S.C. § 274(h)(1). It then exempts several types of services, including data processing, voice messaging, and video programming. 47 U.S.C. § 274(h)(2).
BellSouth, of course, is not a BOC but an RBOC. Yet the government rightly refrains from raising a standing defense on that ground. The injury BellSouth suffered as the sole shareholder of two affected corporations (South Central Bell Telephone Company and Southern Bell Telephone and Telegraph Company) is clearly enough to give it Article III standing. See Franchise Tax Board of California v. Alcan Aluminium Ltd., 493 U.S. 331, 336, 110 S.Ct. 661, 663-664, 107 L.Ed.2d 696 (1990). Besides, as an affiliate of two BOCs, § 274(i)(1), BellSouth is itself affected; it can engage in electronic publishing only by maintaining structural separation from its BOCs.
Bill of Attainder Challenge
We turn first to BellSouth’s challenge under Article I, section 9, clause 3 of the Constitution, which says that “[n]o Bill of Attainder or ex post facto Law shall be passed” by Congress, For the framers of the Constitution the term “bills of attainder” carried a specific meaning: it referred to parliamentary acts sentencing named persons to death without the benefit of a judicial trial. As early as 1810, however, in Fletcher v. Peck, 10 U.S. (6 Cranch) 87, 3 L.Ed. 162 (1810), Chief Justice Marshall noted in dictum that the prohibition on bills of attainder ought to extend to legislation subjecting specified persons to penalties short of death—what the framers called “bills of pains and penalties.” Id. at 138, 3 L.Ed. 162; United States v. Brown, 381 U.S. 437, 447, 85 S.Ct. 1707, 1714, 14 L.Ed.2d 484 (1965). Later in the nineteenth century the Supreme Court confirmed that the legislative punishments foreclosed by the Bill of Attainder Clause include bills of pains and penalties. Cummings v. Missouri, 71 U.S. (4 Wall.) 277, 320, 323, 18 L.Ed. 356 (1866). Moreover, the Court has recognized that not all bills of attainder expressly name their targets; some simply describe them. Brown, 381 U.S. at 442, 85 S.Ct. at 1711-12. In sum, the Court has developed a potentially sweeping definition of forbidden attainders, holding that “legislative acts, no matter what their form, that apply either to named individuals or to easily ascertainable members of a group in such a way as to inflict punishment on them without a judicial trial áre bills of attainder prohibited by the Constitution.” United States v. Lovett, 328, U.S. 303, 315-16, 106 Ct.Cl. 856, 1079, 66 S.Ct. 1073, 90 L.Ed. 1252 (1946). The result is a prohibition triggered when a legislative act meets two tests—first, that it apply with specificity, and second, that it impose punishment.
Even classic attainders seem not only to have specified individuals but also classes-defined as the confederates of a named traitor, as in the case of the attainder against the Earl of Kildare and his associates during the reign of Henry VIII. See Cummings, 71 U.S. at 323-24, 18 L.ED. 356. But the Supreme Court watered down the specificity requirement a bit more when it invalidated two post-Civil War enactments targeting all persons who could not truthfully swear that they had been loyal to the Union during the war. See Cummings; Ex parte Garland, 71 U.S. (4 Wall.) 333,18 L.Ed. 366 (1866). And in Lovett it said generally that specificity is shown if the law applies to “easily ascertain*63able members of a group.” 328 U.S. at 315, 66 S.Ct. at 1079. Since virtually all legislation operates by identifying the characteristics of the class to be benefited or burdened, it is not clear that the specificity requirement retains any real bite. In „any event, it is obviously met here, since § 274’s requirements apply uniquely to the twenty BOCs identified by name in the Act.
We assume; as do the parties, that the Bill of Attainder Clause protects corporations as well as individuals. Although the Supreme Court has yet to address the question directly, it has suggested as much in dictum, see Plaut v. Spendthrift Farm, Inc., 514 U.S. 211, 239 n. 9, 115 S.Ct. 1447, 1463 n. 9, 131 L.Ed.2d 328 (1995) (indicating that Bill of Attainder Clause applies to laws that burden “a single individual or firm”), and comparable constitutional rights have been extended to legal “persons” taking the corporate form. See, e.g., Metropolitan Life Ins. Co. v. Ward, 470 U.S. 869, 881 n. 9, 105 S.Ct. 1676, 1683 n. 9, 84 L.Ed.2d 751 (1985) (equal protection); United States v. Martin Linen Supply Co., 430 U.S. 564, 97 S.Ct. 1349, 51 L.Ed.2d 642 (1977) (double jeopardy). The clause’s coverage clearly seems to include at least closely held corporations, where an attainder would fall on a narrowly circumscribed, easily identified group of flesh-and-blood people.5 Given the parties’ shared assumption, we will not explore the issue further.
At times BellSouth comes close to arguing that the specification requirement ought to be the end of the matter. On this view, the Bill of Attainder Clause bars Congress from singling out a specified class of persons for burdens of any kind, regardless of whether those burdens can be viewed as punishments in any ordinary sense of the term. This was the theme of a famous student essay, Note, The Bounds of Legislative Specification: A Suggested Approach to the Bill of Attainder Clause, 72 Yale L.J. 330 (1962), and traces of the same approach can be found in the Supreme Court’s most extensive discussion (and most expansive application) of the clause, United States v. Brown, 381 U.S. 437, 442, 85 S.Ct. 1707, 1711-12, 14. L.Ed.2d 484 (1965). For example, Brown said in a footnote that “a legislature can provide that persons possessing certain characteristics must abstain from certain activities, but must leave to other tribunals the task of deciding who possesses those characteristics.” Id. at 454 n. 29, 85 S.Ct. at 1718 n. 29 (emphasis added). And Brown’s explanation of the clause “as an implementation of the separation of powers, a general safeguard against legislative exencise of the judicial function, or more simply-trial by legislature,” id. at 442, 85 S.Ct. at 1711-12, certainly lent itself to sweeping application. Nonetheless, even Brown seemed at times to limit itself to punishments, saying, for example, that the clause “reflected the Framers’ belief that the Legislative Branch is not so well suited as politically independent judges and juries to the task of ruling upon the blameworthiness of, and levying appropriate punishment upon, specific persons.” Id. at 445, 85 S.Ct. at 1713. Such a limitation is practically indispensable: given the demise of the requirement that a forbidden attainder fall on named individuals, and the elusive character of. Brown’s own effort to articulate a coherent specificity test to replace that requirement, see 381 U.S. at 455 n. 29, 85 S.Ct. at 1718 n. 29, a definition of attainder that encompassed any burden imposed on specified individuals or groups would cut a broad swath, mowing down much of the Supreme Court’s equal protection jurisprudence at a single stroke.
In any event, whatever Brown’s potential for diluting the punishment requirement, the Supreme Court .has since taken that requirement seriously. It made this emphatically clear in Nixon v. Administrator of General Services, 433 U.S. 425, 97 S.Ct. 2777, 53 L.Ed.2d 867 (1977), where the law at issue burdened a single person.' Despite the statute’s surgical focus on a sole individual, the Court held that “the mere specificity of a law does not call into play the Bill of Attain*64der Clause,” id. at 471 n. 33, 97 S.Ct. at 2805 n. 33, and indeed that Congress had on that occasion singled out “a legitimate class of one,” id. at 472, 97 S.Ct. at 2805.6 It insisted that the burden must be a punishment to qualify as a bill of attainder, and considered three questions in determining whether it was. Id. at 478-84, 97 S.Ct. at 2808-13. Despite Chief Justice Burger’s suggestion that the Nixon precedent itself would become a “class of one,” id. at 544-45, 97 S.Ct. at 2841 (Burger, C.J., dissenting), the Court later formalized the punishment inquiry into a three-part test, asking
(1) whether the challenged statute falls within the historical meaning of legislative punishment; (2) whether the statute, viewed in terms of the type and severity of burdens imposed, reasonably can be said to further nonpunitive legislative purposes; and (3) whether the legislative record evinces a congressional intent to punish.
Selective Service System v. Minnesota Public Interest Research Group, 468 U.S. 841, 852, 104 S.Ct. 3348, 3355, 82 L.Ed.2d 632 (1984) (citations and internal quotation marks omitted). Unlike our dissenting colleague, Dissent at 3, 5, 6, we see no warrant in the precedents for treating Congress’s specification of the BOCs by name as a material element in the punishment analysis. Cf. Brown, 381 U.S. at 461, 85 S.Ct. at 1721 (describing supposed contrast between naming and mere specification as a “distinction! ] without a difference”). We take up each of the three factors in turn.
To begin with, § 274’s restrictions are nothing like the classic attainders known to the framers. As mentioned above, bills of attainder at common law generally entailed execution, although this was typically coupled with other punishments, such as “corruption of blood,” which prevented the attainted party’s heirs from inheriting his property. See Brown, 381 U.S. at 441, 85 S.Ct. at 1711. Bills of pains and penalties, also forbidden by the clause, “commonly imposed imprisonment, banishment, and the punitive confiscation of property.” Selective Service System, 468 U.S. at 852, 104 S.Ct. at 3355.
The case becomes closer when we move from historic antecedents to burdens later found by the Supreme Court to rank as punishments, which have included “legislative bars to participation by individuals or groups in specific employments or professions.” Id. at 852, 104 S.Ct. at 3355. Indeed, the Court’s four major decisions invalidating statutes on Bill of Attainder Clause grounds have all involved legislation preventing specific classes of persons from pursuing certain occupations. Those four cases came in two pairs. The first pair involved restrictions imposed immediately after the Civil War on those who had sided with the South, see Cummings (striking down amendments to Missouri constitution denying right to vote, hold office, teach, or serve as trustee for religious organization to persons who aided or sympathized with the Confederacy), and Garland (striking down federal law requiring attorneys to swear oath that they had never assisted Confederacy as condition of admission to practice in federal courts).7 The second pair involved restrictions on Communist Party members during the Cold War, see Lovett (invalidating law cutting off payment of salaries to three named federal employees who were Party members), and Broum (invalidating law making it a crime for members of Party to serve as officers or employees of labor unions).
Although a statute imposing structural separations on corporations seeking to engage in specific types of commercial activity may be analogous to such traditional employ*65ment debarments, the analogy is very loose indeed. Even if we ignore the BOCs’ freedom under § 274 to enter electronic publishing through structurally separated affiliates, the section is nothing more than a line-of-business restriction, comparable for example to the Glass-Steagall Act’s limitation on the entry of commercial banks into investment banking, 12 U.S.C. §§ 24 (Seventh), 78, or to the cross-ownership restrictions on broadcasters upheld in FCC v. National Citizens Committee for Broadcasting, 486 U.S. 775, 98 S.Ct. 2096, 56 L.Ed.2d 697 (1978). Although membership in the class of commercial banks or broadcasting firms is easy enough to ascertain, no one has suggested that those laws work an unconstitutional attainder. Indeed the Supreme Court in Brown strongly suggested that line-of-business restrictions pose no bill of attainder concerns, distinguishing the statute at issue there, which barred Communists from high office in labor unions, from § 32 of the Banking Act of 1933 (now codified at 12 U.S.C. § 78), a conflict-of-interest statute preventing employees of securities underwriting firms from working for banks that belong to the Federal Reserve System. Brown, 381 U.S. at 453-55, 85 S.Ct. at 1717-18; see also Board of Governors v. Agnew, 329 U.S. 441, 67 S.Ct. 411, 91 L.Ed. 408 (1947) (upholding § 32 of Banking Act, though without addressing bill of attainder issue). Brawn reasoned that each of the invalidated employment prohibitions “inflict[ed] its deprivation upon the members of a political group thought to present a threat to the national security,” 381 U.S. at 453, 85 S.Ct. at 1717, and further contrasted the conflict-of-interest statute as “ineorporat[ing] no judgment censuring or condemning any man or group of men.” Id. at 453-54, 85 S.Ct. at 1717. It is apparent—and will be more so when we examine legislative purpose—that § 274 falls on the nonpunitive side of those lines.
Placing § 274-among the burdens historically forbidden as attainders seems especially dubious because it does not bar the BOCs from electronic publishing but simply ■ requires structural separation. As counsel for BellSouth acknowledged at oral argument, the separated affiliate mechanism permits his client to establish a wholly-owned subsidiary to pursue electronic publishing. This subsidiary could disseminate materials over the telephone lines of BellSouth’s BOC subsidiaries, as long as it was kept separate from them in the ways prescribed by. § 274(b). Indeed, BellSouth itself could enter the electronic publishing business provided it observed the norms of separation from its BOCs. In short, § 274 leaves all the investors with stakes in the BOCs (i.e., the shareholders of the RBOCs) free to . pursue their collective electronic publishing ends, and to aggregate their capital to achieve those ends, subject only to structural separation requirements. While structural separation is hardly costless, neither does it remotely approach the disabilities that have traditionally marked forbidden attainders.
The second criterion asks whether. the challenged legislation, considering the type and severity of the burdens it imposes, can reasonably be said to further nonpunitive legislative purposes. This factor appears to be the most important of the three. See Siegel v. Lyng, 851 F.2d 412, 418 (D.C.Cir.1988) (“The line of Supreme Court law on the Bill of Attainder Clause indicates that legislation will survive Bill of Attainder attack if the statute furthers nonpunitive legislative purposes.”). On the one hand, where an enactment falls outside the historical definition of punishment, the second factor prevents Congress from circumventing the clause by cooking up newfangled ways to punish disfavored individuals or groups, Selective Service System, 468 U.S. at 853-54, 104 S.Ct. at 3355. On the other hand, “[e]ven measures historically associated with punishment-such as permanent exclusion from an occupation-have been otherwise regarded when the nonpunitive aims of an apparently prophylactic measure have seemed sufficiently clear and convincing.” Laurence H. Tribe, American Constitutional Law, § 10-5, at 655 (2d ed.1988).
In fact, apart .from its specific targeting aspect, we find that § 274 has the earmarks of a rather conventional response to commonly perceived, risks of anticompetitive behavior. We have long recognized that structural separation is “a permissible regulatory tool” for ensuring “that no cross-subsidization or *66unfair competitive practices occur.” Computer and Communications Industry Ass’n v. FCC, 693 F.2d 198, 219 (D.C.Cir.1982). We return shortly to the realism of those risks, but pause here to note that BellSouth’s claim of punitive purpose is somewhat undermined by § 274’s placement in an Act that as a whole relieves the BOCs of several of the burdens imposed by the MFJ, particularly by prescribing in § 271 a method whereby the BOCs can achieve a long-sought-after presence in the long-distance market.8
BellSouth advances two arguments in support of its claim that Congress cannot reasonably be said to have enacted § 274 for nonpunitive purposes. First, it says the court’s 1991 removal of the information services prohibition from the MFJ—based on a finding that its removal could reasonably be found to advance the public interest (balancing the risk of BOC discrimination against competing information services ventures with the competitive benefits of BOC entry)— shows that Congress in 1996 had no plausible economic basis for reimposing electronic publishing restrictions. Second, BellSouth points to other local exchange carriers who are not covered by § 274’s proscriptions. To the extent the BOCs pose any anticompetitive threat, says BellSouth, then so do the excluded firms, and their exclusion demonstrates that Congress’s real aim was to punish the BOCs.
As we said earlier, the information services ban was lifted from the MFJ at the behest of the Department of Justice, which had insisted on the ban when the MFJ was being negotiated. Circumstances had changed, the government argued; the information services market had become more competitive, and the BOCs’ ability to discriminate and cross-subsidize had consequently decreased. The district court initially rejected the government’s proposal, United States v. Western Electric Co., 673 F.Supp. 525, 587-97 (D.D.C.1987), but we reversed, saying that the court had used too stringent a standard to evaluate the government’s motion, which no party to the consent decree had opposed. United States v. Western Electric Co., 900 F.2d 283, 292 (D.C.Cir.1990). On remand the district court lifted the information services ban. United States v. Western Electric Co., 767 F.Supp. 308 (D.D.C.1991). We affirmed, noting that under the applicable “public interest” standard “the record before the court was such that any district court rejection of the proposed modification would have been reversible error.” United States v. Western Electric Co., 993 F.2d 1572, 1577-78 (D.C.Cir.1993). That record, we said, contained “persuasive evidence that, despite their local monopoly power, the BOCs will be unable to discriminate against competing information service providers.” Id. at 1579-80. We also concluded that there was powerful evidence to counter any suggestion that the BOCs would be able to use their price-regulated monopolies to subsidize their entry into information services. Id. at 1580-81.
Obviously Congress’s reading of the evidence in 1996 was different from the one arrived at by the Department of Justice in 1987—or by this court in 1993 for that matter. It does .not follow from these conflicts between branches, however, that Congress cannot rationally be said to have pursued nonpunitive purposes in enacting § 274. Certainly our triennial review decisions never suggested that the risks of anticompetitive conduct were so feeble that no one could reasonably assert them except as a smokescreen for some invidious purpose (much less for the specific invidious purpose of “punishing” the BOCs). And we note that § 274 is less severe than the analogous pre-1991 MFJ provision along several dimensions: it applies only to electronic publishing rather than to information services as a whole, it expires after five years rather than continuing indefinitely, and it mandates structural separation, rather than complete exclusion.
BellSouth complains that this reading of the second factor reduces it to little more than a rational basis test, the most anemic form of constitutional scrutiny. If this were strictly true, of course, the Bill of Attainder Clause (as applied to non-suspect classes *67such as the BOCs) would do nothing more than duplicate the Equal Protection Clause. But the Supreme Court’s attainder inquiry is in fact more exacting than , a rational basis test, because it demands purposes that are not merely reasonable but nonpunitive. Punitive - purposes, however rational, don’t count.
BellSouth’s second argument focuses on the fact that § 271 does not cover several large non-BOC local exchange carriers, in particular GTE Corporation.9 GTE, which was never part of the AT&T system, supplies about 18.4 million access lines in 27 states; by comparison, BellSouth supplies about 24.5 million access lines in nine states. ‘ 1996 FCC Statistics of Communications Common Carriers 21 (1997). Although its operations are generally rural, as of 1993 “GTE controlled] local exchange service in the entire state of Hawaii as well as in large portions of the Tampa and Los Angeles markets.” United States v. Western Electric Co., 993 F.2d at 1579. Other non-BOC carriers, such as Southern New England Telephone with about 2.1 million access lines in Connecticut, are also not covered by § 274. 1996 FCC Statistics of Communications Common' Carriers 21 (1997). From this selectivity—which BellSouth labels under-inclusiveness—Bell-South would have us draw an inference of punitive purpose.
But the differential treatment of the BOCs and non-BOCs is neither suggestive of punitive purpose nor particularly suspicious. Because the BOCs’ facilities are generally less dispersed than GTE’s, they can exercise bottleneck control over both ends of a telephone call in a higher fraction of cases than can GTE. The BOCs thus enjoy a materially greater opportunity to shift costs from their electronic publishing pursuits to their rate-regulated local exchange ventures.10 In addition, because GTE (unlike the BOCs) is not the dominant provider of local exchange service in any state except Hawaii, state regulators can use the costs of its local competitors as benchmarks against which to measure whether it is engaging in improper cost allocation. Thus the distinction drawn by Congress seems quite understandable without resort to inferences of punitive purpose.
The third device for identifying a punishment focuses on legislative intent, and in practice appears to differ from the second only in inviting a journey through legislative history. On this point we can be brief. Bell-South simply has not come forward with the kind of “unmistakable evidence of punitive intent which ... is required before a Congressional enactment of this kind may be struck down” as an attainder. Selective Service System, 468 U.S. at 855-56 n. 15, 104 S.Ct. at 3357 n. 15 (quoting Flemming v. Nestor, 363 U.S. 603, 619, 80 S.Ct. 1367, 1377, 4 L.Ed.2d 1435 (I960)). Aside from a few scattered remarks referring to anticompetitive abuses allegedly committed by the BOCs in the past, BellSouth has provided no legislative history even touching on the purposes behind § 274, much less presenting “smoking gun” evidence of congressional vindictiveness.
In sum, we hold that § 274 is not a bill of attainder.
First Amendment Challenge
BellSouth complains that § 274 abridges its constitutional right of free speech by restricting its ability to provide electronic publishing. Clearly the structural separation requirements regulate expressive activity within the scope of the First Amendment. So, as is often the ease in the First Amendment arena, the parties devote much *68of their energy toward disputing the appropriate standard of review. BellSouth argues that § 274 warrants strict scrutiny for two reasons: first, because it singles out named corporations for speech restrictions, and second, because it is content-based. The FCC says that § 274 is a content-neutral regulation and should instead be evaluated under the intermediate standard of review applied by the Supreme Court in its decisions upholding the cable television “must-carry” rules. See Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 114 S.Ct. 2445, 129 L.Ed.2d 497 (1994) (“Turner F) (determining that intermediate scrutiny applied); Turner Broadcasting System, Inc. v. FCC, 520 U.S. 180, -, 117 S.Ct. 1174, 1186, 137 L.Ed.2d 369 (1997) (“Turner IF) (upholding must-carry provisions under intermediate scrutiny). We agree with the FCC that an intermediate level of scrutiny is appropriate.
We begin with BellSouth’s claim that § 274 warrants strict First Amendment review because it targets named corporations. In support of this claim BellSouth cites our decision in News America Publishing, Inc. v. FCC, 844 F.2d 800 (D.C.Cir.1988). That case concerned a provision that prohibited the FCC from granting extensions of temporary waivers of the newspaper/television cross-ownership rules to all then-current holders of such temporary waivers—a class of one, as it turned out. We struck down the provision on First Amendment and equal protection grounds, noting that “[w]here legislation affecting speech appears underinclusive, i.e., where it singles out some conduct for adverse treatment, and leaves untouched conduct that seems indistinguishable in terms of the law’s ostensible purpose, the omission is bound to raise a suspicion that the law’s true target is the message.” Id. at 804-05. Contrary to BellSouth’s assertion, however, News America does not stand for the proposition that statutes singling out particular persons for speech restrictions automatically merit strict scrutiny. In fact we chose among standards only to the extent of saying that the provision could not survive any standard of review more exacting than a rational basis test. Id. at 802, 814.
Only marginally more promising for Bell-South are the Supreme Court’s decisions in Minneapolis Star & Tribune Co. v. Minnesota Commissioner of Revenue, 460 U.S. 575, 103 S.Ct. 1365, 75 L.Ed.2d 295 (1983), and Arkansas Writers’ Project v. Ragland, 481 U.S. 221, 107 S.Ct. 1722, 95 L.Ed.2d 209 (1987). In both cases the Court applied strict scrutiny to invalidate state tax laws that had the effect of disproportionately burdening certain segments of the press. In Minneapolis Star, because the challenged use tax exempted the first $100,000 in ink and paper consumed by a publication in a given year, its bxxrden was borne by fewer than one in twenty newspapers in the state, with two-thirds of it landing on one publisher alone. 460 U.S. at 578-79, 103 S.Ct. at 1368-69. The sales tax struck down in Arkansas Writers’ Project appeared to spread its net more broadly, covering general interest magazines but exempting newspapers and religious, professional, trade, and sports magazines. 481 U.S. at 224, 107 S.Ct, at 1725. In fact, however, “the magazine exemption mean[t] that only a few Arkansas magazines pa[id] any sales tax,” giving the tax a targeting effect comparable to that of the Minneapolis Star tax. Id. at 229, 107 S.Ct. at 1727-28. The Supreme Court has explained the two cases as meaning only that strict scrutiny must be applied to regulations that target a small subset of media organizations in ways that threaten to “distort the market for ideas.” Turner I, 512 U.S, at 660, 114 S.Ct. at 2468 (quoting Leathers v. Medlock, 499 U.S. 439, 448, 111 S.Ct. 1438, 1444, 113 L.Ed.2d 494 (1991)). The Court has expressly declined to draw the broad lesson “that the First Amendment mandates strict scrutiny for any speech regulation that applies to one medium (or a subset thereof) but not others.” Turner I, 512 U.S. at 660, 114 S.Ct. at 2468.11
News America, Minneapolis Star, and Arkansas Writers’ Project all featured some suggestion that the legislature’s differ*69ential treatment of speakers was motivated by the content of their speech. See Turner I, 512 U.S. at 660, 114 S.Ct. at 2468 (“Although there was no evidence that an illicit governmental motive was behind either of the taxes, both were structured in a manner that raised suspicions that their objective was, in fact, the suppression of certain ideas.”).12 It is that suggestion, rather than the act of “singling out” by itself, that triggers strict First Amendment scrutiny, as Turner I made clear. See id. at 658, 114 S.Ct. at 2467 (“[S]peaker-based laws demand strict scrutiny when they reflect the Government’s preference for the substance of what the favored speakers have to say (or aversion to what the disfavored speakers have to say).”).
Here, there is no indication that § 274’s coverage was limited to the BOCs because of any concern about the content of their speech—no indication, in other words, that “the legislature’s speaker preference reflects a content preference.” Id. “So long as they are not a subtle means of exercising a content preference, speaker distinctions of this nature are not presumed invalid under the First Amendment.” Id. at 645, 114 S.Ct. at 2460-61. In addition, Turner I held that “heightened scrutiny is unwarranted when the differential treatment is ‘justified by some special characteristic of the particular medium being regulated.” Id. at 660-61, 114 S.Ct. at 2468 (quoting Minneapolis Star, 460 U.S. at 575, 103 S.Ct. at 1366-67). Congress’s imposition of structural separation on the BOCs because of their status as price-regulated bottleneck monopolies is certainly no more suggestive of any effort to exercise a content preference than were the must-carry provisions upheld in Turner, see Turner I, 512 U.S. at 661, 114 S.Ct. at 2468 (identifying “bottleneck monopoly power” held by cable operators as a special characteristic of the cable medium).
We turn next to BellSouth’s claim that § 274 is expressly formulated in terms of content, and thus requires strict scrutiny. To be sure, § 274 defines the field of expression to which it applies by reference to a set of categories that might in a formal sense be described as content-based. Thus it covers items such as “news,” “entertainment,” and “research material,” and exempts information such as “video programming,” “voice messaging,” and “data processing.” See § 274(h)(1).
Nothing about the provision, however, suggests an underlying purpose to favor or disfavor particular viewpoints, nor does Bell-South advance such a suggestion. The Supreme Court has held that statutes lacking such a purpose are likely to be deemed content-neutral. “[L]aws that confer benefits or impose burdens on speech without reference to the ideas or views expressed are in most instances content-neutral.” Turner I, 512 U.S. at 643, 114 S.Ct. at 2459. Indeed the very breadth of § 274’s categories undermines any claim that Congress adopted a categorical approach out of a desire to favor any particular viewpoint or idea. Further, to a large extent neutrality is now gauged by reference to a statute’s justifications: “Government regulation of expressive activity is content neutral so long as it’is justified without reference to the content of the regulated speech.” Ward v. Rock Against Racism, 491 U.S. 781, 791, 109 S.Ct. 2746, 2753, 105 L.Ed.2d 661 (1989) (emphasis in original) (citation omitted); The goal of remedying bottleneck problems is independent of content and viewpoint.
In summary,; then—despite BellSouth’s twin contentions that § 274 favors certain speakers, and certain types of speech, over others—we hold that intermediate scrutiny is appropriate because here, perhaps even more than in Turner, there is simply no hint that “the government has adopted a regulation of speech because of [agreement or] disagreement with the message it conveys.” Turner I, 512 U.S. at 642, 114 S.Ct. at 2459 (quoting Ward, 491 U.S. at 791, 109 S.Ct. at 2754).
A regulation will be upheld under intermediate scrutiny “if it advances important governmental interests unrelated to the *70suppression of free speech and does not burden substantially more speech than necessary to further those interests.” Turner II, 520 U.S. at-, 117 S.Ct. at 1186 (citing United States v. O’Brien, 391 U.S. 367, 377, 88 S.Ct. 1673, 1679, 20 L.Ed.2d 672 (1968)). The requirement of an important governmental interest is amply met here. The asserted interest underlying § 274 is to promote competition by discouraging discrimination and cross-subsidization by the BOCs. This is not only important, see Turner I, 512 U.S. at 644, 114 S.Ct. at 2460, but “unrelated to the suppression of free speech”; indeed, the interest in preventing truly anticompetitive behavior in the electronic publishing marketplace is an interest in the enhancement of speech. Cf. id. at 663, 114 S.Ct. at 2469-70.
Under intermediate scrutiny, while the government obviously need not meet the most rigorous standard of “narrow tailoring,” it must nonetheless “demonstrate that the recited harms are real, not merely conjectural, and that the regulation will in fact alleviate these harms. in a direct and material way.”. Id. at 664, 114 S.Ct. at 2470. In determining whether the government has made this showing, we owe Congress’s economic judgments considerable deference, so as not to “infringe on traditional legislative authority to make predictive judgments when enacting nationwide regulatory policy.” Turner II, 520 U.S. at-, 117 S.Ct. at 1189; see also id. at-, 117 S.Ct. at 1203-03 (Breyer, J., concurring) (deferring to Congress’s non-economie objectives). Aside from noting the various executive branch (and judicial) positions taken in connection with removal of the MFJ’s restriction, which bear no more weight here than in the bill of attainder analysis, see pp. 64-65 above, BellSouth does not claim that Congress’s apparent concern about anticompetitive risks was unreasonable.
But it does' complain that Congress could have guarded against these risks through less speech-restrictive methods, for instance by imposing non-structural safeguards such as accounting requirements. Intermediate First Amendment scrutiny, however, does not entail a “least restrictive means” analysis. See Turner II, 520 U.S. at ---, 117 S.Ct. at 1199-1200. It is at least plausible that structural separation will more effectively meet the perceived anticompetitive threat than would lesser restrictions, and although BellSouth characterizes the additional burden of structural separation as “enormous” (compared with, for example, special accounting rules), it offers neither detail nor quantitative evidence to support this characterization.
As in its bill of attainder attack, BellSouth points to the exclusion of GTE and other non-BOC local exchange carriers. But as we said in that connection, there are plausible reasons for the exclusion, and, just as BellSouth failed in that context to suggest a punitive purpose, here it equally fails to suggest any intent to favor non-BOCs’ viewpoints over BOCs’. Moreover, since intermediate scrutiny demands that the government “not burden substantially more speech than necessaiy to further [its] interests,” Turner II, 520 U.S. at -, 117 S.Ct. at 1186, it would be odd to strike down, a statute because Congress failed to restrict as much expression as it could have—presumably because of a judgment that the interest justifying a restriction in one context, though “important,” was slightly less so than in the other and therefore, at the margin, outweighed by competing interests (including free speech). See Walsh v. Brady, 927 F.2d 1229, 1238-39 (D.C.Cir.1991) (Williams, J., concurring).
Finally, we note again that § 274 leaves each RBOC free to publish electronically, using the facilities of its BOC subsidiaries, either directly or through a subsidiary, so long as the acting corporation conforms to the statutory separation requirements. Bell-South argues that the separated affiliate mechanism is entirely irrelevant to the First Amendment question, since “[i]t hardly answers one person’s objection to a restriction on his speech that another person, outside his control, may speak for him.” Arkansas Writers’ Project, 481 U.S. at 231, 107 S.Ct. at 1728 (internal quotations omitted). While undoubtedly true in the context of natural persons, this observation carries less weight in the context of controlled subsidiaries. The First Amendment does not normally permit *71the government to justify a prohibition on a corporation’s speech by pointing to the fact that its shareholders remain free to express themselves without restriction; corporations exist in significant part to overcome precisely the sorts of collective action problems that the shareholders in that scenario would then face. But there is no collective action problem here. The investors in BellSouth are free to use whatever portion of their pooled' resources they wish for electronic publishing, subject only to the need for structural separation. This is not to say that structural separation requirements count as merely in-, substantial burdens from a First Amendment perspective—they do not. Cf., e.g., FEC v. Massachusetts Citizens for Life, 479 U.S. 238, 252-54, 107 S.Ct. 616, 624-26, 93 L.Ed.2d 539 (1986) (noting disclosure and, record-keeping requirements entailed by corporation’s use of segregated political contribution fund). But the fact that § 274 leaves RBOCs like BellSouth free to pursue electronic publishing strengthens our conclusion that § 274 does. not restrict substantially more speech than necessary.
* * *
The petition for review is
Denied.
. The order under challenge is Implementation of the Telecommunications Act of 1996: Telemessaging, Electronic Publishing, and Alarm Monitoring Services, FCC No. 97-35 (Feb. 7, 1997). BellSouth's challenge to the order is entirely derivative of its constitutional challenge to the statute, with no claim that the FCC acted outside the scope of its statutory authority.
. AT&T also divested its minority holdings in the Cincinnati Bell Telephone Company and the Southern New England Telephone Company, which are not classified as BOCs in the Act.
. The Act defines "Bell operating company” as follows:
The term "Bell operating company"— •
(A)means any of the following companies: Bell Telephone Company of Nevada, Illinois Bell Telephone Company, Indiana Bell Telephone Company, Incorporated, Michigan Bell Telephone Company, New England Telephone and Telegraph Company, New Jersey Bell Telephone Company, New York Telephone Company, U.S. West Communications Company, South Central Bell Telephone Company, Southern Bell Telephone and Telegraph Company, Southwestern Bell Telephone Company, The Bell Telephone Company of Pennsylvania; The Chesapeake and Potomac Telephone Company, The Chesapeake and Potomac Telephone Company of Maryland, The Chesapeake and Potomac Telephone Company of Virginia, The Chesapeake and Potomac Telephone Company of West Virginia, The Diamond State Telephone Company, The Ohio Bell Telephone Company, The Pacific Telephone and Telegraph Company, or Wisconsin Telephone Company; and
(B) includes any successor or assign of any such company that provides wireline telephone exchange service; but
(C) does not include an affiliate of any such company, other than an affiliate described in subparagraph (A) or (B).
47 U.S.C. § 153(4).
. We note that another RBOC has launched a Bill of Attainder Clause and First Amendment challenge to the "Special Provisions” as a whole. See SBC Communications, Inc. v. FCC, 981 F.Supp. 996 (N.D.Tex.1997).
. At least it would do so if it took the form of a conventional penalty such as a fine. If it look the form of a restriction barring certain closely held corporations from specific lines of business, its effect on flesh-and-blood people would depend on the language of the restriction and on the ability of officers, directors and shareholders to carry on their pursuits outside the named corporations. See below at pp. 65-66.
. See also Plaut v. Spendthrift Farm, Inc., 514 U.S. 211, 239 n. 9, 115 S.Ct. 1447, 1463 n. 9, 131 L.Ed.2d 328 (1995) (”[L]aws that impose a duty or liability upon a single individual or firm are not on that account invalid—or else we would not have the extensive jurisprudence that we do concerning the Bill of Attainder Clause, including cases which say that it requires not merely 'singling out' but also punishment, see, e.g., United States v. Lovett, 328 U.S. 303, 315-18 [106 Ct.Cl. 856, 66 S.Ct. 1073, 1078-80, 90 L.Ed. 1252] (1946), and a case which says that Congress may legislate 'a legitimate class of one,’ Nixon v. Administrator of General Services, 433 U.S. 425, 472 [97 S.Ct. 2777, 2805, 53 L.Ed.2d 867] (1977).”)
. See also Pierce v. Carskadon, 83 U.S. (16 Wall.) 234, 21 L.Ed. 276 (1872) (memorandum opinion striking down West Virginia loyalty oath similar to those invalidated in Cummings and Garland).
. Because we find no attainder, we need not wrestle with the issue of remedy. Severability is largely a matter of legislative intent, and it is doubtful that Congress would have intended the many provisions of the Act beneficial to the BOCs to survive deletion of this burdensome one.
. Like the BOCs, GTE was subject to a consent decree for more than a decade, until the- passage of the Act. The decree, however, permitted GTE to provide information services, subject to structural separation and non-discrimination requirements. See United States v. GTE Corp., 603 F.Supp. 730, 742 (D.D.C.1984).
. While in 1993 we somewhat disparaged the distinction then drawn between GTE and the BOCs on the basis of relative dispersion, United States v. Western Electric Co., 993 F.2d at 1579, we did so solely with respect to the claim of discrimination against competing providers. Our reasoning was that' the BOCs could not easily sort out information services transmissions, or intra-corporate transmissions, on the customer end of a call, as they would have to do in order to discriminate efficiently. Id. That deficiency would not seem to impede cost-shifting, where the regulated monopoly attempts to deceive regulators about which costs belong in the regulated enterprise and which in the other.
. In a similar vein, we have described Minneapolis Star and Arkansas Writers' Project as “likely addressed only to the special complexities of taxation.” Walsh v. Brady, 927 F.2d 1229, 1236 (D.C.Cir.1991).
. In fact, the tax exemption challenged in Arkansas Writers' Project facially excluded certain publications on the basis of their content, rendering it especially suspect. 481 U.S. at 229-30, 107 S.Ct. at 1727-28.