New York State Telecommunications Association, Inc. v. James

101 F.4th 135

Opinion

21-1975-cv New York State Telecommunications Association, Inc. v. James

United States Court of Appeals For the Second Circuit

August Term 2022

Argued: January 12, 2023

Decided: April 26, 2024

No. 21-1975

NEW YORK STATE TELECOMMUNICATIONS ASSOCIATION, INC., CTIA - THE WIRELESS ASSOCIATION, ACA CONNECTS - AMERICA’S COMMUNICATIONS ASSOCIATION, USTELECOM - THE BROADBAND ASSOCIATION, NTCA - THE RURAL BROADBAND ASSOCIATION, SATELLITE BROADCASTING AND COMMUNICATIONS ASSOCIATION, ON BEHALF OF THEIR RESPECTIVE MEMBERS,

Plaintiffs-Appellees,

v.

LETITIA A. JAMES, IN HER OFFICIAL CAPACITY AS ATTORNEY GENERAL OF NEW YORK, Defendant-Appellant.

Appeal from the United States District Court for the Eastern District of New York No. 21-cv-2389, Denis R. Hurley, Judge.

Before: SULLIVAN, NATHAN, and MERRIAM, Circuit Judges.

In 2021, New York enacted the Affordable Broadband Act (ABA), which requires internet service providers to offer broadband internet to qualifying households at reduced prices. A group of trade organizations representing internet service providers sued, arguing that the ABA was impliedly preempted by federal law. The district court agreed with the Plaintiffs’ preemption theories and granted a preliminary injunction barring New York from enforcing the ABA. The parties later requested that the district court enter a stipulated final judgment and permanent injunction.

Although the parties stipulated to judgment, we have appellate jurisdiction because the district court plainly rejected the legal basis for New York’s preemption defenses, all claims have been disposed of with prejudice, the stipulation was designed solely to obtain immediate appellate review and does not circumvent restrictions on our appellate jurisdiction, and New York expressly preserved the right to appeal.

Turning to the merits, we conclude that neither of the Plaintiffs’ preemption theories is availing. First, the ABA is not field-preempted by the Communications Act of 1934 (as amended by the Telecommunications Act of 1996), because the Act does not establish a framework of rate regulation that is sufficiently comprehensive to imply that Congress intended to exclude the states from entering the field. Second, the ABA is not conflict-preempted by the Federal Communications Commission’s 2018 order classifying broadband as an information service. That order stripped the agency of its authority to regulate the rates charged for broadband internet, and a federal agency cannot exclude states from regulating in an area where the agency itself lacks regulatory authority. Accordingly, we REVERSE the judgment of the district court and VACATE the permanent injunction. Judge Sullivan dissents in a separate opinion.

________

JUDITH N. VALE (Barbara D. Underwood, Steven C. Wu, Eric Del Pozo, on the brief) for Letitia James, Attorney General, State of New York, New York, NY, for Appellant.

SCOTT H. ANGSTREICH, Kellogg, Hansen, Todd, Figel & Frederick, P.L.L.C. (Andrew

3

E. Goldsmith, Joseph S. Hall, Alex A.

Parkinson, Kellogg, Hansen, Todd, Figel & Frederick, P.L.L.C., Jeffrey A. Lamken, MoloLamken LLP, Jared P. Marx, Harris, Wiltshire & Grannis, LLP, on the brief), Washington DC, for Appellees.

________

NATHAN, Circuit Judge:

In April 2021, New York enacted the Affordable Broadband Act (ABA), which aims to expand internet access by requiring internet service providers to offer broadband internet to low-income New Yorkers at reduced prices. The Plaintiffs, a group of trade organizations representing internet service providers, maintain that the ABA is impliedly preempted by federal law. We conclude that it is not.

As a threshold matter, we conclude that we have jurisdiction to hear this appeal. Although the parties stipulated to the judgment from which New York appeals, they did so under specific conditions that our case law recognizes as preserving appellate jurisdiction. The district court effectively resolved the Plaintiffs’ preemption claim as a matter of law, by rejecting the legal basis of New York’s preemption defenses; all claims have been disposed of with finality and with prejudice; the parties stipulated to judgment solely to obtain immediate

4

appellate review, without circumventing any restrictions on our appellate jurisdiction; and New York expressly preserved its right to appeal from the stipulated judgment. The parties have not circumvented the final judgment rule but have merely accelerated the process of obtaining the final judgment that became inevitable once the district court reached its legal conclusion.

Turning to the merits, we conclude as follows. First, the Communications Act of 1934 (as amended by the Telecommunications Act of 1996) does not wholly preempt states from regulating the rates charged for interstate communications services, because the Act does not establish a framework of rate regulation that is sufficiently comprehensive to imply that Congress intended to exclude the states from entering this field. Second, the ABA is not conflict-preempted by the Federal Communications Commission’s 2018 order classifying broadband as an information service. That order stripped the agency of its statutory authority to regulate the rates charged for broadband internet, and a federal agency cannot exclude states from regulating in an area where the agency itself lacks regulatory

5

authority. Accordingly, we REVERSE the judgment of the district court and VACATE the order permanently enjoining enforcement of the ABA.

BACKGROUND

I. Legal Background The Communications Act of 1934, 47 U.S.C. § 151 et seq., created the Federal Communications Commission (FCC) and authorized it to regulate all “interstate and foreign communication by wire or radio” and “all persons engaged within the United States in such communication.” Id. § 152(a). Under the Communications Act, communications services are subject to different regulatory regimes depending on how they are classified. For example, radio and mobile phone services are regulated under Title III of the Act, and cable television services are regulated under Title VI. The FCC has the authority to determine the appropriate statutory category for a particular communications service, and its determinations are entitled to deference under Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984). See Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980–81 (2005).

6

Broadband internet has, at different times, alternately been categorized by the FCC as a “telecommunications service” under Title II of the Communications Act, and as an “information service” under Title I. These designations are mutually exclusive, and they come with important regulatory consequences. If broadband is a Title II telecommunications service, then internet service providers (ISPs) are common carriers subject to a variety of statutory obligations and restrictions. For example, common carriers are barred from levying unreasonable charges, 47 U.S.C. § 201(b), or unjustly discriminating in the provision of services, id. § 202(a). Title II also contains a provision that permits the FCC to “forbear from applying any regulation or any provision of” the Act if it determines that the regulation is unnecessary. Id. § 160(a). Once the FCC chooses to exercise this forbearance authority, state and local regulators are preempted and “may not continue to apply or enforce” the relevant regulation. Id. § 160(e). On the other hand, if the FCC designates broadband as a Title I information service, then it is “exempted from common carriage status” under the Act. Mozilla Corp. v. FCC, 940 F.3d 1, 17 (D.C. Cir. 2019). Courts have accordingly held that the FCC lacks the

7

power to impose common carrier obligations on ISPs under Title I. See Comcast Corp. v. FCC, 600 F.3d 642, 655 (D.C. Cir. 2010) (rejecting notion that the FCC’s Title I authority allows it to impose rate regulations on ISPs); Verizon v. FCC, 740 F.3d 623, 655–59 (D.C. Cir. 2014) (concluding that the FCC lacked the statutory authority under Title I to impose net neutrality regulations).

The FCC has reclassified broadband internet on several occasions and did so most recently in 2018. See In re Restoring Internet Freedom, 33 FCC Rcd. 311 (2018). This 2018 Order reclassified broadband internet as a Title I information service and eliminated the FCC’s net neutrality regulations 1 as part of a broader agenda to “end utility-style regulation of the Internet in favor of . . . market-based policies” and adopt a “light-touch” regulatory framework. Id. ¶¶ 2, 207. The 2018 Order also contained a Preemption Directive, which purported to expressly preempt all state or local regulations of ISPs that would “interfere with the federal

1 Net neutrality refers to the principle that ISPs should “treat all Internet traffic the same regardless of source.” Verizon, 740 F.3d at 628. Net neutrality regulations “limit the ability of Internet service providers to interfere with the applications, content, and services on their networks [and] allow users to decide how they want to use the Internet without interference from Internet service providers.” Barbara van Schewick, Network Neutrality and Quality of Service: What a Nondiscrimination Rule Should Look Like, 67 Stan. L. Rev. 1, 4 (2015).

8

deregulatory policy restored in this order.” Id. ¶¶ 194–204. The stated goal was to prevent states and municipalities from implementing the “utility-type” common-carrier regulations that the federal government was eliminating. Id. ¶ 195.

As will be discussed extensively below, the D.C. Circuit considered the legality of the FCC’s reclassification of broadband as a Title I service and the FCC’s authority to issue the Preemption Directive. See Mozilla, 940 F.3d at 18 (D.C. Cir. 2019). In Mozilla, the D.C. Circuit upheld the FCC’s reclassification of broadband as a Title I service. However, the court vacated the Preemption Directive because it was not grounded “in a lawful source of statutory authority.” Id. at 74. Because the FCC chose to reclassify broadband as a Title I service, the court concluded that the FCC could not rely on its Title II forbearance authority to preempt state regulation over broadband internet. II. Factual Background In 2021, the New York State Legislature enacted the Affordable Broadband Act, which aims to provide internet access to the families least able to afford it. In legislative memoranda, the ABA’s sponsors explained that the circumstances of

9

the COVID-19 pandemic had “made it abundantly clear” that broadband internet was “an essential service in its own right.” Joint App’x 100. Legislators noted that internet access had become a de facto requirement for accessing health care, education, and work opportunities. Id. at 101. But despite its indispensable role in contemporary society, reliable internet access remained out of reach for many. The New York State Comptroller cited data from the most recent Census estimate, which found that “more than 1 million, or 13.8 percent of, New York households do not have subscriptions to broadband internet,” and “[o]ne in three low-income households lacks access.” Office of the N.Y.S. Comptroller, Availability, Access, and Affordability: Understanding Broadband Challenges in New York State 1 (2021). The Comptroller report concluded that “these access disparities disproportionately impacted low-income households during the pandemic and may generally present a disadvantage for these New Yorkers and their communities.” Id.

In an effort to address this digital divide, the ABA requires anyone “providing or seeking to provide . . . broadband service in New York state” to “offer high speed broadband service to low-income consumers” at statutorily

10

fixed prices. See 2021 N.Y. Sess. Laws 202–04 (McKinney) (codified at N.Y. Gen. Bus. Law § 399-zzzzz). ISPs must offer one of two broadband plans to all low- income consumers who qualify for certain means-tested governmental benefits. N.Y. Gen. Bus. Law § 399-zzzzz(2). Qualifying consumers must be offered broadband at no more than $15 per month for service of 25 Mbps, or $20 per month for high-speed service of 200 Mbps. Id. §§ 399-zzzzz(2)–(4). This requirement, however, is not absolute. Certain price increases may be allowed every few years, and ISPs that serve 20,000 households or fewer may be exempted if the New York Public Service Commission “determines that compliance with such requirements would result in unreasonable or unsustainable financial impact on the broadband service provider.” Id. §§ 399-zzzzz(3)–(5).

Soon after the ABA’s passage, the Plaintiffs filed suit against the New York State Attorney General, seeking injunctive relief and a declaratory judgment that federal law preempts the ABA and that enforcement of the ABA would violate the Supremacy Clause and the Plaintiffs’ rights under 42 U.S.C. § 1983. The Plaintiffs then moved for a preliminary injunction.

11

In June 2021, the district court granted the Plaintiffs’ motion and preliminarily enjoined enforcement of the ABA. Joint App’x 155. The court concluded that the ABA “triggers field preemption” because it "regulates within the field of interstate communications,” and separately held that “the ABA conflicts with the implied preemptive effect of . . . the FCC’s 2018 Order.” N.Y. State Telecomms. Ass’n v. James, 544 F. Supp. 3d 269, 282, 285 (E.D.N.Y. 2021).

Because a grant of a preliminary injunction is immediately appealable as of right, see 28 U.S.C. § 1292(a)(1), New York initially filed an interlocutory appeal from this order. However, because the district court had reached a legal conclusion that appeared to resolve all of the parties’ claims, the parties later jointly requested that the district court enter a stipulated final judgment and permanent injunction based on the court’s reasoning in its preliminary injunction decision. The district court agreed. It therefore permanently enjoined enforcement of the ABA and entered the parties’ stipulated final judgment, which dismissed the Plaintiffs’ § 1983 claim without prejudice and provided that “[d]efendant reserves the right to appeal this stipulated final judgment,

12

declaration, and permanent injunction.” Joint App’x 156–59. After the stipulated final judgment was entered, the parties jointly moved to withdraw the appeal of the preliminary injunction, and this appeal followed.

DISCUSSION

I. Appellate Jurisdiction Before turning to the merits, we first address whether we have jurisdiction to decide this appeal. Following oral argument, we issued an order directing the parties to submit supplemental briefing addressing whether New York’s stipulation to the entry of judgment deprived us of appellate jurisdiction. All parties maintain that we have appellate jurisdiction. We agree.

The fact that the parties stipulated to judgment does not deprive us of jurisdiction. In general, we lack appellate jurisdiction to review appeals from consent judgments. See LaForest v. Honeywell Int’l Inc., 569 F.3d 69, 73 (2d Cir. 2009) (“Appeal from a consent judgment is generally unavailable on the ground that the parties are deemed to have waived any objections to matters within the scope of the judgment.” (citation omitted)). However, in accordance with nearly all other

13

circuits to have considered the question, 2 we have held that we may nevertheless exercise appellate jurisdiction over claims resolved by a consent judgment when certain factors are met. Our cases have identified four such factors. First, the district court must have “plainly rejected the legal basis” for the appellant’s claim or defense. Ali v. Fed. Ins. Co., 719 F.3d 83, 94 (2d Cir. 2013). 3 Second, all claims must be disposed of with prejudice. Id. Third, the appellant’s consent to final judgment must be “designed solely to obtain immediate appeal of the prior adverse decision, without pursuing piecemeal appellate review.” Id. Fourth, the

2See BIW Deceived v. Loc. S6, 132 F.3d 824, 828 (1st Cir. 1997); Keefe v. Prudential Prop. & Cas. Ins. Co., 203 F.3d 218, 222–23 (3d Cir. 2000); Cohen v. Va. Elec. & Power Co., 788 F.2d 247, 249 (4th Cir. 1986); Downey v. State Farm Fire & Cas. Co., 266 F.3d 675, 682–83 (7th Cir. 2001); Slaven v. Am. Trading Transp. Co., 146 F.3d 1066, 1070 (9th Cir. 1998); Mock v. T.G. & Y. Stores Co., 971 F.2d 522, 527 (10th Cir. 1992); Shores v. Sklar, 885 F.2d 760, 762 (11th Cir. 1989) (en banc), cert. denied, 493 U.S. 1045 (1990). To our knowledge, only the Fifth Circuit has arguably disagreed, see Amstar Corp. v. S. Pac. Transp. Co. of Tex. & La., 607 F.2d 1100 (5th Cir. 1979), but a subsequent Fifth Circuit decision called Amstar into question, see Ybarra v. Dish Network, L.L.C., 807 F.3d 635, 639 (5th Cir. 2015); see also Dorse v. Armstrong World Indus., Inc., 798 F.2d 1372, 1375–77 (11th Cir. 1986).

3In Ali, the district court issued a ruling denying summary judgment and rejecting the third-party plaintiffs’ claims “as a matter of law.” 719 F.3d at 89. The parties then jointly requested that the district court dismiss all pending claims with prejudice, which it did, “in order to obtain immediate appellate review.” Id. at 90. Although in Ali the judgment was a “voluntary dismissal,” from which a plaintiff sought to appeal, the reasoning of that decision applies with equal force to the situation here, where a defendant seeks to appeal after entry of a consent judgment.

14

appellant must have “expressly preserved” the right to appeal. LaForest, 569 F.3d at 74 (2d Cir. 2009); see also Linde v. Arab Bank, PLC, 882 F.3d 314, 324 (2d Cir. 2018) (same). Consideration of these four factors is faithful to the Supreme Court’s mandate that “finality is to be given a practical rather than a technical construction.” Microsoft Corp. v. Baker, 582 U.S. 23, 37 (2017) (citation omitted). Our precedents have not directed that all four factors must be met before we exercise appellate jurisdiction over a voluntarily dismissed claim. Our decision in Ali did not discuss the fourth factor, and our decisions in LaForest and Linde did not address the first three. We need not decide whether each factor is necessary because here all four factors are present.

First, the district court plainly rejected the legal basis for New York’s defense. In its June 11 order granting a preliminary injunction, the district court conclusively held that “the ABA . . . stands as an obstacle to the FCC’s accomplishment and execution of its full purposes and objectives and is conflict- preempted.” N.Y. State Telecomms. Ass’n, 544 F. Supp. 3d at 282. It further held: “Because the ABA regulates within the field of interstate communications, it

15

triggers field preemption. Binding Second Circuit decisions are clear: the Communications Act’s ‘broad scheme for the regulation of interstate service by communications carriers indicates an intent on the part of Congress to occupy the field to the exclusion of state law.’” Id. at 285 (quoting Ivy Broad. Co. v. Am. Tel. & Tel. Co., 391 F.2d 486, 490–91 (2d Cir. 1968)). The district court was only required to find a likelihood of success on the merits in order to grant a preliminary injunction. But the court did not restrict its holding to such tentative terms. Instead, it articulated unequivocal and purely legal conclusions concerning the preemptive effect of federal law, which were in no way tentative nor contingent on further discovery or factual development.

Under our precedents, that practical resolution of the legal question in this case is sufficient to support an appeal from the subsequent final judgment. It is of no consequence that the district court’s conclusion was not technically final, because our inquiry is a pragmatic one. We look to whether the court resolved a claim “in effect” by “plainly reject[ing] [its] legal basis.” Ali, 719 F.3d at 88, 90. In other words, even a ruling that does not formally or technically resolve a claim can

16

suffice, as long as it makes clear that the court has effectively resolved the claim as a matter of law. When we have concluded we lacked jurisdiction to review stipulated judgments it was because we determined that the relevant interlocutory decision did not so plainly resolve a claim as a matter of law. See Empire Volkswagen Inc. v. World–Wide Volkswagen Corp., 814 F.2d 90, 95 (2d Cir. 1987); Palmieri v. Defaria, 88 F.3d 136, 140 (2d Cir. 1996). This case readily meets the standard articulated in Ali, given the district court’s unequivocal conclusions regarding preemption. 4 Even if we were to construe the district court’s legal conclusions in its June 11 order as merely tentative ones because they were resolved in the context of a preliminary injunction, the district court’s July 28 order 5 granting a permanent

4 The definitive legal conclusion reached by the district court in this case was nothing like the tentative predictions or contingent in limine rulings the dissent hypothesizes. See Diss. Op. at 15. Our reasoning here would not allow immediate appeal of those decisions, nor of every preliminary injunction decision. For example, a decision granting a preliminary injunction based on provisional legal analysis, on facts not yet fully developed, or primarily on irreparable harm would be entirely different. In short, the dissent sees a slippery slope only because it misses the guardrails already built into our case law. 5The July 28 judgment was amended on August 10 to correct a clerical error. See Joint App’x 160– 61.

17

injunction confirmed that it definitively rejected the legal basis for New York’s defense. That final judgment determined that federal law is not only likely to, but indeed does, preempt the ABA. The judgment stated that “the Court’s holdings on preemption in the June 11, 2021, memorandum and order resolve the substantive legal issues in this matter” and “[f]or the reasons given in the Court’s June 11, 2021, memorandum and order, the Court declares that [the ABA] is preempted by federal law.” Joint App’x 157. Had the district court determined otherwise, it would have rejected the parties’ stipulation to judgment or accepted it without adopting language declaring that its prior holding “resolve[d] the substantive legal issues in this matter” and unequivocally concluding that the ABA “is preempted by federal law” “[f]or the reasons given” in its earlier preliminary injunction order. Id. Although the district court judgment adopted stipulated language, that adoption reflects the district court’s understanding of the finality of its legal holding in this case. District courts are not rubber stamps. 6

6The dissent suggests that we misconstrue the nature of stipulated judgments, which are not rulings on the merits entitled to preclusive or precedential effect. See Diss. Op. at 12-14. But the dissent may misconstrue the nature of our inquiry here. Whatever the force of this stipulated

18

Second, all claims have now been disposed of with prejudice. Although in the district court the Plaintiffs voluntarily dismissed their § 1983 claim without prejudice, they have subsequently agreed to dismiss the claim with prejudice. See Supp. Br. for Appellees at 3. Doing so eliminated the risk of piecemeal appeals in this matter and cured any defect in finality posed by the § 1983 claim, as “we have allowed a [party] to appeal an adverse ruling disposing of fewer than all of its claims following [its] voluntary relinquishment of its remaining claims with prejudice.” Chappelle v. Beacon Commc’ns Corp., 84 F.3d 652, 653 (2d Cir. 1996); see also Empire Volkswagen, 814 F.2d at 94 (same).

Third, New York’s stipulation to final judgment was designed solely to obtain immediate appellate review of the district court’s underlying legal conclusion and does not invite piecemeal litigation or circumvent limitations on our appellate jurisdiction. Appeals from stipulated judgments are not permitted as a means to circumvent carefully calibrated restrictions on appellate jurisdiction,

judgment in a future case, there is no reason why we cannot look to its language to discern what this district court effectively determined in this case, under our case law concerning appeals from stipulated judgments.

19

such as (for example) the discretionary framework that allows courts to decline to hear appeals from class certification decisions. See Microsoft, 582 U.S. at 35, 38-40. 7 But this is simply not a case in which the parties tried to hoodwink the courts or skip the last leg of any real race. New York clearly was not seeking to circumvent the restrictions on interlocutory appeals, given that it had an appeal as of right from the grant of the preliminary injunction, see 28 U.S.C. § 1292(a)(1), or could have stipulated to the same result pursuant to Federal Rule of Civil Procedure 65(a)(2) (or through uncontested summary judgment practice or trial on stipulated

7 The dissent misunderstands Microsoft to mean that a stipulated-judgment appeal can never be used to “seize additional appellate rights.” Diss. Op. at 19. But that cannot be the rule if, as the dissent concedes, some stipulated-judgment appeals are permissible. Any time parties use this procedure, they are attempting to obtain some form of appellate review otherwise not immediately available. Microsoft concerns a narrower proposition: that parties may not manipulate stipulated judgments in order to circumvent restrictions on what parties may ordinarily appeal. In Microsoft, for example, the Court prohibited parties from using this strategy to force appellate review of a class certification decision that the court of appeals had exercised its discretion to deny. See 582 U.S. at 39-40. Similarly, in the non-precedential summary order cited by the dissent, we held that we lacked jurisdiction over a stipulated-judgment appeal following the grant of a motion to compel arbitration because the appeal would have circumvented the Federal Arbitration Act’s prohibition of appeals from the grant of such motions. See Bynum v. Maplebear, Inc., 698 F. App’x 23, 24 (2d Cir. 2017) (summary order).

20

facts). 8 Nor can it be said that the parties stipulated to a final judgment in order to bypass district court resolution of any open merits questions, given that the district court had already concluded in its June 11 order that federal law preempted the ABA. The parties have not circumvented the final judgment rule but have merely accelerated the process of obtaining the final judgment that became inevitable once the district court reached its legal conclusion. There was simply nothing left to litigate in the district court. New York had argued its case and lost.

Moreover, the stipulated-to dismissal does not “invite[] protracted litigation and piecemeal appeals.” Microsoft Corp., 582 U.S. at 37. If anything, the parties entered the consent judgment to avoid piecemeal adjudication and a needless drain on resources. The procedure here allows one appeal to resolve the issue of

8In fact, as the dissent acknowledges, if New York had appealed from the grant of the preliminary injunction, even in that interlocutory posture we could have determined that the Plaintiffs’ claim was “entirely void of merit” and decided to “award judgment to the appropriate party.” New York v. Nuclear Regul. Comm’n, 550 F.2d 745, 759 (2d Cir. 1977), superseded by rule on other grounds as recognized by Zervos v. Verizon N.Y., Inc., 252 F.3d 163, 170 (2d Cir. 2001). And even if we had not formally done so, a decision from this Court on the purely legal question of preemption in this case would not have left the district court with any room to disagree in subsequent proceedings on remand. In light of this, it is especially puzzling that the dissent suggests that New York circumvented any rules of appellate jurisdiction.

21

preemption in this case with finality, rather than litigating the same legal question once at the preliminary injunction stage and again after final judgment. And with the Plaintiffs having agreed to dismiss their § 1983 claim with prejudice, there will be nothing left for the parties to litigate following this appeal—barring, of course, review of this decision by the Supreme Court. As we said in Ali: “The federal policy against piecemeal appeals is not implicated where an entire case can be decided in a single appeal.” 719 F.3d at 89 (cleaned up). Plainly so here. If we affirm, the case ends. If we reverse, the case also ends.

Fourth, New York expressly preserved its right to appeal in the stipulated-

to final judgment. See Joint App’x 158 (stating that New York “reserves the right to appeal”). Having secured the ability to challenge the district court’s preemption conclusions in this Court, New York did not concede to the district court’s substantive holding, but rather agreed “that, if there was to be such a judgment, it should be final in form instead of interlocutory, so that they might come to this court without further delay.” United States v. Procter & Gamble Co., 356 U.S. 677, 681 (1958) (citation omitted). The matter being appealed—the district court’s

22

purely legal preemption holding—clearly falls within the scope of this express reservation. If, by contrast, New York expressly preserved only its right to challenge the district court’s choice of remedy on appeal and not its broader right to challenge the underlying legal holding, then we could not review the district court’s conclusions regarding preemption. However, New York’s express reservation of its right to appeal does not contain any such proviso and the preemption holding of the district court is unquestionably within the scope of the express reservation.

We recognize that the inquiry into our appellate jurisdiction will not necessarily end with these four factors in every case. Satisfying these factors may not be sufficient to confer jurisdiction if, for example, there is an independent reason for finding that adversity no longer remains between the parties or that the appeal has become moot. But here, we do not identify any additional basis for questioning our jurisdiction. To the contrary, this appeal bears all the hallmarks of a case or controversy: a live and genuine dispute remains between the parties, with material consequences at stake.

23

We are easily satisfied that we have jurisdiction to decide this appeal and we reject the dissent’s contention that the parties’ unremarkable use of a stipulated judgment in the circumstances of this case forever forecloses review of the district court’s decision enjoining New York’s duly enacted law. We turn to that review now. II. Preemption In this case, the Plaintiffs have advanced two theories of implied preemption. 9 First, they contend that the ABA is preempted because federal law occupies the entire field of rate regulations for interstate communications services to the exclusion of the states. Second, the Plaintiffs maintain that the ABA is conflict-preempted by the 2018 Order because the ABA stands as an obstacle to the FCC’s stated policy objective of deregulating ISPs. The district court agreed

9“Federal preemption of a state statute can be express or implied . . . .” SPGGC, LLC v. Blumenthal, 505 F.3d 183, 188 (2d Cir. 2007). “Implied preemption renders a state law inoperative in two circumstances: (1) when the state law ‘regulates conduct in a field that that Congress intended the Federal Government to occupy exclusively,’ (so called ‘field preemption’) and (2) when the state law ‘actually conflicts with federal law,’ (so called ‘conflict preemption’).” In re Jackson, 972 F.3d 25, 33 n.4 (2d Cir. 2020) (quoting English v. Gen. Elec. Co., 496 U.S. 72, 79 (1990)). In contrast, “[e]xpress preemption arises when a federal statute expressly directs that state law be ousted.” Air Transp. Ass’n of Am. v. Cuomo, 520 F.3d 218, 220 (2d Cir. 2008) (cleaned up). The Plaintiffs have not asserted any claim of express preemption in this appeal.

24

with both arguments. We review each of those conclusions in turn, de novo. Critcher v. L’Oreal USA, Inc., 959 F.3d 31, 34 (2d Cir. 2020).

A. Field Preemption Field preemption occurs when Congress manifests an intent to occupy an entire regulatory field to the exclusion of the states. This intent “can be inferred from a framework of regulation ‘so pervasive . . . that Congress left no room for the States to supplement it.’” Arizona v. United States, 567 U.S. 387, 399 (2012) (quoting Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230 (1947)). The Supreme Court has noted that these are “rare cases.” Kansas v. Garcia, 140 S. Ct. 791, 804 (2020). “[B]ecause the States are independent sovereigns in our federal system,” courts “start with the assumption that the historic police powers of the States were not meant to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress.” Medtronic, Inc. v. Lohr, 518 U.S. 470, 485 (1996) (citation omitted).

At the district court, the Plaintiffs argued that the ABA was field-preempted because the Communications Act preempted all state regulation of interstate communications services. That was quite a stunning claim. As amici Internet Law

25

Professors note, “no court ha[d] ever found field preemption of the whole of interstate communications. Instead, courts have evaluated field preemption claims with respect to much narrower subfields . . . .” Internet Law Profs. Br. 13. See, e.g., Freeman v. Burlington Broads., Inc., 204 F.3d 311, 319–20 (2d Cir. 2000) (considering “whether federal law preempts state and local regulation of [radio frequency] interference”); N.Y. SMSA Ltd. P’ship v. Town of Clarkstown, 612 F.3d 97, 105–06 (2d Cir. 2010) (identifying the field as “the regulation of the technical and operational aspects of wireless telecommunications service”).

Moreover, courts in New York and across the country have upheld numerous state regulations of interstate communications services against preemption challenges. See, e.g., ACA Connects v. Frey, 471 F. Supp. 3d 318, 323–26 (D. Me. 2020) (affirming Maine’s authority to restrict broadband providers from disseminating customers’ personal information); People v. Charter Commc’ns, Inc., 81 N.Y.S.3d 2, 3 (N.Y. App. Div. 2018) (affirming New York’s authority to regulate deceptive advertising by broadband providers about their broadband services); Patriotic Veterans, Inc. v. Indiana, 736 F.3d 1041, 1046–54 (7th Cir. 2013) (affirming

26

Indiana’s authority to regulate robocalls); Tex. Off. of Pub. Util. Counsel v. FCC, 183 F.3d 393, 418 (5th Cir. 1999) (affirming Texas’s authority to “impos[e] additional eligibility requirements on carriers otherwise eligible to receive federal universal service support”).

The Plaintiffs’ broad claim was stunning, but not long for this world.

Perhaps recognizing this position was not tenable, they defend only a narrower version on appeal. Instead of defining the field as all “interstate communications services,” they now argue that the relevant field is “rate regulation of interstate communications services.” Appellees’ Br. 34–35 (emphasis added). Because it appears that the Plaintiffs have abandoned their original position, we consider whether Congress has occupied the field of rate regulation of interstate communications services to the exclusion of the states. 10 We proceed by

10As a threshold matter, New York argues that the ABA is a purely intrastate regulation because the ABA’s “price regulation applies only to products offered by companies operating in New York to specified consumers who reside in New York, and it concerns only broadband service to be accessed from computers in New York.” Appellant’s Br. 32–33. However, the law of this Circuit instructs us that the FCC has jurisdiction to regulate communications services if the communications “go from one state to another.” N.Y. Tel. Co. v. FCC, 631 F.2d 1059, 1066 (2d Cir. 1980). This “end-to-end” analysis is the controlling test for whether a regulation is jurisdictionally

27

examining the scope of states’ historic police powers over communications services, the text and structure of the Communications Act, and the relevant case law.

1. The States’ Police Powers When reviewing preemption challenges, courts “start with the assumption that the historic police powers of the States were not to be superseded by [a] Federal Act unless that was the clear and manifest purpose of Congress.” Wyeth v. Levine, 555 U.S. 555, 565 (2009) (citation omitted). This Court has held that “[b]ecause consumer protection law is a field traditionally regulated by the states, compelling evidence of an intention to preempt is required in this area.” Gen. Motors Corp. v. Abrams, 897 F.2d 34, 41–42 (2d Cir. 1990).

In this case, however, the Plaintiffs contend that there should be no presumption against preemption because “[t]here is no historic presence of state law regulating the rates of interstate communications services.” Appellees’ Br. 43.

intra- or interstate, and applying it, we conclude that the ABA is a regulation of interstate communications services.

28

The Plaintiffs’ decision to narrow their argument on appeal does important work here. While New York and its amici cite many historical examples of state regulations of interstate communications services, the Plaintiffs argue that none of them are relevant because they are not rate regulations.

The Plaintiffs have moved the goalposts on the preemption field, but their claim fails anyway. Cable television is an interstate communications service, and when it was lightly regulated under Title I—as broadband internet is today— many states enacted laws that regulated the rates cable companies could charge for their services. See Philip R. Hochberg, The States Regulate Cable: A Legislative Analysis of Substantive Provisions 29–30, 91–96 (1978) (describing cable rate legislation and regulation in Delaware, Hawaii, Kansas, Massachusetts, Minnesota, Nebraska, Nevada, New Jersey, New York, South Dakota, and Virginia), https://perma.cc/Z89E-JTHQ. Among these regulatory regimes, New York’s system was “the most comprehensive,” with robust antidiscrimination provisions and requirements that price increases be approved by state authorities. Id. at 91–93. Nevada also imposed public utility–style regulations on cable

29

providers, including a requirement that rates be “just and reasonable.” TV Pix, Inc. v. Taylor, 304 F. Supp. 459, 460 (D. Nev. 1968) (three-judge court), aff’d, 396 U.S. 556 (1970). And when a group of cable companies challenged the Nevada statute, arguing—as the Plaintiffs do now—that it was preempted by the Communications Act, a three-judge panel unanimously rejected their claim. See id. at 464–65 (“Congress, in enacting the Federal Communications Act of 1934, did not intend absolute preemption of the field to the exclusion of all state regulation.”). That decision was summarily affirmed by the Supreme Court. 396 U.S. 556 (1970).

The Plaintiffs attempt to distinguish TV Pix by arguing that it “did not concern interstate rate regulation.” Appellees‘ Br. 45. That is incorrect. Although the TV Pix opinion describes the community antenna systems as being “essentially a local business,” 304 F. Supp. at 463, that language was not relevant to the field preemption holding. Instead, it was related to the court’s separate holding that the laws did not violate the Dormant Commerce Clause. Id. The TV Pix court stated that there was “no doubt” that the community antenna TV businesses were “engaged in interstate communication, even where, as here, the intercepted signals

30

emanate from stations located within the same State.” Id. at 461 (emphasis added) (quoting United States v. Sw. Cable Co., 392 U.S. 157, 168–69 (1968)).

Based on this history and precedent, we conclude that there is a tradition of states using their police power to regulate rates charged for interstate communications services. Therefore, we proceed “with the assumption” that such powers “were not to be superseded by the [Communications Act] unless that was the clear and manifest purpose of Congress.” Wyeth, 555 U.S. at 565. We turn next to the text of the Communications Act to determine that purpose.

2. The Text of the Communications Act The Plaintiffs’ main textual argument is that § 152 of the Communications Act evinces Congress’s intent to preempt all rate regulations of interstate communications services. Section 152 outlines the jurisdictional boundaries of the FCC and provides that:

(a) The provisions of this chapter shall apply to all interstate and foreign communication by wire or radio . . . which originates and/or is received within the United States, and to all persons engaged within the United States in such communication . . . .

(b) Except as provided in sections 223 through 227 of this title, inclusive, section 276, and section 332 of this title, and subject to the

31

provisions of section 301 of this title and subchapter V–A, nothing in this chapter shall be construed to apply or to give the Commission jurisdiction with respect to (1) charges, classifications, practices, services, facilities, or regulations for or in connection with intrastate communication service by wire or radio of any carrier . . . .

47 U.S.C. § 152 (emphases added).

The Plaintiffs contend that this statute “is how Congress confirmed the FCC’s exclusive jurisdiction over rate-setting for interstate communications services,” though they do not explain how their reading of this text could be limited to rate regulation. Appellees’ Br. 36. They quote Louisiana Public Service Commission v. FCC for the proposition that subsections (a) and (b) “divide the world . . . into two hemispheres—one comprised of interstate service, over which the FCC would have plenary authority, and the other made up of intrastate service, over which the States would retain exclusive jurisdiction.” 476 U.S. 355, 360 (1986). The district court also relied on this language from Louisiana, stating that “[t]he FCC’s jurisdiction would hardly be ‘plenary’ if it loses, to the states’ gain, the right to make rules regarding certain interstate communications services when the FCC alters” the Title under which those services are regulated. N.Y. State

32

Telecomms. Ass’n, 544 F. Supp. 3d at 287. These arguments are flawed for two reasons.

First, the Plaintiffs’ reliance on Louisiana is misplaced. The Plaintiffs argue that the Supreme Court interpreted § 152 as dividing the world of communications into two mutually exclusive hemispheres. But that is in fact the opposite of what the Supreme Court did. The Louisiana Court said the following in reference to § 152:

[W]hile the Act would seem to divide the world of domestic telephone service neatly into two hemispheres—one comprised of interstate service, over which the FCC would have plenary authority, and the other made up of intrastate service, over which the States would retain exclusive jurisdiction—in practice, the realities of technology and economics belie such a clean parceling of responsibility. . . . [B]ecause the same carriers provide both interstate and intrastate service, actions taken by federal and state regulators within their respective domains necessarily affect the general financial health of those carriers, and hence their ability to provide service, in the other “hemisphere.”

476 U.S. at 360 (emphases added). Louisiana made clear that the states continue to have a role in regulating communications services, even if such regulations touch on interstate services. See id. at 375 (“The Communications Act not only establishes dual state and federal regulation of telephone service; it also recognizes that

33

jurisdictional tensions may arise as a result of the fact that interstate and intrastate service are provided by a single integrated system.”). The Supreme Court’s decision in Louisiana strongly undermines, rather than supports, the Plaintiffs’ argument based on the text of § 152.

Second, although we agree that § 152(a) broadly grants the FCC jurisdiction over “all interstate and foreign communication,” nothing in the text suggests that the FCC has exclusive jurisdiction over interstate communication, which is the relevant question for implied field preemption. And the dissent, for its part, never explains how it makes the leap from broad jurisdiction to exclusive jurisdiction. See Diss. Op. at 23-24. The Supreme Court’s decisions on preemption make clear that “the mere existence of a federal regulatory or enforcement scheme . . . does not by itself imply pre-emption of state remedies.” English v. Gen. Elec. Co., 496 U.S. 72, 87 (1990). Thus, “a statute granting regulatory authority over [a] subject matter to a federal agency” is not in and of itself sufficient to find field preemption. Kurns v. R.R. Friction Prods. Corp., 565 U.S. 625, 638 (2012) (Kagan, J., concurring). “Congress must do much more to oust all of state law from a field.” Id; see also

34

Hillsborough Cnty. v. Automated Med. Lab’ys, Inc., 471 U.S. 707, 719 (1985) (“Undoubtedly, every subject that merits congressional legislation is, by definition, a subject of national concern. That cannot mean, however, that every federal statute ousts all related state law.”).

The Plaintiffs nonetheless argue that this statutory language granting federal authority evinces an intent to preempt because Congress used substantially similar language in the Federal Power Act and the Natural Gas Act. See 16 U.S.C. § 824(b)(1); 15 U.S.C. § 717(b)–(c). Those Acts give the Federal Energy Regulatory Commission “exclusive authority” over interstate wholesale electricity sales, Hughes v. Talen Energy Mktg., LLC, 578 U.S. 150, 154 (2016), and “exclusive jurisdiction” over interstate wholesale natural gas sales, Schneidewind v. ANR Pipeline Co., 485 U.S. 293, 300–01, 305 (1988).

Without context, this seems like a compelling argument, and it is one the dissent adopts at face value. See Diss. Op. at 25. But the argument loses its force when one notices that the jurisdictional provisions in the Federal Power Act and the Natural Gas Act were passed after the Supreme Court issued a series of

35

Dormant Commerce Clause decisions holding that “regulation of wholesale rates of gas and electrical energy moving in interstate commerce is beyond the constitutional powers of the States.” Interstate Nat. Gas Co. v. Fed. Power Comm’n, 331 U.S. 682, 689 & n.13 (1947). “[T]he basic purpose of Congress in passing the Natural Gas Act was to occupy this field in which the Supreme Court has held that the States may not act.” Id. at 690 (internal quotation marks omitted); see also Jersey Cent. Power & Light Co. v. Fed. Power Comm’n, 319 U.S. 61, 67–68 (1943) (“The primary purpose of Title II, Part II [of the Federal Power Act] . . . was to give a federal agency power to regulate the sale of electric energy across state lines. Regulation of such sales had been denied to the States . . . .”). In other words, the similar jurisdictional language from the Federal Power Act and the Natural Gas Act does not evince Congress’s intent to preempt the field, because Congress was acting in an area in which it was already established that states were prohibited from regulating.

36

Therefore, nothing in the text of § 152 provides “compelling evidence” of Congress’s intent to occupy the field of rate regulation of interstate communications services. Gen. Motors, 897 F.2d at 41.

3. The Structure of the Communications Act Other provisions of the Communications Act also rebut the Plaintiffs’ claim that the federal government exclusively occupies the field of rate regulation of interstate communications services.

To start, the Communications Act has no framework for rate regulation over Title I services like broadband, let alone one that is “so pervasive . . . that Congress left no room for the States to supplement it.” Arizona, 567 U.S. at 399 (cleaned up). When a service is regulated under Title I, the FCC lacks the express or ancillary authority to impose rate regulations. See Comcast, 600 F.3d at 655 (D.C. Cir. 2010).

The sole grant of regulatory authority within Title I is located at 47 U.S.C.

§ 154(i), which permits the FCC to “make such rules and regulations, and issue such orders, not inconsistent with this chapter, as may be necessary in the execution of its functions.” The Supreme Court has held that this authority is

37

“restricted to [acts] reasonably ancillary to the effective performance of the Commission’s various responsibilities.” Sw. Cable, 392 U.S. at 178. Thus, the Court has vacated FCC regulations of information services unless such regulations are in furtherance of a “statutorily mandated responsibilit[y]” that is rooted in “an express delegation of authority to the Commission.” Comcast, 600 F.3d at 652 (citing Sw. Cable, 392 U.S. at 177–78; United States v. Midwest Video Corp., 406 U.S. 649, 670 (1972) (plurality opinion)). However, neither the Plaintiffs—nor the FCC itself—have ever identified a “statutorily mandated responsibility” in the Communications Act that would permit the use of § 154(i) to impose common carrier requirements such as rate regulation. Cf. Verizon, 740 F.3d at 635–50 (D.C. Cir. 2014) (upholding broadband disclosure rules as ancillary to 47 U.S.C. § 1302).

This absence of regulation is the exact opposite of a federal “framework . . .

so pervasive” that it results in field preemption. Arizona, 567 U.S. at 399 (cleaned up). The Plaintiffs’ position would create a regulatory vacuum in which the federal government has both declined to regulate an industry and simultaneously prohibited states from regulating. Though the Supreme Court has noted that such

38

a vacuum may be constitutionally permissible, “to say that it can be created is not to say that it can be created subtly.” P.R. Dep’t of Consumer Affs. v. Isla Petrol. Corp., 485 U.S. 495, 500 (1988); cf. Sprietsma v. Mercury Marine, 537 U.S. 51, 68–70 (2002) (finding no field preemption based on congressional delegation to agency where statute “does not require the [agency] to promulgate comprehensive regulations covering every aspect” of the asserted field). Congress has not legislated an absence of regulatory authority here.

Furthermore, the Communications Act contains provisions expressly prohibiting states from regulating specific types of communications services, and none covers all rate regulations of interstate communications services. Instead, the Act identifies specific types of communications services, regulates them differently under different Titles, and preempts state regulation of some of them on a case- by-case basis. For example, when Congress passed the Cable Communications Policy Act of 1984, Pub. L. No. 98-549, 98 Stat. 2779, it added Title VI to the Communications Act and expressly forbade state regulation of “the rates for the provision of cable service except to the extent provided under this section and

39

section 532 of this title.” 47 U.S.C. § 543(a) (emphasis added). This provision would be wholly unnecessary if the broader field had already been preempted. Congress similarly included a forbearance provision for Title II services, which prohibits the states from enforcing some Title II regulations if certain prerequisites are met and the FCC concludes that the regulations at issue are unnecessary. Id. § 160. No such regime exists for services regulated under Title I.

There is simply no indication that Congress intended to preempt a field as broad as “rate regulation of interstate communications services.” To the contrary, Congress made explicit its intent to preempt other subfields of interstate communications. Supreme Court precedent is clear that “Congress’ enactment of a provision defining the pre-emptive reach of a statute implies that matters beyond that reach are not pre-empted.” Cipollone v. Liggett Grp., 505 U.S. 504, 517 (1992).

Other provisions of the Communications Act also support our conclusion that rate regulation is not field-preempted. For example, Section 414 contains a “savings clause,” which states that “the provisions of this chapter are in addition to

40

such remedies” that “now exist[] at common law or by statute.” 47 U.S.C. § 414 (emphasis added). And strikingly, § 1302(a) provides:

The Commission and each State commission with regulatory jurisdiction over telecommunications services shall encourage the deployment on a reasonable and timely basis of advanced telecommunications capability to all Americans . . . by utilizing, in a manner consistent with the public interest, convenience, and necessity, price cap regulation . . . or other regulating methods that remove barriers to infrastructure investment.

(emphasis added). The most natural conclusion to draw from all these provisions (and the one that comports with our presumption against preemption) is that Congress intended for the states to retain their regulatory authority over many interstate communications services—and to play a role in regulating the rates charged for such services—unless it said otherwise.

4. Case Law on the Communications Act The final refuge of the Plaintiffs’ case for field preemption is this Court’s decision in Ivy Broadcasting Co. v. American Telephone & Telegraph Co., 391 F.2d 486 (2d Cir. 1968). In Ivy, we drew on the Supreme Court’s decisions in Postal Telegraph-Cable Co. v. Warren-Godwin Lumber Co., 251 U.S. 27 (1919), and Western Union Telegraph Co. v. Boegli, 251 U.S. 315 (1920), to conclude that “questions

41

concerning the duties, charges and liabilities of telegraph or telephone companies with respect to interstate communications service are to be governed solely by federal law and that the states are precluded from acting in this area.” Ivy, 391 F.2d at 491.

The Plaintiffs argue that Ivy’s field preemption holding extends to all interstate communications services—not just telephone and telegraph companies. We disagree. Ivy does not field-preempt rate regulation of broadband internet (or other Title I information services) because the Communications Act subjects those services to an entirely different regulatory regime than telephone and telegraph companies.

Telegraph and telephone services were and continue to be regulated as common carriers under the Communications Act. These services are subject to numerous regulations that do not apply to Title I services like broadband internet. The Ivy court’s field preemption holding was premised on its observation that “Congress has enacted comprehensive legislation regulating common carriers engaged in interstate telegraph and telephone transmission.” Id. at 490 (emphases

42

added). The Court highlighted provisions of the Communications Act that are specific to common carriers: § 201, which “requires communications carriers to furnish communications service upon reasonable request”; §§ 201–02, which prohibit carriers from levying “unreasonable or discriminatory charges, practices, classifications and regulations”; and § 203, which requires carriers to “file tariff schedules with the FCC.” Id. Based on “this broad scheme for the regulation of interstate service by communications carriers,” it concluded that Congress had preempted the field. Id (emphases added).

Moreover, the Supreme Court cases Ivy relied upon—Postal Telegraph-Cable Co. and Western Union Telegraph Co.—also concerned telegraph companies that were regulated as common carriers under the predecessor to the Communications Act. Both of those cases relied on the fact that Congress had subjected carriers to the “rule of equality and uniformity of rates” when concluding they could only be regulated by the federal government. Postal Tel.-Cable, 251 U.S. at 30; see also W. Union Tel. Co., 251 U.S. at 316 (“[T]he provisions of the statute bringing telegraph companies under the Act to Regulate Commerce as well as placing them under the

43

administrative control of the Interstate Commerce Commission so clearly establish the purpose of Congress to subject such companies to a uniform national rule . . . .” (emphasis added)). Ivy’s logic may apply to other communications services with common carrier obligations, but it does not apply to services that are wholly exempt from them. The extensive federal regulation of common carriers that justifies field preemption in Ivy is nowhere to be found for broadband internet.

Reading Ivy to cover all communications services would also conflict with Supreme Court precedent on the Communications Act. In Head v. New Mexico Board of Examiners in Optometry, the Supreme Court warned that “the validity of [a preemption] claim cannot be judged by reference to broad statements about the ‘comprehensive’ nature of federal regulation under the Federal Communications Act.” 374 U.S. 424, 429–30 (1963). The Plaintiffs ask us to hold that the Communications Act exempts all services from state rate regulation—regardless of how those services are regulated under the Communications Act. If we were to do that, we would be making the exact sort of sweeping assumption about the Act

44

that Supreme Court precedent forecloses and that is contrary to the actual statutory analysis by this Court in Ivy.

In sum, neither the text and structure of the Communications Act, the history of this type of regulation, nor relevant precedent support the Plaintiffs’ argument that Congress intended to preempt the field of rate regulation of interstate communications services when it passed the Communications Act.

B. Conflict Preemption In the alternative to their field preemption contention, the Plaintiffs argue that the ABA is conflict-preempted because it stands as an obstacle to the accomplishment and execution of the FCC’s 2018 Order. As discussed earlier, the 2018 Order reclassified broadband internet as a Title I service in order to “end utility-style regulation of the Internet in favor of . . . market-based policies” and adopt a “light-touch regulatory framework.” 2018 Order ¶¶ 2, 106. By moving broadband outside of the more comprehensive regulatory regime in Title II, the FCC surrendered the statutory authority to enact any rate regulations on

45

broadband internet providers. See Comcast, 600 F.3d at 655 (D.C. Cir. 2010); Verizon, 740 F.3d at 650 (D.C. Cir. 2014).

Because the ABA subjects broadband providers to rate regulation—a “centerpiece of common-carrier regulation”—the Plaintiffs argue that it stands as an obstacle to the “federal policy of promoting broadband deployment while preserving an open internet.” Appellees’ Br. 17. We consider whether this agency- driven federal policy preference carries preemptive effect against the states and conclude that it does not.

“The burden of establishing obstacle preemption, like that of impossibility preemption, is heavy: the mere fact of tension between federal and state law is generally not enough to establish an obstacle supporting preemption, particularly when the state law involves the exercise of traditional police power.” In re MTBE Prods. Liab. Litig., 725 F.3d 65, 101–02 (2d Cir. 2013) (cleaned up).

Under well-established principles of administrative law and federalism, “States are not permitted to use their police power” to enact a regulation if “failure of . . . federal officials affirmatively to exercise their full authority takes on the

46

character of a ruling that no such regulation is appropriate or approved pursuant to the policy of the statute.” Ray v. Atl. Richfield Co., 435 U.S. 151, 178 (1978) (cleaned up). However, “a federal agency may pre-empt state law only when and if it is acting within the scope of its congressionally delegated authority.” La. Pub. Serv. Comm’n, 476 U.S. at 374. If Congress has not conferred “power to act” upon an agency, that agency cannot “pre-empt the validly enacted legislation of a sovereign State.” Id. It follows that if an agency has no authority to regulate in a particular field, its policy preferences cannot be a valid basis for regulatory action or preemption. See id. at 374–75 (“To permit an agency to expand its power in the face of a congressional limitation on its jurisdiction would be to grant to the agency power to override Congress.”).

Therefore, the question at the heart of the conflict preemption inquiry is whether the FCC has the statutory authority to enact (or preempt) common carrier–style regulations of broadband under Title I. Our two sister circuits that have considered this question have determined the answer is “no.” Mozilla, 940

47

F.3d at 76–86 (D.C. Cir. 2019); ACA Connects v. Bonta, 24 F.4th 1233, 1241–45 (9th Cir. 2022). We agree.

As discussed earlier, Title II imposes common carrier obligations on telecommunications services, including a requirement that rates be “just and reasonable.” 47 U.S.C. § 201(b). Title II also includes a “forbearance provision” that allows the FCC to decline to enforce some regulations of telecommunications services if it believes regulation is unnecessary and forbearance is in the public interest. Id. § 160(a). If the FCC decides to forbear from imposing a common carrier obligation, the states are prohibited from imposing that same obligation on the telecommunications service. Id. § 160(e). There is little doubt that when the FCC determines that a particular communications service should be subject to the heightened regulatory regime of Title II, it has the concomitant power to preempt state law that conflicts with its regulatory decisions.

In contrast, Title I grants the FCC no authority to impose rate regulations, nor does it contain a forbearance provision similar to Title II. Thus, because broadband is now regulated as a Title I service, the FCC has no congressionally

48

delegated authority to impose or forebear rate regulations. Absent the “power to act,” the FCC has no power to preempt broadband rate regulation. La. Pub. Serv. Comm’n, 476 U.S. at 374; see also Nat’l Ass’n of Regul. Util. Comm’rs v. FCC, 533 F.2d 601, 620 n.113 (D.C. Cir. 1976) (noting a “vital difference between a refusal to use granted power, and an attempt to prevent regulation by others in an area where no ordinary Commission jurisdiction appears to exist”).

Neither the Plaintiffs nor our dissenting colleague attempt to identify a source of statutory authority that gives the FCC the power to preempt anywhere in Title I. Instead, the Plaintiffs argue (and the dissent accepts) that the agency’s threshold decision to recategorize broadband from Title II to Title I is an independent source of preemptive authority because it is an “affirmative exercise of the FCC’s statutory authority” and was done to “prohibit the very ex ante rate regulation that the ABA imposes.” Appellees’ Br. 18 (internal quotation marks omitted); see also Diss. Op. at 27-28.

To be sure, the FCC’s decision on how broadband should be classified is entitled to Chevron deference. Brand X, 545 U.S. at 980–81; Mozilla, 940 F.3d at 18–

49

20 (concluding that the FCC’s decision to reclassify broadband from Title II to Title I in the 2018 Order was lawful). But the fact that the FCC can choose between Title I and Title II does not mean that the FCC can opt to retain its Title II preemption authority after reclassifying broadband as a Title I service. There is a crucial distinction between being able to choose which of two exclusive regulatory regimes applies and being able to pick and choose powers from both regulatory regimes simultaneously. Whereas the former comports with the agency’s statutory authority, the latter contravenes it. See Mozilla, 940 F.3d at 80 (observing that the FCC “cannot completely disavow Title II with one hand while still clinging to Title II forbearance authority with the other”).

The Plaintiffs defend this pick-and-choose approach by arguing that “[t]he FCC’s policy preferences are not separable from the 2018 Order’s classification decision.” Appellees’ Br. 20. Because “the FCC started by reaching the affirmative determination that interstate broadband should not be subject to ex ante rate regulation,” and “[t]he D.C. Circuit [in Mozilla] upheld the FCC’s policy grounds as a reasoned basis for its selection of the regulatory regime to govern interstate

50

broadband,” the Plaintiffs argue that according this policy decision preemptive force would be consistent with the principles of Chevron deference. Appellees’ Br. 20–22.

This approach essentially asks us to apply another layer of deference to a determination that already receives Chevron deference. The Plaintiffs hope that the definitional ambiguity “that permits the Commission to classify broadband under Title I” can somehow “spawn[] a power to preempt with all the might of an express statutory grant of authority.” Mozilla, 940 F.3d at 82. But this Chevron- squared strategy fails for three reasons.

First, contrary to the Plaintiffs’ claims, the FCC’s policy preferences and its classification decision are separable. The FCC did not justify its classification decision solely on policy grounds. It also engaged in statutory interpretation and concluded that “the best reading of the relevant definitional provisions of the Act supports classifying broadband Internet access service as an information service.”

51

2018 Order ¶ 20. The FCC called its statutory analysis “sufficient grounds alone on which to base [its] classification decision.” Id. ¶ 86.

Second, the Plaintiffs’ expansive reading of Chevron has no basis in Chevron itself. Chevron is a case about filling gaps in statutes, “not a magic wand that invests agencies with regulatory power beyond what their authorizing statutes provide.” Mozilla, 940 F.3d at 84. If the Plaintiffs had pointed to some statutory ambiguity in Title I and the FCC had construed that provision as providing it with the power to impose rate regulations, then Chevron might be invoked in favor of preempting the ABA. But the only ambiguity that the Plaintiffs have identified pertains to whether broadband internet is an “information service” or a “telecommunications service.” 47 U.S.C. § 153(24), (53). The FCC has the power to fill that gap, and it can use its policy judgment to choose one category or the other, but it cannot rewrite the Communications Act to change the consequences

52

that flow from that choice. To hold otherwise “would virtually free the Commission from its congressional tether.” Comcast, 600 F.3d at 655.

Third, the Plaintiffs provide no coherent basis for distinguishing our implied preemption analysis from the express preemption analysis in Mozilla, which is persuasive authority. The district court concluded that the D.C. Circuit’s decision in Mozilla did not foreclose a finding of conflict preemption because it struck down the 2018 Order’s express preemption provision and left the question of its implied preemptive effect for another day. The court thus reasoned that the decision “does not preclude or revoke the 2018 Order’s implicit preemptive effect.” N.Y. State Telecomms. Ass’n, 544 F. Supp. 3d at 283.

To be sure, the Mozilla court stated that “it would be wholly premature to pass on the preemptive effect, under conflict or other recognized preemption principles, of the remaining portions of the 2018 Order” because “no particular state law is at issue in this case.” 940 F.3d at 86. However, Mozilla was also clear that the statutory ambiguity that allows the FCC to choose between Title I and Title II is not a freestanding source of preemptive authority. See id. at 82. The

53

Plaintiffs—who do not argue that Mozilla was wrongly decided—fail to explain why the same statutory ambiguity should confer implied preemptive authority when it does not confer express preemptive authority.

Instead, the Plaintiffs contend that Mozilla vacated the Preemption Directive on different grounds—namely, because it tried “to categorically abolish all fifty States’ statutorily conferred authority to regulate intrastate communications.” Appellees’ Br. 26 (quoting Mozilla, 940 F.3d at 86). This argument is also unavailing. Though the scope of the Preemption Directive was one reason why it was unlawful, it was not the sole reason. The Preemption Directive was also vacated because it was not rooted in a relevant source of statutory authority. See Mozilla, 940 F.3d at 78 (“[T]he power to preempt the States’ laws must be conferred by Congress. It cannot be a mere byproduct of self-made agency policy. Doubly so here where preemption treads into an area—State regulation of intrastate communications—over which Congress has expressly ‘deni[ed]’ the Commission regulatory authority.” (emphasis added)). Because implied preemption, like

54

express preemption, “cannot be a mere byproduct of self-made agency policy,” the Plaintiffs’ attempt to distinguish Mozilla must fail. Id.

***

Several of the Plaintiffs in this action vociferously lobbied the FCC to classify broadband internet as a Title I service in order to prevent the FCC from having the authority to regulate them. See Donald Shaw, Amidst Fight to Kill Net Neutrality, Comcast and Other Telecoms Spent $190 Million on Lobbying, Sludge (June 11, 2018), https://perma.cc/5BVU-Y97E. At that time, Supreme Court precedent was already clear that when a federal agency lacks the power to regulate, it also lacks the power to preempt. The Plaintiffs now ask us to save them from the foreseeable legal consequences of their own strategic decisions. We cannot. If they believe a requirement to provide internet to low-income families at a reduced price is unfair or misguided, they have several pathways available to them. They could take it up with the New York State Legislature. They could ask Congress to change the scope of the FCC’s Title I authority under the Communications Act. They could ask the FCC to revisit its classification decision, as it has done several times before.

55

But they cannot ask this Court to distort well-established principles of administrative law and federalism to strike down a state law they do not like.

CONCLUSION

The judgment of the United States District Court for the Eastern District of New York is REVERSED, and the permanent injunction barring enforcement of the Affordable Broadband Act is VACATED.

21-1975 N.Y. State Telecomms. Ass’n, Inc. v. James

RICHARD J. SULLIVAN, Circuit Judge, dissenting:

I respectfully dissent from the majority’s opinion for two reasons. First, I believe that we lack jurisdiction to even hear this appeal. Second, even if we had jurisdiction to reach the merits of the parties’ preemption arguments, I am persuaded that New York’s Affordable Broadband Act (the “ABA”) is preempted by federal law. I. We Lack Appellate Jurisdiction To Review The Stipulated Judgment.

This appeal comes to us in an “unusual posture.” Ali v. Fed. Ins. Co., 719 F.3d 83, 88 (2d Cir. 2013). After New York was preliminarily enjoined from enforcing the ABA, it stipulated to judgment against it, and then appealed that stipulated judgment. This was a strategic move. In the district court’s preliminary injunction order, it stated that the ABA “is conflict-preempted” by federal law, and thus concluded that the challengers were likely to succeed in showing preemption on the merits, as required to obtain a preliminary injunction. N.Y. State Telecomms. Ass’n, Inc. v. James, 544 F. Supp. 3d 269, 282 (E.D.N.Y. 2021) (“NYSTA”). At that point, New York could have appealed the injunction directly under 28 U.S.C. § 1292(a)(1) (in fact, New York initially filed such an appeal, only to later withdraw it). That interlocutory appeal, however, would have been a narrow challenge only

2

to whether the district court “abused its discretion” in granting the injunction, as opposed to a challenge that would produce “a final resolution of the merits” of preemption. Univ. of Tex. v. Camenisch, 451 U.S. 390, 393 (1981). In other words, in appealing the preliminary injunction, New York could not have asked us for judgment on the merits of preemption in its favor – it could have asked us only to dissolve the injunction while it continued to litigate the merits before the district court.

Rather than pursue that limited appeal, New York instead consented to a stipulated judgment in order to take a full appeal on the merits of preemption. That is, it stipulated to a judgment against it and asked the district court to enter a permanent injunction forbidding it from enforcing the ABA as preempted. See J. App’x at 157. The district court obliged, and New York has now appealed the resulting judgment, asking us to award it judgment on the merits with a finding that the ABA is not preempted by federal law.

But this tactic – which I will refer to as a “stipulated judgment appeal” – is generally not permitted as a shortcut to appellate review. Because these appeals are attempts to “evade the final judgment rule,” we allow them in only limited

3

circumstances. Palmieri v. Defaria, 88 F.3d 136, 139 (2d Cir. 1996). 1 In the majority’s view, an appellant can appeal from a stipulated judgment when (1) the district court “plainly rejected the legal basis” for the appellant’s case (either a claim or defense), (2) all claims are disposed of with prejudice, (3) the stipulated judgment is “designed solely to obtain immediate appeal of the prior adverse decision, without pursuing piecemeal appellate review,” and (4) the appellant has “expressly preserved” the right to appeal. Maj. Op. at 13–14 (internal quotation marks omitted).

Though I agree that all of these elements are prerequisites, our precedent requires two more conditions before a party may appeal a stipulated judgment. First, in order to “plainly reject[]” the legal basis for the appellant’s case, id. at 13, the district court’s decision must be a “final ruling” on an issue, as opposed to a

1 Over the years, we have confronted stipulated judgment appeals by both plaintiffs and defendants. For plaintiffs, such appeals usually follow an adverse interlocutory decision in the district court and a voluntary dismissal of all claims under Federal Rule of Civil Procedure 41(a)(2). See, e.g., Palmieri, 88 F.3d at 140. For defendants, stipulated judgment appeals typically involve situations like the one here, in which the appellant received an adverse interlocutory decision below, followed by entry of a judgment by consent – effectively a court-approved settlement. See, e.g., LaForest v. Honeywell Int’l Inc., 569 F.3d 69, 73 (2d Cir. 2009). Though there are subtle distinctions between these two scenarios, they are not relevant to this discussion, and I collectively refer to both types as “stipulated judgment appeals.” See generally Bryan Lammon, Manufactured Finality, 69 Vill. L. Rev. (forthcoming 2024) (manuscript at 23–37) (discussing various attempts to “manufacture[] finality” through voluntary dismissals and stipulated judgments), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4572017 [https://perma.cc/86QK-WMVE].

tentative finding or dicta, Palmieri, 88 F.3d at 139 (emphasis added). In other words, a decision cannot “effectively dismiss[]” a claim when it is only a provisional finding that is “subject to change when the case unfolds.” Id. (quoting Luce v. United States, 469 U.S. 38, 41–42 (1984)). Second, the stipulated judgment appeal cannot be an attempt to circumvent the interlocutory appellate rules already in place. As the Supreme Court has held, if the interlocutory appellate rules preauthorize a narrow right to appeal certain issues, then a litigant cannot use a stipulated judgment to claim the right to appeal additional issues beyond those preauthorized. See Microsoft Corp. v. Baker, 582 U.S. 23, 31–32 (2017) (holding that a litigant cannot use a stipulated judgment to appeal a class certification denial “as a matter of right” (internal quotation marks omitted)).

To invoke our appellate jurisdiction, both conditions must be met. Because neither is present here, I would dismiss the appeal for lack of appellate jurisdiction.

A. The Adverse “Decision” Was Provisional Dicta.

Our precedents make clear that an appellant cannot appeal a stipulated judgment when it suffered only a tentative setback in the district court. In other words, if a district court issues a provisional finding subject to change – such as one that casts doubt on a litigant’s claims only in dicta – then that cannot be an “effective dismissal” of the claims, and no appeal can be taken from a stipulated

5

judgment thereafter. We said as much in Palmieri v. Defaria, where we held that a litigant could not appeal a stipulated judgment when he suffered a tentative evidentiary loss before the district court that was “subject to change at trial.” 88 F.3d at 140.

In Palmeiri, the plaintiff brought copyright claims accusing the defendant of copying his song and sought to prove up that allegation with evidence that the defendant had had access to the disputed song prior to the alleged infringement. See id. at 137. After the defendant moved in limine to exclude that evidence, the district court granted the motion in part, finding that some of the evidence concerning the defendant’s access to the song was inadmissible and reserving for trial whether the rest could be introduced. See id. Disappointed with that ruling, the plaintiff invited the district court to enter final judgment against him so that he could appeal the in limine ruling right away. See id. at 138. The district court did so, and the plaintiff appealed the resulting judgment, challenging the district court’s in limine findings.

Emphasizing that the in limine ruling was merely tentative, we held that the stipulated judgment was not appealable. Though we acknowledged the rule that stipulated judgment appeals are occasionally permitted when the district court

6

had “effectively dismissed [the] case,” id. at 139, we nonetheless held that the in limine ruling was not an “effective dismissal” because it lacked two features: (1) the district court had not “take[n] the position” that the plaintiff’s proof was insufficient as a matter of law, and (2) the in limine ruling was merely tentative and “subject to change at trial in the district court’s discretion.” Id. at 140. In other words, we recognized an additional limit on the “effective dismissal” rule – namely, that the adverse decision below must be a “final ruling” as opposed to one that is merely tentative or conditional. Id. at 139 (“An in limine evidentiary ruling does not constitute a final ruling on admissibility.” (italics added)). 2 Indeed, we emphasized the provisional nature of the in limine ruling throughout our opinion, and even distinguished earlier “effective dismissal” cases because those involved district court orders that “could not be examined again at trial.” Id. at 141 (distinguishing Allied Air Freight v. Pan Am. World Airways, 393 F.2d 441 (2d Cir. 1968)). As we went on to explain, this rule – that a stipulated judgment cannot be appealed when the adverse finding is only tentative – makes

2 Though we have characterized our rule against stipulated judgment appeals as “jurisdiction[al],” Ali, 719 F.3d at 88, we have not explained whether the rule is constitutional or statutory in nature. But see Bryan Lammon, Voluntary Dismissals, Jurisdiction & Waiving Appellate Review, 92 U. Cin. L. Rev. 394, 406 (2023) (arguing that this rule is best understood as a waiver doctrine and warning that treating it as an Article III issue could mean conditional guilty pleas are unconstitutional). Whatever the rule’s origins, it bars New York’s appeal here.

good sense. Though we can take appeals from stipulated judgments following conclusive holdings, “[t]here is no reason to spend scarce judicial resources reviewing a decision that may be changed due to [later] developments.” Id. at 139. We therefore allow a party to proceed to appeal through a stipulated judgment only when the case is effectively dismissed by a “final ruling” on the appealed issue. Id. To hold otherwise would only encourage “piecemeal appeals,” id. at 141, with litigants leapfrogging the district court at the first sign of trouble. The fact that litigants might prefer such shortcuts is of no moment. One can surely imagine situations in which litigants might be discouraged by negative comments from a district judge during an early hearing on a purely legal question, or even where a litigant might dislike the initial district court draw based on unfavorable decisions issued by the assigned judge in other related cases. But those sorts of tentative setbacks are not enough to bypass the district court and the adjudicative process. By first requiring a “final” ruling on an issue, the Palmieri rule prevents attempts to “evade the final judgment rule.” Id. at 139.

For that same reason, New York cannot appeal the provisional findings in the district court’s order granting a preliminary injunction against it. As a threshold matter, there is little dispute that the district court’s preliminary

8

injunction was not a “final ruling” on the merits of preemption. Quite the opposite, “the findings of fact and conclusions of law made by a court granting a preliminary injunction are not binding at trial on the merits.” Univ. of Tex., 451 U.S. at 395. Indeed, we have long recognized that, with respect to preliminary injunction rulings, “[t]he judge’s legal conclusions, like his fact-findings, are subject to change after a full hearing and the opportunity for more deliberation.” Hamilton Watch Co. v. Benrus Watch Co., 206 F.2d 738, 742 (2d Cir. 1953) (emphasis added); see id. (“For a preliminary injunction . . . is, by its very nature, interlocutory, tentative, provisional, ad interim, impermanent, mutable, not fixed or final or conclusive, characterized by its for-the-time-beingness.”). If anything, “[a] decision on a preliminary injunction is, in effect, only a prediction about the merits.” Biediger v. Quinnipac Univ., 691 F.3d 85, 107 (2d Cir. 2012) (internal quotation marks omitted). Thus, just like the in limine ruling in Palmieri, the district court’s preemption analysis was strictly provisional and could not have “effectively dismissed” New York’s case. Palmieri, 88 F.3d at 140.

The majority nevertheless maintains that the district court’s ruling was an effective dismissal because the district court used “unequivocal” language when it said that the ABA “is conflict-preempted.” Maj. Op. at 14–15 (quoting NYSTA,

9

544 F. Supp. 3d at 282). But the tenor of the district court’s language in a preliminary injunction ruling is not enough to render the decision “final.” A strong “prediction” is still only a prediction. Biediger, 691 F.3d at 107. Whatever the tone of the district court’s order, those statements came in a preliminary injunction ruling and were necessarily provisional and “subject to change.” Hamilton Watch, 206 F.2d at 742.

In fact, the district court’s comments about the merits of preemption were, if anything, even less final than the evidentiary ruling in Palmieri, given that the preemption comments here were dicta. Because the district court needed only to find that the ABA was likely preempted in order to grant the preliminary injunction, any more definitive “assessment of the actual merits” of preemption was “dicta.” Fish v. Schwab, 957 F.3d 1105, 1140 (10th Cir. 2020) (internal quotation marks omitted); see also United States v. Hussein, 178 F.3d 125, 129 (2d Cir. 1999) (any finding “not necessary” to granting a preliminary injunction is “dictum”). Palmieri could at least argue that the evidentiary rulings were provisional holdings on admissibility. New York cannot even claim that here. Because the district court’s statements about the ultimate merits of preemption were dicta, they were not even a “decision” to begin with, let alone a final ruling. Carroll v. Lessee of

Carroll, 57 U.S. (16 How.) 275, 286–87 (1853) (“If [a point of law] might have been decided either way without affecting any right brought into question, then, according to the principles of common law, an opinion on such a question is not a decision.”).

This conclusion – that litigants cannot take stipulated judgment appeals from dicta in a provisional order – aligns with our other precedents on this issue. As far as I can tell, none of our past cases (including those relied on by the majority) authorized a stipulated judgment appeal after a district court cast doubt on a litigant’s case through provisional dicta. To the contrary, each of the appellants in those cases sustained an adverse holding that “effectively dismissed” his case. See, e.g., Ali, 719 F.3d at 89 (approving stipulated judgment appeal when the district court held in a partial summary judgment order that appellant’s proffered reading of a contract was foreclosed by the “express language” of the contract (internal quotation marks omitted)); Linde v. Arab Bank, PLC, 882 F.3d 314, 322 (2d Cir. 2018) (approving stipulated judgment appeal after appellant was found liable by a jury); Empire Volkswagen Inc. v. World-Wide Volkswagen Corp., 814 F.2d 90, 94 (2d Cir. 1987)

(approving stipulated judgment appeal of certain claims after district court granted summary judgment on those claims). 3 Attempting to reconcile its decision with Palmieri, the majority posits that the only jurisdictional defect in Palmieri was that the in limine rulings did not “plainly resolve a claim as a matter of law.” Maj. Op. at 16. But that is not what Palmieri actually said. We instead made clear that the in limine rulings could not support a stipulated judgment appeal for two separate reasons: (1) the in limine rulings did not resolve the claim “as a matter of law,” and (2) the in limine rulings were only tentative. Palmieri, 88 F.3d at 140. Indeed, we repeatedly stressed that the in limine rulings were insufficient because they were “subject to change” and not a “final ruling on admissibility.” Id. The majority’s best counter is that the

3In fact, Empire Volkswagen – one of our most-cited cases on stipulated judgment appeals – lends further support to the Palmieri rule against stipulated judgment appeals of provisional findings. There, the defendant moved for summary judgment on several of the plaintiffs’ claims, and the district court granted that motion in part. See 814 F.2d at 93. Even though several claims survived, the plaintiffs believed that the ruling “unduly limited” those claims by “excluding” an important theory of recovery. Id. at 93–94. Consequently, they voluntarily dismissed the surviving claims and attempted to appeal all of the claims from the resulting stipulated judgment. See id. at 94. Significantly, we held that the plaintiffs could appeal the claims that were dismissed at summary judgment but could not appeal the voluntarily dismissed claims. We concluded that, even if the partial summary judgment order limited those surviving claims – and cast doubt on their ultimate success – the district court’s order did not in fact “decide[]” those claims “adversely” to the plaintiffs. Id. It mattered not that the plaintiffs “interpret[ed] . . . [the] partial summary judgment order as an effective dismissal of [those claims].” Id. at 95. The only relevant inquiry was whether the district court had issued a holding that rejected those claims. See id. at 94 (“[W]e will consider[] only those portions of [the] order decided adversely to [the plaintiffs].”).

preliminary injunction ruling here was more definitive than usual, but again that goes nowhere, because “a preliminary injunction . . . is, by its very nature, interlocutory, tentative, provisional, . . . not fixed or final or conclusive, characterized by its for-the-time-beingness.” Hamilton Watch Co., 206 F.2d at 742 (emphasis added).

As a fallback, the majority pivots to the language of the stipulated judgment, in which the district court so-ordered the parties’ stipulation that, “[f]or the reasons given in the Court’s [preliminary injunction] order, the Court declares that [the ABA] is preempted by federal law.” J. App’x at 157. In the majority’s view, the district court “determined” that the ABA was preempted as a matter of law when it signed off on the parties’ stipulated language, which in turn was an effective dismissal of New York’s case. Maj. Op. at 17.

But the majority misconstrues the nature of stipulated judgments. A stipulated judgment cannot “effectively dismiss” a case for the simple reason that a district court does not “determine” anything when it so-orders a stipulated judgment. That is because a stipulated judgment “is not a ruling on the merits of the legal issue.” Langton v. Hogan, 71 F.3d 930, 935 (1st Cir. 1995); see also SEC v. Petro-Suisse Ltd., No. 12-cv-6221 (AJN), 2013 WL 5348595, at *3 (S.D.N.Y. Sept. 25,

13

2013) (“A consent decree is ‘not a ruling on the merits.’” (quoting Langton, 71 F.3d at 935) (alterations omitted)). Instead, a consent judgment is the “result of private bargaining,” Lipsky v. Commonwealth United Corp., 551 F.2d 887, 894 (2d Cir. 1976), that “normally embodies a compromise” in which “the parties each give up something they might have won had they proceeded with the litigation,” Barcia v. Sitkin, 367 F.3d 87, 90 (2d Cir. 2004) (quoting United States v. Armour & Co., 402 U.S. 673, 681 (1971)). In other words, the entry of a stipulated judgment merely invites the district court to sign off on a compromise that the parties reached on their own accord.

Because the language in the stipulated judgment was the product of “consent” rather than a “decision on the merits,” the district court could not have effectively dismissed New York’s case merely by granting the stipulated judgment. HS Equities, Inc. v. Hartford Accident & Indem. Co., 609 F.2d 669, 674 n.8 (2d Cir. 1979) (internal quotation marks omitted). Even though the stipulated judgment contained language declaring that the ABA was preempted, that language was not a finding or a determination by the district court. Indeed, the preemption “declar[ation]” appeared in a portion of the stipulated judgment that was “stipulated and agreed” to by the parties (as opposed to a finding that the

14

district court had to make on its own). J. App’x at 157. The majority’s only response is to suggest that the district court’s “adoption” of the stipulated language reflected the “finality” of the “legal holding” from its preliminary injunction order. Maj. Op. at 17–18. But as already discussed, the district court did not “adopt” or “determine” anything in the stipulated judgment, nor was its earlier finding on preemption “final” or even a “holding.” The district court merely signed off on a compromise that the parties (not the court) reached about the meaning of provisional dicta that appeared in an earlier order. That is not enough to establish finality.

To be clear, none of this means that New York was required to toil in the district court until the conclusion of a trial on the merits. New York could have pursued its interlocutory appeal of the preliminary injunction under 28 U.S.C. § 1292(a)(1) and asked this Court to dissolve it. Alternatively, it could have moved to consolidate the preliminary injunction hearing with an expedited trial on the merits under Rule 65(a)(2), which would have triggered an earlier merits ruling (and with it, an earlier appeal). Better yet, New York could have invited the district court to enter summary judgment against it sua sponte – which, unlike the

15

stipulated judgment, would have required the district court to make “an actual adjudication” on preemption. Lipsky, 551 F.2d at 893.

The majority says it was fine to skip those steps – and to “accelerate[]” the appeal – because it would be “pragmatic.” Maj. Op. at 4, 15. But our “jurisdiction . . . does not entail an assessment of convenience.” Wachovia Bank v. Schmidt, 546 U.S. 303, 316 (2006). Quite the opposite, we enforce our jurisdictional rules “strictly,” Muskrat v. United States, 219 U.S. 346, 356 (1911), and this case illustrates why. By abandoning Palmieri’s teachings, we give the greenlight to “piecemeal appeals.” Palmieri, 88 F.3d at 141. Like the parties here, litigants will forego the relief available under Section 1292(a)(1) – dissolution of a preliminary injunction – to proceed straight to a merits appeal through a stipulated judgment. In limine rulings will invite more of the same. By the majority’s logic, litigants may turn to stipulated judgments merely because a judge makes critical remarks during oral argument or at a premotion conference. There may be worthy occasions for a stipulated judgment appeal, but a district court’s provisional dicta is not one of them.

B. The Stipulated Judgment Appeal Circumvents Preauthorized Rules On Interlocutory Appeals.

In addition to lacking the finality required under Palmieri, the stipulated judgment also runs afoul of the Supreme Court’s decision in Microsoft v. Baker because it was procured by subverting the established regime for interlocutory appeals.

In Microsoft, the Supreme Court held that parties cannot use stipulated judgments to circumvent interlocutory appeal rules that otherwise would foreclose their appeal. See 582 U.S. at 37. There, the plaintiffs brought a putative class action and moved to certify it. Id. at 33. After the district court denied that motion, the plaintiffs sought discretionary interlocutory review under Federal Rule of Civil Procedure 23(f), a special provision under which a plaintiff (or a defendant) can ask the court of appeals to immediately review a denial (or a grant) of class certification. Id. at 34. When the Ninth Circuit declined to hear the appeal, the plaintiffs endeavored to force a mandatory appeal through a stipulated judgment. Specifically, they moved to dismiss their case with prejudice, explaining that once the district court entered final judgment they would then “appeal the order striking their class allegations.” Id. at 35 (alterations and internal quotation marks omitted). As requested, the district court granted the plaintiffs’

stipulated motion to dismiss and directed entry of final judgment. The plaintiffs then appealed the class certification order, arguing that they were appealing from a final judgment under section 1291 – and that the appeals court now had to hear their appeal of the class certification denial. See id. The Ninth Circuit agreed that it had jurisdiction to consider the appeal under section 1291, found that the district court had abused its discretion in striking the class allegations, and remanded the case to the district court for further proceedings on the merits. See id. at 35–36.

The Supreme Court granted certiorari on the jurisdictional question and held that the stipulated judgment was not final – and thus not appealable – under section 1291. See id. at 37. Significantly, the Court reasoned that the judgment could not be final because the plaintiffs had procured it in a bid to “subvert[] the final judgment rule” and the interlocutory review process Congress (in tandem with the Rules Committee) had established. Id. Indeed, Rule 23(f) prescribed a “discretionary regime” under which litigants could ask courts of appeals to review adverse class certification decisions. Id. at 39. But after the Ninth Circuit exercised that discretion and declined to review the district court’s initial certification denial, the plaintiffs sought to force the Ninth Circuit to hear their appeal anyway, even though the established interlocutory rules allowed only for discretionary appeals.

See id. at 40. In other words, the plaintiffs had sought to use a stipulated judgment to manufacture appellate rights (there, mandatory appeals) that neither Congress nor the Rules Committee had preauthorized. Therefore, even though the stipulated judgment was “technical[ly]” compliant – in that it resolved all of the plaintiffs’ claims and left nothing else for the district court to do – it still could not be truly final. Id. at 41 (“[Section] 1291’s firm final-judgment rule is not satisfied whenever a litigant persuades a district court to issue an order purporting to end the litigation.”).

Significantly, Microsoft did not purport to limit this rule – that litigants cannot use stipulated judgments to subvert established interlocutory rules – to class certification appeals. See Trendsettah USA v. Swisher Int’l, Inc., 31 F.4th 1124, 1132 (9th Cir. 2020) (explaining that Microsoft applies when there are “similar statutory restrictions [to Rule 23(f)] that would be adversely affected by permitting voluntary dismissal of claims with prejudice”). Indeed, we ourselves have extended Microsoft to another context in holding that litigants cannot use stipulated judgments to subvert the interlocutory rules on orders deciding motions to compel arbitration. See Bynum v. Maplebear Inc., 698 F. App’x 23, 24 (2d Cir. 2017). As we explained, Congress provided a special mechanism in 9 U.S.C.

§ 16 under which a defendant can immediately appeal an order denying its motion to compel arbitration. Yet Congress provided no such avenue for orders granting those motions. We therefore barred plaintiffs from using stipulated judgments to engineer an appeal of an otherwise unappealable interlocutory order sending plaintiffs’ claims to arbitration. See id. (citing Microsoft, 582 U.S. at 27–28). Other circuits are in accord. See Keena v. Groupon, Inc., 886 F.3d 360, 365 (4th Cir. 2018) (reaching the same result as Bynum under Microsoft); Langere v. Verizon Wireless Servs., LLC, 983 F.3d 1115, 1122 (9th Cir. 2020) (same).

Microsoft thus sets forth a broad rule: whenever Congress or the Rules Committee has preauthorized the right to appeal specific interlocutory orders, a litigant may not employ a stipulated judgment to seize additional appellate rights beyond those preauthorized avenues. If the interlocutory rules provide for only discretionary review of certain orders, then litigants cannot exploit stipulated judgments to secure mandatory review. And if the rules authorize interlocutory review only of orders denying a given motion, then litigants cannot resort to such tactics to obtain appellate review of orders granting those motions. A district court’s entry of an “actual final judgment” is of no moment if that final judgment

20

was procured in a bid to subvert the preapproved interlocutory rules. Microsoft, 582 U.S. at 40 (emphasis and internal quotation marks omitted).

Because New York used a stipulated judgment to expand its preauthorized appellate rights, Microsoft bars our appellate jurisdiction here. Once New York was preliminarily enjoined, it had one preauthorized appellate right: to seek dissolution of the preliminary injunction under section 1292(a)(1). See 28 U.S.C. § 1292(a)(1) (permitting interlocutory appeal of orders “granting . . . injunctions”). Had it taken this route, New York could have argued that the district court abused its discretion in granting the preliminary injunction under the familiar four-factor test; if we agreed, we would then dissolve the injunction and send the case back to the district court for continued litigation on the merits of preemption. See Univ. of Tex., 451 U.S. at 392 (listing the discretionary four-factor test for granting a preliminary injunction). But rather than take that narrow appeal, New York used a stipulated judgment to appeal the ultimate merits of preemption right away – that is, by asking us to issue a “final resolution” on whether the ABA is preempted as a matter of law. Id. That is a “significantly different” inquiry than an appeal seeking dissolution of an injunction under section 1292(a)(1). Id. There is thus no escaping it: section 1292(a)(1) did not preauthorize New York to appeal the

21

ultimate merits of preemption, yet New York has done so anyway through a stipulated judgment.

That is precisely what Microsoft disallowed. And just as in Microsoft, New York’s gambit upsets the “careful calibration” of section 1292(a)(1). 582 U.S. at 31. When Congress passed this provision, it authorized interlocutory appeals of preliminary injunctions “in order to prevent the injustice of burdening a party with a manifestly erroneous decree while the ultimate merits of a dispute are being litigated.” Indep. Party of Richmond Cnty. v. Graham, 413 F.3d 252, 256 (2d Cir. 2005) (emphasis added). In other words, Congress provided a limited appellate right to challenge only the injunction, so that a defendant would not be burdened by an erroneous restraint while it litigated the merits before the district court. If Congress had also desired for enjoined defendants to appeal the “ultimate merits” right away, then it would have authorized as much in section 1292(a). Id. Congress did no such thing, and that alone should foreclose New York’s attempt to secure that appellate right by stipulated judgment here.

For its part, the majority suggests that Microsoft does not apply because we have discretion (under our “pendent appellate jurisdiction”) to reach the merits when we hear an interlocutory appeal of an injunctive order under section

22

1292(a)(1). See San Filippo v. U.S. Tr. Co. of N.Y, 737 F.2d 246, 255 (2d Cir. 1984). 4 But that makes this case more like Microsoft, not less. As already discussed, Microsoft bars parties from using a stipulated judgment appeal to convert a discretionary right to appeal into a mandatory one. See 582 U.S. at 31–32 (explaining that Rule 23(f) gives appellate courts discretion to accept an appeal of a class certification denial and rejecting plaintiffs’ attempt to force an appeals court to hear such an appeal). That is essentially what New York has done here. If it had appealed the preliminary injunction under section 1292(a)(1), then we would have had limited discretion to address the ultimate merits of preemption. But because New York appeals on the basis of its stipulated judgment, it now contends that we must address the ultimate merits of preemption, thereby diminishing the discretion of the Court while enhancing its own. There is no meaningful distinction between what the parties have done here and what the parties did in Microsoft. In both cases the parties used a stipulated judgment appeal to secure

4 To be clear, we can exercise this discretionary power in contexts beyond interlocutory appeals of injunctions; as a general matter, “once we have taken jurisdiction over one issue in a case, we may, in our discretion, consider otherwise nonappealable issues in the case as well, where there is sufficient overlap [between] the appealable and nonappealable issues.” San Filippo, 737 F.2d at 255 (alterations and internal quotation marks omitted).

greater appellate rights than those preauthorized by Congress. As the Supreme Court made clear in Microsoft, that is not permitted. II. The ABA Is Preempted By Federal Law.

Although the lack of appellate jurisdiction should, by itself, be dispositive and compel dismissal of this appeal, I write briefly to respond to the majority’s resolution of the merits question concerning federal preemption of the ABA. To my mind, our precedents make clear that the ABA is both field- and conflict- preempted by federal law.

First, the ABA is field-preempted because the Communications Act preempts all rate regulation of interstate communication services. By its text, the Communications Act grants the FCC authority over “all interstate” communication services – save for a limited set of state-law prohibitions – while leaving to the states the power to regulate intrastate communications. 47 U.S.C. § 152(a)–(b) (defining the interstate and intrastate division); id. § 414 (preserving a limited set of state common-law rules). Thus, the Act prescribes that the FCC has exclusive authority over interstate communications, except for certain areas like consumer protection where states have traditionally exercised power. See, e.g., Head v. N.M. Bd. of Exam’rs in Optometry, 374 U.S. 424, 443–44 (1963) (explaining

24

that the “savings clause” in section 414 preserved state power to regulate interstate radio advertisements). Because rate regulation was not one of those traditional spheres of state authority, only the FCC retains the authority to regulate rates of interstate communications. 5 Indeed, we held as much in Ivy Broadcasting Co. v. American Telephone & Telegraph Co., 391 F.2d 486, 490–91 (2d Cir. 1968). There, we explained that both the Communications Act and its predecessor (the Mann-Elkins Act) manifested “an intent on the part of Congress to occupy the field to the exclusion of state law,” including with respect to the “rates” charged. Id. (internal quotation marks omitted). Though the majority asserts that Ivy Broadcasting meant to say that this preemption covered only the rates of Title II common carriers, we have not so limited Ivy Broadcasting when we have cited it in the intervening decades. See, e.g., Glob. NAPs, Inc. v. Verizon New England, Inc., 454 F.3d 91, 102 n.10 (2d Cir. 2006) (citing Ivy Broad., 391 F.2d at 491) (finding that a state regulatory board had

5The majority offers scant support for its claim that states have historically regulated the rates of interstate communications. See Maj. Op. at 28–29. It offers only an article noting that eleven states oversaw rate regulation of cable during the 1970s. But limited activity in twenty percent of the states is far from a meaningful tradition. Moreover, at the time of that rate regulation, cable was “essentially a local business,” where local operators broadcast to small surrounding regions. TV Pix, Inc. v. Taylor, 304 F. Supp. 459, 463 (D. Nev. 1968). That is quite unlike the modern internet, which virtually always involves interstate communications even for the most routine tasks. I therefore do not see a meaningful tradition of such rate regulation at the state level.

“narrowly sidestepped encroachment on the FCC’s jurisdiction to set rates on interstate communications” without limiting these statements to Title II).

The structure of the Communications Act confirms its preemptive scope.

When Congress defined the FCC’s authority in section 152, it used language – contrasting “interstate” versus “intrastate” authority,” 47 U.S.C. § 152(a)–(b) – that mirrored other statutes where Congress conferred exclusive federal authority. For instance, Congress granted the Federal Energy Regulatory Commission (“FERC”) exclusive authority over interstate electricity sales when it provided that a federal statute “shall apply to the transmission of electric energy in interstate commerce,” but not to “the transmission of electric energy in intrastate commerce.” 16 U.S.C. § 824(b)(1); see Hughes v. Talen Energy Mktg., LLC, 578 U.S. 150, 154 (2016). Congress also used such language in granting FERC “exclusive jurisdiction” over interstate natural gas sales. Scheidewind v. ANR Pipeline Co., 485 U.S. 293, 300–01, 308 (1988); see 15 U.S.C. § 717(b)–(c) (providing that the 1938 Natural Gas Act “shall apply to the transportation of natural gas in interstate commerce” but not to gas sales occurring “within” a state). By employing the same structure here, Congress likewise granted the FCC exclusive domain over rate regulation of interstate communications.

Put succinctly, in passing the Communications Act, Congress enacted a “federal law [that] occupies [the] field of [rate] regulation so comprehensively that it has left no room for supplementary state regulation.” Murphy v. Nat’l Collegiate Athletic Ass’n, 584 U.S. 453, 479 (2018) (internal quotation marks omitted). Because the ABA intrudes into that field, it is preempted, and its enforcement should be enjoined.

Second, the ABA is conflict-preempted because it would “frustrate the purposes” of the FCC’s 2018 decision to reclassify broadband as a Title I service. SPGGC LLC v. Blumenthal, 505 F.3d 183, 189 (2d Cir. 2007). For the purposes of conflict preemption, “[f]ederal regulations have no less preemptive effect than federal statutes.” Id. at 188 (internal quotation marks omitted). Thus, we need not focus on whether Congress intended to “supersede state law” so much as whether the agency meant to do so in issuing the regulations. Fid. Fed. Sav. & Loan Ass’n v. de la Cuesta, 458 U.S. 141, 154 (1982).

Here, there is little doubt that the FCC intended to preempt state laws that, like the ABA, imposed ex ante rate regulation on broadband. Even when the FCC briefly reclassified broadband as a Title II telecommunications service in 2015, it explained that “we do not and cannot envision adopting new ex ante rate

27

regulation of broadband [i]nternet access in the future.” 30 FCC Rcd. 5601, ¶ 451 (2015); see also id. ¶ 382 (“There will be no rate regulation.”). And in 2018, when the FCC returned broadband to its traditional classification as a Title I information service, the agency explained that its decision was driven by “concerns” that even the possibility of “rate regulation” attendant to Title II common carriage status “ha[d] resulted” in “untenable social cost[s] in terms of foregone investment and innovation.” 33 FCC Rcd. ¶¶ 87, 101. To that end, the FCC’s order stated its intent to “end utility-style regulation of the Internet in favor of . . . market-based policies” and a “light-touch” regulatory framework. Id. ¶¶ 2, 207.

In sum, the FCC’s actions and words evince an obvious “purpose[],”

SPGGC, 505 F.3d at 188, to foster openness and investment by sheltering broadband internet service from rate regulation. Because the ABA seeks to impose that very regulation, it is preempted.

For its part, New York insists that the FCC’s 2018 Order cannot preempt state law because the FCC has no power to regulate services when they are classified under Title I, as broadband is now. New York Br. at 50–51. In other words, New York suggests that because the FCC currently lacks power to regulate broadband rates, it cannot prevent states from regulating those rates either.

That argument fails to account for the obvious fact the FCC does have the power to regulate broadband. Just as it did in 2015, the FCC could reclassify broadband as a Title II service and impose ex ante rate regulations on it. Yet the FCC chose not to – a choice that “takes on the character of a ruling that no such regulation is appropriate or approved.” Ray v. Atl. Richfield Co., 435 U.S. 151, 178 (1978). Because “federal officials affirmatively [declined] to exercise their full authority” under the Communications Act in making a discretionary choice, “[s]tates are not permitted to use their police powers to enact such a regulation” in the resulting void. Id.

* * *

At bottom, we cannot hear a stipulated judgment appeal until the district court has issued a final ruling on the appealed issue. Nor can we entertain such an appeal when it is the product of an open attempt to subvert the interlocutory appellate rules. Because this appeal violates both of these precepts, I would dismiss it without reaching the merits of preemption. And even if I had to reach the merits, I would find that the ABA is preempted by federal law, as the majority’s cribbed reading of the Communications Act undermines the authority of the FCC to regulate interstate communications and emboldens states like New York to

29

impose costs on broadband internet service that extend well beyond their borders. For all these reasons, I respectfully dissent from the majority’s opinion.