Symmetry and (Network) Neutrality

Review of Daniel E. Walters, Symmetry’s Mandate: Constraining the Politicization of American Administrative Law, 119 Mich. L. Rev. 455 (2020).


In Symmetry’s Mandate, Daniel Walters constructs a compelling account of a difference between theory and practice in administrative law.1Daniel E. Walters, Symmetry’s Mandate: Constraining the Politicization of American Administrative Law, 119 Mich. L. Rev. 455 (2020). In recent years, several scholars have criticized doctrines, such as Chevron, granting agencies substantial power to make regulatory policy or adjudicate regulatory claims—seemingly, in their view, at the expense of Congress and the judiciary.2See, e.g., Philip Hamburger, Is Administrative Law Unlawful? (2014); see generally Gillian E. Metzger, The Supreme Court, 2016 Term—Foreword: 1930s Redux: The Administrative State Under Siege, 131 Harv. L. Rev. 1, 31–33 (2017) (describing such scholarship). And these scholars have found an audience in the courts, with some notable jurists issuing opinions echoing these critiques.3 See, e.g., Baldwin v. United States, 140 S. Ct. 690, 690–95 (2020) (mem.) (Thomas, J., dissenting from the denial of certiorari); Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149–58 (10th Cir. 2016) (Gorsuch, J., concurring). But Symmetry’s Mandate finds something that may have been lost in the transition from scholarship to doctrine. While many scholars’ arguments against administrative governance sound in separation-of-powers concerns—concerns that would seem to apply across both “regulatory” and “deregulatory” agency action—the doctrinal implementation of these ideas has demonstrated serious concern for possible agency overregulation while evincing little regard for possible agency underregulation.4Walters does, to be sure, criticize both courts and scholars, finding that the scholarly literature contains a “telling silence” regarding deregulatory errors that suggests an “operational[] bias[] toward remedying only one kind of error in the implementation of statutes.” Walters, supra note 1, at 486. In Walters’s terms, these cases seem to tolerate a range of Type II errors (deregulatory errors, or errors in which agencies “fail[] to operationalize statutes”) while remaking administrative law doctrine to address Type I errors (regulatory errors, or errors in which agencies take action beyond the statute’s intended scope).5Id. at 484. And he concludes that such asymmetry both undermines administrative law’s coherence and injects political and partisan considerations into the judicial review of agency action.6Id. at 516.

In this short Essay, I take the opportunity to highlight one further potential asymmetry that may yet emerge from the Supreme Court’s application of Chevron’s many doctrines. Drawing on then-Judge Kavanaugh’s dissental from the D.C. Circuit’s decision affirming network neutrality rules, I suggest that there is at least one vote on the Supreme Court—and perhaps more—for an asymmetric approach to the major questions doctrine.7U.S. Telecom Ass’n v. FCC (USTA II), 855 F.3d 381, 417–26 (D.C. Cir. 2017) (Kavanaugh, J., dissenting from the denial of rehearing en banc). Justice Kavanaugh’s express endorsement of an asymmetric approach to the major questions doctrine seems to have been largely overlooked in the still-growing literature regarding that doctrine. One notable exception includes Daniel Deacon, Judge Kavanaugh and “Weaponized Administrative Law, Yale J. on Regul.: Notice & Comment (July 11, 2018), [] Moreover, I demonstrate how asymmetry in this context is deeply irrational. As applied to network neutrality, the asymmetry has at least one of two effects. One, it might simply favor one large industry over another, subjecting one inter-sector wealth transfer to heightened scrutiny, while treating an analogous wealth transfer—in the opposite direction—deferentially. But the judiciary is not typically in the business of favoring one industrial sector over another.8See, e.g., Brett M. Kavanaugh, Fixing Statutory Interpretation, 129 Harv. L. Rev. 2118, 2120 (2016) (reviewing Robert A. Katzmann, Judging Statutes (2014)) (“[O]ne overarching goal is to make judging a neutral, impartial process in all cases . . . .”). Two, it subjects consumer-protection devices to increased regulatory scrutiny, thereby shifting the costs and burdens of overcoming a regulatory default to those entities—consumers—who can likely least afford to bear them. Hence, in more general terms, Justice Kavanaugh’s unbalanced approach to the major questions doctrine tends to undermine many of the values—accountability and expertise, among others—that agency policymaking has long served.9See Mila Sohoni, King’s Domain, 93 Notre Dame L. Rev. 1419, 1437–39 (2018).

I. Asymmetry and the Major Questions Doctrine

A. Asymmetry and Chevron

In Symmetry’s Mandate, Daniel Walters identifies an asymmetry in the development of a range of administrative law doctrines—scrutiny for agency nonenforcement (vis-à-vis judicial review of enforcement actions), or for orders denying petitions for rulemaking (vis-à-vis judicial review of agency rulemaking)—that seems to disfavor regulation for deregulation (and nonregulation) and thereby destabilizes administrative law’s coherence and continuity (and perhaps legitimacy) across presidential administrations. Among Walters’s examples is Chevron.10Walters, supra note 1, at 495–96. Chevron, of course, is a multipartite doctrine, and one typical hornbook account of Chevron follows three steps, numbered zero through two.11See Tejas N. Narechania, Defective Patent Deference, 95 Wash. L. Rev. 869, 880 (2020). But see Daniel J. Hemel & Aaron L. Nielson, Chevron Step One-and-a-Half, 84 U. Chi. L. Rev. 757, 759–60 (2017) (collecting descriptions of Chevron that range from one step to four). First, at Step Zero, a reviewing court considers whether Congress has granted the agency the authority to administer a statutory program by issuing decisions carrying the force of law.12See United States v. Mead Corp., 533 U.S. 218, 229–31 (2001); see also Thomas W. Merrill & Kristin E. Hickman, Chevron’s Domain, 89 Geo. L.J. 833, 873–89 (2001); Cass R. Sunstein, Chevron Step Zero, 92 Va. L. Rev. 187, 188 (2006). If so, then the court moves on to consider Chevron’s Steps One and Two:

When a court reviews an agency’s construction of the statute which it administers, it is confronted with two questions. First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress. If, however, the court determines Congress has not directly addressed the precise question at issue . . . the question for the court is whether the agency’s answer is based on a permissible construction of the statute.13Chevron U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 842–43 (1984) (footnotes omitted).

Walters focuses specifically on Chevron’s Step Two (requiring courts to assess the reasonableness of an agency’s construction and to defer to any reasonable agency construction), finding that “critics of Chevron seem skeptical that it applies with as much force when agencies commit [a potential] Type II error” through underregulation but seem less skeptical in cases presenting the possibility of Type I error through overregulation.14See Walters, supra note 1, at 495.

As an example, Walters emphasizes Justice Thomas’s recent reconsideration of his majority opinion in Brand X as proof of this asymmetry in practice.15Id. In Brand X, the Supreme Court affirmed the Federal Communications Commission’s decision to treat cable modem service—the broadband internet access service sold by cable companies—as beyond its regulatory powers.16Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967 (2005). But as the Commission has more recently exercised the flexibility accorded to it by Chevron to reregulate (and re-deregulate) broadband carriage by, say, imposing (and then later rescinding) network neutrality rules,17See, respectively, U.S. Telecom Ass’n v. FCC (USTA I), 825 F.3d 674 (D.C. Cir. 2016), and Mozilla Corp. v. FCC, 940 F.3d 1 (D.C. Cir. 2019). Justice Thomas has—consistent with his more recent criticisms of Chevron—signaled some willingness to revisit his own majority opinion in Brand X.18Baldwin v. United States, 140 S. Ct. 690, 690–91 (2020) (mem.) (Thomas, J., dissenting from the denial of certiorari) (“Although I authored Brand X, ‘it is never too late to “surrende[r] former views to a better considered position” ’ . . . . My skepticism of Brand X begins at its foundation—Chevron deference.” (first alteration in original) (quoting South Dakota v. Wayfair, Inc., 138 S. Ct. 2080, 2100 (2018) (Thomas, J., concurring))). In Walters’s view, this inconsistency seems outcome-driven: He suggests that “one might . . . reasonably question whether Justice Thomas’s new thinking about the limits of interpretive maneuvers by agencies will extend to cases,” like the Commission’s decision to re-deregulate broadband carriers, “that introduce [the possibility for] Type II error.”19Walters, supra note 1, at 495.

I agree that Justice Thomas has signaled a deep skepticism of Chevron and Brand X, and that his skepticism may well be more likely to manifest in a decision addressing a putative Type I error.20But cf. Lipschultz v. Charter Advanced Servs. (MN), LLC, 140 S. Ct. 6, 7 (2019) (mem.) (Thomas, J., concurring in the denial of certiorari) (finding it “doubtful [that] a federal policy . . . of nonregulation . . . is ‘Law’ for purposes of the Supremacy Clause” and, hence, preemptive of state regulation). For now, however, I aim to shift focus to a different aspect of Chevron and to a different member of the Court—namely, to Step Zero and to Justice Kavanaugh.

B. Asymmetry and the Major Questions Doctrine

As noted, I focus here on Chevron’s Step Zero and, in particular, the major questions doctrine.21Sunstein, supra note 12, at 236 (describing the major questions doctrine as a Step Zero doctrine). There is, to be sure, some controversy over the major questions doctrine’s true place in Chevron’s framework. See, e.g., Christopher J. Walker, Toward a Context-Specific Chevron Deference, 81 Mo. L. Rev. 1095, 1101 (2016). I can safely set aside such questions for my present purposes. My analysis does not turn on the doctrine’s precise place in Chevron’s larger framework. While there is some dispute over the doctrine’s precise contours, its basic premise is simple: Some matters are so important—so major—that Congress would not have delegated policymaking authority to an agency over such matters without saying so explicitly. In MCI v. AT&T, one of the earliest examples of the doctrine, the Court concluded that the Commission’s power to “modify” rate regulations did not encompass the power to eliminate them entirely, reasoning that it was “highly unlikely” that Congress would delegate to an agency the power to decide whether an entire industry should or should not be subject to rate regulation.22MCI Telecomms. Corp. v. AT&T, 512 U.S. 218, 225–26, 231 (1994). Congress would later grant the FCC such power in express terms. See Daniel T. Deacon, Administrative Forbearance, 125 Yale L.J. 1548, 1569 (2016) (“The FCC’s decades-long efforts to alter the Communications Act’s requirements in the face of newly emerging competition—culminating in the Commission’s defeat in MCI Telecomms. Corp. v. AT&T—is what initially prompted Congress to enact the 1996 Telecommunications Act’s forbearance provision.” (footnote omitted)). More recently, Chief Justice Roberts explained in King v. Burwell that if “Congress wished to assign” a “question of deep ‘economic and political significance’ ” “to an agency, it surely would have done so expressly.”23576 U.S. 473, 486 (2015) (quoting Util. Air Regul. Grp. v. EPA, 573 U.S. 302, 324 (2014)). Finding no such express delegation of policymaking power, the Court declined there to apply Chevron’s framework.24King, 576 U.S. at 486 (“This is not a case for the IRS. It is instead our task to determine the correct reading [of the statute].”).

Can the status of a question—as either major or not—turn on its answer? Based on his views as a member of the Courts of Appeals for the D.C. Circuit, Justice Kavanaugh seems to think so. Explaining how this is so requires a brief detour into the regulatory history of network neutrality (or, colloquially, net neutrality).

1. Network Neutrality

As I noted above, the Commission decided, in 2002, to treat cable modem–based internet access as beyond most of its regulatory powers. It accomplished this by reclassifying the service an “information service.”25Inquiry Concerning High-Speed Access to the Internet over Cable and Other Facilities, 17 FCC Rcd. 4798, 4802 (2002) [hereinafter Cable Modem Order] (Declaratory Rule and Notice of Proposed Rulemaking). The Commission’s regulatory powers under the Telecommunications Act of 1996 depend, in part, on the type of service it seeks to regulate—“information service” or “telecommunications service,” among other regulatory classifications. The differences between the telecommunications service and the information-service designations are vast.2647 U.S.C. § 153(24), (50), (53). “Classified as a []telecommunications service,[] a communications offering may be subject to the full panoply of the [Commission’s] regulatory power—rate regulation, service specification, and more.”27Tejas N. Narechania, The Secret Life of a Text Message, 120 Colum. L. Rev. F. 197, 209 (2020). By contrast, information services are largely beyond the Commission’s reach.28See Comcast Corp. v. FCC, 600 F.3d 642, 649 (D.C. Cir. 2010).

The Commission’s decision to choose the information-service designation was premised, however, on the mistaken impression that, such deregulation notwithstanding, it would nevertheless retain some ancillary powers to require broadband carriers to comply with network neutrality rules—rules that, among other things, prevent carriers from blocking or unreasonably inhibiting consumers’ access to lawful internet content.29See, e.g., Appropriate Framework for Broadband Access to the Internet over Wireline Facilities, 20 FCC Rcd. 14986 (2005) [hereinafter Internet Freedoms Policy Statement] (Policy Statement). In short, network neutrality rules require that broadband carriers—companies like AT&T, Comcast, and Verizon—treat like internet traffic alike, without prioritizing favored providers (including affiliated companies), blocking disfavored ones, or otherwise demanding fees for delivering internet content to its destination.30See generally Tim Wu, Network Neutrality, Broadband Discrimination, 2 J. on Telecomms. & High Tech. L. 141 (2003). Former Federal Communications Commission Chairman Michael Powell explained that such rules offered “an insurance policy against the potential rise of abusive market power by vertically integrated providers.”31Michael K. Powell, Preserving Internet Freedom: Guiding Principles for the Industry, 3 J. on Telecomms. & High Tech. L. 5, 12 (2004). Moreover, then-Chairman Powell explained that these rules would “empower American consumers” and give “[p]ower to the [p]eople” by ensuring that they remained free from broadband carrier “interference.”32Id. at 6–7, 11. But the Commission was wrong about its power to enforce such network neutrality rules under the information-service paradigm.33E.g.Comcast Corp., 600 F.3d 642; Verizon v. FCC, 740 F.3d 623 (D.C. Cir. 2014).

Hence, in 2015, the Commission decided to revisit its classification decision, concluding that broadband carriage was now better characterized as a telecommunications service and that, given this new classification, the Commission might finally promulgate enforceable network neutrality obligations in service of these competition and consumer-protection goals.34Protecting and Promoting the Open Internet, 30 FCC Rcd. 5601, 5644, 5763–65 (2015) [hereinafter Open Internet Order] (Report and Order on Remand, Declaratory Ruling, and Order). It was correct: The Commission’s decision to reclassify the broadband carriage also conferred the power to enforce network neutrality rules against the newly designated telecommunications services that it could not apply to information services.35U.S. Telecom Ass’n v. FCC (USTA I), 825 F.3d 674 (D.C. Cir. 2016).

2. Network Neutrality and the Major Questions Doctrine

It may seem odd that the Commission can so easily change its mind about the statutory classification of a service, thereby dramatically altering the scope of its regulatory powers over that service. But this is precisely the approach that the Supreme Court endorsed in its 2005 Brand X decision, reasoning (consistent with Chevron itself) that agencies could best draw upon specialized technical and legal expertise and be held to account for their policy choices by political leaders and the voting public.36Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 981 (2005) (explaining that agencies “must consider varying interpretations and the wisdom of its policy on a continuing basis,” including “in response to changed factual circumstances, or a change in administrations” (quoting Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 863–64 (1984))); see id. at 1002–03 (explaining that “[t]he questions the Commission resolved in the order under review involve a ‘subject matter [that] is technical, complex, and dynamic’ ” and that “[t]he Commission is in a far better position to address these questions” than the courts (second alteration in original) (quoting Nat’l Cable & Telecommunications Ass’n Inc. v. Gulf Power, Co., 534 U.S. 327, 339 (2002))). As noted above, Brand X affirmed the Commission’s Cable Modem Order, which placed cable modem service largely beyond its regulatory ambit.37Brand X, 545 U.S. at 976.

Before the Cable Modem Order, the Commission treated most broadband carriers—consistent with both agency and judicial precedent—as telecommunications service providers.38See Appropriate Framework for Broadband Access to the Internet over Wireline Facilities, 20 FCC Rcd. 14853 (2005) (Report and Order and Notice of Proposed Rulemaking) (explaining that DSL had been traditionally treated as a telecommunications service before this reclassification order); see also Brand X, 545 U.S. at 1000 (noting the regulatory classification of DSL). But broadband internet access offered over cable television wires was new and thus did not obviously fall inside any existing regulated service.39See Nat’l Cable & Telecomms. Ass’n, Inc. v. Gulf Power Co., 534 U.S. 327, 347 (2002) (Thomas, J., concurring in part and dissenting in part) (calling on the Commission to make a classification decision, and concluding that cable-based internet access might even fall within the “cable service” classification). But see Tejas N. Narechania & Tim Wu, Sender Side Transmission Rules for the Internet, 66 Fed. Commc’ns L.J. 467, 488 (2014) (concluding that the better interpretative view is one that treats broadband carriage as a telecommunications service). I continue to believe that this is the better interpretative view, though I agree that the statutory definitions of these classifications are ambiguous and merit deference under Chevron. Hence, finding sufficient ambiguity in the statutory definitions of these competing services, the Court agreed in Brand X that the Commission may reasonably conclude that broadband carriage was an “information service” largely immune to Commission regulation.40Brand X, 545 U.S. at 987–89.

As I noted above, the Commission sought to undo this decision one decade later. The Commission’s reasoning rested heavily on intervening changes in the market for broadband carriage41Open Internet Order, supra note 34, at 5743–44.—and also on Brand X. The Supreme Court’s opinion there, after all, seemed to suggest that either classification—information service or telecommunications service—might be a reasonable interpretation of an ambiguous statute.42Brand X, 545 U.S. at 981. And in view of new market conditions, the Commission concluded that the telecommunications service classification was a better fit.43Open Internet Order, supra note 34, at 5745.

The broadband carriers’ challenge to the Commission’s Order setting out this new classification and promulgating concomitant network neutrality rules failed on Brand X’s terms.44 U.S. Telecom Ass’n v. FCC (USTA I), 825 F.3d 674, 700–01 (D.C. Cir. 2016). In USTA v. FCC, the U.S. Court of Appeals for the D.C. Circuit explained that the Supreme Court’s decision in Brand X practically compelled a decision in favor of the Commission: Brand X concluded that the statute was ambiguous; Brand X suggested that either classification was reasonable; and the Commission’s factual analysis of new market conditions adequately explained its revised conclusion.45Id.see also Brand X, 545 U.S. at 1001 n.4 (“Any inconsistency [over time] bears on whether the Commission has given a reasoned explanation for its current position, not on whether its interpretation is consistent with the statute.”).

But, on a petition for rehearing en banc, then-Judge Kavanaugh dissented, suggesting that Brand X did not supply the relevant standard. Instead, Kavanaugh contended that the relevant question arose at Chevron’s Step Zero. Chevron, recall, does not apply to questions of deep “economic and political significance.”46U.S. Telecom Ass’n v. FCC (USTA II), 855 F.3d 381, 417 (D.C. Cir. 2017) (Kavanaugh, J., dissenting from the denial of rehearing en banc). And Kavanaugh contended that, though the 2002 Cable Modem Order at issue in Brand X entailed mere “light-touch regulation” that constituted only “an ordinary rule, not a major rule,” the Commission’s 2015 Open Internet Order was a “major rule” not expressly authorized by Congress.47Id. at 425–26 n.5. Importantly, Judge Kavanaugh’s characterization of the Commission’s 2002 Order is rather incomplete, overlooking the Commission’s (mistaken) belief that it could continue to enforce network neutrality norms under the information-service classification it then promulgated. That is, even in 2002, the Commission understood its rule as entailing “common-carrier regulation” of the exact same “vast economic and political significance” as the 2015 Order. Compare id. (suggesting that the Cable Modem Order was an example of mere “light-touch regulation”), with Internet Freedoms Policy Statement, supra note 29 (explaining that the Commission intended to enforce network neutrality obligations on broadband carriers classified information-services providers). And the Court’s 2005 decision in Brand X predates any other decision suggesting that the Commission could not enforce these norms under the information-service paradigm. See Comcast Corp. v. FCC, 600 F.3d 642 (D.C. Cir. 2010); Verizon v. FCC, 740 F.3d 623 (D.C. Cir. 2014). Indeed, at the time Brand X was decided, the Commission had already enforced network neutrality norms against an internet access provider, settling, for $15,000, allegations that it had blocked access to a competing Voice Over IP telephony service provider. See Madison River Commc’ns, LLC, 20 FCC Rcd. 4295 (2005) (Order). Hence, though Chevron’s framework applied in Brand X, he contended that it could not sustain the Commission’s actions under review in USTA. In his view, the Commission’s network neutrality rules purported to regulate a significant aspect of the domestic economy (namely, broadband carriage), and agencies may issue such rules of economic significance only when clearly authorized to do so. So, the Commission was not clearly authorized to issue these rules: “Brand X’s finding of ambiguity by definition means that Congress has not clearly authorized the FCC to issue the net neutrality rule. And that means that the net neutrality rule is unlawful under the major rules doctrine.”48USTA II, 855 F.3d at 426 (Kavanaugh, J., dissenting from the denial of rehearing en banc).

In short, then-Judge Kavanaugh suggested that whether a question is major turns on the agency’s response. The Commission was asked, repeatedly, to resolve the status of broadband internet access.49See, e.g., Nat’l Cable & Telecomms. Ass’n v. Gulf Power Co., 534 U.S. 327, 347 (2002) (Thomas, J., concurring in part and dissenting in part). When it answered “information service,” the Supreme Court said that Chevron’s usual rules applied. But when it answered “telecommunications service,” Kavanaugh suggested that the classification question was too significant for deference to apply. Indeed, as noted, Kavanaugh expressly acknowledged this asymmetry, stating plainly that the major questions limitation on Chevron applies only when an agency offers a “regulatory” answer—but not, as was the case in Brand X, a putatively deregulatory one.

One might wonder whether it was a major step for the FCC to impose even light-touch “information services” regulation on Internet service providers. The answer is no . . . . The FCC’s light-touch regulation did not entail common-carrier regulation and was not some major new regulatory step of vast economic and political significance. The rule at issue in Brand X therefore was an ordinary rule, not a major rule. As a result, the Chevron doctrine applied, not the major rules doctrine. 50USTA II, 855 F.3d at 425 n.5 (Kavanaugh, J., dissenting from the denial of rehearing en banc). See supra note 47 (explaining that Judge Kavanaugh’s characterization is inaccurate, even on its own terms).

Stated in Walters’s terms, Justice Kavanaugh has suggested that the major questions doctrine may help to address a possible Type I error (here, the view that Congress did not mean for broadband carriage to fall within the Commission’s authority over telecommunications services) but not a Type II error (the view that Congress did mean for the service to continue to fall within the Commission’s authority over telecommunications services).51See Deacon, supra note 7.

II. Symmetry and Neutrality

Then-Judge Kavanaugh’s dissental in USTA suggests an additional example of asymmetry in administrative law, alongside the many others Walters catalogues in Symmetry’s Mandate. And this example, too, is significant for its normative implications, both as applied to network neutrality specifically and more generally. Communications scholars disagree over the import and effect of the Commission’s network neutrality rules. Some—even some network neutrality proponents—see the policy primarily as a wealth-allocation device, shifting profits from broadband carriers (companies like Comcast and Verizon) and assigning them, instead, to content providers (companies like Facebook, Google, and Netflix).52See infra notes 53–54. Others (myself included) see it as a competition-, democracy-, and speech-enhancing rule that, among other things, mitigates concerns over potentially anticompetitive and antidemocratic behavior while preserving the internet’s opportunities to advance both free speech and new innovation. But no matter how one views the network neutrality rule, then-Judge Kavanaugh’s asymmetric application of the major questions doctrine to the Commission’s decision seems to make little sense.

A. Network Neutrality as Wealth Transfer

I begin by adopting, arguendo, the view that network neutrality regulation is little more than a wealth-allocation device. Even some advocates of network neutrality rules have suggested that broadband carriers (like Comcast and Verizon) and content providers (like Facebook and Netflix) are “complements of each other in a hugely lucrative business, and the network neutrality debates are nothing more than an argument about how to share the proceeds.”53Yotam Harchol, Dirk Bergemann, Nick Feamster, Eric Friedman, Arvind Krishnamurthy, Aurojit Panda, Sylvia Ratnasamy, Michael Schapira & Scott Shenker, A Public Option for the Core, 2020 Proc. Ann. Conf. ACM Special Int. Grp. on Data Commc’n on Applications, Techs., Architectures, & Protocols for Comput. Commc’n 377, 388. Critics have likewise characterized network neutrality as a mere “battle over the allocation of economic rents.”54See, e.g.Why Net Neutrality Matters: Protecting Consumers and Competition Through Meaningful Open Internet Rules: Hearing Before the S. Comm. on the Judiciary, 113th Cong. 5 (2014) (testimony of Jeffrey A. Eisenach, visiting scholar, Am. Enter. Inst.), []; see also Tim Wu & Christopher S. Yoo, Keeping the Internet Neutral? Tim Wu and Christopher Yoo Debate, 59 Fed. Commc’ns L.J. 575, 589–90 (2007) (Christopher Yoo contending that “a large part of the network neutrality debate can be viewed as nothing more than an intramural fight between the large content providers (like Google) and the large network providers (like Verizon and Comcast),” that “network neutrality would affect the way that the resulting profits would be divided between the Googles and the Verizons of the world,” and that “[a]lthough the division of profits between network providers is crucial to those companies and their shareholders, it is not ultimately a policy issue”). Under this view, advocates and critics agree that bandwidth is a resource that someone must pay for. Comcast, for example, might bear the initial expenses of upgrading an internet access network and then make that additional bandwidth available to whoever needs it (say, Google) at no additional charge (over and above the subscription fees paid by Comcast’s retail consumers). Alternately, Google might find that its services are being delivered too slowly, and so it might pay Comcast to upgrade its facilities, reserving the additional capacity for itself (and thereby freeing the extant bandwidth for other applications, like Netflix). Hence, when bandwidth is scarce, either its suppliers (broadband carriers) or its consumers (say, content providers) might pay for new capacity. But network neutrality rules foreclose the latter route, on the theory, in part, that robust competition among content providers—competition unimpeded by the barriers that the carriers’ termination fees would present, or by practices favoring the carriers’ own affiliated services—will best grow consumer demand for such content as well as for the bandwidth necessary to deliver it. Supply will follow that demand.

Hence, network neutrality rules have the practical effect of assigning such infrastructure costs to broadband carriers: Whenever, say, Comcast chooses to upgrade its network facilities, it must do so out of its own coffers and recoup those expenses (if at all) from its own retail customers.

Likewise, the Commission’s decision to rescind network neutrality rules has the practical effect of assigning these infrastructure costs to content providers: Whenever, say, Netflix finds that it requires more capacity to deliver video content to consumers, Comcast (for example) can demand that Netflix pay for that capacity, and Netflix can recoup those expenses (if at all) from its own subscribers.

In short, if network neutrality rules are best viewed as a wealth-allocation policy, then any network neutrality policy choice is wealth allocative, no matter whether the rules exist or do not exist. It is as much a policy choice to rescind the rules as it is to promulgate them. If the rules exist, then content providers reap the rewards; but if the rules do not exist, then broadband carriers benefit.

Such an analysis, moreover, applies to a wide range of agency action. Agencies, after all, do not issue rules punitively but in order (in many cases) to assign costs to a cheap cost avoider (so that the rule’s benefits outweigh its costs).55See, e.g., Exec. Order No. 12,291, 46 Fed. Reg. 13,193 (Feb. 19, 1981). Consider, for example, a pollution control statute. An agency might issue a regulation imposing significant costs on, say, coal-fired power plants. Alternately, the agency might rescind (or decline to impose in the first instance) that same rule, thereby imposing the substantial costs of pollution on the public.56Cf. Boomer v. Atl. Cement Co., 257 N.E.2d 870, 871–72 (N.Y. 1970). No matter which path the agency chooses—nonregulation, regulation, or deregulation—the agency’s policy choice is of similar effect, imposing the costs of its decision on some affected population.

Against this backdrop, then-Judge Kavanaugh’s claim that the major questions doctrine applies to imposition of network neutrality rules (but not to their rescission) is odd. In either case, the Commission’s policy choice is a regulatory step of similar significance.57See U.S. Telecom Ass’n v. FCC (USTA II), 855 F.3d 381, 417 (D.C. Cir. 2017) (Kavanaugh, J., dissenting from the denial of rehearing en banc). Indeed, the scalar value of the economic consequence seems roughly the same, no matter whether its vector points in the direction of broadband carriers or content providers.58See e.g.Scalar and Vector Quantities, BBC: Bitesize,‌/bitesize/guides/z2b9hv4/revision/6 [] (defining scalar and vector quantities). If the rule is major, then it should be major no matter whether it imposes these rules or whether it rescinds them, no matter whether it imposes those costs on carriers or on content providers. Such symmetry draws from Step Zero’s basic premise: Step Zero (and the major questions doctrine) asks whether Congress has delegated certain policy choices to the agency. Assuming such a delegation—in this instance, assuming the power to choose between the telecommunications and information-service designations—the Step Zero inquiry is neutral as between the available choices. Once Congress has delegated power to the agency, it is up to the agency to give its best answer. In short, the major questions doctrine is about major policy choices, not major policy outcomes.59 See MCI Telecomms. Corp. v. AT&T, 512 U.S. 218, 231 n.4 (1994) (concluding that the FCC could not make a major deregulatory move without express direction from Congress). Given a major policy question, any outcome is major. And given an ordinary policy question, as in Brand X, any outcome is ordinary.

This, of course, was not then-Judge Kavanaugh’s view. He offered no rationale for a rule that operates, instead, at the level of the regulatory outcome, favoring, in this instance, one large industry (broadband carriers) over another (content providers). In my view, there isn’t one.60One alternative view—though not one endorsed by then-Judge Kavanaugh—is that the Court was wrong to defer to the Commission’s interpretation in Brand X. Though it is well outside the scope of this short Essay to undertake a complete defense of Chevron and Brand X, it suffices for now to note that, for various reasons of institutional competence and accountability—the very rationales that undergird the Chevron doctrine—Brand X (rather than Justice Kavanaugh’s dissental) supplies the better standard of review, no matter whether the agency takes a regulatory or deregulatory step. Indeed, this asymmetry seems, instead, to breach the ideal that such matters of “interpretation should be neutral, impartial, and predictable.”61One alternative view—though not one endorsed by then-Judge Kavanaugh—is that the Court was wrong to defer to the Commission’s interpretation in Brand X. Though it is well outside the scope of this short Essay to undertake a complete defense of Chevron and Brand X, it suffices for now to note that, for various reasons of institutional competence and accountability—the very rationales that undergird the Chevron doctrine—Brand X (rather than Justice Kavanaugh’s dissental) supplies the better standard of review, no matter whether the agency takes a regulatory or deregulatory step.

B. Network Neutrality as a Coordination Device

The defects in then-Judge Kavanaugh’s approach to the major questions doctrine loom even larger when taking a broader view of network neutrality’s purpose. After all, former Chairman Powell did not explain the need for network neutrality rules by way of a subsidy to content providers. Rather, as noted, those rules were meant to insure “against the potential rise of abusive market power by vertically integrated providers,” and moreover, to confer “[p]ower to the [p]eople” by freeing consumers from carrier “interference.”62Powell, supra note 31, at 7, 12.

Hence, the wealth-allocation view of network neutrality is rather incomplete. Rather, like other competition policy principles, network neutrality helps to prevent the aggrandization of power in vertically integrated communications providers by reallocating the coordination rights that are typically vested in such firms and by granting them to consumers, who may select the content and applications of their choice without regard to that content’s status as affiliated (or not) with the facilities owner.63See Sanjukta Paul, Antitrust as Allocator of Coordination Rights, 67 UCLA L. Rev. 378 (2020). In short, network neutrality limits the extent to which Comcast (for example) can favor its own proprietary video service at the expense of both external competitors and, importantly, consumers who prefer those outside providers. Without network neutrality rules, firms are free to favor affiliated entities at the expense of both putative competitors and consumer preferences. AT&T might, as a broadband carrier, grant various advantages to affiliated content providers (including, say, HBO).64 See Wireless Telecomms. Bureau, Fed. Commc’ns Comm’n, Policy Review of Mobile Broadband Operators’ Sponsored Data Offerings for Zero-Rated Content and Services (2017),​/DOC-342987A1.pdf []. This is so even where consumers would prefer to watch Netflix instead of HBO (an AT&T affiliate) or Peacock (Comcast’s service). In many regions, this is because broadband carriers’ service terms are not disciplined by competition: Many consumers only have access to one wireline broadband carrier.65See Tejas N. Narechania, Federal and State Authority for Broadband Regulation, 18 Stan. Tech. L. Rev. 456, 477 (2015). Moreover, even where competition exists, switching is not frictionless: It can be onerous to cancel service from one carrier and provision service from another. And so, even where competition exists, carriers may be able to impose incremental burdens on consumers that do not exceed switching costs.66See id.see also Lukasz Grzybowski, Maude Hasbi & Julienne Liang, Transition from Copper to Fiber Broadband: The Role of Connection Speed and Switching Costs, Info. Econ. & Pol’y, Mar. 2018, at 1. The Commission’s network neutrality rules helped to curb this “gatekeeper” power.67See Verizon v. FCC, 740 F.3d 623, 646–47 (D.C. Cir. 2014).

In total, in this two-sided market (where broadband carriers serve both content providers and retail consumers), the costs or benefits of a network neutrality–related policy decision are borne, as explained above, by broadband carriers and content providers—and also retail consumers.68Some have argued that the consumer effect is relatively small, since either Comcast or Netflix will raise consumers’ prices to account for necessary bandwidth investments. While that may be true, such a view assumes a competitive and relatively frictionless market. Broadband carriage is anything but. Hence, absent network neutrality protections, carriers can impose incremental costs on consumers, up to the costs of switching. Stated simply, any policy choice gives rise to important effects on all three of these interested constituencies.

This is notable for several reasons.

For one, network neutrality rules reduce the burdens that consumers might otherwise face to using their preferred internet-based services. Stated similarly, without network neutrality rules, carriers might nudge consumers to use services that they actually like less. If, for example, AT&T makes HBO free but Netflix costly, consumers may watch HBO, even if they’d rather watch Netflix’s content. And the lost value of the difference in preference presents a regressive burden on consumers; it is a greater proportional harm to the average retail consumer. Indeed, the absence of network neutrality rules may be felt most acutely by historically marginalized populations.69See Raymond M. Brescia, Dominance and Disintermediation: Subversive Stories and Counter-Narratives of Cooperation, 27 S. Cal. Interdisc. L.J. 429, 468 & n.238 (2018) (suggesting that network neutrality helps historically marginalized populations develop and distribute counter-narratives to a dominant frame, citing W. Kamau Bell, Opinion, Net Neutrality: Why Artists and Activists Can’t Afford to Lose It, N.Y. Times (Oct. 31, 2017)); see also John McMurria, From Net Neutrality to Net Equality, 10 Int’l J. Commc’n 5931, 5944–45 (2016) (suggesting that “stopping Internet providers from degrading functionality among disenfranchised populations” is one way to “challeng[e] [the] race-neutral thinking in law and policy” that has dominated the network neutrality discourse); see generally id. at 5939 (noting absence of minority-owned cable systems).

I do not, of course, mean to imply that the restoration of network neutrality rules would suddenly solve the problems faced by historically marginalized populations. Several scholars have explained how network neutrality is a second-order problem for the many communities that still lack internet access. See, e.g., Olivier Sylvain, Network Equality, 67 Hastings L.J. 443, 447 (2016); Rachel E. Moran & Matthew N. Bui, Race, Ethnicity and Telecommunications Policy Issues of Access and Representation: Centring Communities of Color and Their Concerns, 43 Telecomms. Pol’y 461, 471 (2019) (explaining that arguments “center[ing] around the role of Internet as a cure to the issues of representation and access . . . rest[] on an unsubstantiated and dangerous assumption that the first-level digital divide of broadband access no longer exists”). Rather, I mean only what I say: The problems that network neutrality rules help to solve may be felt most acutely by these communities whenever such rules are absent.

Hence, network neutrality rules vest decisional power in consumers. As described above, the carriers’ influence over internet traffic may allow them to distort market outcomes. To stay with the same example, AT&T can give HBO an advantage over Netflix, notwithstanding a true preference for Netflix—especially in regions where AT&T faces little to no competition, or where switching from AT&T is costly or difficult. But network neutrality rules return that decisional power to consumers, who otherwise have few mechanisms to make their preferences known. In short, network neutrality rules ensure that consumers have a voice in the market for internet-enabled services, where exit from the broadband-carriage market is impractical.70See generally Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (1970) (explaining consumer economic choices in terms of exiting the market or voicing preferences). Justice Kavanaugh’s suggestion, making it more difficult for the Commission to select a network neutrality–friendly regulatory classification, thus both is inconsistent with Step Zero’s focus on congressional intent to delegate decisional authority and also tends to “amplif[y]” the “well-known concerns regarding agency capture and the collective action problems that hinder organization around interests shared by the public at large.”71Daniel T. Deacon, Essay, Common Carrier Essentialism and the Emerging Common Law of Internet Regulation, 67 Admin. L. Rev. 133, 182 (2015). For a theoretical discussion of why courts should avoid promulgating approaches that, like the asymmetric approach proposed by Justice Kavanaugh, favor powerful entities over less powerful individuals, see generally Aaron Tang, Reverse Political Process Theory, 70 Vand. L. Rev. 1427 (2017).

Finally, network neutrality is a policy that has broad popular support across the ideological spectrum.72Amir Nasr, The American People Broadly Support Net Neutrality, New Am. (Oct. 23, 2017), []. But an asymmetric approach to the major questions doctrine would exert legal pressure on agencies to choose an interpretation that “favors its regulated entities” (and to avoid a classification decision that “disfavor[s] its regulated entities”), notwithstanding countervailing political pressures.73Melissa F. Wasserman, Deference Asymmetries: Distortions in the Evolution of Regulatory Law, 93 Tex. L. Rev. 625, 649 (2015). Where Chevron explains its rule of mandatory deference by reference to political-accountability values—requiring agencies to justify their policy choices to the political branches and the voting public—then-Judge Kavanaugh’s application of the major questions doctrine undermines such values by putting a thumb on the scale in favor of deregulation—even where an agency’s politics and expertise point in another direction.


As described above, Symmetry’s Mandate offers a compelling account of an asymmetry in the practice of administrative law, one that appears to favor deregulation over regulation. Justice Kavanaugh’s proposed approach to the major questions doctrine—proposed during his tenure on the D.C. Circuit and in response to the Obama Administration’s decision to promulgate enforceable network neutrality rules—fits that mold. But this approach, like the others that Walters highlights, makes little sense. Justice Kavanaugh’s approach treats analogous wealth transfers—between industries, or between industries and the public—dissimilarly, and it does so in favor of regulated industries, putting outsiders and the public at a distinct disadvantage in the regulatory process. In short, his approach conflicts with Step Zero’s foundational focus on congressional intent, pays little mind to capture and collective action–related concerns, and undermines agency expertise and accountability values. The judiciary should thus reject Justice Kavanaugh’s doctrinal suggestion in favor of a more, well, neutral application of the major questions doctrine.

*         Robert and Nanci Corson Assistant Professor of Law, University of California, Berkeley, School of Law. I owe many thanks to Jon Gould, Khushali Narechania, and Daniel Walters for thoughtful comments and suggestions on this project. For excellent research assistance, I thank Ziyad Alghamdi and Delia Scoville. I also thank Mollie Krent and the members of the Michigan Law Review for their helpful suggestions and careful edits.