Breyer, J., dissenting

FEDERAL COMMUNICATIONS COMMISSION, et al.,
PETITIONERS v. FOX TELEVISION STATIONS,
INC., et al.

on writ of certiorari to the united states court ofappeals for the second circuit

---

---

Justice Breyer, with whom Justice Stevens, Justice Souter, and Justice Ginsburg join, dissenting.

In my view, the Federal Communications Commission failed adequately to explain why it changed its indecency policy from a policy permitting a single “fleeting use” of an expletive, to a policy that made no such exception. Its explanation fails to discuss two critical factors, at least one of which directly underlay its original policy decision. Its explanation instead discussed several factors well known to it the first time around, which by themselves provide no significant justification for a change of policy. Consequently, the FCC decision is “arbitrary, capricious, an abuse of discretion.” 5 U. S. C. §706(2)(A); Motor Vehicle Mfrs. Assn. of United States, Inc. v. State Farm Mut. Automobile Ins. Co., 463 U. S. 29, 41–43 (1983) ; Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U. S. 402, 420–421 (1971) . And I would affirm the Second Circuit’s similar determination.

I

I begin with applicable law. That law grants those in charge of independent administrative agencies broad authority to determine relevant policy. But it does not permit them to make policy choices for purely political reasons nor to rest them primarily upon unexplained policy preferences. Federal Communications Commissioners have fixed terms of office; they are not directly responsible to the voters; and they enjoy an independence expressly designed to insulate them, to a degree, from “ ‘the exercise of political oversight.’ ” Freytag v. Commissioner, 501 U. S. 868, 916 (1991) (Scalia, J., concurring in part and concurring in judgment); see also Morrison v. Olson, 487 U. S. 654, 691, n. 30 (1988) . That insulation helps to secure important governmental objectives, such as the constitutionally related objective of maintaining broadcast regulation that does not bend too readily before the political winds. But that agency’s comparative freedom from ballot-box control makes it all the more important that courts review its decisionmaking to assure compliance with applicable provisions of the law—including law requiring that major policy decisions be based upon articulable reasons.

The statutory provision applicable here is the Administrative Procedure Act’s (APA) prohibition of agency action that is “arbitrary, capricious, [or] an abuse of discretion,” 5 U. S. C. §706(2)(A). This legal requirement helps assure agency decisionmaking based upon more than the personal preferences of the decisionmakers. Courts have applied the provision sparingly, granting agencies broad policymaking leeway. But they have also made clear that agency discretion is not “ ‘unbounded.’ ” Burlington Truck Lines, Inc. v. United States, 371 U. S. 156, 167–168 (1962) . In so holding, American courts have followed a venerable legal tradition, stretching back at least to the days of Sir Edward Coke and the draining of the English fens. See Rooke’s Case, 77 Eng. Rep. 209, 210, 5 Coke Rep. 99b, 100a (C. P. 1598) (Coke, J.) (members of sewer commission with authority to act according “to their discretio[n]” are nonetheless “limited and bound with the rule of reason and law … and [cannot act] according to their wills and private affections” (quoted in Jaffe, Judicial Review: Constitutional and Jurisdictional Fact, 70 Harv. L. Rev. 953, 954 (1957))).

The law has also recognized that it is not so much a particular set of substantive commands but rather it is a process, a process of learning through reasoned argument, that is the antithesis of the “arbitrary.” This means agencies must follow a “logical and rational” decisionmaking “process.” Allentown Mack Sales & Service, Inc. v. NLRB, 522 U. S. 359, 374 (1998) . An agency’s policy decisions must reflect the reasoned exercise of expert judgment. See Burlington Truck Lines, supra, at 167(decision must reflect basis on which agency “exercised its expert discretion”); see also Humphrey’s Executor v. United States, 295 U. S. 602, 624 (1935) (independent agencies “exercise … trained judgment … ‘informed by experience’ ”). And, as this Court has specified, in determining whether an agency’s policy choice was “arbitrary,” a reviewing court “must consider whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.” Overton Park, supra, at 416.

Moreover, an agency must act consistently. The agency must follow its own rules. Arizona Grocery Co. v. Atchison, T. & S. F. R. Co., 284 U. S. 370, 389–390 (1932) .And when an agency seeks to change those rules, it must focus on the fact of change and explain the basis for that change. See, e.g., National Cable & Telecommunications Assn. v. Brand X Internet Services, 545 U. S. 967, 981 (2005) (“Unexplained inconsistency is” a “reason for holding an interpretation to be an arbitrary and capricious change from agency practice” (emphasis added)).

To explain a change requires more than setting forth reasons why the new policy is a good one. It also requires the agency to answer the question, “Why did you change?” And a rational answer to this question typically requires a more complete explanation than would prove satisfactory were change itself not at issue. An (imaginary) administrator explaining why he chose a policy that requires driving on the right-side, rather than the left-side, of the road might say, “Well, one side seemed as good as the other, so I flipped a coin.” But even assuming the rationality of that explanation for an initial choice, that explanation is not at all rational if offered to explain why the administrator changed driving practice, from right-side to left-side, 25 years later.

In State Farm,a unanimous Court applied these commonsense requirements to an agency decision that rescinded an earlier agency policy. The Court wrote that an agency must provide an explanation for the agency’s “revocation” of a prior action that is more thorough than the explanation necessary when it does not act in the first instance. The Court defined “revocation,” not simply as rescinding an earlier policy, cf. ante, at 10–11, but as “a reversal of the agency’s former views as to the proper course.” State Farm, 463 U. S., at 41 (emphasis added). See also Verizon Communications Inc. v. FCC, 535 U. S. 467, 502, n. 20 (2002) (portion of Court’s opinion joined by Scalia, Kennedy, and Thomas, JJ.) (noting State Farm “may be read as prescribing more searching judicial review” when “an agency [is] ‘changing its course’ as to the interpretation of a statute”); Thomas Jefferson Univ. v. Shalala, 512 U. S. 504, 524, n. 3 (1994) (Thomas, J., dissenting) (similar).

At the same time, the Court described the need for explanation in terms that apply, not simply to pure rescissions of earlier rules, but rather to changes of policy as it more broadly defined them. But see ante, at 10–11. It said that the law required an explanation for such a change because the earlier policy, representing a “ ‘settled course of behavior[,] embodies the agency’s informed judgment that, by pursuing that course, it will carry out the policies … best if the settled rule is adhered to.’ ” State Farm, supra, at 41–42. Thus, the agency must explain why it has come to the conclusion that it should now change direction. Why does it now reject the considerations that led it to adopt that initial policy? What has changed in the world that offers justification for the change? What other good reasons are there for departing from the earlier policy?

Contrary to the majority’s characterization of this dissent, it would not (and State Farm does not) require a “heightened standard” of review. Ante,at 10 (emphasis added). Rather, the law requires application of the same standard of review to different circumstances, namely circumstances characterized by the fact that change is at issue. It requires the agency to focus upon the fact of change where change is relevant, just as it must focus upon any other relevant circumstance. It requires the agency here to focus upon the reasons that led the agency to adopt the initial policy, and to explain why it now comes to a new judgment.

I recognize that sometimes the ultimate explanation for a change may have to be, “We now weigh the relevant considerations differently.” But at other times, an agency can and should say more. Where, for example, the agency rested its previous policy on particular factual findings, see ante, at 3–5 (Kennedy, J., concurring in part and concurring in judgment); or where an agency rested its prior policy on its view of the governing law, see infra, at 7–11; or where an agency rested its previous policy on, say, a special need to coordinate with another agency, one would normally expect the agency to focus upon those earlier views of fact, of law, or of policy and explain why they are no longer controlling. Regardless, to say that the agency here must answer the question “why change” is not to require the agency to provide a justification that is “better than the reasons for the old [policy].” Ante,at 11. It is only to recognize the obvious fact that change is sometimes (not always) a relevant background feature that sometimes (not always) requires focus (upon prior justifications) and explanation lest the adoption of the new policy (in that circumstance) be “arbitrary, capricious, an abuse of discretion.”

That is certainly how courts of appeals, the courts that review agency decisions, have always treated the matter in practice. See, e.g., Pennsylvania Federation of Sportsmen’s Clubs, Inc. v. Kempthorne, 497 F. 3d 337, 351 (CA3 2007); Yale-New Haven Hosp. v. Leavitt, 470 F. 3d 71, 79 (CA2 2006); Citizens Awareness Network, Inc. v. United States, 391 F. 3d 338, 352 (CA1 2004). But see NAACP v. FCC, 682 F. 2d 993, 998 (CADC 1982) (using word “heightened”). The majority’s holding could in this respect significantly change judicial review in practice, and not in a healthy direction. But see, ante, at 1–5 (Kennedy, J., concurring in part and concurring in judgment). After all, if it is always legally sufficient for the agency to reply to the question “why change?” with the answer “we prefer the new policy” (even when the agency has not considered the major factors that led it to adopt its old policy), then why bother asking the agency to focus on the fact of change? More to the point, why would the law exempt this and no other aspect of an agency decision from “arbitrary, capricious” review? Where does, and why would, the APA grant agencies the freedom to change major policies on the basis of nothing more than political considerations or even personal whim?

Avoiding the application of any heightened standard of review, the Court in State Farm recognized that the APA’s “nonarbitrary” requirement affords agencies generous leeway when they set policy. 463 U. S., at 42. But it also recognized that this leeway is not absolute. The Court described its boundaries by then listing considerations that help determine whether an explanation is adequate. Mirroring and elaborating upon its statement in Overton Park, 401 U. S. 402 , the Court said that a reviewing court should take into account whether the agency had “relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” State Farm, supra, at 43; see also Overton Park, supra, at 416.

II

We here must apply the general standards set forth in State Farm and Overton Park to an agency decision that changes a 25-year-old “fleeting expletive” policy from (1) the old policy that would normally permit broadcasters to transmit a single, fleeting use of an expletive to (2) a new policy that would threaten broadcasters with large fines for transmitting even a single use (including its use by a member of the public) of such an expletive, alone with nothing more. The question is whether that decision satisfies the minimal standards necessary to assure a reviewing court that such a change of policy is not “arbitrary, capricious, [or] an abuse of discretion,” 5 U. S. C. §706(2)(A), particularly as set forth in, e.g., State Farm and Overton Park, supra,at2–7. The decision, in my view, does not satisfy those standards.

Consider the requirement that an agency at least minimally “consider … important aspect[s] of the problem.” State Farm, supra, at 43. The FCC failed to satisfy this requirement, for it failed to consider two critically important aspects of the problem that underlay its initial policy judgment (one of which directly, the other of which indirectly). First, the FCC said next to nothing about the relation between the change it made in its prior “fleeting expletive” policy and the First-Amendment-related need to avoid “censorship,” a matter as closely related to broadcasting regulation as is health to that of the environment. The reason that discussion of the matter is particularly important here is that the FCC had explicitly rested its prior policy in large partupon the need to avoid treading too close to the constitutional line.

Thirty years ago, the Court considered the location of that constitutional line. In FCC v. Pacifica Foundation, 438 U. S. 726 (1978) ,the Court reviewed an FCC decision forbidding the broadcast of a monologue that deliberately and repeatedly uttered the expletives here at issue more than 100 times in one hour at a time of day when children were likely to hear the broadcast. Id., at 739. The Court held that the FCC’s prohibition did not violate the First Amendment . But the Court divided 5 to 4. And two Members of the majority, Justices Powell and Blackmun, explicitly noted that the Court “does not speak to cases involving the isolated use of a potentially offensive word … as distinguished from the verbal shock treatment administered by respondent here.” Id., at 760–761 (Powell, J., concurring in part and concurring in judgment) (emphasis added). This statement by two Members of the majority suggested that they could reach a different result, finding an FCC prohibition unconstitutional, were that prohibition aimed at the fleeting or single use of an expletive.

The FCC subsequently made clear that it thought that Justice Powell’s concurrence set forth a constitutional line that its indecency policy should embody. In 1978, the Commission wrote that the First Amendment “severely limit[s]” the Commission’s role in regulating indecency. It added that the Court, in Pacifica,had “relied … on the repetitive occurrence of the ‘indecent’ words in question.” And it said that, in setting policy, it “intend[ed] strictly to observe the narrowness of the Pacifica holding.” In re Application of WGBH Educ. Foundation, 69 F. C. C. 2d 1250, 1254, ¶10.

In 1983, the Commission again wrote that it understood the Court’s decision in Pacifica to rest on the “repetitive occurrence of the ‘indecent’ words in question.” And, again, the Commission explained that its regulation of fleeting or isolated offensive words would reflect Justice Powell’s understanding of the First Amendment ’s scope. In re Application of Pacifica Foundation, 95 F. C. C. 2d 750, 760, ¶¶17–18. In 1987, the Commission once more explained that its “fleeting expletives” policy reflected the Court’s decision in Pacifica. It said that, under its policy, “speech that is indecent must involve more than an isolated use of an offensive word,” adding that “we believe that under the legal standards set forth in Pacifica, deliberate and repetitive use in a patently offensive manner is a requisite to a finding of indecency.” In re Pacifica Foundation, 2 FCC Rcd. 2698, 2699, ¶13 (emphasis added). In another order that same year, the Commission stated that “the First Amendment dicate[s] a careful and restrained approach with regard to review of matters involving broadcast programming”; it then explained, citing Pacifica, that “[s]peech that is indecent must involve more than the isolated use of an offensive word.” In re Infinity Broadcasting, 2 FCC Rcd. 2705, 2705, ¶¶6–7 (1987) (emphasis added). And in 2001, in giving the industry guidance, the FCC once again said in respect to its regulation of indecent speech that it “must both identify a compelling interest for any regulation … and choose the least restrictive means to further that interest.” In re Industry Guidance On Commission’s Case Law Interpreting 18 U. S. C. §1464 and Enforcement Policies Regarding Broadcast Indecency, 16 FCC Rcd. 7999, 8000–8001, ¶3–5.

The FCC thus repeatedly made clear that it based its “fleeting expletive” policy upon the need to avoid treading too close to the constitutional line as set forth in Justice Powell’s Pacifica concurrence. What then did it say, when it changed its policy, about why it abandoned this Constitution-based reasoning? The FCC devoted “four full pages of small-type, single-spaced text,” ante, at 23, responding to industry arguments that, e.g., changes in the nature of the broadcast industry made all indecency regulation, i.e., 18 U. S. C. §1464, unconstitutional. In doing so it repeatedly reaffirmed its view that Pacifica remains good law. In re Complaints Regarding Various Television Broadcasts Between February 2, 2002, and March, 8, 2008, 21 FCC Rcd. 13299, 13317–13321, ¶¶42–52 (2006) (Remand Order). All the more surprising then that, in respect to why it abandoned its prior view about the critical relation between its prior fleeting expletive policy and Justice Powell’s Pacifica concurrence, it says no more than the following:

“[O]ur decision is not inconsistent with the Supreme Court ruling in Pacifica. The Court explicitly left open the issue of whether an occasional expletive could be considered indecent.” In re Complaints Against Various Broadcast Licensees Regarding Their Airing of the “Golden Globe Awards” Program, 19 FCC Rcd. 4975, 4982, ¶16 (2004) (Golden Globe Order). And, (repeating what it already had said), “[Pacifica] specifically reserved the question of ‘an occasional expletive’ and noted that it addressed only the ‘particular broadcast’ at issue in that case.” Remand Order, supra, at 13308–13309, ¶24.

These two sentences are not a summary of the FCC’s discussion about why it abandoned its prior understanding of Pacifica. They are the discussion. These 28 words (repeated in two opinions) do not acknowledge that an entirely different understanding of Pacifica underlay the FCC’s earlier policy; they do not explain why the agency changed its mind about the line that Pacifica draws or its policy’s relation to that line; and they tell us nothing at all about what happened to the FCC’s earlier determination to search for “compelling interests” and “less restrictive alternatives.” They do not explain the transformation of what the FCC had long thought an insurmountable obstacle into an open door. The result is not simply Hamlet without the prince, but Hamlet with a prince who, in mid-play and without explanation, just disappears.

I have found one other related reference to Pacifica, but that reference occurs in an opinion written by a dissenting Commissioner. That dissenter said that the FCC had “ ‘fail[ed] to address the many serious [constitutional] concerns raised’ ” by the new policy, while adding that the new policy was “not the restrained enforcement policy encouraged by the Supreme Court in Pacifica.” Remand Order, supra, at 13331, 13334. Neither that Commissioner in his dissent, nor I in this dissent, claim that agencies must always take account of possible constitutional issues when they formulate policy. Cf. ante, at 12. But the FCC works in the shadow of the First Amendment and its view of the application of that Amendment to “fleeting expletives” directly informed its initial policy choice. Under these circumstances, the FCC’s failure to address this “aspect” of the problem calls for a remand to the agency. Overton Park, 401 U. S., at 420–421.

Second, the FCC failed to consider the potential impact of its new policy upon local broadcasting coverage. This “aspect of the problem” is particularly important because the FCC explicitly took account of potential broadcasting impact. Golden Globe Order, supra, at 4980, ¶11 (“The ease with which broadcasters today can block even fleeting words in a live broadcast is an element in our decision”). Indeed, in setting forth “bleeping” technology changes (presumably lowering bleeping costs) as justifying the policy change, it implicitly reasoned that lower costs, making it easier for broadcasters to install bleeping equipment, made it less likely that the new policy would lead broadcasters to reduce coverage, say by canceling coverage of public events. Ibid. (“[T]echnological advances have made it possible … to prevent the broadcast of a single offending word or action without blocking or disproportionately disrupting the message of the speaker or performer”).

What then did the FCC say about the likelihood that smaller independent broadcasters, including many public service broadcasters, still would not be able to afford “bleeping” technology and, as a consequence, would reduce local coverage, indeed cancel coverage, of many public events? It said nothing at all.

The FCC cannot claim that local coverage lacks special importance. To the contrary, “the concept of localism has been a cornerstone of broadcast regulation for decades.” In re Broadcast Localism, 23 FCC Rcd. 1324, 1326, 1327, ¶¶3, 5 (2008). That policy seeks to provide “viewers and listeners … access to locally responsive programming including, but not limited to, local news and public affairs matter” and to ensure “diversity in what is seen and heard over the airwaves.” That policy has long favored local broadcasting, both as a means to increase coverage of local events and, insofar as it increases the number of broadcast voices, as an end in itself. See, e.g., In re Reexamination of Comparative Standards for Noncommercial Educ. Applicants, 15 FCC Rcd. 7386, 7399, ¶29 (2000) (adopting a system for selecting applicants for broadcast channels that “would foster our goal of broadcast diversity by enabling the local public to be served by differing … licensees”); In re 2002 Biennial Regulatory Review, 18 FCC Rcd. 13620, 13644, ¶¶77, 79 (2003) (“We remain firmly committed to the policy of promoting localism among broadcast outlets. … A … measure of localism is the quantity and quality of local news and public affairs programming”).

Neither can the FCC now claim that the impact of its new policy on local broadcasting is insignificant and obviously so. Broadcasters tell us, as they told the FCC, the contrary. See Brief for Former FCC Commissioners as Amici Curiae 17–19; App. 235–237; Joint Comments of Fox Television Stations, Inc. et al., In re Remand of Section III.B of the Commission’s March 15, 2006 Omnibus Order Resolving Numerous Broadcast Television Indecency Complaints 14–15, http://www.fcc.gov/DA06–1739/joint-networks.pdf (all Internet materials as visited Apr. 7, 2009, and available in Clerk of Court’s case file). They told the FCC, for example, that the costs of bleeping/delay systems, up to $100,000 for installation and annual operation, place that technology beyond the financial reach of many smaller independent local stations. See id., at 14 (“The significant equipment and personnel costs associated with installing, maintaining, and operating delay equipment sufficient to cover all live news, sports, and entertainment programs could conceivably exceed the net profits of a small local station for an entire year”); id., at App. XI. And they ask what the FCC thinks will happen when a small local station without bleeping equipment wants to cover, say a local city council meeting, a high school football game, a dance contest at community center, or a Fourth of July parade.

Relevant literature supports the broadcasters’ financial claims. See, e.g., Ho, Taking No Chances, Austin American-Statesman, June 18, 2006, p. J1; Dotinga, Dirty-Word Filters Prove Costly, Wired.com, July 9, 2004, http://www.wired.com/entertainment/music/news/2004/07/64127; Stations, Cable Networks Finding Indecency Rules Expensive, Public Broadcasting Report, Aug. 4, 2006. It also indicates that the networks with which some small stations are affiliated are not liable for the stations’ local transmissions (unless the networks own them). Ho, supra, at J1; Public Stations Fear Indecency Fine Jump Means Premium Hikes, Public Broadcasting Report, July 7, 2006. The result is that smaller stations, fearing “fleeting expletive” fines of up to $325,000, may simply cut back on their coverage. See Romano, Reporting Live. Very Carefully, Broadcasting & Cable, July 4, 2005, p. 8; see also ibid. (“Afraid to take chances” of getting fined under the FCC’s new policy, “local broadcasters are responding by altering—or halting altogether—the one asset that makes local stations so valuable to their communities: live TV”); Daneman, WRUR Drops Its Live Radio Programs, Rochester Democrat and Chronicle, May 27, 2004, p. 1B (reporting that a local broadcast station ceased broadcasting all local live programming altogether in response to the Commission’s policy change). And there are many such smaller stations. See, e.g., Corporation for Public Broadcasting, Frequently Asked Questions, available at http://www.cpb.org/aboutpb/faq/stations.html (noting there are over 350 local public television stations and nearly 700 local public radio stations that receive support from the Corporation for Public Broadcasting).

As one local station manager told the FCC,

“[t]o lessen the risk posed by the new legal framework … I have directed [the station’s] news staff that [our station] may no longer provide live, direct-to-air coverage” of “live events where crowds are present … unless they affect matters of public safety or convenience. Thus, news coverage by [my station] of live events where crowds are present essentially will be limited to civil emergencies.” App. 236–237 (declaration of Dennis Fisher).

What did the FCC say in response to this claim? What did it say about the likely impact of the new policy on the coverage that its new policy is most likely to affect, coverage of local live events—city council meetings, local sports events, community arts productions, and the like? It said nothing at all.

The plurality acknowledges that the Commission entirely failed to discuss this aspect of the regulatory problem. But it sees “no need” for discussion in light of its, i.e., the plurality’s, own “doubt[s]” that “small-town broadcasters run a heightened risk of liability for indecent utterances” as a result of the change of policy. Ante, at 24–25. The plurality's “doubt[s]” rest upon its views (1) that vulgar expression is less prevalent (at least among broadcast guests) in smaller towns, ante, at 24; (2) that the greatest risk the new policy poses for “small-town broadcasters” arises when they broadcast local “news and public affairs,” ibid., and (3) that the Remand Order says “little about how the Commission would treat smaller broadcasters who cannot afford screening equipment,” while also pointing out that the new policy “ ‘does not … impose undue burdens on broadcasters’ ” and emphasizing that the case before it did not involve “ ‘breaking news.’ ” Ante,at 24–25.

As to the first point, about the prevalence of vulgarity in small towns, I confess ignorance. But I do know that there are independent stations in many large and medium sized cities. See Television & Cable Factbook, Directory of Television Stations in Operation 2008. As to the second point, I too believe that coverage of local public events, if not news, lies at the heart of the problem.

I cannot agree with the plurality, however, about the critical third point, namely that the new policy obviously provides smaller independent broadcasters with adequate assurance that they will not be fined. The new policy removes the “fleeting expletive” exception, an exception that assured smaller independent stations that they would not be fined should someone swear at a public event. In its place, it puts a policy that places all broadcasters at risk when they broadcast fleeting expletives, including expletives uttered at public events. The Remand Order says that there “is no outright news exemption from our indecency rules.” 21 FCC Rcd., at 13327, ¶71 (emphasis added).The best it can provide by way of assurance is to say that “it may be inequitable to hold a licensee responsible for airing offensive speech during live coverage of a public event under some circumstances.” Id., at 13311, ¶33 (emphasis added). It does list those circumstances as including the “possibility of human error in using delay equipment.” Id., at 13313, ¶35. But it says nothing about a station’s inability to afford delay equipment (a matter that in individual cases could itself prove debatable). All the FCC had to do was to consider this matter and either grant an exemption or explain why it did not grant an exemption. But it did not. And the result is a rule that may well chill coverage—the kind of consequence that the law has considered important for decades, to which the broadcasters pointed in their arguments before the FCC, and which the FCC nowhere discusses. See, e.g., Dombrowski v. Pfister, 380 U. S. 479, 494 (1965) (“So long as the statute remains available to the State the threat of prosecutions of protected expression is a real and substantial one. Even the prospect of ultimate failure of such prosecutions by no means dispels their chilling effect on protected expression”); see also Ashcroft v. Free Speech Coalition, 535 U. S. 234, 244 (2002) ; Gibson v. Florida Legislative Investigation Comm., 372 U. S. 539, 556–557 (1963) ; Wieman v. Updegraff, 344 U. S. 183, 195 (1952) (Frankfurter, J., concurring).

Had the FCC used traditional administrative notice-and-comment procedures, 5 U. S. C. §553, the two failures I have just discussed would clearly require a court to vacate the resulting agency decision. See ACLU v. FCC, 823 F. 2d 1554, 1581 (CADC 1987) (“Notice and comment rulemaking procedures obligate the FCC to respond to all significant comments, for the opportunity to comment is meaningless unless the agency responds to significant points raised by the public” (emphasis added; internal quotation marks omitted)). Here the agency did not make new policy through the medium of notice and comment proceedings. But the same failures here—where the policy is important, the significance of the issues clear, the failures near complete—should lead us to the same conclusion. The agency’s failure to discuss these two “important aspect[s] of the problem” means that the resulting decision is “ ‘arbitrary, capricious, an abuse of discretion’ ” requiring us to remand the matter to the agency. State Farm, 463 U. S., at 43; Overton Park, 401 U. S., at 416.

III

The three reasons the FCC did set forth in support of its change of policy cannot make up for the failures I have discussed. Consider each of them. First, as I have pointed out, the FCC based its decision in part upon the fact that “bleeping/delay systems” technology has advanced. I have already set forth my reasons for believing that that fact, without more, cannot provide a sufficient justification for its policy change. Supra, at 11–16.

Second, the FCC says that the expletives here in question always invoke a coarse excretory or sexual image; hence it makes no sense to distinguish between whether one uses the relevant terms as an expletive or as a literal description. The problem with this answer is that it does not help to justify the change in policy. The FCC was aware of the coarseness of the “image” the first time around. See, e.g., Remand Order, 21 FCC Rcd., at 13308, ¶23 (asserting that FCC has always understood the words as coarse and indecent). And it explained the first time around why it nonetheless distinguished between their literal use and their use as fleeting expletives. See, e.g., In re Application of WGBH Educ. Foundation, 69 F. C. C. 2d, at 1254–1255, ¶¶10–11 (discussing First Amendment considerations and related need to avoid reduced broadcast coverage). Simply to announce that the words, whether used descriptively or as expletives, call forth similar “images” is not to address those reasons.

Third, the FCC said that “perhaps” its “most importan[t]” justification for the new policy lay in the fact that its new “contextual” approach to fleeting expletives is better and more “[c]onsistent with” the agency’s “general approach to indecency” than was its previous “categorica[l]” approach, which offered broadcasters virtual immunity for the broadcast of fleeting expletives. Remand Order, supra, at 13308, ¶23. This justification, however, offers no support for the change without an understanding of why, i.e., in what way, the FCC considered the new approach better or more consistent with the agency’s general approach.

The Solicitor General sets forth one way in which the new policy might be more consistent with statutory policy. The indecency statute prohibits the broadcast of “any … indecent … language.” 18 U. S. C. §1464. The very point of the statute, he says, is to eliminate nuisance; and the use of expletives, even once, can constitute such a nuisance. The Solicitor General adds that the statutory word “any” indicates that Congress did not intend a safe-harbor for a fleeting use of that language. Brief for Petitioners 24–25. The fatal flaw in this argument, however, lies in the fact that the Solicitor General and not the agency has made it. We must consider the lawfulness of an agency’s decision on the basis of the reasons the agency gave, not on the basis of those it might have given. SEC v. Chenery Corp., 332 U. S. 194, 196–197 (1947) ; State Farm, supra, at 50. And the FCC did not make this claim. Hence, we cannot take it into account and need not evaluate its merits.

In fact, the FCC found that the new policy was better in part because, in its view, the new policy better protects children against what it described as “ ‘the first blow’ ” of broadcast indecency that results from the “ ‘pervasive’ ” nature of broadcast media. It wrote that its former policy of “granting an automatic exemption for ‘isolated or fleeting’ expletives unfairly forces viewers (including children) to take ‘the first blow.’ ” Remand Order, supra, at 13309, ¶25.

The difficulty with this argument, however, is that it does not explain the change. The FCC has long used the theory of the “first blow” to justify its regulation of broadcast indecency. See, e.g., In re Enforcement of Prohibitions Against Broadcast Indecency in 18 U. S. C. §1464, 5 FCC Rcd. 5297, 5302, ¶¶34–35 (1990). Yet the FCC has also long followed its original “fleeting expletives” policy. Nor was the FCC ever unaware of the fact to which the majority points, namely that children’s surroundings influence their behavior. See, e.g., In re Enforcement of Prohibitions Against Broadcast Indecency in 18 U. S. C. §1464, 8 FCC Rcd. 704, 706, ¶11 (1993). So, to repeat the question: What, in respect to the “first blow,” has changed?

The FCC points to no empirical (or other) evidence to demonstrate that it previously understated the importance of avoiding the “first blow.” Like the majority, I do not believe that an agency must always conduct full empirical studies of such matters. Ante, at 15–16. But the FCC could have referred to, and explained, relevant empirical studies that suggest the contrary. One review of the empirical evidence, for example, reports that “[i]t is doubtful that children under the age of 12 understand sexual language and innuendo; therefore it is unlikely that vulgarities have any negative effect.” Kaye & Sapolsky, Watch Your Mouth! An Analysis of Profanity Uttered by Children on Prime-Time Television, 2004 Mass Communication & Soc’y 429, 433 (Vol. 7) (citing two studies). The Commission need not have accepted this conclusion. But its failure to discuss this or any other such evidence, while providing no empirical evidence at all that favors its position, must weaken the logical force of its conclusion. See State Farm, 463 U. S., at 43(explaining that an agency’s failure to “examine the relevant data” is a factor in determining whether the decision is “arbitrary”).

The FCC also found the new policy better because it believed that its prior policy “would as a matter of logic permit broadcasters to air expletives at all hours of a day so long as they did so one at a time.” Remand Order, 21 FCC Rcd., at 13309, ¶25. This statement, however, raises an obvious question: Did that happen? The FCC’s initial “fleeting expletives” policy was in effect for 25 years. Had broadcasters during those 25 years aired a series of expletives “one at a time?” If so, it should not be difficult to find evidence of that fact. But the FCC refers to none. Indeed, the FCC did not even claim that a change had taken place in this respect. It spoke only of the pure “logic” of the initial policy “permitting” such a practice. That logic would have been apparent to anyone, including the FCC, in 1978 when the FCC set forth its initial policy.

Finally, the FCC made certain statements that suggest it did not believe it was changing prior policy in any major way. It referred to that prior policy as based on “staff letters and dicta” and it said that at least one of the instances before it (namely, the Cher broadcast) would have been actionably indecent under that prior policy. Id., at 13306–13307, 13324, ¶¶20–21, 60. As we all agree, however, in fact the FCC did change its policy in a major way. See ante, at 13. To the extent that the FCC minimized that fact when considering the change, it did not fully focus on the fact of change. And any such failure would make its decision still less supportable. See National Cable, 545 U. S., at 981.

IV

Were the question a closer one, the doctrine of constitutional avoidance would nonetheless lead me to remand the case. See United States v. Jin Fuey Moy, 241 U. S. 394, 401 (1916) (“A statute must be construed, if fairly possible, so as to avoid not only the conclusion that it is unconstitutional but also grave doubts upon that score” (emphasis added)). That doctrine seeks to avoid unnecessary judicial consideration of constitutional questions, assumes that Congress, no less than the Judicial Branch, seeks to act within constitutional bounds, and thereby diminishes the friction between the branches that judicial holdings of unconstitutionality might otherwise generate. See Almendarez-Torres v. United States, 523 U. S. 224, 237–238 (1998) ; see also Solid Waste Agency of Northern Cook Cty. v. Army Corps of Engineers, 531 U. S. 159, 172–173 (2001) ; Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Constr. Trades Council, 485 U. S. 568, 575 (1988) ; Rescue Army v. Municipal Court of Los Angeles, 331 U. S. 549, 571 (1947) ; Ashwander v. TVA, 297 U. S. 288, 345–348 (1936) (Brandeis, J., concurring). The doctrine assumes that Congress would prefer a less-than-optimal interpretation of its statute to the grave risk of a constitutional holding that would set the statute entirely aside. See Almendarez-Torres, supra, at 238 (construction of statute that avoids invalidation best reflects congressional will); cf. United States v. Booker, 543 U. S. 220, 249, 267 (2005) .

Unlike the majority, I can find no convincing reason for refusing to apply a similar doctrine here. The Court has often applied that doctrine where an agency’s regulation relies on a plausible but constitutionally suspect interpretation of a statute. See, e.g., Solid Waste Agency, supra, at 172–174; NLRB v. Catholic Bishop of Chicago, 440 U. S. 490, 506–507 (1979) . The values the doctrine serves apply whether the agency’s decision does, or does not, rest upon a constitutionally suspect interpretation of a statute. And a remand here would do no more than ask the agency to reconsider its policy decision in light of the concerns raised in a judicial opinion. Cf. Fullilove v. Klutznick, 448 U. S. 448, 551 (1980) (Stevens, J., dissenting) (a holding that a congressional action implicating the Equal Protection Clause “was not adequately preceded by a consideration of less drastic alternatives or adequately explained by a statement of legislative purpose would be far less intrusive than a final determination that the substance of” that action was unconstitutional). I would not now foreclose, as the majority forecloses, our further consideration of this matter. (Of course, nothing in the Court’s decision today prevents the Commission from reconsidering its current policy in light of potential constitutional considerations or for other reasons.)

V

In sum, the FCC’s explanation of its change leaves out two critically important matters underlying its earlier policy, namely Pacifica and local broadcasting coverage. Its explanation rests upon three considerations previously known to the agency (“coarseness,” the “first blow,” and running single expletives all day, one at a time). With one exception, it provides no empirical or other information explaining why those considerations, which did not justify its new policy before, justify it now. Its discussion of the one exception (technological advances in bleeping/delay systems), failing to take account of local broadcast coverage, is seriously incomplete.

I need not decide whether one or two of these features, standing alone, would require us to remand the case. Here all come together. And taken together they suggest that the FCC’s answer to the question, “Why change?” is, “We like the new policy better.” This kind of answer, might be perfectly satisfactory were it given by an elected official. But when given by an agency, in respect to a major change of an important policy where much more might be said, it is not sufficient. State Farm, 463 U. S., at 41–42.

For these reasons I would find the FCC’s decision “arbitrary, capricious, an abuse of discretion,” 5 U. S. C. §706(2)(A), requiring remand of this case to the FCC. And I would affirm the Second Circuit’s similar determination.

With respect, I dissent.