Full opinion text
SACK, Circuit Judge: The parties, the district court, and amici have raised a wide variety of interesting legal and policy issues during the course of this litigation. We need not address most of them. We conclude that under principles that are well established in this Circuit, the plaintiffs’ claim against the defendant for “hot news” misappropriation of the plaintiff financial firms’ recommendations to clients and prospective clients as to trading in corporate securities is preempted by federal copyright law. Based upon principles explained and applied in National Basketball Association v. Motorola, Inc., 105 F.3d 841 (2d Cir.1997) (sometimes hereinafter “NBA ”), we conclude that because the plaintiffs’ claim falls within the “general scope” of copyright, 17 U.S.C. § 106, and involves the type of works protected by the Copyright Act, 17 U.S.C. §§ 102 and 103, and because the defendant’s acts at issue do not meet the exceptions for a “hot news” misappropriation claim as recognized by NBA, the claim is preempted. We therefore reverse the judgment of the district court with respect to that claim. The plaintiffs-appellees — Barclays Capital Inc. (“Barclays”); Merrill Lynch, Pierce, Fenner & Smith Inc. (“Merrill Lynch”); and Morgan Stanley & Co. Inc. (“Morgan Stanley”) (collectively, the “Firms”) — are major financial institutions that, among many other things, provide securities brokerage services to members of the public. Largely in that connection, they engage in extensive research about the business and prospects of publicly traded companies, the securities of those companies, and the industries in which those companies are engaged. The results of the research are summarized by the Firms in reports, which customarily contain recommendations as to the wisdom of purchasing, holding, or selling securities of the subject companies. Although the recommendations and the research underlying them in the reports are inextricably related, it is the alleged misappropriation of the recommendations, each typically contained in a single sentence, that is at the heart of the district court’s decision and the appeal here. Each morning before the principal U.S. securities markets open, each Firm circulates its reports and recommendations for that day to clients and prospective clients. The recipients thus gain an informational advantage over non-recipients with respect to possible trading in the securities of the subject companies both by learning before the world at large does the contents of the reports and, crucially for present purposes, the fact that the recommendations are being made by the Firm. The existence of that fact alone is likely to result in purchases or sales of the securities in question by client and non-client alike, and a corresponding short-term increase or decrease in the securities’ market prices. The Firms and similar businesses, under their historic and present business models, profit from the preparation and circulation of the reports and recommendations principally insofar as they earn brokerage commissions when a recipient of a report and recommendation turns to the firm to execute a trade in the shares of the company being reported upon. The defendant-appellant is the proprietor of a news service distributed electronically, for a price, to subscribers. In recent years and by various means, the defendant has obtained information about the Firms’ recommendations before the Firms have purposely made them available to the general public and before exchanges for trading in those shares open for the day. Doing so tends' to remove the informational and attendant trading advantage of the Firms’ clients and prospective clients who are authorized recipients of the reports and recommendations. The recipients of the information are, in turn, less likely to buy or sell the securities using the brokerage services of the reporting and recommending Firms, thereby reducing the incentive for the Firms to create such reports and recommendations in the first place. This, the Firms assert, will destroy their business models and have a severely deleterious impact on their ability to engage in further research and to create further reports and recommendations. In an attempt to preserve their business models, the Firms have increasingly taken measures to seek to prevent or curtail such pre-market — and therefore, from their point of view, premature — public dissemination of their recommendations. As the district court reported in Barclays Capital Inc. v. Theflyonthewall.com (“Fly I”), 700 F.Supp.2d 310 (S.D.N.Y.2010), the Firms have, for example: “communicated to their employees that the unauthorized dissemination of their equity research or its contents is a breach of loyalty to the Firm, undermines the Firm’s creation of revenue, and can result in discipline, including firing,” id. at 319-20; included in their licensing agreements with third-party distributors and in the reports themselves provisions prohibiting redistribution of their content, id. at 320; adopted policies limiting public dissemination of the reports and the information they contain, id.; and employed emerging Internet technology by which the Firms can seek to find the source of such “leaks” and to “plug” them, id. It is not clear from the record the extent to which these efforts are currently effective, but no concern has been expressed to us as to their legality or legitimacy. The Firms instituted this litigation as part of the same endeavor. The first of their two sets of claims against the defendant sounds in copyright and is based on allegations of verbatim copying and dissemination of portions of the Firms’ reports by the defendant. The Firms have been entirely successful on these copyright claims. See Fly /, 700 F.Supp.2d at 328 (“Fly no longer disputes ... that it infringed the copyrights in [seventeen of the Firms’ reports].... [J]udgment shall [therefore] be entered for the [Firms] on their claims of copyright infringement.”). Although the extent to which the Firms’ success on the copyright claims has alleviated their overall concerns is not clear, their victory on these claims is secure: Fly has not challenged the resulting injunction on appeal. Appellant’s Br. at 61. What remains before us, then, is the second set of claims by the Firms, alleging that Fly’s early republication of the securities recommendations that the Firms create — their “hot news” — is tortious under the New York State law of misappropriation. The district court agreed and granted carefully measured injunctive relief. It is to the misappropriation cause of action that this appeal and therefore this opinion is devoted. BACKGROUND We find little to take issue with in the district court’s careful findings of facts, to which we must in any event defer. We therefore borrow freely from them. The Firms and their Research Reports The Firms are multinational financial entities that provide a variety of asset management, sales and trading, investment banking, and brokerage services to institutional investors, businesses of various sizes, and individuals. Among their many activities, the Firms compile research reports on specific companies whose securities are publicly traded, on industries, and on economic conditions generally. They disseminate such reports and accompanying trading recommendations to clients, such as hedge funds, private equity firms, pension funds, endowments, and individual investors. The reports, which vary in format, range from a single page to hundreds of pages in length. They typically include data analysis, qualitative discussion, and the recommendation. In the process of producing and disseminating the reports, the Firms employ hundreds of research analysts and spend hundreds of millions of dollars annually. In preparing a company report, an analyst will gather data related to its business, and may visit its physical facilities, converse with industry experts or company executives, and construct financial or operational models. The analyst then uses that information in light of his or her expertise, experience, and judgment to arrive at formal projections and recommendations regarding the value of the company’s securities. This litigation concerns the trading “Recommendations,” a term which the district court defined as “actionable reports,” i.e., Firm research reports “likely to spur any investor into making an immediate trading decision[] Recommendations upgrade or downgrade a security; begin research coverage of a company’s security (an event known as an ‘initiation’); or predict a change in the security’s target price.” Fly I, 700 F.Supp.2d at 316. The better known and more respected an analyst is, the more likely that a recommendation for which he or she is primarily responsible will significantly affect the market price of a security. Most Recommendations are issued sometime between midnight and 7 a.m. Eastern Time, allowing stock purchases to be made on the market based on the reports and Recommendations upon the market opening at 9:30 a.m. Timely receipt of a Recommendation affords an investor the opportunity to execute a trade in the subject security before the market has absorbed and responded to it. The Firms typically provide complimentary copies of the reports and Recommendations to their institutional and individual clients using a variety of methods. The Firms then conduct an orchestrated sales campaign in which members of their sales forces contact the clients the Firms think most likely to execute a trade based upon the Recommendation, with the understanding that continued receipt of reports and Recommendations may be made contingent on the generation of a certain level of trading commissions paid to the Firm. The Firms contend that clients are much more likely to place a trade with a Firm if they learn of the Recommendation directly from that Firm rather than elsewhere, and estimate that more than sixty percent of all trades result from Firm solicitations, including those highlighting Recommendations. It is from the commissions on those trades that Firms profit from the creation and dissemination of their reports and Recommendations. They assert that the timely, exclusive delivery of research and Recommendations therefore is a key to what they frequently refer to as their “business model.” Theflyonthewall. com The defendant-appellant Theflyonthewall.com, Inc. (“Fly”) is, among other things, a news “aggregator.” For present purposes, “[a]n aggregator is a website that collects headlines and snippets of news stories from other websites. Exam-pies include Google News and the Huffing-ton Post.” Tony Rogers, “Aggregator,” About.com Guide, available at http:// journalism.about.com/od/journalism glossary/g/aggregatordefinition.htm (latest visit Jan. 4, 2011). Understanding that investors not authorized by the Firms to receive the reports and Recommendations are interested in and willing to pay for early access to the information contained in them — especially the Recommendations, which are particularly likely to affect securities prices — several aggregators compile securities-firm recommendations, including the Recommendations of the Firms, sometimes with the associated reports or summaries thereof, and timely provide the information to their own subscribers for a fee. Fly is one such company. It employs twenty-eight persons, about half of whom are devoted to content production. It does not itself provide brokerage, trading, or investment-advisory services beyond supplying that information. Typical clients of the Firms are hedge funds, private equity firms, pension funds, endowments, and wealthy individual investors. By contrast, Fly’s subscribers are predominately individual investors, institutional investors, brokers, and day traders. These customers purchase one of three content packages on Fly’s website, paying between $25 and $50 monthly for unlimited access to the site. In addition to maintaining its website, Fly distributes its content through third-party distributors and trading platforms, including some, such as Bloomberg and Thomson Reuters, that also separately provide authorized dissemination of the Firms’ Recommendations. Fly has about 3,300 direct subscribers through its website, and another 2,000 subscribers who use third-party platforms to receive the service. Fly characterizes itself as a source for breaking financial news, claiming to be the “fastest news feed on the web.” Fly I, 700 F.Supp.2d at 322 (internal quotation marks omitted). It advertises that its “quick to the point news is a valuable resource for any investment decision.” Id. Fly has emphasized its access to analyst research, saying that its newsfeed is a “one-stop solution for accessing analyst comments,” and brags that it posts “breaking analyst comments as they are being disseminated by Wall Street trading desks, consistently beating the news wires.” Id. at 322-23 (internal quotation marks omitted). The cornerstone of Fly’s offerings is its online newsfeed, which it continually updates between 5:00 a.m. and 7:00 p.m. during days on which the New York Stock Exchange is open. The newsfeed typically streams more than 600 headlines a day in ten different categories, including “hot stocks,” “rumors,” “technical analysis,” and “earnings.” One such category is “recommendations.” There, Fly posts the recommendations (but not the underlying research reports or supporting analysis) produced by sixty-five investment firms’ analysts, including those at the plaintiff Firms. A typical Recommendation headline from 2009, for example, reads “EQIX: Equinox initiated with a Buy at BofA/Merrill. Target $110.” Id. at 323. Fly’s headlines, including those in the “recommendations” category, are searchable and sortable. Users can also subscribe to receive automated e-mail, pop-up, or audio alerts whenever Fly posts content relevant to preselected companies’ securities. Fly publishes most of its recommendation headlines before the New York Stock Exchange opens each business day at 9:30 a.m. Fly estimates that the Firms’ Recommendation headlines currently comprise approximately 2.5% of Fly’s total content, down from 7% in 2005. According to Fly, over time it has changed the way in which it obtains information about recommendations. Some investment firms, such as Wells Fargo’s investment services, will send Fly research reports directly as soon as they are released. Others, including the plaintiff Firms, do not. Until 2005, for recommendations of firms that do not, including the plaintiff Firms, Fly relied on employees at the investment firms (without the firms’ authorization)' to e-mail the research reports to Fly as they were released. Fly staff would summarize a recommendation as a headline (e.g., “EQIX initiated with a Buy at BofA/Merrill. Target $110.”). Sometimes Fly would include in a published item an extended passage taken verbatim from the underlying report. Fly maintains that because of threats of litigation in 2005, it no longer obtains recommendations directly from such investment firms. Instead, it gathers them using a combination of other news outlets, chat rooms, “blast IMs” sent by people in the investment community to hundreds of recipients, and conversations with traders, money managers, and its other contacts involved in the securities markets. Fly also represents that it no longer publishes excerpts from the research reports themselves, and now disseminates only the Recommendations, typically summarizing only the rating and price target for a particular stock. The Firms’ Response to The Threat Posed by Fly and Other Aggregators Because the value of the reports and Recommendations to an investor with early access to a Recommendation is in significant part derived from the informational advantage an early recipient may have over others in the marketplace, most of the trading the Firms generate based on their reports and Recommendations occurs in the initial hours of trading after the principal U.S. securities markets have opened. Such sales activity typically slackens by midday. The Firms’ ability to generate revenue from the reports and Recommendations therefore directly relates to the informational advantage they can provide to their clients. This in turn is related to the Firms’ ability to control the distribution of the reports and Recommendations so that the Firms’ clients have access to and can take action on the reports and Recommendations before the general public can. The Firms have employed a variety of measures in an attempt to stem the early dissemination of Recommendations to non-clients. Most of them have either been instituted or augmented relatively recently in response to the increasing availability of Recommendations from Fly and competing aggregators and news services. The Firms describe these steps as follows: The Firms have made a “very substantial and costly effort to study the unauthorized dissemination of their research reports and ... to plug the leaks they have found.” Merrill Lynch, for example, has: (a) worked with third-party vendors to limit access to Merrill Lynch clients; (b) employed an internal security program to detect breaches of security; (c) investigated Merrill Lynch employees, including a review of cell phones, for leaks to third parties; (d) internalized Merrill Lynch’s email subscription system; (e) identified and blacklisted websites that seek to post links to Merrill Lynch content; and (f) created unique signature URLs when links to research are sent to clients so that clients’ usage can be monitored and abuse tracked, [citation to record] (describing breach control as an “all-consuming task”). Barclays and Morgan Stanley have undertaken comparable measures to protect their research. Each Firm has a restrictive media and communications policy intended to preserve the time-sensitive value of Recommendations for their clients. The policies provide that any disclosure of equity research to the press occurs only after expiration of a prescribed period of time, and even then it is limited to entities that use the research as part of contextual news reporting and analysis. Appellees’ Br. at 13 (citations omitted). As outlined above, the district court also cataloged these efforts, emphasizing their increasing intensity “in recent years.” It is not clear from the record, however, the extent to which these efforts increased in response to the actions of Fly and others similarly disseminating the Recommendations on Internet-borne services, nor does the record disclose how successful the measures have been. Fly has not challenged the legality of the Firms’ anti-dissemination efforts in these proceedings. The Complaint and Pre-Trial District Court Proceedings In 2004, the Firms identified Fly as one of several entities systematically publishing the Recommendations without the Firms’ permission. Others doing the same included larger and better-known news outlets with far broader audiences, such as Bloomberg, Dow Jones, and Thomson Reuters. All of them regularly post short headlines reporting Recommendations soon after they become available. The Firms nonetheless focused their legal actions in this regard on Fly. In March and April 2005, the Firms complained to Fly that its publication of the Firms’ Recommendations in February and March of that year infringed the Firms’ copyrights and was tortious under New York State’s “hot news” misappropriation doctrine. The Firms demanded that Fly cease and desist. Fly’s counsel responded in April and May 2005, representing that Fly had altered its reporting practices so that it no longer obtained the Recommendations from research reports sent by employees of the Firms, instead gathering the information from independent, public sources. Fly continued posting the Firms’ Recommendations. On June 26, 2006, the Firms filed this suit naming Fly as the sole defendant. The Firms assert two causes of action in their complaint: copyright infringement based on Fly’s extensive excerpting of 17 research reports released in February and March 2005, and “hot news” misappropriation based on Fly’s continual electronic publication of the Firms’ Recommendations. The gravamen of the latter claim is that the aggregate widespread, unauthorized reporting of Recommendations by Fly and other financial news providers— including better known, better financed, more broadly accessed outlets — has threatened the viability of the Firms’ equity research operations. The Firms allege that this unauthorized distribution allows clients and prospective clients to learn of Recommendations from sources other than the Firms before the Firms’ sales staff can reach out to them to solicit their business, thereby reducing the ability of research to drive commission revenue. This, they assert, seriously threatens their ability to justify the expense of maintaining their extensive research operations. On August 16, 2006, Fly answered, raising several affirmative defenses, including “fair use” and protections purportedly afforded to it and its dissemination of news by the First Amendment. On May 18, 2009, after completion of discovery, the Firms and Fly cross-moved for summary judgement. The district court (Denise L. Cote, Judge) denied the summary judgment motions on November 6, 2009. The Firms then waived their claims for actual damages, and the court set the case for a bench trial. The Trial and The District Court Decision In a joint pre-trial order dated February 12, 2010, the parties stipulated to, among other things, the district court’s jurisdiction and the identification of the issues presented for trial. Joint Pre-Trial Order (Dkt. No. 167), Barclays Capital Inc. v. Theflyonthewall.com, No. 06-cv-4908 (S.D.N.Y. April 21, 2010) (the “Joint Pretrial Order”). The parties also agreed that: The following affirmative defenses previously asserted by Defendant are not to be tried: Defendant’s publication of daily news from firms in the financial industry, including Plaintiffs, is constitutionally protected by the First Amendment to the U.S. Constitution. Joint Pre-trial Order at 5 (emphasis in original). The district court read this to mean that Fly had waived any First Amendment defenses to the Firms’ “hot news” misappropriation claim. See Bar-clays Capital Inc. v. Theflyonthewall.com (“Fly II”), 700 F.Supp.2d 310, 352-54 (S.D.N.Y.2010) (Opinion and Order Denying Stay). Fly also abandoned the “fair use” copyright-infringement defense, thereby effectively conceding liability on the copyright claim. An injunction “which restrains Fly from further infringement of ‘any portion of the copyrighted elements of any research reports’ generated by Barclays Capital or Morgan Stanley,” Fly I, 700 F.Supp.2d at 331, was entered and, so far as we know, remains in effect. In the pre-trial order, the Firms contended that they satisfied all five “elements” of the tort purportedly identified in NBA, 105 F.3d at 845, although the Firms did not explicitly refer to that case. Joint Pre-Trial Order at 3. Fly appeared to concede that the Firms generate their Recommendations at great expense and that the Recommendations are time-sensitive, but disputed the other three “elements” of the misappropriation claim. Id. at 4. At a four-day bench trial in early March of last year, the witnesses for the plaintiffs were primarily Firm executives responsible for or familiar with a Firm’s research activities. The defendant called, inter alios, Fly employees to testify, including Fly’s President and majority owner, Ron Etergino. Inasmuch as Fly had effectively conceded liability for copyright infringement, the primary issues at trial were (1) the scope of remedies for copyright infringement, (2) whether Fly was liable for “hot news” misappropriation and, if so, (3) the appropriate remedy. On March 18, 2010, the district court issued its Opinion and Order, deciding for the plaintiffs on both the copyright-infringement and the “hot news” misappropriation claims. It awarded the plaintiffs statutory damages and attorney’s fees related to the copyright infringement claim. As part of its judgment in favor of the plaintiffs on the misappropriation claim, the court entered an order, inter alia, enjoining Fly from reporting Recommendations for a period ranging from thirty minutes to several hours after they are released by the plaintiffs. See Fly I, 700 F.Supp.2d at 348; Permanent Injunction (Dkt. No 138), Barclays Capital v. Theflyonthewall.com, No. 06-cv-4908 (S.D.N.Y. March 18, 2010) (the “Permanent Injunction”). Relying upon one of two — or arguably three — iterations of NBA’s multi-factor “test,” the district court concluded that for a misappropriation claim under New York law to survive federal copyright law preemption, and for the plaintiff to succeed on the claim, the plaintiff is required to demonstrate that: (i) [it] generates or gathers information at a cost; (ii) the information is time-sensitive; (iii) a defendant’s use of the information constitutes free riding on the plaintiffs efforts; (iv) the defendant is in direct competition with a product or service offered by the plaintiffs; and (v) the ability of other parties to free-ride on the efforts of the plaintiff or others would so reduce the incentive to produce the product'or service that its existence or quality would be substantially threatened. Fly I, 700 F.Supp.2d at 334-35 (quoting NBA, 105 F.3d at 845). The district court concluded that the first two “elements” — the cost of generating information and time-sensitivity — were not disputed by Fly and in any case were easily met. Id. at 335-36. The district court decided with respect to the third factor, “free riding,” that, “[i]n essence, [it] exists where a defendant invests little in order to profit from information generated or collected by the plaintiff at great cost.” Id. at 336. According to the court, “Fly does no equity research of its own, nor does it undertake any original reporting or analysis.” Id. at 336. In deciding in the Firms’ favor on this issue, the district court rejected Fly’s argument that its efforts in the collection, aggregation, and dissemination of information were sufficient to avoid a finding of free-riding, on the ground that efforts contributed nothing to the actual Recommendations that Fly provided to its subscribers. Id. at 336-37. The court also disagreed with Fly’s argument that its gathering of the Recommendations from public sources renders that information freely available for all: “[T]he fact that others also engage in unlawful behavior does not excuse a party’s own illegal conduct.” Id. at 337. In concluding that the fourth factor, direct competition, was present, the district court relied on its finding that both Fly and the Firms were engaged in “disseminating Recommendations to investors for their use in making investment decisions,” that production and distribution of the reports was among the Firms’ “primary businesses,” and that the companies used similar distribution channels. Id. at 339-40. The court also thought significant Fly’s then-recent attempts to link its subscribers to discount brokerage services, which in the district court’s view had the potential to further draw commission revenue away from the Firms. The district court rejected Fly’s contention that our decision in NBA required the court to find “head-to-head competition in a primary market,” concluding that neither Fly’s lack of brokerage and investment-advisory services nor its role as a news aggregator was inconsistent with a finding of direct competition. Id. at 340. The court appeared to conclude that Fly’s other activities were immaterial, so long as Fly, like the Firms, was engaged in the business of disseminating Recommendations. Finally, the district court concluded that the fifth factor, sufficiently reduced economic incentives, was present. The court found that “common sense and the circumstantial evidence about the plaintiffs’ business model make the Firms’ contentions about [their] reduced incentives utterly credible.” Id. at 342. The Finns had asserted that they had been forced to cut their analyst staffs and budgets significantly during the previous five years, in significant measure although by no means exclusively because of competition from unauthorized redistributions of their Recommendations. They acknowledged, as did the court, that there were unrelated substantial causes for the contraction during this period, including the then-recent recession and accompanying stock-market collapse, and the April 2003 Global Research Analyst Settlement. Fly sought to portray the Firms’ evidence of reduced economic incentives, which was based almost entirely on the testimony of the Firms’ own research executives, as speculative and self-serving. The district court concluded to the contrary (1) that the executives’ testimony was credible despite their employment by the Firms, (2) that the Firms did not need to demonstrate actual harm, but rather merely show that harm would occur if Fly and others were allowed to continue their conduct, and (3) that the precise impact of the recent recession and the Global Research Analyst Settlement was irrelevant, because the mere showing that Fly and others like it significantly affected the Firms’ incentives was sufficient to establish the fifth factor, even if other events also contributed to the reduction in incentives. Id. at 342-43. Having concluded that the Firms had established the tort of “hot news” misappropriation, the district court entered a permanent injunction barring Fly from reporting a Recommendation until either (a) half an hour after the market opens, if the report containing the recommendation was released before 9:30 a.m., or (b) two hours after release, if the report was released after 9:30 a.m. This time period represented roughly the midpoint between what Fly and the Firms, respectively, requested. Perhaps because Fly purported to waive its First Amendment defenses, the district court’s opinion contains no explicit discussion of First Amendment doctrine beyond the court’s reference, in consideration of the propriety of injunctive relief, to “public policy considerations,” and the balancing of “the public interest in unrestrained access to information.” Id. at 344. Similarly, although in the court’s thorough recitation of the history of the law of “hot news” misappropriation, it explained in some detail the role of Copyright Act preemption of state tort law, it did not expressly consider whether “hot news” misappropriation was preempted by federal copyright law in this case. Instead, it adopted as determinative NBA’s ruling that a narrow form of the “hot news” misappropriation tort survives preemption, and it applied language from that decision indicating the tort’s limitations by virtue of preemption doctrine. Postr-Trial Procedural History Fly filed a notice of appeal on April 9, 2010. Four days later, it moved before the district court to stay or modify the injunction pending that appeal. In support of its motion, Fly argued (1) that it was likely to succeed on the merits on appeal, specifically with regard to the direct competition and reduced incentive elements of the misappropriation claim; (2) that it would suffer irreparable harm from the operation of the injunction as customers cancelled their subscriptions; and (3) that the injunction’s curtailment of First Amendment protected speech required a finding of irreparable harm. On May 7, 2010, the district court denied Fly’s motion on the grounds (1) that Fly was not likely to succeed on the merits of its appeal; (2) that Fly had shown evidence of only two instances of customers cancelling their subscriptions because of the injunction; and (3) that Fly had waived its First Amendment arguments prior to trial. See Fly II, 700 F.Supp.2d at 349-56. Fly thereupon moved in this Court for a stay of the injunction and an expedited appeal. On May 19, 2010, a panel of this Court granted the motion. On appeal, Fly argues principally that (1) the district court erred in finding that the plaintiffs established “hot news” misappropriation under New York law, specifically in that the plaintiffs failed to prove time-sensitivity, free-riding, direct competition, and reduced incentives; (2) that the district court’s injunction violates Fly’s free-speech rights under the First Amendment; (3) that the district court’s finding of “hot news” misappropriation violates the Copyright Clause of the Constitution and the Copyright Act; (4) that the district court failed to apply the proper standard in granting injunctive relief; and (5) that the injunction is unreasonably overbroad. DISCUSSION I. Standard of Review “When reviewing a judgment following a bench trial in the district court, we review the court’s findings of fact for clear error and its conclusions of law de novo.” Tiffany (NJ) Inc. v. eBay Inc., 600 F.3d 93, 96 (2d Cir.), cert, denied, — U.S.-, 131 S.Ct. 647,178 L.Ed.2d 513 (2010). II. Viability of the “Hot News” Misappropriation Tort Amici Google, Inc. and Twitter, Inc., referring to the “hot news” misappropriation tort as an “end-run” around the Constitution’s Copyright Clause and Supreme Court precedent, and arguing that their position is supported by “[ijmportant public policy concerns,” urge us to “repudiate the tort.” Brief for Google, Inc. and Twitter, Inc. as Amici Curiae Supporting Reversal at 3, Barclays Capital Inc. v. Theflyonthewall.com, No. 10-1372-cv (2d Cir. June 22, 2010). We need not address the viability vel non of a “hot news” misappropriation tort under New York law. Were we to do so, though, plainly we would be bound by the conclusion of the previous Second Circuit panel in NBA that the tort survives. See, e.g., United States v. Jass, 569 F.3d 47, 58 (2d Cir.2009) (explaining the binding nature of one panel opinion on a subsequent panel of the same circuit); Meacham v. Knolls Atomic Power Lab., 461 F.3d 134, 141 (2d Cir.2006) (similar), rev’d on other grounds, 554 U.S. 84, 128 S.Ct. 2395, 171 L.Ed.2d 283 (2008). We are therefore without the authority to “repudiate” that view. Were we indeed called upon to consider the continued viability of the tort under New York law, perhaps we would certify that issue to the New York Court of Appeals. The issue we address, however, is federal preemption. As a federal court, we answer that question ourselves. III.Copyright Act Preemption A. National Basketball Association v. Motorola, Inc. National Basketball Association v. Motorola, Inc., 105 F.3d 841 (2d Cir.1997), appears to be the only judicial decision— surely the only decision binding upon us— that addresses directly the preemption issue raised in this appeal. There, defendant Motorola, Inc. produced and sold (or otherwise provided) to members of the public a telephonic pager called SportsTrax. Motorola’s co-defendant, STATS, Inc., supplied statistical information about National Basketball Association (“NBA”) professional basketball games. The information was transmitted to SportsTrax pagers owned or leased by Motorola and STATS customers roughly simultaneously with the playing of the games. NBA, 105 F.3d at 843. The information included “(i) the teams playing; (ii) score changes; (iii) the team in possession of the ball; (iv) whether the team is in the free-throw bonus; (v) the quarter of the game; and (vi) time remaining in the quarter.” Id. at 844. The information [was] updated every two to three minutes, with more frequent updates near the end of the first half and the end of the game. There [was] a lag of approximately two or three minutes between events in the game itself and when the information appeared] on the pager screen. Id. SportsTrax gathered the information for the service by employing persons who would watch the games on television or listen to accounts of them on the radio and supply the information to STATS’s host computer. The computer compiled, analyzed, and formatted the data for retransmission. The information was then sent to FM radio stations which retransmitted them to the subscribers’ individual SportsTrax pagers. Id. The NBA itself also publicly disseminated similar, and therefore to some extent competitive, information. As Judge Winter wrote for the NBA panel: [T]he NBA does provide, or will shortly do so, information like that available through SportsTrax. It now offers a service called “Gamestats” that provides official play-by-play game sheets and half-time and final box scores within each arena. It also provides such information to the media in each arena. In the future, the NBA plans to enhance Gamestats so that it will be networked between the various arenas and will support a pager product analogous to SportsTrax. SportsTrax will of course directly compete with an enhanced Gamestats. Id. at 853. The district court whose decision was on appeal in NBA had found for the plaintiff on its New York-law “hot news” misappropriation claim arising out of the defendants’ taking, redistributing, and profiting from the facts generated by the NBA in the course of the playing of NBA games. The district court therefore had entered a permanent injunction against the defendants, but stayed that injunction pending appeal. Id. 1. NBA Preemption Analysis. a. Copyright Act The NBA panel began its analysis by noting that prior to the 1976 amendments to the Copyright Act, the Act contained no express provisions as to the circumstances under which the federal copyright law preempted state law. The 1976 Amendments changed that. Title 17 U.S.C. § 301, enacted in 1976, sets forth a two-part test to determine whether a state-law claim is preempted by the Copyright Act, with a further “extra elements” exception we discuss below. Such a claim is preempted (i) if it seeks to vindicate “legal or equitable rights that are equivalent” to one of the bundle of exclusive rights already protected by copyright law under 17 U.S.C. § 106 — -the “general scope requirement”; and (ii) if the work in question is of the type of works protected by the Copyright Act under 17 U.S.C. §§ 102 and 108 — the “subject matter requirement.” NBA, 105 F.3d at 848 (quoting 17 U.S.C. § 301). The NBA panel observed that “[t]he subject matter requirement” — the second factor in a preemption analysis — “is met when the work of authorship being copied or misappropriated 'falls within the ambit of copyright protection.’ ” Id. at 849 (quoting Harper & Row, Inc. v. Nation Enters., 723 F.2d 195, 200 (2nd Cir.1983) (brackets omitted), rev’d on other grounds, 471 U.S. 539, 105 S.Ct. 2218, 85 L.Ed.2d 588 (1985)). In deciding whether a state-law claim is preempted by the Copyright Act, then, it is not determinative that the plaintiff seeks redress with respect to a defendant’s alleged misappropriation of uncopyrightable material — e.g., facts — contained in a copyrightable work. “Copyrightable material often contains uncopyrightable elements within it, but Section 301 preemption bars state law misappropriation claims with respect to uncopyrightable as well as copyrightable elements,” if the work as a whole satisfies the subject matter requirement. NBA, 105 F.3d at 849; see also id. at 850 (quoting ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1453 (7th Cir.1996)). In NBA, facts about what transpired during broadcasted NBA basketball games thus fell within the subject matter of copyright for the purpose of the court’s preemption analysis, even though the games themselves were not copyrightable. Id. at 848^49 (“Although game broadcasts are copyrightable while the underlying games are not, the Copyright Act should not be read to distinguish between the two when analyzing the preemption of a misappropriation claim based on copying or taking from the copyrightable work.”). Turning to the other preemption element, the NBA panel thought it clear that what the NBA was seeking to protect fell within the “general scope of copyright.” Title 17 U.S.C. § 106, which states that the general scope of copyright, “affords a copyright owner the exclusive right to: (1) reproduce the copyrighted work; (2) prepare derivative works; (3) distribute copies of the work by sale or otherwise; and, with respect to certain artistic works, (4) perform the work publicly; and (5) display the work publicly. See 17 U.S.C. 106(l)-(5).” Computer Assocs. Int’l, Inc. v. Altai Inc., 982 F.2d 693, 716 (2d Cir.1992). “Section 301 [of the Copyright Act] thus preempts only those state law rights that ‘may be abridged by an act which, in and of itself, would infringe one of the exclusive rights’ provided by federal copyright law,” id. (quoting Harper & Row, 723 F.2d at 200), i.e., “acts of reproduction, performance, distribution or display,” id. (internal quotation marks omitted). The claim of tortious behavior in NBA was indeed for the acts of reproduction, distribution, and display of facts by the defendants of material taken from the copyrighted broadcasts. The NBA panel therefore concluded that the plaintiffs tort claim was within the general scope of copyright. The court was thus satisfied that both preemption factors were met. b. Extra-Element Test Having decided that the two preliminary factors counseled in favor of preemption, the NBA panel observed: [C]ertain forms of commercial misappropriation otherwise within the general scope requirement will survive preemption if an “extra-element” test is met. As stated in Altai: But if an “extra element” is “required instead of or in addition to the acts of reproduction, performance, distribution or display, in order to constitute a state-created cause of action, then the right does not lie ‘within the general scope of copyright,’ and there is no preemption.” Altai 982 F.2d at 716 (quoting 1 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 1.01[B] at 1-14-15 (1991)). NBA, 105 F.3d at 850; see also Harper & Row, 723 F.2d at 200 (“[W]hen a state law violation is predicated upon an act incorporating elements beyond mere reproduction or the like, the rights involved are not equivalent and preemption will not occur.”). It is with respect to the “extra elements” that the NBA Court proffered a three-factor analysis: “We ... find the extra elements — those in addition to the elements of copyright infringement — that allow a ‘hotnews’ claim to survive preemption are: (i) the time-sensitive value of factual information, (ii) the free-riding by a defendant, and (iii) the threat to the very existence of the product or service provided by the plaintiff.” Id. at 853 (emphasis added). i. International News Service v. Associated Press The NBA Court briefly summarized the Supreme Court’s seminal 1918 “hot news” decision, International News Service v. Associated Press, 248 U.S. 215, 39 S.Ct. 68, 63 L.Ed. 211 (1918) (“INS ”): INS involved two wire services, the Associated Press (“AP”) and International News Service (“INS”), that transmitted news stories by wire to member newspapers. Id. INS would lift factual stories from AP bulletins and send them by wire to INS papers. Id. at 231 [39 S.Ct. 68]. INS would also take factual stories from east coast AP papers and wire them to INS papers on the west coast that had yet to publish because of time differentials. Id. at 238 [39 S.Ct. 68]. The Supreme Court held that INS’s conduct was a common-law misappropriation of AP’s property. Id. at 242 [39 S.Ct. 68]. NBA 105 F.3d at 845. INS itself is no longer good law. Purporting to establish a principal of federal common law, the law established by INS was abolished by Ene Railroad Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), which largely abandoned federal common law. But, as the NBA panel pointed out, “[b]ased on legislative history of the 1976 [Copyright Act amendments], it is generally agreed that a ‘hot-news’ INS-like claim survives preemption.” NBA 105 F.3d at 845 (citing H.R.Rep. No. 94-1476 at 132). The House of Representatives Report with respect to the preemption provisions of the 1976 Copyright Act amendments commented in this regard: “Misappropriation” is not necessarily synonymous with copyright infringement, and thus a cause of action labeled as “misappropriation” is not preempted if it is in fact based neither on a right within the general scope of copyright as specified by [17 U.S.C. § ] 106 [specifying the general scope of copyright] nor on a right equivalent thereto. For example, state law should have the flexibility to afford a remedy (under traditional principles of equity) against a consistent pattern of unauthorized appropriation by a competitor of the facts (i.e., not the literary expression) constituting “hot” news, whether in the traditional mold of [INS], or in the newer form of data updates from scientific, business, or financial data bases. H.R. No. 94-1476 at 132, reprinted in 1976 U.S.C.C.A.N. at 5748 (footnote omitted), quoted in NBA 105 F.3d at 850. The House Report thus anticipated that INS-like state-law torts would survive preemption. It did not itself create such a cause of action or recognize the existence of one under federal law. It allowed instead for the survival of such a state-law claim. The NBA Court thus used INS as a description of the type of claims — “INS-like” — that, Congress has said, are not necessarily preempted by federal copyright law. Some seventy-five years after its death under Erie, INS thus maintains a ghostly presence as a description of a tort theory, not as precedential establishment of a tort cause of action. ii. Moral Dimensions One source of confusion in addressing these misappropriation cases is that INS itself was a case brought in equity to enjoin INS from copying AP’s uncopyrightable news. In that context, the INS Court emphasized the unfairness of INS’s practice of pirating AP’s stories. It condemned, in what sounded biblical in tone, the defendant’s “reaping] where it ha[d] not sown.” INS, 248 U.S. at 239, 39 S.Ct. 68. The Court said: This defendant ... admits that it is taking material that has been acquired by complainant as the result of organization and the expenditure of labor, skill, and money, and which is salable by complainant for money, and that defendant in appropriating it and selling it as its own is endeavoring to reap where it has not sown, and by disposing of it to newspapers that are competitors of complainant’s members is appropriating to itself the harvest of those who have sown. Stripped of all disguises, the process amounts to an unauthorized interference with the normal operation of complainant’s legitimate business precisely at the point where the profit is to be reaped, in order to divert a material portion of the profit from those who have earned it to those who have not; with special advantage to defendant in the competition because of the fact that it is not burdened with any part of the expense of gathering the news. The transaction speaks for itself, and a court of equity ought not to hesitate long in characterizing it as unfair competition in business. Id. at 239-40, 39 S.Ct. 68 (emphasis added). This dicta has been absorbed by New York misappropriation law: New York courts have noted the incalculable variety of illegal practices falling within the unfair competition rubric, calling it a broad and flexible doctrine that depends more upon the facts set forth than in most causes of action. It has been broadly described as encompassing any form of commercial immorality, or simply as endeavoring to reap where one has not sown; it is taking the skill, expenditures and labors of a competitor, and misappropriating for the commercial advantage of one person a benefit or property right belonging to another. The tort is adaptable and capacious. Roy Exp. Co. Establishment of Vaduz, Liech. v. Columbia Broad. Sys., Inc., 672 F.2d 1095, 1105 (2d Cir.1982) (citation and alteration omitted). And it has been reflected in the rhetoric of federal district courts applying New York law. See, e.g., Fly I, 700 F.Supp.2d at 336 (quoting INS); NBA v. Sports Team Analysis & Tracking Sys. (“NBA SDNY”), 939 F.Supp. 1071,1075 (S.D.N.Y.1996) (quoting INS), rev’d, NBA, 105 F.3d 841. The NBA Court also noted that the district court whose decision it was reviewing had “described New York misappropriation law as standing for the ‘broader principle that property rights of commercial value are to be and will be protected from any form of commercial immorality’; that misappropriation law developed ‘to deal with business malpractices offensive to the ethics of [ ] society’; and that the doctrine is ‘broad and flexible.’ ” NBA, 105 F.3d at 851 (brackets in original) (quoting NBA SDNY, 939 F.Supp. at 1098-1110) (internal citation omitted). But Judge Winter explicitly rejected the notion that “hot news” misappropriation cases based on the disapproval of the perceived unethical nature of a defendant’s ostensibly piratical acts survive preemption. The Court concluded that “such concepts are virtually synonymous [with] wrongful copying and are in no meaningful fashion distinguishable from infringement of a copyright. The broad misappropriation doctrine relied upon by the district court is, therefore, the equivalent of exclusive rights in copyright law.” NBA, 105 F.3d at 851 (deeming preempted the broad theory of misappropriation embodied in Metropolitan Opera Ass’n v. Wagner-Nichols Recorder Corp., 199 Misc. 786, 101 N.Y.S.2d 483 (N.Y. County Sup. Ct.1950), affd, 279 A.D. 632, 107 N.Y.S.2d 795 (1st Dep’t 1951)). No matter how “unfair” Motorola’s use of NBA facts and statistics may have been to the NBA — or Fly’s use of the fact of the Firms’ Recommendations may be to the Firms — then, such unfairness alone is immaterial to a determination whether a cause of action for misappropriation has been preempted by the Copyright Act. The adoption of new technology that injures or destroys present business models is commonplace. Whether fair or not, that cannot, without more, be prevented by application of the misappropriation tort. Indeed, because the Copyright Act itself provides a remedy for wrongful copying, such unfairness may be seen as supporting a finding that the Act preempts the tort. See id. iii. Narrowness of the Preemption Exception The NBA panel repeatedly emphasized the “narrowness” of the “hot news” tort exception from preemption. See id. at 843, 848, 851, 852 (using the word “narrow” or “narrowness” five times). Although our discussion of preemption in NBA did not focus on the importance of maintaining the uniform nationwide scheme that the Copyright Act, with its 1976 preemption amendment, 17 U.S.C. § 301, provides, we later underscored it. In Krause v. Titleserv, Inc., 402 F.3d 119, 123 (2d Cir.2005), we declined to limit protection for copyrights held by “owners” of computer programs to those with formal title to such programs. The first reason we gave was that title may depend on state law that differs from one state to another. The result would be to undermine some of the uniformity achieved by the Copyright Act.... If [the relevant section of the Copyright Act] required formal title, two software users, engaged in substantively identical transactions might find that one is liable for copyright infringement while the other is protected by [the section], depending solely on the state in which the conduct occurred. Such a result would contradict the Copyright Act’s “express objective of creating national, uniform copyright law by broadly preempting state statutory and common-law copyright regulation. ” Community for Creative Non-Violence v. Reid, 490 U.S. 730, 740, 109 S.Ct. 2166, 104 L.Ed.2d 811 (1989); see also 17 U.S.C. § 301(a). Id. at 123 (emphasis added). Indeed, central to the principle of preemption generally is the value of providing for legal uniformity where Congress has acted nationally. See, e.g., Paneccasio v. Unisource Worldwide, Inc., 532 F.3d 101, 113 (2d Cir.2008) (“The purpose of ERISA preemption is to ensure that all covered benefit plans will be governed by unified federal law, thus simplifying life for employers administering plans in several states, because a patchwork scheme of regulation would introduce considerable inefficiencies in benefit program operation.” (internal quotation marks and brackets omitted)). This is a pressing concern when considering the “narrow” “hot news” misappropriation exemption from preemption. The broader the exemption, the greater the likelihood that protection of works within the “general scope” of the copyright and of the type of works protected by the Act will receive disparate treatment depending on where the alleged tort occurs and which state’s law is found to be applicable. The problem may be illustrated by reference to a recent case in the Southern District of New York. In Associated Press v. All Headline News Corp., 608 F.Supp.2d 454 (S.D.N.Y.2009), the court sought to determine whether there was a difference between New York and Florida “hot news” misappropriation law in order for it to analyze, under choice-of-law principles, which state’s law applied. Judge Castel observed that “[n]o authority has been cited to show that Florida recognizes a cause of action for hot news misappropriation. Then again, defendants have not persuasively demonstrated that Florida would not recognize such a claim.” Id. at 459-60. It appears, then, that the alleged “hot news” misappropriation in All Headline News Corp. might have been permissible in New York but not in Florida. The same could have been said for the aggregation and publication of basketball statistics in NBA and the same may be said as to the aggregation and publication of Recommendations in the case at bar. To the extent that “hot news” misappropriation causes of action are not preempted, the aggregators’ actions may have different legal significance from state to state — permitted, at least to some extent, in some; prohibited, at least to some extent, in others. It is this sort of patchwork protection that the drafters of the Copyright Act preemption provisions sought to minimize, and that counsels in favor of locating only a “narrow” exception to Copyright Act preemption. c. Three- and Five-Part “Tests” Before concluding that the NBA’s claim was preempted, the NBA panel set forth in its opinion — twice—a five-part “test” for identifying a non-preempted “hot news” misappropriation claim. The district court in this case, when applying NBA structured its conclusions-of-law analysis around NBA’s first iteration of the “test”: We hold that the surviving “hot-news” INS-Hke claim is limited to cases where: (i) a plaintiff generates or gathers information at a cost; (ii) the information is time-sensitive; (iii) a defendant’s use of the information constitutes free-riding on the plaintiffs efforts; (iv) the defendant is in direct competition with a product or service offered by the plaintiffs; and (v) the ability of other parties to free-ride on the efforts of the plaintiff or others would so reduce the incentive to produce the product or service that its existence or quality would be substantially threatened. We conclude that SportsTrax does not meet that test. NBA 105 F.3d at 845; see Fly I, 700 F.Supp.2d at 334-35 (quoting the passage but omitting the first fifteen prefatory words). But the panel restated the five-part inquiry later in its opinion: In our view, the elements central to an INS claim are: (i) the plaintiff generates or collects information at some cost or expense, see [Financial Information, Inc. v. Moody’s Investors Serv., 808 F.2d 204, 206 (2d Cir.1996) (“FII”)]; INS, 248 U.S. at 240, 39 S.Ct. 68; (ii) the value of the information is highly time-sensitive, see FII, 808 F.2d at 209; INS, 248 U.S. at 231, 39 S.Ct. 68; Restatement (Third) Unfair Competition, § 38 cmt. c.; (iii) the defendant’s use of the information constitutes free-riding on the plaintiffs costly efforts to generate or collect it, see FII, 808 F.2d at 207; INS, 248 U.S. at 239-40, 39 S.Ct. 68; Restatement § 38 at cmt. c.; McCarthy, § 10:73 at 10-139; (iv) the defendant’s use of the information is in direct competition with a product or service offered by the plaintiff, FII, 808 F.2d at 209, INS, 248 U.S. at 240, 39 S.Ct. 68; (v) the ability of other parties to free-ride on the efforts of the plaintiff would so reduce the incentive to produce the product or service that its existence or quality would be substantially threatened, FII, 808 F.2d at 209; Restatement, § 38 at cmt. c.; INS, 248 U.S. at 241, 39 S.Ct. 68 (“[INS’s conduct] would render [AP’s] publication profitless, or so little profitable as in effect to cut off the service by rendering the cost prohibitive in comparison with the return.”). NBA, 105 F.3d at 852. Throughout this litigation the parties seem to have been in general agreement that the district court and we should employ a five-part analysis taken from the NBA opinion. It is understandable, of course, that counsel and the district court did in this case, and do in other comparable circumstances, attempt to follow our statements in precedential opinions as to what the law is — which we often state in terms of what we “hold.” But that reading is not always either easy to make or technically correct. As Judge Friendly put it in colorful terms: “A judge’s power to bind is limited to the issue that is before him; he cannot transmute dictum into decision by waving a wand and uttering the word ‘hold.’ ” United States v. Rubin, 609 F.2d 51, 69 (2d Cir.1979) (Friendly, J., concurring), quoted in Pierre N. Leval, Judging Under the Constitution: Dicta about Dicta, 81 N.Y.U. L.Rev. 1249, 1249 (2006). See also generally Leval, supra (containing seminal discussion of judicial use of the term “holding”); id. at 1256 (“A dictum [i.e., a conclusion or point of view in an opinion that is not a holding] is an assertion in a court’s opinion of a proposition of law [that] does not explain why the court’s judgment goes in favor of the winner.”); Judith M. Stinson, Why Dicta Becomes Holding and Why it Matters, 76 Brook. L.Rev. 219, 219 n. 2 (2010) (collecting authorities addressing difficulties with judicial use of the term “hold”). It is axiomatic that appellate judges cannot make law except insofar as they reach a conclusion based on the specific facts and circumstances presented to the court in a particular appeal. Subordinate courts and subsequent appellate panels are required to follow only these previous appellate legal “holdings.” The NBA panel decided the case before it, and we think that the law it thus made regarding “hot news” preemption is, as we have tried to explain, determinative here. But the Court’s various explanations of its five-part approach are not. Indeed, we do not see how they can be: The two five-part “tests” are not entirely consistent, and are less consistent still with the three-“extra element” test, which also appears later in the opinion: We therefore find the extra elements— those in addition to the elements of copyright infringement — that allow a “hotnews” claim to survive preemption are: (i) the time-sensitive value of factual information, (ii) the free-riding by a defendant, and (iii) the threat to the very existence of the product or service provided by the plaintiff.” Id. at 853. For example, the fifth of the five factors in the first iteration of the test is that “the ability of other parties to free-ride on the efforts of the plaintiff or others would so reduce the incentive to produce the product or service that its existence or quality would be substantially threatened.” NBA, 105 F.3d at 845 (emphasis added). The second iteration is similar, but adds a quotation from INS which can be read to make the factor far more difficult to demonstrate: that the conduct “would render [the plaintiffs] publication profitless, or so little profitable as in effect to cut off the service by rendering the cost prohibitive in compariso