In Part I of this analysis of the FCC’s Report and Order on “Preserving the Open Internet,” I reviewed the Commission’s justification for regulating broadband providers. In Part II, I looked at the likely costs of the order, in particular the hidden costs of enforcement. In this part, I compare the text of the final rules with earlier versions. Next, I’ll look at some of the exceptions and caveats to the rules—and what they say about the true purpose of the regulations.
In the end, the FCC voted to approve three new rules that apply to broadband Internet providers. One (§8.3) requires broadband access providers to disclose their network management practices to consumers. The second One (§8.4) prohibits blocking of content, applications, services, and non-harmful devices. The third One (§8.5) forbids fixed broadband providers (cable and telephone, e.g.) from “unreasonable” discrimination in transmitting lawful network traffic to a consumer.
There has of course been a great deal of commentary and criticism of the final rules, much of it reaching fevered pitch before the text was even made public. At one extreme, advocates for stronger rules have rejected the new rules as meaningless, as “fake net neutrality,” “non neutrality,” or the latest evidence that the FCC has been captured by the industries it regulates. On the other end, critics decry the new rules as a government takeover of the Internet, censorship, and a dangerous and unnecessary interference with a healthy digital economy. (I agree with that last one.)
One thing that has not been seriously discussed, however, is just how little the final text differs from the rules originally proposed by the FCC in October, 2009. Indeed, many of those critical of the weakness of the final rules seem to forget their enthusiasm for the initial draft, which in key respects has not changed at all in the intervening year of comments, conferences, hearings, and litigation.
The differences—significant and trivial—that have been made can largely be traced to comments the FCC received on the original draft, as well as interim proposals made by industry and Congress, particularly the framework offered by Verizon and Google in August and a bill circulated by Rep. Henry Waxman just before the mid-term elections.
Compare, for example, the final text of the transparency rule with the version first proposed by the FCC.
Subject to reasonable network management, a provider of broadband Internet access service must disclose such information as is reasonably required for users and content, application and service providers to enjoy the protections specified in this part. (Proposed)
A person engaged in the provision of broadband Internet access service shall publicly disclose accurate information regarding the network management practices, performance and commercial terms of its broadband Internet access service sufficient for consumers to make informed choices regarding use of such services and for content, application, service and device providers to develop, market and maintain Internet offerings. (Final)
The final rule is much stronger, and makes clearer what it is that must be disclosed. It is also not subject to the limits of reasonable network management, Rather than the vague requirement of the draft for disclosures sufficient to “enjoy the protections” of the open Internet rules, the final rule requires disclosures sufficient for consumers to make “informed choices” about the services they pay for, a standard more easily enforced.
By comparison, the final rule comes close to the version that appeared in draft legislation circulated but never introduced by Rep. Henry Waxman in October of 2010. It likewise reflects the key concepts in the Verizon-Google Legislative Framework Proposal from earlier in the year.
As the Report makes clear (¶¶ 53-61), the transparency rule has teeth. Though the agency declines for now from making specific decisions about the contents of the disclosure and how is must be posted, the Report lays out a non-exhaustive list of nine major categories of disclosure, including network practices, performance characteristics, and commercial terms, that must be included. It’s hard to imagine a complying disclosure that will not run to several pages of very small text.
That generosity, of course, may be the rule’s undoing. As anyone who has ever thrown away a required disclosure from a service provider (mortgage, bank, drug, electronic device, financial statement, privacy, etc.) knows full well, information “sufficient” to make an informed choice is far more information than any non-expert consumer could possibly absorb and evaluate, even if they wanted to. The more information consumers are given, the less likely they’ll pay attention to any of it, including what may be important.
The FCC recognizes that risk, however, but believes it has an answer. “A key purpose of the transparency rule,” the Commission notes (¶ 60), “is to enable third-party experts such as independent engineers and consumer watchdogs to monitor and evaluate network management practices, in order to surface concerns regarding potential open Internet violations.”
Perhaps the agency has in mind here organizations like BITAG, which has been established by a wide coalition of participants in the Internet ecosystem to develop “consensus on broadband network management practices or other related technical issues.” Or by consumer watchdogs, perhaps the agency imagines that some of the public interest groups who have most strenuously rallied for the rules will become responsible stewards of their implementation, trading the acid pens of political rhetoric for responsible analysis and advocacy to their members and other consumers.
We’ll see. I wish I shared the Commissions confidence that, “for a number of reasons” (none cited), “the costs of the disclosure rule we adopt today are outweighed by the benefits of empowering end users and edge providers to make informed choices….” (¶ 59). But I don’t. Onward.
The final version of the blocking rule (§8.5) consolidated the Content, Applications and Services and Devices rule of the original draft. The final rule states:
A person engaged in the provision of fixed broadband Internet access services, insofar as such person is so engaged, shall not block lawful content, applications, services or non-harmful devices, subject to reasonable network management.
A more limited rule applies to mobile broadband providers, who
[S]hall not block consumers from accessing lawful websites, subject to reasonable network management, nor shall such person block applications that compete with the providers’ voice or video telephony services, subject to reasonable network management
Much of the anguish over the final rules that has been published so far relates to a few of the limitations built into the blocking rule. First, copyright-reform activists object to the word “lawful” appearing in the rule. “Lawful” content, applications, and services do not include activities that constitute copyright and trademark infringement. Therefore, the rule allows broadband providers to use whatever mechanisms they want (or may be required to) to reduce or eliminate traffic that involves illegal fire-sharing, spam, viruses and other malware, and the like.
A provider who blocks access to a site selling unlicensed products, in other words, is not violating the rules. And as the agency finds it is “generally preferable to neither require not encourage broadband providers to examine Internet traffic in order to discern which traffic is subject to the rules” (¶ 48), there will be considerable margin of error given to providers who block sites, services, or applications which may include some legal components.
On this view, though the FCC otherwise contradicts it—see footnote 245 and elsewhere—a complete ban on the BitTorrent protocol, for better or worse, might not be a violation of the blocking rule. Academic studies have shown that over 99% of BitTorrent traffic constitutes unlicensed file sharing of protected content. Other than inspecting individual torrents, which the agency disfavors, how else can an access provider determine what tiny minority of BitTorrent traffic is in fact lawful?
A second concern is the repeated caveat for “reasonable network management,” which gives access providers leeway to balance traffic during peak times, limit users whose activity may be harming other users, and other “legitimate network management” purposes.
Finally, disappointed advocates object to the special treatment for mobile broadband, which may, for example, block applications, services or devices without violating the rule. There is an exception to the exception for applications, such as VoIP and web video, that compete with the provider’s own offerings, but that special treatment doesn’t keep mobile providers from using “app stores” to exclude services they don’t approve. (See ¶ 102)
Of course even the original draft of the rules included the limitation for “reasonable network management,” and refused to apply any of the rules to unlawful activities. The definition of “reasonable network management” in the original draft is different, but functionally equivalent, to the final version.
The carve-out for mobile broadband, however, is indeed a departure from the original rules. Though the Oct. 2009 Notice of Proposed Rulemaking expressed concern about applying the same rule to fixed and mobile broadband (see 13, 154-174), the draft blocking rule did not distinguish between fixed and mobile Internet access. The FCC did note, however, that different technologies “may require differences in how, to what extent, and when the principles apply.” The agency sought comment on these differences (and asked for further comment in a later Notice of Inquiry). Needless to say, they heard plenty.
Wireless broadband is, of course, a newer technology, and one still very much in development. Spectrum is limited, and capacity cannot easily be added. Those are not so much market failures as they are regulatory failures. The FCC is itself responsible for managing the limited radio spectrum, and has struggled by its own admission to allocate spectrum for its most efficient and productive uses—indeed, even to develop a complete inventory of who has which frequencies of licensed spectrum today.
Adding additional capacity is another regulatory obstacle. Though mobile users rail against their providers for inadequate or unreliable coverage, no one, it seems, wants to have cellular towers and other equipment near where they live. Local regulators, who must approve new infrastructure investments, take such concerns very much to heart. (There is also rampant corruption and waste in the application, franchising, and oversight processes at the state and local levels, a not-very-secret secret.)
The FCC, it seems, has taken these concerns into account in the final rule. Its original open Internet policy statements—from which the rules derive—applied only to fixed broadband access, and the October, 2009 draft’s inclusion of mobile broadband came as a surprise to many.
The first indication that the agency was considering a return to the original open Internet policy came with the Verizon-Google proposal, where the former net neutrality adversaries jointly released a legislative framework (that is, something they hoped Congress, not the FCC, would take seriously) that gave different treatment to mobile. As the V-G proposal noted, “Because of the unique technical and operational characteristics of wireless networks, and the competitive and still-developing nature of wireless broadband services, only the transparency principle would apply to wireless at this time.”
The Waxman proposal didn’t go as far as V-G, however, adding a provision that closely tracks with the final rule. Under the Waxman bill, mobile providers would have been prohibited from blocking “lawful Internet websites”, and applications “that compete with the providers’ voice or video communications services.”
So the trajectory of the specialized treatment for mobile broadband is at least clear and, for those following the drama, entirely predictable. Yet the strongest objections to the final rule and the loudest cries of betrayal from neutrality advocates came from the decision to burden mobile providers less than their fixed counterparts. (Many providers offer both, of course, so will be subject to different rules for different parts of their service.)
At the very least, the advocates should have seen it coming. Many did. A number of “advocacy” groups demonized Google for its cooperation with Verizon, and refused to support Waxman’s bill. (It should also be noted that none of the groups objecting to the final rules or any interim version ever actually proposed their own version—that is, what they actually wanted as opposed to what they didn’t want.)
3. Unreasonable discrimination
The final rule, applicable only to fixed broadband providers, demands that a provider not “unreasonably discriminate in transmitting lawful network traffic over a consumer’s broadband Internet access service.” (§ 8.7, and see ¶¶ 68-79 of the Report).
Though subtle, the difference in language between the NPRM and the final rule are significant, as the FCC acknowledges. The NPRM draft rule noted plainly that “a provider of broadband Internet access service must treat lawful content, applications, and services in a nondiscriminatory manner.”
The difference here is between “nondiscrimination,” which prohibits all forms of differential network treatment, and “unreasonable discrimination,” which allows discrimination so long as it is reasonable.
The migration from a strict nondiscrimination rule (subject, however, to reasonable network management) to a rule against “unreasonable” discrimination can be seen in the interim documents. The Verizon-Google proposal, which called for a “Non-Discrimination Requirement,” nonetheless worded the requirement to ban only “undue discrimination against lawful Internet content, application, or service in a manner that causes meaningful harm to competition or to users.” (emphasis added)
Rep. Waxman’s draft bill, likewise, would have applied a somewhat different standard for wireline providers, who “shall not unjustly or unreasonably discriminate in transmitting lawful traffic over a consumer’s wireline broadband Internet access service,” also subject to reasonable network management.
Over time, the FCC recognized the error of its original draft and now agrees “with the diverse group of commenters who argue that any nondiscrimination rule should prohibit only unreasonable discrimination.” (¶ 77)
As between the suggested limiting terms “undue,” “unjust” and “unreasonable,” the FCC chose the latter for the final rule. Though many have complained that “unreasonable” is a nebulous, subjective term, it should be noted that of the three it is the only one with understood (if not entirely clear) legal meaning, particularly in the context of the FCC’s long history of rulemaking and adjudication.
The earliest railroad regulations, for example, which also provided the beginning of the FCC’s eventual creation and authority over communications industries, required reasonable rates of carriage, and empowered the Interstate Commerce Commission to intervene and eventually set the rates itself, much as the FCC later did with telephony.
One lesson of the railroad and telephone histories, however, is the danger of turning over to regulators decisions about what behaviors are reasonable. (Briefly, regulatory capture often ends up leaving the industry unable to respond to new forms of competition from disruptive technologies, with disastrous consequences.)
The V-G proposal gets to the heart of the problem in the text I italicized. Despite the negative connotations of the word in common use, “discrimination” isn’t inherently bad. As the Report makes clear, in managing Internet access and network traffic, there are many forms of discrimination—which means, after all, affording different treatment to different things—that are entirely beneficial to overall network behavior and to the consumer’s experience with the Internet.
The draft rule, as the FCC now admits (see ¶ 77 of the Report), was dangerously rigid. If any behavior should be regulated, it is the kind of discrimination whose principal purpose is to harm competition or users—though that kind of behavior is already illegal under various antitrust laws.
For one thing, users may want some kinds of traffic – e.g., voice and video – to receive a higher priority over text and graphics, which do not suffer from latency problems. Companies operating Virtual Private Networks for their employees may likewise want to limit Web access to selected sites and activities for workers while on the job.
A strict nondiscrimination rule would have also discouraged or perhaps banned tiered pricing, harming consumers who do not need the fastest speeds and the highest volume of downloads to accomplish what the want to online. (Without tiered pricing, such consumers effectively subsidize power users who, not surprisingly, are the most vociferous objectors to tiered pricing.)
Discrimination may also be necessary to manage congestion during peak usage periods or when failing nodes put pressure on the backbone. Discrimination against spam, viruses and other malware, much of which is not “lawful,” is also permitted and indeed encouraged. (See ¶ 90-92.)
By comparison, the Report notes three (¶ 75) types of provider discrimination that are of particular concern. These are: discrimination that harms competitors (e.g., VoIP providers of over-the-top telephone service, such as Skype or Vonage, that competes with the provider’s own telephone service), “inhibiting” end users from accessing content, services, and applications of their choice (but see the no-blocking rule, above, which already covers this), and discrimination that “impairs free expression,” including slowing or blocking access to a blog whose message the broadband provider does not approve.
On that last point, however, it’s important to note that Congress has already given broadband providers (and others) broad freedom to filter and otherwise curate content they do not approve of or which they believe their customers don’t want to see. Under Section 230 of the Communications Decency Act,
“No provider or user of an interactive computer service shall be held liable on account of . . . any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected.”
The goal of Section 230 was to immunize early Internet providers including CompuServe and Prodigy from efforts to exercise editorial control over message boards whose content was provided by customers themselves. But it gives providers broad discretion in determining what kind of content it believes its customers don’t want to see. So long as the filtering is undertaken in “good faith” (e.g., not with the intent of harming a competitor), there is no liability for the provider, who does not, for example, become a “publisher” for purposes of defamation law.
The FCC (¶ 89) acknowledges the limit that Section 230 puts on the discrimination rule.
On the harm to competitors prong, the FCC waffles (see ¶ 76) on whether “pay for priority”—the bugaboo that launched the neutrality offensive in the first place, actually constitutes a violation of the rules. While a broadband provider’s offering to prioritize the traffic of a particular source for a premium fee “would raise significant cause for concern,” the agency acknowledges that such behavior has occurred and thrived for years in the form of third party Content Delivery Networks. (See footnote 236) CDNs are allowed. (More on CDNs in the next post.)
So in the end the discrimination rule doesn’t appear to add much to the blocking rule or existing antitrust laws. Discrimination against competing over-the-top providers would violate antitrust. Blocking or slowing access to disfavored content is already subject to the blocking rule. And interfering with “free expression” rights of users is already significantly allowed by Section 230.
What’s left? “The rule rests on the general proposition,” the agency concludes (¶ 78), “that broadband providers should not pick winners and losers on the Internet,” even when doing so is independent of competitive interests. What exactly this means—and how “reasonable” discrimination will be judged in the course of enforcing the rules—remains to be seen.
Next: The exceptions and what they say about the real purpose of the rules