This Thursday, the Federal Communications Commission (FCC) is poised to reclassify ISPs like Verizon and Comcast that connect us to the Internet as common carriers and adopt strong net neutrality rules to protect users, innovation, and free speech online. This is great news.
The rules are the result of a year-long process where more than four million Americans, including dozens of lawmakers, called upon the agency to create clear bright-line rules that ban blocking, throttling, and paid prioritization. The agency listened, and now the finish line is in sight.
But we’re not quite there. Over the next four days, the agency should strengthen four key provisions to ensure that its rules don’t invite any unnecessary confusion or create gaps in protection.
1. Revise the General Conduct Rule.
The FCC should establish a strong general conduct rule so that we don’t have to repeat this process year after year.
According to the FCC Fact Sheet describing the proposed rules, the FCC intends to adopt a “a general Open Internet conduct standard that ISPs cannot harm consumers or edge providers.” This rule would allow the agency to evaluate practices on a case by case basis that are not already banned by the bright-line rules. That is good, and allows the FCC to address problematic ISP practices as they emerge.
While a rule that investigates whether a practice reduces competition, innovation, or free speech asks the right questions, it is not specific enough to be administrable without further clarification. If we want to effectively protect against problematic ISP behavior in the future, we need a clear, simple rule that is easily applicable and makes it feasible for users, start-ups, and small businesses to bring complaints – a rule that provides certainty to the market and keeps the costs of regulation low.
The stakes are high. A vague, unbounded rule will create an ineffective regime that leads to unnecessary confusion, requires costly litigation, provides opportunities for FCC overreach, and tilts the playing field in favor of large, established companies that can afford long and costly proceedings at the FCC.
The agency’s current experience illustrates the costs of reinventing the wheel. In January 2014, the Court of Appeals for the D.C. Circuit struck down the FCC's Open Internet rules. More than one year, many roundtables, and over 4,000,000 comments later, the FCC understands how important it is to ban practices that increase the costs of innovation and free speech online – just like it did in 2010, when it first adopted net neutrality rules. The general conduct rule should embody these insights; it should not force us to learn the same lessons over and over again.
Thus, the rule should evaluate ISP practices only based on whether they:
preserve user choice;
do not discriminate against applications or classes of applications (i.e. whether they are “application-agnostic”); and
keep the costs of application innovation and speech low.
After two long proceedings, we know that these are the factors that have fostered application innovation and free speech in the past; they are the factors that network neutrality rules should protect. Having seven or more factors, many of which sound quite vague, as the FCC seems to be proposing, injects uncertainty into the market. We won't know how practices will be judged or how the FCC will weigh the different factors against each other. With three clear factors, market participants will have more guidance.
The FCC used this approach for its 2010 nondiscrimination rule. It should adopt the same approach now. Market participants are already familiar with these factors, and the rule has worked well in the past.
Analysis of Proposed Network Neutrality Rule, pp. 10-19.
Ex Parte Letter of Vimeo, Cogent, Contextly, Distinc.tt, Dwolla, Engine Advocacy, Kickstarter, OpenCurriculum, and Tumblr, pp. 2-3.
Ex Parte Letter of Electronic Frontier Foundation, pp. 1-2.
2. Ban zero-rating against a fee.
The FCC should protect against the next big threat to innovation and free speech online.
The rules should explicitly prohibit ISPs from charging application providers for zero-rating (i.e. not counting applications against users’ monthly bandwidth caps). Since 2010, zero-rating has become common around the world. ISPs in the US have all declared interest in the practice.
Allowing ISPs to charge application providers for zero-rating creates exactly the same problems for start-up innovation, small businesses, and free speech as charging them to be in the fast lane. ISPs will have an incentive to adopt low bandwidth caps to make paying for zero-rating more attractive, which in turn harms users and providers who can’t pay for zero-rating. Thus, banning fast lanes, but allowing zero-rating creates an important gap in protection.
What’s more, failing to prohibit zero-rating against a fee would be a step back from the 2010 rules. The 2010 order banned the practice – according to Verizon itself, which recently admitted that it sued the FCC over the 2010 order for this very reason. Seven other countries – Chile, the Netherlands, Slovenia, Canada, Norway, Germany, and Austria – have already found zero-rating to be a violation of network neutrality. The FCC should do the same.
van Schewick, Network Neutrality and Zero-rating, pp. 1-5.
Ex Parte Letter of Union Square Ventures.
Ex Parte Letter of Civil Rights Groups and Reform Organizations.
Letter of 36 scholars to FCC and FTC.
3. Adopt a strong legal foundation.
The FCC should adopt a jurisdictional foundation that will be upheld in court.
The FCC intends to classify the ISPs that connect us to the Internet as common carriers under Title II of the Communications Act. This is good, and it’s the only way to adopt strong network neutrality rules that will be upheld in court.
However, according to the FCC Fact Sheet, the FCC intends to go beyond classifying the service that we, the users, buy when we buy access to the Internet (i.e. the “consumer-facing side of that service”). It says that by transporting data across their network to and from users, ISPs are also offering a telecommunications service to anybody to whom their subscribers connect online (a so-called “sender-side” service).
Classifying that service is unnecessary and harmful. It raises a host of legal problems, and is unlikely to be upheld in court. When the FCC explored this approach last year, almost all participants in the debate – from the ISPs over public interested groups to the tech companies – agreed this approach would be likely to be struck down in court, and said so in filings to the FCC. Such widespread agreement is rare, and should be a red flag.
While there is a reading of the case that suggests that the D.C. Circuit invited the FCC to classify this “sender-side” service, this imagined service does not exist. The service does not meet key parts of the definition of a telecommunications service. Application providers like Tumblr don’t have a relationship with their users’ ISPs. They don’t pay those ISPs “a fee” to have Tumblr’s data transported across the ISPs’ access networks to their users. In fact, the rules’ ban on access fees (whether for access to users or for preferential treatment) prohibits ISPs from charging such a fee. Nor is the transmission of Tumblr’s data on an end user’s access network offered to Tumblr “directly.” When I call a colleague in L.A., my colleague’s local phone provider helps terminate my call, but that provider is not “offering” me a telecommunications service “directly” when doing so. Internet service is no different.
While Verizon v. FCC discussed the existence of such a service, it did so in a different context. The court did not engage in a classification analysis. But even if it had done so, the Supreme Court made clear in Brand X that a court’s prior interpretation of an ambiguous statute does not bind the agency. It’s the agency’s job to engage in classification decisions involving ambiguous terms, and it receives deference for those decisions, even if a court previously has taken a position on this issue. Here, the court did not even engage in a classification analysis, but even if it had done so, the Brand X analysis would apply here.
What’s more, assuming that ISPs offer a service to any provider that interacts with their users runs counter to the way the Internet operates and might set a dangerous precedent by making every Internet user and application provider a customer of ISPs around the world.
On the Internet, we can reach everyone without needing to interact with their ISP. That’s the beauty of the Internet and it should stay that way.
Ex Parte Letter of Google.
Ex Parte Letter of Free Press.
van Schewick & Schierenbeck, Comments on Mozilla's Proposal.
4. Fix the Interconnection problem.
The FCC should clarify that last-mile ISPs can’t use practices related to interconnection to evade the FCC’s network neutrality rules.
Most of us experienced net neutrality as a problem when our video streaming buffered in a spinning wheel of death. Netflix’s performance improved after it started to pay large ISPs for interconnection. But connections by Level 3, Cogent, and other interconnecting providers that refuse to pay are still congested, harming every user and provider whose traffic enters these networks via these providers. This order should address that problem – not kick it down the road.
ISPs can block, discriminate, or charge access fees when data enters their networks (i.e. at the point of interconnection) or when data is traveling over those networks. The impact on users and websites is the same either way, as is the harm to innovation and free speech. Thus, banning blocking, throttling, and paid prioritization on the access network, but not at the point of interconnection, allows ISPs to easily evade the rules by engaging in these practices at the point of interconnection instead. To fix this problem, the FCC should clarify that last-mile ISPs can’t use practices related to interconnection to evade the FCC’s network neutrality rules – that access fees are illegal, no matter where they are charged.
van Schewick, Analysis of Proposed Network Neutrality Rules, pp. 19-21.
Ex Parte Letter of Open Technology Institute.