Archives For Internet service provider

In the face of an unprecedented surge of demand for bandwidth as Americans responded to COVID-19, the nation’s Internet infrastructure delivered for urban and rural users alike. In fact, since the crisis began in March, there has been no appreciable degradation in either the quality or availability of service. That success story is as much about the network’s robust technical capabilities as it is about the competitive environment that made the enormous private infrastructure investments to build the network possible.

Yet, in spite of that success, calls to blind ISP pricing models to the bandwidth demands of users by preventing firms from employing “usage-based billing” (UBB) have again resurfaced. Today those demands are arriving in two waves: first, in the context of a petition by Charter Communications to employ the practice as the conditions of its merger with Time Warner Cable become ripe for review; and second in the form of complaints about ISPs re-imposing UBB following an end to the voluntary temporary halting of the practice during the first months of the COVID-19 pandemic — a move that was an expansion by ISPs of the Keep Americans Connected Pledge championed by FCC Chairman Ajit Pai.

In particular, critics believe they have found clear evidence to support their repeated claims that UBB isn’t necessary for network management purposes as (they assert) ISPs have long claimed.  Devin Coldewey of TechCrunch, for example, recently asserted that:

caps are completely unnecessary, existing only as a way to squeeze more money from subscribers. Data caps just don’t matter any more…. Think about it: If the internet provider can even temporarily lift the data caps, then there is definitively enough capacity for the network to be used without those caps. If there’s enough capacity, then why did the caps exist in the first place? Answer: Because they make money.

The thing is, though, ISPs did not claim that UBB was about the day-to-day “manage[ment of] network loads.” Indeed, the network management strawman has taken on a life of its own. It turns out that if you follow the thread of articles in an attempt to substantiate the claim (for instance: here, to here, to here, to here), it is just a long line of critics citing to each other’s criticisms of this purported claim by ISPs. But never do they cite to the ISPs themselves making this assertion — only to instances where ISPs offer completely different explanations, coupled with the critics’ claims that such examples show only that ISPs are now changing their tune. In reality, the imposition of usage-based billing is, and has always been, a basic business decision — as it is for every other company that uses it (which is to say: virtually all companies).

What’s UBB really about?

For critics, however, UBB is never just a “basic business decision.” Rather, the only conceivable explanations for UBB are network management and extraction of money. There is no room in this conception of the practice for perfectly straightforward pricing decisions that offer pricing that differs by customers’ usage of the services. Nor does this viewpoint recognize the importance of these pricing practices for long-term network cultivation in the form of investment in increasing capacity to meet the increased demands generated by users.

But to disregard these actual reasons for the use of UBB is to ignore what is economically self-evident.

In simple terms, UBB allows networks to charge heavy users more, thereby enabling them to recover more costs from these users and to keep prices lower for everyone else. In effect, UBB ensures that the few heaviest users subsidize the vast majority of other users, rather than the other way around.

A flat-rate pricing mandate wouldn’t allow pricing structures based on cost recovery. In such a world an ISP couldn’t simply offer a lower price to lighter users for a basic tier and rely on higher revenues from the heaviest users to cover the costs of network investment. Instead, it would have to finance its ability to improve its network to meet the needs of the most demanding users out of higher prices charged to all users, including the least demanding users that make up the vast majority of users on networks today (for example, according to Comcast, 95 percent of its  subscribers use less than 1.2 TB of data monthly).

On this basis, UBB is a sensible (and equitable, as some ISPs note) way to share the cost of building, maintaining, and upgrading the nation’s networks that simultaneously allows ISPs to react to demand changes in the market while enabling consumers to purchase a tier of service commensurate with their level of use. Indeed, charging customers based on the quality and/or amount of a product they use is a benign, even progressive, practice that insulates the majority of consumers from the obligation to cross-subsidize the most demanding customers.

Objections to the use of UBB fall generally into two categories. One stems from the sort of baseline policy misapprehension that it is needed to manage the network, but that fallacy is dispelled above. The other is borne of a simple lack of familiarity with the practice.

Consider that, in the context of Internet services, broadband customers are accustomed to the notion that access to greater data speed is more costly than the alternative, but are underexposed to the related notion of charging based upon broadband data consumption. Below, we’ll discuss the prevalence of UBB across sectors, how it works in the context of broadband Internet service, and the ultimate benefit associated with allowing for a diversity of pricing models among ISPs.

Usage-based pricing in other sectors

To nobody’s surprise, usage-based pricing is common across all sectors of the economy. Anything you buy by the unit, or by weight, is subject to “usage-based pricing.” Thus, this is how we buy apples from the grocery store and gasoline for our cars.

Usage-based pricing need not always be so linear, either. In the tech sector, for instance, when you hop in a ride-sharing service like Uber or Lyft, you’re charged a base fare, plus a rate that varies according to the distance of your trip. By the same token, cloud storage services like Dropbox and Box operate under a “freemium” model in which a basic amount of storage and services is offered for free, while access to higher storage tiers and enhanced services costs increasingly more. In each case the customer is effectively responsible (at least in part) for supporting the service to the extent of her use of its infrastructure.

Even in sectors in which virtually all consumers are obligated to purchase products and where regulatory scrutiny is profound — as is the case with utilities and insurance — non-linear and usage-based pricing are still common. That’s because customers who use more electricity or who drive their vehicles more use a larger fraction of shared infrastructure, whether physical conduits or a risk-sharing platform. The regulators of these sectors recognize that tremendous public good is associated with the persistence of utility and insurance products, and that fairly apportioning the costs of their operations requires differentiating between customers on the basis of their use. In point of fact (as we’ve known at least since Ronald Coase pointed it out in 1946), the most efficient and most equitable pricing structure for such products is a two-part tariff incorporating both a fixed, base rate, as well as a variable charge based on usage.  

Pricing models that don’t account for the extent of customer use are vanishingly rare. “All-inclusive” experiences like Club Med or the Golden Corral all-you-can-eat buffet are the exception and not the rule when it comes to consumer goods. And it is well-understood that such examples adopt effectively regressive pricing — charging everyone a high enough price to ensure that they earn sufficient return from the vast majority of light eaters to offset the occasional losses from the gorgers. For most eaters, in other words, a buffet lunch tends to cost more and deliver less than a menu-based lunch. 

All of which is to say that the typical ISP pricing model — in which charges are based on a generous, and historically growing, basic tier coupled with an additional charge that increases with data use that exceeds the basic allotment — is utterly unremarkable. Rather, the mandatory imposition of uniform or flat-fee pricing would be an aberration.

Aligning network costs with usage

Throughout its history, Internet usage has increased constantly and often dramatically. This ever-growing need has necessitated investment in US broadband infrastructure running into the tens of billions annually. Faced with the need for this investment, UBB is a tool that helps to equitably align network costs with different customers’ usage levels in a way that promotes both access and resilience.

As President Obama’s first FCC Chairman, Julius Genachowski, put it:

Our work has also demonstrated the importance of business innovation to promote network investment and efficient use of networks, including measures to match price to cost such as usage-based pricing.

Importantly, it is the marginal impact of the highest-usage customers that drives a great deal of those network investment costs. In the case of one ISP, a mere 5 percent of residential users make up over 20 percent of its network usage. Necessarily then, in the absence of UBB and given the constant need for capacity expansion, uniform pricing would typically act to disadvantage low-volume customers and benefit high-volume customers.

Even Tom Wheeler — President Obama’s second FCC Chairman and the architect of utility-style regulation of ISPs — recognized this fact and chose to reject proposals to ban UBB in the 2015 Open Internet Order, explaining that:

[P]rohibiting tiered or usage-based pricing and requiring all subscribers to pay the same amount for broadband service, regardless of the performance or usage of the service, would force lighter end users of the network to subsidize heavier end users. It would also foreclose practices that may appropriately align incentives to encourage efficient use of networks. (emphasis added)

When it comes to expanding Internet connectivity, the policy ramifications of uniform pricing are regressive. As such, they run counter to the stated goals of policymakers across the political spectrum insofar as they deter low-volume users — presumably, precisely the marginal users who may be disinclined to subscribe in the first place —  from subscribing by saddling them with higher prices than they would face with capacity pricing. Closing the digital divide means supporting the development of a network that is at once sustainable and equitable on the basis of its scope and use. Mandated uniform pricing accomplishes neither.

Of similarly profound importance is the need to ensure that Internet infrastructure is ready for demand shocks, as we saw with the COVID-19 crisis. Linking pricing to usage gives ISPs the incentive and wherewithal to build and maintain high-capacity networks to cater to the ever-growing expectations of high-volume users, while also encouraging the adoption of network efficiencies geared towards conserving capacity (e.g., caching, downloading at off-peak hours rather than streaming during peak periods).

Contrary to the claims of some that the success of ISPs’ networks during the COVID-19 crisis shows that UBB is unnecessary and extractive, the recent increases in network usage (which may well persist beyond the eventual end of the crisis) demonstrate the benefits of nonlinear pricing models like UBB. Indeed, the consistent efforts to build out the network to serve high-usage customers, funded in part by UBB, redounds not only to the advantage of abnormal users in regular times, but also to the advantage of regular users in abnormal times.

The need for greater capacity along with capacity-conserving efficiencies has been underscored by the scale of the demand shock among high-load users resulting from COVID-19. According to OpenVault, a data use tracking service, the number of “power users” and “extreme power users” utilizing 1TB/month or more and 2TB/month or more jumped 138 percent and 215 percent respectively. Meaning that now, in total, power users represent 10 percent of subscribers across the network, while extreme power users comprise 1.2 percent of subscribers.

Pricing plans predicated on load volume necessarily evolve along with network capacity, but at this moment the application of UBB for monthly loads above 1TB ensures that ISPs maintain an incentive to cater to power users and extreme power users alike. In doing so, ISPs are also ensuring that all users are protected when the Internet’s next abnormal — but, sadly, predictable — event arrives.

At the same time, UBB also helps to facilitate the sort of customer-side network efficiencies that may emerge as especially important during times of abnormally elevated demand. Customers’ usage need not be indifferent to the value of the data they use, and usage-based pricing helps to ensure that data usage aligns not only with costs but also with the data’s value to consumers. In this way the behavior of both ISPs and customers will better reflect the objective realities of the nations’ networks and their limits.

The case for pricing freedom

Finally, it must be noted that ISPs are not all alike, and that the market sustains a range of pricing models across ISPs according to what suits their particular business models, network characteristics, load capacity, and user types (among other things). Consider that even ISPs that utilize UBB almost always offer unlimited data products, while some ISPs choose to adopt uniform pricing to differentiate their offerings. In fact, at least one ISP has moved to uniform billing in light of COVID-19 to provide their customers with “certainty” about their bills.

The mistake isn’t in any given ISP electing a uniform billing structure or a usage-based billing structure; rather it is in proscribing the use of a single pricing structure for all ISPs. Claims that such price controls are necessary because consumers are harmed by UBB ignore its prevalence across the economy, its salutary effect on network access and resilience, and the manner in which it promotes affordability and a sensible allocation of cost recovery across consumers.

Moreover, network costs and traffic demand patterns are dynamic, and the availability of UBB — among other pricing schemes — also allows ISPs to tailor their offerings to those changing conditions in a manner that differentiates them from their competitors. In doing so, those offerings are optimized to be attractive in the moment, while still facilitating network maintenance and expansion in the future.

Where economically viable, more choice is always preferable. The notion that consumers will somehow be harmed if they get to choose Internet services based not only on speed, but also load, is a specious product of the confused and the unfamiliar. The sooner the stigma around UBB is overcome, the better-off the majority of US broadband customers will be.

With Matt Starr, Berin Szoka and Geoffrey Manne

Today’s oral argument in the D.C Circuit over the FCC’s Net Neutrality rules suggests that the case — Verizon v. FCC — is likely to turn on whether the Order impermissibly imposes common carrier regulation on broadband ISPs. If so, the FCC will lose, no matter what the court thinks of the Commission’s sharply contested claims of authority under the Telecommunications Act.

The FCC won last year before the same court when Verizon challenged its order mandating that carriers provide data roaming services to their competitors’ customers. But Judge Tatel, who wrote the Cellco decision is likely to write the court’s opinion overturning the Net neutrality rules — just as he wrote the court’s 2010 Comcast v. FCC opinion, thwarting the FCC’s first attempt at informal net neutrality regulation.

Over an extraordinary two-hour session, Judges Tatel and Silberman asked a barrage of questions that suggest they’ll apply the same test used to uphold the data roaming rule to strike down at least the non-discrimination rule at the heart of the Open Internet Order — and probably, the entire Order.

Common Carrier Analysis

The Communications Act explicitly prohibits treating services that are not regulated under Title II as common carriers. Title II regulates “telecommunications services,” such as landline telephone service, but broadband is an “information service” regulated under Title I of the Act, while wireless is regulated under Title III of the Act (as a “radio transmission”).

In Cellco, the court ruled that the FCC’s data roaming rule did not impermissibly classify mobile providers as common carriers even though it compelled wireless carriers to let other companies’ subscribers roam on their networks. Here, the Open Internet Order effectively forces ISPs to carry traffic of all “edge” providers in an equal, non-discriminatory manner. While these might seem similar, the two mandates differ significantly, and Tatel’s analysis in the data roaming case may lead to precisely the opposite result here.

Tatel’s data roaming opinion rested on a test, derived from decades of case law, for determining what level of regulation constitutes an impermissible imposition of common carrier status:

  1. “If a carrier is forced to offer service indiscriminately and on general terms, then that carrier is being relegated to common carrier status”;

  2. “[T]he Commission has significant latitude to determine the bounds of common carriage in particular cases”;

  3. “[C]ommon carriage is not all or nothing—there is a gray area [between common carrier status and private carrier status] in which although a given regulation might be applied to common carriers, the obligations imposed are not common carriage per se” because they permit carriers to retain sufficient decisionmaking authority over their networks (by retaining programming control and/or the authority to negotiate terms, for example); and

  4. In this gray area, “[the FCC’s] determination that a regulation does or does not confer common carrier status warrants deference” under the Supreme Court’s Chevron decision.

In Cellco, the court determined that the data roaming rule fell into the gray area, and thus deferred to the FCC’s determination that the regulation did not impose common carrier status. The essential distinction, according to the court, was that carriers remained free to “negotiate the terms of their roaming arrangements on an individualized basis,” provided their terms were “commercially reasonable.” Rather than impose a “presumption of reasonableness,” the Commission offered “considerable flexibility for providers to respond to the competitive forces at play in the mobile-data market.” Thus, the court held, the data roaming rule “leaves substantial room for individualized bargaining and discrimination in terms,” and thus “does not amount to a duty to hold out facilities indifferently for public use.”

The Open Internet rules, by contrast, impose a much harsher restriction on what ISPs may do with their broadband networks, barring them from blocking any legal content and prohibiting “unreasonable” discrimination. Judges Tatel and Silberman repeatedly asked questions that suggested that the Order’s reasonable discrimination rule removed the kind of “flexibility” that justified upholding the data roaming rule. By requiring carriers to “offer service indiscriminately and on general terms” and to “hold out facilities indifferently for public use” (to quote the D.C. Circuit’s test), the rule would go beyond the “gray area” in which the FCC gets deference, and fall into the D.C. Circuit’s definition of common carriage. If that’s indeed ultimately where the two judges wind up, it’s game over for the FCC.

The Open Internet Order requires broadband ISPs to make their networks available, and to do so on equal terms that remove pricing flexibility, to any edge provider that wishes to have its content available on an ISP’s network. This seems to be Judge Tatel’s interpretation of ¶ 76 of the Order, which goes on at length about the reasons why “pay for priority” arrangements would “raise significant cause for concern” and then concludes: “In light of each of these concerns, as a general matter, it’s unlikely that pay for priority would satisfy the ‘no unreasonable discrimination’ standard.” So… legal in principle, but effectively banned in practice — a per se rule dressed up as a rule of reason.

If that isn’t, in effect, a requirement that ISPs hold out their networks “indifferently for public use,” it’s hard to imagine what is — as Tatel certainly seemed to think today. Tatel’s use of the term “indiscriminately” in Cellco almost hints that the test was written with the FCC’s “no discrimination” rule in mind.

The FCC tried, but failed, to address such concerns in the Open Internet Order, by arguing that broadband providers remained free to “make individualized decisions” with the only customers that matter: their subscribers. Today, the agency again insisted that restricting, however heavily, a broadband provider’s ability to negotiate with an edge provider (or the backbone providers in between) is irrelevant to the analysis of whether the FCC has illegally imposed common carriage. But if that argument worked, the D.C. Circuit would not have had to analyze whether the data roaming rule afforded sufficient flexibility to carriers in contracting with other carriers to provide data roaming services to their customers.

Similarly, the FCC failed today, and in its briefs, to effectively distinguish this case from Midwest Video II, which was critical to the Cellco decision. here, the Supreme Court court struck down public-access rules imposed on cable companies as impermissible common carrier regulation because they “prohibited [cable operators] from determining or influencing the content of access programming,” and “delimit[ed] what [they could] charge for access and use of equipment.” In other words, the FCC’s rule left no flexibility for negotiations between companies — the same problem as in the Open Internet Order. The FCC attempted to distinguish the two cases by arguing that the FCC was restricting an existing wholesale market for channel carriage, while no such market exists today for prioritized Internet services. But this misses the key point made, emphatically, by Judge Silberman: it is the FCC’s relentless attempt to regulate Net Neutrality that has prevented the development of this market. Nothing better reveals the stasis mentality behind the FCC’s Order

Perhaps the most damning moment of today’s arguments occurred when Verizon’s lawyer responded to questions about what room for negotiation was left under the unreasonable discrimination rule — by pointing to what the FCC itself said in Footnote 240 of the Order. There the FCC quotes, approvingly, comments filed by Sprint: “The unreasonable discrimination standard contained in Section 202(a) of the Act contains the very flexibility the Commission needs to distinguish desirable from improper discrimination.” In other words, the only room for “commercially reasonable negotiation” recognized by the FCC under the nondiscrimination rule is found in the limited discretion traditionally available to common carriers under Section 202(a). Oops. This #LawyerFail will doubtless feature prominently in the court’s discussion of this issue, as the FCC’s perhaps accidental concession that, whatever the agency claims, it’s really imposing common carrier status — analogous to Title II, no less!

Judges Tatel and Silberman seemed to disagree only as to whether the no-blocking rule would also fail under Cellco’s reasoning. Tatel suggested that if the non-discrimination rule didn’t exist, the blocking rule, standing alone, would “leave substantial room for individualized bargaining and discrimination in terms” just as the data roaming rule did. Tatel spent perhaps fifteen minutes trying to draw clear answers from all counsel on this point, but seemed convinced that, at most, the no-blocking rule simply imposed a duty on the broadband provider to allow an edge provider to reach its customers, while still allowing the broadband provider to negotiate for faster carriage on “commercially reasonable terms.” Silberman disagreed, insisting that the blocking rule still imposed a common carrier duty to carry traffic at a zero price.


Ultimately the distinction between these two rules under Cellco’s common carriage test may not matter. If the court decides that the order is not severable, striking down the nondiscrimination rule as common carriage would cause the entire Order to fall.

The judges got into an interesting, though relatively short, discussion of this point. Verizon’s counsel repeatedly noted that the FCC had never stated any intention that the order should be read as severable either in the Order, in its briefs or even at oral argument. Unlike in MD/DC/DE Broadaster’s Assoc. v. FCC, the Commission did not state in the adopting regulation that it intended to treat the regulation as severable. And, as the DC Circuit has stated, “[s]everance and affirmance of a portion of an administrative regulation is improper if there is ‘substantial doubt’ that the agency would have adopted the severed portion on its own.”

The question, as the Supreme Court held in K Mart Corp. v. Cartier, Inc., is whether the remainder of the regulation could function sensibly without the stricken provision. This isn’t clear. While Judge Tatel seems to suggest that the rule against blocking could function without the nondiscrimination rule, Judge Silberman seems convinced that the two were intended as necessary complements by the FCC. The determination of the no-blocking rule’s severability may come down to Judge Rogers, who didn’t telegraph her view.

So what’s next?

The prediction made by Fred Campbell shortly after the Cellco decision seems like the most likely outcome: Tatel, joined by at least Silberman, could strike down the entire Order as imposing common carriage — while offering the FCC a roadmap to try its hand at Net Neutrality yet again by rewriting the discrimination rule to allow for prioritized or accelerated carriage on commercially reasonable terms.

Or, if the the court decides the order is severable, it could strike down just the nondiscrimination rule — assuming the court could find either direct or ancillary jurisdiction for both the transparency rule and the non-discrimination rule.

Either way, an FCC loss will mean that negotiated arrangements for priority carriage will be governed under something more like a rule of reason. The FCC could try to create its own rule.  Or the matter could simply be left to the antitrust and consumer protection laws enforced by the Department of Justice, the Federal Trade Commission, the states and private plaintiffs. We think the latter’s definitely the best approach. But whether it is or not, it will be the controlling legal authority on the ground the day the FCC loses — unless and until the FCC issues revised rules (or Congress passes a law) that can survive judicial review.

Ultimately, we suspect the FCC will have a hard time letting go. After 79 years, it’s clearly in denial about its growing obsolescence.

By Geoffrey Manne & Berin Szoka

As Democrats insist that income taxes on the 1% must go up in the name of fairness, one Democratic Senator wants to make sure that the 1% of heaviest Internet users pay the same price as the rest of us. It’s ironic how confused social justice gets when the Internet’s involved.

Senator Ron Wyden is beloved by defenders of Internet freedom, most notably for blocking the Protect IP bill—sister to the more infamous SOPA—in the Senate. He’s widely celebrated as one of the most tech-savvy members of Congress. But his latest bill, the “Data Cap Integrity Act,” is a bizarre, reverse-Robin Hood form of price control for broadband. It should offend those who defend Internet freedom just as much as SOPA did.

Wyden worries that “data caps” will discourage Internet use and allow “Internet providers to extract monopoly rents,” quoting a New York Times editorial from July that stirred up a tempest in a teapot. But his fears are straw men, based on four false premises.

First, US ISPs aren’t “capping” anyone’s broadband; they’re experimenting with usage-based pricing—service tiers. If you want more than the basic tier, your usage isn’t capped: you can always pay more for more bandwidth. But few users will actually exceed that basic tier. For example, Comcast’s basic tier, 300 GB/month, is so generous that 98.5% of users will not exceed it. That’s enough for 130 hours of HD video each month (two full-length movies a day) or between 300 and 1000 hours of standard (compressed) video streaming. Continue Reading…