Archive for July, 2014

The Net Neutrality Debate Reduced To A Cocktail Napkin

In 1974, economist Arthur Laffer famously sketched a curve on a cloth napkin at a restaurant to inform tax policy. More recently, economist Alex Tabarrok employed a similar tactic to inform patent policy. Some economic problems simply defy prose.

The “Laffer Curve,” which depicted a simplified relationship between tax rates and tax revenues, was designed to assist Ford Administration officials in determining the revenue-maximizing tax rate. Given the complexity in implementing a net-neutrality regime, FCC Chairman Tom Wheeler similarly could benefit from a simplified chart to determine the investment-maximizing Internet regulation. Having typed for 20-odd years and thereby having lost all fine-motor skills, I apologize in advance for my artwork.

Net neutrality napkin

Depending on which of the two graphs best depicts the likely investment response by the Internet ecosystem to changes in regulation, the Chairman could support a heavy- or light-touch approach to implementing a net-neutrality regime.

The FCC recognizes the importance of investment in shaping Internet policy, as expressed by its recent Open Internet Notice of Proposed Rulemaking:

We seek comment on the current role of the Internet’s openness in facilitating innovation, economic growth, free expression, civic engagement, competition, and broadband investment and deployment. Particularly, we seek comment on the role the open Internet rules have had in investment in the broadband marketplace—networks and edge providers alike. We are similarly interested in understanding the role that the open Internet may play in the promotion of competition or in identifying barriers to infrastructure investment that an open Internet may eliminate or lessen. 

As is customary in Washington, extreme voices have dominated the net-neutrality debate, emphasizing what investment they consider to be paramount.

On one extreme, the netizens advocate heavy-handed regulation to protect investment at the “edge” of the network by content and application providers. Writing in the Washington Post, Nancy Scola articulated this edge-centric ethic, suggesting that without strict net neutrality rules in place, “young entrepreneurs eventually give up. And knowing of their struggles, the next generation of would-be online entrepreneurs doesn’t even bother trying.”

On the other extreme, the libertarians advocate that the FCC play no role in regulating the Internet. They fret that even the slightest regulation will discourage investment at the “core” of the network by Internet service providers (ISPs). As a reflection of this core-centric ethic, the House Judiciary Committee held a hearing last week on why antitrust laws, enforced by the Federal Trade Commission, were sufficient to protect edge providers.

The libertarian’s view was rejected by the D.C. Circuit in Verizon v. FCC. The Court determined that some regulation is justified to promote the “virtuous cycle” of investment by edge providers, which in turn spurs investment at the core:

The Commission’s finding that Internet openness fosters the edge-provider innovation that drives this ‘virtuous cycle’ was likewise reasonable and grounded in substantial evidence . . . This conclusion finds ample support in the economic literature on which the Commission relied . . . The record amassed by the Commission contains many similar examples, and Verizon has given us no basis for questioning the Commission’s determination that the preservation of Internet openness is integral to achieving the statutory objectives set forth in Section 706 [of the Telecommunications Act of 1996].

The Court went on to explain that ISPs “could act in ways that would ultimately inhibit the speed and extent of future broadband deployment.” The Court also determined that ISPs, if left to their own devices in establishing pay-for-priority arrangements, would not internalize any “resultant harms to innovation” by edge providers. Finally, the Court found that ISPs “have the technical and economic ability to impose such [investment-deterring] restrictions.”

But how much regulation is needed to trigger this “virtuous cycle”? The answer is that the optimal amount of regulation should maximize the sum total of investment at the edge and the core. In the spirit of neutrality, the FCC should consider a dollar of investment at the edge to generate the same social benefit as a dollar of investment at the core. Although the “virtuous cycle” theory embraced by the Court suggests that edge investment could be given more weight in the FCC’s calculus, there are positive spillovers associated with core investment as well, and there is no objective way to determine the relevant weights.

Because most adults, including economists, cannot visualize what the sum of two curves looks like, a napkin is unavoidable. Scroll back up to the top of this blog. In both graphs, the vertical axis is the amount of investment, and the horizontal axis is the amount of regulation. The red curve depicts edge investment as a function of regulation (more is better), and the blue curve depicts core investment as a function of regulation (less is better). The black curve depicts the sum total of core and edge investment for any level of regulation.

The graph in the left pane depicts a world in which both edge and core investment is sensitive to the amount of regulation, but edge investment is more so. Thus, while core investment slowly declines with more FCC oversight, edge investment drops dramatically as regulation is withdrawn. In this world, the optimal policy is a heavy-handed approach, such as common carriage under Title II, which maximizes total investment (and happens to maximize edge investment).

The graph in the right pane depicts a world in which neither type of investment is sensitive to the level of regulation except at the extremes. Core investment is generally robust unless regulators overreach, and edge investment can be counted on except when regulators fall asleep. In this case, the optimal policy is a light-touch approach, such as Section 706, which maximizes total investment at the expense of a little edge and a little core investment (relative to their maximum levels).

How should the Chairman decide which of these two worlds best captures the facts on the ground? There is scant empirical evidence of the “elasticity” of edge investment with respect to a change in regulation, as the Internet is a nascent industry and the regulatory protections for edge providers have hardly changed. In comparison, the elasticity of core investment to changes in other types of regulation, such as mandatory unbundling, has been empirically studied. But again, the type of rules being contemplated here—such as a no-blocking rule or the equal treatment of packets—are a bit novel. Indeed, the D.C. Circuit was persuaded by “the absence of evidence that similar restrictions of broadband providers had discouraged infrastructure investment.”

It defies logic to believe that an app developer operating out of her parent’s garage is going to throw in the towel because NetflixGoogle, or some other edge behemoth entered into a voluntary, for-pay arrangement with an ISP; there are other, more significant obstacles standing in her way. Further, the service needs of a startup are very different from the needs of an established content provider with tens of millions of customers. This suggests that edge investment will not drop precipitously in response to a modest reduction in regulation. If I am right, then the right pane of the graph should guide policy.

Notwithstanding this intuition, in the absence of empirical evidence about likely investment responses, the FCC Chairman should preserve his flexibility. A light-touch approach, in which the agency draws on its authority under Section 706, would do just that it. If I am wrong about what motivates an edge entrepreneur, and if edge investment instead were to fall precipitously under a light-touch approach—that is, the left pane captures the real world—then the agency could always apply the “commercially reasonable” standard under Section 706 in a more restrictive manner (so long as it permitted some room for bargaining between edge providers and ISPs).

In contrast, starting with a heavy-handed Title II approach could risk substantial core investment without generating any offsetting incremental investment at the edge. Under the “mother-may-I” approach of Title II, it is much harder to ratchet down regulation. For example, Title II could entail a rate-setting process for edge- or customer-facing offerings by ISPs; once a regulated rate is established, a new proceeding would be needed to adjust it. Because Title II would limit the FCC’s flexibility in the face of such uncertainty over investment responses, the better approach is the light-touch of Section 706.

Under a light-touch approach, both core and edge providers would be asked to make a modest sacrifice in exchange for the greater good. Barring the unlikely event in which the left panel depicts reality, total investment would be maximized, and the regulator’s flexibility to adjust the rules in response to investment reactions would be preserved.

Now can I have my napkin back?


Twitter: @halsinger

Leave a comment

In Its Dealings With ISPs, Netflix Is Holding A Powerful Card

By producing compelling online content and interfacing directly with its customers, Netflix is holding a powerful card—and I’m not talking about its Emmy-award-winning show. Rather than playing this card, Netflix is asking the Federal Communications Commission (FCC) to intervene in its dealings with Internet service providers (ISPs). Before delving into Netflix’s potential counter-strategy and the need (if any) for regulatory intervention, a bit of background is in order.

FCC Chairman Tom Wheeler announced last week that the agency is launching a new investigation of “interconnection” agreements, which as the name suggests, govern connections between Internet networks. Interconnection has taken center stage since Netflix struck deals with Comcast and Verizon in February; prior to those direct connections with ISPs, Netflix paid “transit” providers such as Cogent and Level 3 to obtain access to the ISPs’ networks. Presently, interconnection arrangements are governed by private contracts.


Interconnection issues are not implicated in the FCC’s pending Open Internet proceeding, which addresses the treatment of traffic within an ISP’s network, as opposed to on its doorstep. Yet some companies are trying to conflate the issues and leverage the religious fervor and grassroots political machines of the net neutrality movement. Adding to the drama is Netflix’s suggestion that it was forced to accept terms for direct connections at gunpoint; according to some accounts, the counter-parties were purposefully degrading the quality of the connection with Netflix until Netflix coughed up some cash.

Is there a constructive role for the regulator? The Progressive Policy Institute (a D.C. think tank with which I am affiliated) held a conference on interconnection last month, and invited Wharton Professor Kevin Werbach to make the case for FCC regulation of interconnection. At the conference and in his writings, Werbach cited some classic interconnection showdowns, including Comcast-Level 3 and Verizon-Cogent, as the basis for intervention: A benevolent referee could resolve these disputes quickly, Werbach argues, and get traffic flowing to Internet customers.

To evaluate the potential benefits of intervention, I looked into these disputes and was surprised by what I found: Based on historical frequencies, the likelihood of a dispute between networks is rare, and the likelihood of a dispute leading to a service outage for consumers is even rarer. By my count, there have been just six major interconnection disputes since 2002 (five of which involved Cogent)—about one every other year—and the average number of days without service across these disputes was close to zero. In other words, even when these disputes occur, traffic generally continues to flow pursuant to a standstill agreement while the dispute is worked out. Unless something has radically tipped the balance of power in the Internet ecosystem, history suggests that the benefits of the FCC’s intervening in these affairs—in terms of forgone service outages—are likely small.

On the other side of the ledger, inserting the FCC into these negotiations could impose significant costs on society. For example, mandatory interconnection at regulated rates could undermine the incentive of ISPs to expand or enhance broadband networks. Some have blamed mandatory roaming for certain wireless operators’ decision not to build out in high-cost areas (but rather rely on roaming). Moreover, the FCC’s assistance could discourage access-seeking networks, including transit providers and some large content providers such as Google, from expanding their networks into last-mile services. According to this cost-benefit analysis, the FCC should stay out of these affairs.

Two other considerations should give the Chairman pause about intervening on interconnection. First, for customers who are hooked on Netflix exclusive content, such as House of Cards or Orange Is the New BlackNetflix is the “must-have” network. I could access the Internet at super-fast speeds through my cable operator or my telephone provider (and soon through my mobile device), but there is no good substitute for what Netflix is producing. So if push came to shove, and my ISP started fooling with my Netflix connection, I would consider switching ISPs to see whether Frank Underwood maintains his presidency or Piper Chapman gets out of jail. Although this choice in super-fast connections is not available to all customers—by the FCC’s latest count, nearly three-quarters of U.S. households are served by two or more wireline ISPs with download speeds of at least 6 Mbps—the choice is available to enough households to make the ISPs think twice about degrading Netflix.

Second, Netflix has a potent counter-strategy that, if deployed, could be significantly more powerful in its dealing with ISPs than regulation: By charging its subscribers different prices based on their ISP, Netflix can gently steer its customers to “low-priced” ISPs—that is, ISPs that charge low or no interconnection fees. For example, Google Fiber, an ISP with a limited national footprint, recently announced that it would abstain from charging Netflix (or any content provider) an interconnection fee.

Like a credit card, Google Fiber is best understood as a “platform provider” that connects end users with content providers. When certain credit cards sought to impose relatively higher fees on merchants, merchants countered by imposing surcharges (or discounts) on the merchandise to steer customers to the lower-priced cards. Some large banks responded by imposing a “no-surcharge rule” on merchants, forcing merchants to charge the same price for goods regardless of which card was used. Following the abolition of no-surcharge rules in Australia (a similar movement is afoot in the United States), the number of merchants surcharging payment card transactions has increased steadily over time, leading to a significant reduction in merchant transaction fees.

Applying that lesson here, Netflix could charge Google’s customers a discount (say $6.99 per month as opposed to its standard $7.99 charge) for Netflix service. Alternatively, Netflix could charge customers of a high-priced ISP a surcharge (say $9.99 per month). By revealing to its subscribers the identity of the low-priced ISP, this counter-strategy could temper the interconnection charge of the high-priced ISP. Unlike cable networks, which rely on the cable operator to interface with the video customer, Netflix and other online providers are customer-facing and thus wield significantly greater bargaining power in their dealings with the platform provider—as long as they are willing to use it.

I asked a Netflix spokesperson at a recent Aspen Institute event whether Netflix has contemplated this counter-strategy. His answer, which begins at about 1:50:32 on the video, was (1) he has at least considered it, but (2) the interconnection fee charged by ISPs to date was “so small” in relation to Netflix’s content costs that a surcharge would not make sense. Admittedly, my question was tough, but this answer does not engender much sympathy for Netflix’s plight.

Before seeking further regulatory intervention, Netflix should avail itself of all potential counter-strategies in its dealings with ISPs. To do anything less is to ask the FCC to carry your water. As Frank Underwood put it, “There is but one rule: Hunt or be hunted.” Netflix is holding a powerful trump card that potentially obviates the need for regulation, but it seems disinclined to use it. Until Netflix has gone on the hunt and failed, the Chairman should shelve interconnection rules and focus his attention on the Open Internet rules now pending before him.


Leave a comment

What’s Missing from the Net Neutrality Debate? A Principle for Designing Good Remedies

The reaction from netizens was swift and fierce: Chairman Wheeler’s proposal to permit paid prioritization on the Internet—with an offer to stamp out discriminatory conduct on a case-by-case basis—was considered a betrayal of President Obama’s net-neutrality pledge. Protesters gathered in front of the Federal Communications Commission (FCC), and petitions made the rounds on Twitter.

The grassroots campaign for reclassifying Internet services from “information” to “telecommunications” served its purpose: The Chairman has essentially changed his position and has put the agency on the path to embracing a more invasive “Title II” approach. In economic terms, this means that the regulator could establish a “zero price” for paid prioritization. And when its price goes to zero, priority delivery will cease to exist.

This is not the agency’s first attempt to regulate paid prioritization out of existence. In a recent decision to toss the FCC’s original Open Internet rules, the D.C. Circuit said that interfering with market negotiations by setting a zero price amounted to “common carriage” regulation, which was legal only if the FCC reclassified Internet service (and then established a regulated rate of zero).

The Court also articulated a less-invasive path for regulating such arrangements, in which Internet service providers (ISPs) and content providers could voluntarily negotiate the terms for priority delivery. The FCC could serve as a backstop to adjudicate disputes if negotiations broke down and discrimination was to blame. Importantly, the Court signaled that the FCC could invoke this alternative approach under its existing (section 706) authority without reclassifying ISPs.

How should the FCC choose between these two alternatives? Economic efficiency dictates that the regulator should select the least restrictive remedy that addresses the potential harm—that is, select the most targeted protection against the potential that ISPs could favor certain content providers over others in paid-prioritization offerings.

This least-restrictive principle is sorely missing from the current version of the net neutrality debate. Perhaps it is overlooked because it is so simple, or perhaps it’s being studiously ignored because Title II proponents (led by a handful of law school professors) don’t think like economists. But even lawyers understand that if two remedies equally mitigate a potential harm, and if the first uniquely introduces a new risk, the second is clearly preferred.

To drive home this least-restrictive principle, consider a simple analogy in your kitchen. There is small chance of starting a fire when preparing dinner for the kids. To protect against that harm, you can adopt one of two remedies: Purchase a fire extinguisher from your neighborhood hardware store, or install a fire hydrant equipped with a fire hose from a contractor.

There is no doubt that both remedies will put out a fire. Aside from offending aesthetic sensibilities, the fire-hydrant solution opens up the possibility that your toddler will flood the kitchen. Because the extinguisher gets the job done, and because there is zero risk of ancillary harm, you head to Home Depot.

To further drive home this principle, consider how the Department of Justice selects among possible remedies when reviewing a merger. According to its Policy Guidelines to Merger Remedies, the Antitrust Division looks for a surgical strike:

A remedy carefully tailored to the competitive harm is the best way to ensure effective relief. Before the Division will conclude that a proposed remedy is acceptable, the relief must effectively address each of the Division’s competitive concerns. There should be a close, logical nexus between the proposed remedy and the alleged violation—and the remedy should fit the violation and flow from the theory or theories of competitive harm. Effective remedies preserve the efficiencies created by a merger, to the extent possible, without compromising the benefits that result from maintaining competitive markets.

In other words, when choosing among remedies to address a potential competitive harm, don’t toss out the good with the bad.

Unfortunately, the least-restrictive principle is completely ignored by Title II proponents, who seek to prohibit an entire category of business agreements that could be abused. They are not sensitive to the possibility that some of these priority-delivery arrangements may actually be good for consumers and both parties to the transaction, while leaving no other parties worse off. And they are not sensitive to the possibility that, once reclassified, the agency could impose other common-carriage obligations on ISPs—for example, regulating broadband access prices—in a way that undermines their incentive to invest. Is it any wonder why 28 Internet CEOs just issued a letter saying that Title II regulation could threaten their business plans?

By the same logic, Title II proponents would counsel you to install a fire hydrant (with matching hose) in your kitchen. Damn the aesthetics and don’t worry about the floods (or horrible water pressure in your very cool, new soaker shower). Problem is that these folks are not thinking like good economists.


Follow me on twitter @halsinger

Leave a comment

Has The FCC Made Its Case To Restrict Certain Bidders In The Broadcast-Spectrum Incentive Auction?

FCC Chairman Tom Wheeler recently signaled that his agency is considering certain bidding restrictions for the upcoming broadcast-spectrum auction that are specifically targeted at the two largest nationwide providers. At some ill-defined point in the auction, the restrictions reportedly would be imposed on any bidder that has more than one third of the available “low-band” spectrum in a market.

And guess who holds more than one-third of “low-band” spectrum in any particular market? AT&T and Verizon. As a result of the proposed restrictions, between 40 and 50 percent of the spectrum blocks in a given band plan would be off limits for the two mobile broadband companies best positioned to battle cable modem providers. Is this a good thing?

The best policy justification for bidding restrictions in an auction is the presence of monopoly power. The theory is that a monopolist is willing to pay more to cement its position than a rival is willing to pay to displace the monopolist. Although the auction might cause the monopolist to surrender a good portion of its profits to the auctioneer, at the end of the day, consumers are still beholden to monopoly prices. (The second best justification for a restriction is that there is something special about “low-band” spectrum—without it, smaller carriers cannot compete effectively. I have rebutted this justification here.)

A review of the evidence suggests that no wireless carrier is exercising monopoly power—that is, setting prices above competitive levels or restricting output.

Recent price cuts in response to T-Mobile’s “Uncarrier” initiatives and no-contract plans have put downward pressure on wireless margins. In February, AT&T cut its Mobile Share shared data plan prices (with 10 GB of data) to $160 per month for four phone lines; in response, Verizon matched that pricing in April. In March, AT&T also cut the price of its smaller shared data plan (with 2 GB of data) by $15 to $65 per month for one phone line. These pricing episodes are hardly consistent with the notion of monopoly power.

Perhaps these recent price cuts mask a longer trend of rising prices? Not so. According to the FCC’s 2013 Wireless Competition report, competition is robust:

  • Monthly average revenue per unit (“ARPU”) for wireless service declined from $48.04 in 2006 to $46.63 in 2012; wireless voice revenue per minute has declined from $0.06 to $0.05 over the same period.
  • Voice revenue per minute in the United States ($0.033) is less than one third of the European average.
  • U.S. mobile subscribers talked an average of 945 minutes per month on their mobile phones in 2011, compared with 134 minutes in Japan and 170 minutes in Western Europe.
  • And the United States has the second least concentrated market structure in a Bank of America survey of ten countries, behind only the United Kingdom.

After 408 pages of excruciating detail on the state of wireless competition, the FCC is hard-pressed to identify any data consistent with monopoly power. And without a showing of monopoly power, the social benefits of these bidder restrictions are likely insignificant.

On the other hand, unwarranted restrictions can inflict significant losses on society in three important ways.

First, assuming AT&T even shows up to the auction, prices on the restricted blocks will be significantly less than the prices in the non-restricted blocks. Although there is some chance that prices in the non-restricted blocks could be higher (due to the artificial scarcity created by the restrictions), the FCC is exposing itself and the taxpayer to a considerable risk of diminished auction revenues—revenues needed to fund deficit reduction, build-out of an interoperable public-safety network, and other priorities enumerated in The Middle Class Tax Relief and Job Creation Act of 2012. And auction revenues are needed to compensate broadcasters interested in giving up their spectrum. The amount of the broadcast spectrum that will be available for reallocation to wireless broadband will depend critically on the broadcasters’ perception of auction prices; the law of supply dictates that there will be less spectrum available for sale the lower the expected price.

Second, by setting aside valuable spectrum, the FCC is creating an attractive opportunity for firms to engage in regulatory arbitrage. Set asides will encourage firms not interested in building networks but instead buying spectrum to flip it later for a windfall. History has shown that set-aside spectrum sits fallow for years, staving off the sort of broadband deployment that Congress desires. Competition for the arbitrage opportunity leads to wasteful “rent seeking” activity, which represents another loss. Allowing the carrier that will ultimately deploy the spectrum to purchase it immediately and directly (rather than through a wasteful and superfluous middleman) is clearly the more efficient choice.

Third, efficiency dictates that spectrum goes to the wireless carriers that value it the most. If an incumbent carrier facing a spectrum crunch is willing to pay more for the next chunk of available spectrum than an entrant, assigning the spectrum to the entrant represents a misallocation of society’s resources. Relatedly, a significant challenge facing the FCC is injecting competition into the broadband marketplace. According to the FCC’s most recent data, a full 19 percent of U.S. homes were beholden to a single provider of broadband service (including wireless operators) capable of delivering download speeds of 10 Mbps. Wireless broadband could impose significant discipline on cable operators in these pockets if the FCC opens up the spectrum spigot to all firms, as opposed to parsing out thin slices to smaller companies.

In sum, the FCC has failed to meet its evidentiary burden for the use of bidding restrictions as currently proposed for the upcoming incentive auctions. There is no compelling evidence of monopoly power in the wireless sector. And there has been no attempt to prove that smaller carriers need access to “low-band” spectrum to compete effectively against their larger competitors. Until those burdens are met, the FCC should let the auction blocks fall where they may.

Follow me on twitter @halsinger

Leave a comment

Connections Between Communications Networks: Should The FCC Breathe Life Into Internet Middlemen?

The NetflixComcast arrangement has brought fresh accusations of so-called “net neutrality” violations. The flames were stoked when Netflix’s CEO, in a blog posting, bemoaned the plight of “intermediaries” such as Akamai, Cogent, and Level 3; he advocates new “net neutrality” rules to prevent Internet service providers (ISPs) from charging a toll for interconnection to these Internet middlemen. Some commentators fear that, if direct connections between ISPs and content providers such as Netflix proliferate, Internet middlemen “will become mere resellers of access” or “may be cut out of the loop altogether.”

In light of the D.C. Circuit’s decision in Verizon v. FCC, which rejected the FCC’s effective prohibition of payments from content providers to ISPs, a widespread acceptance is emerging for case-by-case review of discrimination complaints by a content provider against an ISP. In this middle ground, pay-for-priority deals would be tolerated but the arrangements would be policed for abuses by the FCC. Should the same protections be extended to resolve interconnection disputes involving Internet middlemen?

Although I accept the case for regulatory protections of content providers, the FCC may not be the right venue to deal with the antitrust and political aspects of interconnection disputes involving Internet middlemen. Before explaining why, however, a brief refresher on how we got here is in order.

The term “net neutrality” was first coined by Tim Wu to describe a world in which all packets on the Internet are equal and should be treated that way by ISPs. This vision implies no favoritism of an individual content provider’s packets over another’s. And the strong form of net neutrality implies that if there is priority treatment of a content provider’s packets, it should be priced at zero.

Peering arrangements between networks have always been outside the purview of net neutrality, because the FCC had expressed the view for years that the backbone market was competitive and not in need of regulation. In their dealings with ISPs, Internet middlemen such as Cogent carry the packets of myriad content providers, making discrimination vis-à-vis an individual content provider impractical. Moreover, unlike “settlement-free” peering, exchanges of unequal amounts of traffic between two networks have involved positive prices for years, which is also inconsistent with the zero-pricing principle of net neutrality. Accordingly, the FCC did not extend net-neutrality protections to peering arrangements in its Open Internet Order (see footnote 209).

Flash forward to the NetflixComcast arrangement, which bypasses the middlemen. To the extent that such direct connections become the new norm, these intermediaries may find themselves marginalized, or even obsolete. What role, if any, should the FCC play in preventing such a development? Stated differently, should ISPs be forced to deal with these middlemen at regulated interconnection rates?

The answer to this question can be informed by a limiting principle that should define the scope of the FCC generally: Is there some important social objective not recognized by antitrust laws for getting the FCC involved in these affairs?

Setting aside any of the particulars of the NetflixComcast arrangement, an ISP could make life difficult for these middlemen by requiring content providers to purchase data transport or content delivery services (“middle-mile services”) from the ISP as a condition of getting access to its customers (or getting a working connection). As a variant of this strategy, an ISP could refuse to supply terminating access to middlemen, forcing the content provider to deal directly with the access provider.

It is not clear how additional protections for these middlemen, above and beyond those afforded by antitrust laws, would advance any important social objective: Why is it a good thing to promote standalone providers of these middle-mile services via regulatory protection? Unlike content providers, who generate positive spillovers (information and artistic content can be viewed as “public goods”) and thus cannot be expected to monetize their investment, the Internet middlemen are more akin to resellers of a homogenous product (data packets), and are likely to capture the entire value-added of these services.

The best justification I can conceive for the FCC’s providing a backstop for these middlemen is that it is better for providers of these services to be independent from the provision of broadband access, because their independence ensures that content providers will get “better” treatment in middle-mile services. But the question of whether this “better” treatment is a worthy social objective is really a political judgment that should come from Congress, not the FCC.

To the extent that independent firms can provide middle-mile services at a lower cost than ISPs, basic Coasian economics predicts that market forces should ensure their survival. Why would Comcast, in its “make or buy” decision, exclude Akamai from the market if doing so would impose higher costs on Comcast? Over the last three months, Akamai has outperformed the Nasdaq index (a 30% return versus virtually no return), indicating that financial markets are not discouraged by these direct, pay-for-priority developments.

Moreover, to the extent any exclusionary conduct by an ISP leads to higher prices or lower output, the antitrust laws offer all the protection these middlemen need. Competition laws were designed to prevent, among other things, a dominant firm from leveraging its market power from one market (broadband access) into another (data transport or content delivery service). Because these cases could get a serious hearing in an antitrust court, middlemen do not need any additional regulatory protection.

In what appears to be a plea for regulatory intervention, Cogent’s CEO recently claimed that ISPs “are attempting to leverage their monopoly on broadband residential Internet connections to increase their profits by imposing tolls on traffic requested by their customers and delivered by other Internet service providers.” Although such comments likely caught the telecom regulator’s attention, by articulating an antitrust claim, Cogent made a pretty good case for why no FCC involvement was needed.

The case for preemptive protections for Internet middlemen is ultimately based in politics, rather than economics. Either we let the market naturally develop and let middlemen live or die on the economic merits, or Congress bans vertical integration in this space because it produces a political outcome we cannot tolerate. There is no added economic benefit for the FCC to dicker around the margins.


Follow me on twitter @halsinger

Leave a comment