By Francisco Beneke*
Regulations that ban paid prioritization have been discussed in different countries around the world. Prioritizing Internet content is either allowing certain data to travel faster, to not count it towards an end user’s maximum consumption cap, or any other form of preference that an Internet Service Provider (ISP) can contract with a content provider. The first country that adopted a ban on paid prioritization was Chile in 2010. Regulation of this type was also adopted in Mexico as part of a reform package in 2012 in the telecommunications sector, which included constitutional reforms that make access to the Internet a constitutional right. Other countries that prohibit prioritization include the Netherlands and Ecuador. The last country to issue such a prohibition has been the United States amidst a heated debate. The nascent trend of regulation and the adoption of such a rule by influential countries in their regions and worldwide can be a catalyst for many other countries to discuss and eventually adopt similar regulations. Therefore, it is important to sketch and analyze the main arguments that surround the debate.
In the US, the Open Internet rules adopted by the Federal Communications Commission (FCC) ban paid prioritization by means of classifying broadband Internet suppliers as common carriers under Title II of the Communications Act. The reclassification means that the services provided as a common carrier are excluded from Federal Trade Commission (FTC) jurisdiction. This is a particularity specific to the United States legal system and, therefore, it is not something that must necessarily be an issue in other countries. Having clarified this, we can move on to the substantive arguments of the debate.
Before the FCC adopted on a divided vote (3-2) the Open Internet rules, a group of scholars wrote a letter to the FTC Commissioners requesting them to advocate against the rules. They argued that the ban on paid prioritization amounts to a per se prohibition and that there is not sufficient evidence that supports the main justifications for instituting such a broad-sweeping rule: the high likelihood of significant harm and the low likelihood of both overlooking possible pro-competitive justifications and deterring pro-competitive behavior.
Another group of scholars reacted to the aforementioned letter and wrote to the Commissioners supporting the ban. On their part, they argued that there is enough evidence on the net benefits to competition that a bright-line prohibition on paid prioritization can have. They point to the fact that the United States Court of Appeals for the D.C. Circuit found the FCC’s position as reasonable and grounded in substantial evidence in Verizon v. Federal Communications Commission, 740 F.3d 623 (D.C. Cir. 2014). In that occasion, the court stroke down the FCC’s ban but on the ground that it had relied on an inadequate statutory basis. The classification of broadband services under Title II solves that problem.
The opponents of the Open Internet rules favor a rule-of-reason kind of approach to paid prioritization. They point to an FTC report in 2007, which states that the broadband industry is a dynamic one and that it is moving towards more competition. The report also points out that it is not clear if ISPs have a clear incentive to discriminate against data from non-affiliated content providers. Furthermore, even if such discrimination takes place, the FTC argues that it is not possible to know a priori if the net effect on consumer welfare will be negative. For the academics that oppose the regulations, price discrimination will lower the costs of content providers that do not need to use a fast lane (such as email services) and only firms that rely on greater speeds need to bear a higher price (e.g. Skype).
Supporters of the regulations, on the other hand, believe that the price discrimination schemes will take an anticompetitive turn because of the gatekeeper position of ISPs.
The argument behind the claim that the effects of paid prioritization need not be anticompetitive relies on competition between ISPs as a force that will discipline the market. There are at least two reasons why on this instance the trust in market forces can be misplaced. One of them has to do with information asymmetries. The bargaining power of the ISP can be higher compared to the content provider even if it does not have a dominant market share. The ISP can bargain hard by lowering the download speeds from, for example, Netflix without fearing that its customers are going to divert to other ISPs. The reason is that it will be hard for end users to distinguish who is to blame for the lower quality of the service and, therefore, they will have a weaker incentive to switch to other ISPs. The second point has to do with switching costs. Even if one assumes that customers have perfect information, it is still costly to switch from one ISP to another because of contract commitments and brand loyalty, among other factors. This is why the supporters of the ban talk about a terminating monopoly or gatekeeping position from the part of ISPs.
In addition, the supporters of the ban argue that a rule-of-reason approach cannot be relied on because, at least in the US, the hurdle to prove exclusionary conduct is high, which will discourage many administrative and judicial complaints. Such an approach would also mean higher administrative costs in adjudicating disputes, which would be avoided by a bright-line ban on paid prioritization.
Both sides of the debate are also divided on the effects that the regulations will have on innovation. For the supporters of the Open Internet rules, the ban will promote innovation from content providers. For the opponents, forbidding paid prioritization will chill innovation on the part of ISPs.
The innovation argument of the supporters’ side is built on the idea of a virtuous cycle that flows in the following direction: when new content or apps are developed by edge providers, more people use the internet, which in turn pushes ISPs to innovate and increase their broadband capacity. In the supporters’ letter to the FTC words: “if the next Facebook has to pay for an Internet fast lane, the next Mark Zuckerberg might go into investment banking instead of creating the next big new thing on the Internet”.
The opponents of the ban do not explain in detail why innovation would be hindered by the regulations but implicit in their stance is that the price discrimination mechanism and the increased profits it can bring might act as an incentive to develop faster broadband technology (an argument akin to what patent rights do to innovators).
The issue of the effects on innovation is complex and it is hard to make a good prediction on how the innovation in these markets really works. Regarding the virtuous cycle described by the supporters of the ban, the causality can feasibly run in a different way. It might be that investments in broadband capacity and innovations that increase traffic speeds incentivize innovation on the content providers’ side. This may be true at least for developers of apps and content who rely ever increasingly on the speed of networks. On this respect, the course of the innovation cycle is ambiguous from a prospective point of view.
Regarding the opponents’ argument on innovation effects, extracting revenue from content providers as an incentive to improve the network will come as a trade-off with innovation from the content side. In other words, paid prioritization could, in theory, promote innovation from ISPs but at the cost of less innovation from edge providers. On this instance, this is an effect that one can predict with higher certainty because paid prioritization will decrease the appropriability of benefits on the side of developers of apps, web pages, and so on.
The argument of the opponents’ side also makes sense only if the content providers’ demand for higher speeds is more inelastic than the end consumers’ demand. That is, if for every dollar less charged to consumers the ISP can charge more than a dollar to edge providers. In this case, the price discrimination scheme will allow ISPs to increase both their revenues and their incentive to innovate. However, the lower elasticity from the side of content providers can be an argument against the desirability of the trade-off between innovation from ISPs and edge providers from a dynamic perspective.
One reason to think that ISPs face a less elastic demand on the content providers side is because of the terminating monopoly that was explained above. Whereas a certain market share would not grant an ISP market power in the end user market, it could do so in the market for paid prioritization. The reason is that if an ISP is of a certain size, the content provider cannot reasonably succeed if it cannot have access to the network’s users, which in turn weakens its bargaining position against the ISP. Therefore, it is not necessary that the latter enjoy what is conventionally believed to be a dominant market share in order to extract supra-competitive rents from edge providers.
One issue that will not be analyzed in depth in this post is the discussion around the narrowness of antitrust law’s objectives and their inadequacy to protect what is at stake in net neutrality. In this respect, the position of dissenters of the ban has been misunderstood. They do not advocate for antitrust laws to solve the issue from a consumer welfare point of view, but rather for an antitrust law-like approach. They propose a rule-of-reason framework in which all relevant policy objectives must be weighed. Therefore, the criticism that consumer welfare is too narrow to justify the benefits and harms of paid prioritization is not applicable.
Finally, another point of dissention is the freedom of ISPs to manage the traffic on their network. Paying for transiting in a high-speed lane can be viewed as a way of increasing efficiency in the market. Why? Because every content provider that develops an attractive app imposes an externality on previous firms by congesting the highway. Therefore, it would only be fair if a new content provider internalizes part of the cost it causes. Nevertheless, the argument has a weak point. Even without being able to charge content providers for a toll, the ISP can still manage traffic on its network in an efficient way without the incentives of charging monopoly fees to edge providers.
To conclude, although both sides have valid arguments, a rule-of-reason approach to solve the problem causes more costs than benefits. The reliance on a case-by-case ex post solution to the problem can be misplaced because it does not account for the difficulty of developing a system that adequately deters harmful conduct. An ex post law enforcement approach would make more sense if there is reason to believe that the market will generally perform well, which is the argument the opponents of the ban make. However, for the reasons stated above, there is stronger evidence that supports the theory of a terminating monopoly held by ISPs. One can also look into other platform markets where there appeared to be competition but in the end the platforms did enjoy market power against one side of the platforms. Specifically, I am referring to the credit card payment market. Visa and MasterCard where successfully prosecuted for anticompetitive conduct against merchants that accept credit cards, followed by another ruling against American Express on similar grounds. The resemblance of the arguments in the credit card cases and on the debate around the Open Internet rules is hard to miss, and the former sets an example that tilts the balance in favor of banning paid prioritization.
*Co-editor, Developing World Antitrust
Developing World Antitrust 