Network neutrality and quality of service: what a nondiscrimination rule should look like.

• Author: van Schewick, Barbara
• Position: II. Proposals for Nondiscrimination Rules C. Case-by-Case Approaches 4. Problems with Case-by-Case Adjudication d. Strategic Incentives of Policymakers and Big Stakeholders through D. More Nuanced Rules 2. Substantive Approaches a. The First Approach ii. Allowing Discrimination Among Classes of Applications That Are Not Alike C. User Choice, p....

d. Strategic incentives of policymakers and big stakeholders

In spite of these considerable social costs associated with general or ambiguous nondiscrimination standards, the strategic incentives of legislators or regulators who consider adopting network neutrality rules and of the big stakeholders on both sides of the debate are aligned in favor of such a scheme. (280) Stakeholders cannot agree which discriminatory behavior is acceptable today; it is unlikely that they will be able to do so in the future. There are large, well-financed entities on both sides of the network neutrality debate. Any substantive decision would take on either the large, well-financed, well-organized, and politically influential network providers (e.g., in the United States, AT&T, Verizon, Comcast, and Time Warner) or big providers of Internet applications, content, or services such as Google and Amazon. Under these circumstances, adopting a very general or ambiguous nondiscrimination rule today constitutes an attractive compromise, since the controversial question is not decided one way or the other. (281)

The legislator or regulator can reap any immediate benefits associated with adopting network neutrality rules (282) while avoiding the immediate political costs of taking on powerful interests on one side of the debate. (283) While an ambiguous or general nondiscrimination rule that is applied case by case is more difficult and costly to apply and enforce in the future, these costs will not be borne by the entity adopting the rule. (284) If the nondiscrimination rule is adopted through legislation, it will most likely be enforced by a regulatory agency (e.g., in the United States, by the FCC). Even if the nondiscrimination rule is adopted by a regulatory agency such as the FCC through administrative rulemaking, it may be enforced by future members of the agency (e.g., in the case of the FCC, by future Commissioners) or by another entity within the agency (e.g., the FCC's Enforcement Bureau). The social costs of this type of rule will not be borne by the entity adopting the nondiscrimination rule, either. (285)

Big stakeholders support this type of nondiscrimination rule because each side can claim a win (or at least a nonloss) and gets a second chance to influence the ultimate decision over the legality of specific practices in the context of individual adjudications in the future. While adjudications are costly, big stakeholders have the resources to play the case-by-case game and prevail in future adjudications. Given these incentives, it is not surprising that the proposals for a general or ambiguous nondiscrimination rule described above emerged from industry negotiations at the FCC (286) and in Congress, (287) or, as in the Verizon-Google legislative framework proposal, from direct negotiations between two big stakeholders on opposite sides of the debate. (288)

4. More Nuanced Rules

   A final group of proposals would adopt more nuanced rules that specify in advance which differential treatment is and is not allowed. Like the standards-based approaches discussed above, these proposals recognize that some forms of discrimination are socially beneficial, while others are socially harmful. Contrary to the standards-based approaches, however, these proposals define in advance what constitutes acceptable and unacceptable discrimination to avoid the social costs associated with leaving the decision about specific discriminatory conduct to future case-by-case adjudications.

   Out of the three proposals in this category, only one--ban application-specific discrimination, but allow application-agnostic discrimination (Part II.D.2.b)--accurately distinguishes socially beneficial from socially harmful discrimination. This is the rule that policymakers should adopt. By contrast, the other two proposals--ban discrimination that is not disclosed (Part II.D.1) and ban discrimination that does not treat like traffic alike (Part II.D.2.a)--do not adequately protect the values that network neutrality rules are intended to protect and should be rejected.

   1. Formal approaches: ban discrimination that is not disclosed

   The first set of approaches in this group bans discrimination that is not disclosed, distinguishing between socially beneficial and socially harmful practices using the formal criterion of whether the network provider disclosed the differential treatment. Alternatively, a network neutrality regime might allow blocking or discrimination but require Internet service providers to disclose any blocking or discrimination that occurs. (289) In January 2014, the Court of Appeals for the D.C. Circuit struck down the Open Internet Order's rules against blocking and discrimination but upheld the disclosure rule. (290) Thus, until the FCC adopts new network neutrality rules, the current network neutrality regime in the United States constitutes an example of this approach.

   In 2009, the European Union adopted this approach following the review of its regulatory framework for telecommunications services. (291) The European Universal Service Directive neither requires network providers to impose restrictions on users' use of applications nor prevents them from doing so. (292) It does, however, require Internet access service providers to inform their customers about any limits on access to or the use of services and applications, and about any traffic management measures and their impact on service quality. This information must be disclosed in the terms of the contract and when practices change.

   This approach is based on the idea that if a network provider discriminates against an application that users would like to use, users can switch to another network provider that does not discriminate against the affected application. The threat of switching, proponents of this approach assume, will discipline providers. (293)

   In line with this reasoning, participants in the network neutrality debate often assume that the viability of disclosure rules as a substitute for substantive regulation solely depends on the amount of competition in the market for Internet access services. After all, if there is no competition, there will be no other providers that consumers can switch to in response to discriminatory conduct, making it impossible for them to discipline providers. Based on this reasoning, participants in the debate often assume that mandatory disclosure alone will be sufficient to discipline wireline providers in Europe or in countries like Canada, where the market for wireline Internet access is generally more competitive than in the United States. (294) Similar arguments are often made for mobile Internet access, where users often have a choice between three or more competitors. (295)

   These arguments fail to recognize that the market for Internet services is characterized by a number of factors--incomplete customer information, product differentiation in the market for Internet access and for wireline and wireless bundles, and switching costs--that limit the effectiveness of competition and reduce consumers' willingness to switch. Rules that require network providers to disclose whether and how they interfere with applications and content on their networks reduce the problem of incomplete customer information, though only to some extent. They do not remove any of the other problems. As a result, they still leave the network provider with a substantial degree of market power over its customers, enabling it to restrict some applications and content on its network without losing too many Internet service customers. (296) They also do not affect the cognitive biases, cognitive limitations, and externality problems that lead users to underestimate the benefits of switching providers compared to what would be in the public interest. Thus, even if there is competition in the market for Internet access services, disclosure cannot replace substantive regulation as a tool to discipline providers. (297)

   a. Problems with disclosure-only network neutrality regimes

   Disclosure can only discipline providers if there is effective competition. (298) In order for disclosure to have a disciplining effect, customers need to realize that the network provider is discriminating against an application they want to use. They need to be able to switch to another provider that meets their needs and does not impose a similar restriction, and they need to be able to do so at low cost. Even if there is competition in the market for Internet access services, these conditions will often not be met.

   i. Consumers' incomplete knowledge, cognitive limitations, and cognitive biases

   First, even with disclosure, users' decision to switch will suffer from incomplete knowledge, cognitive limitations, and cognitive biases. Users may not realize that their network provider is interfering with their application. (299) An application's bad performance may have many reasons (e.g., bad application design, insufficient server capacity, network congestion, problems on the network of another Internet service provider), and network provider interference will not necessarily be the first explanation that comes to mind. (300) Even if users consider that possibility, many will lack the expertise to investigate the cause of the bad performance. (301) While mandatory disclosure of discriminatory practices is intended to address this problem, experience with disclosure requirements in other contexts shows that disclosure is usually less effective at informing consumers than would be necessary for disclosure to have the intended effect. (302) Consumers often do not read disclosures, and in many cases, those who read them do not understand them. (303) For those who read and understand the disclosure, knowing which practices their network provider engages in will not necessarily allow them to make an informed decision. Many users lack the technical expertise to understand how the disclosed...