The Case for gTLD Auctions: A Framework for Evaluating Domain Name Policy

Abstract

The Internet Corporation for Assigned Names and Numbers (ICANN) is charged with designing and implementing policies for names, which function as address identifiers for computers on the Internet. Domain names can have tremendous economic value, as do the registration services that control the databases that resolve domain names into numerical Internet addresses. Economic value is partly the consequence of scarcity in the domain name space. Incumbents strive to perpetuate scarcity so as to preserve the values of their domain names and name services. ICANN has struggled to establish a coherent domain name policy, especially for generic Top Level Domains (gTLDs), for almost the entirety of its 5-year existence. In this policy paper we first present a framework for evaluating Internet domain name policy in general and gTLD policy in particular. We demonstrate that the domain name system (DNS) is what economists call a good and not a good. We then show that, because of what economists call effects, root service, the part of the domain name system that handles the TLDs, is a natural monopoly. At the first level of analysis, we establish that gTLD policy ought to take into account these two fundamental economic facts: (1) domain name service is a private good, and (2) root service is a natural monopoly created by networking effects. We next argue for a specific set of conclusions about gTLD policy. In particular, we demonstrate that there is a compelling case for allowing the market to operate in the creation of new gTLDs. This could be accomplished through a variety of mechanisms, including a rule of first occupation or through an auction. Although the creation of gTLDs should allow for the operation of market forces, it does not follow that ICANN itself should act as a profit-maximizer. Instead, we reason that, because ICANN is a non-profit corporation and because it is the trustee for a natural monopoly, ICANN ought to act in the public interest. We conclude that ICANN should structure the expansion of the root in a way that insures the stability and efficiency of root service. We offer a specific proposal for an auction of new gTLDs, and show that this approach offers substantial advantages over current domain name policy. Our conclusions are reinforced by a set of comparisons between the policy questions faced by ICANN as both a participant in and regulator of the DNS and with analogous policy questions faced by market participants and regulators in other sectors of the telecommunications system. In particular, we argue that there are important insights to be gleaned and lessons to be learned by comparing domain name service with broadcasting and telephone service.