From the ICANN does good department, ICANN has released a report by Charles Rivers Associates on the economic relationship between registries and registrars. Although we have not fully assesed the report as a whole yet, here is a quick summary of some interesting conclusions.
The report correctly characterizes ICANN's registry regulation as a price cap regime with vertical separation of the registry and registrar functions. It notes that vertical separation effectively prevents discrimination by registries against registrars, but notes that such separation can actually increase consumer costs by introducing two profit margins rather than one. It also constrains business model innovation.
The report notes that the benefits of integrating registries and registrars are: creating the potential for more innovative business models, increased incentives to provide registrar services to new gTLDs (independent registrars know that 95% of their business is going to come from .com, .net and .org and may show little interesting in marketing and promoting new TLDs), and reduced double margins.
In essence, the report recommends that it would be ok to relax or eliminate the vertical separation requirement when there are no price caps on the registry. But most major registries (.com, .net, .org, .info, .mobi, etc.) operate under a price cap.
The case for eliminating the vertical separation is strongest in the case of a single-owner TLD, e.g., if IBM Corporation operated a .ibm.
The report notes that a for large company interested in having its own TLD for use by its employees, a requirement that registry and registrar functions be separated would be “especially inefficient” because it would require such a company to go through third parties to register its own employees. The report also highlights the security concerns inherent in separation, especially if equal access requirements were also in force. If a major business rival to the single owner (say, HP in our IBM example) obtained ICANN accreditation as a registrar it would have access to the firm's registry.
It also recommends a hybrid regime that would allow registrars to own registries and registries to own registrars, but prevent either registry from serving as a registrar for its own registry. This model minimizes the main risk of integration — discrimination against unaffiliated registrars — but also minimizes many of the benefits of integration. Mainly it simply allows existing players to diversify their position in the DNS market.
All in all the report takes a fairly cautious approach to its recommendations and its main benefits lie in the introduction of more rigorous economic analysis into the ICANN policy dialogue. We will review the whole report in greater detail and report anything of note later.