CircleID

I was expecting something that would discuss the unique properties, if any, of the domain name market, and the types of inventory theoretically available to allocate, and the expected outcomes for the various types of auctions, and some showing that for some desired policy goals, whether greatest gain to seller or lowest loss to buyer, or something entirely different, the expected outcomes.

This would assist the better informed, bottom-up, stakeholder-driven, consensus policy making.

Is the domain name market indistinguishable from the spectrum market? If name spaces are distinguishable from units of spectrum, and the curvature of the earth suggests that spectrum is, due to physics, only locally unique, a property that would be if true here if multiple name space roots were assumed to be true. I suspect that assumption is not widely shared, at least by nominal seller and buyers, as that would have some affect on value and price.

It is of course desirable, trendy even, to impute to some rather banal market the exotic properties of another market, and the spectrum market is full of exciting players, the crowns of national governments, the capitalist barronies, robber and otherwise, survivors of regulation, mergers, acquisitions and deregulation in the broadcast consumer content and the broadcast consumer access markets, the churches, both the civic religions of reform and equity, and the deists thumping their texts, and a vibrant press making sure that the exciting narrative excludes organized labor, social poverty, gender, and anything else that distracts from a good story in the corporate media.

But are we few, we grungy few, from VeriSign Global Registry Services down to start-up would-be registries, actually that interesting, that exciting? The Fourth Estate is not convinced. We don't get above the fold in the dailies or the first minutes of the broadcast news unless (a) someone recycles the fabricated quote that Al Gore invented the internet, or (b) hacker wolves pull down another sick corporate caribou, or (c) it's red-baiting season somewhere.

Assume for the moment that the spectrum allocation model is applicable. In New Zealand^{1 & 2}, spectrum is held jointly by Maori and Pekeha polities, under the Treaty of Waitangi, and for all the failures in the implementation of that model, it attempts to benefit Maori and Pekeha not just Pekeha. In other territorial jurisdictions (see "physics", above) there are other conditions. Which specific set of conditions of the spectrum allocation model are applicable? The first ICANN Board to approve new generic Top-Level Domains (gTLDs) adopted a model like New Zealand's, four of the initial seven new gTLDs, .info, .biz, .name and .pro were non-sponsored, and three of the initial seven new gTLDs, .aero, .coop, and .museum, were sponsored. For all the failures in the implementation of that model, it attempts to benefit speculators and communities, not just speculators.

So, which of single-sided or multi-sided, and if single-sided, which of single bid or multi bid, and if single bid, which of sealed first price or sealed second price, and if multi bid, which of buy-out, all-pay, dutch, english, or reverse, are we to use? Here's a graphic:

Does the same answer offer reasonable expectations for the single letter in the legacy TLDs allocation, assuming that goes forward, and for late-phase irreconcilable multi-bidder string contention in the present proposed new gTLD process? Is the answer the same where there are two or more high-cap bidders? Where there is one high-cap bidder and one or more low-cap bidders? Where there are two or more low-cap bidders? Finally, since ICANN exists to transform a public single-source contract market into a competitive private and public market, what are the rules on VeriSign? None? If none, how do "auctions" remain no worse than neutral, and not regressive, towards that fundamental policy goal?

The consultant's letter, for it is little more than that, which waxes lyrical on spectrum auctions, and auctions as theory, leaves me just as ignorant after reading it quite a few times as I was when "a study" was first proposed. Which kind of auction and why?

Bottom line to staff. The pie is not done. The crust is soggy and the filling is cold. Place back in oven. Check oven is lit. Else start a Policy Development Process (PDP). It can't be worse than the Fast Flux PDP, and some of us have given the matter a bit of thought.

¹ Attorney-General v New Zealand Maori Council (No 1) [1991] 2 NZLR 129 and (No 2) 147 (Radio Spectrum)
² New Zealand Maori Council v Attorney-General [1992] 2 NZLR 576 (Broadcasting Assets)

By Eric Brunner-Williams, Mathematician. Visit the blog maintained by Eric Brunner-Williams here.

Related topics: Registry Services, ICANN, Internet Governance, Policy & Regulation, Top-Level Domains