Tim Lee draws our attention to a serious, serious problem in draft spectrum reform legislation:
Under the procedure envisioned by House Republicans, companies (like Google, Microsoft or Intel) that want to see more unlicensed spectrum would bid on that option. If their bids added up to more than the top bid for an exclusive license, then that block of spectrum would become available for everyone to use. This is a recipe for systematically under-valuing unlicensed spectrum because each individual supporter of unlicensed spectrum has an incentive to “free ride.” The auction procedure is likely to systematically under-value unlicensed spectrum.
While Tim believes that we need more licensed spectrum, for reasons he has explained in an earlier post, he also believes that assigning property rights to the entire spectrum would be a mistake and that auctions aren’t the way to determine which part of the spectrum should remain wild and free.
But auctions are not a good way to decide whether a particular block of spectrum will be licensed or unlicensed. Rather, the auction mechanism described in the discussion draft is a way of pretending to take the unlicensed option seriously while stacking the deck against spectrum actually going to unlicensed uses.
Tim has also argued that unlicensed spectrum proposals with a lot of strings attached should be opposed. As always, Tim is eminently sensible on this subject, though I am sympathetic to Yochai Benkler’s argument that we don’t need to assign property rights to the spectrum, which Tim suggests goes a bit too far.
In “Overcoming Agoraphobia,” Benkler writes the following:
The rhetorical effects of treating spectrum as “a resource” obscure the more important choice to be made with respect to radio communications: whether to regulate them by centralizing control of wireless communications or, alternatively, by establishing a means of allowing users to coordinate their wireless transmissions multilaterally. Once we understand that the question is how to regulate the use of equipment, not of “a resource,” we will be able to recognize that we have alternative regulatory models in our society. In the case of cars or networked computers, which involve similar coordination problems, our social choice has not been to give a small number of users an exclusive license or property right to control an input essential to effective use of the equipment. Instead, in the case of automobiles, we have chosen to allow anyone to buy and use the equipment, subject to certain “rules of the road” that allow equipment users to coordinate their use and avoid interference.
This essay was published in 1998, and Benkler’s latest take on the idea of a spectrum commons was published in 2002 and included a cautionary note:
My conclusion is that open wireless networks are likely to be better at optimizing the ability of users to communicate without wires than could spectrum property based systems. Our relative lack of actual experience with either open wireless networks or spectrum property-based systems suggests, however, that it is too early to recommend a wholesale transition to one new approach or the other on the basis of this qualitative conclusion. Instead, the last part of the article proposes a series of regulatory moves that could permit the emergence of a long-term and large scale market test of the two alternatives, a test that within a few years could provide us with substantial practical experience to help answer the question of whether the optimal approach would be to adopt spectrum property based systems, open wireless systems, or some combination of the two approaches.
Benkler’s cautious, incremental approach is very much in tune with Tim’s take.
There is one source of solace: if we really do screw up our spectrum policy, we could turn to light-based data transmission.