Panel discussion: Commons versus property for spectrum allocation

A panel on property versus commons:

Gregory "Deputy Director, Stanford Institute for Economic Policy Research" Rosston: companies should be allowed to bid to operate "private commons," just as Disney World (!) is a private park.

Ed: Er, sure. So all of the rights that we have in public (speech, search and seizure, etc), will exist only at the sufferance of the market and private contract law?

Stuart "Yale Law" Benjamin: If the transaction cost for private entities is too high, that could be an artifact of the fact that we make small allocations to private entities — maybe we should allocate large chunks of spectrum to private entities. We should wait a decade to see what shakes out. Corporations may attempt to create anticompetitive/anti-interop technology. But what about the private power: one company could control everything? My answer: If we auction off enough spectrum, we'll get hundreds of competing companies. We can limit predatory powers with spectrum caps and band managers who control the spectrum without providing the service.

Howard "Berkeley Public Policy Group" Shelanski: Yochai's right: spectrum isn't property, it's the right to speak. (Ed: Amen!) The only thing that matters is displacement costs. Both sides want to do the same thing: maximize the value of speech. They use different yard-sticks, though. What is the value of different speech? In a perfect market, there would be no divergence between the audience value of the speech and the speaker's value of the speech. Those who bid the most would be those who had the most burning need to speak or the hungriest audience. In a commons approach, you still need regulation — the distinction between commons and property rights is who is doing the regulations. In a commons model, if a and b's communication displaces x and y's communication, you'd want to know whether a and b were having a more valuable conversation. If you rely on parties to reveal the amount of interference they experience and the capacity they need, they will report this strategically, and you need someone to monitor their truthfulness.

Dave "Former CTO of the FCC" Farber: I have a half-assed knowledge of policy and law, so I can only speak as a technologist. There's more in common between the commons and property rights than differences. We should experiment with the two models — but there's almost no research capability left in this country, all the field labs are gone. The only labs we have left are at universities, which can help, but can't do it all. We could always have MSFT buy the whole thing, and then I could tesitfy again against them. We need REAL experiments, not experiments on paper, and then more discussion. Without knowing what works in the field, there's no basis for dialog. When we designed the Internet, we didn't design security, so it's impossible to control things like DoS attacks — it's dangerous to ignore security in open wireless nets.

Yochai Benkler: The Internet functions on a commons model in exactly the way that detractors of the commons have criticized. In that framework, over a decade, more capacity has been added — we're being asked to adopt a property model that the market has already rejected. Property can't exist without a commons: propertizing the roads and the sidewalks would make owning houses untenable.

You may need to regulate to have a commons network, but the same is true of a property network. Investment will come from the machine makers — it's credible that they will make investments as large as those who would make networks.

We can get investement in experimentation by creating a commons that's credible for a long period (a decade is forever for a machine-maker) — that's all the stability an investor needs.

Farber: Yes, but we need research before we plunge in — this isn't simple stuff.


Audience member: In a property model, why would I invest in a technology without some guarantee that I'll have access to the spectrum to use it in? What if the pricing of the spectrum turns out to be in excess of what I can afford.

Ed "Chief Engineer of the FCC" Thomas: The FCC has already sold off all the good spectrum. We put out a notice of inquiry offering to dedicate a huge swatch of high-frequency spectrum to the commons, and no one came forward to advocate for that.

Benkler: I don't want a command-and-control system: the commons model doesn't entail command-and-control. Spectrum-as-commons only requires a minimal set of etiquette and no more. I'm the only one here who really believes in technology — how did we choose Ethernet over token ring? Let people build the technology and see what happens!

Bob "VisiCalc" "Connectivity" Frankston: How do we allow experminention in low-value applications that can grow up? We can connect local experiments with wire networks, and these are very low-value and low-risk! Individual efforts make a difference. How do you give individuals the power to make a difference in a property model? You don't create a horseless carriage by mechanizing the horse.

Harold Dempsest: A member of the commons has a private right to control his equipment — in a property model, you have the private right to control some spectrum. We could start off with a 100% property system or a 100% commons system: which of these two is more likely to move expeditiously toowards the correct mix of commons and non-commons. There's a fundamental problem in moving from 100% commons, since you've got all these free-rider problems. A 100% property system will move towards a commons as people buy swaths of spectrum to use as a commons. So, how do we start the ball rolling to get to the right place.

Ed Thomas: 802.11 works — it wasn't dictated, it wasn't a policy, it grew because it works. There's a ton of spectrum available for low-powered experimentation (<1 Watt), on the same rules as the 2.4GHz WiFi band -- you can't interfere with others and you must accept all interference.

Audience member: We need a 21st Century definition of the public interest. Spectrum policy today undermines speech and the market. The 21st Century definition should define how spectrum serves the public.

Dave "Radio Userland" Winer: Technology is a constant process of building on things that drive adoption: SGML, HTML, XML. 802.11b is on fire — I want it in every device I own.

Audience member: We know how to get from private property to commons — every all-you-can-eat restaurant is a private commons; when Disneyland switched from ticket-books to all-you-can-ride, they created a commons. I don't know how to convert a commons to private property.

Lessig: Couldn't the FCC convert a commons to property?

Audience member: Packets in the air aren't like packets in a wire. On an oversusbscribed T1 that accepted by packet, someone else's packet wouldn't get through. I've seen no evidence that in a wireless world, packets will behave this way.

Duncan "Sky Pilot" Davidson: We're using WiFi and multihop mesh nets as well as spatial diversity. I raised $30MM in venture money. What's the FCC doing to help or hurt this phenomenon? We're trying to compete with DSL — at high frequency we can't get through trees, so it's no use to us. Give us access to the unused TV spectrum!

(Ed: I just got an Interesting People mailing list post that Farber very unobtrusively sent from his pager down on the stage)