Moot Court: Resolved: A True Coasian Would Favor A Spectrum Commons.
Michael "Chairman of the FCC" Powell couldn't make it this today. The judges are a Ninth Circuit Judge, a Nobel prize winner in economics, and a prof of Business Economics.
Lessig and Benkler are on the pro side, and Hazlett and Faulhaber are on the con.
Lessig: (I have chills — feels like being back at the Supreme Court!) Coase wrote a remarkable article, written in a time when the government actually chose which speakers could speak and how. He called it a clash of the doctrine of the freedom of the press. The allocation of spectrum is just like the allocation of any other rivalrous resource — it should therefore be allocated according to a property system like any other market.
Kozinski: Property system, not a commons?
Lessig: That's right.
Kozinski: You're sure you're not arguing the wrong side.
Lessig: I'm sure. Coase's arguments reflected the state of the art at the time. Property was the best way to allocate spectrum in 1959. But it's the wrong answer today. Not because property does no good — in fact, it does a great deal of good. This should not be taken to imply that administrative allocations are inevitably worse — a market has costs, and if those costs exceed the value, then markets result in misallocation. Coase's insight — most prescient — is that spectrum is not in its nature rivalrous. It's not a thing at all. Colors, sounds correspond to frequency.
Demsetz: Why does that make it not a thing?
Lessig: Because it wouldn't make sense to allocate property rights in colors or notes. There are two kinds of costs which property creates here.
Demsetz: Why are we led to a better understanding of the problem by saying that spectrum is not a thing?
Lessig: You can talk about it as though spectrum was a thing, but by talking about it like a thing, you naturally move to the conclusion that it should be governed in a property model, but that's contingent if the value of property rights exceeds their cost.
Kozinksi: In common law we were deemed to own property above and below, but that no longer held once airplanes came about — you can't stop a 747 from flying over your house at 35,000'. Is your house any more of a thing than spectrum? A house is nothing at all — a property right in a house defines my rights in respect of other people. It's the right to stop people from entering my house, destroying my house.
Lessig: Correct. The proper legal structure for regulation shouldn't turn on the physical nature of the thing you're regulating. Coase believed that considering spectrum as a property caused a short-circuit in thinking.
Kozinski: Aren't we just debating the scope of property right? Not property rights themselves?
Lessig: We're asking first whether there should be a property or commons right or some mix in spectrum, and second, what the scope of those rights should be. Two technical facts have radically changed the analysis. These two additional technical facts should be considered to determine whether Coase would still advocate a property right in their light.
Demsetz: Something has changed the cost-benefit calculus of property rights in spectrum. Coase's views on whether spectrum is a thing isn't furthering your argument.
Smith: What are the actions that people can take with this thing or non-thing?
Lessig: The question is first: what are we trying to achieve. It's not to minimize interference, it's to maximize the output. There are two factors that have changed in respect of that maximum. There's a presumption that carving up spectrum doesn't reduce its value. But the opportunity to have a wide bandwidth of spectrum creates value that is destroyed when it is carved up — just like a racehorse is only value so long as you don't carve it into 35 piece, so too spectrum ahs value that is only present when the spectrum is intact. If carving spectrum into property has an externality into the value as a whole, then the property system could not be as valuable as its cost.
There is the reduction in value in the carving up of the thing and a transaction cost that is insurmountable in reassembling it.
The burdens of property are twofold: On the Internet you have facilitated coordination, but not facilitated allocation — QoS mechanisms are complex, burdensome and cumbersome and not worth using because their costs outweigh their benefits. Secondly, spectrum as a resource was a good analogy in 1959, but today, there are a number of dimension that get traded off dynamically, as protocols, frequencies and so on change. When you introduce property, the frequency allocation from moment-to-moment is burdened by a transaction that may or may not be important in respect of the technically optimal allocation.
Demsetz: Change is costly. Coase would have told the FCC that if he'd asked me. Would he have still argued for the superiority of taking away the power to handle change away from the FCC>
Lessig: Absolutely. Everyone agrees that the FCC is the wrong group to manage change.
Kozinski: Someone needs to enforce limits. Who? And what limits?
Lessig: Think of a highway. It functions in this sense as a commons. In many highways, this is not a problem. The regulation that governs it regulates the type of devices you can bring onto the highway. No tanks, no go-carts, and no going over 80mph. The rules need to be as minimal as possible, but no regulation about where you're going or what route you take.
Kozinski: But you have to widen highways from time to time. How do you widen the spectrum highway. Is there really no limit to the width of this highway?
Lessig: No — but given the uncertain state of the future of spectrum (Ed: lost it). Coase in 1959 concluded that the problem was that signals interfere. This problem can be solved by delimiting rights. He's right that experience should guide whether one model triumphs over another. But the insight is that the delimiting of the rights can be accomplished by the market or by tech, each of which tries to allow maximum communication. The failure to examine the tech leads people to a simple analysis that is unfaithful to Coase.
Faulhaber: Coase saw a problem and proposed a solution. In its simplest components, the problems were: a bureaucratic and political allocation process and exclusive use licenses. This was hugely inefficient. It substituted the judgement of bureaucrats for the judgement of the markets.
Demsetz: Do we use regulators elsewhere?
Faulhaber: But rarely to allocate resources.
Kozinksi: Big chunks of commons (sidewalks) make private property valuable.
Faulhaber: The central planning model made the FCC the US Spectral GOSPLAN. He proposed a market allocation, but with exclusive use licenses.
Demsetz: Would Coase have felt this way if someone could conclusively demonstrate that there was a way to use spectrum without crowding?
Faulhaber: If you could wave a wand and make it nonscarce, then markets lose their power.
Demsetz: If you get the FCC out of the picture would there still be scarcity?
Kozinski: How about the Internet? Or cable? The less the FCC was involved, the more channels we got. I had channels coming out of the wazoo.
Faulhaber: Many commons advocates use the Internet to bolster their case — a voluntary org, the IETF that sets standards and we abide by them.
Kozinski: It's more than that. It started as a commons. If we'd been present at the net's birthing, we probably would have believed that in a decade, it would be clogged beyond use. But as far as I can tell, it's getting bigger faster! Sometimes capacity creates demand which creates capacity.
Faulhaber: Re. capacity. Capacity increased to overcome the World Wide Wait. Capacity is a priced item — it's not free. Every connection to the net costs, by the month or what have you (Ed: what about interchanges?).
Kozinski: So, you have an unregulated resource that is capable of overuse that solves the problem by charging users for its use. This cuts the other way. The other side proposes to turn spectrum into a wireless Ethernet, as we did with 2.4GHz, and the very availability of free space will give people the incentive to come in and create.
Faulhaber: I'm not arguing against any commons. There's a place for commons, within an overarching property rights regime. But there is a problem that's solved in the Internet — interference — that's not solved in spectrum. Likewise protocols; if my protocols are noncompliant, they don't interfere, they are ignored. But in spectrum it may be advantageous to step on my competitors with noncompliant protocols.
Kozinski: Couldn't you have noninterference regulation with everything else being free?
Faulhaber: So is it more efficient to solve interference through property or regulators?
Demsetz: That question must be faced in two contexts: one, if there is no scarcity, no congestion, and the other, in which there is congestion. I'm not sure that the two sides in this dispute have focused their arguments on these two contexts.
Faulhaber: If there is congestion, if spectrum can fill up (at least in the long run), or if you're spending a lot of money to avoid congestion (then there's an opportunity cost in the use of the spectrum), even then there's room for a commons. The noninterfering easement will permit the commons to operate in a way that elides some of the transaction costs that Lessig referred to.
If it turns out that there's more spectrum than we can ever use, then the marginal cost in a property world would fall near to zero.
Smith: I'm not sure I've heard anyone say anything yet that would indicate that they know what they're talking about. In 1959, Coase wrote about an interference problem. Has there been a change in the technology that impacts that, and how do these two sides differ on that?
Faulhaber: UWB, agile radio and other dynamic allocation technologies that bounce across the exclusive use boundary can make use of fallow spectrum.
Smith: How does that differ from having n long lines and a switching system that can put calls on lines depending on which one is free. You still have property rights in that system — a right to act. Property rights are not separable from an action. What actions are effected by technology and property rights?
Faulhaber: If we have agile phones, I can do things that I couldn't do in 1959 — when I would have needed a single channel for my sole use. But today, we have (illegal) technology that can use others' spectrum in an opportunistic way without interfering with them. Some uses still (and will continue) to require exclusive use. The noninterference easement allows you to use my spectrum so long as you don't interfere with me.
Demsetz: The issue seems to be that if you auction off spectrum rights, that this will raise the cost of noninterfering uses. Do you believe that if there was no congestion, that auctioning off frequencies would hamper the ability to apply the commons technique.
Faulhaber: No — because in that situation the value would fall to zero, so the cost of enforcement would exceed the benefit you'd get from the enforcement, so you'd get a commons anyway.
Benkler: The property right created by the commons is the property right to use my equipment in ways that allow me to communicate. In order to implement a property regime defined by frequencies, then most radios have to be prohibited. A series of parameters about what can and can't be done, who can radiate with what equipment, etc needs to be transacted. In 1959, it was too costly to employ techniques for differentiating between signals other than used division based on code, but today, we can do many other kinds of division. Frequency is no longer the thing that prevents 2 or 3 or 1MM people from communicating with each other without interfering with each other.
Smith: This is like a system where lines are so plentiful that any call can be completed.
Benkler: Even more than that. Many techniques are possible if you use small parts of a million lines that each used to have to be a circuit. There are many dimensions other than frequency on which we can (Ed: lost it).
Kozinski: Is spectrum unlimited?
Benkler: We don't know, but there's only one good model for scaling capacity with demand without diminishing capacity, and that only applies to latency insensitive applications. The claim isn't that there's no limit, but that capacity scales better if the devices aren't encumbered in how they can reconfigure to adapt to their environment — i.e., UWB.
Demsetz: Are there advantages to exclusive use at high power?
Benkler: Yes, but that's an inefficient method for communicating. This isn't a freebie, but there's a market in the devices that drives investment in them.
Demsetz: Wouldn't the market forces drive spectrum property owners to merge their spectrum to accomplish this?
Benkler: Would a system based on rights to use devices subject to some etiquette improve welfare? In that case, we need to ask about transaction costs. You need to discover a lot about your neighbors and transact a lot of business with them. The question is, are these transactions easier in a property or a commons regime.
There is a market system in devices and services.
Demsetz: How would the commons respond to the need to have exclusive, high-powered use of a band?
Kozinski: What happens when someone wants to drive an ambulance down your freeway?
Benkler: We can build protocols that recognize and respond to emergencies. In the case of enstating a permanent property right, it would have to be accomplished through regulation, but it would be possible.
Demsetz: Would the converse be possible? Could a property model allow for a commons?
Benkler: Easements would help. But if UWB required 10GHz for perfect efficiency, you'd have to negotiate with 10,000 individuals to acquire the spectral rights you require — some of whom would have businesses that were threatened by UWB.
Hazlett: The case is overwhelming that property rights are much more important today than they were in 1959.
Kozinski: Aren't property rights really grabby? Property owners are litigious and intolerant of uses that don't really impact them or their use of their property.
Hazlett: Nobody is talking about an actual commons — free entry without government regulation. People are talking about power limits and other policing because there's a congestion problem today just as there was in 1959.
(…bathroom break…)
The property system allows flexible use in all blocks — high AND low powered.
Kozinski: But Stanford has no property right in WiFi and yet here we have a perfectly good network, even without that right.
Hazlett: Yes, they do have a property right.
Kozinski: If you call that a property right, then you've joined the other side.
Demsetz: Stanford can't sell access to the spectrum, it can't prosecute people who radiate over its property line. It's not a property right.
Demsetz: Does this have to be accomplished by auction?
Hazlett: Not at all. Coase thought that was simplest, but he was wrong. Priority of use would've been simpler.
Smith: You want to allow anyone to do anything that doesn't cause uncompensated interference.
Hazlett: That's right.
