Eli Noam: The Fourth Way for spectrum
By Eli M. Noam
Published: May 29 2003 19:56 | Last Updated: May 29 2003 19:56
Richard Epstein - The first, best way for the spectrum: privatisation
Thomas W. Hazlett - Building blocks for spectrum efficiency
In a previous column - "The Third Way for spectrum", March 13 2003 - I argued against the pure property rights system for the radio-magnetic frequencies that is advocated by some reformers, and also against the free-access “commons” approach advanced by others. My three fellow columnists - Richard Epstein, Thomas Hazlett, and Lawrence Lessig, all prominent and respected proponents of the two rival perspectives - disagreed learnedly with me and even more with each other.
The areas of commonality, however, permit a synthesis. No, they require it. This is not the subject of idle academic speculation, but one of collective head-scratching on both sides of the Atlantic and Pacific. All of us agree that an overhaul of the present system of allocating spectrum is a key task for the new economy. But what should take its place? As it turns out, the three approaches we advanced can coexist.
The Third Way for spectrum
Should the electromagnetic spectrum be auctioned off to the highest bidders or treated as common property, free for all to use? Eli Noam, Lawrence Lessig, Thomas Hazlett and Richard Epstein offer conflicting views. Go there
Let’s start by recognising that most frequencies are used only intermittently by those who hold licences, whether private or governmental. Other frequencies are used intensely in some places but are quiet elsewhere, as any twisting of the radio dial demonstrates. Yet one could use these frequencies for low-power, low-range usage of, say, cellular phones. This would suggest that regulatory restrictions are in the way, and that much more multiple-usage of frequencies should take place.
A reformed spectrum system would therefore include the following rule. 1. The holders of frequency licences can use their spectrum for all purposes, and lease it to others. In some cases, a payment or auction might be required first. This should satisfy the property rights proponents.
This arrangement works best, however, where frequencies and users are clearly matched and transaction costs are low. It is similar to landowners charging an admission fee to their property. But it breaks down for aircraft flying over the land. It is simply not practical for thousands of aircraft owners to negotiate with millions of landowners for this right. Therefore, we do not let property rights to stand in the way of overflights.
It is similar with information streams transmitted over the air. In the past, the limited state of technology required that an information stream would occupy a specific frequency, and be protected from interfering use of the same frequency by others. But with today’s technology, the information can be transmitted in little chunks over many slivers of frequencies and reassembled at the receiver end. This opens up entirely new ways of using spectrum. Instead of dedicating well-defined frequency bands to specific usage types, such as taxicab dispatch or walkie-talkies, and slicing up the bands to specific users, we can now program smart transmitters known as “software defined radios” to seek frequencies that are unused at that moment, transmit for a short period on them, and then move to another unoccupied frequency. Hence, the second requirement of the system establishes a “right of way” access to frequencies that lie fallow at certain times or locations. 2. Unused spectrum can be utilised without a licence, subject to limitations of transmission power and time. This would correspond to the “commons” approach, in which open access exists and where technology expands supply, at least for a while.
In technical terms this assures that the spectrum is used in a fairly balanced way, shifting traffic from congested frequency bands to others, and squeezing vastly more usage out of it.
New economy policy forum
Archive of recent articles and explanation of how the forum works
Go there
This leaves, however, the problems of how to compensate the legacy licence holder for the use of the right of way, and how to discourage a wasteful use of frequencies if access is open and there is no price to pay for it. Nothing would then stop every teenager from starting his or her own multi-channel TV transmitter (unless one would ration usage). The better alternative would be to add a further element to the system: 3. Unlicensed users must pay a usage fee. This fee, set by a spectrum agency – which would be needed under any arrangement to police spectrum abuse - would be periodically adjusted to the demand. If demand is low relative to spectrum availability the price would be zero, thus creating the free “commons”, at least temporarily, in the growth phases of new spectrum applications, when they should be as unburdened as possible.
How might this fee be collected in practice? Low-power devices would not be included. But for those radiating a stronger signal, usage could be metered by a chip in the transmission equipment, and monthly totals transmitted to sites run by credit card companies, which would take care of the payment. The allocation of the collected revenues would be set according to the sampled usage of different frequency bands and distributed to the licence holders. Thus, private licence holders who do not fully utilise their frequency would benefit, as would governmental agencies that use their frequencies only intermittently but do not want to give them up in case of emergencies.
Nothing in this system would preclude a private or governmental licensee or property holder from creating a similar proprietary collection system, or a directly negotiated arrangement. But such a system is likely to cover only a few frequencies. The spectrum “right of way” system is the default after the privately negotiated arrangements. A good comparison, made by Charlie Firestone of the Aspen Institute, is that of copyright holders in music. They can negotiate for royalties directly with each radio station and piano lounge, or be part of a blanket licence set by a collective performing rights organisation, or have the government set a compulsory licence payment for play over the internet.
Which system will prevail? I suspect they all will. There are too many uses and circumstances for one size to fit all. We do not start with a clean slate. The extensive needs of government for spectrum use will always require a mixed system. Nor is there a need for uniformity, any more than exists for land use. Nor is it clear that government has the right to prevent the use of unused spectrum if this does not interfere with other users. The system of spectrum rights of way, property rights and usage charges would combine the best of both worlds and allow evolution in response to future technological and business developments.
Hopefully, our debate can reach further convergence to a mixed system. One can only hope that purism on either side will not prevail and help extend the present anachronistic system.
The writer is professor of economics and finance at Columbia University and director of its Columbia Institute for Tele-Information
...........................................................................................................................................
Richard Epstein - The first, best way for the spectrum: privatisation
We should all be grateful to Eli Noam for continuing his earlier dialogue on the structure of property rights in the electromagnetic spectrum. I think that his new proposal is a clear improvement over his earlier suggestions for the free-floating use of parts of the spectrum but on balance I think that the property rights proposal contains within it enough flexibility to answer his basic concerns.
It is true that the level of use varies with different components of the spectrum and that regulatory restrictions are largely to blame for the resulting misallocations. It follows that Prof Noam is right when he says current holders should be treated as the owners of property rights, with the options to shift use patterns and to engage in sales, leases or subdivisions without limitation so long as the boundary conditions with respect to other spectrum holders are respected. Indeed, under the property rights model there is no need for any central authority to decide which portions of the spectrum count as underused and which do not. The owners have every incentive to figure out how to get the last ounce of use from their slice of the spectrum.
That pattern of use, moreover, could easily take into account the new technologies that allow information to be transferred in little slivers over multiple frequencies before they are reunited by their recipients.
Prof Noam thinks that these smart elements pose a problem to the system of property rights, not dissimilar to that found with the rise of the aeroplane, which would have been grounded from day one if individual landowners had been able to assert exclusive rights of ownership to the upper airspaces. But the spectrum does not present that situation. The airlines in question could have found no path through the skies under the traditional property rights regime, either in the age of the biplane or that of the jet aircraft.
But there are no multiple blockades that arise in the current configuration. The holder of any block of spectrum need not create leases for specific frequencies but could create systems of "fill-in" rights for the entire band that exploit the advantages of the "software defined radios" of which Prof Noam speaks. Indeed the system allows for two improvements over the general commons.
First, subject to antitrust limitation, two or more spectrum holders could combine their holdings as joint lessors to open up wider opportunities for this new technology. Second, the lessor or lessors could prioritise these rights, enabling high-demand users to pay higher fees for more secure access to the spectrum if this were economically desirable and technically feasible. There is, in short, no high transaction cost story that bears even a passable resemblance to the overflight case.
Similarly, there is no reason to adopt a position that allows unused spectrum to be adopted without a licence, because there is no reason to leave any portion of the spectrum unused. Private owners could have the full spectrum and then balance the traffic by ordinary pricing without the need for the state to set up any system of access rules (of which first-come-first-served could be inefficient).
It also avoids the question of how to compensate the legacy licence-holders, which should be done, if at all, for reasons that do not relate to the new technology. We could therefore just make the licences perpetual without any charge; or we could charge some royalty for revenues obtained no matter what the stream of use; or we could auction them off, with or without some credit to the existing holder against future bids. But the transition issues should not blind us to the truth that there is no real advantage in leaving any portion of the spectrum unowned.
If we privatise in a sensible way, we do not have to figure out collectively what usage fee to charge these unlicensed users, because that set of users is again empty. Prof Noam notes the comparison with pools for licences for copyrighted music and the like; but that is just the point. Those pools were not created by government but by private holders of rights using contractual devices.
The contractual sophistication is there; the hold-out problems do not exist if there are fluid technologies and multiple owners. We have no more reason to fear the full ownership of the spectrum than we have to fret that there are no unowned plots of land on the most exclusive shopping street in Beverly Hills. Get the right set of private property rights and the conundrums of the commons can, in this case at least, be avoided by sensible contracts that exploit the possibilities of sophisticated technology.
...........................................................................................................................................
Thomas W. Hazlett - Building blocks for spectrum efficiency
This synthesis is provocative and may prove highly productive. If so, I declare Prof. Noam eligible to receive a thick stack of Spectrum Coupons. Let us now assure such script prove valuable, both for Eli’s sake and the consuming public’s.
As I read, Eli’s 3-pronged approach:
1. Liberalizes spectrum use for current licenses, giving them free reign in the radio space allocated for the limited uses now permitted.
2. Allows low-power uses of spectrum without fees or rules (other than power limits).
3. Allows non-interfering high power use of licensed bandwidth (see No. 1), upon payment of a fee set by the government.
The first of these measures would deliver spectrum markets, pushing valuable wireless services to customers. The second is already official government policy, both with respect to unlicensed bands set aside for various low-powered devices and ultra-wide band technologies using extremely limited wattage beneath wireless services up and down the dial. The third is not a practical policy and will stymie efforts for serious pro-consumer breakthroughs.
Eli is right to note that conflicts in unlicensed spectrum use arise and that a rationing device will help avert tragedy of the commons. Yet the proper tools are needed. Today, the government regulates unlicensed use, but over-consumption develops in some bands (say, 900 MHz), with virtually no use in others (say, the barren PCS unlicensed band). Similarly, bureaucratic access fees will be either too high, resulting in under-use, or too low, with over-use. Wasted spectrum either way.
Adjusting prices to find the efficient level, as Prof. Noam suggests, is a marvellous idea –exactly why private owners are required. With incentives to maximize value, spectrum owners search for the delicate balance between use and conservation. This involves far more than updating prices as demands are revealed, but investing in value-yielding innovations. It creates mechanisms to capture efficiency gains both through spectrum sharing and via the creation of co-ordinated national networks.
The expensive and complex shifting of millions of analog cell phone users (stuck in an obsolete technology mandated by the FCC) to advanced digital systems was achieved smoothly and seamlessly for customers – no chaos, policy paralysis, or more spectrum. Wireless telephone operators, with license flexibility yielding the closest thing to private frequency rights in the U.S., carefully orchestrated the transition of airwaves, handsets, and new equipment. Cellular systems depend on sharing and re-using bandwidth, creating reliable wide area networks, all in a decentralized, competitive, property model.
But only tiny parcels are available to the market – just 6 per cent of prime frequencies (under 3 GHz) according to a recent FCC paper. Eli’s proposals fail to unleash large blocks of bandwidth into the hands of private owners. This is the policy reform that will trigger a process of creative destruction in wireless. Once property rules are in place, markets will function. Additional non-interfering uses over, under, around, or through these spectrum properties will then prove no more difficult to affect than a rule deciding that airplanes need not compensate real estate owners for incidental use of the sky above. |
