Economists have always been suspicious of unlicensed radio spectrum. The idea just shouldn’t work. Unlicensed users share the underlying resource — spectrum — without paying for it. Theory says that every user will strive to maximize his own benefits by increasing his consumption until the resource as a whole is so over-burdened as to be useless. This is the so-called "tragedy of the commons."
The fact that unlicensed spectrum functions as well as it does, notwithstanding an ever-growing body of users, must be a constant irritation to the economists. Perhaps that is why they keep producing papers that propose economics-based regulatory schemes to avert the supposed tragedy.
The FCC released three such today, prepared by the same core group in its Office of Strategic Planning and Policy Analysis.
Two of the papers consider alternative "spectrum etiquettes" for granting access to unlicensed spectrum. In general, a spectrum etiquette is a set of procedures through which contending users can minimize interference. ("Listen-before-talk" and "Defer-to-emergency-traffic" are two simple examples.) These papers propose more complex versions that take into account both the user’s willingness to pay and his tolerance of incoming interference. In some versions, each user sends out this information for an "instant auction" that determines who gets access next. The authors claim that their models, appropriately tweaked, can double the usage of a given amount of spectrum, compared to present methods.
One of these papers also notes a perverse consequence: increasing unlicensed capacity will attract more users from licensed services, which tend to cost more, and hence will increase unlicensed congestion and interference.
The third paper looks at decisions on how to divide up spectrum between licensed and unlicensed uses. The wrong way, say the authors, is the way we do it now: interested parties communicate their views to the FCC, which does its best to accommodate everybody. This method suffers from a tendency of the parties to exaggerate how well they would use the spectrum. Working with certain-to-be-distorted information, the FCC makes less-than-optimal decisions. The remedy? Put your money where your mouth is. The authors propose a modified form of auction in which the interested parties bid to have various blocks of frequencies allocated respectively for licensed or unlicensed use.
There is an obvious problem in a scheme that entails bidding by unlicensed users: since the spectrum will eventually be available to all, each company can sit out the auction, let other unlicensed proponents do the bidding, and then step in afterwards to use the spectrum. According to the authors’ analysis, however, the need to outbid the advocates of licensed spectrum for at least some frequencies will overcome the inclination to take a free ride.
Papers like these have a science-fiction quality, at least for a non-economist like me. A science-fiction author assumes someone else will have solved the hard problems — faster-than-light drive, generation spaceships, whatever it is — and concentrates on the story whose characters use those accomplishments. In these papers, too, the authors bypass the pesky practicalities of how to make things work: implementing a spectrum etiquette that pools other users’ information; managing a chaotic and ever-changing group of unlicensed users to implement a common bidding strategy; and so forth.
I have no doubt that the results are mathematically valid, subject to their assumptions. And I accept the authors’ claims of successful tests, via people at computer terminals role-playing spectrum users. But what are we to do with the results? If this is all an academic exercise, very interesting, and let’s get back to work. But if the results are intended as a plan for action by the FCC, then a lot of very hard detail remains to be filled in.
