Circular Logic of "Scarcity"
In November, the Federal Communications Commission (FCC) of the United States published the report written by the Spectrum Policy Task Force (SPTF). Summarizing a half year of of extensive research and discussion, the report is indeed impressive in its deep understanding of digital wireless technologies and its call for for bold reforms. Particularly noteworthy is the FCC's commitment to abandon the command and control approach that regulates usage by licensing. It is also remarkable that the FCC recognized the efficiency of the commons approach that share a wide band and the overlay usage of spectrum by different kinds of terminals.
However, the SPTF's conclusion is a half-hearted compromise of exclusive rights model and commons model. According to the exclusive rights model, only 384 kbps can be available using the expensive networks of 3rd-generation mobile phones, while the commons model enables wireless LAN to realize up to 108 Mbps by much cheaper wireless routers. Why should we divide spectrum rights into these two kinds when the commons is clearly more efficient than exclusive rights? SPTF claims that the exclusive rights are more efficient than commons where "scarcity is relatively high and transaction costs associated with market-based negotiation of access rights are relatively low" (p.38).
This justification is incomprehensible. Scarcity means shortage of spectrum capacity, which is severely restricted by the old model of exclusive rights. In fact, alleged scarcity is created by the poor spectrum management. It can be overcome by packet radio technologies that can supply much more capacity than demand if they are deployed in sufficient bandwidth. Thus it is a circular logic to justify exclusive rights by the scarcity that is created by exclusive rights.
Standard economics tells us that commodities should be supplied as private goods based on property rights when they are excludable (externality is small) and rival (marginal costs of use are high) in economist's jargon (cf. Stiglitz). A member of SPTF excused that, with the standard terminology, scarcity meant rivalry. Strange semantics. Moreover, rivalry is not the sufficient condition for private goods. If the spectrum is rival but non-excludable, it should be supplied as commons. A typical example of commons is the traffic on the road. The road is rival because traffic jam will arise if too many cars are on a road, but it is non-excludable because assigning property rights of the road to drivers is inefficient. Thus roads should be maintained as commons monitored by police.
Exclusive rights approach is similar to the policy to allot a lane for each car. It is inefficient to divide a band into many narrow bands owned by licensees because packet radio technologies can utilize the spectrum much more efficiently by sharing the whole band with many terminals. Oddly enough, SPTF rejects the old paradigm based only on frequency, saying "the development of frequency-agile technology has created the potential for development of services and uses that are not tied to specific frequency bands" (p.19). They prescribe power control, microcells, software-defined radio (SDR), smart antennas, and time-sharing technologies to enable various types of signals to co-locate in a band. Their excellent understanding of new technology is inconsistent with their poor justification of exclusive rights.
Furthermore, property rights are not necessary to control rivalry. Spread spectrum technology makes it possible for every packet to avoid interference identifying each other by its spread code. SDR and cognitive radio technologies can avoid interference by changing modulation systems or by time division. In fact, all new digital technologies are based on commons model; it is even possible to substitute ultra-wide band for radar and GPS. Therefore it is pointless to secure exclusive rights for newly allocated bands.
Confusion of Economic and Political Problems
I suspect that such strange argument about scarcity is a rhetoric to cover up the political problem - incumbent's vested interests. SPTF recommends that the bands below 5 GHz should be allocated as private goods because of "high level of incumbent use" (p.38). This proposal confuses (economic) efficiency and (political) difficulty. Indeed the resistance of incumbents will make the transition to more efficient usage difficult, but it has nothing to do with the efficiency determined by technology. Exclusive rights model is politically easy but inefficient, while commons model is politically difficult but efficient. The mission of SPTF must be the prescription of the efficient solution. Political solution should be left to the Congress.
Of course the strategy of transition to more efficient technologies is important. I agree with SPTF that spectrum policy must "provide incentives for users to migrate to more technologically innovative and economically efficient uses of spectrum" (p.15). However, their "market-oriented" policy is a two-edged sword because exclusive rights supply strong incentives but authorize the incumbents to exclude other party's more efficient usage. The "Big Bang auctions", proposed in a working paper written by economists of Office of Plans and Policy of the FCC, are likely to make things worse. If the spectrum were sold by high price, the "owner" of spectrum would maximize its value by monopolizing it. This is rational for individual users, but it will lead to socially inefficient outcomes. Even worse, such policy is irreversible; once spectrum was given away to incumbents, spectrum commons will be lost forever because incumbents will never open it.
Instead of such dangerous auctions, I propose spectrum buyouts by which the FCC takes back spectrum and opens it without license. Such reverse auctions can supply incentives for incumbents to sell their legacy bands and migrate to wireless Internet. This policy is reversible because it is easy to set property rights in commons. Moreover, reverse auctions will be very cheap because the winners will be the least efficient users, whose valuation of spectrum will be much lower than that of winners of ordinary spectrum auctions who are the most efficient users (see my discussion paper for more details).
SPTF rejects the wholesale conversion to commons model, because some technologies rely on high-power propagation (p.40). It is true that commons band and private band will coexist, but the threshold of 5 GHz is arbitrary. The criterion should be determined by the excludability (efficiency of exclusion) of a band. Above 3 GHz, commons model is absolutely superior to exclusive rights model, so all spectrum should be opened without license. Exclusion might be justified in the extremely lower band (probably below 30 MHz) where high-power propagation is economical, but in fact no digital radio technology is likely to be implemented in such low band. Therefore the easement of overlay usage should be enforced in the band below 3 GHz. Exclusive rights will make it difficult to enforce easement because the "owners" of spectrum will resist the easement which infringes their property rights. Command and control approach might be better to enforce easement.
Summing up, the SPTF report confuses technical, economic, and political problems. The condition on which legacy band and commons band coexist should be investigated again from scientific point of view. Scarcity cannot be the ground to justify exclusive rights. Roads, parks, and streetlights are supplied as commons while they are scarce. The whole argument of "spectrum rights" does not make sense because its fundamental assumption for exclusive rights is wrong. Before recommending spectrum policy reforms, economists at the FCC had better read undergraduate textbooks of public economics.
