Ostensibly the purpose of licencing spectrum use is because bandwidth is a fixed natural resource, and if we just let people use it for whatever they wanted, the bandwidth would be squandered and used inefficiently. In the case of wireless voice and data, it would mean that either providers would not be able to make any service guarentees because you never know what other people might be doing with the spectrum, or providers would find ways of holding on to bandwidth to meet service guarentees, which is just as bad as the current situation except it would then be first-come, first-serve instead of having a market-based control mechanism.

I don't believe that we should completely abandon the concept of licencing bandwidth for a fee, but instead look at the way in which we licence bandwidth. The current model made sense when the only commercial use for the spectrum was broadcast radio, television, or fixed frequency two-way radio. Licencees are assigned fixed bands of spectrum for periods of several years at a time. This makes sense for traditional radio and TV, because otherwise broadcasters would be constantly changing channels. Traditional broadcasters need a fixed presence in order to effectively provide their service.

However, this is not how cell phones and newer wireless modems work. Cell towers controls a set of data channels, which are a subset of the licenced band. When a phone wants to make a call or transmit data it asks the tower, which assigns it a channel to use. It's a bit more complicated than that, but the point is that cellular-based services do not require the use of a fixed channel in order to function, and mobile stations (phones) are only assigned channels as needed. So if cellular providers don't use all their channels all the time, why are we licencing large bands of spectrum for long periods of time?

As mentioned earlier, the obvious answer is reliability. If access to a minimum number of channels can't be guarenteed, it would be extremely difficult to make reasonable service guarentees. However, it should be possible to create an automated system where providers can licence individual channels for very short periods of time. A sort of micro-licencing system if you will. Under such a regieme, licences could range anywhere from free to extremely expensive, depending on factors such as demand, channel size, broadcast area, and how much of the available spectrum you're already using. During peak usage, the most popular providers will enter an automated bidding process to ensure they have enough channels. When almost no one is using the band at all, even licenced hobbiests will be able to access the spectrum for whatever they want.

Through these means, Canada would be able to automatically collect the true financial value of the spectrum at all times, new-comers won't be locked-out of the market until the next spectrum auction comes around, the wireless oligopoly could be broken, and the spectrum would be used as efficiently as technology allows. Of course, the technology for this sort of micro-licenced use of spectrum doesn't exist yet as far as I know, but if a trial of even a small band of prime spectrum was attempted, I'm sure you would find takers right away. This may not have been what Nadir Mohamed, CEO of Rogers Communications, meant when he suggested that usage-based billing was the way of the future, but he may have been quite right when he said it.