Are auctions the best way to allocate mobile spectrum?

As the communications regulator prepares to reallocate over 300 MHz of “prime” radio spectrum, David Havyatt asks if it is time to review the management of spectrum for mobile services.

Following the Federal election, we have been told we face a new political “paradigm”, borrowing the term Thomas Kuhn used to refer to periods of normal science between revolutionary upheavals.   

Fifty years ago, J.K. Galbraith anticipated Kuhn by a few years when hecoined the phrase “Conventional Wisdom” to refer to “ideas which are esteemed at any time for their acceptability”. 

In doing so he noted that in social comment “the test of audience approval, far more than the test of truth, comes to influence comment.”

In telecommunications policy, today's 'conventional wisdom' is an unqualified appeal to competition and reliance on markets.

The Federal Government's policy response to managing the radiocommunications spectrum is a good example.  The first object of the legislation is to maximise the public benefit by ensuring the efficient allocation and use of spectrum.

Globally, spectrum management came from an early recognition of the military potential of radio and the national defence interest in managing interference. 

With the advent of public broadcasting, the management of interference then became a commercial issue, and there was a consequent decision to regulate the content of broadcasting services in their use of the public asset of spectrum.

The move to recognise spectrum as a property right rather than a Government resource was a product of the sixties. Ronald Coase in 1959 wrote an essay that in 1960 morphed into his famous theory of externality (the Coase Theorem).  In 1964 Any Rand, a particularly virulent breed of libertarian, railed against government control of spectrum and how treating it as a property right would remove the basis for government to control broadcast content.

From these two seeds a global movement argued that spectrum allocation should be market based, primarily on the grounds of efficiency.  It should be put to its “highest and best use” by those prepared to pay for it.

Simultaneously the market for telecommunications services globalised. Spectrum planning extended from being merely about the management of interference to being about achieving benefits from scale of production and “roaming” by aligning the spectrum allocations in different jurisdictions. 

In Australia we have seen the ACMA conduct two recent “Spectrum Tune-Ups” over the issue of the reallocation of the 2.5 GHz band (2500 to 2690 MHz) and the planning of the so-called “Digital Dividend”.   

The eagerness to reallocate this spectrum stems from the growing popularity of mobile services. 

In the USA President Obama has recently asked the FCC to find yet another 500 MHz for wireless data. The ever expanding need of wireless broadband for new spectrum derives from the disjunction between Moore’s Law and Cooper’s Law.  

Moore's Law predicted that the capacity of microprocessors doubles every two years. But spplications have a tendency to utilise all this resource (under what I call Havyatt’s Law - “What Intel giveth, Microsoft taketh away".)

Cooper's Law states that the capacity of wireless networks doubles every two and a half years. 

To keep pace with the demands of applications, mobile data networks need more spectrum. 

The two frequency bands in question have been chosen by the ACMA since international forums have determined they are the appropriate bands for future mobile services. This means that there will be economies of scale (and experience) in production and the ability to roam. 

The ACMA plans to use its centralised planning powers over the 2.5 GHz band to “reallocate it”.  The Minister plans to use his powers over broadcasting spectrum to redesignate part of the broadcasting band (the dividend from compressing TV broadcasts as digital) for potential use by mobile applications.

At the end of this two step centralised planning exercise, globally and nationally, the final regulatory process says “now we will use an auction to allocate these licences because markets are efficient.”  

But are they? When we dive down to the technical level, the question must be asked.

There are four bands in use for mobile services in the 3GPP family (GSM and UMTS/HSPA): 850 MHz, 900 MHz, 1.8 GHz and 2.1 GHz.  As a consequence of earlier allocation processes, these are all divided between the operators (in only 850 is it only divided between two of them). Each of the operators using a band builds separate radio transmission equipment in each band. 

There are small engineering losses at the edge of each allocation. There are bigger economic losses as a consequence of what the theorists call “monopolistic competition”: the most successful provider needs to build new infrastructure to serve customers in areas where other operators still have unused capacity. 

I would argue then that the process of allocating each band across the whole market as it is released is inherently inefficient. (Let alone the environmental consideration that much fewer antennae would be required if operators were using bigger lumps of the bands than the ‘bits and pieces’ approach). 

But the conventional wisdom remains that each new allocation needs to be auctioned.

And here comes the most egregious part of the whole process. The ethos of competition policy included the sale of Government owned utilities. It did so because Governments are greedy rent extractors from monopolies they own, not because the utilities have excessive demands for capital.

But lurking in the background on spectrum auctions is the rent extracting part of Government. Their interest in auctions is the money they raise, the rent they can extract.  This is the exact opposite of the idea of efficiency that the “market” is supposed to deliver.

There are a host of other issues to be resolved, including properly resolving the issue of licence tenure. But it does seem pretty odd that a process of spectrum planning that works perfectly well in broadcasting and has been used to create the “digital dividend” can’t be applied to the analogous telecommunications services.

But to reach that conclusion you have to challenge the conventional wisdom.  Maybe that is possible in the new political paradigm.