Competing for spectrum
I had an entertaining time at RaComms'09 last week.
I learnt that spectrum management is really very easy, at least according to some. Various presentations had a flavour of “spectrum management is easy, all you have to do is the following things for the telecommunications/radio/mobile/fixed link/satellite/astronomy/regional community”. The challenge being of course that the lists was in direct conflict.
Clearly a lot of the noise is about the future Digital Dividend (700MHz), with both TV and radio still interested as well as the mobile operators. But similar contentions arise between mobiles and news gathering at 2.5GHz, and between WAS operators and satellite operators in 3.6GHz. Even the discussion about 400MHz had disputes between various users and a big concern about "spectrum squatting" by Government users.
A couple of really interesting prersentations by the radio and television sectors on the prospects for further digital radio services and television developments to 3D and super HD. The funy part was that no one mentioned the NBN once and what the implication of fibre everywhere might be to the desirability of wireless based services. However the question is starting to get some coverage in the industry media.
But the biggest bit was the bid by the mobile operators colectively (as AMTA) and individually (as Telstra) trying to argue for both the Digital Dividend (700MHz) and the ENG (2.5GHz) spectrum for mobiles. The latter presentation was written up in both Communications Day and the AFR under the theme that "Australia is at great risk of lagging behind the rest of the world on LTE deployment."
The presentation by Telstra's Dr Tony Warren was funny, but on detailed analysis was also highly fanciful. What amounted to special pleading was masked by some impressive figures that did an excellent job of hiding the fact that this was a biased reconstruction of history and relied on dodgy forecasts.
The selective history came with an attempted comparison between European 3G auctions and those in Australia. The thesis advanced was that the Europeans got it wrong by creating an “artificial scarcity” by having four lots for six bidders whereas Australia got it right so we didn’t bankrupt the industry. It is nonsense because;
• there isn’t enough 2GHz spectrum to create six lots,
• the Australian experience of narrow lots has resulted in some odd allocations,
• the outrageous prices paid for 3G in the UK were actually preceded by the ridiculous prices paid in Australia for the 1.8GHz spectrum in the infamous “One.Tel auction”,
• and the number of bidders in Australia was reduced from five to four by Telecom NZ entering a JV with Hutchison.
As it transpired these four winners of the 2GHz spectrum co-operated to build only two physical networks.
We were entertained by a series of interesting forecasts. These included a reference to an assertion from Telstra’s own Hugh Bradlow that at some point in the future three-quarters of internet access would be by wireless devices. This begs the question of three-quarters of what? If it is just the number of devices I think he is seriously underestimating that – given the already high use of connections through WiFi in public and private spaces. If he means by downloaded data he is seriously mistaken because of the existing and projected price differences.
We were also entertained to some data from Ovum forecasting both connections and dollars of revenue for wireless data into the future. Over at Unwired they remember well the same sets of industry analysts claiming a few years ago that there was no market for wireless data. Dr Warren informed us the chart’s obvious decline in ARPU represented the effect of competition, whereas a more studious approach might conclude that competitive intensity is not increasing over that time, but the combined effects of increasing economies of scale and the experience curve effect in device cost resulted in price declines to attract more marginal consumers. To put it simply, the quantity demanded has increased (and price declined) not because the demand curve shifts up but because the supply curve shifts down.
We were also entertained with a slide that purported to be a projection of future spectrum requirements based on an ITU-R study. However this started with the assumption that the services in 2010 are being supported using the entire available spectrum, including the mobile (850MHz, 900MHz, 1.8GHz and 2GHz) and WAS (2.3GHz, 3.4GHz) spectrum. This ignores the fact that the WAS spectrum is currently lightly used, half the 850 MHz is unused (in metro) and at least two-fifths and probably more of the 1.8 GHz is unused.
But the funniest parts were the sources that Dr Warren relied upon for vindication of his views. The GSMA had written to the G20 leaders to argue the case for co-ordinated action on spectrum, and they were most proud about the endorsement they got from Gordon Brown. Given the state of the British economy this is like a telco relying on an endorsement from Jodee Rich! Later Dr Warren quoted approvingly from Lord Carter’s Digital Britain report, conveniently forgetting that he isn’t as full in his praise of Carter’s earlier work at Ofcom in structurally separating BT. (When challenged on this he did say Lord Carter has got some things right - I agree, but which bits).
But ultimately it is the twin claims that failure to move quickly (or more quickly than the existing plan of 2012) on reallocating the 2.5GHz and of imposing any competition limits will impede LTE deployment that need to b tested. The practical reality is that LTE can and will be deployed in many bands, not just 2.5GHz. Telstra itself is proud of its unusual use of 850MHz for W-CDMA. Lower frequencies like the unused 1.8GHz will actually get better returns.
The issue of competition limits is far wider than just the creation of artificial demand. A big issue with using the 2.5GHz is that out of the 190MHz there may only be 2*60MHz arrangement for LTE. Telstra is right, that is probably best sold as one lot. But should it be sold to the guy who already dominates every other access technology in the market?
The rush to allocate happened in the 2GHz band as well. Australia was about the third market to auction the spectrum. Despite that rush the first network deployment wasn’t till 2003. Despite that rush the leading 3G network actually operates at 850MHz.
A measured approach of making the reallocation together with the Digital Dividend and together with any re-allocation decisions of existing 15 year licences will achieve a far better outcome.
The meaning of words and pictures
Another bit of humour was the need to point out the disclaimer on Telstra’s 21Mbps speed claim because, according to Dr Warren, apparently Australian consumers are stupid. By this I think he means that the ACCC shouldn’t be constraining the Telstra speed claims because no consumer would believe that is the speed they get. I suppose this begs the question of why you would make the speed claim in the first place if consumers didn’t believe it.
It was also interesting to note that this was now a new disclaimer. The original said that typical speeds range from 550Kbps to 8Mbps. I believe the new disclaimer states that these will be typical speeds in regional areas, but 550K to 3Mbps is typical for metro. My apologies to Telstra if I didn’t get that right, but it was small print and I don’t have the best eyesight.
Later Dr Warren presented an impressive chart of 3D columns going through different iterations of technology and describing their peak speeds. He again drew our attention to the disclaimer. The fact that the columns were not drawn to scale belied the impressive gesture he had used to describe the burgeoning capacity. (Admittedly these columns would not have worked in a linear scale, but a logarithmic scale would have allowed them to be drawn to scale).
The conference provided both a very practical view of the challenges in spectrum management, but also a very instructive view of the ways various participants frame public policy discussion. This ranged from heavy focus on impacts on national income, to selling the sizzle of new services, through to the benefits of sheer outright perserverence.