Yet more MSM NBN politicking

It's been a while since I had a chance to unpick egregious NBN commentary – not that the stories have gotten any better, but because I've been extremely busy!

But this piece at Fairfax is too much for me to let pass:

Don't look now, the white elephants are multiplying

Let's just pick off some hit points.

“The government's plan to roll out high-speed fibre-optic telephone cable to the home is now estimated by most experts to cost $90 billion to $110 billion, compared with the government's latest guesstimate of $45 billion.”

Fibre-optic telephone cable? That should set the tone.

And “most experts” don't estimate $90 to $110 billion – that, dear friends, is the Coalition's number. I don't know why the columnist, Kenneth Davidson, decided not to source the number correctly.

The government's $45 billion is the whole network, not just the fibre network – it includes the satellite services and fixed wireless services to the last 7% of the population. NBN Co's calculation for the fibre is currently in the order of $37 billion.

“There are about 11 million fixed lines in Australia. Fibre has a life of about 25 years. In order to fund the capital cost and depreciation of the network, every fixed-line user would have to pay $9400 a year to the NBN, based on a return of 10 per cent a year.”

“Fibre has a life of 25 years”? Oh dear. Mr Davidson has mistaken the depreciation life of the fibre for its operational life. The depreciation life – which gets revised upwards from time to time – is a mere financial fiction. The operational life of the cable could easily reach 60 years; as I've reported before, Corning has tested fibres which, after twenty years of exposure to mud, flood, heat and snow, showed no observable deterioration.

“By 2020, the US will probably have developed low earth orbiting satellites capable of picking up and sending wireless signals at a fraction of the cost of the Gillard government's scheme.”

Wow: let's party like it's 1998, when I was being instructed by boosters to get excited about Iridium, the planned constellation of low-Earth-orbit satellites that would usher in a new world of communications, etcetera.

Iridium got launched – but not until after a bankruptcy, for a host of different reasons, and a subsequent restructure that meant some of the planned constellation was abandoned. As it happens, the late 1990s saw this happen more than once, with Teledesic, Orbcomm, ICO Global and Globalstar all suffering a similar fate.

Some of the constellations are there – but you might notice that we're not all using satellite communications in preference to our land-based services.

I would add that I'd just as soon have my services delivered by an Australian company, and pretend futilely that at least some of my communications aren't subject to the Patriot Act.

When so many errors are fashioned into a stick to beat the government, it's hard to believe that the opinion was nothing more than politicking. The problem is, few readers are in a position to pick up the errors.

Mr Davidson Responds 

Update: I'm told Mr Davidson has responded to criticisms, and his response is posted here at Dropbox.

Note that I can't vouch for the veracity of the response. If it's a hoax, I have no way of knowing.

So here's the nub of Mr Davidson's case:

The major difference is between my statement that the capital cost of ftth will be in the order of $90 billion (the Turnbull estimate) rather than the $46 billion (the official NBN estimate) 

The difference is explained by the fact that my NBN critics claim that the ‘least average cost’ of the underground tail connection from the NBN pit to the customer is $4,000 dollars. The highest average cost is $12,000. Based on 11 million connections the total cost of the tail connections ranges from $44 billion to $132 Billion dollars

I don't exactly know where Mr Davidson sources his data but I suspect it's a mangling of data presented on page 207 of the NBN Implementation study. Yes, the peak price does reach between $11,000 and $12,000 per premises.

The $11,000-$12,000 cost per premises is not the cost across the whole rollout – that's the bit of the graph that's been misunderstood here. The graph says that for the easy premises, in metro, near the exchange and so on, the connection cost is low, whereas for someone in a low density area, at the edge of the network, in a country town, the cost is high.

It's quite clearly described in the Implementation Study as a “marginal cost curve”. The $12,000 cost of connecting Joe Bloggs in Dubbo does not result in a $12,000 cost of connecting Jenny Basketweaver in Balmain.