Coalition's NBN analysis needs more analysis

The Coalition's background discussion paper on competing NBN policies raises important questions over the current policy, but overlooks some of its own problems.
Written by Josh Taylor, Contributor

Shadow Communications Minister Malcolm Turnbull has laid out a stinging analysis of potential situations that could lead to the current National Broadband Network (NBN) rollout costing up to AU$94 billion, but there are many assumptions in the pricing of the Coalition's own policy that go untested.

The background document, released with the Coalition's alternative AU$29.4 billion NBN policy last week, claims that in a perfect storm scenario, where several things with the NBN rollout go awry, it could cost up to AU$94 billion to roll out the fibre-to-the-premises (FttP) NBN, rather than the AU$44.1 billion outlined in NBN Co's corporate plan.

The first factor is that revenue coming in to NBN Co over the rollout will not be as high as NBN Co has forecast in its corporate plan. NBN Co has forecast that the average revenue per user (ARPU) will go from AU$22.46 per month up to AU$62.11 per month by 2021, when the build is scheduled to be completed. Turnbull has said that this is unrealistic, because it assumes 9 percent real growth year on year over that time, and would mean that fixed-line telecommunications' share of the gross domestic product (GDP) would increase by between 60 percent and 70 percent.

Turnbull argued that there is no evidence that people will be willing to pay more for fibre-to-the-premises products, as NBN Co has suggested. His view is that ARPU would remain the same, and grow with the economy at an annual rate of approximately 3.5 percent.

The second factor is the cost of deploying fibre to each premises. The corporate plan puts an average capital expenditure for passing each premises at AU$2,400. Turnbull stated in the paper that he doesn't think this is what is happening, because NBN Co has refused to release the current cost per premises for rolling out the NBN.

"The only certainty is that if the FttP rollout was on budget, NBN Co would be saying so."

He also suggested that the complications with deploying fibre to every premises in a multi-dwelling unit (MDU) would also add to the cost of the rollout. Turnbull said that the cost per premises is likely to blow out by AU$12.9 billion, with each premises costing AU$3,600 to connect. He based this assumption on a Macquarie Bank analysis, as well as industry reporting on the cost of Telstra upgrading South Brisbane over to fibre.

The third factor is the wireless-only take-up. NBN Co has assumed that 13 percent of households will be wireless only by 2021. Turnbull pushed this up to 25 percent.

The final factor is a delay in the rollout, with the network not finished until 2025. A four-year delay would push the required funding from AU$14.9 billion to AU$58.4 billion, Turnbull claimed.

If all four of these factors take place, then Turnbull claimed that the cost of building the network would be AU$94 billion.

It is a perfect storm scenario, but Turnbull suggested that it is a "much more likely forecast" than that undertaken in NBN Co's corporate plan.

Turnbull has also sought to address the argument that to go to fibre to the node (FttN) now, and then fibre to the premises later, would be more costly. He said in the paper that the cost of going to FttP in three years' time would require AU$3,633 per premises, assuming that FttN effectively halves the cost of the FttP upgrade. He said this would be less than an immediate FttP upgrade, which he costed at AU$3,755 — higher than the original estimate of AU$3,600 per premises that he estimated NBN Co's current FttP deployment at.

On this basis, he estimated that a decade-long deferral of going FttP would end up saving AU$12.7 billion. The paper does not take into account the operating expenditure of maintaining the copper component of the network.

When the discussion paper shifts its focus to the Coalition's fibre-to-the-node policy, it makes many assumptions, including being able to reduce the cost per premises from AU$3,600 down to AU$2,700. It assumes the rollout can be completed on time by 2019, and it will take no more than a year after the September election to shift from a fibre-to-the-premises deployment to a fibre-to-the-node rollout. It also takes on the previously outlined assumptions about wireless-only households and ARPU growth at 3.5 percent year on year.

The policy does not outline how long it expects to negotiate with Telstra to gain access to the last section of its copper network, or the associated cost for obtaining that copper, but Turnbull has stated his confidence that a deal with Telstra can be reached quickly, without Telstra asking for more money. This overlooks Telstra already pushing out the NBN deployment through its original negotiations with NBN Co. It is also unclear whether Telstra shareholders would need to vote again on any new deal, which could further delay accessing the copper for fibre to the node.

Turnbull has also said that line tests will be performed on each line to ensure that premises in fibre-to-the-node areas can achieve download speeds of at least 25Mbps, and remediation work or fibre replacement done to those premises that can't reach 25Mbps. There is nothing in the document that states how these tests will be conducted, at what cost, nor how many premises are expected to need either remediation or to be switched to fibre to the premises.

While Turnbull has said that he won't write NBN Co a blank cheque — as he said the Labor government has done — it is difficult to see how costs can be restrained, while at the same time promising to deliver fibre to the premises for every household that cannot get at least 25Mbps by 2016 under Turnbull's own modelling for the cost of delivering fibre to each premises.

As The Register noted on Monday, the policy document also appears to ignore the cost of maintaining Telstra's copper. The policy document suggests that claims of Telstra's copper maintenance costing AU$750 million per year are slightly less than the AU$1 billion of ongoing annual capital expenditure by NBN Co after the network is completed, but this cost was not broken down in the Coalition's final costing of its proposal.

It is clear that the question mark over the Coalition's policy is always going to be the copper component of the FttN network. The comparisons to BT's and AT&T's FttN network upgrades overlook the fact that those companies already own the networks that they are upgrading. Without any clear signals from Telstra over the state of its copper network, or at what price it is willing to hand it over to a potential Coalition government after September, the cost of the Coalition's proposal is not fully known.

It would be a perfect storm scenario of an entirely different kind if Telstra was willing to hand over a copper network in pristine condition for no extra money.

Editorial standards