Sunday, 18 August 2013

NBN: disputing the four Turnbull "Stress Tests". Pure fantasy.

Turnbull created 5 spreadsheets, published on 09-Apr-2013, that he asserted contained "reasonable" assumptions and allowed him to claim a fantastical figure of "$94 billion" as the Coalition estimate of the NBN Co Corporate Plan:
  • sustained 3.5% pa real growth
  • 40% increase in Fibre build costs, to $3600.
  • 25% of households opt for Wireless (4G) only, a 50% increase.
  • Fibre construction takes 50% longer to finish in 2025, not 2021.

So "reasonable" that a large staff of highly competent professional staff at NBN Co, over a couple of years, and covering all relevant disciplines, completely missed them. Nothing like any of Turnbull's assumptions have appeared in any of the Risk Analyses published by NBN Co.

That's as likely as Turnbull being able to pick holes in QANTAS or Commonwealth Bank business plans: Not likely in the least, yet he continues the charade and attacking the professional competence and integrity of the whole of the NBN Co staff. That's magnificent arrogance from Turnbull: "everyone is wrong but me"!

Turnbull is now widely spruiking "My figures must be right because nobody has challenged them!".

Which is utterly wrong: this is why I was given short shrift by his staffer. Because I'd hassled them and questioned their figures and assumptions since April. And they are fully aware of it. But enough of his lies.

The five "stress tests" (my phrase) are four independent "worst-case scenarios" that are then rolled into one colossal OMG! scenario, the probability of which isn't given, but is likely to be a million to one.

The problem for Turnbull is that he both insists that his figures are completely up to date, and he didn't revise his published spreadsheets when NBN Co released their latest update on 19-April.

The NBN Co results of April refute all Turnbull's inane and fantastical propositions.

Sustained 3.5% pa real growth: WRONG. Growth is 9% or higher.

The actual financial results of NBN Co are 2-5 years ahead of forecasts, depending on what variable is examined. ARPU was $38 for March, with AVC (line access charges) at $30, around 20 months ahead and CVC (Volume) 50% higher than the ABS national average.

The NBN Co Corporate Plan is being shown to be conservative.
At over 50,000 premises connected, the results are statistically significant, the mean and variance within a small fraction (<0.5%) of the whole 12M premises figures.

The real AVC revenue of ~$30, especially 31% selecting 100/40Mbps vs 18% planned, should deeply trouble Turnbull: it's double the pathetic $16 "one price fits all, take what you get and like it" of his Copper/Node Plan.

This is the real reason Turnbull can't get his rate of growth up for the Copper/Node Plan: he can't charge enough and can't match consumer preference to products.

40% increase in Fibre build costs, to $3600: WRONG: 40% is $450 not $1200

The Professional experts within NBN Co released the current mass rollout costs of constructing Fibre connections. They are in two parts:

  • Fibre to the street, past the premises: $1100-$1400 Estimate at Completion (EAC), and
  • Fibre from the street into the premise: $1100, final.
The only cost that can vary is the Fibre past the premises, the costs of running Fibre into premises are very well known and dominated by the fixed Telstra 'PSAA' contracts for lead-ins and access.

A 40% blowout in Fibre construction costs, should it occur, could only affect the fibre past the premises component. It's completely incredible to think that the prime contractors signed these large contracts so far out. 10-15% might be possible, as reflected in the ranged EAC estimate.

25% of households opt for Wireless (4G) only, a 50% increase: WRONG not 12.5% but 6.25% decrease.

This is a based in profound ignorance of the market. As I've outlined before, we know from the Sandvine data the distribution of broadband demand is exponential: 1% of users consume 10% of downloads, while the low 50% consume just 6.4% (six point four).

The lowest 25% will be on both the entry-level plan and consume 2% of download data.

The high uptake of  high-end services is perfectly in-line with this.

The April NBN Co figures showed that much higher propensity to pay higher prices than forecast, 31% vs 18% paid $38 wholesale for 100/40Mbps and 39% not 51% paid $24 wholesale for 12/1Mbps entry-level.

The AVC ARPU is around $30/mth, reducing the entry-level by 8.3%, increases the ARPU to $30.50, while reducing AVC revenue by 6.2%.

CVC revenue accounts for around 15% of total Revenue. This will be reduced by 0.16% (8.3% of users who in total contribute 2% of download). That's enough to be a rounding error.

Turnbull calculated a 12.5% reduction in Revenue in 2021, it can at most be half that.
We are yet to see the effect of 1Gbps services, at $150/mth, being offered and

Fibre construction takes 50% longer to finish in 2025, not 2021: WRONG, no effect on CapEx

Mike Quigley when interviewed by Parliament demolished this point, so I don't have to.

Quigley said, as I recall: "The NBN Co Plan will be cash-flow positive by then, delaying the project by 50% will NOT affect CapEx or Funding required".

Aggregate Position: NIL variance outside budgeted Contingency.

With two of the four "stress tests" of no affect and the other two out by a factor of 2- and 3-times respectively we can safely say the "$94 billion"is completely bogus and fantastical.

Instead of a 1 in a million chance, there might be a 1 in 10,000 chance of a 10% budget over-run.

This is in-line with what you'd expect of a highly professional and independent group of experts who do this every day for their living.

Just to rub salt into the wounds, the April NBN Co document discloses they have a 10% contingency and that it is, as yet, untouched.

Worse to come for Turnbull: what if Fibre is a "River of Gold?"

When we have the first full-year results from NBN Co, we will know just how conservative the forecasting was. Already we know that consumers have an almost double propensity to pay higher AVC charges than anticipated.

The other figure is the 30%/year growth in Download demand. At 45Gb/mth, NBN Co users are already 50% ahead of the Dec 2012 ABS national average figures (series 8153).

If this is showing pent-up demand, then volume growth will exceed the 70% long-term average in the ABS data series.

This time next year we will know how things are shaping up.

If data demand does continue to run at the long-term ABS average, or higher, then revenues will be significantly ahead of forecasts, easily by 25%, or $1.5-$2 billion in 2021, enabling NBN Co to self-fund much earlier, to both reduce CapEx & Funding and increase the capital intensity, to speed up the mass rollout.

This more likely scenario, is borne out by the data currently available.


  1. Thank you for your contributions to the NBN debate. I like many others are not in a position to fully understand the technical information behind the policies, your posts go a long way in enhancing my and others knowledge on the subject. Keep up the good work.

    1. Nait,

      You're very welcome. I'm glad my writing has been of some value to you.


  2. why on earth would Turnbull assume that under labor FTTP users will opt for 4G ? the fibre connections are close to weather proof (unlike FTTN with copper) lower latency than wireless and will be a couple of digits faster for those people actually opting for a higher tier FTTP plan.

    1. Tim,

      _good_ question: Cost, Speed, Availability.

      The Coalition thesis is that "Fibre is expensive" and people will get sick of waiting.

      And they have an assumption that Telstra is going to go after low-end Fibre users aggressively.

      I don't necessarily agree with their assumptions..


  3. Hi Steve,

    Have you redone the NBN debt figures based on the current 9% growth? I wonder how much lower it will be than the current $37.4 billion max funding?

    Even if an extra 10% is spent on construction...


    1. David,

      No, I haven't gotten to reworking the figures. Great suggestion, though.
      It's a bit tough getting inside all the NBN Co figures - they'd prefer to keep costs and other figure confidential.

      I'll keep this in mind, see if I can think of a valid, simple way through.


  4. Hi,
    i was just wondering where you got your information (stats, percentages, etc.) from? I would love to see your sources, as I am quite interested in the topic.

    1. Old Spice,

      Links are in the piece:
      - 9th April NBN statement has most data
      - Link there to Turnbull "Background"

      Quigley comment can be found in pg5 (Question 53) of Australian Partlt. doc:


  5. Steve, if your analysis of earnings is correct - which I believe it is - then logically doesn't that increase the likelihood of the Libs switching their position on the NBN should they win the election? After all, why would anyone want to persist with green fridges and copper when fibre installations would be proving themselves to be little goldmines?

    One therefore has to acknowledge the possibility that FTTH might still happen, though of course how it is controlled and charged for is a worry. Telstra was also a "rivers of gold" proposition and the Libs sold it. They have a track record of selling gold-egg-laying gooses to vested interests.

    Of course if they did switch to an FTTH rollout, that would certainly make Turnbull look the biggest goose of all, but that wouldn't be of concern to the Libs - all gooses are expendable, profitable or not.

    1. Dedalus,

      I don't think there's a scenario where the Libs would go FTTH. They need the NBN to fail and for their claims of "Worst Government _ever_" and "Great big Waste" to be borne out.

      I don't believe they care about wasting $30 billion or sacrificing our collective future/competitiveness. It seems only about ideology to me.

      Depressing view I know. I'd like a better scenario.



Note: only a member of this blog may post a comment.