NBN cost benefit notes

There will be 7 million new fibre connected premises in the world this year, for a total of 50 million up from 43 million. Most of these will be in Asia.

Not many in the scheme of things, but this will be an interesting number to track in the next 20 years.

“The NBN will also facilitate a major reallocation of capital in the telecommunications industry, which has historically been dominated by high Public Switched Telephone Network voice revenues.

The decline of voice revenue is already underway, and is expected to accelerate with the advent of high quality Voice over Internet Protocol (VOIP) under the NBN. However, declining voice revenues are expected to be substituted by increasing broadband revenues, as business models continue to shift from toll calling charges to access charges.”

Telcos make lots of money from the telephone and the line charges they make you pay. Going forward we will stp paying these but pay more for internet. Many of us went this way last century.

NBN Co’s objectives:”

  1. The network should be designed to provide an open access, wholesale only, national network, covering all premises;
  2. The technologies utilised should be fibre to 93 per cent of premises (including Greenfields developments), fixed wireless to 4 per cent of premises (delivering at least 12Mbps), and satellite to 3 per cent of premises;
  3. The pricing principles to be employed should ensure uniform, national wholesale pricing accessible on non-discriminatory terms; and
  4. The network expected rate of return should be in excess of current public debt rates.


The NBN Co product set will be offered as follows:

  • A uniform product construct across fibre, wireless and satellite, featuring the same four product components across each access network, and based on the technology-agnostic Ethernet Bitstream framework.
  • A 12Mbps downstream and 1Mbps upstream entry-level offer across all three access technologies (i.e. fibre, wireless & satellite), at the same price (network ubiquity).
  • An initial fibre product suite with committed speed options of up to 100Mbps and peak speed options of up to 1Gbps (performance certainty and speed).”

I like these a lot, but it’s hard when you try to match satellite and fibre. The idea that everyone in the nation should pay the same, that 93% of premises have fibre and that anyone can use NBN service is fantastic.

By December 2011: “Capability to fulfil, activate and assure an increased number of products with multiple RSPs and up to 19 per cent of premises passed.”

Over 20% of the fibre rollout possible within a year, or maybe this just refers to the ability to accept those connection.

Underlying assumptions: “the most significant … growth in speeds and demand and hence revenue.”

The question here is not whether demand for speed will increase, or whether prices will come down, it’s what the total pool of money spent will do over time – and that goes back to end users choosing to spend more money on things that require fibre and less on other things such as voice calls.

The stated internal rate of return is also dependant on the completion of the Telstra deal, which has a material impact on construction costs.

Part of the reason for the delayed cost-benefit is that exposing the Telstra deal benefits gives Telstra more negotiating power. The deal is not yet done.

The internal rate of return does not take account of any external benefits anticipated from the NBN to the economy, productivity or social outcomes.

The economic benefits to society make the rest of the the returns look pathetic. It’s impressive that the project stands alone, but if Australia (or NZ) are serious about competing then these fibre projects are a must. There is much to lose and much to gain.

The number of GB/month in June 2010 is an average of fixed downloads (9.2 GB/month) and wireless downloads (1.2 GB/month).

Does your wireless device download 1.2GB/sec per month? That’s the average in Australia

NBN Co has also referenced usage projections against a number of independent sources:

  • Historic ABS data, showing a 36 per cent per annum compound average growth rate from 2000 – 2010.
  • Cisco international forecasts for internet traffic to increase at a 34 per cent compound average growth rate from 2009 – 2014.


In NZ Southern Cross is reported to have said it’s 35% a year for the last 10 years. Everybody is forecasting much higher CAGR going forward, as better plans and higher available speeds release the pent up demand.

A conservative approach has been taken with regard to long-term data usage growth, factoring in considerations including saturation of usage, slowing growth in online hours and increasing delivery of content on multicast applications. This sees growth fall to ~8 per cent p.a. from FY2030, well below the baseline increase in access speeds forecast.

Every cost-benefit or Investment Memorandum incudes the words “A conservative approach”, and it usually isn’t. However I disagree with their findings and see no reason why a 30 year trend of 35% growth would suddenly stop in 2030.

“At the end of the contribution and deployment period, the total capital expenditure (capex) is estimated by NBN Co to be $35.7 billion…. The Agreement between NBN Co and Telstra, worth an expected value of $9 billion, in net present value terms(NPV).. ..$13.8 billion decommissioning and infrastructure payments by June 2020”

It seems like $35.7 bilion of capex and $13.8 billion in opex, but the NPV for both is of course lower as it is spent over time. These are scary numbers, but NBN pays for itself, just over a very long time. And don’t forget to add in the economic benefits.

A capital weighted WACC has been derived at 10 per cent-11 per cent over the 30-year period.    This has been factored into the business model.

NBN are using a multi-staged discount rate for the project, as the risk changes over time. However the chart below shows that the average cost of capital is above 10% until 2025, so I struggle to see how they can have an average of 10-11%. But I do like the idea that the various stages of investment are at a declining risk.

And finally, once again, there is nothing on international capacity, as it is out of mandate.

With 10.9m (current, growing each year at 1.6%)  premises on 100Mbps connections, that’s a combined capacity of 1090 Terabits/sec. That’s about 200 times the current unlit international capacity for Australia and New Zealand, so if all unlit capaicty went to Australia the contention ratio wuld be 200-1. That’s about 10 times to large – in NZ Crown Fibre  wants about 20-25 I think. So it implies that we need 1090/20=54.5 terabits/sec of capacity from Australia by the time the build is complete, in about 2020.

But we’ve under-estimated the combined capacity number, as a large percentage of households will choose the 1 Gbps plans instead of 100Gbps (and ther are other adjustments up and down to make). However you look at it though, there will be a huge demand for international capacity that Pacific Fibre will only partially meet.

Published by Lance Wiggs