Creative maths taints NBN Strategic Review's revenue argument

Careful reading of the NBN Strategic Review's financial projections suggests its authors have taken gross liberties to ensure the revised NBN model stacks up better on paper than the current rollout.
Written by David Braue, Contributor

One of the most contentious issues in the National Broadband Network (NBN) debate has been the forecasts about just how much revenue the various models will produce — and, as a result, what sort of internal rate of return (IRR) each network topography can produce.

Yet, while the NBN Strategic Review (SR) seemed to triumphantly debunk earlier revenue forecasts and a planned IRR of 7.1 percent as over-ambitious – and suggested that the proposed multi-technology alternative offered better return on the government's investment – close examination of the figures in the review suggest that this conclusion has only been possible through egregious manipulation of the truth.

Play around enough with the numbers, and you can get any answer you want. Image: Public domain, Tungsten

Table 0-2, on page 17, summarises the arguments by presenting comparative financial models for the six different scenarios considered in the report. The report's authors concede that all numbers in the table “are estimates and subject to change”, but they also use differing terms of analysis to paint Scenario 6 – which, conveniently, is the model promoted for some time by communications minister Malcolm Turnbull – as being superior to any other model.

Cumulative revenue between FY2011 and FY2021, the table tells us, will be $10 billion in the existing rollout (Scenario 1), $16b under Scenario 4 (expanding HFC's footprint) and $18b under Scenario 6.

At some levels, this makes sense: it is, after all, already widely accepted that the fibre-to-the-node (FttN) element of Scenario 6 could be rolled out more quickly than a fully fibre-to-the-premise (FttP) deployment – and that revenues under an FttN model would therefore ramp up faster than under FttP.

This might explain the revenue difference expressed in the SR – if, that is, you ignore two highly inconvenient facts.

The first is a matter of timing: the SR's revenue forecasts only extend until the end of FY2021, even though the same table shows that the current rollout wouldn't be complete until the end of CY2024 – halfway into FY2025.

In other words, the painfully apples-and-oranges analysis ignores three and a half years of what would be the current NBN's most important revenue-earning years in what can only be an attempt to paint Scenario 6 as delivering higher revenues.

These 42 months are crucial to FttP and FttN models for entirely different reasons.

For FttP – which may take longer to reach its maximum rollout speed but will be generating more revenues when it does – they will represent 42 months when the network rollout is racing at maximum speed to complete its project.

Premises-passed figures will not only be increasing rapidly during this time, but it is reasonable to assume that customer activations will be at an equally high takeup given that demand for high-speed services will be significantly higher than it is now. Oh, and because customers will have to switch when Telstra disconnects their copper services.

Revenue forecasts only extend until the end of FY2021, even though the same table shows that the current FttP rollout wouldn't be complete until the end of CY2024 – halfway into FY2025...The analysis ignores three and a half of what would be the current NBN's most important revenue-earning years in an attempt to paint Scenario 6 as delivering higher revenues.

The SR presents but then ignores its own revenue projections for this period, suggesting in table 2-21 (p56) that the Revised Outlook for an all-FttP deployment would see the network deliver $4.2b in FY22, $5.0b in FY23, and $5.8b in FY24 (there are no figures for the first half of FY25, but given the SR's projection of completion at end of CY2024 we can assume the NBN would generate at least $2.9b more revenues over those six months).

In other words, the 42 months missing from Table 0-2 represent a time when an all-FttP model will be at its revenue-generating prime. By then, early adopters, who might initially take lower-cost entry-level services in the next few years, will likely be upgrading to higher-revenue 100Mbps, 500Mbps or 1Gbps services.

This will not be happening under FttN, which by NBN Co's own admission does not support the same type of high-revenue services, guaranteed performance levels and overall performance as FttP. A Scenario 6 rollout will see revenues peak, and then begin to decline rapidly as retail competition pushes down prices for commoditised services and inadequately-serviced customers find their own fibre.

The SR's modelling completely ignores the changes in those crucial years – most likely because the preferred Scenario 6 will be complete by the end of CY2020. The revenue figures presented in the table, then, are based on comparing a complete multi-technology NBN's revenue contribution within its first year of full operation, with the revenue generated by an incomplete, FttP-only NBN for which $15b of peak revenue has simply not been counted.

That has not, it should be noted, stopped the Strategic Review from counting capital expenditure under Scenario 1 through FY2024 – further skewing the cost-benefit equation against FttP and its $56b capital expenditure.

In other words, the review has presented Scenario 1 capital expenditures of $56b vs revenues of $10b, versus Scenario 6 capital expenditure of $33b and revenues of $18b for the multi-technology mix. This may support the case for Scenario 6 on paper, but if you factor in the missing $15b the figures look much different.

Capital expenditure of $56b on revenues of $25b, after all, actually compares pretty favourably with capital expenditures of $33b on revenues of $18b – and would keep the IRR for Scenario 1 from being the disaster it's made out to be.

Clearly, however, this would not serve the purpose of the SR, which was always expected to actively discredit the current rollout.

Ignoring FttP revenue growth is one problem, but the omission of those 3.5 years is important for Scenario 6 for another reason: previous NBN Co modelling has suggested that a predominantly FttN network will suffer significant revenue decline over time, as customers abandon its relatively-limited services for third-party alternatives.

Remember that NBN Co's advice to the incoming minister, which Turnbull never wanted to be made public and tried to discredit upon its release, showed this revenue decline could be as much as 30 percent. This was later confirmed by the deputy secretary of the Communications Department, Ian Robinson.

This should have had a direct impact on the revenue projections contained in the Strategic Review – especially during the lost years from FY2022 to FY2025 when the SR apparently fails to represent Scenario 6's revenues.

There's an air gap between NBN Co's earlier 30 percent worst-case scenario and the 2.5 percent worst-case scenario outlined in the SR's modelling. If revenue decline was anything like the former number, the decline would obliterate the 5.3 percent IRR the SR has assigned to the government's model.

Indeed, the entire SR seems to completely ignore NBN Co's earlier warnings about revenue shortfalls on a non-FttP network. Granted, the review does not advocate a 100-percent FttN solution so the precipitous drop forecast earlier might be moderated somewhat – but it is optimistic to suggest that cutting the amount of FttP by 75 percent, and the revenue implications of such a move, would only cut revenues by 2.5 percent per year.

Just to recap: just four months ago, NBN Co warned that implementation of the Coalition's NBN policy could cut $1.5 billion per year from NBN Co's revenues in FY2021 because customers won't be paying as much or getting as advanced services as under FttP.

There's an air gap between NBN Co's earlier 30% worst-case scenario and the SR's 2.5% worst-case scenario. A revenue decline of anything approaching that magnitude would obliterate the 5.3% IRR the SR has assigned to the government's model...[and] it is the height of optimism to suggest that cutting the amount of FttP by 75%, and the service and revenue implications of such a move, would only cut revenues by 2.5% per year.

Now, we have a Strategic Review, framed and paid for by the Coalition government, that says a network nearly half comprised of FttN services – it is important and fair to remember that one-quarter of Australians will still get higher-revenue FttP under Scenario 6 – will generate revenues nearly twice as high as an FttP-only model by 2021.

And we have a series of revenue projections that ignore over three years in which an FttP-only model would still be ramping up – at the same time as an FttN-heavy model would have already started losing customers that had already outgrown the limited capabilities of services whose speeds even NBN Co can't guarantee.

The Strategic Review does not talk about the period between FY2021 and FY2025 because it knows that those years will see the technological limitations of Scenario 6 become increasingly evident. It also knows that NBN Co will, by that point, already be planning to upgrade its multi-technology mix to FttP; remember that even NBN Co chairman Ziggy Switkowski has admitted that the Coalition's network will likely need to be upgraded five years after it's complete – just one year after the full FttP rollout would be complete under the Revised Outlook scenario.

The net result of all this is to overstate the revenues from an FttN model while understating the revenues from an FttP model. Conveniently, this understates the IRR for an FttP model and overstates the IRR for the multi-technology model.

It's a curious bit of accounting, and highlights just how ready the authors of the Strategic Review were to make the Coalition's model come out on top. And, more than anything, it's a reminder of how careful the industry must be when evaluating the varying claims around the best way forward for the NBN.

What do you think? What can we read from the figures in the Strategic Review table? And, are you comfortable with the review's findings?

Editorial standards