Beyond The NBN Business Model

Beyond The NBN Business Model

Malcolm Turnbull is the Shadow Communications Minister and has had a beef with the Labor government’s National Broadband Network (NBN) project ever since it was minted. He addressed the American Chamber of Commerce In Australia (AMCHAM) last week on the problems with the NBN and how they might be fixed by a Coalition government.

What have we learned about the National Broadband Network this week?

Well, courtesy of the Australian Financial Review’s story about the hidden costs of the NBN on Monday and its aftermath, we’ve learned there are many reputable, independently-minded economists and analysts who share the Coalition’s view that this has been hugely understated by the Government up to now — perhaps by $10 billion — by excluding on-Budget interest costs.

We’ve learned that the costs of the NBN that Labor does admit to — those of building and operating the network through to the point where it is financially self-sustaining — are going to be higher than the Government previously thought, even with the rosy assumptions baked into the latest Corporate Plan.

We’ve learned that by by mid-2013 the NBN will have received $7.5 billion from the Government in capital subscribed as equity.

And finally, mid-2013, near what the Coalition hopes will be the end of six years of Labor Government over which Senator Conroy and other ministers have said literally hundreds of thousands of words about the huge economic and social value of universal fast broadband, the NBN’s vaunted fibre network will have 54,000 customers. That’s if NBN Co actually achieves its milestones for a change.

But against that paltry number there will still be in excess of a million households — and perhaps as many as two million — across the nation, in both regional and remote areas and the suburbs of our major cities, which cannot get a fixed line broadband connection that meets the basic requirements of average 2012 internet users.

Incredibly, many of areas where these households are situated are not on the NBN’s current three year rollout roadmap.

Many Australians who most urgently need upgraded broadband have been waiting since a nationwide network upgrade was first proposed by Telstra in 2004. We all know the subsequent history, where first the Coalition and then Labor were unable to strike a deal with Telstra – make no mistake, everyone bears some responsibility for today’s situation, including the regulator.

But the abandoned policies of the past are not relevant today. What is relevant is whether Labor’s current approach is the most rational and cost-effective way to achieve the objective all sides of politics broadly agree on – which is to ensure all Australians have access to fast broadband at affordable prices as soon as possible.

I believe this week’s additions to the information we all have about this area of public policy confirms that it is not. Let me discuss these developments one by one, and then give a brief outline of our alternative approach.

On-Budget And Off-Budget Costs

What precisely will Labor’s NBN cost if it is completed?

Well, the revised NBN Co Corporate Plan assumes that to reach a point where the fibre, fixed wireless and satellite networks are complete, NBN Co is cashflow positive and there is no further need for public funds requires peak funding requirement of $44.1 billion in 2021, compared to $40.9 billion in the old plan.

Note both these figures include interest on borrowings paid by NBN Co.

Another approach to estimating the cost of the NBN is to put a number on what the physical network will cost — the capex needed to complete it.

This figure is now $37.4 billion, up from $35.9 billion in the old plan — although count me skeptical whether the proposed network could be built for that price given the experience so far. In any case, this fails to account for funding costs or operational losses, and is therefore somewhat misleading.

Both the old and new corporate plans assume the Commonwealth funds NBN Co with a mix of about 67 per cent equity — shares the Commonwealth buys in NBN Co — and 33 per cent debt borrowed in the name of NBN Co (but, inevitably, certain to be valued by the market as though it is guaranteed by the Commonwealth — as though these debt instruments are Government bonds).

The new plan shows the amount of equity invested by the Commonwealth rising from $27.5 billion to $30.4 billion. The amount of debt NBN Co will service increases from $13.4 billion to $13.7 billion. As noted, the interest on this debt appears on the NBN Co income statement.

Now, every dollar being spent by the Government or NBN Co on this project is borrowed money. But when Labor talks about the cost, it excludes any interest on the money it borrows to subscribe as equity.

Senator Conroy attempted to distort this issue by claiming we want the interest cost of these borrowings to be paid by NBN Co. We don’t — they belong on the Budget, where they are. All we want is for him to tell us how much these interest costs are. How much have they been up to now? And how much will they be over the life of the investment?

Because, by any measure, these are part of the total cost of the NBN to taxpayers.

To claim otherwise is nonsense.

If an individual takes out a margin loan to buy shares, of course they need to include the interest costs on that loan when they tally up whether they made a good investment or not.

Why would the investment taxpayers are being compelled to make in NBN be judged any differently?

The Government’s strategy from the beginning has been to hide the true costs of the NBN and exempt it from Parliamentary and public scrutiny – even from FOI. This is not good enough when the investment and risks for taxpayers are so huge.


Our estimate of another $10 billion in hidden subsidies for the NBN is before any takeup or revenue shortfalls and any network construction blow-outs or delays. We know that assumption will not hold true – Wednesday’s release of revised figures both for how many people can connect to NBN at various points in time, and how many will, underscored that.

The problems with the pace of the NBN rollout so far have gotten plenty of airing this week, and so I won’t go over them in detail.

Perhaps the key figure is that the NBN’s fibre network (the raison d’être of the whole exercise, and under the current plan responsible for servicing 93 per cent of homes and businesses) will reach only one in four of the 1.3 million households the Government originally claimed by June 2013.
NBN Co originally expected to have 566,000 paying customers using its three networks (fibre, wireless and satellite) by then. Now it estimates only 92,000.

There are some curious figures in the expenses side of the revised Corporate Plan. The one that sticks out is in December 2010, NBN Co projected $3.8 billion of what it calls ‘other’ operating expenses between the start and the end of construction — that is, from 2010-11 until 2020-21 if the schedule is met. In the new plan NBN Co projects $7.9 billion of ‘other’ expenses.

From NBN Co’s previous financial reporting to Parliament we know most of these expenses — 96 per cent of them in the six months ended on 31 December 2011, according to the Government’s report to the Parliamentary Joint Committee on the NBN — are for employee costs, corporate travel, lawyers and consultants, IT, and office space.

On the face of it these non-network-related overhead costs have blow out by $4.1 billion – they are now expected to be more than double the estimate signed off by the NBN Co Board and the Government less than two years ago.

We know NBN Co is an exceptionally generous payer by industry standards, but we didn’t realize it was that generous! Or is it headcount that is going to rise twice as high as previously expected? Or other corporate costs?

Senator Conroy owes all of us an explanation for exactly what is going on here.


But revenue is the real issue for the next generation of broadband networks — around the world, we see governments and telcos hesitating or pulling back from investments in fibre to the home, and even from less costly network upgrades, because they just can’t see paying customers and demand to justify these investments.

Uncertain future demand in combination with extremely large, long-lived and irreversible investment requirements is the conundrum. It hasn’t been solved entirely satisfactorily anywhere, not even in the very different economies that Senator Conroy likes to talk about such as Korea and Singapore, which have relatively high penetration levels for very fast broadband.

And if one considers NBN Co’s total assumed revenues over the period prior to the 2013 election are now estimated at only $20 million, compared to $205 million in the old Corporate Plan, we might surmise that NBN Co doesn’t feel very certain or confident about this issue either.

Broadband and telecommunications revenue projections boil down to two things — how many people use it, or take-up, and what they on average pay each month — or ARPU, in the elegant language of telecommunications.

On take-up, it was interesting to hear David Thodey yesterday say that Telstra has seen the number of wireless-only households rise to 14 per cent, and anticipates it will rise further.

That means it has already exceeded the NBN Co’s assumption of 13 per ccent today and 16 per cent by 2040.

It is hard to calculate a figure for the ‘wholesale’ share of current ARPU.

But we know that since 2003 the industry’s basic fixed line voice revenues have been stagnant in nominal dollars at around $11bn a year.

Think about that — in 2003, there were just 470,000 fixed line broadband customers compared to 5.3 million in 2010.

In September 2003 the ABS began breaking down subscribers between dial-up and ADSL — roughly half of all customers were still on dial-up then.

At the time, the ABS was still measuring the amount of data downloaded in MBs — for ease of comparison, I will tell you the total amount downloaded was 4,600 terabytes. As of December 2010, the ABS showed that over 350,000 terabytes were downloaded in Australia.

So when Senator Conroy says there’s a broadband revolution just around the corner, don’t be fooled — it has been underway for over a decade.

And the single most salient point of that revolution is that despite all the investment and upgrades and technological advances, despite the increases in usage, in speed, in penetration, in downloads and services — none of that has coincided with a rise in the amount of money people are willing to spend on fixed line communications.

So what does the NBN think of all this? Well for a variety of reasons it has artificially low ARPU in the current years, with revenue per customer hovering around $20 a month. The main reason why this ARPU is so low is that the NBN was forced to drop its Connectivity Virtual Circuit charges — which have been described as a proxy for usage charges — until RSPs can attract a critical mass of customers in each fibre serving area and thus buy a reasonable amount of capacity to smooth out among users.

By 2020, however, the NBN is forecasting that revenue will increase to $60 a month and then increasing to $100 a month to 2039.

How will the NBN achieve this? Well, it will largely tried to square the circle on a long running issue between infrastructure providers and service providers. That is — while service providers have driven consumption and usage, they have not been the parties who have had to stump up the bill when it comes to upgrading infrastructure. Of course vertical integration fixes that problem but leads to other perverse incentives in the market, which are well known in Australia.

So getting beyond the vertical integration model requires a new business model. In the NBN’s case, the model involves charging customers a premium not just for increasing speeds but for increasing usage, via its CVC charges. In fact, the current corporate plan has substantially brought forward the usage charges it intends to impose. In the 2010 corporate plan, CVC charges were to account for only 26 per cent of revenue by FY2025 (p.116) while in this plan, the CVC and other products such as multicast approach almost 50 per cent of revenues by 2020.

What are the threats to this model? The most important threat is that it is extracting a premium where none has ever existed before. In Telstra’s response to the ACCC’s first consultation paper (available online here, p.14), its showed that a customer buying a 100mbps package in 2012 with a 30 per cent annual usage increase would see their CVC prices increased from $1 a month to $50 a month by 2025. This is a heroic assumption given the trends of the past decade.

Even the willingness to pay a premium for speed is heroic. We know that in very advanced markets like Japan and Korea, customers have been churning back from higher to lower speeds to save a very small price — in the case of Korea, a case study I have looked at, customers have been churning from 100mbps packages down to 50mbps packages to save the equivalent of around $3.

So when Mike Quigley says wholesale prices will fall in both real and nominal terms, what is he actually talking about? Well, that is based on an assumption that users will continue to keep downloading more data and continue using higher speeds. And yet, charging premiums to do both of these things threatens to slow or halt altogether the natural progression of consumers doing these things. That has been the experience of the past decade and it has been the experience in mature markets.

The irony here is that Stephen Conroy keeps warning that there may be applications and services that will require everyone to have 100mbps speeds. And yet the NBN business model is designed to subsidise consumers getting on the network now — when there are few services or applications that justify it. It will then attempt to defy the economics of telecommunications in a decade or so — in other words, when there is a much higher chance there may be applications consumers will need higher bandwidths for.

The Way Forward

All of this leads to the question: If Labor’s NBN business model is so flawed, then what is the right business model?

My proposition is simple though: The best business models and the greatest innovations occur not in Canberra but in the organized chaos of the market place. When I look around and see the many representatives from our best telcos and finance industry, I would wager that those guys have forgotten more about telecoms than all the regulators and politicians in Canberra could ever hope to learn.

I know what it is like to run a business. I know what it is like to fail, to know the heart break and disappointment of that failure, and what it is like to dust yourself off and try again. I never had the luxury of issuing corporate plans like the NBN has done this week, of announcing huge cost blowouts and delays, of telling shareholders they have to stump up more money, without fear or failure or that there would be serious consequences.

The NBN risks becoming a seriously undisciplined monopolist, which becomes better at threatening its shareholders and gaming regulators than it is on delivering to its customers. Indeed, just before the NBN withdrew its latest SAU from the ACCC, one industry participant was quoted that “the amount of discretion being sought by NBN boss Mike Quigley would make even the Telstra of old blush.” (The Australian, online here)

But I also recognise that infrastructure investment is hard to get right and we need to get the market settings right to maximise the benefits of the continuing broadband revolution. As the NBN project continues to flounder — as indeed does the Government itself – there is an increasing focus on what the Coalition would do in power. It is a healthy sign that this is occurring. Unlike Senator Conroy or the NBN, we do not attack fair scrutiny of our plans nor do we seek to browbeat those covering this area of policy.

But I would just like to finish by saying this. People get dispirited by the partisanship of this debate — not just for its bitterness but for years we seem to have spent going over the same issues in the most circuitous fashion. I understand from talking to industry how difficult uncertainty over the future can be.

Well, by June 2013, the NBN is now projected to be pass just 341,000 houses by fibre. That is just 2 per cent of the total 12.2 million houses it will supposedly passed by 2021 – in other words, the ships crew will barely be starting to move the gang planks by the time of the next election.

Our key pledge is to conduct a cost-benefit analysis.

That is not just because we think it will result in a more efficient rollout. But because it will finally give us a document which all politicians should be able to accept as impartial. If Senator Conroy were truly to stand by his convictions, he would go to the next election with that cost-benefit analysis to show to the electorate.

Republished with permission from Malcolm Turnbull MP

Image: Brendon Thorne/Getty Images