When the current version of the National Broadband Network was unveiled in April 2009 many people in the telecommunications industry were delighted.
Delighted that Telstra’s vertical integration would come to an end. Delighted almost all of the NBN would be fibre to the premises, an expensive but capable architecture often seen as the “end game” for last mile networks And delighted that NBN Co’s mission would be strictly limited to building and operating wholesale networks open to any service provider on fair and equal terms.
The Labor Government’s 2009 broadband plan unabashedly threw reform-era orthodoxy under the bus. It had no place for out-dated ideas about quantifying and comparing costs and benefits, or making cross-subsidies transparent, or eliminating artificial barriers to market entry.
But it also apparently provided, at a stroke, a simple resolution for several complex problems – how to upgrade broadband quality, how to deliver city-standard services in the bush, and how to level the playing field for service providers sharing bottleneck infrastructure. Hence the many positive reactions in the industry.
Of course simple and popular solutions to complicated problems don’t come cheap. The original $43 billion price for the Government’s 2009 NBN was more than nine times the contribution taxpayers were going to make to the 2007 version it replaced.
But for insiders favourably disposed towards the policy, and indeed much of the public, even this lofty sum was defensible given the benefits the NBN was heralded as providing. Implicit in such views, I suspect, was a hope that ‘somebody else’ would ultimately pay.
That was the beginning of a public debate over broadband economics that has become increasingly detached from reality as it has evolved:
We’ve seen fervent supporters of the NBN argue fibre will generate such sizeable economic and social externalities that nobody should even contemplate trying to measure them.
We’ve been told initial NBN plans offered by ISPs in late 2011 (when barely a thousand users, not all of them paying, were connected to the NBN fibre) offered definitive proof prices wouldn’t rise under the new monopoly.
We’ve heard claims that e-health, e-education, smart grids and sundry other applications – most available today, some needing very narrow bandwidth – depended on the NBN.
Repeated often enough, falsehoods and distortions degrade a public debate. My colleagues and I are now routinely accused of nit-picking, negativity or worse if we dare raise the question of whether the NBN offers value for money or will ever make a profit. We apparently just don’t comprehend that this project serves a higher purpose: nation building.
Well, when politicians use the words “nation building,” reach for your wallet!
While I can tell you from experience that canvassing these matters doesn’t win popularity contests in the blogosphere, I’m going to revisit the economics of Labor’s NBN. My focus will be on the lessons we can learn that are most relevant to the Coalition’s policy.
What have we learned since 2009?
But first let me observe that today’s landscape is quite different to 2009. The three years since Labor’s current policy was unveiled may not have seen much fibre deployed by NBN Co or many Australians get better broadband, but they have yielded a lot of relevant information.
One of the most pertinent trends since 2009 is a marked slowing in deployments of fibre in most OECD markets, as analyst Rupert Wood at Analysis Mason points out. He argues without a ‘killer app’ for FTTP other than video, consumers won’t pay a premium for super-fast broadband, and hence carriers are pursuing less capital-intensive alternatives if they can.
Wood is also critical of FTTP as a carrier upgrade strategy in competitive markets for a second reason – the sheer time to construct and bring them to market, which he argues exposes incumbents to damaging incursions by 4G/LTE or cable rivals.
NBN Co has no such concerns since it will have no direct competitors. But Australians who have waited for years for better broadband would not be thrilled to hear this critique of FTTP – although they may already be onto it given the glacial pace of progress since 2007.
If we set NBN Co and other public-sector-led investments to one side, the real dilemma with FTTP is this: the costs of the upgrade substantially exceed the value consumers place on what it delivers. That is why private investors have been so hesitant. What’s increasingly clear is that in a post-GFC world this applies to many network industries: operators in water, electricity, cable TV and so on are grappling with variants of the problem, although nowhere is the gap between upgrade costs and value to typical users as large as for FTTP.
Of course prices for many utilities reflect decisions by regulators rather than consumers. In Australia we’ve seen considerable public anger over repeated increases in regulated prices for electricity and water. The pollsters tell us consumers remain enthusiastic about the NBN – but will they remain so if, over time, prices stop falling in real terms or start to rise?
For the NBN, recent events in the retail electricity market offer an interesting reference point. Electricity prices fell in real terms from the mid 80s to the mid 90s, and were then broadly stable until 2007. Since then a huge wave of transmission and distribution capex (some reversing earlier neglect, some less clearly needed, all earning a regulated ROI) and the inevitable resulting price hikes have returned inflation-adjusted electricity prices to roughly where they were in the early 1980s. But don’t try explaining that to enraged consumers.
Perhaps in part mindful of that experience, Paul O’Sullivan, whose company Optus was an early cheer leader for the NBN, has recently argued NBN capex may be excessive – with the attendant implications for retail margins and consumer prices. He wants the ACCC empowered to judge whether NBN Co investment plans are economically efficient. Of course regulators were not able to effectively exercise such oversight in electricity.
A related observation is that it will be consumers rather than taxpayers in the front line if the NBN ‘s financial performance lags. Of course most consumers are taxpayers, and vice versa, but the difference matters – it is far more politically feasible for a Government owned monopoly to lift charges on the former than seek a legislative appropriation paid for by the latter to cover unanticipated financial shortfalls.
Three years is also long enough to render some judgments on NBN Co. One takeaway is that NBN Co has issues meeting milestones, even those it sets for itself. While it is still early in this vast project, and there is usually a plausible excuse, NBN Co has missed or is on track to miss the vast majority of its targets for revenue, customers and network construction over the first two years of its 2011-2013 Corporate Plan
Proof are the many thousands of Australians in new housing developments currently without broadband because the Government and NBN Co dithered on their plan for these areas, and NBN Co then was unable to execute when given responsibility.
The NBN Co Corporate Plan states 132,000 premises in Greenfield estates will be NBN customers by June 2012. But according to the Government’s six-monthly progress report to Parliament last week only 110 such premises had been connected by 31 December 2011 – while developer applications covering 109,988 premises were parked on a waiting list.
We should be cautious about projecting NBN Co’s future from early stumbles, but the stumbles don’t inspire confidence it would be able to complete the rollout nine and a half years from last month, which is what’s been promised.
Sticking to the same area, a final observation about what we’ve learned in the past three years involves our rapidly evolving understanding of NBN Co’s character and likely behaviour.
There are still a few people who argue that compared to Telstra as incumbent, NBN Co is a gift to the industry from a benevolent Government; a civic-minded (and rather better-paid) reincarnation of the PMG committed to peace, love and open access to world-class networks on equivalent and transparent terms.
They obviously didn’t have to negotiate with it over the wholesale broadband agreement.
On the contrary, NBN Co is the new monopolist in town and Michael Quigley has signalled his readiness to use whatever leverage is available to reach the goals he and his team have been set by the NBN Co Board and Shareholder Ministers.
In this vein over the past 18 months we’ve seen NBN Co test the boundaries of ‘wholesale only,’ attempt to block regulatory oversight in as many areas as possible (from points of interconnect to capital spending), and even float (then hastily abandon) an audacious plan to seek the ACCC’s consent to real price rises of up to 5 per cent annually for most NBN products.
We should expect to see more of this. And if you don’t like it, don’t blame Mr Quigley or his executives – blame the NBN Co Board and Shareholder Ministers.
How On Earth Did We Get Here?
If Labor’s commitment to a version of the NBN that will cost more and take longer to build than any conceivable alternative sounds like a risky and potentially ugly situation to you, given the backdrop I’ve just described, then you are not alone.
Percy Allan – a respected former Treasury Secretary of NSW who served two Labor and two Liberal premiers, and is not prone to public hyperbole – recently bluntly described the NBN as a potential financial disaster if take-up levels do not meet its projections.
At the heart of the issue is a very basic problem: Between December 2008 and April 2009 the Labor Government decided to build a fibre to the premises network to 93 per cent of the population with barely any analysis, and without evaluating any other options for achieving its implicit objectives – which seem to have been structurally separating Telstra and ensuring all Australians receive access to very fast broadband. This decision was presented to the public as a fait accompli.
Since that time various reports and analyses have studiously ignored whether NBN 2.0 is the best approach – they have simply sought to provide an ex post facto rationale for it.
Along the way the Government has locked itself into three pledges about the network which seem incapable of simultaneous realisation:
First it decided to extend FTTP to 93 per cent of Australian premises. That is a vastly wider coverage (1.5 – 2.5X) than next-generation networks in comparable countries. As a consequence the NBN is the most expensive broadband project in the world.
Labor has also promised prices for entry-level packages comparable to current ADSL2 pricing. This will be the most popular service rung for the first decade of the NBN – so this is an implicit cap on potential revenue.
Finally, Labor said taxpayer funds invested in the NBN will earn 7 per cent.
It’s the combination of these commitments that is motivating the more totalitarian aspects of Labor’s NBN, such as the determination to utterly stamp out rival infrastructure, the pressure on the ACCC to sign off on investment plans and pricing ‘flexibility,’ and the ham-fisted attempts to stop Telstra and Optus marketing 4G/LTE to departing fixed line customers.
If one assumes the one pledge here which cannot be broken is to keep entry prices the same as today, then either the NBN design must be changed to reduce capex or a large chunk of capital will have to be written off as a social investment.
Of course the Government and NBN Co don’t accept there is a problem, and claim the 2011-2013 Corporate Plan released in December 2010 proves it.
It forecasts NBN Co can meet all three objectives by spending $37.3 billion to completely build out its FTTP, fixed wireless and satellite networks, migrating over 9 million customers, and earning enough revenue during the rollout to cover its operating expenses (for 2010 to 2021, revenues are projected at $23.6 billion and opex at $23.1 billion.) The levered funding requirement for NBN Co peaks at $40.9 billion in 2021.
That’s a neat story. But there are numerous discordant notes in the Plan. Let me address them, and in each case try and draw out a lesson for the Coalition’s policy.
Revenues per user
At a time when fixed line voice and broadband revenues per customer are steady or falling in nominal terms in most advanced economies, and fixed line revenue to GDP is sliding, NBN Co’s financial projections forecast sustained real growth for this part of the telecommunications market in Australia.
Perhaps scepticism about such heroic projections are why all but a dozen paragraphs and a handful of charts were blacked out from the 14 pages an obviously intrigued Greenhill Caliburn team devoted to revenue when it reviewed the Corporate Plan for the Government.
If we make some rough assumptions about what Telstra’s wholesale fixed line revenues might be today if its fixed line network was in a separated group, and about economic growth over the next decade, we see NBN Co’s business plan is essentially predicated on wholesale fixed-line revenues remaining at their current level of approximately 0.25 per cent of GDP until at least the mid 2020s.
To make this happen NBN Co intends to increase monthly wholesale ARPU from $37 in 2012 to $64 in 2023, a rise of more than 6 per cent annually, with the CVC (a charge for usage volume) driving the rise . That compares with a real fall of 7 per cent per year in the cost of services over Telstra’s copper network since 2000.
So Australians who have been used to paying less and less for broadband are now going to find themselves paying more, in both real and nominal terms.
That seems at odds with empirical evidence suggesting consumer spending on communications services (defined broadly to include pay TV, postal services and so on) has for the past decade been bounded as a share of total expenditure, and is increasingly being diverted to mobile services.
If NBN is able to use its market power to hold fixed line wholesale revenues steady where they otherwise would have fallen, that means either less growth for other types of products or services, or narrower margins for retail service providers.
The great imponderable for all next-generation fixed networks is revenue; nobody knows what new applications might emerge and what consumers will pay for them. This is why so many incumbents are choosing the path of FTTN and putting less capital at risk
To be candid, any Coalition broadband plan will face similar uncertainties about the level of future demand – which is why deferring expensive and irreversible commitments such as FTTP, to the extent that this is possible, is simply prudent management, not a bloody-minded refusal to face the future as our hang-the-cost critics like to allege.
Labor’s alternative strategy for mitigating demand-side risk is to invest anyway, but create an option of using market power to raise prices and recoup costs if this proves to be necessary. That is why there must not be any fixed line rivals.
If NBN Co does break the glass and press that button, nobody should delude themselves that regulation or regulators can prevent prices rising. As Optus and a few others have pointed out, a careful reading of NBN Co’s SAU reveals that it to offers virtually no binding guarantees about prices (despite the claims of NBN Co’s devotees). Don’t take my word for it, read the most recent Optus submission about the SAU to the ACCC.
Likewise, Ovum’s David Kennedy has expressed his view that price controls would be scrapped in the blink of an eye if NBN Co financial sustainability was at stake.
So in the end the NBN’s key protection against revenue uncertainty given its huge capital commitments will be its ability to gouge consumers down the track. No one should be in doubt about this.
ARPU or revenue per user is only one part of the revenue equation. The other is the number of users who subscribe out of the population which could.
Again, NBN Co has been fairly aggressive in its estimates – after empty premises are accounted for, it only expects 13 per cent of households and businesses to go mobile-only by 2021. In parts of the US the ratio is reportedly 25 per cent already.
Clearly one driver of take-up will be the extent to which NBN Co is required by its shareholders to use its market power to maximise returns. Higher prices will reduce penetration. Price elasticity for various groups of consumers will determine by how much, but if 4G/LTE is as competitive with fixed lines (and therefore becomes a realistic substitute for more users) as carriers such as Verizon expect, the fall could be substantial.
Another significant factor in Australia will be that households do not watch most TV over fixed line connections, unlike the US or parts of Europe, and this is likely to continue to be the case. The economics of the NBN appear likely to be problematic for pay TV for some time. Ironically this means FTTP is even less necessary for most users, since video is its main application.
Indeed Cisco’s forecasts of increasing consumer demand for bandwidth rely ever more heavily on video traffic. In the words of its annual forecast, released in June 2011: “Internet video is now 40 per cent of consumer Internet traffic, and will reach 61 per cent by the end of 2015, not including…video exchanged through P2P file sharing. The sum of all forms of video will continue to be approximately 90 per cent of global consumer traffic by 2015.”
There is nothing wrong with video traffic, but we to ask whether building a network with unique capabilities to carry video to consumers is more important to Australians than rectifying broadband black spots, or achieving ubiquitous national coverage earlier than the 2020s.
NBN Co’s evident enthusiasm for a regulatory regime that permits it to fully recover capital it invests from users is what finally set off alarm bells for other players, and rightly so. From the outset the Coalition has made the point that a heavily capitalised monopoly is not going to reduce prices – the reverse is more probable.
The high cost of NBN’s network largely reflects the choice of FTTP and its rollout beyond the 50-60 per cent or so of premises that are commercially viable in typical markets.
The McKinsey cost curve tells us that getting from the 61st to the 93rd percentiles of Australian premises ranked by cost-to-pass accounts for almost half the $28 billion NBN has budgeted for fibre (few outside NBN Co expect the FTTP rollout to cost that little.
In Europe or North America the experience over the past decade has consistently been that the capex for FTTP is about 3 to 4 times more per premise than the capex required for FTTN or FTTC network where the fibre terminates a kilometre or less from the customer premises.
In the United States, for example, the two leading telcos – Verizon and AT&T – faced a competitive threat when cable companies started to offer broadband and voice as well as pay TV over their HFC networks around the turn of the century.
Verizon chose to build an FTTP network (FiOS) to selected markets – its rollout eventually covered 54 per cent of its legacy copper footprint. AT&T built an FTTN network (uVerse) to a larger part of its territory. Each proceeded to offer similar triple play packages of video, broadband and voice. At present each has a similar ARPU of around $US145 for users on their next-generation network. But AT&T’s cost a third or less to build.
Speeds higher than the bulk of NBN users will purchase prior to 2025 are today available from FTTC – in the UK BT offers 80 megabits per second down/10 megabits per second up – and new technologies for squeezing more out of copper are being developed all the time.
The key lesson for Coalition policy is abundantly clear – it is critical to extend fibre further into the field, but defer the cost of going all the way to the premises. If an FTTx approach can reduce copper loops to under 1000 meters, there is no reason to expect any material reduction in potential revenues available at fair market prices compared to an FTTP rollout.
But there are other takeaways from a review of the NBN’s approach and FTTP or FTTx deployments in other economies.
One is the importance of focusing resources where they are needed – and not overbuilding modern infrastructure for the sake of ‘neatness’. Yet this is what we see confirmed by a 28 March entry on NBN Co‘s rollout blog which describes work preparing for rollout of the NBN in Crace, in the ACT.
What the NBN Co’s intrepid blogger fails to mention is that Crace is already served by TransACT’s FTTP network – while literally millions of households in other parts of Australia have been waiting patiently for years for upgraded broadband.
Another takeaway is that ‘inside the house’ costs for FTTP can mount and become a serious economic factor. These costs tend to fall over time, and vary from place to place, but in the UK ‘inside the house’ expenses have been estimated to be 20 per cent of total FTTP costs.
The Telstra fibre rollout in South Brisbane likewise indicates this expense is likely to be material – Telstra this week confirmed industry reports that it is using up to a day of technician labour at each residential premise, adding up to $1000 to the cost of the cutover.
It is not clear NBN Co has budgeted for this expense anywhere. If it has, it would be instructive for someone from NBN Co to explain where and how.
If not then presumably RSPs will be expected to carry it (in which case customers will end up paying through installation charges). It is difficult to see NBN Co avoiding a contribution or subsidy, given the potential impact on take-up rates of thrusting this expense onto service providers already facing reduced margins.
A final observation on investment is that NBN’s heavy capital spending will not cease with the rollout. Both the Greenhill Caliburn review and NBN Co’s Corporate Plan show its completed networks require remarkably high ongoing capex to sustain them, presumably reflecting upgrades of the active electronics and in due course replacement of the fibre.
By the 2020s, the Corporate Plan shows ‘replacement’ capex escalating to around $1.5 billion annually – double most estimates of the current cost of maintaining Telstra’s copper (the most costly and remote parts of which will still have to be maintained at that time). Greenhill Caliburn estimates in total $14 billion in ‘replacement’ capex will be incurred by NBN Co by 2028 for a network that by then will only on average be about ten years old.
Analysis of the 2012-2021 period (where NBN Co has made public more detailed projections) show that ‘replacement’ investment averages an extraordinary $115 each year for each premise with an active service. So much for another frequently repeated NBN myth – that FTTP will drastically curtail network upkeep costs.
Compared to capex and revenue, opex projections for the NBN should have been a breeze. Yet as I pointed out in December, NBN Co’s current Corporate Plan shows $23.6 billion in operating expenses over the period 2010 to 2021, with approximately $15 billion of that identified as payments to Telstra.
This is a puzzle, because when financial market analysts work out the payments that will be made to Telstra and Optus over this period in return for migrating their customers and shuttering their networks, and for NBN Co’s leases over Telstra infrastructure, they come up with considerably higher nominal figures – usually around $6 billion higher. By their reckoning, if the rollout is on time, NBN Co will have to pay Telstra about $19 billion and Optus about $2 billion in round numbers.
We don’t know the exact actual timing of these payments because they will depend on the rollout schedule. But if NBN Co then needs a further $8.5 billion to cover ‘real’ direct and indirect opex, as the Corporate Plan projects, then there is a $6 billion unexplained hole.
The rest of opex is peanuts beside the Telstra and Optus payments. But investigation of this area does yield a further lesson – and that is to be careful believing the hype of NBN devotees.
A key advantage often cited for FTTP over FTTC/VDSL2 is reduced energy usage, given GPON is not powered. But in practice this saving is minimal compared to the vast gap in capital costs. According to Verizon, annual central office power usage is 32 kWh for a DSL line and 12 kWh for a FIOS line.
At 25 cents per kWh, a move from copper to FTTP therefore cuts annual energy costs per line from $8 to $3.
Across the 8 million premises forecast to be connected to the fibre when the NBN Co rollout is complete, FTTP therefore saves $40 million a year in power costs. That isn’t meaningless, but it is a solid $15 million less than the sum NBN Co spent on lawyers last year (and just 2 per cent of the annual interest on $28 billion).
Over the past four and a half years Labor has had two entirely different broadband policies. The first at least vaguely resembled the ALP broadband policy offered at the 2007 election.
(When we take a look back to that time, it turns out Senator Conroy released 578 words on the substance of that policy prior to the election, or 1271 words if we count his description of how it was going to be funded by raiding the Future Fund – how ironic to hear his constant calls for more detail from us!).
In the past three years, over a billion dollars has been spent and another billion committed to deliver NBN 2.0 to about 5000 premises across Australia, only half connected over fibre.
Julia Gillard and Senator Conroy now assure us the rollout is poised to go into “high gear”. Very soon, if we believe them, Simon Hackett will no longer be able to speak of the “Nearly Begun Network”.
Having passed 3 premises per working day with its fibre rollout between July 2011 and March 2012, NBN Co must now smoothly accelerate to the 6000 premises per working day it needs to deliver on its schedule and the Government’s promises to the electorate.
But even if it does so and adheres to its existing schedule and budget, which I doubt, the economics of the NBN will be tenuous for years to come. And if there is any slippage at all, they will become disastrous.
To me that makes this a very risky project. The Coalition’s policy is, as you know focussed on achieving a comparable outcome (ubiquitous very fast broadband) but achieving it sooner in terms of rollout, cheaper in terms of cost to taxpayers, and more affordably in terms of consumers. All of that follows from taking a pragmatic and technological neutral approach.
But above all, at the front of our priorities is reducing risks for taxpayers and risks for consumers