Thursday, 8 November 2012

Fiber scepticism = climate change denial?

Polar bear annoyed by fibre sceptics
Paul Budde, the Australian telecoms analyst and FTTH enthusiast, has a feisty blog post up. In it he draws parallels between fibre-sceptics and climate change deniers, saying they're both mud-slingers who get far more attention than they deserve.

Paul complains: "So far we have not come across any NBN sceptics who have dared to move into the area of what the extent of the NBN’s potential – the digital economy, intelligent infrastructure, big data, clouds computing, e-health, e-government, smart grids and so on".

This is a surprising allegation, since there are plenty of people out there expressing doubts about these beneficial externalities, and I know Paul has read at least one such paper - the one I wrote (together with my brother). This addressed exactly the issue of superfast's potential, and spent some pages on why its contribution to e-health and smart grids has been much overstated. I certainly don't expect everyone to have read this paper, but I happen to know Paul did read it - he wrote quite a long response to it. Still, perhaps it wasn't interesting enough to stay in his memory.

Never mind, let's move on to some of Paul's other points. I'll set aside his unsupported assertion that less than 5% of commentators feel the NBN is a waste of money. His key points (paraphrased) seem to be:
  • 'We need to embrace the digital economy'
  • 'The NBN could create of millions of new jobs'
  • 'It will reduce healthcare costs by $30 billion and energy costs by $2 billion'
  • '119 countries around the world have national broadband plans, and everywhere but in Australia there is bipartisan support'

The Digital Economy
Certainly we need to embrace the digital economy - it's a bit of a motherhood statement. The critical question is whether we need superfast broadband to do so. The US, the economic heart of the internet and home of most of the major internet application companies, doesn't actually have particularly good broadband infrastructure. Akamai ranks it 9th worldwide for average connection speed. The UK has the largest internet contribution to GDP of any G-20 country, but is only just above the middle of the G20 pack in terms of average connection speed (among the countries tracked by Akamai). Thus there seems to be very little connection between broadband speed and strength of the digital economy. Rolling out fibre is simply not a magic economic wand.

New jobs
Paul's suggestion that the NBN could create "millions of new jobs" is certainly a bold claim to make without supporting evidence - total employment in Australia is a little over 11 million, so to add 'millions' by improving broadband speeds would be quite something. Paul's claim is all the bolder when you realise that of the 11 million, 3 million work in sectors such as agriculture, mining, public admin & safety, and health care & social assistance, all areas where the NBN is unlikely to create many new jobs.

Cost savings - healthcare
Indeed, if Paul is to be believed, there may in fact be job losses, since he tells us there will be huge cost savings in healthcare and electricity. Paul hasn't provided us with sources, so we'll need to do some guesswork in looking at these. He's been using the $30bn healthcare saving number since at least 2009, and I suspect it's from a 2007 KPMG report on the cost-benefit of electronic health records. This is not itself available online, but it underpins a key November 2008 Booz report on the potential of eHealth (see p35), for National Health & Hospitals Reform Commission.

The vital point here is that the number is not the benefit of NBN, but rather the benefit of eHealth. The vast majority of the benefits of eHealth do not depend on superfast to the home. They can be achieved via good broadband to medical premises (which generally already have it), and basic broadband to homes. Indeed, while the Booz report highlights the importance broadband, it absolutely does not highlight fast universal broadband as a requirement for the eHealth benefits. Indeed, it discusses Germany's substantial eHealth plan based on DSL. To say the $30bn eHealth benefit depends on NBN is simply wrong.

Cost savings - energy
Let's have a look at the $2bn energy cost saving. Again, Paul hasn't provided a source for it, but it is a number he's been using since 2010, when he said it stemmed from "savings from a duplicated comms network and double installation". Even then he didn't provide a source, so we can't directly review its basis.

However, there are plenty of figures out there for the cost of installing smart grids. Enel, the Italian electricity company, has installed 33m smart meters, at a unit install cost of €13 (derived from this). Scaling this to Australia's 8.7m households, we get a total cost of A$136m. While it's plausible that some joint roll-out of the NBN and smart meters might save some of this, clearly a percentage of $136m is not going to make much of a dent in the purported $2bn saving.

So perhaps most of it comes from avoiding a 'duplicated comms network'? The trouble is that smart grids absolutely do not require the bandwidth of NBN. They can work perfectly well on existing telecoms networks, and indeed the successful Enel network (completed in 2006, long before superfast broadband) does exactly that.

International bipartisan support
This leaves Paul's comment that there are "119 countries around the world [that] have national broadband plans, and everywhere but in Australia there is bipartisan support". Deliberately or otherwise, this statement gives the impression that Australia is in good company with its broadband plan - nothing could be further from the truth.

Firstly, in renationalising a substantial part of the telecoms indusrty, Australia has taken a highly unusual step. The great majority of countries (including all other OECD countries) that have chosen to support superfast have done so via commercial players. Even those that have taken parts of the network into government hands have done so on a far more limited scale.

Secondly the NBN is an extreme outlier in terms of the level of government investment:

Government spend from OECD

Given that the Australian government is spending so much more than other governments, surely it's not unreasonable to focus on costs, and to ask if the benefits will really outweigh them?

Oh, and for the record, I am a believer in climate change!


  1. I welcome debate on the NBN Robert and yes you are right I did read your report with interest and also commented on it.

    As late as June this year the Opposition Leader Tony Abbott repeated for the zillionth time that the NBN is a ‘waste of money’, a bit of an approval on his previous lines ‘I will kill the NBN’. The Shadow Treasurer is on the record for many similar comments. If that doesn’t mean that they are NBN deniers than so be it, give it another name.

    But in particular you are questioning me on the social and economic benefits. There is no way that anybody can put any exact or even indicative figures on this. Who would have predicted 3 or 5 years ago that the smartphones would have such an impact on the need for higher speed broadband with now already more than 4 devices linked in the home to the fixed wireless connection.

    Did we know when electricity was introduced for the production of light that this would lead to the industrial revolution with motors, fridges, computers and so on? Should we have waited till somebody invented those devices before we start building that infrastructure? Natural gas changed the way people cooked and heated. Would Google, Facebook, Apple, Amazon have become such giants without having a broadband infrastructure in place and are those developments not guiding us toward what else we can expect if we upgrade the infrastructure?

    All of these developments have created massive productivity gains, changed people lifestyle and have greatly contributed to the society and economy that we now have. Did those people who took the bold decisions to start rolling out these infrastructures know what was laying ahead of them, I don’t think so.

    Didn’t all of these developments not created millions of new jobs? You can question if broadband falls into that same category, but that would than clearly separate my vision from yours.

    I have not come across any serious arguments that indicate that online alternatives are not going to save costs in doctors and nurses – patient time, or travelling time from patients. Monitoring old age people or chronicle ill people from there homes is again an area where I believe not many people are questioning the benefits of online alternatives.

    Smart Grids are part of the M2M revolution were literally billions of sensors and devices are being connected, they produce data that needs to be stored, processed and analysed in real time. This requires reliable, low latency, secure, affordable high capacity networks, like mobile networks eventually they require fibre very deep into the neighbourhoods.

    On the international side, I only wanted to indicate that now 119 countries have committed that they will develop broadband policies for all of the above mentioned reasons.

    You are mixing this up with how you than actually do this. In my work for the UN I would be foolish if I would use the Australian example as the only way to do this, there are many ways to skin the cat. Is the Australian model ideal, no I don’t think so.

    However, the current rhetoric re a cheaper NBN and a faster implementation without linking that to some first principles is in my opinion similar to saying I can build roads cheaper if we simply create dirt tracks.

    I not just criticised the Opposition but also the current government for not having clear policies on this and I have indicated that the current NBN Co business plan is flawed:

    There certainly are ways to save costs and I have also at numerous occasions indicated that using the current HFC and ADSL2+ networks longer would be a better solution. As long as any technology plans include policies for the upgrade of these networks to fibre when demand becomes more apparent:

    Once again thank for reacting to my article and for your valuable contribution to the debate .


  2. Paul -

    Thanks very much for taking the time to respond, but it seems to me your note contains the very same issue that caused me concern in your original post - that is, failing to distinguish between the benefits of superfast and the benefits of basic broadband.

    For example, you say "I have not come across any serious arguments that indicate that online alternatives are not going to save costs in doctors and nurses". I agree with you - but the critical question re NBN is will superfast speeds bring incremental benefits. If they don't, then eHealth is (largely) a red herring as far as the NBN debate goes.

    It may be true to say that broadband has created millions of jobs (though I would strongly doubt it has done so in Australia alone). More generally, I am a huge believer in the benefits of broadband. But even if we accept that, there is simply no reason to believe that superfast will do the same.

    Cars have been a huge boost to the economy and society, but that doesn't mean Ferraris are. If I were to buy a Ferrari, it would have various benefits - it would get me to the shops, and might even have a handy lick of speed to get us to Maternity when my wife goes into labour. But that doesn't mean that buying a Ferrari wouldn't be a huge waste of money.

    Of course this is just a metaphor - it doesn't prove that the NBN is a waste of money. However, it does suggest that it is overly simplistic to say that because broadband brought benefits, superfast must bring substantial incremental benefits.

    Your examples of the applications supposedly enabled by broadband are instructive - Amazon was founded in 1994, Google in 1998. iTunes was launched in 2001, Facebook in 2004. Broadband wasn't launched until 1998. Even by 2001 there were only 8m broadband households in the US.

    By coincidence, there are about 8m FTTH households in the US today. Of course, as far as superfast goes, there are many markets well ahead of the US, such as Korea and Japan.

    In other words, FTTH penetration is already well past the level that broadband penetration was when many of the giants of the broadband era got started. It is one of the reasons to be very worried about the returns on superfast investment that there seems to be a dearth of applications that have developed that really need that level of infrastructure.

    People say of FTTH 'Build it and they will come'. We already have built it in many markets, but no-one (or no key applications) have shown up.

    Thanks again for commenting, and for pointing me to your comments re HFC and ADSL2+, which I hadn't previously seen. On that at least we are in agreement!


  3. The benefit of so-called "superfast" over "regular" broadband is that with more road lanes, you can support more traffic. That is, you can send more data, more reliably, more consistently and with less latency. Any industry that involves the sending of any sort of information will be revolutionised, to a lesser or greater extent.

    Let's use a different analogy. You could consider it the same way you could consider money. That is, internet bandwidth is a currency. Someone can live on $50 a day, no doubt, but with $500 a day, what more could that person do? There's any number of things that person could do, but it's up to them, their own needs, wants and imagination. Can you honestly claim that person's quality of living wouldn't rise?
    With inflation, prices rise - so in the future there's no way that $50 a day will be enough. With the internet, it's obvious that bandwidth requirements are rising and will continue to rise, especially for multi-device households, so we will need greater bandwidth in the future (if not already).

    The other thing is that you need "equality" or consistency to gain the real benefits of the internet. If someone (on fibre) is able to upload at 10Mb/s, but someone (on ADSL) is only able to download at 1Mb/s, then clearly there's a bottleneck at the latter end. With FTTH for the vast majority of the population, that bottleneck is widened, and so the potential of FTTH can be truly realised. So saying that "this one person already has FTTH but we haven't seen the benefits gained from that" ignores the fact that "everyone else is on ADSL". It's a worthless argument.

    But you can't just look at the bandwidth alone. You are restricting your vision. You also have to consider the inherent stability gained with fibre. For example, there are elderly people living in residential homes; if someone were to have a heart attack they could press a buzzer, alerting a helper who would be able to provide prompt assistance. With an unstable DSL connection, that alert might not get through immediately or even at all, and the consequences of that are fatally obvious. Connection stability is not just desirable, it is necessary. With a century-old copper connection that has degraded significantly, it's impossible to guarantee that stability. With a brand new fibre optic network, you can largely guarantee it.

    So you can't just look at the average person and how they might not realise the full potential of fibre optics (they might just see it as a way to download movies faster), you HAVE to look at the people that might benefit from it most. And if you can see them, then a compelling argument for FTTH simply becomes clear.

  4. commensurate -

    Thanks for taking the time to write extensive comments.

    Much of what you say seems to be based on an assumption that more bandwidth must be good. But in lots of areas we (society) take the view that there are diminishing returns.

    We don't invest much these days in building faster airliners or cars. We build six-lane highways but rarely twelve-lane highways and so on. The key question is when does the incremental cost of bandwidth outweigh the benefits?

    As you say, FTTH is not yet widespread in some countries, and it is possible that critical mass is needed to see the benefits. But it is widespread in a number of places, such as Korea, Hong Kong and Japan. And yet, those places don't seem to have generated widespread socially beneficial applications for that bandwidth.

    Re your medical alert application, that most certainly does not need FTTH. I know people who use it without even broadband - it generally runs off the traditional voice network. That network has been used for decades for 911/999 services without anyone sounding too worried about its reliability.

    Thanks again for your comment,


  5. Hi,

    I could turn that argument upside down: much of what you say seems to be based on an assumption that more bandwidth is not beneficial and will never be beneficial. It is safe* to assume that, given current trends, more bandwidth IS beneficial.
    When I say safe, I'm talking from a risk management perspective. Can Australia afford to be left behind, especially when other countries around the world are beginning to embrace fibre optics? Will Australia always be the Lucky Country, or will we have to grasp our fortunes with our own hands?

    The incremental cost of, say, FTTH over FTTN is small, if not negative (it depends on the timeframe you look at - the cost of active power, climate control and maintenance of the eroded copper network over time). Recently, there was also that article about how, as contractors' skills, experience and processes have improved, it actually costs less to install fibre in the last mile than it will to install roadside cabinets.

    The benefit of FTTN over ADSL(2) is small to negligible, and would NOT justify the cost. The benefit of FTTH over FTTN is extremely large, and would future-proof Australia for perhaps a century. FTTN has a speed of "up to" 80Mbps... if you're right next to the node. While now the highest speed of FTTH is 100Mbps, the current theoretical limit of a single fibre strand is 10Gbps, it will just require an upgrade of hardware but not a replacement of the fibre. And it is well-established that FTTN is, for many countries, simply a temporary stepping stone on the way to FTTH. It would be madness to insist on building the intermediate first when the cost of directly building the end result is incremental and the benefit is exponential.

    While the cost per capita might be larger, the cost per unit of land area is actually smaller in Australia than in other countries around the world. When you're rolling out infrastructure, I think cost per unit of land area is a more appropriate measurement - it's about cost efficiency. I mean, if your argument is based on cost per capita, you would be more justified arguing against those rural folks that only make up 4% of Australian premises being serviced by billion-dollar satellites. But, of course, the NBN is also about bringing about a baseline equality of service.

    South Korea, Hong Kong and Japan are technological and/or economic juggernauts. I'm not saying that there's a direct cause-and-effect, but if you give people the right tools...

    The medical alert application was used to demonstrate the benefits, or the necessity, of a stable connection. Unless you replace the copper, it will not have a stable connection for the foreseeable future. But if you're going to replace the copper, why not simply use fibre which costs less and is of more benefit for longer?

    The only criticism I can have about the NBN is that it is unnecessarily servicing more fortunate areas (i.e. with good ADSL2 connections, or HFC, etc.) while other areas are stuck with a poor wireless connection, a poor copper connection, or no connection at all. However, I also recognise that it is likely a technological and economic imperative. Building a network has to be holistic - you can't just pick and choose locations, that would be grossly inefficient.

  6. Commensurate -

    I'm not saying bandwidth will never be beneficial. What I am saying is that the applications with wider social benefits (that is, the ones that might justify government intervention) don't need FTTH speeds. I didn't set out all the reasons why I think this in this post, but have a look in the 'papers and presentations' section if you're interested.

    I would say other countries are shifting their emphasis away from FTTH. The US is rapidly decelerating its FTTH roll-out, long before widespread coverage has been achieved. In Europe the emphasis is very much shifting to FTTC approaches.

    Do you have a source for the claim that it's cheaper to roll out FTTH than FTTC? That's the first time I've heard that, and certainly very many sources think FTTH is roughly 3x more expensive than FTTC.

    The evidence is that FTTC does deliver extremely close to its advertised speeds. Have a look at the Ofcom stats, where actual customer speeds were within a Mbps or two of the advertised rate, thereby giving a substantial uplift over ADSL2+.

    You say that rolling out FTTN is madness - it may be, but that is exactly what most of Europe is doing right now.

  7. Rob,

    You say "don't need", but that is a personal value judgment and it is also based on the assumption that things will not change. The bandwidth limit of FTTH (10Gbps) is not even in the same ballpark as that of FTTN ("up to" 80Mbps). And yet the increased investment of FTTH over FTTN is small if not negative; as long as you don't neglect the significantly larger upkeep required for FTTN (power, climate control, maintenance), i.e. the operating cost. No commercial enterprise ever neglects the operating cost over the lifespan of a project in its cost estimates or its determination of profits, and I think it would be foolish of us to. I

    "that is, the ones that might justify government intervention"
    It seems that our primary disagreement is ideological. I believe that some natural monopolies exist, and it is in a country's best interests if that monopoly was publicly-owned, or regulated to the extent that the public best benefits. Base infrastructure such as the physical communications network is one such natural monopoly. You talked about diminishing returns earlier; a similar concept is "economies of scale": the total cost of one large project is smaller than many small projects. Additionally, "doubling up" on infrastructure is a waste of space, a waste of labour and a waste of materials - all premium resources; it is not only costly, not only unsustainable, but it does not benefit the consumer or society in the least.

    While competition is good, it is the means and not the end; the end is that which benefits the consumer, and society, most. With the NBN, the competition will be at the retailer level. So to begin with, the justification for government intervention is to clean up the mess of our lazy communications industry and enable consumer-beneficial retailer competition. I mean, alternatively, the government could perhaps nationalise all telecom infrastructure... but that would make a lot of people very unhappy. So the government is taking a different, better approach: building its own, superior network (especially as existing telecoms companies are cherrypicking or refusing to) and levelling the playing field.

    Now, on FTTN/FTTC vs FTTH/FTTP, it comes down to three things: The state of the copper, the simple maths, and the accounting.

    The original copper is a century old, and the infrastructure is barely maintained - these days technicians are tasked with fixing the symptoms rather than eliminating the causes.

    The increased investment of FTTH over FTTN is small (especially over, say, a 20 year timeframe), but as said, the increased /potential/ benefit of FTTH over FTTN is large. Whether "actual" benefit approaches potential benefit is debatable, but I would argue that it's all relative to individual needs. Indeed, should FTTN be built initially, FTTN will need to be upgraded to FTTH in the future, but it would obviously cost more to do two upgrades in succession than one big upgrade, because: not all the infrastructure would carry over; unnecessary infrastructure would need to be removed; upkeep of FTTN is higher than that of FTTH;

    Finally, it comes down to accounting. "Cost" is only relevant versus the revenue made. In other words, return on investment. The NBN is being built using borrowed money, and will repay itself over its lifespan. Cost is therefore not an issue, as the return on investment, given conservative predictions, remains positive. The borrowed money that is spent on the NBN cannot be spent elsewhere, because that money could not be repaid; and I'm sure you understand this, so our arguing over costs effectively comes down to cost vs potential benefit, in which case FTTH is the clear winner.

  8. Regarding the reference of FTTN vs FTTH costs, I'm sure you're already aware of it:
    Could you provide sources that show that FTTH (GPON) is roughly 3x more expensive than FTTC? Could you provide sources that do not ignore the upkeep cost of the network over its lifespan?

    I have not seen the Ofcom stats, but as we are talking about FTTN with its advertised speed of "up to 80Mbps", are you making the claim that actual customer speeds of FTTN will be between 78-80Mbps?

    On your final point: If Amy jumped off a cliff, then Eugene jumped off the same cliff, should Audrey, having witnessed these two, jump off the cliff as well?

  9. For FTTH vs FTTC costs, WIK (a highly respected European consultancy) say that FTTH costs roughly 3x as much in Germany (p20). Ofcom, the UK regulator, says the costs per year are 3-4x (p4).

    You say "No commercial enterprise ever neglects the operating cost over the lifespan of a project in its cost estimates or its determination of profits". Can you explain then why European carriers are (to a very large degree) choosing FTTC over FTTH, if FTTH is, as you claim, cheaper?

    Re FTTC speeds, according to the Ofcom stats (p11), the BT FTTC customers on 'up to 76 Mbps' are actually getting 66 Mbps. Not the entire amount advertised (and not quite 'within a Mbps or two' as I had remembered), but enough for more than a dozen simultaneous HDTV streams, which should be sufficient for most households for many many years to come.

    Moreover, VDSL technology is very rapidly improving (BT only offered this tier a year ago), and it's reasonable to expect that these speeds will improve shortly. Much higher speeds are being trialled.

    The RoI on NBN Co is only positive if you assume Australian consumers are prepared to pay more every year for their broadband. In practice most countries have had to price superfast at or even below standard broadband speeds in order to get meaningful take-up. This is what's happened in Korea, Japan and Hong Kong.

  10. Well, I can't argue with those numbers, but I do wonder as to two things which have not been specified: The time frame considered, and whether operating costs have actually been included in the "cost per year".

    Sure, that is easily explained. The time frame considered by a corporation is shorter, and the payback period must be much shorter, than for a public infrastructure project. Corporations generally don't make investments that take 20 years to make a modest return, you see. From a business perspective, you can "milk" the most out of your customers by offering them an incremental, higher-value upgrade now, and letting them bear the cost of future operating costs and maintenance. And this explains precisely why the market has failed, especially in Australia: companies cherry-pick dense urban, affluent locations to provide higher-value services, in order to generate the most profit. The NBN, thankfully, isn't about generating the most profit, but about providing the best possible service to all Australians.

    How "many" many years to come? Will it even be sufficient until 2020, when the NBN is scheduled to finish completion? I doubt it will be sufficient for major applications to 2016. And at what price? The NBN offers a guaranteed speed (100/40) greater than the advertised speed of FTTN, at a wholesale price of $38 a month.,nbn-co-wholesale-prices-to-start-at-24-a-month.aspx
    And I'm sorry, but your "should be sufficient" belongs in this list:

    Let's imagine that VDSL technology is improving, and the gains are relative... Well, so is, I imagine, FTTH technology. For the same relative gain, say, the absolute gain in FTTH performance must be greater than that for FTTN performance. Under no circumstance is FTTN able to compete with FTTH in any performance benchmarks.

    Except that... they won't have to pay for separate telephony services, so in most circumstances they would not be paying more. Additionally, the NBN (internet and voice) will be competing with mobile (internet and voice), not fixed-line, since the copper will be decommissioned, so anyone who needs more than 3GB a month and/or a reliable connection free of congestion must opt for the NBN option.

    There are also two other benefits that FTTH provides over FTTN - the much better upload speeds, critical to high quality video communications and business functions, and the fact that it does not use the existing copper that is well past its use-by date. And the NBN is not just about broadband internet, it's about completely overhauling the nation's telecommunications infrastructure, including telephony services.

    FTTH is the cutting-edge, but it is not new technology, it's tried-and-true. If Australia had already embarked upon the path to FTTN, then there might be some justification in pursuing FTTN to its end (the cost of changing the approach and the physical network may indeed be immense), but as Australia is essentially beginning from scratch and looking towards the future, FTTH is the best solution going forward.