Telecommunication Journal of Australia Article: Broadband: what is the problem to be solved, and is NBN solving it?.

Synposis:

A striking feature of the Rudd-Gillard Government’s NBN policy – to spend over $50 billion of taxpayers’ money on a 100 Mbps fibre to the premises network serving over 10 million premises – is the lack of clarity about the problem this policy is designed to solve. In this article I argue that the NBN was announced for political reasons rather than policy reasons.

After briefly explaining that point, the paper then discusses the policy issues – by identifying and assessing several 'candidate problems' that NBN might be thought to be designed to address, and argues that the policy as is presently being pursued is not a good solution to any of these.

The paper concludes with some suggestions for a more sensible approach.

 

Introduction

A striking feature of the Rudd-Gillard Government’s NBN policy – to spend over $50 billion of taxpayers’ money on a 100 Mbps fibre to the premises network serving over 10 million premises – is the lack of clarity about the problem this policy is designed to solve. In this article I argue that the NBN was announced for political reasons rather than policy reasons.

After briefly explaining that point, I want to turn to the policy issues – by identifying and assessing several 'candidate problems' that NBN might be thought to be designed to address. I will argue that the policy as is presently being pursued is not a good solution to any of these.

I will conclude with some suggestions for a more sensible approach.

There is significant lack of clarity about the problem the NBN is designed to solve. This is because the genesis of NBN was to solve a political rather than a policy problem.

'Candidate problems' that NBN might be thought to be designed to solve include : low broadband penetration; high broadband pricing; facilitating useful applications; fixing broadband blackspots; and poor competition in fixed line services.

But examining each of these 'candidate problems' reveals that:

1. International experience suggests that extensive fibre networks do not correlate with high broadband penetration.

2. Price is a key driver of taken up. Under NBN, prices will be higher than under other industry models and that will dampen take up.

3. There is little evidence to suggest what 100Mbs might need to be used for, or to support building fibre to ten million premises rather than a much smaller number of premises, namely key institutions such as schools and hospitals.

4. International experience and studies show that where consumers have a choice of speeds, most do not take up the highest speeds, or are prepared to pay only very low premium for those highest speeds.

5. Broadband black spots are a problem, but NBN does not offer a timely or a cost effective way to fill these gaps.

6. NBN is not necessary for the structural separation of Telstra.

7. NBN will be a wholesale monopoly – ensuring monopoly pricing is at the heart of the new broadband market.

These conclusions help identify objectives to which broadband policy should be directed. These include :

1. Achieving ubiquitous broadband access

2. Filling broadband black spots quickly

3. Targeting key institutions such as schools and hospitals for higher speeds

4. Flexibility in technology choices

5. Economic sustainability.

NBN Announced for Political Reasons

The precursor to Australia’s current NBN policy was the proposal put to the Australian public by the then Labor Opposition in the 2007 election: to build a 12 Mbps fibre to the node network to 98% of the population, with public expenditure of $4.7 billion.

The implementation of this policy, largely during 2008, was a failure. The Rudd Government conducted a tender process to find a private sector partner with which it would jointly build the network. Telstra lodged a non-compliant bid and the Government cancelled the tender after tenderers had submitted their bids. A National Audit Office inquiry found that the process was never likely to succeed. The tender wasted $30 million including $17 million of taxpayers’ funds.

By early 2009, Broadband Minister Conroy and then Prime Minister Rudd found themselves with an acute political problem.  They had taken a policy to the 2007 election which they now could not deliver.  Their solution was a political strategy of ‘double or nothing.’  They would abandon their initial plan of spending $4.7 billion and building a 12 Mbps fibre to the node network; now it was to be ‘shock and awe’ with a 100 Mbps $43 billion fibre to the premises network to serve 90 per cent of the population (along with wireless and satellite to serve the rest).  The announcement was made in April 2009 and it achieved its political objective – the failure of Labor’s fibre to the node policy was largely ignored by the media, and all talk was instead of the new 100 Mbps network. (Prime Minister et al 2009)

Other commentators, from political perspectives different to my own, offer a similar analysis. For example, a former union officer, and a member of Kim Beazley’s advisory panel that oversaw the introduction of competition in 1991, Kevin Morgan, sees the policy in precisely these terms. Computerworld magazine summarised Morgan’s position as follows in August 2010:

'The $43 billion national broadband network (NBN) is a ‘purely political cover up’ of the failure of the original FTTN proposal and a repeat of decades-old, failed nation-building technology Labor policy, according to independent telco consultant and long-time trade unionist, Kevin Morgan.' (Computerworld 2010)

The problems NBN might be thought to be addressing

The Rudd-Gillard Government has never been particularly clear about which policy problem the NBN is designed to address. In this section, I consider several ‘candidate problems’ – and argue that the NBN in its present form is not an effective solution to any of them

Low Broadband Penetration

It appears that Broadband Minister Conroy sees the NBN as a solution to low broadband penetration.

Consider, for example, his December 2010 media release headed ‘Latest OECD Statistics Reinforce the Need for the NBN.’

Noting that Australia had now dropped to eighteenth place in the latest figures just released by the OECD (reporting the position as at June 2010), Conroy said this showed Australia was 'lagging behind world leaders in access to broadband' and was 'further evidence that Australia has lacked vital investment in fixed-line broadband infrastructure' (Conroy 2010)'

The OECD statistics measure the number of fixed line broadband services per 100 people in the country. 

In other words, Senator Conroy was arguing that his strategy of investing in a fibre network would drive up the number of Australians taking a broadband service – and achieve an improvement on the June 2010 numbers he was quoting in his release. 

There is now more recent OECD data, current as at June 2011(OECD Broadband Statistics 2011). These numbers have Australia slipping down the ranking with Australia twenty first (out of thirty four) with 24 fixed broadband subscriptions per 100 Australians.

Is Conroy right to assume that increased fibre provision is correlated with increased broadband takeup? The data suggest otherwise.

The OECD data show that the highest ranked country (the Netherlands) had 38.5 broadband subscribers per 100.

Interestingly, of the 38.5 broadband subscribers per 100 in the Netherlands, only 1.3 were fibre. The majority were DSL (21.2 of the 38.5), and the balance were cable.

What did these numbers show for another European country, Denmark, recently singled out in a submission by Senator Conroy’s Department to a Parliamentary inquiry on the NBN:

'Denmark is considered by the OECD to be among the best in Europe in terms of the sophistication in e-government services with 84 per cent of the 20 basic public services for citizens online. This is supported by Denmark’s performance as a leader in terms of broadband penetration and frequent Internet users' (DBCDE 2011).

Denmark ranks third on the OECD’s penetration list.  But again, a closer examination shows that only a small share of its services are fibre:  5 per 100, compared to 21.9 DSL and the balance cable.

So ‘low fibre’ countries can have very high broadband penetration. In fact, only 10 nations in the OECD survey have fibre for greater than 2 per cent of connections.

Interestingly, the highest fibre country, Japan, does not do particularly well on penetration. Some 60% of Japan’s broadband services are fibre (of Japan’s 27 broadband services per 100, 16.4 are delivered over fibre) – the highest percentage of all the countries in the OECD survey.

Yet Japan ranks only sixteenth in the OECD broadband rankings.

In short, building an expensive high fibre network is neither necessary nor sufficient to achieve a high penetration of broadband services.

If our policy objective is to have as many Australians as possible using broadband services, the data suggest that spending big on a national fibre network offers no certainty of meeting that objective.

High Broadband Pricing

The next ‘candidate problem’ which NBN might be designed to solve is high broadband pricing.  It is politically uncontentious that Australia today has high broadband pricing by world standards.  As Prime Minster Gillard has stated, 'Australians have one of the most expensive broadband systems in the world' (Hansard, 25 November 2010).

The assumption inherent in this statement – that high broadband prices keep broadband take up lower than it otherwise would be – is one I agree with, and one supported by the data.

In my 2009 book Wired Brown Land (Fletcher 2009) I traced the history of fixed broadband pricing – and takeup - in Australia.  I found that Telstra, as the monopoly supplier of consumer DSL, initially set very high prices - $89 per month for a 512 Kbps service for a customer who also took a voice service - when it launched the service in 2000. In turn, take up was initially low – by December 2003, the number of services had reached only 241,000. However, in early 2004 Telstra sharply dropped prices (at the same time as Optus entered the market.)

In 2004 growth rocketed ahead and by the end of that year there were over one million ADSL services; by the end of 2005 there were over two million.

In the two years 2004 and 2005 over 1.6 million Australian households decided to take up a DSL service. They clearly saw value in having a broadband connection – but previously it had been too expensive. So 1.6 million households had to wait to receive a service which Telstra could readily have delivered to them five years earlier, had it chosen to do so.

Unfortunately, the way the NBN has been set up gives NBN Co the same kind of monopoly pricing power which Telstra has previously enjoyed. The result is likely to be a repeat of the problem we experienced with the introduction of DSL: NBN Co will have the means and the motive to keep broadband prices higher than they would be under another market structure, and in turn take up will be lower.

Already NBN Co’s behaviour is consistent with the expectation of monopoly behaviour and high prices. This is evident from considering the pricing proposed by NBN Co for approval by the ACCC. The first version (in a discussion paper issued by NBN Co) would have given NBN Co the right to increase prices, for all but the entry level service, by the inflation rate plus five per cent each year (NBN Co 2011a). This met with a storm of protest. 

The second version, contained in the Special Access Undertaking lodged by NBN Co with ACCC for its approval, reduces this to a permitted annual price increase equal to half the inflation rate. (NBN Co 2011b)  This at first blush seems a little better – but it would still involve a sharp break from the steady price falls which have characterised broadband and telecommunications pricing over the last decade.

For example, in relation to PSTN services the ACCC price index declined 38 per cent between 1998 and 2010, inclusive. For broadband, the ACCC’s index showed a decline of 15.7 per cent for the same period (Ovum 2011).  

An additional reason to fear that NBN will keep prices high is that the NBN policy involves a huge capital expenditure – all of which will have to be recouped to meet the Rudd-Gillard Government’s requirement of a 7 per cent rate of return.  The hugely expensive network design necessitates high prices to end users. 

RBS telecommunications analyst, Ian Martin, has recently published a paper in this journal showing the NBN will have to increase revenue per customer by 5.7 per cent a year to meet its corporate plan (Martin 2011).

The policy of rising prices which NBN Co apparently intends to pursue would be at odds with world trends in broadband pricing, and could only be achieved due to NBN Co’s monopoly status.

If history is any guide, price increases will mean only one thing: broadband take up will stay disappointingly low.

Low Broadband Speeds Are Holding Back Useful Applications

A third possibility is that the NBN is designed to solve a problem that useful high-speed applications are being denied to millions of Australians.

If this is the problem the NBN is solving, however, it is surprising that there is so little evidence of the widespread existence of these applications.

I participated in the recent inquiry into the NBN by the House of Representatives Infrastructure and Communications Committee. I was struck by how few examples we saw of applications which actually require the speeds that NBN will deliver.

Let me quote from the Coalition Members’ Dissenting Report:

The Committee was provided with many fascinating and encouraging examples of applications which could be delivered if there was widespread availability of high speed broadband services.

There was however a conspicuous lack of evidence of a need for applications needing the speeds that a FTTH network can deliver (100Mbps) as opposed to speeds of say 10-20 Mbps which can be delivered over existing DSL and HFC networks to many Australians.

Some witnesses effectively argued that the applications which would require 100 Mbps are not yet known. This view was put by the CSIRO for example: The future transformative impact of broadband communications, including Internet access, is, to some extent, unknown.

Coalition Members welcome the intellectual honesty of this submission.

We are not persuaded however that it makes sense to spend so many billions of taxpayers’ money on a venture the impact of which is unknown.

The failure to demonstrate specific applications requiring 100 Mbps was notable in many different fields of activity. (House of Representatives Committee on Infrastructure and Communication 2011).

Those fields of activity included telemedicine, education, business and government services.

A related point is whether applications requiring 100Mbps are needed in 10 million or more homes – as opposed to the much more limited numbers of institutions such as universities, schools, hospitals and libraries.

For example, there are some 10,000 schools in Australia; some 6300 have a fibre connection already (Australian Information and Communications Technology in Education Committee 2011).  There are some 1,300 hospitals.  A strategy of connecting key institutions directly to fibre would be several orders of magnitude cheaper than connecting every home (not least because a high proportion of such institutions already have a fibre connection) – but would deliver significant benefits.

The other issue concerning speed is the extent to which consumers are willing to pay for very fast speeds.

A report of the Australian Communications and Media Authority notes that 35 per cent of Australian Internet users had speeds in the 1.5Mbps to 8Mbps range and another 35 per cent in the 8Mbps to 24Mbps range.

Subscriptions to access speeds greater than 24Mbps represent a relatively small proportion of the overall market at 10 percent of household Internet subscribers and 3 per cent of business/government subscribers, at December 2010 (ACMA 2011)

This tends to suggest a limited appetite to purchase high speed broadband, as more than ten per cent of households (those connected to the Telstra and Optus HFC networks) would be able to purchase those speeds if they wanted to. That being said, it is of course true that in Australia only a minority of customers are served by networks able to deliver speeds in excess of 24Mbps. The real test is that, where very fast speeds are available, what is the propensity of consumers to take up and pay for those higher speeds?

When we look to those overseas markets in which such speeds are more widely available than in Australia, the experience of customer take up of high speed services is not encouraging.

A very detailed study prepared for the US Federal Communications Commission has found that households see little value in faster speeds:

Our empirical results show that reliability and speed are important characteristics of Internet service. The representative household is willing to pay $20 per month for more reliable service, $45 for an improvement in speed from slow to fast, and $48 for an improvement in speed from slow to very fast. The latter finding indicates that very fast Internet service is not worth much more to households than fast service (emphasis added) Rosston et al 2011).

This finding is consistent with evidence to the Parliamentary inquiry I mentioned earlier, that there are few applications for which consumer will want – or pay for – 100Mbps.

Ofcom, the UK regulator, has recently published a report which indicates that while 57 percent of UK residences have access to super-fast broadband (25 Mbps or over), only two per cent of the population subscribed to and received super-fast broadband. In other words, less than four percent of those who had access to super-fast broadband actually subscribed to it (Ofcom 2011). The evidence suggests therefore that only a small proportion of consumers will pay for higher speeds.

This raises serious questions about why we are incurring the massive cost of engineering the network to provide all homes and businesses with such speeds. It also rings very loud alarm bells for the economics of the NBN, based as it is on optimistic assumptions concerning the take up of its highest speed offerings.

Filling in Australia’s Pervasive Broadband Blackspots

A fourth 'candidate problem' is the fact that Australia has pervasive broadband blackspots.

The lack of any broadband, or the availability of only inadequate service, is a key issue for a significant minority of Australians. Many in rural and regional Australia do not have an adequate service, as they are too far from an exchange. There are problems too in metropolitan Australia, particularly many outer suburban areas, as broadband over DSL is not available due to the use of RIMS and pair gains.

The Implementation Study estimated that some eight percent of premises cannot receive a DSL service (McKinsey & Company and KPMG 2010).

Unfortunately, under the Rudd-Gillard Government’s NBN policy, many Australians will be waiting a decade or more. While NBN may solve lack of broadband eventually, many of those in need will have to wait years for their home to be connected to the NBN.

The Howard Government, by contrast, recognised the need for rapid action prioritised towards filling in blackspots. In 2007 it awarded a contract to the OPEL consortium to build a national WiMAX network, initially offering 6 Mbps peak speed but rising to 12 Mbps peak speed. Had that contract not been cancelled by the incoming Rudd Government, rural and regional Australians would already be receiving improved broadband services.

Poor competition in fixed line services

One of the arguments which the Government uses to justify its NBN policy is that it achieves the ‘structural separation’ of Telstra. This is argued to be a good idea because it will mean the Australian telecommunications industry is no longer dominated by a vertically integrated incumbent – in turn increasing competition in telecommunications. In other words, Telstra’s vertical integration is another ‘candidate problem’ for which the NBN could be proposed as a solution.

It is widely agreed that in the fixed line telecommunications industry, competition is not as strong as it should be – and consumers lose out as a result.

A key reason for this is Telstra’s dominance.  As a vertically integrated incumbent, it enjoys great advantages over its smaller, retail only fixed-line competitors.  Because it so much bigger than its competitors (in fixed line, Telstra serves around ten times as many customers as its next biggest competitor, Optus), Telstra enjoys unmatched economies of scale.  Because it has so many products sharing many of the same network elements – fixed line voice, broadband, mobile, pay television – Telstra also enjoys economies of scope.

Structural separation solves many of these problems. It ends Telstra’s vertical integration and reduces its scale and scope advantages over competitors. Accordingly, structural separation is a desirable policy objective – and one supported by both sides of politics.

Unfortunately, the Rudd-Gillard government’s method of achieving structural separation is stupendously expensive and unnecessarily complicated. Separation could be achieved simply by splitting Telstra into two companies, one to own and operate its existing access network (which could then be upgraded to deliver a higher standard of broadband services) and one to own and operate its retail fixed line business together with its other businesses such as mobile. This is essentially what has occurred with Telecom New Zealand, with effect from late 2011.

Even more troublingly, the Rudd-Gillard Government’s version of structural separation comes with some very nasty anti-competitive side effects which largely offset the in principle benefits that structural separation would be expected to deliver. Let me mention three of them. First, under NBN the possibility of competing networks being built is effectively prohibited. The Rudd-Gillard Government has passed laws requiring that any rival networks must be built to the same technical standards as NBN and be subject to the same access arrangements.

Secondly, Telstra and Optus are being paid to cease retail operations over their HFC cables, which pass nearly one third of Australian households, and have the potential to deliver broadband services at the same 100 Mbps speed which is promised over the NBN. This is in substance an arrangement where an industry entrant (NBN Co) pays existing participants to cease their operations. It would be illegal under section 45 of the Trade Practices Act, as a 'contract, arrangement or understanding [which] has the purpose, or would likely to have the effect of substantially lessening competition' if the Gillard Government had not passed a law to expressly authorise it.

Thirdly, the arrangements also seek to suppress competition between the NBN fixed line service and wireless networks. The contract between NBN Co and Telstra contains an undertaking by Telstra that it will not market its wireless service as a competitive offering to customers as they migrate from the copper network to NBN.

What are the real problems that should be addressed?

In summary, we can draw several conclusions from our examination of whether the NBN is a solution to existing broadband policy problems:

1. International experience suggests that extensive fibre networks do not correlate with high broadband penetration.

2. Price is a key driver of take up. Under NBN, prices will be higher than under other industry models and that will dampen take up.

3. There is little evidence to suggest what 100Mbs might need to be used for, or to support building fibre to ten million premises rather than a smaller subset to key institutions.

4. International experience and studies show that where consumers have a choice of speeds, the majority do not take up the highest speeds, or are prepared to pay only very low premium for those highest speeds.

5. Broadband black spots are a problem, but NBN does not offer a timely or a cost effective way to solve this problem.

6. NBN is not necessary for the structural separation of Telstra.

7. NBN will be a wholesale monopoly – ensuring monopoly pricing is at the heart of the new broadband market.

We can use these conclusions to start to identify broadband policy directions which should be pursued. Let me offer some brief concluding thoughts on this point.

First, there is a critical distinction between ubiquitous broadband and super-fast broadband. The concept of ubiquitous service at reasonable – rather than lightning – speeds underlay the comments of many of the witnesses who appeared before the Parliamentary Committee Inquiry into the Role and Potential of the National Broadband Network. Witness after witness made the point that having broadband available to all households and business was more important than the particular speed being offered.

This suggests that the priority in the short to medium term should be to focus on those areas of the country where broadband is clearly inadequate – that is, to fill in the black spots as quickly as possible. This was the rationale for the Coalition's including within our Broadband policy at the 2010 election an allocation of $750 million to remediate DSL. We expected that this would allow a rapid filling in of black spots in many parts of the country.

The NBN approach on the other hand is essentially 'super-fast broadband arriving super slowly'.

When considering the building of a network to all Australian homes and residences at vast expense, the speed that is necessary should be carefully considered. Higher speed means higher cost. Is the vast expense of the Rudd-Gillard Government’s NBN (currently sitting at around $60 billion, and only likely to increase if past Government-run infrastructure projects are any indication) justified? Are the incremental benefits of 100 Mbps – over other speeds that could be delivered more cost-effectively – so compelling as to justify the incremental cost? Without a cost-benefit study, we cannot know – but the available evidence I have cited earlier gives us grounds to be sceptical.

It does make sense, however, to target fibre to key institutions with a need for higher speeds and capacity – such as schools, universities and hospitals. Much evidence before the Parliamentary Committee Inquiry into the Role and Potential of the National Broadband Network supported this conclusion.

A more flexible approach to a network build is also required. The Rudd-Gillard Government has set an artificial objective that a particular percentage of premises need to be served by fibre as opposed to other broadband technologies.

Finally, economic sustainability is a vital consideration. The danger with NBN is that it will be economically unstainable and it will collapse. This has the potential to leave those Australians with inadequate broadband stranded for even longer.

The Coalition is well advanced on planning our alternative approach to broadband policy, as is clear from detailed statements made by Opposition Broadband Spokesman Malcolm Turnbull. Ubiquitous access at reasonable speeds, filling broadband black spots more quickly, fibre to key institutions and economic sustainability are key principles which will inform the Coalition’s approach to broadband policy should we come to government at the next election.

 

References

 Australian Communications and Media Authority (ACMA). 2011. The Internet service market and Australians in the online environment, Australian Government, Canberra, July 2011, p. 20. Available from: http://www.acma.gov.au/webwr/_assets/main/lib310665/the_internet_service_market_in_australia.pdf

 Australian Information and Communications Technology in Education Committee. 2011. Submission number 124 to House of Representatives Committee on Infrastructure and Communication Inquiry into the Role and Potential of the NBN, p.2

 Computerworld. 2010. NBN Pie in the Sky: Morgan, 20 August 2010. Available from: http://www.computerworld.com.au/article/357748/nbn_pie_sky_morgan/

 Conroy, Senator the Hon Stephen. 2010. ‘Latest OECD Statistics Reinforce the Need for the NBN’ Media Release from the Minister for Broadband, Communications and the Digital Economy, 7 December 2010. Available from: http://www.minister.dbcde.gov.au/media/media_releases/2010/111

 Department of Broadband, Communications and the Digital Economy. 2011. Submission No.215 to House of Representatives Committee on Infrastructure and Communication Inquiry into the Role and Potential of the NBN, p.23

 Fletcher, Paul. 2009. Wired Brown Land? Telstra’s Battle from Broadband, New South Wales Press

 Hansard. 2010. 25 November 2010, p.3816. Available at: http://parlinfo.aph.gov.au/parlInfo/genpdf/chamber/hansardr/2010-11-25/0114/hansard_frag.pdf;fileType=application%2Fpdf

 House of Representatives Committee on Infrastructure and Communication. 2011. Inquiry into the Role and Potential of the NBN, Report p.317

 Martin, Ian. 2011. ‘A significant gap in the NBN corporate plan’. Telecommunications Journal of Australia 61 (3): 51.1-51.11. Available from: http://tja.org.au.

 McKinsey & Company and KPMG. 2010. Implementation Study for the National Broadband Network, 5 March 2010, prepared for the Department of Broadband, Communications and the Digital Economy, p.12

 NBN Co. 2011a. ‘NBN Co Discussion Paper: Introducing NBN Co’s Special Access Undertaking’, July 2011, p.33

 NBN Co. 2011b. ‘NBN Special Access Undertaking in respect of the NBN Access Service’, 5 December 2011, p.45

 OECD. 2011. Broadband Statistics. ‘OECD Fixed (wired) broadband subscriptions per 100 inhabitants, by technology June 2011’. Available from: http://www.oecd.org/document/54/0,3746,en_2649_34225_38690102_1_1_1_1,00.html

 Ofcom. 2011. ‘Communications Market Report: UK’ 4 August 2011, pp.246-249

 Ovum. 2011. ‘Unfinished Business – 20 years of competition in Australia’s telecommunications sector' November 2011, p.29 and p.35.

 Prime Minister et al. 2009. ‘New National Broadband Network’ Joint Media Release Prime Minister, Treasurer, Minister for Finance, Minister for Broadband, 7 April 2009. Available from: http://www.minister.dbcde.gov.au/media/media_releases/2009/022,

 Rosston G; Savage S J; Waldman D M. 2011. Household Demand for Broadband Internet Service, Final Report to the Broadband.govTask Force, Federal Communications Commission, 3 February 2011, p iii. Available from: http://siepr.stanford.edu/system/files/shared/Household_demand_for_broadband.pdf