Monash Business Review
Fibre to the home or the node? What price the service? Whatever the answers, the government must buy back Telstra, keep the infrastructure and sell off the wholesale and retail divisions, says Nicholas Beaumont.
Demand on the internet is soaring. Cisco Systems believe that internet traffic will nearly double every two years until 2011. Much of this growth is attributable to increased traffic in video data needing huge speed and capacity.
To remain internationally competitive and raise Australians’ quality of life, Australia needs to replace its existing copper telephony network with a high speed Optical Fibre Network (OFN). The enormous capacity of Optical Fibre (OF) will facilitate myriad private and business uses and provide significant social and economic benefits to Australian individuals and companies.
OFN’s limitless capacity makes possible applications such as videoconferencing, Voice Over Internet Protocol (VOIP) and access to any of the world’s television stations or online game environments such as Second Life. It will boost business by speeding online transactions and providing new business models and services. It can help besotted grandparents keep in touch with family via email and videolinks. New applications such as YouTube and Facebook will change the way people communicate and socialise.
The Australian government should own, build and control a Fibre to the Home (FTTH) network capable of delivering at least 100 megabits per second (Mbps) to homes and businesses. To do this, it will have to take over Telstra, Australia’s dominant, near-monopoly, telephony provider that owns Australia’s copper, coaxial and optical telephony network. The government should retain the network and sell off Telstra’s retail and wholesale businesses. The capital required is large, but comparable with Australia’s uncontroversial spending on roads.
Both major political parties have promised to spend several billion dollars on the alternative OFN, a Fibre to the Node (FTTN) system. This proposal is fatally flawed if only because Telstra will retain monopoly ownership of the copper connecting nodes and premises and use this power to hamstring its competitors.
An OFN has a distinctive cost structure: very high capital costs, low maintenance and administrative costs and negligible usage costs. Negligible usage costs imply that marginal cost of use should be zero, in particular, that phone calls within Australia should be free. There should be a free and open market for entrepreneurs providing equipment or services based on the network. This policy would reduce business costs, maximise social utility, impel the Australian economy and better connect Australia to the rest of the world.
Several questions arise:
• Should Australia choose Fibre to the Node (FTTN) or Fibre to the Home (FTTH)?
• What will the network be used for and will the demand justify the outlay?
• Should the network be owned and/or operated by a private company (such as Telstra), a consortium or the Federal government?
• What is the most appropriate way of charging users?
The aircraft’s ‘black box’ was invented in Australia but commercialised elsewhere. Australia is not good at managing or commercialising technological innovations, perhaps because until about 1983 the Australian economy was protected from international competition and few politicians were qualified or interested in science or technology. If this continues, Australia will become less competitive.
Other examples of Australian governments’ technological policy mismanagement include:
Digital radio: now available in UK and US but not Australia except experimentally. Valuable digital spectrum has been given gratis to incumbent radio networks: a sure and certain way of stultifying innovation and new services.
Digital television: policy may well have been written by the incumbents. Governments, driven by fear of the proprietors’ political influence, have preserved the incumbents’ oligopoly, killing consumers’ access to new technology and cheaper prices.
Water resource management: Governments appear to be much more concerned with the rights of people who possess water entitlements (often obtained for free). There is no sense of overall optimisation or recognition that what happens in headwaters affects downstream residents.
Old industries: Governments support old industries with political clout (especially automotive and agricultural). Australia should be manufacturing lasers and computer chips not stamping out car parts designed by overseas owed companies.
A FTTH network has essentially infinite capacity and zero marginal cost of use. Social utility will be maximised if the marginal charge for use is zero. Free national phone calls are an obvious benefit for businesses and citizens. Building such a network is technically possible.
FTTN is an unsatisfactory halfway house. The network’s speed and capacity is severely restricted by the old, copper technology used to connect premises to nodes. These limitations would make some high bandwidth applications (eg viewing television programs) difficult or impossible. Any FTTN scheme would grant Telstra a monopoly on the ‘last mile’. Telstra would be constrained only by the Australian Competition and Consumer Commission (ACCC).
Unconstrained, Telstra could charge excessive fees and do its utmost to hamper organisations providing services such as VOIP that threaten its revenues. Telstra habitually draws out and appeals legal challenges to its pricing policies and market power. It has been accused of changing its technology primarily to make competitors’ complementary technology unworkable. The consequent lack of competition will stifle innovation manifest in new products and services.
The network should be used to give Australians free national phone calls and fast access to the internet. Cheap communication enhances democracy, social welfare, community involvement, and public discussion. Governments should encourage vigorous competition for equipment (handsets, mobiles, modems, PDAs, home networks etc) and services (internet access, e-commerce, movies on-line, video phones, always on telephony, e-education, e- health, mobile commerce, local commerce, VOIP, internet services etc) that use the infrastructure.
The Australian constitution allows the Federal Government to make laws pertaining to “Postal, telegraphic, telephonic, and other like services” (section 51 v) and “The acquisition of property on just terms from any State or person for any purpose in respect of which the Parliament has power to make laws” (51 xxxi). The latter section ostensibly allows the government to acquire Telstra’s infrastructure. Ascertaining the infrastructure’s value and “just terms” would be difficult and certainly entail protracted court proceedings. These difficulties would be exacerbated by Australian governments rarely having a majority in the senate (the upper house). A conventional takeover is likely to be cleaner and perhaps less politically hazardous then a compulsory acquisition.
The sums involved are superficially staggering. Telstra has a share market value of about $A60 billion; the total initial acquisition outlay would comprise that $60bn, a takeover premium of about $5-10bn and the $15-20bn cost of building a FTTH network, a total of $80-90bn (about 7 per cent of Australia’s Gross National Product). The government might be consoled by arbitrage profits: Telstra’s share price has declined appreciably in real terms since some tranches were sold. The estimated $15-20bn needed to establish the network looks modest compared with Australia’s expenditure on roads, highways, and subdivisions which totalled $38.31bn from 2003 to 2005 inclusive.
The government would sell the retail and wholesale elements of Telstra retaining only the infrastructure. Financial analysts such as Morgan Stanley reportedly to believe that breaking up Telstra would increase its total stockmarket value by 20 per cent, but it is unclear how much each component would be worth.
There would be appreciable savings if all telephony operations (except those in thinly populated areas) were transferred from copper to OF. The copper comprising the old network and Telstra’s 5,000 local exchange buildings could be sold. The substantial maintenance costs of the copper network ($600m per annum) would be eliminated and administrative costs would be reduced.
The capital cost of replacing the Australia copper network with fibre (assuming access to Telstra’s ducts) is not precisely known but according to finance journalist Alan Kohler may be estimated to be “…between $1,200 and $1,500 per home connected, depending on the density.” For commercial operators the cost is critically dependent on the ‘take rate’ (the proportion of premises that subscribe to the service; we have assumed a 100 per cent take rate because the proposed cost-reducing change will be compulsory, cheap and seamless). Other estimates with a 100 per cent take rate are $US1,465 and $US1,200-$2,400.
Beside the CAPEX, administering and maintaining the network will cost an estimated $50-$100 per annum per premise. This is appreciably less than the cost of maintaining Australia’s copper network, estimated as $600m pa. Some administrative functions that, for copper network, require a technician’s visit (a ‘truck roll’) can be performed remotely in a FTTH network.
The government should build the network and maximise public utility by charging a modest annual administrative fee of (say) $50-$100 – but nothing for actual use. This accords with the zero marginal cost of using the network.
An entrepreneur building a FTTH would have to charge customers serious money to recover a $15-20bn investment locking out poorer households, reducing take rates, and thereby increasing the capital cost per customer. Interminable arguments at the ACCC and the courts over competitors’ access and what the network owner should charge competitors for network use would continue. It would be especially iniquitous for Telstra to own and control the network while competing with other telephony companies using it; vital infrastructure should be available on equal terms to all potential users. A huge FTTH project would be risky for a private investor as it might conceivably be superseded by a new (wireless?) technology.
Many authors point out that FTTH investments are likely to benefit from as yet unknown ‘killer applications’ and/or that reliance on wire-based technologies rules out many new applications and opportunities. All authors agree that fibre’s operating expenses, comprising maintenance and administrative costs, are substantially less than those of wire or cable. Installing FTTH when housing estates are built is much cheaper than retrofitting and some new Australian housing estates have been built with FTTH. The successful installation of OF in a greenfields housing estate in Victoria in 2006 reportedly increased the value of homes.
The number of occupied private dwellings in Australia in 2006 was 7,596,000. An unknown number of these are apartments in one building and would not require separate connection to an OFN. In June 2007 there were 2,011,770 actively trading businesses in Australia. Many associated businesses would operate out of one office and one building could contain the offices of many businesses. Assuming that there were 5,000,000 Australian premises and that the cost per premise was $US1,500, the total cost of the national network would be a relatively modest $8.5bn assuming that Telstra’s ducts were available. This estimate is much lower than others in the public domain.
The exact capital cost of building an Australian FTTH network depends on policy choices and presently unknown costs. A network covering all Australian premises no matter how remote would be prohibitively expensive: connecting between perhaps 95 per cent and 98 per cent is more realistic. The total cost depends heavily on this choice as capital costs per premise increase with declining density. Another imponderable is whether the installer will have access to Telstra’s ducts and (if so) Telstra’s charges for access and use. Given that VOIP will immediately cannibalise Telstra’s cash cow telephony business, the prospects of cooperation seem extremely remote. The initial cost would be reduced if local councils improbably allowed fibre to be strung on telephone poles instead of being buried. The capital expenditure required is declining with technology improvements, components’ declining costs and deskilling of installation procedures. Avoiding digging trenches by threading fibre through sewers and/or drilling though suitable soils is sometimes practical.
Until 1997, Australia’s telephony was run by a government department that emphasised social goals such as a universal service obligation rather than business goals. Between 1997 and 2002 a conservative government privatised Telstra in stages through three public offerings of shares. The privatisation was not ‘clean’; Telstra remains encumbered by obligations to provide basic services cheaply to poor households and rural communities. In retrospect, the Telstra privatisation was a mistake. The share price has retreated by up to 50 per cent from the first tranche’s sale price.
Telephony infrastructure is a natural monopoly; there is a good case for its public ownership.
Had Telstra remained government owned, government decisions on an OFN would have been better informed and less constrained. Australia has been pilloried because, compared with other OECD countries, its telephony and broadband services have higher than average costs (a local call is charged at 20 cents but the marginal cost is assumed to be no more than 1 cent) but lower than average capacities.
Lingo of the FTTH
OF: Optical Fibre is a technology that allows digital data to be transferred at very high speeds, facilitating new applications such as YouTube, Facebook, videoconferencing and access to global TV stations.
OFN: An Optical Fibre Network.
FTTH: A Fibre to the Home network takes the optical fibre technology directly into the home or the business.
FTTN: A Fibre to the Node network terminates the optical fibre technology at nodes which then service 200 to 300 premises through the coaxial cable or copper wire (twisted pair) presently used to carry telephone conversations and data. Although schemes such as Asymmetric Digital Subscriber Line (ADSL) and ADSL2+ have been used to multiply copper’s bandwidth, the number of subscribers and the distance from the node impact on speed.
Mbps: Megabits per second
CAPEX: Capital expenditure
OPEX: Operating expenditure
Telstra: Australia’s dominant, near-monopoly, telephony provider
Broadband: Signifies any technology allowing rapid transfer of data
Bandwidth: Normally expressed in megabits per second (Mbps), it is synonymous with capacity
MBR subscribers: to view full academic paper email email@example.com
Public access: www.mbr.monash.edu/full-papers.html (six month embargo applies)
Dr Nicholas Beaumont is Senior Lecturer in the Department of Management, Faculty of Business and Economics, Monash University.